). The first term on the right-hand side is of the through-space type. It would be present even if there were no hyperconjugation, that is, even if the nonbonding orbitals were strictly localized on C(1) and C(2), respectively (c, = 1, cz = c, = 0). The second and third terms would vanish in this limit and are of the through-bond type. The action of the operator 9: on a n-symmetry orbital on its right is to rotate it by 90" around the z axis in the clockwise sense when viewed against the direction of z (cf. Equation 4.13); for instance,
Thus, the through-space term makes a positive contribution to the z compoand the two through-bond terms make nent of the atomic vector <Xe1911~b>. negative contributions. I t is impossible to evaluate the sign of the resultant without at least a rough numerical evaluation of the opposing contributions: the through-space term contains the overlap integral S,, as a relatively small multiplier while the through-bond terms contain the small coefficient c, or c, as a multiplier, but there are more of them. In a minimum basis set approximation, a computation shows that the through-bond terms dominate. Inclusion of the two-electron part of the spin-orbit operator in the calculation reduces the magnitude of the result by a factor of about two. hut does not change its sign. Note that the opposed signs of the through-space and through-bond contributions are a result of the nodal properties of the orbitals x,, and xb. As is seen in Figure 4.25a, they have a node separating the main part of the orbital from the minor conjugatively delocalized part. However. the presence of the node
Example 4.11: The magnitude of the total spin-orbit coupling strength can change dramatically as a function of biradical conformation (Carlacci et al., 1987). The throughspace part can be roughly approximated by where w is the acute angle between the axes of the p orbitals containing the unpaired electrons and S is their overlap integral. The overlap S is related to the resonance integral between these two orbitals and thus to the coefficient of the in-phase combination of the two hole-pair functions in the So wave function, Co.+.The factor sin w originates from the operator 9~in the matrix element which rotates the orbital xhaccording to Equation (4.13) before its overlap with x,, is taken. No such simple generally valid approximation is currently available for the through-bond part of the spin-orbit coupling strength, which results from the delocalization of even the most localized form of the nonbonding orbitals into the a skeleton, onto nearby carbon atoms p located between the radical centers. This produces nonzero coefficients on the p orbitals on these atoms and permits them to contribute to the sum in Equation (4.12). A similar mechanism operates when atoms carrying lone pairs, such as oxygen. are located between the radical centers. The low-lying electronic states of biradicals of this type are more numerous and the spin-orbit coupling is less likely to be properly described by the simple model that led to this equation. The availability of additional states involving promotion from the lone pairs appears to make spinorbit coupling particularly effective. Note that the placement of lone-pair carrying or heavy atoms into positions that do not lie between the two radical centers cannot be expected to have much effect, since then only one of the most localized nonbonding orbitals has significant coefficients on their orbitals.
<X,,lefillxb>.
As discussed in Example 4.9, the probability of the intersystem crossing to the singlet will be determined by the weight of the singlet spin function in the "impure" triplet at a time when the molecular motion brings its geometry to a region in which the singlet-triplet splitting increases significantly, so that a decision between an essentially pure triplet and an essentially pure singlet state must be made. If it occurs, further nuclear motions will be dictated by the So surface, and if it does not, the biradical will continue its conformational motion in the impure triplet state. The geometries at which the singlet-triplet splitting becomes large are those at which the covalent interaction between the two localized singly occupied orbitals, described in the simple model by the quantity y, is large. At many of these, the spin-orbit coupling is also large. At these geometries, the Sosurface will typically slope steeply toward a product minimum, with no barrier in the way, and the probability that a newly formed singlet molecule might escape the likely fate of falling into this abyss is undoubtedly minimal. The somewhat startling conclusion, that intersystem crossing in triplet biradicals can and normally does produce closed-shell singlet products and not floppy singlet biradicals, was initially suggested by Closs. It has been gradually gaining recognition (De Kanter and Kaptein, 1982; Wagner, 1989; Wagner et al., 1991). Indeed, intersystem crossing caused by random spin relaxation at the two independently acting radical centers, which can be accelerated by the addition of paramagnetic impurities, and which would be expected to generate floppy singlet biradicals at a variety of geometries, yields different product ratios (Scaiano, 1982). Geometries expected to be favorable for intersystem crossing are those at which both the covalent interaction y as well as the spin-orbit coupling strength become large. This requires that the two most localized singly occupied orbitals overlap enough to establish a significant covalent perturbation of the biradical, and that they are mutually oriented in such a manner that they still overlap after the action of the angular momentum operator 8' on one of them, that is, after a 90" rotation of the p orbital around an axis that passes through its center (particularly if it is located on an atom of a high atomic number). If the orbitals overlap through space, the likely singlet reaction is covalent bond formation between the two radical centers, as exemplified in Figure 4.26a for n. bond formation in 1,2-biradicals by planarization, and in Figures 4.26b and c for disrotatory and conrotatory a bond formation in 1,n-biradicals by radical recombination. It also occurs when they interact in a through-bond fashion, for example, in I,Cbiradicals, in which case the likely singlet reactions are disproportionation or fragmentation, depending on the details of the geometry as indicated in Figures 4.26~4 and e. In order to understand the rate of intersystem crossing and the nature of the products, it is therefore essential to consider not only the shape of the T potential energy surface, which determines the occupancy of the various conformer minima and the frequency with which the various molecular ge-
Examples of biradicaloid geometries expected to be favorable for spinorbit coupling: overlap leading to covalent interaction y between the localized nonbonding orbitals (full lines) and nonzero overlap after the action of the angular momentum operator k on one of them (dotted lines); a) partial double-bond twist, b) disrotatory, and c) conrotatory ring closure, d) disproportionation, and e) fragmentation. Figure 4.26.
ometries are visited, but also the magnitude of the spin-orbit coupling element and of the singlet-triplet splitting as a function of geometry. To summarize the result of the simple theory, an admixture of Socharacter into the triplet wave function of a biradical is expected to be large at those geometries at which covalent interaction and spin-orbit coupling are both simultaneously large, that is, at which the most localized singly occupied orbitals overlap sufficiently before as well as after a 90" rotation of one of the important p orbitals. A jump to an essentially pure singlet, followed by motion on the So surface, occurs upon excursion of the biradical geometry into an area of strong covalent perturbation (large singlet-triplet splitting, singlet stabilization), with a probability dictated by the degree of admixture of singlet character into the triplet wave function. The simple theory of spin-orbit coupling in biradicals has been found useful for the interpretation of the lifetimes of triplet biradicals as a function of their structure and conformation (Johnston and Scaiano, 1989; Adam et al., 1990) and of the stereochemistry of their reaction products (Chapter 7).
4.4 Pericyclic Funnels (Minima) The global term "pericyclic funnel" will be used to refer to the funnel or funnels in the S, surface that occur at the critically heterosymmetric biradicaloid geometries reached near the halfway point along the path of a thermally forbidden pericyclic reaction, and the minima in S, that are encountered along one-dimensional cuts along reaction paths that miss the conical intersections (in particular, those along high-symmetry paths, which pass
through the geometry of a perfect biradical, and in which the S I S otouching is therefore avoided). This region of the S, surface was usually called "pericyclic minimum" because it was thought that the touching is most likely weakly avoided everywhere. Thanks to the extensive recent work of the groups of Bernardi, Olivucci, and Robb (Bernardi et al., 1990a-c. 1992a-c; Olivucci et al., 1993, 1994a,b) this is now known not to be so (cf. Section 4.1.3). and for quite a few reactions the exact location of the bottom of this funriei has been determined in considerable detail. An understanding of the electtonic states in the region of the pericyclic funnel is of fundamental signitlc,~ncein many photoreactions such as ground-state-forbidden cycloadditionc, electrocyclic reactions, or sigmatropic rearrangements, and it appears assential to discuss these states in some detail. We skiall base the discussion on an analysis of a simple four-electron-fourorbital ibodel, exemplified by H, (the "20 x 20 CI model"), which contains all the dssential ingredients (Gerhartz et al., 1976, 1977), as shown by the more redent calculations on actual molecules (two of these are discussed in more deliail in Section 6.2.1). Although it might thus appear that an understanding~of the electronic states of a tetraradical is necessary, we shall see that alreddy an understanding of the states of biradicals and biradicaloids at the level k>f the two-electron-two-orbital ("3 x 3 CI") model (Section 4.3) is immen$ely helpful, although not quite a substitute for the full 20 x 20 CI descriptign. For those familiar with valence-bond theory, an even simpler "2 x 2" VB model (Bernardi et al., 1988, 1990b) will be described, and the strengths and weaknesses of the 3 x 3 and 2 x 2 models will be compared.
Pericyclic brocesses involving interactions of four electrons in four overlapping orbitab arranged in a cyclic array (A, C, D, B, labeled clockwise along the perimetqr) correspond in general to the switching of two bonds originally connecting oms A with B and C with D, in the reactants to connect atoms A with C an B with D in the final product (Scheme 1). Examples are the conversion i f butadiene into cyclobutene (cf. Section 4.2.2) or norbornadiene (15) ingo quadricyclane (16):
Analogous processes in which the final product has atom A attached to D and atom B attached to C (cross-links in the perimeter) are also known (Scheme 1) and could be called cross-pericyclic (the term "cross-cycloaddition" is relatively common in the literature). An example is the conversion of 1,3-butadiene to bicyclobutane (17). If the new bonds are formed in a suprafacial way, the process is referred to as x[2, + 2,]. The singlet states of the hypothetical H, molecule at triply right tetrahedral geometries calculated by Gerhartz et al. (1977) can be thought of as a simple model for the electronic states involved. The results of model 20 x 20 CI calculations on H, led to the proposal (Gerhartz et al., 1977) that the [2, + 2,] and x[2, + 2,] processes proceed through the same pericyclic funnel in the S, state. This has been supported by recent more realistic calculations (Olivucci et al., 1993, 1994b), and the two processes are therefore best discussed together. To simplify a relatively complex situation, we shall however first pretend that only a relatively high-symmetry version of the [2, + 2,] path is accessible, in which the four orbitals in question (modeled by the four H atoms) are located at the corners of a rectangle. Halfway along this reaction path, they are located at the corners of a square. At this point we shall find a fairly strongly avoided touching of the S, and So states and a "pericyclic minimum" in the S, state. Of course, in H,, this is not at all a minimum, since several types of distortion, such as an increase in the size of the square, lower the energy. A realistic example of such a path is the one used in the correlation diagrams in Figures 4.7 and 4.8. Subsequently, we shall relax the condition that all four H atoms have to be in the same plane, and we shall find a downhill path that leads from this pericyclic "minimum" to an S , S , conical intersection as the square is puckered and the two diagonals, AD and BC, shortened. A real molecule is likely to follow a lower-symmetry path straight to this "cross-bonded" pericyclic conical intersection. Once on the So surface, it will end up as a [2, + 2,], as an x[2, + 2,1 product, or as the starting material. Figure 4.27 is based on the results of a full 20 x 20 C1 calculation for H, and shows the energies calculated for the lowest singlet states of the reaction H, + Hi 2HH' with geometries corresponding to the hypothetical best ground-state rectangular path. The nomenclature used to label the states refers to the whole H, system or "supermolecule." The singly excited S state correlates smoothly from left to right without a barrier; the calculation produces, however, a broad minimum at the biradicaloid geometry. We shall see
If the new bonds are formed in a suprafacial way, the process is referred to as [2, + 2,]. The singlet states of the hypothetical H, molecule at square geometries cgculated by Gerhartz et al. (1976) can be thought of as a simple model for the/ electronic states involved.
Schematic state correlation diagram for the H, + Hi e 2HH' reaction along a rectangular path. Solid and broken correlation lines refer to states of different symmetry (by permission from Michl, 1977).
Figure 4.27.
(Section 5.4.2) that this corresponds to the excimer minimum. The avoided crossing between the ground state G and the doubly excited state D produces the "pericyclic minimum" in the excited-state and the barrier in the groundstate surface. In the nomenclature of the simpler 3 x 3 CI model (BonaCiCKouteckp et at., 1987), square H,corresponds to a perfect biradical close to the "pair" biradical limit (uncharged 4N-electron perimeter), with near degeneracy between the T and So states and between the S, and S, states, and a fairly large gap between the two pairs of states. This agrees well with the relatively large degree to which the intended So-S, touching in Figure 4.27 is avoided. However, in another respect the simplified 3 x 3 CI description fails, as it renders incorrectly the order of the states within the nearly degenerate pairs (it is inherently incapable of placing the D state below the S state, as indicated below: Example 4.12). The ordering of states at the geometry of the pericyclic minimum provided by the full four-electron-four-orbital treatment (20 x 20 CI) may be rationalized only when we combine the two simpler and intuitively more easily grasped models, 3 x 3 CI (MO) and 2 x 2 VB. The comparison will clearly illustrate the strengths and the shortcomings of the two simplified models. Unlike the full 20 x 20 CI model, which contains all the possible MO configurations and all the possible VB structures, the 3 x 3 model only
contains one dot-dot (covalent) structure and two hole-pair (zwitterionic) structures based on the most localized nonbonding MOs. It confines two of the four electrons into a nonpolarizable core (analogous to the most stable MO of square cyclobutadiene), and ignores the availability of the antibonding MO (analogous to the least stable MO of square cyclobutadiene). It will be particularly poor at geometries at which one or both of the two MOs that have been ignored are energetically close to the two nonbonding orbitals that are being considered (e.g., at tetrahedral geometries) or those of large squares. The 2 x 2 VB model considers all four electrons explicitly but permits only the two possible covalent structures. By ignoring all zwitterionic structures, it gives up all hope of describing the excited states allowed for absorption from the ground state, the excimer state, and any states with separated charges. In H,,the energies of structures with separated charges are quite unfavorable, but in real organic molecules they will be stabilized by dynamic correlation with the electrons of the other bonds present, and they can be quite competitive with the excited covalent structures, particulady when the opposite charges are not far from each other. Figure 4.28 shows the energies EG,ED,and Es of the G, D, and $ states of H, at the square geometries obtained from the full 20 x 20 calculhtion as a function of the square size. In the limit of a very small square, the ordering of EG< Es < ED,the same as obtained from the simple 3 x 3 CI description. (Cf. Example 4.12.) In the limit of an infinitely large square, the ordering is EG = ED< Es, the same as in the simple 2 x 2 VB description, with four H atoms coupled into a covalent overall singlet in two different ways. According to Figure 4.28, neither simple description is adequate for square sizes of practical interest. It is perhaps not necessary to go to a full 20 x 28 C1, but
Calculated state correlation diagram for H, at square ge/ometriesas a function of the size of the square. Solid and broken correlation lines4 respectively. refer to states of different symmetry (by permission from Michl, 197711
Figure 4.28.
certainly more than three configurations are needed when the MO approach is used, and the VB approach definitely requires zwitterionic structures. Hence, arguments based on the simplified models of bonding have to be interpreted with due care.
Example 4.12: A 3 x 3 CI calculation for H, takes into consideration the configurations
= + xz - X 3 - x4) and +3 = which can be constructed from the MOs + x2 + x3 + x,) is $xl - x2 - x3 + x,). The most stable MO = neglected in this example: it is always doubly occupied and forms a nonpolarizable core. At square or nearly square geometries the system is a perfect biradical or a hom*osymmetric biradicaloid. respectively. and the energy ordering of the three singlet states may be obtained from Figure 4.19 or 4.20:
For an explicit calculation of the energy order, Slater rules yield the CI matrix elements
We are now ready to give up the fiction that a high-symmetry path along rectangular geometries will be followed in the excited-state pericyclic process and to explore downhill paths from the square version of the pericyclic minimum in H4 thus far considered. It was recognized early on (Gerhartz et al., 1977) that the 20 x 20 CI wave function of the G state of a square array of four orbitals exhibits predominantly singlet local coupling along the perimeter and predominantly triplet local coupling across the diagonals, and that the opposite is true of the covalent part of the wave function of the D state. Naively, in the 2 x 2 V B approximation, one could say that the G state is reached by bringing two ground-state H, molecules together side by side, and the D state by placing them across each other. These are indeed the state correlations found in the full calculation. Figure 4.29 shows cuts through the potential energy surfaces of H4at the 20 x 20 CI level along the already discussed path of rectangles and along the path of triply right tetrahedra, which result when the two diagonals of a square are pulled apart in opposite directions perpendicular to the plane of the square but their lengths are kept unchanged. Along the latter path, the G state increases and the D state decreases in energy, and the two become degenerate when the geometry of a regular tetrahedron is reached. Afterward, the D state represents the So surface. Ultimately, if the lengths of the diagonals were permitted to adjust to the ground-state equilibrium values in Hz, the energy would drop much more, and the ordinary ground state of two Hz molecules would result. As this situation is approached, the zwitterionic contributions to the D state are reduced in size, and the 2 x 2 VB description, which ignores them completely, would be quite adequate. Already at the geometry of the G-D state crossing (So*, state touching), the 2 x 2 VB model is quite good in a qual-
where the last line follows from symmetry arguments. Thus the C1 problem factorizes into a 2 x 2 problem and a I x I problem. Using the ZDO approximation, one has I, = I3 and J2, = J,, = Jz3. Thus K& = [(JZ2+ J3J/2 J2,]/2 = 0 and the state energies can be written down immediately:
Since in the ZDO approximation Ki3 vanishes, the S and D states are degenerate. Interactions with additional configurations are needed in order to stabilize D to such an extent that ED< Es.
Figure 4.29. Schematic state correlation diagram for the H, + Hi e 2HH' reaction along the rectangular path (x) and along the tetrahedral path for R, = R, ( y ) .
itative sense. In contrast, the 3 x 3 CI model, incapable of describing more than a single covalent function, is unsuitable. One can expect other downhill paths from the square "pericyclic minimum" in the D state, in which advantage will be taken of the diagonal singlet coupling, but the two diagonals will not remain of equal length. For instance, a distortion of the square into a rhombus will shorten one and extend the other diagonal. A continuation of this path will not produce two H, molecules but only one, plus two hydrogen atoms, in the right electronic state for subsequent barrierless recombination into a second H2molecule. In an organic molecule, this path would correspond to the formation of a I ,4-biradical ready to collapse with the formation of an additional bond. The effects of this type of distortion are also well described qualitatively by the 2 x 2 VB model. (See Example 4.13.) Now, however, even the 3 x 3 CI model predicts an S,-So touching, since unequal diagonal interactions in a square represent a heterosymmetric perturbation S. (Cf. Section 4.3.) When S is allowed to reach a critical value, a heterosymmetric biradicaloid with an So-S, degeneracy will result. The perturbation due to the introduction of resonance integrals across a conjugated perimeter was discussed in some detail in Sections 2.2.3-2.2.7 in connection with the perimeter model. Thus, a diagonal bond in cyclobutadiene or a 1,5 bond in octagonal cyclooctatetraene is expected to convert these perfect biradicals into heterosymmetric biradicaloids, to split the degeneracy of the nonbonding orbitals, and to lead toward an S,-So touching (Howeler et al., 1989). The polarizing perturbation 6 introduced in this manner can be quite large; in going from cyclooctatetraene to pentalene, 6 reaches So and then exceeds it, s o the S , and So states touch (Doehnert and Koutecky, 1980). It is clear that the critical value 6, expected from the simple formula Equation 4.6 will not be quite right, since the 3 x 3 CI model ignores the effect of the second covalent V B structure. In a similar vein, one would not expect the analogous formula for the exact position of the state touching that is derived from the 2 x 2 V B model (Example 4.13) t o be exactly right either, since the zwitterionic contributions to the D state are now left out. This inexactness should be particularly damaging when one considers the effects of electron-donating and electron-withdrawing substituents. Qualitatively, however, it is clear from the consideration of the states of H, that an energy decrease in S , and a S,-So touching are to be expected upon distortion of the square geometry of the initially diagonal interactions across the perimeter. Even in real organic molecules, the "bottom" of the pericyclic funnel is therefore to be sought at lower symmetry geometries containing such distortions. Since one or the other of a square's diagonals can be shortened, one can expect that two o r possibly even a larger number of conical intersections will flank the "pericyclic minimum," producing rather intricate topological structure for the resulting "pericyclic funnel."
Example 4.13: in the 2 x 2 VB model (Bernardi et al., 1988, 1 M b ) the ground- and excited-
where Q is the Coulomb energy and T the total exchange energy given by the London formula (cf. Eyring et al., 1944):
K,, K,, and K , are defined (Bernardi et al., 1988, 1990b) in terms of two center VB exchange integrals* as follows:
if orbitals 1.2 and 3,4 are coupled in the reactants while orbitals 1,3 and 2,4 are coupled in the product. (Cf. Scheme I .) At the conical intersection the two states are, by definition, degenerate ( E , = E,), and the total exchange energy T must vanish. Therefore the following two independent relations must hold: K, KR
Since the behavior of the exchange integrals K, is easily predicted as a function of the geometrical coordinates, the structure corresponding to a conical intersection can be predicted by distorting the molecular structure of the system in such a way as to satisfy those two equations. (Cf. Section 6.2.1 .) In practical applications of the simple H, model there will be other bondforming electrons in addition to the four electrons involved in the bondswitching o r bond-crossing process. These additional bonds will prevent many of the total or partial fragmentation processes expected for H,. For instance, the uninterrupted downhill slope of the D surface toward four separated atoms (cf. Figure 4.28) will not exist in the case of the [2, + 2,] cycloaddition of two ethylenes, since the C-C a bonds of the two ethylenes will prevent it. With due caution, however, it should be possible to transfer much of the information obtained from the 20 x 20 CI model to [2, + 2,] and x[2, + 2,l reactions in general, particularly information concerning the *The two-center exchange integral K, has the same interpretation as in the Heitler-London tmatment of H,and can be written as K, = (ij1g)ji) + 2s&ilhlJJ,where (ijlglji) is the usual two-electron-exchange repulsion integral, (i(A(nis the one-electron integral, and S, the overlap integral between orbitals iand j .
physical nature of the various excited states, and qualitative understanding of the reasons for the ways in which their energies change with the molecular geometry. (Cf. Section 6.2.1 ;Chapter 7.)
4.4.2 Spectroscopic Nature of the States Involved in Pericyclic Reactions In discussing certain photochemical reactions, such ils photocycloadditions, it is of interest to relate the electronic states at the pericyclic minimum to the states of the two reacting fragments that come together side by side, in the present model case, to the states of two H2 molecules. It then turns out that in the limit of infinite intermolecular separation, the G state corresponds to a combination of two ground-state fragments H2 + H,, the S state to a combination of a ground-state fragment H, with a singly excited fragment H:, and the D state to an overall singlet coupling of two triplet-excited fragments H: + H:. Thus, at infinite separation, the S state can also be labeled exciton state and the D state triplet-triplet annihilation state. There actually are two exciton states, represented simply as H,H: 2 H,*H2;S is the lower of these, and the higher one is labeled Sf in Figure 4.27. In the limit of infinite internuclear separation, two additional charge-transfer or ion-pair states can be identified, which can be represented simply as H,@H,@ + H,@[emailprotected] Figure 4.27 they are labeled I and 1'. At finite separations, these four zero-order terms interact pairwise and produce states S, S t , I, and I' of strongly mixed exciton-charge-transfer character, where the V B functions are ionic and contain terms such as A-BC")llOo A-BCoL)(+) (exciton) or A"'B-CIl' ' (CT). Of the four states. S is greatly stabilized compared to the others, through exciton as well as CT interactions-that is, exactly by the factors responsible for the stability of excimers and exciplexes. (Cf. Sections 5.4.2 and 5.4.3.) The S state has a minimum at the biradicaloid square geometry that may be referred to as the excimer minimum. A strong absorption band in the near-1R region corresponding to a transition between the S and S' states of the fluorene excimer has been observed by photodissociation spectroscopy (Sun et al., 1993). Because of the especially high symmetry and because all the electrons are involved in binding interactions, the conditions in H, are evidently optimal for the formation of an excimer minimum. The situation will be different for photocycloadditions in which ~cbonds are converted into a bonds and were additional electrons in mutually repelling closed shells make an intimate approach of the two reactants difficult. In these cases, the excimer minimum will be flatter and will occur only at larger nuclear separations, where the molecules are barely touching. (Cf. Section 5.4.2 and Figure 6.7.) The D state with its pericyclic funnel is not related to the four low-energy ionic states. It originates from a mixing of the triplet-triplet annihilation wave function with totally symmetric higher excited configurations and
charge-transfer terms and acquires partial ionic character as the intermolecular separation decreases. In contrast to the excimer minimum, the pericyclic minimum will generally occur at a geometry at which the old bonds in the perimeter are half-broken, the new bonds in the perimeter half-established, and a strong diagonal interaction introduced-that is, roughly halfway along the reaction path. The relative location of the excimer minimum and the pericyclic funnel can thus vary over a wide range. (Cf. Section 6.2.3 and Figure 6.17.) They can represent two separate topological features in the S, surface, but it is also conceivable that they lie above each other and are not both present in the S, surface. (Cf. Figure 7.27.)
Supplemental Reading General Michl, J. (1972). "Photochemical Reactions of Large Molecules." M o l . Photochem. 4, 243, 257, 287. Michl, J. (1974). "Physical Basis of Qualitative M O Arguments in Organic Photochemistry," Fortschr. Chem. Forsck. 46. I . Michl. J. (1978). "The Role of the Excited State in Organic Photochemistry," in Excited States in Qrtantrtr?~Chemistry; Nicolaides, C.A., Beck, D.R., Eds.; Riedel Publ.: Dordrecht. Michl, J., BonaEiC-Kouteckq, V. (1990). Electronic Aspects of Organic Photochemistry; Wiley: New York. Simons. J. (1983), Energe,ric Principles oj'Chemicrr1 Rec~ctions;Jones and Bartlett Publ.: I3oslon.
Potential Energy Hypersurfaces Klessinger. M . (1982). Elektrortc~n.strrrktrrrorgcrnisc.her Molckiile; Verlag Chemie: Weinheim. Mezey, P.G. ( 1987). Potentic11Etrcjrgy lfyl>c*r.~u~firc.c.s; Elsevier: Amsterdam. Salem, 1,. (1982). Elc,c.rrorrs in Chemical Rc,rrc~riotts:First Princ.iples; Wiley: New York.
Spin-Orbit Coupling and Hyperfine Interactions Gould. I.R.. Turro, N.J.. Zimmt, M.B. (1984). "Magnetic Field and Magnetic Isotope Effects on the Products of Organic Reactions"; Adv. Phys. Org. 20, 1. Khudyakov, I.V., Screbrennikov. Y.A.. Turro, N.J. (1993). "Spin-orbit Coupling in Free Radical Reactions on the Way to Heavy Elements." CAem. Rel.. 93. 537. Michl, J. (1991). "The States of an Electron Pair and Photochemical Reactivity," in Theorc~tic~rl crnd Comprrtcrrio~rcrlModelsSor Organic. Chemistry; Formosinho. S.J., et al., Eds.; Kluwer: Dordrecht. Richards, W.G.. Trivedi, H .P., Cooper, D . L . (198 1 ), Spin-Orbit Corrpling in Molecules; Clarendon Press: Oxford. Salem, L., Rowland, C. (1972). "The Electronic Properties of Diradicals," A n g e ~ , Chem. . 1111. Ed. EngI. 11, 92.
Salikhov, K.M., Molin, Y. u. N., Sagdeev, R.Z., Buchachenko, A.L. (1984), Magnetic and Spin Effects in Chemical Reactions; Elsevier: Amsterdam.
Johnston, L.J.. Scaiano, J.C. (1989). "Time-Resolved Studies of Biradical Reactions in Solution." Chem. Rev. 89, 521.
Steiner, U.E., Ulrich, R. (1989). "Magnetic Field Effects in Chemical Kinetics and Related Phenomena," Chem. Rev. 89.51.
Michl, J. (1977), "The Role of Biradicaloid Geometries in Organic Photochemistry," J. Photochem. Photobiol. 25. 141.
Michl, J. (1992). "Singlet and Triplet States of an Electron Pair in a Molecule-a Model," J. Mol. Struct. (Theochem) 260, 299.
BonaCiC-Kouteckf, V. (1983), "On Avoided Crossings between Molecular Excited States: Photochemical Implications," Pure Appl. Chem. 55,213.
Platz, M.S., Ed. (1990). Spectroscopy of Curbenes and Biradicals; Plenum: New York.
Devaquet, A. (1975). "Avoided Crossings in Photochemistry," Pure Appl. Chem. 41,455.
Gerhartz, W..Poshusta, R.Il., Michl, J. (1976). "Excited Potential Energy Hypersurfaces for H, at Trapezoidal Geometries. Relation to Photochemical 2s + 2s Processes." J. Am. Chem. Soc. 98,6427.
Devaquet, A., Sevin, A,, Bigot, B. (19781, "Avoided Crossings in Excited States Potential Energy Surfaces," J. Am. Chem. Soc. 100,479. Salem, L.. Leforestier, C., Segal, G., Wetmore. R. (1975). "On Avoided Surface Crossings," J. Am. Chem. Soc. 97.479.
Gerhartz. W.. Poshusta. K.11.. Michl. J. (1977). "Excited Potential Energy Hyperwrf:~ces for H,.2. 'Triply Right' (C'!,) Tetrahedral Geometries. A Possible Relation to Photochemical 'Cross-Bonding' Processes." J. A~rr.Clrc~rrr.Soc.. 99. 4263.
Atchity, G.J., Xantheas, S.S., Ruedenberg, K. (1991) "Potential Energy Surfaces near Intersections", J. Chem. Phys. 95, 1862.
Lippert. E., Rettig, W., Bonatif-Koutecky, V., Heisel, F., Miehe, J.A. (1987), "Photophysics of Internal Wisting," Adv. Chem. Phys. 68, 1.
Bernardi, F.,Olivucci, M., Robb, M .A. ( 1990), "Predicting Forbidden and Allowed Cycloaddition Reactions: Potential Energy Topology and Its Rationalization." Acc. Chem. Res. 23, 405.
Rettig, W. (1986). "Charge Separation in Excited States of Decoupled Systems." Angrw. Chem. Int. Ed. Engl. 25,971.
Herzberg, G., Longuet-Higgins, H.C. (1963). "Intersection of Potential Energy Surfaces in Polyatomic Molecules," Disc. Faraduy Soc. 25.77. Longuet-Higgins, H.C. (1975). "lntersection of Potential Energy Surfaces in Polyatomic Molecules," Proc. Roy. Soc. London A344, 147. Teller, E. (1937), "The Crossing of Potential Surfaces", J. Phys. Chem. 41, 109. Teller, E. (1%9), "Internal Conversion in Polyatomic Molecules," Israel J. Chem. 7, 227.
Correlation Diagrams Bigot, B., Devaquet, A., Turro, N.J. (1981), "Natural Correlation Diagrams. A Unifying Theoretical Basis for Analysis of n Orbital Initiated Ketone Photoreactions." J. Am. Chem. Soc. 103.6. Michl, J. (1974). "Photochemical Reactions: Correlation Diagrams and Energy Barriers," in Chemical Reactivity and Reaction Paths; Klopman, G., Ed.; Wiley: New York. Pearson, R.G. (1976). Symmetry Rules for Chemicul Reactions; Wiley: New York. Woodward, R.B., Hoffmann, R. (1970), The Conservation of Orbital Symmetry; Verlag Chemie: Weinheim.
Biradicals and Biradicaloids BonatiC-Koutecky, V., Kouteck9, J.. Michl, J. (1987), "Neutral and Charged Biradicals. Zwitterions, Funnels in S, and Proton Translocation: Their Role in Photochemistry, Photophysics and Vision," Angew. Chem. Int. Ed. Engl. 26, 170. Borden. W.T. (1982). Dirudic.e~Is;Wiley: New York. Johnston, L.J. (1993). "Photochemistry of Radicals and Biradicals," Chem. Rev. 93, 251.
Electronically excited states have only a short lifetime. In general, several processes are responsible for the dissipation of the excess energy of an excited state. These will be discussed in the following sections. For this purpose it is useful to distinguish between photophysical and photochemical pathways of deactivation, although such a distinction is not always unequivocal. (Cf. the formation of excimers, Section 5.4.2.) The present chapter deals with photophysical processes, which lead to alternative states of the same species such that at the end the chemical identity of the molecule is preserved. Photochemical processes that convert the molecule into another chemical species will be dealt with in later chapters.
5.1 Unimolecular Deactivation Processes The excess energy taken up by light absorption can be dissipated through unimolecular processes either as radiation (emission) or by radiationless transitions. It can also be transferred to other molecules through bimolecular processes. The relative importance of these various processes depends on the molecular structure as well as on the surroundings of the molecules.
5.1.1 The Jablonski Diagram The various unimolecular photophysical processes may be envisaged in a rather illuminating way with the help of the Jablonski diagram shown in Fig-
Figure 5.1. Jablonski diagram. Absorption (A) and emission processes are indicated by straight arrows (F = fluorescence, P = phosphorescence), radiationless processes by wavy arrows (IC = internal conversion, ISC = intersystem crossing, VR =
vibrational relaxation).
ure 5.1. This diagram schematically displays the singlet ground state So, the excited singlet states S, and S,, as well as the triplet states TI and T,. For polyatomic molecules the spacing between the vibrational levels decreases rapidly with increasing energy, and the density of states increases very rapidly as the vibrational energy increases. Rotational levels of the various vibrational states have therefore been omitted for the sake of clarity. Standard convention shows absorption and emission processes as straight arrows and radiationless processes as wavy arrows. Bimolecular photophysical processes and photochemical processes are not shown in the Jablonski diagram. They provide different possible pathways for the deactivation of excited states and will be discussed in later sections. Emission or luminescence is referred to as fluorescence or phosphorescence, depending on whether it corresponds to a spin-allowed or a spinforbidden transition, respectively. Similarly, radiationless transitions between states of the same multiplicity and of a different multiplicity are known as internal conversion (IC) and intersystem crossing (ISC), respectively. From Figure 5.1 it can be seen that a molecule can reach an excited vibrational level of the electronically excited state S, either by the absorption
of a photon of appropriate energy or by internal conversion from one of the vibrational levels of a higher electronic state such as S,. In liquid solutions vibrational rc.laxution (VR)to the vibrational ground state (or more accurately, to a Boltzmann distribution over the vibrational levels corresponding to thermal equilibrium) is very rapid, and the excess vibrational energy is converted into heat through collisions with solvent molecules. From the zero-vibrational level of the S, state the molecule can return to the ground state S, by fluorescence (F), or it can reach the triplet state T, by intersystem crossing (ISC), and after loss of excess vibrational energy it can return to the ground state Soby phosphorescence. Radiationless deactivation from S, to So can occur via internal conversion (IC) and subsequent vibrational relaxation (VR). Radiationless deactivation from TI to Socan occur by intersystem crossing (ISC) followed by vibrational relaxation. In many instances intersystem crossing is fast enough to compete with fluorescence or even with vibrational relaxation. As indicated in Figure 5.1, higher excited triplet states can be reached under such circ*mstances, and subsequent internal conversion and vibrational relaxation then proceeds in the triplet manifold. Higher excited triplet states are also accessible from TI via triplet-triplet absorption. Yet another pathway is shown in Figure 5.1. This corresponds to the thermal activation of TI and reverse intersystem crossing into S,. This process gives rise to E-type delayed fluorescence, which derives its name from the circ*mstance that it was first detected for eosin, in contrast to P-type delayed fluorescence, initially observed for pyrene. The latter will be described in Section 5.4.5.5.
5.1.2 The Rate of Unimolecular Processes If the spontaneous emission of radiation of the appropriate energy is the only pathway for a return to the initial state, the average statistical time that the molecule spends in the excited state is called the natlrral radiative lifetime. For an individual molecule the probability of emission is time-independent and the total intensity of emission depends on the number of molecules in the excited state. In a system with a large number of particles, the rate of decay follows a first-order rate law and can be expressed as where I, and I are the intensities of emitted radiation immediately after excitation and at a later time t , respectively (cf. Figure 5.2); &,, is the rate constant and has the dimension of reciprocal time. The quantity is the mean natural lifetime of the excited state (in s). k, in Equation (5.2) is the rate constant for spontaneous emission and is given by the Einstein prob-
Each process competing with spontaneous emission reduces the observed lifetime t relative to the natural lifetime to.In the case where only unimolecular processes i with rate constant ki compete with emission, one has
Figure 5.2.
rate law.
ability of spontaneous emission A,,. lifetime in s can be estimated as
where C, is the wave number of the absorption maximum and f the oscillator strength of the corresponding electronic transition (Strickler and Berg, 1962). Since emission shows the same dependence on the transition moment as absorption, it follows that the emission decay time is inversely proportional to the integrated intensity of the absorption. Using iJ = 40,000 cm-' for a multiplicity-allowed transition and a singletand f = I or f = triplet transition, respectively, the natural radiative lifetimes of an excited singlet and triplet state may be estimated from Equt~tion(5.3) as t,, = s and to= I s, respectively. The radiative lifetime is nearly temperature independent, but depends to some extent on the environment. Thus, the solvent can influence the radiative lifetime either by means of its effect on the transition moment (cf. Section 2.7.2) or through its refractive index. (See Forster, 1951.) From the above estimate of the lifetime of a singlet state the rate constant of fluorescence kF = lltOis obtained as kF = lo6-10' s-I, depending on the value of the oscillator strength f. Phosphorescence, on the other hand, is a spin-forbidden process with a much smaller rate constant. k. = 10-2-l@ S I.
According to Equation (5.3), the rate constant of IR emission is much smaller than k,, since 3 is correspondingly small. Vibrational energy is therefore ordinarily transferred to the surroundings through radiationless dissipation and converted to translational, librational, and rotational energy by collisions. At sufficiently high pressures the collisional frequency is loi3s-I, so the vibrational relaxation rate constant kvR is expected to be kvR 5 lot3s- I. In fact, a value kvR = (4 2 1) x 1012s- I has been measured by picosecond spectroscopy for the vibrational relaxation of the first excited singlet state of 9,lO-dimethylanthracene (Rentzepis, 1970) and many similar values have been measured for other molecules. In solid solutions at low temperatures, vibrational relaxation may be slower by several orders of magnitude. The rate of internal conversion (IC), a radiationless transition between isoenergetic levels of different states of the same multiplicity, may be of the same order of magnitude or even faster than vibrational relaxation. It depends, however, on the energy separation AE,, between the zero-vibrational levels of the electronic states involved (energy gap law, see Section 5.2.1). Similar relations hold for intersystem crossing transitions between states of different multiplicity, which are slower by 4-8 orders of magnitude. In conclusion, the following rough estimates of rate constants may be s-I for S,-.S, and given: For internal conversion for instance k,, = 1012-1014 klc < 108 s-I for SI-SO, whereas for intersystem crossing typical values are ksT = I@-10" s- for Sl-TI and kTs = lo4-10-I s - for TI-SO. Intersystem crossing in carbonyl compounds is particularly fast. In biradicals, where T, and So are very nearly degenerate, kTs tends to be high, often in the range 106- I08 (Johnston and Scaiano, 1989).
5.1.3 Quantum Yield and Efficiency A useful quantity in the description of photophysical and photochemical processes is the quantum yield @. The quantum yield ajof a processj is defined as the number n, of molecules A undergoing that process divided by the number nQ of light quanta absorbed, that is,
the ratio of the rate dn,,ldt of the process j to the intensity of the absorbed radiation I,,, = dnddt.* If the quantum yield is not constant over the entire reaction time, these two definitions of the total and the differential qirantum yield agree only after extrapolation to the time t = 0. For many purposes it will be useful to distinguish between quantum yield a, related to the absorbed radiation and efficiency qj related to the number of molecules in a given state. may be defined by
njlnA.
Table 5.1 Relations Between Quantum Yield, Lifetime, and Rate Constant of Unimolecular Photophysical Processes
which is the ratio of the number n, of molecules that take a specific reaction path j to the number nAlof molecules in the precursor state A*, Analogously, the efficiency may be expressed as the ratio of a process involving a given excited state to the rate of production of that state. If z k i is the sum of rate i
constants for all processes under consideration, the probability that any one of the molecules will yield the product j is given by
quantum yield of fluorescence is given as the ratio of the observed to the natural lifetime. Expressions for the quantum yield of internal conversion a,, and of intersystem crossing @ ,,, which populates the T, state, may be derived in a similar way. If the TI state is deactivated only by first-order processes such as phosphorescence and intersystem crossing with rate constants k, and kTs, the quantum yield of phosphorescence is given according to Equation (5.8) by
of a reaction R is given as the product of the efficienThe quantum yield cies of all steps necessary to reach the product R. Therefore, for a one-step reaction @R
whereas for a reaction with one metastable intermediate Z (5.9b) For the selective excitation of the zero-vibrational level of the S, state tlabr = 1; in general this is also valid for higher energy radiation. For a twophoton process, where two photons are absorbed simultaneously, correspondingly rl.bs= 0.5. If only processes that obey a purely exponential rate law such as fluorescence, internal conversion, and intersystem crossing with rate constants k,, kIc, and klsc are involved in deactivating the singlet state S,, the quantum yield of fluorescence may be written according to Equations (5.8) and (5.9) as
where the appropriate expression for the efficiency qsT for triplet state formation has been used. In the same way, an expression for the quantum yield aTS of the radiationless deactivation of the TI state may be obtained. All these relationships are collected in Table 5.1.
where the natural lifetime $ and the observed lifetime ts of the singlet state are given by Equations (5.2) and (5.4). Thus, under these circ*mstances the
Table 5.2 Observable Photophysical Parameters and their Relationship to Rate Constants of Various Photophysical Processes and Sources of their Information* Photophysical Parameter Fluorescence quantum yield Phosphorescence quantum yield Triplet formation quantum yield Singlet lifetime Triplet lifdime
* If the absorbed intensity is measured per volume V, Equation (5.6) takes the form @, = (dn,ldr)ll., = (d~,ldr)/(l.,,V),where c, is the concentration of A .
a),.
k,, klc + ks, + k, k~ AS, ( ~ T S+ k ~ (klc ) + ksr + k,) ksr klc + k,- + kF I klC+ k, + kF 1 k.,., + k,
TI.
Source Fluorescence spectrum Phosphorescence spectrum TI-T,, absorption Fluorescence decay Phosphorescence decay
*These relations hold assuming that second-order processes such as quenching and photochemical reactions may be neglected.
The quantum yields of fluorescence and phosphorescence, aFand a,, may be determined experimentally by means of a fluorescent standard such as a rhodamine B solution whose @, is independent of the exciting wavelength within a wide range. Lifetimes tF and tpare also experimentally accessible through time-resolved fluorescence measurements (phase method or single-photon counting) or by measuring the time dependence of phosphorescence. (Cf. Rabek, 1982.) In Table 5.2 the observable quantities and their relationship to rate constants are collected.
5.1.4 Kinetics of Unimolecular Photophysical Processes From the five relationships collected in Table 5.2, the five rate constants kF, and kTs of unimolecular photophysical processes may in principle be obtained if all observable quantities are known. Simplifying assumptions are often possible. For instance, if internal conversion S,*S, is negligible = 1 or
5.1.5 State Diagrams Data obtained from spectroscopic measurements can be collected in a Jablonski diagram. In this way a state diagram is obtained for a molecule, and this can be extremely useful in discussing its photochemistry. As an example the state diagram of biacetyl is displayed in Figure 5.3. From the 0-0 transition in the absorption or fluorescence spectrum, as well as in the phosphorescence spectrum, the energies E(S,) and E(Tl) of the lowest singlet and triplet states are obtained. From the fluorescence quantum yield @, = 0.01 it is concluded that only 1% of the excited molecules emit; the remaining 99% are transferred into the TI state, assuming that radiationless deactivation Sl-So and photochemical reactions are negligible. From the fluorescence decay one has k, = lo5 s-I = I/$, and the rate constant for intersystem crossing ksT is obtained from the relationships collected in Table 5.2 as
4', = 0.25 means that 25% of the molecules in the TI state phosphoresce with k, = 1.2 x lo2 s-I. The remaining 75% return nonradiatively to the ground state, with
is obtained. Example 5.1 : For benzene in EPA at 77 K @, = 0.19, @, = 0.18, and t, = 6.3 s have been measured (Li and Lim. 1972); from the integrated area under the absorption curve k , = Ilt, = 2 x IW s - I is obtained. With the simplifying assumption @b + @, =. 1 the relationships collected in Table 5.2 yield
according to Equation (5.12).
results. Using qTs=
is obtained from Equation (5.1 I), and finally Equation (5.12) yields Figure 5.3.
son. 1%3).
Figure 5.4. Jablonski energy diagram a) of benzophenone and b) of I-chloronaphthalene (by permission from Turro, 1978).
Figure 5.4 displays state diagrams of other molecules of interest. (Cf. Example 6.6.)
by v and v' of different electronic states n and m of the same multiplicity. which may have quite different energies at their respective equilibrium geometries. Often, however, the designation internal conversion is used in a wider sense encompassing vibrational relaxation as well. It then denotes radiationless transitions Sf,-S, or T,,-TI from a higher excited singlet or triplet state into the lowest excited singlet or triplet state, respectively, as well as the radiationless transition Sl-So from the first excited singlet state into the vibrationally equilibrated ground state. The raditionless processes Sf,-S1 and T,,-TI are usually so fast that lifetimes of higher excited states are very short and quantum yields of emission from higher excited states are very small. In the vast majority of cases luminescence is observed exclusively from the lowest excited state. This socalled Kasha's rule is ofcourse relative in that what is observed depends on the sensitivity of the detector. For benzenoid aromatics, fluorescence from higher excited states in addition to fluorescence from the lowest excited state was observed for the first time in 1969 (Geldorf et al., 1969), whereas the fluorescence from the S, state of azulene had been known for quite some time. (See below.) The radiationless tri~nsitionS,-S,,, however, is in general much slower than Sf,-S,. For instance, most aromatic compounds fluoresce, and the internal conversion SI-SO contributes at most partially to the deactivation of the first excited singlet state. Thus, the cascade of nonradiative conversions of higher excited states normally ends at the S, state, and does not lead to the ground state So.
5.2 Radiationless Deactivation The electronic excitation of a molecule in general produces a state whose equilibrium geometry differs from the ground-state equilibrium geometry. If the excitation energy is sufficiently high, molecular vibrations of the higher state will also be excited. Reactions of "hot" molecules generated in this way will be discussed in Chapter 6. Before a polyatomic molecule can move over potential energy barriers along a reaction path, the initially introduced vibrational energy is quickly distributed among the various normal vibrations and can be partially or completely dissipated by collisions. This is particularly true for condensed phases where the exchange of vibrational energy with the environment is so fast that thermal equilibrium can be reached in a time as short as lo-'' s. (Cf. Maier et al., 1977.)
5.2.1 Internal Conversion According to Figure 5.1 the term internal conversion (IC) is used for a transition s:-~: or T;-T: between two isoenergetic vibrational levels denoted
From the oscillator strength f = 1.89 of the absorption band of anthracene (1) at 39,700 cm - I the natural lifetime of the excited state can be estimated from Equation (5.3) as t,, = 0.5 x 10- s. Since no fluorescence from this higher excited state S,, is observed, the actual lifetime must be smaller at least by a factor of lo-.'. so one has
for the rate constant of internal conversion. In the case of pyrene (2) the environment-dependent rate constant for the deactivation of the S, state is approximately 106 s-I, and since deactivation occurs practically exclusively by fluorescence and intersystem crossing, one has
Thus, typical rate constants for the two nonradiative processes Sf,-Sl and St-So differ by a factor of 1V-10' (Birks, 1970).
From experimental results it has been concluded that for aromatic hydrocarbons the radiationless transition Sl-SO is negligible if the energy difference AE(Sl - SO)between S, and So states is larger than 60 kcallmol. For molecules with low-lying singlet states such as tetracene (3) (AE = 57 kcall mol) and its hom*ologues, however, it becomes increasingly more important, and it accounts for more than 90% of the Sl state deactivation in the case of hexacene (4) (AE = 40 kcallmol) (Angliker et al., 1982). These observations can be summarized by the relationship which shows the dependence of the rate constant klc on the energy gap AE(Sl - So)and which is referred to as the energy-gap law (Siebrand, 1967). a is a proportionality constant, approximately equal to 4.85 eV-' for benzenoid aromatics. Exceptions to the rules described are observed in the case of azulene (5) and its derivatives (Beer and Longuet-Higgins, 1952). where for the internal conversion S p S , the rate constant k I C ( S p S I= ) 7 X loRs - , is exceptionally small. This may be at least in part due to the large energy gap, AE(S, - S,) .= 40 kcallmol. The radiationless transition from the first excited singlet state into the ground state, on the other hand, is extremely fast with k,c(Sl-So) = 1012s - I , although the energy gap AE(Sl - SO)is of the same order of magnitude as AE(S, - S,). This is consistent with computational results which demonstrate the existence of an S,-S,, conical intersection that can be reached from the Franck-Condon region with almost no barrier (Bearpark et al.. 1994). Another class of compounds that exhibit S2-S,, fluorescence are the thiocarbonyl compounds (Maciejcwski and Steer. 1993; cf. Section 6.1.5.3).
vation of the lowest excited singlet and triplet states, that is, the intersystem crossing S,-TI and TI-SO, with rate constants k, and kTs. The transition SlmTl can take place either by direct spin-orbit coupling of S, to the higher vibrational levels of TI or by spin-orbit coupling to one of the higher states TI,followed by rapid internal conversion TI,-TI. The ratedetermining step is the spin inversion, and rate constant values ksr are in the range lo7to 10" s ' and depend on the extent of spin-orbit coupling as well as on the energy gap between the states involved. According to the selection rules for intersystem crossing known as El Suyed's rules (El Sayed, 1963) transitions are allowed, while transitions are forbidden. These selection rules can be related to spin-orbit coupling with the help of the Fermi golden rrrle. (Cf. Section 5.2.3.) The values ksT .= lo6 s- I for naphthalene and ksT .= 10' s - ' for 1-bromonaphthalene (Birks, 1970), the difference of which can be explained through the heavy atom effect, also indicate clearly the influence of spin-orbit coupling. The fact that the rate
5.2.2 Intersystem Crossing Transitions from singlet to triplet states and vice versa become possible through spin inversion. Of particular importance is the radiationless deacti-
Figure 5.5. Relationship between the energy gap AE(T, - So)and the logarithm of the rate constant k,, of intersystem crossing in aromatic hydrocarbons (data from Birks. 1970).
constants ksT and kTs can differ by a factor of up to loY,so that, for instance, km = 0.41 s-' for naphthalene, can also be related to the energy gap. From Figure 5.5 it can be seen that the kTs values of aromatic hydrocarbons show a similar dependence on the energy gap AE(T, - So)between the T I and S,, states to the one given in Equation (5.13) for klc. Example 5.3: Intersystem crossing is temperature dependent in some anthracene derivatives. Thus, the rate constant for 9.10-dibromoanthracene (6) may be written as
with an activation energy E,, = 4 kcallmol. Triplet-triplet absorption spectra show that T, is energetically higher than S, by 4-5 kcal/mol. It can be concluded that due to the large energy gap between S, and TI,intersystem crossing does not occur into one of the higher vibrational levels of TI, but with a small activation energy instead predominantly into T? (Kearvell and Wilkinson, 1971).
The observation that the rate constants k, for the S,-T intersystem crossing in anthracene (1) and pyrene (2), k,(anthracene) .- 108 s - ' and k,,(pyrene) = 106 s-', differ by a factor of 100 may be explained along similar lines. In pyrene a direct transition into a vibrationally excited level of TI occurs, the energy gap between the zero-point vibrational levels being AE(S, - TI) .- 30 kcal/mol (Dreeskamp et al., 197% whereas in anthracene the transition leads to the nearly isoenergetic T, state (Almgren, 1972).
5.2.3 Theory of Radiationless Transitions The probability of radiationless transitions between different states is particularly great when the potential energy surfaces of the corresponding states touch or come at least very close to each other. In the framework of the Born-Oppenheimer approximation, radiationless transitions from one surface to another are impossible. (See, e.g., Michl and BonaCiC-Kouteckq, 1990.) It is therefore necessary to go beyond the BornOppenheimer approximation and to include the interaction between different electronic molecular states through the nuclear motion in order to be able to describe such transitions. Using the time-dependent perturbation theory for the rate constant k,,, of a transition between a pair of states one arrives at
where & is the perturbation operator, 9,and 0, are the wave functions of the initial and the final state, and Q, is the density of states, given by the number of energy levels per energy unit in the final state at the energy of the initial state (Bixon and Jortner, 1968).* Equation (5.15) is referred to as the Fermi golden rule for the dynamics of transitions between states. In the case of internal conversion between states of equal multiplicity & = ANis the kinetic energy operator of the nuclei. In the case of weak coupling the matrix element of the perturbation operator can be split into an electronic part PIc and a contribution due to the vibrational terms which, with the help of further simplifying assumptions, can be written as the Franck-Condon overlap integral: In the case of intersystem crossing transitions between states of different multiplicity, an additional spin-orbit coupling term As, has to be considered. From the perturbational expansion it follows that the contribution
The rate of intersystem crossing can be increased by the presence of paramagnetic molecules such as oxygen as well as by molecules with heavy atoms such as halogen or organometallic compounds. These are concentration-dependent effects. In the case of oxygen the rate constant of a radiationless singlet-triplet transition can be written as
due to the spin-orbit interaction is dominant. Here qiand Yf are wave functions of different multiplicity (singlet and triplet wave functions) and is the electronic part of the interaction integral. The El-Sayed selection rules [Equation (5.14)] follow from Equation (5.17). (Cf. Example 1.8 and Section 4.3.4.) -
molar solution of 0, one has where k z = 109-1010mol-I s-I. For a k>[0,] = lo7-I@ s-I, so the oxygen effect becomes noticeable when ksr = 10' or smaller (Stevens and Algar, 1967).
The density of states is approximately given by the number of ways of distributing A E over the normal modes of vibration. Because of many low-frequency modes, the number of these overtones and combinations is enormous: there are as many as -3 x 10' states per cm-I in the St-T, case ( A E = 8,500 cm - I ) of benzene with 30 normal modes of vibration.
Figure 5.6. Nonradiative conversion in polyatomic molecules. Due to the difference k,, k,,,,, and the tranin state density of the initial state Vli and the final state q,. sition is practically irreversible.
The influence of the energy gap between the states involved on the rate constant of radiationless transitions may be clarified with the aid of Equations (5.15) and (5.16) as follows: According to Figure 5.6, the density of states Q, in the final state increases with an increasing energy gap. The electronic excitation energy, which has to be converted into vibrational motion of the nuclei, increases at the same time. The larger the difference in vibrational quantum numbers, the smaller the overlap of the vibrational wave functions. Thus, the expected increase of the rate constant k of a radiationless transition with increasing energy gap A E due to the density of states is overcompensated by a decrease due to increasingly unfavorable FranckCondon factors. Theoretical arguments lead to an exponential dependence of the Franck-Condon factors on the energy gap (Siebrand, 1966; Englman and Jortner, 1970). According to Equation (5.15)- the rate constant for a transition from the higher to the lower state is larger than that of the reverse transition. This is because for a given energy the density of states Q, of the lower state is larger than the density of states of the higher state. as can be seen from Figure 5.6. Also, the rate-determining transition between approximately isoenergetic levels of close-lying states is followed by a very fast dissipation of the vibrational energy. This prevents the back reaction and makes radiationless transitions into lower states virtually irreversible. (Cf. Figure 5.6.) Example 5.4: Figure 5.7 gives a schematic representation of the potential energy curves for two states of a diatomic molecule. Depending on the relative positioning of these curves different probabilities for radiationless transitions result because the Franck-Condon factors can differ appreciably. If the energy difference between the two states is large. as is generally the case for Soand S,, the zerovibrational level (v' = 0) of S, overlaps with a higher vibrational level (e.g., v = 12) of Soin a region close to the equilibrium geometry, where the kinetic energy of the nuclear motion is large and the probability x:, is small. The overlap integral <X,J~,,.> is therefore close to zero (Figure 5.7a) and the FranckCondon factor makes the transition very unlikely. If the states are energetically closer to each other, such as S2and S,, or if the potential energy curves cross,
Figure 5.7. Franck-Condon factors for radiationless transitions between different potential energy curves of a diatomic molecule; a) for a large and b) for a small energy gap, such as those observed, for instance, between S, and Soor between S, and S,, respectively, and c) for the case that the potential energy curves (e.g., S, and TI)cross. as may be the case for S, and T,, the overlap <x,Jx,,> between the zero-vibrational level of the higher state and the isoenergetic vibrational level of the lower state will be much larger (Figure 5.7b and c). Radiationless transitions between these states are therefore much more likely. According to Figure 5.7, the rate of radiationless transitions depends not only on the energy gap, but also on the equilibrium geometries of the states involved. Thus, rigid n systems such as condensed aromatic hydrocarbons possess only a relatively small energy gap between S, and So,but the bonding characteristics in the ground state and in the first excited (a,*) state differ so little that the potential energy surfaces of these states are nearly parallel. Since the amplitudes of the vibrational wave functions X, of the higher excited vibrational levels of S, oscillate very quickly around zero near the equilibrium geometry, integration results in very small Franck-Condon factors. Therefore. in these systems fluorescence can compete with radiationless deactivation.
Example 5.5: In radiationless transitions from the triplet state to the ground state of aromatic hydrocarbons, the excess electronic energy goes predominantly into the CH stretching vibrations. Being of high frequency (C .= 3,000 cm-I), they are widely spaced and much smaller vibrational quantum numbers are required than for other normal modes of vibration whose wave numbers are at most half
this size. Indeed, the triplet lifetime of naphthalene increases from t = 2 s to t = 20 s on perdeuteration. The energy gap AE(T, - S,,)is virtually the same for the deuterated and the undeuterated compound, but higher numbers of vibrational quanta are necessary to overcome this gap because the frequencies of the CD vibrations are smaller by roughly 30%. Hence, the Franck-Condon factors are much less favorable for the deuterated than for the protiated compounds (Laposa et al., 1%5).
Emission of a photon from an electronically excited state is referred to as luminescence. Fluorescence and phosphorescence can be differentiated depending on whether the transition is between states of equal or different multiplicity and hence spin-allowed or spin-forbidden. (Cf. Section 5.1.1 .) Thus, for molecules with singlet ground states fluorescence constitutes a pathway for deactivating excited singlet states whereas phosphorescence is observed in the deactivation of triplet states.
5.3.1 Fluorescence of Organic Molecules According to the results presented in the last section, a fraction of the energy that a malecule acquires through absorption of a light quantum is dissipated in condensed phases very rapidly via radiationless deactivation and thermal equilibration. In general, the rate of energy loss by emission is comparable with the rate of radiationless deactivation only for the lowest excited singlet state S, or the lowest triplet state T,. As a rule, therefore, the energy of the emitted radiation is lower than that of the absorbed radiation by the amount of energy that has been dissipated nonradiatively. The resulting emission is of longer wavelengths than the absorbed light. As a further consequence of thermal equilibration the intensity distribution in fluorescence and phosphorescence spectra is independent of the exciting wavelength. The shapes of absorption and emission bands are determined in the same way by Franck-Condon factors. The shift of the emission maximum with respect to the absorption maximum, which is referred to as Stokes' shifr, increases with the increasing difference between the equilibrium geometries of the ground and the excited states. This is schematically shown for a diatomic molecule in Figure 5.8; the maximum intensity of absorption will be observed for the vertical transition v = 0 + v' = n, whereas emission will occur after vibrational relaxation, with the highest probability for the transition form v' = 0 to a different ground-state vibrational level, v = m.
Often the 0-0 transitions of absorption and emission do not coincide, and a " 0 4 gap" results. This is referred to as anomalous Stokes shifr and is due
Stokes shift; a) definition and b) dependence on the difference in equilibrium geometries ofground and excited states. Shown is the probability distribution in various vibrational levels, which is proportional to the square of the vibrational wave function (adapted from Philips and Salisbury, 1976).
Figure 5.8.
to different intermolecular interactions in the ground and excited states. The energy of the excited-state molecule decreases during its lifetime through reorientation of the surrounding medium, so fluorescence is shifted to longer wavelengths. As an example the difference of the 0-0 transitions in the absorption and in the fluorescence of p-amino-p'-nitrobiphenylis shown in Figure 5.9 as a function of solvent polarity. In solid solutions at low temperatures the motion of solvent molecules can be slowed down to such an extent that no reorganization occurs and the anomalous Stokes shift disappears. Frequently, the fluorescence spectrum is the mirror image of the absorption spectrum, as exemplified for perylene in Figure 5.10. This spectral symmetry is due to the fact that the excited-state vibrational frequencies, responsible for the fine structure of the absorption band, and the ground-state vibrational frequencies. which show up in the fluorescence band, are often
Figure 5.9.
in absorption and fluorescence of p-amino-p'-nitrobiphenylin benzeneldioxane as a
function of the dioxane content (by permission from Lippert, 1%6).
quite similar. According to Figure 5.8 the intensity distribution given by the Franck-Condon factors for absorption and emission are also comparable. This is particularly true in those cases where electronic structures and equilibrium geometries of the ground and excited states differ very little, as for instance in n-n* transitions of delocalized n systems. In biphenyl, however, which is less twisted around the central single bond in the excited than in the ground state (cf. Section 1.4. I ) , the absorption and fluorescence spectra differ appreciably, and vibrational structure is observed only in emission. Another example of the mirror-image relation between absorption and fluorescence spectra is provided by anthracene (Figure 5.1 I). In this particular case the situation is complicated by vibronic coupling; the 'Laand 'L, bands overlap in the absorption spectrum, and the location of the 'L, origin has been inferred from the substituent effects evident from the MCD spectrum (Steiner and Michl, 1978). The relationship between the quantum yield of fluorescence and molecular structure is determined to a large extent by the structural dependence of the competing photophysical and photochemical processes. Thus, for most rigid aromatic compounds fluorescence is easy to observe, with quantum yields in the range 1 > @, > 0.01. This may be explained by the fact that the Franck-Condon factors for radiationless processes are very small because changes in equilibrium geometry on excitation are small; thus internal conversion becomes sufficiently slow. (Cf. Calzaferri et al., 1976.) A fundamental factor that determines the fluorescence quantum yield is the nature of the lowest excited singlet state-that is to say, the magnitude of the transition moment between S,,and S,. If the SO+Sl transition is symmetry forbidden, as in benzene, k,: is small compared to x k i for the comI
Absorption and fluorescence spectrum of perylene in benzene (by permission from Lakowicz, 1983).
Figure 5.10.
Figure 5.11.
Turro, 1978).
peting processes and only minor quantum yields of fluorescence are observed. If due to substitution the intensity of the S,,+S, transition becomes larger, as in aniline, the fluorescence yield increases. Most compounds whose lowest excited singlet state is an (n,n*) state exhibit only very weak fluorescence. The reason is that due to spin-orbit coupling, intersystem crossing into an energetically lower triplet state is particularly efficient. (Cf. Section 5.2.2.) Heavy atoms in the molecule (e.g., bromonaphthalene) or in the solvent (e.g., methyl iodide) may favor intersystem crossing among (n,n*) states to such an extent that fluorescence can be more or less completely suppressed. At low temperatures, photochemical deactivation and energy-transfer processes involving diffusion and collisions become less important, and low-frequency torsional vibrations that are particularly efficient for radiationless deactivation are suppressed, so fluorescence quantum yields increase. For instance, for trans-stilbene (7), @, = 0.05 at room temperature, but @, = 0.75 at 77 K. If the stilbene chromophore is fixed in a rigid structural frame as in 8, @, equals 1 .O independent of temperature (Sharafy and Muszkat, 1971; Saltiel et al., 1968).
Figure 5.12. Qualitative state diagram for the fluorescence quenching of benzene by radiationless transition into one of the higher vibrational levels of the isomeric
benzvalene. The back reaction is a hot ground-state reaction.
1990 for leading references) thus involves no special mechanism of nonradiative decay.
The quantum yield of fluorescence from unsaturated compounds is independent of the exciting wavelength, unless photochemical reactions originating from higher excited singlet states or intersystem crossing compete with internal conversion. A good example is benzene in the gas phase at low pressure (less than 1 torr): on excitation of the S, state at I = 254 nm fluorescence occurs with a quantum yield of @, = 0.4. This decreases if higher vibrational levels of this state are excited, and at A < 240 nm no emission can be detected at all. It is assumed that a radiationless transition is possible from the higher vibrational levels of the S, state of benzene into very highly excited vibrational levels of the ground state of the isomeric benzvalene (9), as shown in the schematic representation in Figure 5.12. Benzvalene can either be stabilized by vibrational relaxation or can undergo a hot groundstate reaction and return to the benzene ground state (Kaplan and Wilzbach, 1%8). Calculations shdw that a funnel in S, (conical intersection of S, and So),separated from the vertical geometry by a small barrier to isomerization, provides a mechanism for an ultrafast return to So, as soon as the molecule has sufficient vibrational energy to overcome the barrier (Palmer et al., 1993; Sobolewski et al., 1993). The so-called "channel 3" effect (see Riedle et al.,
By measuring the fluorescence intensity of sufficiently dilute solutions as a function of the exciting wavelengths afluorescence excitation specrrum is obtained. Such a measurement represents a remarkably sensitive method for investigating the absorption spectrum. With I, and IFbeing the intensities of the absorbed light and of the light emitted as fluorescence, respectively,
where I, is the intensity of the exciting light, E the extinction coefficient, c. the concentration of the sample, and d the optical path length. (Cf. Section 1.1.2.) If the absorbance A = ~ c isd smaller than say 0.2-0.3, a series expansion of the exponential function ec" = 1 - x + . . . , with neglect of the higher powers of x, yields
and as a result, for a given @, and I,, the fluorescence intensity reflects the wavelength or wave-number dependence of the extinction coefficient E.
5.3.2 Phosphorescence Because internal conversion and vibrational relaxation are very fast, phosphorescence corresponds to a transition from the thermally equilibrated lowest triplet state TI into the ground state So and the phosphorescence spectrum is approximately a mirror image of the So+T, absorption spectrum, which is spin forbidden and therefore difficultto observe because of the low intensity. This mirror-image symmetry is evident from the singlet-triplet absorption and phosphorescence spectra of anthracene shown in Figure 5.1 1. In general the TI state is energetically below the S, state, and phosphorescence occurs at longer wavelengths than fluorescence, as shown in Figure 5.11. Since the transition moment of the spin-forbidden T,-*So transition is very small, the natural lifetime t,Pof the triplet state is long. Consequently, radiationless processes can compete with phosphorescence in deactivating the T, state. Of particular importance are collision-induced bimolecular processes (cf. Section 5.4), and phosphorescence of gases and liquid solutions is relatively difficult to observe (Sandros and Backstrom, 1962). An exception is biacetyl with a very short TI lifetime t:. (Cf. Figure 5.3.) Phosphorescence spectra are commonly measured using samples in solvents or mixed solvents that form rigid glasses at 77 K (such as EPA = ether-pentane-alcohol mixture). (Cf., however, Example 5.6). The natural lifetime of the triplet state t,P = Ilk, may be estimated from the observed lifetime and the quantum yields of fluorescence and phosphorescence. According to Equation (5.1 1)
Figure 5.13. Various possiblities for the disposition of the lowest singlet and triplet (n,lc*) and (n,n*) states of organic molecules (adapted from Wilkinson, 1968).
If internal conversion and all energy-transfer processes and photochemical reactions are negligible, q, = 1 - QF, and one obtains 1 - aF -
s and many seconds. The natural lifetime $ varies between The rate constant ksT of intersystem crossing depends on the energy gap AEsT between the singlet and triplet states and in particular on spin-orbit coupling. The difference in the magnitude of spin-orbit coupling contributes greatly to the fact that the quantum yield as,of triplet formation is small for aromatic hydrocarbons, but nearly unity for carbonyl compounds. (Cf. the El-Sayed rules, Section 5.2.2.) In discussing the energy gap dependence of the intersystem crossing rate one has to note that the triplet state closest to
S, can be one of the higher triplet states T,, such as is the case for anthracene (cf. Example 5.3), and that in molecules with lone pairs of electrons either TI or T, may be of the same type as S,. This is illustrated in Figure 5.13, where various possibilities for the energy order of the (n,n*) and (n,n*) states are displayed schematically. Not all the situations shown are equally probable; it is, for instance, very unlikely that the singlet-triplet splitting of the (n,fl) states could be appreciably larger than that of the (n,llL)states, as shown in Figure 5.13f.
The relative position of the excited states of benzophenone (10) is shown in Figure 5.13a. Intersystem crossing leads from a '(n,*) state to a '(n,n*) state; this type of transition is favored by spin-orbit coupling to such an extent that as, = I and no fluorescence is observed. If, however, S, and TI are (n,n*) states and are disposed as in Figure 5.13b, a, is so small that practically no phosphorescence can be observed although the molecule has rr-** transitions. The relative disposition of (n,llr) and (n,n*) states may be changed by solvent effects and the rate constants of the various photophysical and photochemical processes can be drastically altered. Thus lone pairs of electrons in molecules such as quinoline (11) may be stabilized by hydroxylic solvents to such an extent that the (n,lt*) states become higher in energy than the (n,*) states, as shown in Figure 5.13d and e. Phosphorescence would predominate in the former case, fluorescence in the latter. As a result of the heavy-atom effect or the effect of paramagnetic molecules such as 02,which both enhance S,l+T,, absorption (cf. Section 1.3.2). phosphorescence TI+Stl as well as the rate constants k,., and k.,., of intersystem crossing will be Favored. The consequence according to Equation (5.1 I ) is an increase in qlsc, whereas r], will increase or decrease depending on which of the two processes, radiationless deactivation of the triplet state or phosphorescence, is more strongly favored. Frequently, an increase in the
quantum yield a, of phosphorescence by the heavy-atom effect is observed. For example, free-base porphine (12) exhibits practically only fluorescence in a neon matrix and practically only phosphorescence in a xenon matrix (Figure 5.14). It has been concluded that the phosphorescence rate constant is enhanced by at least two orders of magnitude by the heavy-atom effect (Radziszewski et al., 1991).
Example 5.6: The heavy-atom effect can be utilized for measuring phosphorescence in solution. In Figure 5.15 the luminescence spectrum of 1.4-dibromonaphthalene
Figure 5.14. Luminescence of free-base porphine (12) in N e (4 K) and X e (12 K). The former is totally dominated by fluorescence (left) and the latter by phosphorescence (right) (by permission from Radziszewski et al., 1991).
Figure 5.15.
Luminescence spectrum of 1.4-dibromonaphthalene in acetonitrile with and without purging with NZ(by permission from Turro et al., 1978).
fluorescence excitation spectra and has been discussed in detail in Section 5.3.1. The use of a heavy-atom solvent such as ethyl iodide and of an intense light source permits the direct measurement of S,+T, transitions by this method. This technique may be used even when the compound itself is nonphosphorescent by having present a phosphorescent molecule of lower triplet energy that is excited by energy transfer (see Section 5.4.5) from the nonphosphorescent triplets. Example 5.7: The phosphorescence excitation spectrum may be used to decide whether the lowest triplet state is a '(n,n*) or a '(n,n*) state, if measurements are carried out in two different solvents with or without the heavy-atom effect, respectively. The intensity of the phosphorescence excitation spectrum will in general increase due to the heavy-atom effect if T, is a '(n,n*) state as in p-hydroxyacetophenone. However, no intensity increase is observed if T, is a '(n,fl) state as in benzophenone, since spin-orbit coupling associated with a n+n* transition is already so strong that the additional solvent effect is negligible. The spectrum in Figure 5.17 illustrates the intensity increase expected for a '(n,n*) state.
Figure 5.16. Luminescence spectra of naphthalene and triphenylene in 1,2dibromoethane (by permission from Turro et al.. 1978).
is shown as an example of the "internal heavy-atom effect." The spectra of naphthalene and triphenylene in Figure 5.16 demonstrate the "external heavyatom effect" due to the solvent I .2-dibromoethane. For such measurements it is important to use highly purified and deoxygenated solvents since otherwise phosphorescence is too weak to be detected, as is evident from Figure 5.15 (Turro et al., 1978).
When a, is independent of the excitation wavelength the extreme sensitivity associated with emission spectroscopy can be utilized to obtain So-T absorption spectra by measuring phosphorescence excitation spectra (Marchetti and Kearns, 1967). The principle of the method is the same as for
Figure 5.17. Phosphorescence excitation spectrum of p-hydroxybenzophenone in an ether-toluol-ethanol mixture with (-) and without (---) the addition of ethyl iodide. The curves are normalized in such a way that the S T excitation of the '(n,lr*) state shows about the same intensity in both solvents (by permission from Kearns and Case, 1%6).
5.3.3 Luminescence Polarization A light quantum of appropriate energy can be absorbed by a molecule fixed in space only if the light electric field vector has a component parallel to the molecular transition moment. If the directions of the transition moment and of the electric field vector form an angle 9,the absorption probability is proportional to cos29.(Cf. Section 1.3.5.) The light quanta of luminescence are also polarized, with the intensity again proportional to cosZq. The polarization direction of an electronic transition may be determined by measurement of the absorption of polarized light by aligned molecules. (Cf. Michl and Thulstrup, 1986.) Orientation may be achieved in a number of ways. When single crystals are used or when the molecules of interest are incorporated into appropriate single crystals, a very high degree of orientation can be obtained if the crystal structure is favorable. Other methods accomplish orientation by embedding the molecules in stretched polymer films (polyethylene, PVA; Thulstrup et al., 1970)or in liquid crystals; further possibilities are orientation by an electric field (Liptay, 1963), or, in the case of polymers such as DNA, by a flow field (Erikson et al., 1985). The relative polarization directions of the different electronic transitions of a molecule may be determined by exciting with polarized light and analyzing the degree of polarization of the luminescence, referred to briefly as luminescence polurizution. When a solution of unoriented molecules is exposed to plane-polarized light of a wavelength appropriate for a specific electronic transition, only those molecules that have their transition moment oriented parallel to the electric field vector absorb with maximum probability. Using this selection process, known as photoselection, an effective alignment of the excited molecules is achieved. Only excited molecules can emit, and the directions of the transition moment of emission [in general M(S,-*S,) or M(T,-*S,,)] and the transition moment of absorption M(S,+S,,) will form an angle a. If rotation of the excited molecules during their lifetimes is prevented by high viscosity of the solution, the emission will also be polarized. The degrt~ec?f'polurizc~tion is defined as
while the degree of anisotropy, which sometimes leads to simpler formulas (cf. Michl and Thulstrup, 1986), is defined as
Here Ill and I , are the intensities of the components of the emitted light parallel and perpendicular to the electric vector of the exciting light, respectively. The curves P(A) or P(5) are called the polarization spectrum. Depending on whether the measurement is carried out with constant excitation
wavelength A, or at constant luminescence wavelength A2, different polarization spectra result. Using excitation light of fixed wavelength A, a fluorescence polarization spectrum (FP) or a phosphorescence polarization spectrum (PP), respectively, is obtained. Measurement at a constant emission wavelength yields an absorption-wavelength-dependent polarization spectrum of either fluorescence [AP(F)I or phosphorescence [AP(P)]. The relationship between the degree of polarization P and the angle a between the transition moments of absorption and emission is given by
Ideally, P can assume values ranging between P = 0.5 (corresponding to a = 0") and P = -0.33 (corresponding to a = 90'). If a molecule possesses a symmetry axis of order n > 2 and absorption and emission are isotropically polarized in a plane perpendicular to this axis, P = const = 0.14 (Dorr and Held, 1960). The value of P is affected by the overlap of different electronic transitions and also reflects the orientational dependence of the transition moment on vibronic mixing with nonsymmetrical vibrations. Under favorable conditions it is therefore possible not only to determine the relative directions of electronic transition moments from the polarization spectra but also to locate bands hidden in the absorption spectrum and to ascertain the symmetry of the vibrations giving rise to the fine structure. Energy migration (repeated intermolecular energy transfer) causes depolarization; such measurements have therefore to be carried out in dilute solutions.
In ~ i i u r 5.18 e the absorption and emission spectra of azulene are shown. The anomalous fluorescence of azulene from the S, state is easy to recognize. The AP(F) spectrum exhibits a deep minimum at 33,900 cm-I. The small peak in the absorption spectrum at the same wave number is therefore not duk to vibrational structure but rather to another electronic transition, the polarization of which had been predicted by PPP calculations. Figure 5.19 shows all four types of polarization spectra of phenanthrene. FP becomes negative at the vibrational maxima of the fluorescence; the most intense vibration is not totally symmetric, in contrast to the one which shows up weakly. For all absorption bands, AP(P) = -0.3. The polarization direction of phosphorescence is perpendicular to the transition moments of all n-n* transitions lying in the molecular plane and is therefore perpendicular to the molecular plane. One observes PP = -0.3 as well, with a modulation due to vibrations. The observed phosphorescence polarization direction may be accounted for by the fact that singlet-triplet transitions acquire their intensity by spin-orbit coupling of the Sostate with triplet states and particularly, of the T, state with singlet states. (Cf. Section 1.3.2.) Under usual conditions the phosphorescence is an unresolved superposition of emissions from the three components of the triplet state, which in the absence of an external magnetic field are described
Figure 5.18. Absorption (A) and fluorescence spectrum (F) of azulene in ethanol at 93 K. FP and AP(F) denote the polarization spectrum of fluorescence and the excitation-polarization spectrum of fluorescence, respectively (by permission from DOrr, I%@.
by the spin functions 8,, a,, and 63, [cf. Equation (4.7)] and have the same symmetry properties as the rotations A,,,fi,, and fi: about the molecular symmetry axes. In general they belong to different irreducible representations of the molecular point group and therefore mix with different singlet states under the influence of the totally symmetric spin-orbit coupling operator. The transition moments of the different admixed singlet states then determine the intensity and polarization direction of emission from the various components of the TI state. In the case of phenanthrene. the triplet component with the transition moment perpendicular to the molecular plane contributes most of the states with pointensity. since spin-orbit coupling of this component to '(u,lr*) larization direction perpendicular to the plane predominates over the spin-orbit coupling of the other components. The TI-S, transition steals its intensity from such S,+S, transitions. In Figure 5.20, the emission spectrum of triphenylene is shown: P = 0.14 in the FP spectrum due to the threefold symmetry axis perpendicular to the molecular plane, which is also the polarization plane for all n-n* transitions.
Figure 5.19. Absorption (A) and emission spectra ( F and P) of phenanthrene in ethanol at 93 K. FP and PP denote the polarization spectra of fluorescence and phosphorescence, AP(F) and AP(P) the excitation polarization spectra of fluorescence and phosphorescence, respectively (by permission from M r r , 1966).
The phosphorescence polarization direction is again perpendicular to the molecular plane and is modified by out-of-plane vibrations.
If the excited molecules merely return to their ground state by radiative or nonradiative processes, no permanent orientation remains after the photoselective irradiation is terminated. However, when the excited molecules undergo a permanent chemical change, photoselection leads to a lasting alignment of that portion of the reactant molecules that remain when the irradiation is interrupted. Sometimes, the product molecules are aligned as well, depending on the degree of correlation between the average orientation of the parent reactant and the daughter photoproduct molecules in space (Michl and Thulstrup, 1986). Samples oriented by photoselection have been used for studies of molecular anisotropy by polarized absorption spectroscopy. For instance, the
that involve the transfer of excitation energy from one molecule to another. These processes are generally referred to as quenching processes. The suppression of emission by energy-transfer processes is in particular referred t o a s luminescence quenching (quenching in the strict sense). If it is not the deactivation that is of principal interest during a bimolecular process but rather the excitation of the energy acceptor molecule, the process is referred to a s sensitizution. States that otherwise would be accessible only with difficulty o r even not at all may be populated through sensitized excitation.
5.4.1 Quenching of Excited States Fluorescence quenching is a very general phenomenon that occurs through a variety of different mechanisms. All chemical reactions involving molecules in excited states can be viewed as luminescence quenching. Such photochemical reactions will be dealt with in later chapters. Photophysical quenching processes that d o not lead t o new chemical species can in general be represented as
Figure 5.20. Emission spectra (F and P) of triphenylene in ethanol at 93 K. FP and PP denote the polarization spectra of fluorescence and phosphorescence, respectively (by permission from D6rr, 1966).
symmetries of all IR-active vibrations of free-base porphine (12) have been measured on a sample photooriented with visible light in a rare-gas matrix (Radziszewski e t al., 1987, 1989). Photoorientation of the major and the minor conformer of 1,3-butadiene in rare-gas matrices was used t o determine the directions of the IR-transition moments in both, which revealed that the latter is planar in these media (i.e., s-cis and not gauche), and yielded the average relation between the orientation of the s-cis reactant and that of the s-trans photoproduct (Arnold et at., 1990, 1991).
where M' is the ground state o r another excited state of M. According t o whether the quencher Q is a molecule M of the same kind o r a different molecule self-quenching o r concentration quenching can be distinguished from impurity quenching by some other chemical species. Most intermolecular deactivation processes are based on collisions between an excited molecule M* and a quencher Q. They are subject to the Wigner-Witmer spin-conservation rule according to which the total spin must not change during a reaction (Wigner and Witmer, 1928). Example 5.9: In order that the products C and D lie on the same potential energy surface as the reactants A and B, the total spin has to be conserved during the reaction. The spins S, and S, 01' the reactants may be coupled according to vector addition rules in such a way that the total spin of the transition state can have the following values: (SA
In addition t o monomolecular processes such as emission and radiationless deactivation there are very important bimolecular deactivation mechanisms
+ s, - 1).
Similarly, the spins S, and S,, of the products may be coupled to give one of the total spin values (Sc
+ S,).
+ Sn). (Sc + Sn -
A reaction is allowed according to the Wigner-Witmer spin-conservation rule if the reactants can form a transition state with a total spin that can also be obtained by coupling the product spins, that is. if the two sequences above have a number in common.
Thus, for S, = S, = 0 the reaction is allowed if S,. = S,, = 0, but not if S,. = I, S, = 0 since in the latter case S, and S, can be coupled only to the total spin (S,. + S,) = IS,. - S,I = 1. Therefore, the singlet-singlet energy transfer
is allowed. Similarly, the triplet-triplet energy transfer
is seen to be allowed. For the reaction of two molecules in their triplet states one has S, = S, = 1 and the total spin can take the values 2, 1, and 0. Triplettriplet annihilation Figure 5.21.
MO scheme of excimer and exciplex formation.
thus gives one molecule in a singlet state while the other one may be in a singlet, a triplet, or a quintet state.
Except for some long-distance electron-transfer and energy-transfer mechanisms bimolecular deactivation involves either an encounter complex (M* . . . Q), or an exciplex (MQ*) or excimer (MM)*. An exciplex or an excimer has a binding energy larger than the average kinetic energy (312)kT and represents a new chemical species with a more or less well-defined geometrical structure corresponding to a minimum in the excited-state potential energy surface. This is not true for an encounter complex in which the components are separated by widely varying distances and have more random relative orientations. Encounter complexes, exciplexes, and excimers can lose their excitation energy through either fluorescence or phosphorescence, by decay into M + Q* which corresponds to an energy transfer, by electron transfer to give Mm + Q" or M@ + Q@,by internal conversion, and by intersystem crossing (Schulten et al., 1976b). All these processes lead to quenching of the excited state M* and are therefore referred to as quenching processes.
occupied. This gives rise to the relative minimum of the excimer '(MM)*on the excited-state potential energy surface. (Monomer Fluorescence)
. .
.
In Section 4.4 it was shown for the (H2 + H?) system that interaction and q M results in exciton between locally excited states described by qMM.
5.4.2 Excimers Frequently, it is observed that an increase in the concentration of a fluorescent species such as pyrene is accompanied by a decrease in the quantum yield of its fluorescence. This phenomenon is called self-quenching or concentration quenching and is due to the formation of a special type of complex formed by the combination of a ground-state moIecule with an excited-state molecule. Such a complex is called an excimer (excited dimer) (Forster and Kaspar, 1955). Whereas for a system M + M of two separate ground-state molecules all interactions are purely repulsive except for those yielding the van-der-Waals minima observed in the gas phase, stabilizing interactions are possible according to Figure 5.21 if one of the molecules is in an excited state and the hom*o and LUMO of the combined system are only singly
Schematic representation of the potential energy surfaces for excimer formation and of the difference between monomer fluorescence and excimer fluorescence (adapted from Rehm and Weller, 1970a). Figure 5.22.
in Figure 5.22. From this diagram it is evident that excimer fluorescence is to be expected at longer wavelengths than monomer fluorescence and that the associated emission band should be broad and generally without vibrational structure, being due to a transition into the unbound ground state. The spectrum of pyrene in Figure 5.23 is the perfect confirmation of these expectations. Excimer formation is observed quite frequently with aromatic hydrocarbons. Excimer stability is particularly great for pyrene, where the enthalpy of dissociation is AH = 10 kcallmol (Fdrster and Seidl, 1965). The excimers of aromatic molecules adopt a sandwich structure, and at room temperature, the constituents can rotate relative to each other. The interplanar separation is 300-350 pm and is thus in the same range as the separation of 375 pm between the two benzene planes in 4,4'-paracyclophane (13), which exhibits the typical structureless excimer emission. For the higher hom*ologues, such as 5,S'-paracylophane, an ordinary fluorescence characteristic of p-dialkylbenzenes is observed (Vala et al., 1965).
Calculations based on a wave function corresponding to Equation (5.22) also indicate a sandwich structure and an interplanar distance of 300-360 pm (Murrell and Tanaka, 1964).
5.4.3 Exciplexes Figure 5.23. Absorption (..,) and fluorescence spectrum (-) of pyrene a) lo-' moll L in ethanol, b) 10-2 mol/L in ethanol, and c) absorption and emission (---) of crystalline pyrene (adapted from Farster and Kaspar, 1955).
states. The magnitude of the interaction is a measure of the energy transfer from one molecule to the other. (Cf. the Forster mechanisms of energy transfer, Section 5.4.5.) Interaction with ion-pair or CT states further stabilizes the lower of the exciton states, so the excimer can be described by a wave function of the form Whether the stabilization is given by a plus or minus combination depends on the orientation of M* relative to M. (For a more detailed treatment see Michl and BonaCiC-Koutecky, 1990.) The potential energy surfaces of the ground state M + M and the ercited state M* + M resulting from these arguments are represented schematically
%o different molecules M and Q can also form complexes with a definite stoichiometry (usually I: I). If complexation is present already in the ground state and leads to CT absorption (Section 2.6), which is absent in the individual components, the complex is referred to as a charge-transfer or donoracceptor complex. If, however, the complex shows appreciable stability only in the excited state, it is called an exciplex (excited complex): 'M' (Monomer Fluorescence)
The spectra of anthracene-dimethylaniline shown in Figure 5.24 exemplify the new structureless emission at longer wavelengths due to the formation of an exciplex. This fluorescence is very similar to excimer fluorescence. Contrary to the situiition for excimers, one component of the exciplex acts predominantly as the donor (D); the other one acts as the acceptor (A).
Fluorescence and exciplex emission from anthracene in toluene (3.4.10 ' mol/L) for various concentrations c., of dimethylaniline (by permission from Weller, 1968).
Figure 5.24.
If use is made of this fact in the notation. one obtains instead of Equation (5.22) the wave function
Figure 5.25. Exciplex formation by charge transfer a) from the donor to the excited acceptor and b) from the excited donor to the acceptor. Exciplex emission is indicated by a broken arrow (by permission from Weller, 1%8).
Exciplexes and excimers appear to be involved in many photochemical processes; in particular, they are probably involved in many quenching and charge-transfer processes and in many photocycloadditions.
(CT) excited state (cf. Section 4.4) and is by far the most important one, and the exciplex corresponds to a contact ion pair. Experimentally the chargetransfer character is revealed by the high polarity, with exciplexes from aromatic hydrocarbons and aromatic tertiary amines having dipole moments p(AD)* > 10 D (Beens et al., 1967; cf. Section 1.4.2). From a simple MO treatment it follows that the electron transfer leading to exciplex formation can occur either from an excited donor to an acceptor or from a donor to an excited acceptor. (See Figure 5.25.) In both cases, the singly occupied orbitals of the resulting exciplex correspond to the hom*o of the donor and the LUMO of the acceptor. Neglecting solvent effects, the energy of exciplex emission is therefore given by where IP, and EAAare the ionization potential of the donor and the electron affinity of the acceptor, and C describes the Coulomb attraction between the components of the ion-pair D@A@.
5.4.4 Electron-Transfer and Heavy-Atom Quenching According to the following scheme, electron transfer, intersystem crossing and energy transfer can all compete with fluorescence in deactivating an exciplex '(DA)* formed from the molecules A and D.
In polar solvents, polar exciplexes (contact ion pairs) dissociate into nonfluorescent radical ions (loose ion pairs or free ions) due to the stabilization of the separated ions by solvation. It has been observed that exciplex emission decreases with increasing polarity of the solvent and that at the same
time the free-radical ions DO and A@can be identified by flash spectroscopy (Mataga, 1984). Assuming the existence of a quasi-stationary state the rate constant of an exothermic electron-transfer reaction can be written as kdiR,k,, k-dirr,and k-, are the rate constants for diffusion, forward electron transfer from D to A, for the dissociation of the encounter complex to A and D, and for the back electron transfer from A o to DO, respectively. The terms with k - , have been neglected since k-, e k, can be assumed for exothermic reactions. According to the Marcus theory (1964) a relationship of the form exists for adiabatic outer-sphere electron-transfer reactions* between the free enthalpy of activation AGt and the free enthalpy of reaction AG. A in Equation (5.25) is the reorganization energy essentially due to changes in bond distances and solvation. As a consequence of this relation, the rate constant should first increase with increasing exothermicity until the value AG = -A is reached and then decrease again. The dependence of log kc on AG obtained in this way is shown in Figure 5.26 by the dashed curve; the region of decreasing rate for strongly exergonic electron-transfer reactions (AG < A) is referred to as the Marcus inverted region. The temperature dependence of electron-transfer rate constants is interesting. In the normal region, it shows an activation energy as predicted from simple Marcus theory. In the inverted region, the activation energy is very small or zero. This agrees with the quantum mechanical version of the theory (Kestner et al., 1974; Fischer and Van Duyne, 1977), which makes it clear that the transition from the upper to the lower surface behaves just like ordinary internal conversion. Studies of the fluorescence quenching in acetonitrile have shown that the electron-transfer reaction
Figure 5.26. Dependence of the rate constant log k for electron-transfer processes on the free reaction enthalpy AG, Rehm-Weller plot (-) and Marcus plot (---) for I = 10 kcal/mol (adapted from Eberson, 1982).
leading to a radical ion pair is diffusion controlled if the free enthalpy of this reaction is AG I- 10 kcal/mol, and that in contrast to the Marcus theory it remains diffusion-controlled even for very negative values of AG. Therefore, Rehm and Weller (1970b) proposed the empirical relationship AGt = AGl2
which adequately describes the experimentally observed data, as can be seen from the solid curve in Figure 5.26. The free enthalpy of electron transfer AG can be estimated according to Weller ( 1982a) from the equation where E;il(D) and E.sd(A)are the half-wave potentials of the donor and acceptor, AE,,,(A) is the (singlet or triplet) excitation energy of the acceptor, and AEcoUlis the Coulombic energy of the separated charges in the solvent in question. The authors suggested that the disagreement with the Marcus theory is due to fast electron transfer occurring via exciplex formation (Weller, 1982b), as shown in the following reaction scheme: Encounter complex (d = 700 pm)
* Redox processes between metal complexes are divided into outer-sphere processes and inner-sphere processes that involve a ligand common to both coordination spheres. The distinction is fundamentally between reactions in which electron transfer takes place from one primary bond system to another (outer-sphere mechanism) and those in which electron transfer takes place within a primary bond system (inner-sphere mechanism) (Taube, 1970).
Experimental values of k,E,:;indicate that dissociation of weakly solvated dipolar exciplexes into a strongly solvated radical ion pair requires charge separation against the Coulomb attraction as well as diffusion of solvent molecules during resolvation (Weller, 1982b). More recent investigations of rigidly fixed donor-acceptor pairs, for example, of type 14 with different acceptors A (Closs et al., 19861, and also of free donor-acceptor systems (Gould et al., 1988) have shown, however, that rates of strongly exothermic electron-transfer reactions in fact decrease again, as is to be expected for the Marcus inverted region. The deviation of the original Weller-Rehm data from Marcus theory at very large exothermicities may well be due to the formation of excited states of one of the products, for which the exothermicity is correspondingly smaller, and to compensating changes in the distance of intermolecular approach at which the electron transfer rate is optimized. Recently a number of covalently linked porphyrin-quinone systems such as 15 (Mataga et al., 1984) or 16 (Joran et al., 1984) have been synthesized in order to investigate the dependence of electron-transfer reactions on the separation and mutual orientation of donor and acceptor. These systems are also models of the electron transfer between chlorophyll a and a quinone molecule, which is the essential charge separation step in photosynthesis in green plants. (Cf. Section 7.6.1 .) Photoinduced electron transfer in supramolecular systems for artificial photosynthesis has recently been summarized (Wasielewski, 1992).
occurring via a highly excited triplet exciplex -'(MO,)** that undergoes internal conversion to the lowest triplet exciplex '(MO,)* and decays into the components:
The net reaction consists of a catalyzed intersystem crossing which is spin allowed as opposed to the simple intersystem crossing (Birks, 1970). Triplet states may also be quenched by oxygen, but triplet-triplet annihilation (cf. Section 5.4.5.5) seems to be the predominant mechanism.
5.4.5 Electronic Energy Transfer If an excited donor molecule D* reverts to its ground state with the simultaneous transfer of its electronic energy to an acceptor molecule A, the process is referred to as electronic. energy trunsfer: The acceptor can itself be an excited state, as in triplet-triplet annihilation. (Cf. Section 5.4.5.5.) The outcome of an energy-transfer process is the quenching of the emission or photochemical reaction associated with the donor D* and its replacement by the emission or photochemical reaction characteristic of A*. The processes resulting from A* generated in this manner are said to be sensitized. Energy transfer can occur either radiatively through absorption of the emitted radiation or by a nonradiative pathway. The nonradiative energy transfer can also occur via two different mechanisms-the Coulomb or the exchange mechanism.
5.4.5.1 Radiative Energy Transfer Radiative energy transfer is a two-step process and does not involve the direct interaction of donor and acceptor: Heavy-atom quenching occurs if the presence of' a heavy-atom-containI SC
ing species enhances the intersystem crossing '(MQ)* +'(MQ)* to such an extent that it becomes the most important deactivation process for the exciplex. Since the triplet exciplex is normally very weakly bound and dissociates into its components, what one actually observes in such systems is Luminescence quenching by oxygen appears to be a similar process with = ?02. The process is diffusion-controlled. and it may be thought of as
The efficiency of radiative energy transfer, frequently described as "trivial" because of its conceptual simplicity (Forster, 1959). depends on a high quantum efficiency of emission by the donor in a region of the spectrum where the light-absorbing power of the acceptor is also high. I t may be the dominant energy transfer mechanism in dilute solutions, because its probability decreases with the donor-acceptor separation only relatively slowly as compared with other energy-transfer mechanisms. When donor and acceptor are identical and emission and absorption spectra overlap sufficiently, radiative
trapping may occur through repeated absorption and emission that increases the observed luminescence lifetime.
tor absorption, respectively, are normalized to a unit area on the wave-number scale, that is
5.4.5.2 Nonradiative Energy Transfer The nonradiative energy transfer is a single-step process that requires that the transitions D*+D and A-A* be isoenergetic as well as coupled by a suitable donor-acceptor interaction. If excited-state vibrational relaxation is faster than energy transfer, and if energy transfer is a vertical process as implied by the Franck-Condon principle, the spectral overlap defined by
is proportional to the number of resonant transitions in the emission spectrum of the donor and the absorption spectrum of the acceptor. (Cf. Figure 5.27). The spectral distributions iD(s)and zA(fi)of donor emission and accepEmission
This clearly reflects the fact that J is not connected to the oscillator strengths of the transitions involved. The coupling of the transitions is given by the interaction integral
=
J VffirVldt
(5.30)
where fit involves the electrostatic interactions of all electrons and nuclei of and Yf = AVDVA. the donor with those of the acceptor, and Y, = AT,,*, are antisymmetrized product wave functions of the initial and the final state. The total interaction B may be written as the sum of a Coulomb and an exchange term; thus in the two-electron case Y, = +t[1(+ !v),A , 2() ~ D * ( 2 ) ~ A ( l ) land lfi = [VD(l)vA*(2)- vD(2)vA*( 1)I/@, and
*It,
6 = (6'
- 0") = L J V D * ( ~ ) V A ( ~ ) ~ ~ ' W D ( ~ ) ~ Y A * ( ~ ) ~ ~ I ~ ~ Z (5.31) - J%*( 1 ) ~ ~ ( 2 )~f D i' (~)VA )dt,dtzI *(I where the second integral. the exchange term. differs from the Coulomb term in that the variable I and 2 are interchanged on the right.* The functions t#, are spin orbitals and contain a space factor @, and a spin factor a or B. (Cf. Section 1.2.2.) The Coulomb term represents the classical interaction of the charge dis1) = lei $,, ( I I@,,( I ) and Q,(2) = 1c.l @A(2)@A (2) and may be tributions Q,( expanded into multipole terms: dipole-dipole, dipole-quadrupole. etc. With MDand MAdenoting the transition moments of the two molecules, and at not too small distances RA,,between the donor and the acceptor, the dipole term, which dominates for allowed transitions, may be written as p(dipole-dipole) - MdM,lRiD
(5.32)
and is thus related to experimentally measurable quantities (Fdrster, 1951). The exchange interaction, which is responsible for example for the singlet-triplet splitting, is a purely quantum mechanical phenomenon and does not depend on the oscillator strengths of the transitions involved. The exchange integral
- -
Figure 5.27. Schematic representation of the spectral overlap J and its relation t o the emission and absorption spectrum.
vA.
I t should be noted that here and yt,,. denote wave functions o f the excited states o f A and D. respectively, and not the complex conjugate of yr, and yf,,.
5.4
representing the interaction of the charge densities QYI) = lei @Iy(l)@A"(l) and 9 3 2 ) = [el @,,(2)@,(2) vanishes if the spin orbitals v, and 1/,, or VA-:and Vl, respectively, contain different spin functions. Since the charge densities Q: and Q: depend on the spatial overlap of the orbitals of D and A. the exchange interaction decreases exponentially with increasing internuclear distances. similarly as the overlap. For forbidden donor and acceptor transitions the Coulomb term vanishes and the exchange term will predominate. If both transitions are allowed and the distance is not too small, the dipoledipole interactions will prevail. The higher multipole terms are important only at very short distances, where, however, the essential contributions arise again from the exchange interaction, unless it vanishes due to the spin symmetry. Time-dependent perturbation theory yields for the rate constant of nonradiative energy transfer
.
(Cf. the Fermi golden rule, Section 5.2.3.) The density of states eE(number of states per unit energy interval) is related to the spectral overlap J, and using the relations for (3 given above the expressions derived by Forster (1951) and Dexter (1953) for the rate constant of energy transfer by the Coulomb and the exchange mechanism, respectively, may be written as
BIMOLECULAR DEACTIVATION PROCESSES
From Equation (5.3 1) it is seen that spin integration yields a nonvanishing Coulomb interaction only if there is no change in spin in either component. Thus ID*
+ 'A
+-
'D
+ 'A*
and are fully allowed. However, triplet-triplet energy transfer )D*
+ 'A
+-
'D
+ 3A*
is forbidden. Nevertheless, it is sometimes observed, because the spin-selection rule, which strongly reduces the magnitude of kET,also prolongs the lifetime of 3D* to such an extent that the probability of energy transfer can still be high compared with the probability of deactivation of 3D*(Wilkinson, 1964).
5.4.5.4 The Erchange Mechanism o f Nonradiatiue
Enetgy Transfer According to Equation (5.39, energy transfer by the exchange mechanism is a short-range phenomenon since the exchange term decreases exponentially with the donor-acceptor separation RAD.Since it requires the intervention of an encounter complex (D* - A) it is also called the overlap or collision mechanism. The Wigner-Witmer spin-selection rules (Section 5.4.1) apply, and the spin-allowed processes e
and
ID*
f, and f, are the oscillator strengths of the donor and acceptor transition, respectively, L is a constant related to an effective average orbital radius of the electronic donor and acceptor states involved, and J is the spectral overlap.
5.4.5.3 The Coulomb Mechanism o f Nonradiative
Enetgy Transfer Energy transfer according to the Coulomb mechanism, which is also referred to as the Forster mechanism, is based on classical dipoledipole interactions. From Equation (5.32) the interaction energy is seen to be proportional to Rid, with RADbeing the donor-acceptor separation, and is significant at distances up to the order of 10 nm, which is large but less than the range of radiative energy transfer. Introduction of reasonable numerical values into the Forster equation for the rate constant [Equation (5.34)] leads to the expectation that k,, can be much larger than the diffusion rate constant kdiw
291
+ 'A
-*
ID
e
+ 'A*
and are referred to as singlet-singlet and triplet-triplet energy transfer, respectively. The singlet-singlet energy transfer is also allowed under the Coulomb mechanism, which, as a long-range process, in general predominates. Collisional singlet-singlet energy transfer is therefore likely to be rare and observable only under special conditions-for example, with biacetyl as a quencher, since this shows only a very weak absorption in the UVIVIS region (Dubois and Van Hemert, 1%4), and when it is intramolecular (Hassoon et al., 1984). Triplet-triplet energy transfer, on the other hand, is a very important type of energy transfer observed with solutions of sufficient concentration. It has been established as occurring over distances of 1-1.5 nm, comparable with coilisional diameters. Steric effects were shown to be significant; for in-
stance, the introduction of gem-dimethyl groups into the diene used as a quencher reduces the rate constant of fluorescence quenching of diazabicyclooctene (17) by a factor of 3 4 . Apparently, it is important which regions of the molecules touch in the encounter complex (Day and Wright, 1969).
Intramolecular triplet-triplet energy transfer in compounds of the type D-Sp-A, where Sp is an appropriate spacer, such as trans-decalin or cyclohexane, has also been studied and a relation between charge transfer and energy transfer has been shown to exist. As depicted in Figure 5.28, electron transfer can be symbolized as electron exchange between the LUMOs of donor and acceptor, hole transfer as electron exchange between the hom*os, and triplet-triplet energy transfer as a double electron exchange involving both the hom*os and the LUMOs. This simplified view suggests that the probability of triplet-triplet transfer should be proportional to the product of the probabilities of hole transfer and electron transfer. Indeed, a remarkably good proportionality between rate constants k,, of triplet-triplet energy transfer and the product k,k, of rate constants of electron transfer and hole transfer has been observed experimentally (Closs et al., 1989). This proportionality has also been found in ab initio calculations indicating that the observed distance dependence of electron transfer, hole transfer, and triplet-triplet energy transfer is determined not only by the number of intervening CC bonds but also by an angular dependence of the through-bond coupling (Koga et al., 1993). Triplet-triplet energy transfer is most important in photochemical reactions. It is utilized to specifically excite the triplet state of the reactant. This process is referred to as photosensitization and the donor 'D* is called a triplet sensitizer jSens*. For efficient triplet sensitization the sensitizer must absorb substantially in the region of interest, its intersystem crossing effi-
,
so - - - - - - - So So-- - - .So - -.-- --
----
Sens
M
Se ns
M
.- - -
so-- - - so
---
Sens
M
Figure 5.29. Three possible situations for luminescence quenching: a) only triplet excitation can be transferred from the sensitizer Sens to the molecule M (triplet quenching), b) singlet as well as triplet transfer is possible, and c) M can capture singlet excitation from Sens and transfer triplet excitation back to Sens.
Figur* 5.30. Triplet-triplet energy transfer from biacetyl in benzene to various acceptors. Rate constant k, as a function of the triplet energy ET of the acceptor (by permission from Lamola, 1968). Figure 5.28. Frontier orbital representation of electron exchange in a) electron transfer, b) hole transfer, and c) triplet-triplet energy transfer (adapted from Closs et a]., 1989).
1
I
,
294
1'1 I( ) l UI'HYSICAL PROCESSES
5
ciency must be high, and its triplet energy must be higher than that of the acceptor. The desired relationship of the energy levels of the sensitizer Sens and the acceptor molecule M is depicted in Figure 5.29a. If the energy levels are disposed as shown in Figure 5.29b, singlet-singlet and triplet-triplet energy transfer are both possible. The situation shown in Figure 5.29c, however, would allow singlet-singlet energy transfer from the sensitizer to the molecule M and subsequent triplet-triplet energy transfer back to the sensitizer. When the triplet energy E., of the acceptor is about 3.5 kcallmol or more below that of the donor, triplet-triplet energy transfer is generally diffusion controlled, that is, kET = kdiWWhen both triplet energies are the same, then only the 0-0 bands overlap and kETis smaller by a factor of lo2;it decreases even further with increasing E,. This is evident from the results of Lamola (1968) shown in Figure 5.30, where the rate constant kET of the triplet-triplet energy transfer of biacetyl is plotted against the triplet energy ETof various acceptors. Example 5.10: Terenin and Ermolaev (1956) were the first to observe triplet-triplet energy transfer by measuring the benzophenone-sensitized phosphorescence of naphthalene in a rigid glassy solution at 77 K. The relative energies of the lowest singlet and triplet states of these molecules are shown in Figure 5.31. Excitation of the I(n,n*) state and subsequent intersystem crossing produce the TI state of benzophenone, which may transfer the excitation energy to the T I state
5.4
BIMOLECULAR DEACTIVATION PROCESSES
295
of naphthalene. The use of a filter prevented direct population of the TI state of naphthalene via excitation of its S, state followed by intersystem crossing.
5.4.5.5 Triplet-Triplet Annihilation Energy transfer can also occur between two molecules in their excited states; this phenomenon is most common for two triplet states due to their relatively long lifetimes. According to the Wigner-Witmer spin-selection rules the following triplet-triplet annihilation processes are allowed by the exchange mechanism:
where the double asterisk is used to denote higher excited states. The first process is most important and for most organic molecules the combined triplet energy is sufficient to excite one of them into an excited singlet state. The donor and the acceptor molecules are identical in many cases, and, as shown for the (Hz + H,) system in Section 4.4, the acceptor molecule will undergo internal conversion and finally reach the lowest excited singlet state, so triplet-triplet annihilation may be summarized by the equation If 'M* is fluorescent, triplet-triplet annihilation produces delayed fluores3M*(Parker, 1964). This phenomenon was first studied for pyrene, and was therefore dubbed "P-type delayed fluorescence" in contrast to the "E-type delayed fluorescence" discussed in Section 5.1.1. cence, reflecting the long lifetime of
-_--_---_-_ -- -- -- --Filter -E
I
TI
Example 5.11: Triplet-triplet annihilation (TTA) between different molecules A and X (heteroTTA) may produce the excited singlet state of either A or X. This reaction is utilized in a general and synthetically useful method of generating singlet oxygen according to
-
sen^ + hv 'Sens* -, 'Sens* 'Sens* + IO, -+ Sens + 10,
Figure 5.31. Relative energies of the lowest singlet and triplet states of benzophenone and naphthalene, and triplet sensitization of naphthalene; direct excitation of the higher-lying naphthalene S, state is prevented by a filter.
Strongly absorbing dyes such as Rose Bengal or methylene blue are usually used as photosensitizers. (Cf. Section 7.6.3.) Singlet oxygen generation is used in photodynamic tumor therapy-for instance, with porphyrins as photosensitizers. (Cf. Dougherty. 1W2.) The disadvantage of natural porphyrins such as hematoporphyrin is that they are of low chemical stability and have an absorption spectrum similar to that of
296
5
PHOTOPHYSICAL PROCESSES
5.4
BIMOLECULAR DEACTIVATION PROCESSES
hemoglobin. Therefore, expanded porphyrins such as (22lcoproporphyrin 11 tetramethylester (18) (Beckman et al., 1990) and octaethyl[22lporphyrin(2.2.2.2) (19) (Vogel et a]., 1990) have been synthesized.
Delayed fluorescence from a very-short-lived upper excited singlet state S t populated by hetero-TTA has been observed for the first time using the system A = anthracene and X = xanthone (Nickel and Roden, 1982). An energy-level diagram for this system is shown in Figure 5.32, and the corrected spectrum of the delayed fluorescence of anthracene and xanthone in trichlorotrifluoroethane is depicted in Figure 5.33. The band at 36,00040,000 cm-' has been assigned to the S,+St delayed fluorescence of anthracene produced by Ti'
+
of a solution of Figure 5.33. Corrected delayed fluorescence spectrum (-) moYL) and xanthone (X, [XI = 2 x lo-' moVL) anthracene (A, [A] = 7 x in 1.1.2-trichlorotrifluoroethane at 243 K. For comparison spectra of delayed fluorescence of A alone (-----)and of X alone (---..-) as well as absorption spectra of A (------) and X(-..---)are also shown (by permission from Nickel and Roden, 1982). TT. It is almost exactly a mirror image of the S,+S: anthracene spectrum.
absorption band in the
5.4.6 Kinetics of Bimolecular Photophysical Processes In the absence of photochemical reactions an excited-state molecule M* can be deactivated by emission, by radiationless decay, or by quenching:
Figure 5.32. Energy-level diagram of anthracene (A) and xanthone (X). Double lines denote the three different triplet pairs for which TI'A processes are indicated; asterisks mark the delayed fluorescence (DF) resulting from heteroTTA (by permission from Nickel and Roden, 1982).
The quantum yield of fluorescence is then k,:
@: k, + k, + kq[Q] =
PH( )'I.( )PHYSICAL PROCESSES
5
298
5.4
BIMOLECULAR DEACTIVATION PROCESSES
299
while in the absence of quencher, one has from Equation (5.10)
If k, k _ , , then kq(obs) = k,k,,$k-, = k,K, where K is the equilibrium constant for the formation of the complex; k,(obs) will be independent of solvent viscosity. Finally, if kq and k - , are of the same order, then k,(obs) will be less than
Hence, the ratio aF/@P of the quantum yield of fluorescence without quencher t o that with quencher is given by
kclip
@F - kF -
@$
+ k, + k,[QI k,
+ k,
=
I
+
kq[Q1 k,: + k,
=
I
+
zFk,,[Q]
(5.37)
where t, is the excited-state lifetime in the absence of quencher. Equation (5.37) is known a s the Stern-Volmer equation (Stern and Volmer, 1919). If @,/@P is plotted against the quencher concentration [Q] a straight line of slope k,t will result. Hence if t is known, the quenching rate constant k, is immediately obtained, and if k, is known t may be determined. Because the direct measurement of lifetimes is easily performed nowadays, it is also useful to transform the Stern-Volmer relation into the form
Quenching of two excited states, one of which is inaccessible to the quencher, will yield a Stern-Volmer plot that is not a straight line but rather a curve bent toward the x axis. An upward curvature, concave toward the y axis a s shown in Figure 5.36, is observed if combined dynamic and static quenching occurs. Static quenching results from the formation of a nonfluorescent ground-state complex between fluorophore and quencher o r from the statistical presence of the quencher next to the excited molecule (Sun e t al., 1992a). In the case of purely static quenching the Stern-Volmer plot yields a straight line with a slope equal to the equilibrium constant K = [MQ]/[M][Q], whereas tJt8 = I , since the lifetime of the uncomplexed molecules is unperturbed by the presence of the quencher. Example 5.12:
which is obtained from the expression for t, given in Table 5.2 and from the corresponding expression that takes into account bimolecular quenching. For many systems kq is of the order 101° L m o l l s-I, that is, close t o the diffusion-controlled rate constant k,,,. This suggests that in these cases quenching is s o rapid that the rate-determining step is the actual diffusion of the molecules t o form an encounter complex. Thus, &dim
M*
+Q e k
(M*
. ..
Figure 5.34 shows the Stern-Volmer plot for triplet quenching of the photochemical addition of benzaldehyde to 2,3-dimethyl-2-butene (cf. Section 7.4.3) by piperylene. The ratio of the quantum yield of oxetane formation without quencher to that with quencher is plotted against the quencher concentration [Q]. From the resulting straight line it may be concluded that the reaction proceeds via a single reactive state which is assigned as '(n,n*). In the case of diffusion-controlled quenching, k, should depend on solvent viscosity in the same way as k,,,, that is. k, = aT/q. A plot of k,(obs) for triplet
k Q ) A M + Q
I
where k, is the actual quenching rate constant within the complex. Under steady-state conditions kdin[M*I[Q] = [(M* . . . Q)I(kq
+ k-1)
and from
the observed quenching rate constant -00
1
I
1
2
3
Piperylene
is obtained. If k, 9 k-,, then kq(obs) = k,,, and the observed quenching rate that is equal to the diffusion rate will be dependent on solvent viscosity.
LI
I
ImollLl
Figure 5.34. Stern-Volmer plot for triplet quenching of the oxetane formation from benzaldehyde and 2,3-dimethyl-2-butene by piperylene for different olefin concentrations (by permission from Yang et al., 1%7).
5
PHOTOPHYSlCAL PROCESSES
5.5
301
ENVIRONMENTAL EFFECTS
quenching of valerophenone by 2.5-dimethyl-2,4-hexadieneagainst llq does not give a straight line (Figure 5.351, so in this case energy transfer cannot be solely diffusion-controlled; that is, not every encounter complex between an excited donor and an acceptor molecule results in energy transfer. The Stern-Volmer plot of fluorescence quantum yields for the quenching of 10-methylacridinium chloride by the nucleotide guanosine-5'-monophosphate shown in Figure 5.36 provides an excellent example in which both static and dynamic quenching occur. Since lifetime measurements were available, the contribution of static quenching could be separated off; the Stern-Volmer plot of the purely dynamic quenching was obtained using t,lt$.
5.5 Environmental Effects For a deeper understanding of photophysical processes it is sometimes very useful to study the influence of the environment. The results of some such studies will be briefly mentioned in the following sections. Figure 5.35. Comparison of the theoretical and the observed relation between k,, and I/q for triplet quenching of valerophenone by 2.5-dimethyl-2,Chexadiene (adapted from Wagner and Kochevar, 1968).
5.5.1 Photophysical Processes in Cases and in Condensed Phases In solution vibrational relaxation is extraordinarily fast, and usually, thermal equilibrium is reached before emission or other photophysical or photochemical processes can come into play. In low-pressure vapors, however, time intervals between collisions will be large enough for other processes to be able to compete with vibrational relaxation. "Hot" reactions in excited states and particularly in the ground state, with the molecule excited to a
\\ I hv
2
6
L rnrnollL
8
Excited state
10
GMP
Figure 5.36. Fluorescence quenching of 10-methylacridinium chloride by guanosine-5'-monophosphate and separation of the static and dynamic quenching cohtributions (by permission from Kubota et al., 1979).
Figure 5.37. Schematic representation of the potential energy surfaces of "hot" excited-state reactions.
302
5
PH0.I ()PHYSICAL PROCESSES
higher vibrational state than would correspond to thermal equilibrium, will become important. The excess excitation energy may be sufficient to overcome barriers to chemical reactions and to enable the molecule to undergo processes not accessible in thermal equilibrium. This situation is illustrated in Figure 5.37 by means of schematic potential energy surfaces. It is easy to see from this representation that hot excited-state reactions are dependent on the excitation wavelength; a reaction is possible only if excitation is into a vibrational level higher in energy than the barrier. The conversion of benzene into benzvalene discussed in Section 5.3.1 is an example of such a wavelength-dependent photoreaction.
5.5.2 Temperature Dependence of Photophysical Processes Photochemical reactions competing with photophysical processes often require thermal activation in order to cross small barriers in the S, or T, state. (See Section 6.1 .) In general, such reactions may be suppressed at low temperatures and photophysical processes would then be favored. This is by far the most important temperature effect on photophysical processes although the following effects should also be considered. In discussing photophysical processes conformers can be considered as different species since in the excited state the conversion of one conformer into another is frequently slower than many of the competing processes. (Cf. Example 7.13.) Also, conformational changes are often much slower in the excited state than in the ground state, because formal single bonds of conjugated systems can show higher n-bond orders in the excited state than in the ground state. This is true, for instance, for polyencs. 'The conformational distribution in the ground state is temperature dependent, and rates of photophysical or photochemical processes characteristic for a particular conformer will become temperature dependent as well. There is, however, also evidence for just the opposite case: in the (ad) excited state of certain oligosilanes, conformational changes are faster than in the ground state and are driven by large differences in the energies of conformers that are of nearly equal energy in the ground state (Sun et al., 1992b). Another type of temperature dependence that may be thought of as competition between temperature-dependent and temperature-independent processes has been observed for substituted anthracenes. It is due to the fact that the T, state of anthracene is energetically very close to the lowest excited singlet state S,. When T, is just above S, the intersystem crossing S,-T, will be associated with a barrier Ea. The rate constant of this transition can be expressed in the form of a normal Arrhenius equation k = Ae -"Iu and the intersystem crossing becomes temperature dependent. This situation is found in 9- and 9.10-substituted anthracenes. (Cf. Example 5.3.) The fluorescence quantum yield @, increases strongly with decreasing temperature. In other substituted anthracenes as well as in anthracene itself T2
5.5
ENVIRONMENTAL EFFECTS
303
is just below S,; intersystem crossing is facile and without a barrier, resulting in the quantum yield of fluorescence being lower and temperature independent. A case in which the fluorescence rate constant increases with temperature has been described recently (Van Der Auweraer et al., 1991). The fluorescence from TICT (twisted internal charge transfer) states of donor-acceptor pairs that may be considered as strongly heterosymmetric biradicaloids (see Section 4.3.3) is strongly forbidden since the donor and acceptor orbitals avoid each other in space nearly perfectly at the orthogonally twisted geometry and the overlap density is nearly exactly zero. Vibrational activation populates levels in which the twist angle deviates from 90" and that carry larger transition moments to the ground state. A similar observation was made for highly polar intramolecular exciplexes.
5.5.3 Solvent Effects Polar solvents and in particular hydrogen bonds are able to stabilize polar molecules. This is also true for the excited states of a molecule that may have quite different polarities, and solvent variation may possibly change the energy order of the excited states as well as transition moments. (Cf. Section 2.7.2.) As an example we recall quinoline, whose solvent-dependent fluorescence quantum yield has been discussed in Section 5.3.2. Example 5.13:
p-Dimethylaminobenzonitrile (20), which has been studied by Lippert (1%9), is a classical example of the solvent effect on the ordering of excited states.
Whereas the ground state and locally excited benzene states are of lower polarity, there is another excited state of 20 consisting of an NR2 group as donor and the C,H,CN group as acceptor that has a dipole moment y = 12 D and that can be described as a TlCT state with intramolecular charge transfer (Grabowski and Dobkowski, 1983) stabilized by twisting around the single bond. (Cf. Section 4.3.3.) In nonpolar solvents fluorescence from a locally excited state is observed. In polar solvents, however, the TICT state is stabilized to such an extent that it becomes the lowest excited state, and fluorescence from this state is observed. By means of picosecond spectroscopy it has been shown that in nonpolar solvents, relaxed fluorescence sets in 7-20 ps after the exciting flash due to vibrational relaxation, whereas relaxed fluorescence from the TlCT state is characterized by an appreciably longer inhibition time of = 40 ps associated with a geometry change and reorientation of the polar solvent molecules in the field of the large excited-state dipole moment (Struve et al., 1973).
5
PHOTOPHYSICAL PROCESSES
5.5
ENVIRONMENTAL. EFFECTS
305
excited state and the CT state have different symmetries s o that the crossing of the potential energy surfaces is not avoided. If variations in the distance between H,O molecules and the donor-acceptor system are taken to be characteristic for solvation changes and hence for variations in solvent polarity, the schematic diagram of Figure 5.39 is obtained which shows the solvent dependence of the fluorescence of donoracceptor pairs. According to this diagram the energy of the exciplex fluorescence decreases with increasing solvent polarity. Example 5.14: c- Solvent approach
Using rigid bichromophoric systems with donor and acceptor components sterically fixed by spacers, the dependence of electron transfer on the geometry as well as on solvation has been studied. According to Figure 5.40 exciplex fluorescence to be expected from the diagram in Figure 5.39 can be observed for system 22a in polar solvents, whereas electron transfer does not occur in solvents of low polarity, and only fluorescence of the donor component is seen,
Dependence of the light-induced electron transfer on donor-acceptor separation R and upon solvation characterized by the distance between solvent molecules and donor-acceptor system (by permission from Ramunni and Salem, 1976).
Figure 5.38.
Solvent effects may be discussed in a particularly illuminating way in the case of electron-transfer reactions. Figure 5.38 gives a schematic three-dimensional representation of the potential energy surfaces responsible for photoinduced electron transfer and their dependence upon donor-acceptor separation and upon the approach of solvent molecules (Ramunni and Salem, 1976). The diagram is based on model calculations for a system consisting of NH, as the donor, cyanoethylene a s the acceptor, and two H,O molecules. These molecules were arranged in such a way that the locally
Solvent polarity
a : X = H , Y =COOC2H3 b:X= Y-CN
a:X=COOMe,Y b:X- Y -CN
=
H
-
Figure 5.39. Schematic representation of the fluorescence of donor-acceptor systems as a function of solvent polarity (by permission from Pasman et al., 1985).
Figure 5.40.
Fluorescence maxima of compounds 22a (e),22b (o),and 23 (0) as a function of solvent polarity (by permission from Pasman et al., 1985).
1
306
5
l7i( V1'OI'HYSICAL PROCESSES
and this agrees with the fluorescence from compound 23. For larger donoracceptor separ;~tions.a s in 21a, only donor fluorescence independent of solvent polarity is observed. Irrespective of solvent polarity and donor-acceptor separation. systems 21b and 22b. with inherently better acceptor moieties, undergo extremely fast electron transfer (k, > lou1s I ) and show the expected exciplex fluorescence (Figure 5.40). Solvent reorganization is believed t o provide a significant barrier for long-range electron transfer from the donor t o the weaker but not t o the stronger acceptor (Pasman e t al., 1985). A relation has been derived that provides useful design criteria for systems t o display optimally fast charge separation across a given distance that is also virtually independent of both medium effects and temperature (Kroon e t al., 1991). From this relation it is predicted that a D-bridge-A system which is barrierless in a nonpolar solvent like nhexane will stay barrierless in any other solvent. If it is not barrierless in nhexane it will never become so whatever more polar solvent is used.
Other possible solvent effects depend on the heavy-atom effect, which may favor intersystem crossing, and on the viscosity, which may change the rate of diffusion and hence the collisional triplet-triplet energy transfer. These effects have already been mentioned in previous sections.
307
SUPPLEMENTAL READING
Cirldwell, R.A. (1984). "Intersystem Crossing in Organic Photochcniical Intermediates," I'rtrc? Appl. Chant. 56. l 167. Freed, K.F. (1978). "Radiationless Transitions in Molecules," Ace. Chem. Rcs. 11.74. Robinson. G.W., Frosch, K.P. (1962). "Theory of Electronic Energy Relaxation in the Solid Phase." J . Ch'rn. Plrys. 37. 1962. Robinson, G.W., Frosch, R.P. (1963), "Electronic Excitation Transfer and Relaxation," J . Chem. Phys. 38, 1187. Siebrand, W. (1967). "Radiationless Transitions in Polyatomic Molecules," J . Chem. Phys. 46; 440; 47, 241 1. Teller, E. (1960). "Internal Conversion in Polyatomic Molecules," Israel J . Chem. 7, 227.
Luminescence Lakowicz, J.R. (1983). Princ.ip1e.vof Hrrorc,.sc~c,nc.c~ Spc,ctro.scopv; Plenum Press: New York, 1-ondon. Parker. C.A. (1969). Photolrrrrtine.sc~~~~i~~c~ in Solrttior~;Elsevier: London. Turro, N.J., Liu. K.-C., Chow, M.-F., Lee, P. (1978), "Convenient and Simple Methods for the Observation of Phosphorescence in Fluid Solutions. Internal and External Heavy Atom and Micellar Effects," J . Photochem. Photohiol. 27, 523.
Polarization
Supplemental Reading Barltrop, D.A.. Coyle, J.D. (1975). Excited Stcites in Orgctrtic ('1tertzi.stry; Wiley: London.
Dorr. F., Held, M. (1960), "Ultraviolet Spectroscopy using Polarized Light," Angew. Chem. 72, 287.
Birks, J.B. (1970). Photophysics of Aromutic Molecrtles; Wiley: New York.
Michl, J., Thulstrup, E.W. (1986), Spectroscopy with Polurized Light: VCH Inc.: New York.
Birks. J.B., Ed. (1973175). Orgcrriic. Molec.iilar Pltotoph,ysic~.s.Vol. 1-2. Wiley: New York.
Thulstrup, E.W.. Michl, J. (1989). Elemcntury Pol~rizutionSpectrosc,opy; VCH Inc.: New York.
r CRC Press: Boca Gilbert, A., Baggot, J. (1991), Es.sc~tttic11.sof M o l e c r ~ l ~Pliotoc~hc~rrzistrv; Raton.
1-amola, A.A. (1971-76). Creution crnd Detection (?/'Excitrtl Sttitc,~.Vol. 1-4: Marcel Dekker: New York. crrtd Photophysics, Part 1 Rabek, J.F. (1982-91). Expcrirnc~rrtcilMctltods irt Pliotocltc~itristr.~ and 2. Photochc~rrlistrycrttd Pltotoph.~sic.s;CRC Press: Boc;~Raton, FL.
Excimers and Exciplexes Beens; H.. Weller. A. (1975). in Orgcirtic. Molec.rtlnr Photoplty.sic~.s,Vol. 2. Birks, J.B., Ed.; Wiley: London.
+ 21 I'hotocycloadditions,"
Simons. J.P. (1971). Pliotochemistry urid Spectroscopy: Wiley: 1-ondon.
Caldwell, R.A.. Creed, D. (198O), "Exciplex Intermediates in 12 Acc.. Chern. Res. 13, 45.
Turro, N.J. (1978). Moc/c*rrr Molc,c~rilctrPltotoc~hrrrtistr?~; Benjiunin: Menlo Park.
Davidson. R.S. (1983). "The Chemistry of Excited Complcxcs: a Survey of Reactions,"
Time-resolved Laser Spectroscopy
Ac/18.Pltys. Org. C'ltc,rrl. 19. l .
Fiirstcr, Th. (1969). "Exciniers." A t t g c , ~ C'ltc~m. . Ittt. Ed. ErtgI. 8 . 333.
Fleming, G.R. (1986). Chemic~ctlApplicctrions (?fUltrctfirst Sprc~trosc~opy: Oxford University Press: Oxford. c?fPic.osc,c,ondSpc.c.rrosc~pyto Cherttistry. NATO Eisenthal, K.B.. Ed. (1983). Applic~~tiorls AS1 Series: Reidel: Dordrecht.
Gordon, M., Ware, W.R.. Eds. (1975). The E.rciplex: Academic Press: New York. Lim. E.C. (1987). "Molecular Triplet Excimers." Acc. Chem. Res. 20. 8. Weller, A. (1982). "Exciplex and Radical Pairs in Photochemical Electron Transfer," Pitre Appl. Chem. 54. 1885.
Rentzepis, P.M. (1982), "Advances in Picosecond Spectroscopy," Scic,nce 218, 1183.
Electron Transfer
Radiationless Deactivation Avouris, P., Gelhart. W.M., 1'1-S;~yctl. M.A. (1977). "Nonl.:~cli;\livcl~lcctronicKcl;~xation 1111cl(*r ('t>lli\i,>n
1:1t*,-
(',>~lcli~io '' ~( '~ I I \, , I I I
/?,,I,.
77. 793.
Closs. G.L.. C;llcatcrra. L.T.. Grcen. N.J.. Penfield. K.W.. Miller, J.R. (1986). "Distance, Stereoelectronic Effects. and the Miircus Inverted Region in Intramolecular Electron Trilnsfer in Organic Radical Anions." J . Pliys. Clicrt~.90. 3673.
308
5
PHOTOPHYSICAL PROCESSES
CHAPTER
Eberson, L. (1987), Electron Transfer Reactions in Organic Chemistry; Springer: Berlin, Heidelberg, New York. FOX, M.A.. Ed. (1992). "Electron Transfer: A Critical Link between Subdisciplines in Chemistry," thematic issue of Chem. Rev. 92, 365-490. Fox, M.A., Chanon, M., Eds. (1988). Plzotoindrtced Electron Trctnsfc,r, Vol. 1-4, Elsevier: Amsterdam.
Photochemical Reaction Models
Gould, I.R., Farid, S. (1988). "Specific Deuterium Isotope Effects on the Rates of Electron Transfer within Geminate Radical-Ion Pairs," J. Am. Chem. Soc. 110, 7883. Kavarnos, G.J. (1993). Fundamentals of Photoinduced Electron Transfer; VCH Inc.: New York. Kavarnos, G.J., Turro, N.J. (1986), "Photosensitization by Reversible Electron Transfer: Theories, Experimental Evidence, and Examples," Chem. Rev. 86, 401. Marcus, R.A., Sutin, N. (1985), "Electron Transfers i n Chemistry and Biology," Biochirn. Biophys. Actcr 81 1. 265. Mattay, J., Ed. (1990-92). "Photoinduced Electron Transfer, Vol. I-IV," Chemistry. 156, 158. 159, 163; Springer: Heidelberg.
Topics of Current
Salem. L. (1982). Elc,c.trons in Chcmic~crlRt,uctiorl.s:First I'rinc.ip1e.s; Wiley: New York. Weller, A. (1968). "Electron Transfer and Complex Formation in the Excited State," Pure Appl. Chem. 16, 115.
Energy Transfer Closs, G.L., Johnson, M.D., Miller, J.R., Piotrowiak, P. (1989). "A Connection between Intramolecular Long-Range Electron, Hole, and Triplet Energy Transfers," J. Am. Chem. Soc. 111, 3751. Lamola, A.A. (1969). "Electronic Energy Transfer i n Solution: Theory and Applications" in Techniques of Organic Chemistry 14; Weisberger, A., Ed.; Wiley: New York. Turro, N.J. (1977), "Energy Transfer Processes." Pure Appl. Chrm. 49,405. Yardley, J.T. (1980), Introduction to Molecular Energy Transfer; Academic Press: New York.
Ground-state reactions are easily modeled using the absolute reaction-rate theory and the concept of the activated complex. The reacting system, which may consist of one or several molecules, is represented by a point on a potential energy surface. The passage of this point from one minimum to another minimum on the ground-state surface then describes a ground-state reaction, and the saddle points between the minima correspond to the activated complexes or transition states. For a theoretical discussion of photochemical reactions at least two potential energy surfaces are required: the ground-state surface with reactant and product minima as well as the excited-state surface on which the photoreaction is launched. Furthermore, the theoretical model must be capable of describing what happens between the time of light absorption by a molecule in its electronic ground state and the appearance of the product molecule, also in its electronic ground state, and in thermal equilibrium with its surroundings.
6.1 A Qualitative Physical Model for Photochemical Reactions in Solution A starting point for the discussion of experimental results in mechanistic photochemistry is the knowledge of the shapes of the ground-state (S,,)and
310
I'HO'fOCHLMICAI> KFA(1ION MODELS
first excited-state (S,) singlet surfaces and the lowest triplet-state surface (TI).Three steps can then be distinguished: First, minima and funnels in S, and TI have to be located. Then, it must be estimated which minima (funnels) are accessible, given the reaction conditions, and which ones will actually be populated with significant probabilities. Finally, from the shape of the ground-state surface So it must be determined what the products of return from these important minima (funnels) in S, and TI will be. This simple model ignores molecular dynamics problems as well as information on additional excited states, density of vibrational levels, vibronic coupling matrix elements, etc., which are required in more sophisticated advanced applications.
6.1.1 Electronic Excitation and Photophysical Processes According to the Franck-Condon principle, light absorption is a "vertical" process. Consequently, immediately after brief excitation with broad-band light the geometry of the system is virtually identical with that just before excitation. Subsequently, however, the motions of the nuclei are suddenly governed by a new potential energy surface, so the geometry of the system will change in time. In solution, the surrounding medium will act as a heat bath and efficiently remove excess vibrational energy. In a very short time, on the order of a few (5-50) picoseconds, thermal equilibrium will be established and the molecule will be sitting in one or another of the numerous minima in the excited-state surface. (Cf. Section 4.1.3.) If, due to the initial kinetic energy of the nuclei, the molecule can cross barriers before reaching this minimum, the process is referred to as a reaction of "hot molecules in an excited state" or simply as a hot c~xcited-stuterecrction. (See Section 6.1.3.) If the initial excitation was not into the lowest excited state of given multiplicity, a fast crossing to that state will occur via internal conversion, that is, typically to the S, or TI state (Kasha's rule, cf. Section 5.2.1). In some cases a funnel in S, is accessible and internal conversion from S, to Socan be so fast that the first thermal equilibration of the vibrational motion in these molecules is achieved in a minimum in the So state. Such a process is referred to as a direct reccc.rion. Here as well. the excess kinetic energy of the nuclei may take the molecule over barriers in the So state into valleys other than the one originally reached; analogous to the above-mentioned reactions, such processes are referred to as hot grorrnd-stare reacrions. Finally, in the presence of heavy atoms or in other special situations (cf. Section 4.3.4) intersystem crossing may proceed so fast that it is able to
6.1
A QlJALlTATlVE PHYSICAL MODEL
311
compete with vibrational relaxation. The first thermally equilibrated species formed may then be found in a minimum on the T, surface, even if the initial excitation was into S,. TI may also be reached via sensitization, that is, by excitation energy transfer from another molecule in the triplet state. (Cf. Section 5.4.5.) In one way or another, picoseconds after the initial excitation, the molecule will typically find itself thermally equilibrated with the surrounding medium in a local minimum on the S,, TI, or So surfaces: in S, if the initial excitation was by light absorption, in TI if it was by sensitization or if special structural features such as heavy atoms were present, and in So if the reaction was direct. Frequently, the initially reached minimum in S, (or TI)is a spectroscopic minimum located at a geometry that is close to the equilibrium geometry of the original ground-state species, so no net chemical reaction can be said to have taken place so far. Slower processes are the next to come into play. The most important among these are: First, thermally activated motion from the originally reached minimum over relatively small barriers to other minima or funnels, representing the adiabatic photochemical reaction proper. Second, intersystem crossing that takes the molecule from the singlet to the triplet manifold and thus eventually to a new minimum in TI. Third, fluorescence or phosphorescence that return the molecule to the ground-state surface So. Fourth, radiationless conversion from S, or TI to So,which is usually not competitive unless the energy gap AE(Sl - So)or AE(Tl - So)between the S, or TI and So states in the region of the minimum is small. (Cf. Section 5.2.3.) In this case, a "hot" molecule is produced, and a hot ground-state reaction can take place before removal of the excess energy. Other processes are also possible. These are, for example, further photon absorption (cf. Example 6. I), or either excitation or deexcitation via energy transfer, such as triplet-triplet annihilation or quenching. (Cf. Section 5.4.) A complex reaction mechanism that consists either of a two- or three-quantum process. depending on the temperature, will be discussed in Example 7.18. Example 6.1:
The photochemical electrocyclic transformation of the cyclobutene derivative 1 to pleiadene (2) does not proceed from the thermalized S, and TI states. At room temperature, the reaction cannot compete with fluorescence or intersystem crossing (ISC) due to high intervening barriers. However, these can be overcome by excitation to higher states, populated starting from Soeither by excitation with a single photon of sufficiently high energy (A < 214 nm) or in a rigid glass at 77 K by subsequent two-photon excitation. In the latter case U V absorption followed by intersystem crossing yields the long-lived TIstate. This intermediate can absorb visible light corresponding to a TI+-T, transition (Castellan et al., 1978). The situation is illustrated in Figure 6.1. Experimentally, it was not possible to distinguish between a reaction from one of the higher ex-
312
PHOTOCHEMICAL REACTION MODELS
!
i
6.1
A QUALITATIVE I'HYSICAL MODEL ~
I[
a]
b)
I
Figure 6.2. Schematic representation of the jump from an excited-state surface (S, or TI)to the ground-state surface a) without chemical conversion and b) in the region
of a "continental divide" with partial conversion.
Figure 6.1. Photophysical processes and photochemical conversion of
6b,lob-dihydrobenzo[3,4]cyclobut[I,2-ulacenaphthylene. The numbers given are quantum yields of processes starting at levels indicated by black dots (by permissign from Castellan et al., 1978). cited S, or T, states and a reaction of vibrationally "hot" molecules in one of the lower excited states, possibly even Sl or TI.
excitation at least for some of the molecules, and the process is considered photochemical. The description given so far is best suited for unimolecular photochemical reactions. As mentioned above, if the process is bimolecular, both components together must be considered as a "supermolecule." The description given here remains valid except that some motions on the surface of the supermolecule may be unusually slow since they are diffusion-limited.
6.1.2 Reactions with and without Intermediates
No matter whether the eventual return to Sowill be radiative or radiationless, its most important characteristic is the geometry of the species at the time of the return, that is, the location of the minimum in S, or TIfrom which the return occurs. If the returning molecule reaches a region ofthe Sosurface that is sloping down to the starting minimum as shown in Figure 6.2a, the whole process is considered photophysical, since there is no net chemical change. If the return is to a region of S,, that corresponds to a "continental divide" (cf. Figure 6.2b) or that clearly slopes downhill to some other minimum in s,, a net chemical reaction will have occurred as a result of the initial
Figure 6.3 shows a schematic representation of two surfaces of the ground state (S,) and of an excited state (S, or TI)and of various processes following initial excitation. The thermal equilibration with the surrounding dense medium requires a sojourn in some local minimum on the surface for approximately I ps or longer. This may happen for the first time while the reacting species is still in an electronically excited state (Figure 6.3, path c), or it may occur only after return to the ground state (Figure 6.3, path a). In the former case, the reaction mechanism can be said to be "complex," in the latter, "direct." Direct reactions such as direct photodissociations cannot be quenched. The first ground-state minimum in which equilibrium is reached need not correspond to the species that is actually isolated. Instead, it may correspond to a pair of radicals or to some extremely reactive biradical, etc. (Figure 6.3, path k). Ncvertheless, this seems to be a reasonable point at which the photochemical reaction proper can be considered to have been
0.1
A QUALI'IXI'IVE PHYSICAL MODEL
315
Example 6.2: I'assage of a reacting system to the S,,surface through a funnel in the S, surface
is common in organic photochemical reactions, but the quantitative description of such an event is not easy. Until recently, it was believed that the funnels mostly correspond to surface touchings that are weakly avoided, but recent work of Bernardi, Olivucci, Robb, and co-workers (1990-1994) has shown that the touchings are actually mostly unavoided and correspond to true conical intersections. This discovery does not have much effect on the description of the dynamics of the nuclear motion, since the unavoided touching is merely a limiting case of a weakly avoided touching. Only when the degree of avoidance becomes large, comparable to vibrational level spacing, does the efficiency of the return to So suffer much. The expression "funnel" is ordinarily reserved for regions of the potential energy surface in which the likelihood of a jump to the lower surface is so high that vibrational relaxation does not compete well. The simplest cases to describe are those with only one degree of freedom in the nuclear configuration space. In such a one-dimensional case, the probability P for the nuclear motion to follow the nonadiabatic potential energy surface
Figure 6.3. Schematic representation of potential energy surfaces of the ground state (S,,) and an excited state (S, or TI) and of various processes following initial excitation (by permission from Michl. 1974a).
completed, even though the subsequent thermal reactions may be ofdecisive practical importance. If return t o So from the minimum in S , o r T , originally reached by the molecule is slow enough for vibrational equilibration in the minimum to occur first, the reaction can be said to have an excited-sttrte intermediate. The sum of the quantum yields of all processes that proceed from such a minimum, that is, from an intermediate, cannot exceed one. Those minima in the lowest excited-state surface that permit return t o the ground-state surface so rapidly that there is not enough time for thermal equilibration are termed "funnels." (Usually. they correspond to conical intersections, cf. Section 4.1.2.) By definition, direct photochemical reactions without an intermediate proceed through a funnel. Sometimes, a funnel may be located on a sloping surface and not actually correspond t o a minimum. Since different valleys in S,, may be reached through the same funnel depending on which direction the molecule first came from, the sum of quantum yields of all processes proceeding from the same funnel can differ from unity. The reason for the difference between an intermediate and a funnel is that a molecule in a funnel is not sufficiently characterized by giving only the positions of the nuclei-the directions and velocities of their motion are needed a s well (Michl, 1972).
Figure 6.4. Schematic representation of potential energy surfaces So and S, as well as So+S, excitation (solid arrows) and nuclear motion under the influence of the potential energy surfaces (broken arrows) a) in the case of an avoided crossing, and b) in the case of an allowed crossing (adapted from Michl, 1974a).
316
PHUrOCHEMICAL KEACTION MODELS
(i.e., to perform a jump from the upper to the lower adiabatic surface) is given approximately by the relation of Landau (1932) and Zener (1932): P = exp [ ( - n2/h)(AEZIvAS)]
Here, AE is the energy gap between the two potential energy surfaces at the geometry of closest approach, AS is the difference of surface slopes in the region of the avoided crossing, and v is the velocity of the nuclear motion along the reaction coordinate. Thus, the probability of a "jump" from one adiabatic Born-Oppenheimer surface to another increases, on the one hand, with an increasing difference in the surface slopes and increasing velocity of the nuclear motion, and on the other hand, with decreasing energy gap AE. The less avoided the crossing, the larger the jump probability; when the crossing is not avoided at all, AE = 0 and therefore P = 1. This is indicated schematically in Figure 6.4. In the many-dimensional case, the situation is more complicated (Figure 6.5). The arrival of a nuclear wave packet into a region of an unavoided or weakly avoided conical intersection (Section 4.1.2) still means that the jump to the lower surface will occur with high probability upon first passage. However, the probability will not be quite 100%. This can be easily understood qualitatively, since the entire wave packet cannot squeeze into the tip of the cone when viewed in the two-dimensional branching space, and some of it is forced to experience a path along a weakly avoided rather than an unavoided crossing, even in the case of a true conical intersection. An actual calculation of the S,+S, jump probability requires quantum mechanical calculations of the time evolution of the wave packet representing the initial vibrational wave function as it passes through the funnel (Manthe and
Figure 6.5. Conical intersection of two potential energy surfaces S, and So; the coordinates x, and x, define the branching space, while the touching point corresponds to an (F - 2)-dimensional "hyperline." Excitation of reactant R yields R*, and passage through the funnel yields products P, and P, (by permission from Klessinger, 1995).
Koppel, 1990), or a more approximate semiclassical trajectory calculation (Herman, 1984). In systems of interest to the organic photochemist, ~ i r ~ ~ t i l t a neous loss of vibrational energy to the solvent would also have to be included, and reliable calculations of quantum yields are not yet possible. I t is perhaps useful to provide a simplified description in terms of classical trajectories for the simplest case in which the molecule goes through the bottom of the funnel, that is, the lowest energy point in the conical intersection space. The trajectories passing exactly through the "tip of the cone" (Figure 6.5) proceed undisturbed. They follow the typically quite steep slope of the cone wall, thus converting electronic energy into the energy of nuclear motion. This acceleration will often be in a direction close to the x, vector, that is, the direction of maximum gradient difference of the two adiabatic surfaces. (Cf. Example 4.2.) Trajectories that miss the cone tip and go near it only in the twodimensional branching space have some probability of staying on the upper surface and continuing to be guided by its curvature, and some probability of performing a jump onto the lower surface and being afterward guided by it. However, in this latter case, an amount of energy equal to the "height" of the jump is converted into a component of motion in a direction given by the vector X, (direction of the maximum mixing of the adiabatic wave functions, cf. Example 4.2), which is generally not collinear with x, and is often approximately perpendicular to it (Dehareng et al., 1983; Blais et al., 1988). After the passage through the funnel, motion in the branching space, defined by the x,, xz plane, is most probable. Of course, momentum in other directions that the nuclei may have had before entry into the funnel will be superimposed on that generated by the passage through the funnel.
Figure 6.6. Schematic representation a) of the transition state of a thermal reaction and b) of the conical intersection as a transition point between the excited state and the ground state in a photochemical reaction. Ground- and excited-state reaction paths are indicated by dark and light arrows, respectively (adapted from Olivucci et al., I994b).
ti. 1
A bifurcated reaction path will probably result, and several products are often possible as a result of return through a single funnel. (Cf. Figure 6.5.) In summary, the knowledge of the arrival direction, the ~nolecularstructure associated with the conical intersection point, and the resulting type of molecular motion in the x,, x, plane centered on it provide the information for rationalizing the nature of the decay process, the nature of the initial motion on the ground state, and ultimately. after loss of excess vibrational energy, the distribution of product formation probabilities. Depending on the detailed reaction dynamics, the sum of the quantum yields of photochemical processes that proceed via the same funnel could be as low as zero or as high as the number of starting points. To appreciate the special role of a conical intersection as a transition point between the excited and the ground state in a photochemical reaction, it is useful to draw an analogy with a transition state associated with the barrier in a potential energy surface in a thermally activated reaction (Figure 6.6). In the latter, one characterizes the transition state with a single vector that corresponds to the reaction path through the saddle point. 'The transition structure is a minimum in all coordinates except the one that corresponds to the reaction path. In contrast, a conical intersection provides two poss~blelinearly independent reaction path directions.
Often. the minimum in S, or TI that is originally reached occurs near the ground-state equilibrium geometry of the starting molecule (Figure 6.3. path c); the intermediate corresponds to a vibrationally relaxed approximately vertically excited state of the starting species, which can in general be identified by its emission (fluorescence or phosphorescence, Figure 6.3, path d). Quenching experiments can help decide whether the emitting species lies on the reaction path or represents a trap that is never reached by those molecules that yield products. However, often the minimum that is first reached is shallow and thermal energy will allow the excited species to escape into other areas on the S, or TI surface before it returns to So(Figure 6.3, path e). This is particularly true for the TI state due to its longer lifetime. In the case of intermolecular reactions the rate also depends on the frequency with which diffusion brings in the reaction partner. The presence of a reaction partner may provide ways leading to minima that were previously not accessible, for example, by exciplex formation. Other possibilities available to a molecule for escaping from an originally reached minimum are classical energy transfer to another molecule, absorption of another photon (cf. Example 6.1), triplet-triplet annihilation (cf. Example 6.4), and similar processes.
A QUALII'A'rIVEPHYSICAL MODEL
'(n,n*) state (S,)to the CC double bond, whereas intersystem crossing into the '(n,n*) state (T,) prevents oxetane formation. The '(n,n*) state is quenched by triplet energy transfer to the diene, which then undergoes a sensitized transcis isomerization to 7 (Hautala et al., 1972).
Example 6.4:
A well-studied example of a photoreaction involving excimers is anthracene dimerization (Charlton et al., 1983). Figure 6.7 shows part of the potential energy surfaces of the supermolecule consisting of two anthracene molecules. Singlet excited anthracene ('A* + 'A) can either fluoresce (monomer fluores-
Example 6.3:
For reactions proceeding from the S, state. intersystelli crossing is frequently a dead end. Thus. irradiation of truru-2-methylhexadiene (4) in acetone (3) yields the oxetanes 5 and 6 through stereospecific addition of the ketone in its
319
Figure 6.7. Schematic potential energy curves for the photodimerization of anthracene (adapted from Michl. 1977).
320
PHOTOCHEMICAL REACTION MODELS
cence hv,), intersystem cross to the triplet state ('A* + 'A), or undergo a bimolecular reaction to form an excimer I(AA)*, which can be identified by its fluorescence (hv,). Starting from the excimer minimum and crossing over the barrier, the molecule can reach the pericyclic funnel and proceed to the ground-state surface leading to the dimer (24%) or to two monomers (76%). In this case intersystem crossing to the triplet state does not necessarily represent a dead end, because two triplet-excited anthracene molecules may undergo a bimolecular reaction to form an encounter pair whose nine spin states are reached with equal probability. (Cf. Section 5.4.5.5.) One of these is a singlet state '(AA)**that can reach the pericyclic funnel without forming the excimer intermediate first. The experimentally observed probability for the dimer formation via triplet-triplet annihilation agrees very well with the spin-statistical factor (Saltiel et a)., 1981). From Example 6.4 it can be seen that the molecule may also end up in a minimum o r funnel in S, o r T, that is further away from the geometry of the starting species. This then corresponds t o a "nonspectroscopic" minimum o r funnel (Figure 6.3, minimum f) such a s the pericyclic funnel of the anthracene dimerization in Figure 6.7, o r even to a spectroscopic minimum of another molecule o r another conformer of the same molecule (Figure 6.3, minimum i). Reactions of the latter kind can sometimes be detected by product emission (Figure 6.3, path j). (Cf. Example 6.5.) In many photochemical reactions return t o the ground-state surface S, occurs from a nonspectroscopic minimum (Figure 6.3, path g) that may be reached either directly o r via one o r more other minima (Figure 6.3, sequence c,e).
6.1.3 "Hot" Reactions Even in efficient heat baths, there are usually several short periods of time during a photochemical reaction when the reactant vibrational energy is much higher than would be appropriate for thermal equilibrium. One of these occurs at the time of initial excitation, unless the excitation leads into the lowest vibrational level of the corresponding electronic state. The amount of extra energy available for nuclear motion is then a function of the energy transferred t o the molecule in the excitation process, that is, a function of the exciting wavelength. Further, such periods occur during internal conversion (IC) o r intersystem crossing (ISC), when electronic energy is converted into kinetic energy of nuclear motion, o r after emission that can lead t o higher vibrational levels of the ground state. The length of time during which the molecule remains "hot" for each of these periods clearly depends on the surrounding thermal bath. In ordinary liquids at room temperature it appears to be of the order of picoseconds. (Cf. Section 5.2.1 .) Although short, this time period permits nuclear motion during lo2-1W vibrational periods. During this time, a large fraction of the mol-
ecules can move over potential energy barriers that would be prohibitive at thermal equilibrium. Since the vibrational energy may be concentrated in specific modes of motion, a molecule can move over barriers in these favored directions and not move over lower ones in less-favored directions. Chemical reactions (i.e., motions over barriers) that occur during these periods are called "hot. " Example 6.5:
Reactions of hot molecules are known for the ground state as well as for excited states. Thus, after electronic excitation of 1,Cdewarnaphthalene (8) in a glass at 77 K emission of 8 as well as of naphthalene (9) produced in a photochemical electrocyclic reaction is observed. The intensity of fluorescence from 9 relative to that from 8 increases if the excitation wavelength is chosen such that higher vibrational levels of the S, state of 8 are reached. This suggests that there is a barrier in S,. (Cf. Example 6.1.) If some of the initial vibrational energy is utilized by the molecule to overcome the reaction barrier before it is lost to the environment, 9 is formed and can be detected by its fluorescence. If the energy is not sufficient or if thermal equilibrium is reached so fast that the barrier can no longer be overcome, fluorescence of 8 is observed. The situation is even more complicated because at very low temperatures a quantum mechanical tunneling through the small barrier in S, of 8 occurs (Wallace and Michl, 1983).
Thermal o r quusi-equilibrium reactions, as opposed to hot ones, can be usefully discussed in terms of ordinary equilibrium theory a s a motion from one minimum t o another, using concepts such as activation energy, activation entropy, and transition state, as long as the motion remains confined t o a single surface. "Leakage" between surfaces can be assigned a temperature-independent rate constant. In this way jumps from one surface to another through a funnel may be included in the kinetic scheme in a straightforward manner.
Example 6.6: For the photoisomerization of 1.4-dewarnaphthalene (8) to naphthalene already discussed in Example 6.5 the following efficiencies could be measured in an N, matrix at 10 K (Wallace and Michl, 1983): q1.(8)= 0.13, qsT(8)= 0.29, qR. = 0.14, qRd+ qrcl = 0.44, qJ9) = 0.047, and qrs(9) = 0.95. Here, the indices Ra and Rd refer to the adiabatic and diabatic reaction from the singlet state, ret refers to the nonvertical return to the ground state of 8, and TS refers to the intersystem crossing to the ground state of 9. M;~kingthe plausible assumption that triplet 8 is converted to triplet 9 with lo()%efficiency. these data
1
Figure 6.8. Schematic representation of the potential energy surfaces relevant for the photochemical conversion of 1.4-dewarnaphthalene to naphthalene. Radiative and nonradiative processes postulated are shown and probabilities with which each path is followed are given (by permission from Wallace and Michl, 1983).
may be combined to produce the graphical representation shown in Figure 6.8. Here the probabilities with which each path is followed after initial excitation of a molecule of 8 into the lowest vibrational level of its IL, (S,) state are shown. Using the values r,.(8) = 14.3 ? I ns and T , (9) = 177 ? 10 ns for the fluorescence lifetimes, rate constants can be derived f~.omEquation (5.8) for the various processes. Thus, one obtains k,. = qlltl = 9.1 x 10" s ' for the fluorescence of 8 and k, = (q,, + qWd+ vrr,)ltk= 4.1 x 10' s I for the rate of passing the barrier.
6.1.4 Diabatic and Adiabatic Reactions The description of the course of photochemical processes outlined s o far implies that there is a continuous spectrum of reactions from unequivocally diahatic ones that involve a nonradiative jump between surfaces t o unequiv-
I
/\
QLII~LI1AI IVE PHYSICAL MODEL
323
ocally udic1hatic ones that proceed on a single surface (Forster, 1970). The dicrhatic liinit is represented by reactions in which return to the ground state occurs at the geometry of the starting material. The rest of the reaction then occurs in the ground state and the reaction must be a hot ground-state one; otherwise the ground-state relaxation would regenerate the starting material, resulting in no net chemical change. The other extreme, a prrrely udiahutic reaction, is represented by reactions that proceed in an electronically excited state all the way to the ground-state equilibrium geometry of the final product. Here the excited state converts to the ground state via emission of light o r by a radiationless transition. In both these limiting cases, the return to the ground state occurs from a spectroscopic minimum at an ordinary geometry. In addition to such minima the lowest excited states tend t o contain numerous minima and funnels a t biradicaloid geometries, through which return to the ground state occurs most frequently. Most photochemical reactions then proceed part way in the excited state and the rest of the way in the ground state, and the fraction of each can vary continuously from case to case. (Cf. Figure 6.3, path a.) It is common to label "adiabatic" only those reactions that produce a "spectroscopic" excited state of the product (cf. Figure 6.3, path h), s o distinction between diabatic and adiabatic reactions would appear to be sharp rather than blurred. But this is only a n apparent simplification, since it is hard to unambiguously define a spectroscopic excited state. A second obvious problem with the ordinary definition of adiabatic reactions is the vagueness of the term "product." If the product is what is actually isolated from a reaction flask at the end, few reactions are adiabatic. (Cf. Example 6.7.) If the product is the first thermally equilibrated species that could in principle be isolated a t sufficiently low temperature, many more can be considered adiabatic. A triplet Norrish 11 reaction is diabatic if an enol and an olefin are considered as products. It would have to be considered adiabatic, however, if the triplet 14-biradical, which might easily be observed, were considered the primary photochemical product. (See Section 7.3.2.) Example 6.7:
For a photochemical conversion to be adiabatic. the excited-state surface has to exhibit an overall downhill slope from reactant to product geometries and must not contain unsurmountable barriers or local minima by which the reacting molecules get trapped and funneled off to the ground-state surface before reaching the product geometry. For instance, the photochemical conversion of 1.4-dewarnaphthalene to naphthalene is so strongly exothermic that the pericyclic minimum is without doubt shallow enough to facilitate efficient escape on the excited surface. (Cf. Figure 6.8; for a semiempirical calculation of the potential energy curves, see Jug and Bredow. 1991.) The result is that a portion of the reactants proceeds along the adiabatic path. An example of this behavior
324
PHOTOCHEMICAL KWC'I'ION M0l)t:LS
is the photochemical decomposition of the benzene dimers 10 and 11, which yields excited benzene efficiently. Since the benzene excited state lies at 110 kcaVmol and the wavelength of the exciting light was 335 nm, corresponding to 85.3 kcallmol, the result clearly demonstrates that a part of the chemical energy stored in the reactant is utilized in the adiabatic process to generate electronic excitation in the product (Yang et al., 1988).
The situation is different for the valence isomerization of the metacyclophanediene 12 to the methano-cis-dihydropyrene 13, which also proceeds adiabatically (Wirz et al., 1984). In this case reactant and product hom*oS correlate with each other in the same way as the corresponding 1,UMOs. The reaction is therefore allowed in the ground state. Nevertheless, the excited state has no correlation-imposed barrier, although the photochemical reaction should be forbidden by the simple Woodward-Hoffmann rules. (See also Example 7.18.) Other examples for adiabatic reactions may be found among triplet reactions, as funnels do not exist in the triplet surface, and return to the singlet ground state requires spin inversion. The ring opening of 1,4-dewarnaphthalene to naphthalene is an example of such a case. (See Figure 6.8.)
6.1.5 Photochemical Variables In addition to concentration there are essentially four reaction variables that can be relatively easily controlled and that may have a considerable effect on the course of a photochemical reaction; these are the reaction medium and temperature, and the wavelength and intensity of the exciting light. In addition, magnetic field and isotope effects may come into play.
6.1.5.1 m e Effecto f the Reaction Medium Medium effects can be divided into two classes: those that directly modify the potential energy surfaces of the molecule, such a s polarity o r hydrogen bonding capacity, affecting through strong solvation in particular the (n,n*) as opposed to the (n,n*)state energies, and those that operate in a more subtle manner. Examples of the latter are microscopic heat conductivity,
which determines the rate of removal of excess vibrational energy, the presence of heavy atoms, which enhance rates of spin-forbidden processes, o r viscosity, which affects diffusion rates and thus influences the frequencies of bimolecular encounters. Through these, it may also control triplet lifetimes o r the competition between monomolecular and birnolecular processes. Very high solvent viscosity, encountered in crystalline and glassy solids, also effectively alters the shape of potential energy surfaces by making large changes in molecular geometry either difficult o r impossible. This increases the probability that the excited molecule will not greatly change the initial geometry and will eventually emit light rather than react. The structural dependence of biradicaloid minima discussed in Section 4.3.3 on an example of twisting of a double bond A=B can be extended to take solvent effects into account. Not only the nature of the atoms A and B but also polar solvents and counterions affect the stability of zwitterionic states and states of charged species. Then, depending on the solvent, a biradicaloid minimum can represent either an intermediate o r a funnel for a direct reaction. Example 6.8:
If light-induced electron transfer is the crucial step in a photochemical reaction, the solvent dependence expected for this process (cf. Section 5.5.3) may carry over to the whole reaction. An example is the reaction of I-cyanonaphthalene (14) with tlonor-substituted acetic acids such asp-methoxyphenylacetic acid (15).
ti. l
A
QUALI'TA'rIVE PHYSICAL MODEL
I n polar solvents such as acetonitrile, electron transfer occurs followed by proton transfer from the radical cation to the radical anion with concurrent loss of CO?.The radicals collapse to addition products such as 16 or 17. Alternatively, the radical pair may escape the solvent cage to give, after hydrogen abstraction from a suitable hydrogen source, the reduction product 18. In nonpolar solvents such as benzene, however. electron transfer is not possible; only exciplex emission and no chemical reaction are observed (Libman, 1975).
6.1S.2 Temperature Effects Temperature can have an essential effect on the course of a photochemical reaction because it can affect the rates at which molecules escape from minima in S, or TI. At very low temperatures, even barriers of a few kcal/mol are sufficient to suppress many photochemical processes more or less completely. Processes such as fluorescence, which were too slow at room temperature, may then be able to compete. (Cf. Example 6.5.) At times, the excited molecule has the choice of reacting in two or more competing ways. and the competition between them can also be temperature dependent. An example is the temperature dependence of the diastereoselectivity and regioselectivity of the cycloaddition of' menthyl phenylglyoxylate 19 to tetramethylethylene (20) and to ketene acetal 21 (Buschmann et al., 1991):
Temperature dependence of the diastereoselectivity of oxetane formation (22: @, 23: A, 24: m) by cycloaddition of menthyl phenylglyoxalate (19) to tetramethylethylene (20) and ketene acetal (21) (hy permission from Huschmann et al., 1991).
Figure 6.9.
temperatures the benzene oxide predominates in the benzene oxide-oxepin equilibrium, and only the photochemistry of this tautomer is observed. At higher temperatures, however, the photochemical reactions of the oxepin (26), present in the equilibrium, prevail due to its considerably higher extinction coefficient E at the wavelength of irradiation (Holovka and Gardner, 1%7).
6.1.5.3 Effectsof Wauelength and Intensity of the Exciting Light Figure 6.9 shows a plot of In klk' versus T-I, where k and k' are the overall formation rate constants for the major and the minor oxetane isomer, respectively. The diastereomeric excess (% dc) can either increase or decrease with decreasing temperature. Changes in temperature generally affect ratios of conformer concentration of the starting ground-state molecules and also their distribution among individual vibrational levels, thus codetermining the nuclear configuration of the average molecule just after excitation. The equilibrium between valence tautomers is also affected by temperature changes, and a temperature effect on the photochemical reactivity may result. 'I'hus, irradiation of benzene oxide (25) at room temperature produces the furan 27, whereas irradiation at -80°C gives 11% 27. 74% phenol and 15% benzene. At lower
Changes in light wavelength determine the total amount of energy initially available to the excited species. As pointed out in Section 5.2.1, in dense media part of this energy is rapidly lost and after 10- "-10 - l 2 s the molecule reaches either one of the minima on the S , or T I surfaces, or in the case of a direct reaction through a funnel, one of the minima on the S,, surface. Since internal conversions need not be "vertical" the probability that one or another minimum is reached may change drastically as the energy of the starting point changes. In general, one can imagine that a higher initial energy will allow the molecule to move above barriers that were previously forbidden. and that additional minima will become available. Whether the additional energy is actually used for motion toward and above such barriers or whether it is used for motion in "unproductive" directions and eventually lost as heat could be a sensitive function of the electronic state and vibra-
328
PHOTOCHEMICAL KEACI'ION M0I)I::LS
6.1
A QUALII'A'I'IVI; I'I IYSICAL. MOI)I
329
If rotational barriers ;ire sufficiently high in the ground state a s well a s in the excited state, different conformers can exhibit different photochemistry. Thus, if the spectra ofconlbrmers are suft'icienlly different, ;IS in the case of substituted hexntriencs, the photoreiictions of one o r another of the conformers can be observed depending on the excitation wavelengths.
Figure 6.10. General state diagrams for reactions a) from a thermalized S, state, b) from a "hot" S, state, and c) from a higher excited state S,,.
tional level reached through initial excitation. According t o Figure 6.10 three cases can be distinguished in principle. The simple model does not, however, warrant a differentiation between, for example, a hot St and an isoenergetic but less hot S, state in a large molecule. In these large molecules the density of vibronic states is high, the Born-Oppenheimer approximation poor, and mixing of states likely to be extremely efficient. The effect of barriers on the S, surface has already been discussed in Example 6.5 for the conversion of I ,Cdewarnaphthalene to naphthalene. A different kind of wavelength dependence is observed for the photochemical reaction of thiobenzophenone (28) with acrylonitrile (29). Depending on the wavelength of the exciting light, products are obtained that are produced either from the TI state reached by ISC from the S t state (A > 500 nm) o r from the S, state (A = 366 nm). These results were demonstrated by appropriate quenching experiments (De Mayo and Shizuka, 1973). In this case, bimolecular reactions in the S, state are possible because the energy gap between the first two singlet states, the (n,n*) and the (n,n*) state, amounts t o = 50 kcallmol for thiones, which is unusually large; internal conversion S p S l is therefore slow. (Cf. Section 5.2.1 .)
Example 6.9: The photocyclization of(Z)-1,3J-hexatriene represents a well-studied example of the influence of ground-state conformational equilibrium on the product composition (Havinga, 1973). The unsubstituted triene exists mainly in the strans-s-trans conformation and therefore cannot undergo photocyclization; only 2-E isomerization has been observed. With the 2.5-dimethyl derivative the s-cis-s-cis conformation predominates and photocyclization yields dimethylcyclohexadiene ant1 some dimethylvinylcyclobutene. Finally, with the 2methyl derivative the s-cis-s-trans conformation is preferred, and a relatively good yield of methylbicyclo[3. I .O]hexene as well as methylvinylcyclobutene and methylhexatriene-1.2.4 is found. These results are collected in Scheme I.
Scheme 1
The product composition may also be determined by the wavelength dependence of a photostationary state. If bicyclononadiene (30) is irradiated with 254-nm light, cyclononatriene (31) is formed. On further irradiation, it converts slowly to the tricyclic product 32 at the expense of 30 and 31. If 300-nm light is used, no cyclononatriene (31) is found; only slow formation of the tricycle 32 is observed. This wnvelength dependence has been explained by a spatial orientation of the n syhtem of triene 31 as a Mobius array of orbitals (33) that favors the conrotatory ~.ecyclizationto form 30 to such an extent that the quantum yield for closure of 31 is about 20 times that for the opening of 30. This explains why at 254 nln the photostationary statc consists of;rhout 40% 30 and 60% 31, although the extinction coefficient of 30 is about 30 times that of 31. At 300 nm the ratio of extinction coefficients drops to about 0.02, and the photostationary state is greatly displaced in favor of the bicyclonon;L d'lene. thus allowing the less efficient formation of tricyclononene 32 to proceed from the diene (Dauben and Kellog, 1971). A third example, the photochemical valence isomerization of the tricycle 34 to form I-cyanoheptalene (35). is worthy of mention. It proceeds at 254 nm with a quantum yield nearly 100 times that at 365 nm. In this case, the rate of reaction depends on the initial concentration, but not on the instantaneous concentration. This is an indication that the isomerimtion proceeds via an exciplex
6.1
A QUALITATIVE PHYSICAL MODEL
with ;In clect~.c)llic;~lly cxcitccl colitirlllin;u~t. wl~oscconccl~tratiollrem;rins constant. I t is quite possible that similar eSSccts occur also with other wavelengthdependent photoreactions (Sugihara et al.. 1985).
carbene dimerization. Tetraphenylethylene (40). which also is the major product under flash-photolysis conditions, appears to result predominantly from dimerization of two triplet carbenes (Turro et al., 1980a).
6.1.5.4 Magnetic-Field Effects
Light intensity at the usual levels seldom has an effect on the primary photochemical step if all other variables are kept constant, although overall results may be considerably affected since it can control the concentration of the reactive intermediates. However, it will affect the outcome of a competition between primary one-photon and two-photon processes. These are particularly relevant in rigid glasses where triplets have a long lifetime and quite a few of them are likely to absorb a second photon. The additional energy can permit motion to new minima on the excited-state surfaces. thus leading to new products. Example 6.10: Irradiation o f diphenyldiazomethane (36) with a laser at 249 nm yields fluorene (37). 9.10-diphenylanthracene (38), and 9,IO-diphenylphenanthrene (39) i n varying amounts depending on the laser intensity or the initial concentration. With conventional lamp excitation these products are not observed. These findings have been explained by the following scheme: With increasing laser intensity the relative yields o f products formed by second-order reactions (3840) increase with respect to the production of fluorene (37) from ;I first-order reaction. I t is assumed that fluorene and diphenylanthracene originate from the singlet carbene. whereas diphenylphenanthrene (39) arises from either a triplet-singlet carbene dimerimtion or a triplet-triplet
Reactions involving radical pairs or biradicals can be affected significantly by a magnetic field imposed from the outside. In these intermediates, the singlet and triplet levels can lie so close together that even the very weak perturbation represented by the magnetic field can be sufficient to affect their mixing, and thus the outcome of the photochemical reaction. The fundamentals needed for the understanding of these phenomena have been discussed in Section 4.3.4. (For recent reviews, see Steiner and Ulrich, 1989, and Khudyakov et al.. 1993; see also Salikhov et al., 1984.)
6.1.5.5 Isotope Effects Diff'erent isotopes of iin element differ in their weight nncl possibly in other nuclear properties, such as magnetic moment. The mechanism of action of the ponderal effects is generally familiar from ground-state chemistry, and can be understood on the basis of differences in zero-point vibrational energies. (Cf. also Example 5.5.) In rare cases involving tunneling. the mass of the tunneling particle is affected by isotopic substitution. (Cf. Melander and Saunders, 1980.) In photochemical processes, contributions from both of these mechanisms can be quite lai-ge, since the reactions often involve very small barriers. In addition, those photochemical reactions that proceed by radical pair or biradical intermediates offer an opportunity for very large isotope effects based on differences in the magnetic moments of the individual isotopes. As outlined in Section 4.3.4, in these intermediates the singlet and triplet levels can be very close in energy, and the hyperfine coupling mechanism of singlet-triplet coupling plays a role in determining the nature of the products.
332
PHOTOCHEMICAL REACTION MODELS
This mechanism is exquisitely sensitive to nuclear magnetic moments. (Cf. Salikhov et al., 1984; Doubleday et al., 1989.)
6.2 Pericyclic Reactions Pericyclic minima and funnels that can be easily predicted by means of correlation diagrams are of great importance in concerted pericyclic photoreactions. However, there may be additional minima and barriers on the excitedstate surfaces that affect or even determine the course of the photoreaction.
6.2.1 Two Examples of Pericyclic Funnels The recognition of a leak in the lowest excited singlet potential energy surface halfway along the path of ground-state-forbidden pericyclic reactions is due to Zimmerman (1966). He argued at the Huckel level of theory, at which cont'igurational wave functions do not interact. The correlation diagram drawn Ibr the high-symmetry reaction path then exhibits a touching of S,,, S,, and Sz surfi~ces(Figure 6.1 la), since (I+,, [I), .,, and @)=I:are all degenerate at the pericyclic geometry. Van der Lugt and Oosterhoff (1969) took into account electron repulsion at the PPP level for the specific example of the conversion of butadiene to cyclobutene and noted that along the high-symmetry path the surface touching becomes avoided (Figure 6.1 I b), with a minimum in the S , surface rather than the S, surface as drawn in the older Longuet-Higgins-Abrahamson (1965) and Woodward-Hoffmann (1969) correlation diagrams. This "pericyclic minimum" results from an avoided crossing of the covalent G and D
Schematic correlation diagrams for ground-state-forbidden pericyclic reactions; a) HMO model of Zimmerman (1%6), b) PPP model of van der Lugt and Oosterhoff (1%9). and c) real conical intersection resulting from dii~gonalinteractions. The two planes shown correspond to the hom*osymmetric (y)and heterosymmetric (6) case. Cf. Figure 4.20.
configurations (a,and @I=;:). Van der Lugt and Oosterhoff pointed out that it will provide an efficient point of return to S,, and a driving force for the reaction. Subsequently, from an ab initio study of a simple four-electron-fourorbital model problem (H,), Gerhartz et al. (1977) concluded that the pericyclic "minimum" thitt results if symmetry is retained in the correlation diagram is not a minimum along other directions, specifically the symmetrylowering direction that increases the diagonal interactions in the perimeter, stabilizing the S, statc and destabilizing the S,, state (Section 4.4.1). They identified an actual S,-SO surface touching (conical intersection) in H,, in a sense returning to the original concept of Zimmerman (1966), but at a less symmetrical geometry. They proposed that analogous diagonal bonding in the pericyclic perimeter is a general feature in organic analogues of the simple Hz + Hz reaction they studied, in which the simple minimum actually was not a minimum hut a saddle point (transition structure) between two conical intersections (Figure 6.1 Ic). Assuming, however, that in the lowsymmetry organic analogues the surface touching probably would be weakly avoided, they conlinucd to retkr to the global region of the pericyclic funnel, including the diagonally distorted structure, as "pericyclic minimum." They proposed that becausc of the diagonal distortion at the point of return to the S,, surface, the mechanism now accounted not only for the ordinary 12, + 2,] pericyclic reactions, such as the butadiene-to-cyclobutene conversion, but also for x[2, + 2,] reactions, such as the butadiene-to-bicyclobutane conversion. Although the exact structure at the bottom point of the pericyclic funnel remained unknown, it was clear from the experiments that it is a funnel in the sense of being extremely efficient in returning the excited molecules to the ground state. Finally, in a remarkable series of recent papers, Bernardi, Olivucci, Robb and their collaborators (IWO-1994) demonstrated that the S,-S, touching actually is not avoided even in the low-symmetry case of real organic molecules, and they confirmed t he earlier conjectures by computing the actual geometries of the funnels (conical intersections) in the S , surface at a reasonable level of ab initio theory. They also pointed out that still additional reactions can proceed through the same pericyclic funnel, such as the cistrans isomerization of butadiene. Sections 6.2. I. I and 6.2.1.2 describe the pericyclic funnel in more detail for two particularly important and illuminating examples: the cycloaddition of two ethylene molecules and the isomerization of butndiene. We rely heavily on the results of recent ad initio calculations (Bernardi et id., IC)!90u;Olivucci et al., 1993).
Figure 6.11.
6.2.1.1 The Cycloaddition of Two Ethylene Molecules First, we consider the face-to-face addition of two ethylene molecules. The orbital correl;~tioncli;~! r;1111 li)r thc I~igh-hvn~rnctl-y path ;~n:~logous to tho rec-
6.2
tangular H z + Hz path discussed in Section 4.4.1 wils given in Figure 4.15. The resulting configuration and state correlation diagriims are essentially the same as those for the electrocyclic ring closure of hutadiene discussed in Section 4.2.3. At the halfway point (rectangular geometry, four equal resonance integrals along the perimeter), the situation corresponds to a perfect biradical, analogous to the n system of cyclobutadiene. (Cf. Section 4.3.2.) The order of the states is T below G (S,,), followed hy D (S,) and S (Sz) at higher energies. As discussed in Section 4.4.1 for the model case of H,, one can use either of the simplified approaches, 3 x 3 CI or 2 x 2 VB, to understand the nature of the distortion that leads from the rectangular geometry to the conical intersection of S, with S,,. The use of the former permits a discussion of substituent effects, and we shall consider it first. From Figure 4.20b it is apparent that a perturbation (3 produces a heterosymmetric biradicaloid and promises to lead to thc critically heterosymmetric situation (i,= c'i,,). whcre S,, and S, ;we degcnc~iteand u real conical inlersection results. f;rom 1;igur-e 6.12 i t is seen that ;I diagonal interiiction differentiates the otherwise degenerate energies of thc localized nonbonding orbitals of the cyclobutadiene-like perfect biradical and thus produces a 6 perturbation. This suggests that the conical intersection may be reached by a rhomboidal distortion of the pericyclic geometry. which decreases the length of one diagonal and increases that of the other and therefore corresponds to either an enhanced 1,3 or an enhanced 2.4 interaction. For unsubstituted ethylenes, these two cases are equivalent by symmetry, and the pericyclic funnel consists of two conical intersections separated by a transition state iit the rectangular geometry as shown in Figure 6.13 (Hernardi et al., I99oa.h; Bentzien iind Klessinger, 1994). In the case of the [2 + 21 photocycloaddition ol' substituted ethylenes, the two situations remain equivalent for head-to-head addition (1,2 arrangement of the substituents) but are no longer equivalent for head-to-tail addition (1.3 arrangement of the substituents), since the latter affects the energy difference between the two localized forms of the nonbonding orbitals. Ac2.L interaction
1.3 ~nteraction
Figure 6.12. The effect of di;igonal interactions on the energy of the degenerate loc;~liir.ednonhonding orbitals of ;I cyclohut;idicnc-like pcrlkct hiradical.
PEKICYCLIC REACTIONS
Figure 6.13. Pericyclic funnel region of ethylene dimerization, showing two equivalent conical intersections corresponding to 1.3 and 2,4 diagonal interactions and the transition state region at rectangular geometry (a = 0). The curves shown for a = 0 correspond to the van der Lugt-Oosterhoff model (by permission from Klessinger, 19951.
cess to onc of the otherwise equivalent conical intersections will thus be favored. Electron-donating substituents will reinforce the effect of the diagonal interaction in the case of 1,3 substitution and will counteract it in the case of 2.4 substitution. The opposite will be true of electron-withdrawing substituents (BonaCiC-Koutecky et al., 1987). The overall effect of substituents on the kinetics and product distribution in [2 + 21 photoaddition of olefins is however complicated by the possible intermediacy of an excimer or an exciplex, which will typically be more stable for the head-to-head arrangement of substituents. This is discussed in more detail in Section 7.4.2. Depending on the extent of diagonal interaction, the rhomboidal distortion may lead to preferential diagonal bonding in contrast to pericyclic bonding after return to the ground state, resulting in an x[2, + 2,] addition product rather than a [2, + 2,l product. Examples of x[2 + 21 cycloaddition reactions also are discussed in Section 7.4.2. In terms of the 2 x 2 V B model, the argument for rhomboidal geometries as favored candidates for a conical intersection (Bernardi, 1990a,b) goes as follows. (Cf. Figure 6.14.) The exchange integral KR (cf. Example 4.13), which corresponds to the spin coupling in the reactants, will decrease rapidly with decreasing 1.2 iind 3.4 overlap as the intrafi-agment CC distance
PHOTOCHEMICAL REACTION MODELS
Figure 6.14. Schematic representation of possible geometries at which the conical intersection conditions may be satisfied. Dark double-headed arrows represent K , , light arrows K,,,and broken arrows K , contributions (adapted from Bernardi et a].,
1990a). increases. The exchange integral K p , which corresponds to the product spin coupling, will increase rapidly with increasing 1,3 and 2,4 overlap, as the interfragment distance is decreased. K x will be quite small for rectangular geometries. While one might satisfy Ku = K, at these geometries, Ku = K x can never be satisfied. Figure 6.14 gives a schematic representation of possible geometries on the (F - 2)-dimensional hyperline, where the conditions imposed by the equations of Example 4.13 may be satisfied and El = E,, = Q. The Coulomb energy Q will be repulsive in the region of structures b through d in Figure 6.14 because of interference of the methylene hydrogen atoms. Thus, in agreement with the argument based on the two-electrontwo-orbital model, the favored region of conical intersection for the [2 + 21 cycloaddition is found at the rhomboidal geometries.
6.2.1.2 The Isomerization ofButadiene Monomolecular photochemistry of butadiene is rather complex. Direct irradiation in dilute solution causes double-bond cis-trans isomerization and rearrangements to cyclobutene, bicyclo[l. l .O]butane, and I-methylcyclopropene (Srinivasan, 1968; Squillacote and Semple, 1990). Matrix-isolation studies have established another efficient pathway, s-cis-s-trans isomerization (Squillacote et al., 1979; Arnold et al., 1990, 1991). Scheme 2 illustrates the phototransformations of s-cis-butadiene. Until recently, it was not clear whether the reaction paths leading from the two conformers to the various observed photoproducts are different already on the S, surface or whether they do not diverge until al'ter return to S,,. The latter alternative is now believed to be correct. As suggested by the results obtained for the H, model system (20 x 20 C1 model, see Section 4.4). a simultaneous twist about all three CC bonds provides the combination of perimeter and diagonal interactions (Figure 6.15) needed to reach a funnel ( S , S , degeneracy). Return to S,, through this geometry accounts for the formation of both cyclobutene and of bicyclo[l.l.O]butane (Gerhartz et al., 1977). The twist around the central CC bond was first invoked by Bigwood and BouC (1974). Both the 3 x 3 CI and the 2 x 2 VB models account for the presence of this pericyclic funnel in S,. Recent extensive calculations
Scheme 2
(Olivucci et al., 1993, 1994b; Celani et al., 1995) put these conjectures on firm ground and suggest very strongly that all the different photochemical transformations of butudienes listed in Scheme 2, including cis-trans isomerization, cyclization. bicyclization, and s-cis-s-trans-isomerization, involve passage through a common funnel located in a region of geometries characterized by a combination of peripheral and diagonal bonding (Figure 6.1%-f). Moreover, contrary to previous expectations, these calculations demonstrate that in the funnel region the S,-S,, gap not only is very small but vanishes altogether in a large region of geometries. The funnel therefore
Figure 6.15. n-Orbital interactions in butadiene; a) planar geometry, p AOs, b) high-
symmetry disrotatory pericyclic geometry, peripheral interactions along the perimeter, and ckf) geometries of two equivalent funnels; diagonal interactions are shown in c) and d); peripheral inreractions in e) and f).
I'li( )'l'OCHI:bI ICAL. KEAC'I'ION MODELS
338
corresponds to a true S,-S,, touching. that is, u conical intersection of the kind expected for a critically heterosymmetric biradicaloid geometry. For unsubstituted butadienes, the 1.3- and 2,4-diagonal interactions are equivalent (Figure 6 . 1 5 ~and d), as are the two possible movements of the terminal methylene groups above or below the molecular plane. Thus, the pericyclic funnel consists of a total of four equivalent conical intersection regions, communicating pairwise through the two high-symmetry disrotatory "pericyclic minima" investigated by van der Lugt and Oosterhoff (1969). The funnel can be reached from either of the vertically excited geometries, s-cis- and s-!runs-butadiene, after efficient crossing from the initially excited optically allowed B symmetry (S) state to the much more covalent A symmetry (D) state. Three points of minimal energy within the funnel, dubbed s-trans, s-cis, and central conical intersection, have been identified within the conical intersection region, as shown in Figure 6.16. However, the lowest-energy geometries within the funnel region are probably never reached in the S, state, since rapid return to S,,most likely ensues
6.2
8
PERlCYCLlC REACTIONS
339
once the edge of the funnel region has been reached. There are two stereochemically different pathways connecting the spectroscopic minima with the central funnel region. As expected from the classical Woodward-Hoffmann correlation diagrams (Section 4.2. I ) , the disrotatory pathway is energetically mare favorable than the conrotatory one. (Cf. Sections 7.1.3 and 7.5.1.) In the absence of molecular dynamics calculations, it is impossible to make quantitative statements about the probability with which the various product geometries will be reached on the S,, surface once the molecule has passed through the funnel. Bonding is possible both by increasing further the strength of the diagonal interaction to yield bicyclobutane (6 perturbation, x , in the branching space) and by increasing the strength of either pair of opposite-side pericyclic interactions to yield either cyclobutene or butadiene (y perturbation, x, in the branching space). In the latter case, an s-cis or an s-trans product may result, depending on the direction of twist along the central CC bond, and either a cis or a trans geometry may result at the double bonds in suitably labeled butadienes. This ambiguity accounts for the lack of stereospecificity in the ring opening of labeled cyclobutenes, which is believed to occur through the same funnel (Bernardi et al., 1992~).
6.2.2 Minima at Tight and Loose Geometries
5-
cisoid
central
s - transoid
Figure 6.16. The funnel region for butadiene isomerization: a) cross section of the excited-state potential energy surface for P = a:, where a, and a2 correspond to rotations about the two C==€ bonds of butadiene and /3 to rotation about the single bond and b) optimized geometries of the three points of minimal energy within the funnel (by permission from Olivucci et a]., 1993).
I n Section 4.1.3 the notion of loose. and tight biradicaloid geometries has been introduced using the H, molecule as an example. For a pericyclic reaction both of the cyclic conjugated orbitals, which are nonbonding for the biradicaloid geometry, are localized in the same region of space; the biradicaloid geometry of a pericyclic minimum is thus a tight one. In the case of ethylene dimerization it can be represented by formula 41, which is isoconjugate with cyclobutadiene and has two nonbonding orbitals. However, in addition there are a number of open-chain biradicaloid geometries (e-g., 42 and 43) for which the nonbonding orbitals (a and b) are localized in different regions of space, corresponding to loose biradicaloid geometries. They can be obtained from the pericyclic geometry by disrupting the cyclic conjugation through suitable bond stretching and twisting, and are favored in the So and T, states but not in S,. Here, the pericyclic minimum is evidently at tight geometries unless steric effects prevent it. In addition, substituents can stabilize other geometries, for example, by conjugative interactions. From the pericyclic minimum in S, all conformational changes will lead toward unfavorable loose geometries, thus requiring considerable energy. The molecule has little freedom for motion such as bond rotation and the reaction will be stereospecific. Besides, the duration of its sojourn in this minimum is undoubtedly very short, particularly if a diagonal distortion of the type discussed in Section 4.4.1 permits an approach to an area of S , S , touching, and in the limit. the molecule could pass through such a funnel during a
340
PHOTOCHEMICAL KEAC'fION MOI>EIS
single vibrational period. It will then find itself high on the S,)surface at an unfavorable biradicaloid geometry, and the steepest slope is likely to lead to a minimum that corresponds either to the starting material or to the pericyclic product. It is also conceivable that the relaxation on the So surface will actually first complete the diagonal bond, producing a biradical that can then close a ring ("crossed cycloaddition," cf. Section 7.4.2).
Similar but reversed arguments are also valid for the T I surface, if the pericyclic process involves an uncharged perimeter (i.e., if the number of electrons and AOs in the perimeter are equal). In this case loose geometries are more favorable, since there is now no need for a cyclic array of orbitals or any other rigid geometrical requirement. Biradicaloid minima at these geometries will typically allow considerable freedom of motion such as bond rotation, in contrast to the situation in S,. Also, return to S,, is spin forbidden and may be relatively slow, possibly permitting detection by direct observation. The ground state is then reached at a biradicaloid geometry, preferentially at geometries at which ISC is fast. (Cf. Section 4.3.4.) Subsequently, little if any additional stereochemical information is lost before a new bond is formed, or before the biradical can fall apart, as indicated in the following scheme:
butene should reach the So surface via the geometry of twisted butadiene and finally relax to ordinary ground-state butadiene in a nonstereospecific manner. Example 6.1 1:
The photochemistry of butadiene (Hammond et al.. 1964; Srinivasan. 1968) serves as a good exalnple for the demonstration that minima at tight and loose geometries lead to different products. On direct irradiation in eyclohex;ine or ether, a singlet reaction takes place, believed to pass through the diagonally distorted pericyclic ~ninimumat tight geometries, and gives cyclobutene and bicyclobutane, depending on the mode of relaxation on the S,, surface. (Cf. Section 4.4. I.) The sensitized reaction, however, proceeds via the triplet minimum at loose geometries and produces the addition products 44-46. (Cf. Section 7.4.1 .)
6.2.3 Exciplex Minima and Barriers
The electron-repulsion arguments for the preference of the triplet state for a loose biradicaloid geometry over a tight geometry, and for the resulting nonconcerted reaction course, only apply if the number of atomic orbitals involved in the pericyclic perimeter is the same as the number of electrons delocalized in the perimeter. In charged perimeters (cf. Figure 2.20c), the two differ, both the singlet and the triplet prefer the tight cyclic geometry, and both processes are expected to be concerted and stereospecific, as in the ring opening of oxirane or the ring closure in diphenylamine (Section 7.5.1, cf. Michl and BonaCiC-Kouteckv, 191N): Section 5.2.3.) As long as loose biradicaloid geometries are more favorable, pericyclic processes will not occur. For example, the T , surface in Figure 4.8 goes downhill from the cyclobutene to the butadiene geometry along the disrotatory reaction path. What is not shown in this two-dimensional diagram is its path going even further downhill in ;I direction toward the loose biradicaloid geometry of twisted butadiene. 'I'hus, after ISC. triplet-excited cyclo-
Photodimerizutions and photocycloadditions ;ire important examples of bimolecular reactions. For such reactions an encounter complex has to be first formed, which in the following will be treated as a supermolecule. Correlation diagrams can be constructed for this supermolecule in the usual manner and can be utilized to discuss the course of the reaction. This was demonstrated in Chapter 4 for the exploration of pericyclic minima using H, as an example. The formation of the encounter complex will in general be diffusion controlled. When two (or more) molecules D and A form a sufficiently stable aggregate in the ground state, this aggregate will be a new chemical species commonly referred to :is a CT con~plexand can be identified by its characteristic long-wavelength absorption. (Cf. Section 2.6.) Stable complexes can form in excited states even when this is not possible in the ground state. These exciplexes or excimers can freq~entlybe identified by their character-istic fluorescence. (Cf. Sections 5.4.2 and 5.4.3.) In view of the formation of an cxciplex. grc)uncl-sttire-forbidden photocyclo;tclditions typiciilly invol\,c :in intermediiite, E:'\ and return through ;I I'unnel.
342
PHoT( )CHEMI ['Al, REACTION MODELS
6.2
PERlCYCLlC REACTIONS
P*. A representative correlation diagram for this case is that of the photodimerization of anthracene shown in Figure 6.4. The flat excimer minimum (E*) will occur at larger intermolecular separations, with the molecules barely touching. It corresponds to the minimum in the surface of the S state in the case of H, shown in Figure 4.27. Its displacement to rectangular geometries is due to the existence of additional bonds not present in H,. The pericyclic funnel (P*), however, will still occur halfway through the reaction, that is, at a diagonally distorted geometry in which the old bonds are half broken and the new bonds half established. If the initial components A can approach in two different orientations, leading to two possible ground-state products By and B! (e.g., syn and anti, or head-to-tail and head-to-head, see Section 7.4.2), the following kinetic scheme will result:
Which of several possible products will be formed and whether photocycloaddition will occur at all will depend on the depth of the various minima on the excited surface S,, on the height of the barriers between them, and on the partitioning between starting materials and products upon return to S,,. Additional complications result if the diagonally distorted funnels give rise to products of crossed cycloadditions. Formation of a bound excimer ordinarily proceeds without an activation barrier other than that imposed by the need for diffusion to bring the reaction partners together in a medium of nonzero viscosity. It is then reasonable to assume preferential formation of the most stable excimer. This is also reasonable if an equilibrium is established among the various possible excimers. The relative stabilities of the excimers or exciplexes can be estimated from orbital interactions or from experimental data (exciton splitting and CT interactions, cf. Section 5.4.2). In the absence of steric complications, the syn head-to-head excimer is usually preferred. (Cf. Section 7.4.2.) According to Figure 6.17. the height of the barrier between the excimer minimum and the pericyclic funnel depends both on their depths and on the relative placement of the two excited-state surfaces S and D. The depth of the diagonally distorted pericyclic funnel is determined by the nature of the biradical; its dependence on molecular structure, on the head-to-head and head-to-tail orientation of the components, and on reaction medium can be discussed using the principles outlined in Section 4.4.1. A qualitative model for predicting the height of the barrier as a function of the location of the avoided crossing on the reaction coordinate was proposed by Caldwell(1980). It is based on an estimate of the crossing between tht corrtl:lfion l i n c of thr- t1011hlvexcited configur;~tionD and the ground
Figure 6.17. Schematic correlation diagram for photodimerization, showing the crossing ($) that determines the barrier between excimer minimum (E*) and pericyclic minimum (P*),an earlier crossing (---), and a later crossing (...) (by permission from Caldwell. 1980). Effects of a likely diagonal distortion are not shown.
state G with the correlation line of the singly excited state S. (See Figure 6.17.) Due to the interpretation of the D state as 'A* + 'B* (cf. Section 4.4.2). the energy gap AEE between the potential energy surfaces D and S at infinite separation of the reactants A and B may be expressed as
AE,
=
E,A
+ E!
-
EA s
(6.1) where EG 5 E! has been assumed for the singlet excitation energies. The initial slope with which the D state descends from the separated partners toward the pericyclic minimum can be estimated using PMO arguments for the interaction of two triplet states. Assuming that on1y hom*o-hom*o and LUMO-LUMO interactions are important, the energy change during an addition reaction is given by
where n and h are the HMO coefficients of the hom*o (subscript HO) and the LUMO (subscript LU) of molecule A and B, respectively. and the new bonds formed are between positions g in A and a in B. According to Figure 6.17 an early crossing with the S-S correlation line will bring about a low barrier, and therefore high reactivity, due to the large distance r, between the adducts at the crossing and the correspondingly small bonding interaction. The lateness of the crossing may be conveniently defined with respect to the Et level instead of the S-S line, since in that case the crossing con-
344
PHOTOCHEMICAL KEACTION MODELS
dition is simply AE,(r,) = -AE,. This condition allows combination of Equations (6.1) and (6.2) into the relationship
for the resonance integral B(I;.) of the newly formed bonds. The conditions for to be small, that is, for the reaction to start early and for the barrier to be low, are: low triplet energies, high singlet excitation energy E t , and large HMO coefficients of hom*o and LUMO at reacting positions. In agreement with this argument, photodimerizations of olefins are quite facile when a value IP(rc)l < 20 kcal/mol is obtained for the resonance integral from Equation (6.3), whereas for l@(r,-)(> 24 kcal/mol no dimerization has been observed. (Cf. Section 7.4.2.)
6.2.4 Normal and Abnormal Orbital Crossings The initial problem in constructing a correlation diagram for a thermally forbidden pericyclic reaction is in the identification of those orbitals that are most responsible for the ground-state energy barrier. In particular, it is necessary to actually identify the originally bonding orbital of the reactant, say @ i , which becomes antibonding (or at least nonbonding) in the product, and the originally antibonding reactant orbital, say which becomes bonding in the product, in order to completely specify the "characteristic configuration." The second step in the argument consists of an assignment of states of both the reactant and the product to specific configurations, and may require a consultation of spectroscopic data and some CI calculations. In many cases @, is the hom*o and @, the I,UMO ofthe reactant, and its they cross along the reaction coordinate, GI becomes the LUMO and @, the hom*o of the product. This kind of crossing will be referred to as "normal orbital crossing." Although the crossing may be avoided in systems of low symmetry, the arguments do not change. An "abnormal orbital crossing" occurs if at least one of the crossing orbitals and @, is neither the hom*o nor LUMO of the reactant. In this case, already at the stage of the configuration correlation diagram, more or less significant barriers have to be expected on the S, and TI surfaces. This becomes apparent from the following argument: The characteristic , of the reactant will correlate with the configuration @l,i configuration, @ of the product and this will be the lowest-energy configuration likely to remain at approximately constant energy along the path. If the crossing is of the normal type, this configuration most often predominates in the lowest excited singlet and triplet states S, and TI of both the reactant and the product, with the result that no barriers due to correlation are imposed starting on either side. If instead, either due to abnormal orbital crossing or due to configuration interaction effects, it predominates in one of the higher excited
+,,
states, other configurations necessarily represent the lowest excited state,
S, or TI. This then rises along the reaction path until it meets the state primarily represented by the characteristic configuration, and barriers are imposed on the excited-state surfaces. In the case of many aromatics, the lowest singlet excited state is not represented by the hom*o-LUMO excited configuration ('La) but rather by a +,, and Q i - I,A ('L,, cf. Section 2.2.2), and a barrier is exmixture of @ pected in the excited state S,. Since the difference in the S, and S, energies is usually quite small the resulting barrier will be small. (See Example 6.12.) In the triplet state, no barriers of this origin are expected since even in these cases the triplet represented by the hom*o-+LUMO excitation is lowest.
Example 6.12: Figure 6.18a gives a schematic representation of the orbital correlation diagram for the thermally forbidden conversion of one alternant hydrocarbon into another one. The following configuration correlations are easily verified from this diagram: ct)" @l-I,
-9
@';7;: +
+ -*
@;-I.
I-?'
@';z):
Figure 6.18. Excited-state barriers: a) orbital correlation diagram for a thermally forbidden conversion of an alternant hydrocarbon; b) the corresponding configuration ;~nd state correlation diagram for the case that the HOMChLUMO exciti~tiondoes not represent the longest-wavelength absorpt ion.
6.2
where (I,and cD' may refer either to reactant or to product. The ground state -I.. which i n i~roniatichydrocarbons as well as the mixture of cD,-_,. and represents the 'L, state. correlates with high-lying doubly excited states. Assuming that both in the reactant and the product the 'L, state is below the 'La state, which corresponds to the hom*o-+I,UMOexcit;~lion, the state correlation diagriun shown in 1:igurc 0.IXh is obtained, which irlrcady takes into consideration the avoided crossings between singlet states o f equal symmetry. From the derivation of this diagram it is evident that a barrier in the S, state is to be expected in all those cases i n which the lowest excited singlet state is not represented by the hom*o-2LUMO excitation, but is symmetrical with respect to the symmetry element that is being conserved during the correlation, like the ground state. A n example of the situation discussed here is given by the conversion of dewarnaphthalene into naphthalene. (Cf. Example 6.5.) The corresponding correlation diagram was already shown in Figure 6.8.
I'ERICYCLIC REACTIONS
347
that interacts strongly with the antibonding combination (a*)will be If #i is the hom*o of the original molecule and #, the LUMO. the orbital crossing is normal; otherwise it is abnormal. Example 6.13: 120rcyclobutenopheni~nthrcnc(50) two dill'crcnt photochcmici~lreaction pathways are conceivable (Michl, 1974b); the electrocyclic opening o f the cyclobutene ring to form 51 and the cycloreversion reaction to give phenanthrene and acetylene. The HMO coefficients of the 2 and 2' positions in biphenyl have opposite signs in the hom*o and equal signs in the LUMO. Thus, the hom*o
From the generitl form of the correlation diagram in Figure 6.18 it is evident that since barriers can be overcome with the aid of thermal energy the initial excitation need not be into that state that is represented by the characteristic configuration. In this case low temperature can prevent the reaction from taking place. Excitation of higher vibrational levels of one-andthe-same absorption band can enable the nuclei to move over the barrier in a hot reaction. In the case of an abnormal orbital crossing, a straightforward consideration of frontier orbitals would be misleading. Examples of high barriers due to abnormal orbital crossing have been observed in the electrocyclic ring opening of cyclobutenoacenaphthylene (47) and similar compounds (cf. Example 6.1 and Figure 6. I ) (Michl and Kolc, 1970; Meinwald et al., 1970). An example of a low barrier due to abnormal orbital crossing that can be overcome using thermal energy has been reported for the cycloreversion reaction of heptacyclene 48 to two acenaphthylene molecules (49) (Chu and Kearns, 1970).
In deciding whether a normi11 or an abnornial orbital crossing is to be expected for an electrocyclic process, the two-step procedure for constructing qualitative orbital correlation diagrams described in Section 4.2.2 has proven very useful. The least-bonding MO of the reactant that interacts strongly with the bonding combination (a)of the AOs on the two carbons originally joined by a single bond will be Q,,and the least-antibonding MO
Figure 6.19. State correlation diagram for the fragmentation o f cyclobutenophenanthrene. The straight arrows indicate absorption o f light of a given wavelength; the wavy lines indicate how the barriers i n S, (full lines) and in TI (broken lines) can be overcome when sufficient energy is available (by permission from Michl, 1974b).
348
PHOTOCHEMICAL KEACI'ION M0L)CIS cannot interact with the bonding combination and the L U M O cannot interact with the antibonding combination o f the AOs o f the original cr bond during the disrotatory opening of the cyclobutene ring in 50. This results in an abnormal orbital crossing. This reaction has i n fact so far not been detected. Experimentally known, however, is the cycloreversion to give phenanthrene and acetylene. This is assumed to be concerted in the singlet state and stepwise in the triplet state as indicated in Figure 6.19. The orbital correlation diagram for the concerted reaction has been derived in Example 4.5 and exhibits a normal orbital crossing. (See Figure 4.16.) Since the 'I,, band that is represented by the hom*o-+LUMO transition corresponds to an excitation from S,, to S2in biphenyl as well as i n phenanthrene, a barrier i n S, results, as shown for the general case in Figure 6.18. However, the triplet reaction is expected to be endothermic and the molecules will sooner or later again collect in the T, minimum for the starting geometry. Also indicated in Figure 6.19 is the nonconcerted p;~thwayfor which a barrier in 1'' is to he expected, since the TI is o f i t n-n* niiture while the chiu-;~cteristicconfiguration for reaching the openchain minimum i s of ;Ic ~ c P nature. (Cf. Figure 6.20.)
avoided and results in a barrier separating two minima on the TI surface (Figure 6.20). The minimum, which is essentially represented by the '(n,n*) configuration, is a spectroscopic minimum, into which initial excitation occurs, either by energy transfer (triplet sensitization) or indirectly by intersystem crossing (ISC) from the singlet manifold. The minimum represented by the '(0,d) configuration, on the other hand, is a reactive minimum, from which the actual dissociation takes place. Similarly, the characteristic state for an a cleavage in a ketone triplet is of (0.8)nature, whereas the excitation is into the (n,n*) state. The spectroscopic minimum preserves the excited molecule until it can escape toward the reactive minimum. Its role is particularly crucial in bimolecular processes, where this escape has to wait for diffusion to introduce a reaction partner. A high barrier separating the spectroscopic minimum and the reactive minimum is to be expected if the orbitals @, and which represent the characteristic configuration, do not interact with the orbitals involved in the electronic transition into the lowest excited state; the crossing between the potential energy curve that goes up in energy along the reaction coordinate and the one that comes down in energy would then not be avoided. Examples of this situation are the dissociation of an aromatic CH bond in toluene or the cleavage of a CC bond in a ketone that is rather far away from the carbonyl group. No doubt such reactions will show minima in the S, or TI surface. But they are separated by unsurmountable barriers, so these reactions cannot be observed.
6.3 Nonconcerted Photoreactions 6.3.1 Potential Energy Surfaces for Nonconcerted Reactions Most known photochemical processes are not pericyclic reactions. Even in many of these cases correlation diagrams can be helpful in estimating the location of minima and barriers on excited-state surfaces. (Cf. Section 4.2.2.) The derivation of these correlation diagrams, however, is often more difficult, not only because of lack of symmetry, but also because it may be difficult to identify any one excited state as the characteristic state, particularly in large molecules. For example, many of the cwd" excited states of toluene will have some contribution from the v r r * bond orbital excitation in one of the three C-H bonds in the methyl group. Often, however, it is sufficient to distinguish between rr* and n* orbitals in order to decide whether a barrier is likely to occur. For instance, the T, surface of toluene along the path of nuclear geometries that leads to dissociation to C,H,CH2. + H- can be viewed as originating from interaction of a locally excited '(n,n*)configuration of the benzene chromophore and a locally excited '(0.d) configuration of the CH bond. The former is of lower energy at the initial geometries, as is the latter at the final geometries; somewhere along the way they intend to cross, but the crossing is
Figure 6.20. bhcrgies of zclcctcd st;llcb tlur-ing dis\oci;~tionof ;I h c r ~ ~ y l('ti i c hol~d in toluene as ;Ifiinction of thc rc;lction coonlin;~tc(hy pcrnlis\io~lI.roln Michl. 1073;1).
6.3
However, when the orbitals of the characteristic configuration can interact with the orbitals involved in excitation, as is the case for the cleavage of the benzylic C-H bond in toluene or the a CC bond in a ketone, the crossing will be avoided and the barrier will be lowered depending on the interaction. The magnitude of the interaction is rather difficult to predict without a detailed calculation. In the particular case of the photolytic dissociation of the benzylic CH bond in toluene, the barrier separating the two minima is so high that the reaction proceeds on absorption of a single photon only if light of sufficiently short wavelengths is used. With light of longer wavelengths another photon is needed to overcome the barrier. (Cf. Johnson and Albrecht, 1968 as well as Michl, 1974a.) Example 6.14: Except for the relief of ring strain, similar harrier heights are to be expected for the triplet-sensitized photochemical ring opening of benzocyclobutene (52) to form o-xylylene (53) and for the dissociation of the benzylic C--C bond in ethylbenzene leading to a benzyl and methyl radical. Since the energy of the characteristic configuration may be lowered by additional substituents, this may favor the interaction with the orbitals involved in the initial excitation and thus reduce the barrier height. In fact, for a,a,a'af-tetraphenylbenzocyclobutene (54)a photochemical ring opening has been observed. whereas under similar conditions of low-temperature irradiation benzocyclobutene is inert (Quinkert et al., 1969: Flynn and Michl, 1974).
NONCONCEKTED PHOTOREACTIONS
351
to be expected on the basis of orbital symmetry, the interaction between these configurations is likely to be strong, so the barrier resulting from the avoided crossing may not be high. The results of ab initio calculations for hydrogen abstraction by ketones according to
represented in Figure 6.21, in fact show small barriers in the '(n,n*) and 3(n.n*) excited states, which indicate that the MOs apparently "remember" the natural or intended correlation. (Cf. Figure 4.18.) Another barrier on the T, surface is observed for hydrogen abstraction by ketones whose 3(n,n*) state is of lower energy than the '(n,n*) state. From Figure 4.17 it is seen that the >(n,n*)state goes up in energy while the '(n,n*) state comes down in energy along the reaction coordinate or at least remains more or less constant (Figure 6.21), and a crossing of the corresponding potential energy surfaces will occur at a geometry intermediate between that of the reactant and a biradical. For nonplanar arrangements of the nuclei this crossing will be avoided and will produce a barrier. This amounts to 5 kcallmol in the case of the hydrogen abstraction in the naphthyl
In the case of dissociation of the various methyl-substituted anthracenes to form the corresponding anthrylmethyl radicals the relative magnitude of the interaction between the a orbitals of the characteristic configuration and the n and n* orbitals involved in the excitation may be estimated quite easily. Assuming that the interaction is proportional to the square $,of the LUMO coefficient of the carbon atom adjacent to the methyl group, the lowest barrier is expected for the Pmethyl derivative and the highest one for the 2-methyl derivative, as may be seen from the squared coefficients given in formula 55.
1.3
Bo"
50 H
\
151
122~90"
'GJ
In Section 4.2.2 it has been shown that due to intended or natural orbital correlations, crossings may occur in the configuration correlation diagram that are avoided in the state correlation diagram. If no avoided crossing is
109 ,H H-C 109 5O vH VH
H
Figure 6.21. Sti~tecor1~1:rtion diagram for the photochemical hydrogen abstraction, as calculated for the system formirldehyde + methane (by permission from Devaquet et ;)I.. 1978).
352
PHOTOCHEMICAL REAC'TION M0L)EU;
ketone 56 to form 57, an example in which the '(n,n*) is higher in energy by 9 kcallmol than the '(n,n*) state. Besides, this is an abstraction of the 6-H atom that makes a nonplanar reaction path very likely (De Boer et al., 1973).
The picture becomes somewhat more complex if the a cleavage of ketones is considered.
If all atoms involved in the reaction lie in the same plane, the unpaired electron of the acyl radical may be either in an orbital that is symmetric with respect to this plane or in an orbital that is irntisymmctric-that is, either in 21 a or in a n orbital, whereas only a cr orbital is avirilable for the unpaired electron of radical K. Instead of just one singlet and one triplet covalent biradicaloid structure (Figure 4.5). there are now two of each. which may be denoted as I - % , , ,and , , '.'B,,,,,, respectively. Similarly, there are also different zwitterionic structures to be expected. The increase in complexity and the number of states that results from the presence of more than two active orbitals on the atoms of a dissociating bond has been formalized and used for the development of a classification scheme for photochemical reactions ("topicity"), as is outlined in more detail in Section 6.3.3. As shown in Figure 6.22 in the example of the formyl radical, the a acyl radical prefers a bent geometry with the unpaired electron in an approxi-
1
1SO0
ii HCO
Figure 6.22. Qualitative representation of the energy o f the o and n formyl radical as a function of the HCO angle.
mately sp"ybrid AO that has some s character and thus is energetically more favorable than a pure p AO. In contrast, the n acyl radical prefers a linear geometry, since it has its unpaired electron in the n* orbital; the p,. A 0 starts empty and overlaps with the doubly occupied n,, A 0 to form a doubly occupied MO. Since at linear geometries three electrons in one n system are equivalent to three electrons in two p AOs orthogonal to this n system, the aand the n acyl radical states have the same energy, that is, are degenerate. In discussing the rciiction it is helpfill to use bond dissociation and bond angle variation in the reaction product as independent coordinates. In this way the potential energy surfaces of the type shown in Figure 6.23 for the a-cleavage reaction of formaldehyde are obtained. Such a three-dimensional diagram is difficult to construct without calculations, and an initial analysis can be based on a conideration of the reactions to linear and bent products separately as shown in Figure 6.24. Sometimes the two reaction paths are plotted superimposed, with only the a states of the acyl radical shown for the bent species and only the n states shown for the linear species (Salem, 1982). but such plots can be easily mistrnderstood and we avoid them. We shall return to these issues in connection with applications of these diagrams
Figure 6.23. Potential energy surfaces for the a-cleavage reaction of formaldehyde as a function of the CH, distance and the OCH, angle; formaldehyde states are in the front left corner, correlation to the front right corner corresponds to cleavage of the a bond with bond angles kept constant, and correlation from here to the right rear corner corresponds to linearization of the formyl radical. Correlation between states of formaldehyde and of the linear biradical results from a cross section through these surfaces approximately long the diagonal from the front left to the rear right corner (by permission from Reinsch et al., 1987).
6.3
Figure 6.24. State correlation diagram for the n cleavage of saturated ketones. The path to a bent acyl radical is shown on the right, that to 21 linear acyl radical to the left. The right-hand part of the diagram corresponds to the front face of the threedimensional representation in Figure 6.23; the left-hand part corresponds to the cross section along the diagonal in Figure 6.23.
180
aOCH 1°1
120 110
300
110
Rcn Ipml
300
NONCONCERTED PHOTOREACTIONS
355
Figure 6.26. Potential energy curves for the a-cleavage reaction of formaldehyde. Part a) represents a cross section through the surfaces of Figure 6.23 approximately along the QOCH = 130" line, forming a bent acyl radical, and part b) a cross section along the minimum energy path on the (n,n*)excited surfaces S, and T;,,forming a linear acyl radical.
in Section 7.2.1. It is important to remember that the two parts of Figure 6.24 do not represent different reactions but rather correspond to different reaction paths on the same potential energy surfaces. This becomes particularly evident from the contour diagrams shown in Figure 6.25. For the potential energy surfaces of Figure 6.23 these diagrams show the reaction path to the bent acyl radical, which is the minimum energy path in the ground state and in the -'(n,n*) state (Soand T,), and the alternative reaction path to the linear acyl radical, which is more favorable in (n,n*) states ( S t ,and T,). Cross sections through the surfaces of Figure 6.23 along these reaction paths are depicted in Figure 6.26. Figure 6.26a corresponds to the right-hand side of the correlation diagram shown in Figure 6.24, and Figure 6.26b to the lefthand side.
6.3.2 Salem Diagrams
Contour diagrams of the potential energy surfaces for the a-cleavage reaction of formaldehyde shown in Figure 6.23. The broken lines indicate the reaction paths to the bent ( A ) and the linear ( B ) acyl radical ( b y permission from Reinsch et ;)I.. 1987). Figure 6.25.
..
In f;ivorable cases state correlation diagrams of the type shown in Figure 6.2 1 and Figure 6.24 may be obtained simply from the leading VB structures of the reactant and the biradici~loidproduct. The molecular plane is chosen as the symmetry elenlent-that is. one considers coplan;u- reactions, and any deviation from coplanarity may possibly be taken into account as an additional perturbation. As tr orbitals are symmetric and n orbitals antisymmetric with respect to the molecular plane, symmetries of the various states may be obtained by simply counting the number of a and n electrons in a VB structure. A single structure is in general not sufficient to describe an electronic state accurately. but all contributing V B structures have the same symmetry, so the inclusion of just some of them is sufficient to establish a correlation diagram. The use of VB structures in the consideration of photochemical reaction paths was pioneered by Zimmerman (1969). The use of
PHOTOCHEMICAL REACTION M0L)EL.S
Figure 6.27. Salem diagram for hydrogen abstraction by a carbonyl compound. The biradicaloid product states are denoted by '.'B (dot-dot, covalent)and Zl,z(hole-pair, zwitterionic),respectively (by permission from Dauben et at., 1975).
VB correlation diagrams for this purpose was developed by Salem (1974) and collaborators (Dauben et al., 1979, and they are commonly referred to as Salem diugrtrnzs. The photochemical hydrogen abstraction by carbonyl compounds, which has already been discussed in the last section, will be used to illustrate the procedure. The left-hand side of Figure 6.27 shows the dominant VB structures of the ground state and of the excited (n,n*) and (n,n*) states of the reactants. Only the electrons directly involved in the reaction are considered. In the ground and the (n,n*) states, the numbers of a and n electrons are 4 and 2, respectively, and these states are symmetric relative to reflection in the mirror plane. In the (n,n*) states, both numbers are equal to 3. and these states are antisymmetric. The right-hand side shows the analogous VR structures for the various states of the biradicaloid system formed as a primary product. Dot-dot (covalent) states are denoted by % and ' B and holepair (zwitterionic) states by 2, and Z2, using the nomenclature introduced in Section 4.3.1. The energy of these states depends on the nature of radical centers A and B as has also been discussed in Section 4.3.1. In the present case both radical centers are C atoms, so the dot-dot states are clearly lower in energy than the charge-separated hole-pair ones. Since the electron count gives 30, 3n for the dot-dot and 40, 2n for the hole-pair states, correlation lines can be drawn immediately as shown in Figure 6.27. Thus, a diagram is obtained that is basically identical to the results represented in Figures 4.18 and 6.21 except for the barriers revealed by natural orbital correlations.
6.3.3 Topicity The differences between the correlation diagrams for the dissociation of the H-H or Si-Si single bond (Figure 4.5) and for the dissociation of the C--C
single bond in the Norrish type I process (Figure 6.24) are striking. In particular, in the former case, only the triplet state is dissociative, while in the latter case, a singlet state is as well. We have already seen in Section 4.3.3 that the correlation diagram for the dissociation of a single bond can change dramatically when the electronegativities of its termini begin to differ significantly. Now, however., both bonds to be cleaved are relatively nonpolar. The contrast is clearly not due to the differences in the properties of silicon and carbon, either. Rather, it is due to the fact that in one case a double bond is present at one of the termini of the bond that is being cleaved. This causes an increase in the number of low-energy states and changes the correlation diagram. This role of the additional low-energy states has been formalized in the concept of topicity (Salem, 1974; Dauben et al., 1975), which was generalized by its originators well beyond simple bond-dissociation processes. We believe th:lt this concept is most powerful and unambiguous in the case of reactions in which only one bond dissociates, and shall treat this situation only. Simple bond-dissociation reactions are classified as bitopic, tritopic, tetratopic, etc., according to the total number of active orbitals at the two terminal atoms of the bond. Active orbitals at an atom (AOs or their hybrids) are those whose occupiincy is not always the same in all V B structures that are important for the description of the low-energy states that enter the correlation diagram for the bond-dissociation process. These are obviously the two orbitals needed to describe the bond to be cleaved, so the lowest possible topicity number ib two, but possibly also other orbitals located at the two atoms: those containing lone pairs out of which excitation is facile (e.g., oxygen 2p), those that are empty such that excitation into them is facile (e.g., boron 2p), and those participating in multiple bonds adjacent to the bond being cleaved (e.g., carbonyl in the rx-cleavage reaction). I t is now clear that the dissociation 01' the H-H bond and of the Si-Si bonds in saturated oligosilanes are bitopic processes. The Norrish type I a cleavage is an example of a tritopic process, sometimes subclassified further as a a(a,n) tritopic process to specify that one terminal carries an active orbital of a symmetry and the other two active orbitals, one of o and one of n symmetry. Other examples of tritopic reactions are the dissociation of the C-N bond in a saturated amine and the dissociation of the C--0 bond in a saturated alcohol (the oxygen 2s orbital is very low in energy, doubly occupied in all important stiltes, and not counted as active). Figure 6.28 shows the correlation diagram for the dissociation of the C--0 bond in methanol. Diagrams for such simple dissociations are not necessarily of use in themselves, not only since these reactions require high-energy excitation and are relatively rarely investi~dtedby organic photochemists. but primarily because the lowest excited states of fully saturated organic molecules are usually of Rydberg ch:tracter while only valence states are shown in the diagrams. Thus. at least one side of the dii~gram is quite unrealistic. Experimental (Keller el ;ti.. 1992; .lensen et a1.. 1993) and theol-elical (Y;II--
Supplemental Reading
75';)
\C S m G , d:
\"
loo*
"g2,
P3 q)
I k ~ u h c n .W.G.. S;~lcni.I... '1'11rro. N.J. (1975). " A CI ; ~.s. s' "~ l l c a t i oonf I'horochemical Re;~clions," Ac.c.. ('hc,rn. Rc,s. 8 . 41. Dougherty. R.C. (1971). "A Perturbation Molecular Orbital Treatment of Photochemical Reactivity. The Nonconservation o f Orhital Symmetry in I'hotochcmic:~l Pericyclic Re. 93. 7187. actions." J. Am. C l i c ~ i i.Yo(,. Fijrster. Th. (1970). "Diabatic and Adiabatic Processes in Photochemistry," Pure Appl. Clrc~rn.24, 443. Gerhartz, W., Poshusta. R.D., Michl, J. (1976). "Excited Potential Energy Hypersurfaces for H, at Trapezoidal Geometries. Relation to Photochemical 2s + 2s Processes," J. A m . ('ltc,rn. Soc.. 98. 6427. Michl, J. (1974). "Physical I h s i s of Qualitative MO Arguments in Organic I'hotochemistry," Fortsc.l~r.Clrenr. Forsclr. 46, I. Michl, J., BonaCiC-Kouteckg. V. (19901, Elc~c.trotiicAspc,c~tsof Orgtrnic. Phofochemistry; Wiley: New York. Salem. 1,. (1976). "Theory of Photochemical Reactions." Scic,ricx, 191. 822. Salem. L. (1982). E1c~c.trotr.sin Clrc~~nic.ul Recrc~1ion.s:First Princ.iplc,.s; Wiley: New York.
Figure 6.28. VB structure (dotted lines) and state (full lines) correlation diagram for the dissociation of the C--0 bond in methanol (by permission from Michl and BonaEiC-Kouteckv, 1990).
kony, 1994) investigations of Rydberg-valence interitctions in the processes CH,SH -+ CH, + SH and CH,SH -+ CH,S + H exemplify this situation. Although diagrams such as that of Figure 6.28 are useless for the understanding of molecules such as CH,OH or CH,SH themselves, they are still useful for the derivation of orbital correlation diagrams for bond dissociation reactions of molecules with unsaturated chromophores, using the stepwise procedure described in Section 4.2.2. Thus, Figure 6.28 forms the basis for the understanding of reactions such as the photo-Fries and photo-Claisen rearrangements discussed in Section 7.2.1; it applies to the ring opening reactions of oxiranes, etc. Examples of tetratopic reactions are the C-N bond dissociation in azo compounds, discussed in Section 7.2.2, C-X bond dissociation in alkyl halides, and the 0--0 bond dissociation in peroxides. Examples of pentatopic reactions are the dissociation of the C-X bond in vinyl halides, of the C=C bond in ketenes, and of the C=N bond in diazoalkanes. An example of a hexatopic bond dissociation is the fragmentation of an alkyl azide to a nitrene and Nz.A verification of the topicity rules at a semiempirical level was reported (Evleth and Kassab, 1978). and a detailed description of the electronic structure aspects of bond dissociations characterized by various topicity numbers, with references to the original literature. has appeared recently (Michl and BonatiC-Kouteckjj, 1990).
Turro, N.J.. McVey, J.. Ramamurthy, V., Cherry, W., Farneth, W. (1978), "The Effect of Wavelength on Organic Photoreactions in Solution. Reactions from Upper Excited States," Chc,m. Rc,v. 78. 125. Zimmerman, H.E. (19761, "Mechanistic and Explaratory Organic Photochemistry," Science 191, 523. Zimmerman. H . E . (1982). "Some Theoretical Aspects of Organic Photochemistry," Acc. Chem. Res. 15, 312.
CHAPTER
Organic Photochemistry
Examples of photoreactions may be found among nearly all classes of organic compounds. From a synthetic point of view a classification by chromophore into the photochemistry of carbonyl compounds, enones, alkenes, aromatic compounds, etc., or by reaction type into photochemical oxidations and reductions, eliminations, additions, substitutions, etc., might be useful. However, photoreactions of quite different compounds can be based on a common reaction mechanism, and often the same theoretical model can be used to describe different reactions. Thus, theoretical arguments may imply a rather different classil'ication, based, for instance, on the type of excitedstate minimum responsible for the reaction, on the number and arrangement of centers in the reaction complex, or on the number of active orbitals per center. (Cf. Michl and BonaCik-Kouteckq, 1990.) Since it is not the objective of this chapter to give either a complete review of all organic photoreactions or an exhaustive account of the applicability of the various theoretical models, neither of these classifications is followed strictly. Instead, the interpretation of experimental data by means of theoretical models will be discussed for selected examples from different classes of compounds or reaction types in order to elucidate the influence of molecular structure and reaction variables on the course of a photochemical reaction.
7.1
7.
CIS-TRANS ISOMERIZATION OF DOUBLE BONDS
Cis-trans Isomerization of Double Bonds
Cis-trans photoisomerizations have been studied in great detail and can serve as an instructive example for the use of state correlation diagrams in discussing photochemical reactions. They have been observed for olefins, a7c*r;iethines,and azo compounds.
71. I .1 Mechanisms of cis-trans Isomerization Po4srlde cis-trans isomerization mechanisms of isolated double bonds can using ethylene as an example. A state correlation diagram for is shown in Figure 4.6. (Cf. also Figure 2.2.) This diagram the strongly avoided crossing that gives rise to a biradicaloid minlrnum in the S, surface (Cf. Section 4.3.2). This is surely not a minimum with I.-spect to distortions other than pure twisting (see below). The diagram also ci'splays the biradicaloid minimum in the TI surface, where T I and S, are ~ieitrlydegenerate. rding to the MO model, both the hom*o and the LUMO are singly in the T, state by electrons of parallel spin. The minimum in this state {hen arises from the fact that the destabilizing effect of the n* MO (LUYO) is stronger than the stabilizing effect of the n MO (hom*o) if overlap oflthe AOs is taken into account. For the biradicaloid geometry, however, t)ie overlap vanishes and its destabilizing effect disappears. (Cf. Section 4.3.2.) From the state energies of a hom*osymmetric biradicaloid shown in Figure 4.20, the energy difference between the S, and TI states is expected to be copstant; Figure 4.6 indicates that this is true to a good approximation. Since forla twisting of the double bond the internuclear distance is more or less fixed by the abond. there are in this case no lome geometries that could favor the 'I", state. Similar arguments apply to the S, state. Since twisted ethylene is a hom*osymmettic biradicaloid, this state is described by configurations that correspond tolcharge-separated structures in the VB picture, such as
t
\0 /
1
c-c;
-
M
\8
,c-c;
0 ,
(cf. Sectio 4.3.1 ). It has been proposed that charge separations of this type, which ma bring about very polar excited states if the symmetry is perturbed (srqdden pollrrizcrtion effect, cf. Section 4.3.3). may also be important in photochemical reactions (Bruckmann and Salenl, 1976). but no experimental evidence supports this so far. At thejlevel of the 3 x 3 CI model, heterosymmetric perturbations 6 that introducd an energy difference between the two localized nonbonding orbitals of the twisted double bond may reduce the s,-s,,gap to zero and provide a very efficient relaxation path. (Cf. Section 4.3.3.) Calculations show that pyramidalization of one of the methylene groups alone does not represent a
Figure 7.1. Geometry of the S,-So conical intersection of ethylene calculated at the CASSCF level, indicating the possibility of cis-trans isomerization and [ I ,2] hydrogen shift; a) side view, b) Newman projection with localized nonbonding orbitals (Freund and Klessinger, 1995).
perturbation S sufficient for reaching a critically heterosymmetric biradicaloid geometry (real conical intersection). However, this is accomplished by an additional distortion of one of the CH bonds toward the other carbon atom (Ohmine, 1985; Michl and BonaCiC-Kouteckq, 1990). The geometry of the resulting conical intersection calculated at the CASSCF level is shown in Figure 7.1. Thus it is most likely that ethylene cis-trans isomerization and 11.21 hydrogen shift occur via the same funnel. (Cf. Section 6.2.1.) Two essentially different cis-trans isomerization mechanisms may be derived from the state correlation diagram of ethylene: I. Through absorption of a photon, the molecule reaches the biradicaloid minimuni in S, via one of the excited singlet states; return from this minimum to the ground state S, at the pericyclic funnel region close to the orthogonal geometry can then lead to either of the planar geometries. Heterosymmetric perturbations that introduce an energy difference between the two localized nonbonding orbitals of the twisted form reduce the So-S, gap, possibly to zero, and provide very efficient relaxation paths. (Cf. Section 4.3.3.) 2. Energy transfer from a triplet sensitizer produces the TI state; return to the ground state So occurs again for a twisted geometry. T, has a minimum near 0 = 90". but spin-orbit coupling is inefficient at this geometry (cf. Section 4.3.4). and the return to S,, most likely occurs at a smaller twist angle. The advantage of sensitization is that it is readily applicable to monoolefins, which require very high energy radiation for singlet excitation. Furthermore, competing reactions of the S, state such as valence isomerizations, hydrogen shifts, and fragmentations are avoided. Given suitable reaction partners, the TI state of the olefin may also be reached via an exciplex and a radical ion pair (see Section 7.6. I), which may undergo ISC and subsequent reverse electron transfer (Roth and Schilling, 1980).
'
364
ORGANIC PHOTOCHEMISI'UY
Additional mechanisms that have been established may be termed photocatalytic. One of these is the Schenck mechanism, which involves an addition of the sensitizer (Sens) to the double bond and formation of a biradicaloid intermediate that is free to rotate about the double bond and subsequently collapses to the sensitizer and olefin: Sens
7.1
365
CIS-TRANS ISOMERIZATION OF DOUBLE BONDS
In protic solvents, strained truns-cycloalkenes such as methylcyclohexene (3) give an adduct according to the following scheme (Kropp et al., 1973):
hv + senso
Obviously, the reaction can proceed according to this mechanism even if the triplet energy of the sensitizer is below that of the olefin (Schenck and Steinmetz, 1962). Instead of the sensitizer, a photolytically generated halogen atom can also add to the olefin and produce a radical, which may then rotate about the double bond:
7.1.2 Olefins Cis-trans isomerization of simple nonconjugated olefins is difficult to achieve by direct irradiation because such high energy is required for singlet excitation (A < 200 nm). Cis-trans isomerization via the S, state has been observed for 2-butene (1) (Yamazaki and CvetanoviC, 1969); at higher concentrations a photochemical [2 + 21 cycloaddition comes into play. (See Section 7.4.1 .) Cis-trans isomerizations of cycloalkenes clearly demonstrate the influence of ring strain: trans-cycloalkenes are impossible to make with unsaturated small rings; with six- and seven-membered rings the trans products can be detected at low temperatures (cf. Bonneau et al., 1976; Wallraff and Michl, 1986; Squillacote et al., 1989), whereas larger rings give transcycloalkenes that are stable at room temperature. trtrns-Cyclooctene (2) has been obtained on direct as well as on sensitized irradiation (Inoue et al., 1977). Enantioselective cis-trans isomerization with very high optical purities (64%) has been obtained for cyclooctene (2) by triplex-forming sensitizers (Inoue et al., 1993).
From the state correlation diagram of cis-trans isomerization in Figure 4.6, it is seen that the crossing is strongly avoided with a large energy difference A E between S,, and S, for the biradicaloid geometry ( 8 = 90"). The energy gap AE will be reduced at less symmetrical geometries, where the electronegativities of the two termini of the double bond will differ. (Cf. Section 4.3.3.) Still, in simple olefins the result could well be a biradicaloid minimum rather than a funnel, and this suggests the occurrence of an intermediate. In the earlier literature this intermediate is referred to as phantom stcrte 'P* (Saltiel et al., 1973). Therefore the fraction P of molecule$ in the biradicaloid minimum that reach the ground state of the truns-olefin is independent of the initial isomer. At a given wavelength A the cis and the trans isomers c and t are formed in a constant ratio, and after a certain peiiod of time a photo.sttrlioncrr~~ sttrte (PSS) is reached. The composition of tue photostationary state is given by
where &(A)is the extinction coefficient at wavelength A and Q, is the iuantum yield of the photochernical conversion. (See Example 7.1 .) If thdre are no competing reactions one has 4,,, = /? and
Complete conversion rnay be achieved only if it is possible to chqose solvent and excitation wavelength such that only one of the isomers absorbs and E = 0 for the other one. Example 7.1: The rate of a simple photoreaction of species X is given by
! I
I
where the fraction 4 is obtained by multiplying the total absorbed iadiation intensity &, by the ratio A,(A)IA,,,(A) of the absorbances at the corre onding wavelength 1. Photostationary equilibrium is reached if the rates of is-trans and trans-cis conversion are equal, and for direct irnidiation
"P I
7.1
From Lambert-Beer's law, A&)
=
&,(A)[X]d,and one has
@,-,~,.(A)[cl= @,,,.~,(A)[tl which then yields Equation (7.1). In the case of a sensitized reaction. the rate of formation of triplet-excited molecules 'X* is proportional to the product k:.,['S*I[X] of the rate constant for energy transfer and the concentrations of sensitizer and molecule X in the ground state, and kg, and k;, replace E, and E, in the foregoing equations. Sensitization of alkenes by carbonyl compounds, which is likely to proceed via an exciplex, can be accompanied by oxetane formation. (See Section 7.4.4.)
7.1.3 Dienes and Trienes The singlet cis-trans isomerization mechanisms of conjugated olefins appear to be quite different from those of isolated olefins. Direct irradiation of butadienes is known to yield a mixture of cis-trans isomerization and cyclization photoproducts. (Squillacote and Semple, 1990; Leigh, 1993) All these photochemical transformations are now believed to involve passage through a common funnel. (Cf. Scheme 2 in Section 6.2.1.2.) In agreement with expectations from correlation diagrams (Figure 4.6; see also Figure 7.3 below), the initially excited optically allowed state of B symmetry (S) is depopulated in 10 fs owing to fast internal conversion to the nearby state of A symmetry (D) (Trulson and Mathies, 1990). Calculations (Olivucci et al., 1993; Celani et al., 1995) in fact yield a IB,2A, conical intersection near the bottom of the B state minimum. The energy of the 2A, state drops rapidly as one proceeds further toward the pericyclic funnel at a disrotatory geometry in which all three CC bonds are considerably twisted, and pericyclic as well as diagonal interactions are present (Section 6.2.1 and Figures 6.15 and 6.16). The calculations suggest that after crossing from the B state it might be possible, but 10 kcallmol more costly energetically, to reach the funnel region on the A surface by an initially conrotatory motion followed by a backward twist of one of the double bonds. This path, however, appears to have no chance to compete with the barrierless disrotatory approach toward the funnel region described earlier. It is probably reasonable to assume that the excited-state motion is initially dominated by the slope of the B and A surfaces, which points in the disrotatory way and toward the diagonally-bonded pericyclic funnel, and to assume that the acquired momentum is kept after the jump to S,. This would point in the direction of bicyclobutane. Unless the surface jump occurs right at the cone tip, it generates an additional momentum in the x, direction, that is, along the y perturbation coordinate, toward cyclobutene and the original as well as cis-trans isomerized hutadiene. Whether further facile motions on
-
CIS-TRANS ISOMERIZATION OF DOUBLE BONDS
367
the So surface follow, in particular s-cis-s-trans interconversion, before all excess vibrational energy is lost, is hard to tell. The case of s-trans-butadiene has been investigated computationally in less detail. The disrotatory pathway is again favored and enters the pericyclic funnel region of conical intersections at a much larger twist angle along the central C-C bond. This pathway is again barrierless and is steeper than the s-cis pathway. After the jump to S,, diagonal bonding to bicyclobutane appears more likely and peripheral bonding to cyclobutene quite unlikely. It seems probable that the molecule will enter the So state with a higher velocity along the s-trans isomerization path, suggesting a higher efficiency for the s-transjs-cis than for the s-cisjs-trans isomerization. In fact, it is known that in 2,3-dimethylbutadiene the s-transjs-cis isomerization is 10 times more efficient than the s-cis-s-trans isomerization (Squillacote and Semple, 1990). Triplet-sensitized cis-trans isomerization is frequently more efficient. Thus, the direct excitation of 1,3-pentadiene (piperylene), which has been studied in great detail, results in cis-trans isomerization with very low quantum yields (a,.,, = 0.09, a,,. = 0.01) and very small quantities of dimethylcyclopropene as a side product, whereas when benzophenone is used as sensitizer, a,,, = 0.55 and a ,,. = 0.44. This efficient quenching of higher-energy triplet states by conjugated dienes is utilized in mechanistic studies for identifying the excited state responsible for a photochemical reaction. Computational results similar to those for butadiene have been obtained for the singlet excited-state cis-trans isomerization of c-is-hexatriene around its single and double CC bonds (Olivucci et al., 1994a). For each of these torsional modes, a reaction pathway has been found on the D excited state (A symmetry). Both lead over a small barrier to funnels that are reached already at about 60" twist angles. In fact, the fluorescence excitation spectrum of very cold cis-hexatriene in supersonic jet is consistent with the existence of two very fast radiationless decay channels with a barrier below I kcallmol (Petek et al., 1992). As in butadiene, the funnels are located at biradicaloid geometries with pericyclic and diagonal interactions, as indicated in Figure 7.2, which also schematizes some of the possible bond-forming processes that are expected to occur along different ground-state relaxation paths after the S,-So return. The reaction pathways both for single- and double-bond isomerization enter the funnel region at less than one-third of the way toward the products, suggesting that the majority of excited-state molecules should decay back to the ground-state reactant. In addition, the excited-state barrier for singlebond isomerization is smaller than that for double-bond isomerization, reducing the quantum yield for cis-trans isomerization (since the product of s-cis-s-trans isomerization is just a different conformer, not a new cis-trans isomer of the reactant). Thus the low experimental value of = 0.034 reported for the quantum yield of trans-hexatriene from cis-hexatriene (Jacobs
ORGANIC PHOTOCHEMISTRY
7.1
CIS-TRANS ISOMERIZATION OF DOUBLE BONDS
7.1.4 Stilbene The cis-trans isomerizittion of stilbene (4) has been thoroughly studied. The potential-energy diagram for a rotation about the CC double bond that preserves a twofold symmetry axis is shown in Figure 7.3.
Figure 7.2. Schematic representation of the conical intersection structure for cistrans isomerization of cis-hexatriene and of some of the possible bond-making processes that might occur along different ground-state relaxation paths (by permission from Olivucci et al., 19944.
and Havinga, 1979) is easily rationalized. 2-Vinylbicyclo[l . l .O]butane (Datta et al., 1971) and bicyclo[3.l.O]hexene (Jacobs and Havinga, 1979) are also found in small yields, as might be suspected from the geometry of the funnel , (Figure 7.2) and by analogy to butadiene. --
As in ethylene, at a twist angle of about 90" the doubly excited state, responsible for the biradicaloid minimum in S,, lies below the singly excited state. In Figure 4.6 it is fairly widely separated from So;in olefins with more extensive conjugation, however, the gap AE at the twisted geometry is smaller and the possibilities for reaching a critically heterosymmetric biradicaloid geometry richer (cf. Section 4.3.3). As a result, the biradicaloid minimum may actually turn into a funnel: In the case of stilbene, any possible intermediate in the cis-trans process is found to have a lifetime of less than 150 fs (Abrash et al., 1990). Nevertheless, a transient spectrum different from those of cis- and trans-stilbene has been observed upon excitation of
- -
Example 7.2: From the quantum yields of the direct cis-trans photoisomerization of 2.4-hexadienes it is seen that a one-step conversion of the trans-trans isomer into the cis-cis isomer or vice versa does not occur. A common intermediate can therefore be excluded, and at least two different intermediates have to be assumed as indicated in Scheme I (Saltiel et al., 1970).In contrast to this singlet reaction with very fast return to So,the triplet reaction is characterized by intermediate lifetimes long enough to allow for interconversion. The result of the benzophenone-sensitized photoisomerization of2,4-hexadienes is therefore in agreement with a common triplet intermediate for the isomerization of both double bonds (Saltiel et a]., 1969).
Figure 7.3. a) Schematic state correlation diagram for the cis-trans isomerization of stilbene along the symmetric path (by permission from Orlandi and Siebrand, 1975); b) modified Orlandi-Siebri~nddiagram (by permission from Hohlneicher and Dick, 1984).
7.1
cis-stilbene (Doany et al., 1985). presumably due to vibrationally hot excited trans-stilbene. In contrast to the situation in ethylene, the model proposed by Orlandi and Siebrand (1975) suggests a maximum at 8 = 90" along the symmetrical reaction path for the singly excited state because of conjugative stabilization of the planar configurations (8 = 0" and 8 = 180"). absent in ethylene. The result is a barrier in S, separating the spectroscopic minimum of the trans isomer (8 = 180°), but not that of the less stable cis isomer (Greene and Farow, 1983), from the biradicaloid minimum at 8 = 90". The situation is similar in the case of long-chain polyenes, which also show fluorescence (Hudson and Kohler, 1973). As a consequence of the barrier in the S, state, the quantum yield for the trans-cis isomerization of stilbene is temperature dependent; at lower temperatures fluorescence becomes increasingly important as a competing process; and a, becomes nearly unity at temperatures below 100 K (Figure 7.4). If intersystem crossing can be neglected, one has The rate constant for the isomerization of trans-stilbene in the S, state is also affected by solvent viscosity and has served as a favorite prototype for the investigation of solvent dynamics in fast monomolecular kinetic processes (Saltiel and Sun, 1990). Quantum chemical calculations essentially confirm the above simple model for the trans-cis isomerization of stilbene. Two-photon excitation spectra and more recent calculations indicate, however, that contrary to what is shown in Figure 7.3a, it is not the lowest but rather a higher excited A state of trans-stilbene that correlates with the lowest excited singlet state
CIS-TRANS ISOMERIZATION OF DOUBLE BONDS
371
at the twisted biradicaloid minimum. This results in several avoided crossings and a barrier in the lowest excited singlet state as shown in Figure 7.3b (Hohlneicher and Dick, 1984). Other authors have proposed that the barrier in S, does not result from an avoided crossing, and that the predominantly doubly excited state does not become S, until after the barrier has been passed and the perpendicular geometry is nearly reached (Troe and Weitzel, 1988). It is difficult to make any definitive a priori statements about the potential energy surfaces of a molecule of this size at this time, particularly since the optimal reaction path almost certainly preserves no symmetry elements. Substituted stilbenes have also been extensively studied (Saltiel and Charlton, 1980; see also Saltiel and Sun, 1990). On direct irradiation the formation of dihydrophenanthrene (5, DHP) accompanies the cis-trans isomerization of stilbene. Thus, one has where /3 = 0.4 is the fraction of molecules in the biradicaloid minimum that reach the ground state of trans-stilbene.
Using azulene as a triplet quencher it has been shown that triplet states are not involved in the cis-trans isomerization of stilbene on direct excitation. Triplet-sensitized cis-trans isomerization, however, is observed and proceeds in both directions through a minimum in the triplet potential energy surface at a twisted geometry, often referred to as the triplet "phantom" state jP*. From Example 7.1 the photostationary state may be written as with quantum yields and
Figure 7.4. Temperature dependence of the quantum yield @,+, for trans-cis isomerization (A), of the quantum yield a, of fluorescence (0).and of the fluorescence lifetime (*)of stilbene (by permission from Saltiel and Charlton, 1980).
where y is the fraction of molecules in the triplet minimum that form the trans isomer. Since for stilbene /?and y are both close to 0.4, it is probable that the singlet and the triplet minima are located at similar geometries (Saltiel and Charlton, 1980). Insertion of the expressions for a, and a, yields
ORGANIC PHOTOCHEMISTRY
7.1
CIS-TRANS ISOMERIZATION OF DOUBLE BONDS
Figure 7.5. Variation of the photostationary state in cis-trans isomerization of stilbene with triplet energy E, of the sensitizer (by permission from Saltiel and Charlton,
1980). Figure 7.6. Calculated potential energy curves'for a twist in the double bond in
C H d H , (---) and CH,=NH,@ The ratio kfET/kT may vary over a wide range. Vertical excitation of transstilbene and of cis-stilbene requires sensitizer triplet energies of ET > 50 kcaVmol and ET > 57 kcal/mol, respectively. When ET of the sensitizer is sufficiently higher than the triplet energies of both the cis and trans olefin, triplet energy transfer is diffusion-controlled; that is, practically every encounter results in energy transfer. Sensitizers with a triplet energy above that of trans-stilbene but below that of cis-stilbene transfer triplet energy only to the trans isomer, and kjTis very small. According to Equation (7.3, ([~]l[t]),~then becomes very large. The dependence of ([cll[tl),ss on ET of the sensitizer shown in Figure 7.5 is thus easy to understand; the fact that sensitization occurs even for ET < 50 kcal/mol has been explained by "nonvertical" energy transfer into the triplet minimum 3P* (Hammond and Saltiel, 1%3), which presumably corresponds to transitions originating from a vibrationally excited ground state. Other mechanisms have been discussed as well (Saltiel and Sun, 1990).
7.1.5 Heteroatom, Substituent, and Solvent Effects As indicated in Section 4.3.3, substituent as well as environmental effects on photochemical cis-trans isomerization can be discussed in terms of the dependence of the energy gap and consequently also of the S,+So transition rate on the electronegativity difference 6. This is determined by the nature of the atoms on the double bond, by substituents, and by the solvent. If 6 is
(-)
(by permission from BonaCiC-Koutecky et al.,
1984).
equal to the critical quantity Go the energy gap disappears and the S, and So surfaces touch. According to Figure 7.6 this situation is just attained for the twisted formaldiminium ion CHFNH,@. In this case a thermal equilibrium in S, is not to be expected at the biradicaloid minimum, which corresponds to a funnel instead. Both the conversion of the trans excited state to the cis ground state and the conversion of the cis excited state to the trans ground state would then proceed with dynamical memory and the sum of the quantum yields a,,, and cP,-, might therefore reach the limiting values of zero and 2 (Michl, 1972; see also the conversion of azomethines below). These processes are very important for the understanding of the rapid deactivation of excited states of triphenylmethane and rhodamine dyes (Rettig et al., 1992). In cyanine dyes (Momicchioli et at., 1988), for example, 6 can go beyond the critical value do, and it can become very large in TlCT molecules. (Cf. Section 5.5.2 and 5.5.3.) In the case of the retinal Schiff base (6) the efficiency of cis-trans isomerization of the double bond between C-1 l and C-12 is considerably enhanced by polar solvents on the one hand and by protonation of the Schiff base on the other hand (Becker and Freedman, 1985). This is rather important because 6 is the chromophore of rhodopsin and this isomerization represents one of the primary steps in vision.
7.1
CIS-TRANS ISOMERIZATION OF DOUBLE BONDS
375
7.1.6 Azomethines Syn-anti isomerization about a C=N double bond is intrinsically more complicated than cis-trans isomerization of a C=C double bond. This is due to the fact that w n * excitations have to be discussed in addition to n+n* excitations, and because syn-anti isomerization can be effected by either of two linearly independent kinds of motion or their linear combination, namely twisting and in-plane inversion at the nitrogen atom. (See Figure 7.7.) It is believed that in simple azomethines thermal isomerizations occur through inversion, while photochemical isomerizations proceed along a twisting path (Paetzold et al., 1981). This has been confirmed by quantum chemical calculations of the potential energy surfaces of the ground state and the lowest excited states of formaldimine in the two-dimensional subspace defined by the twisting and linear inversion motions (Bona6C-Koutecky and Michl, 1985a). Selected cuts through these surfaces for different dihedral angles are displayed in Figure 7.8. Whereas the ground state prefers planar geometries
Figure 7.7. The syn-anti isomerization of formaldimine a) through in-plane inversion and b) by rotation. a is the CNH valence angle and 8 the torsional angle.
Figure 7.8. Dependence of the energy of the lowest states of formaldimine on the valence angle a, shown for selected values of the twist angle 0 (by permission from BonaCiC-Koutecky and Michl, 1985a).
(6 = 0" or 180°), orthogonal geometries (8 = 90") are preferred by the T, and S, states that correspond to n-n* excitation. The very small energy gap between S, and Sofor orthogonal geometries in the region 100" < a < 120" can be viewed as a consequence of a conical intersection at a valence angle a = 106.5". Thus, vertical excitation into the S, state should be followed by vibrational relaxation to an orthogonal twisted geometry and to the funnel in S,, followed by a very rapid radiationless relaxation to So,leaving little opportunity for fluorescence or intersystem crossing. Back in the S,, state, the molecule should vibrationally relax rapidly to one of the two symmetry-equivalent planar forms of the imine ("syn" and "anti") with equal probability, and one should have = = 0.5. However, if a significant fraction of the excited molecules reaches the region of the conical intersection without having lost dynamical memory of their original geometry, syn or anti, both quantum yields of isomerization may deviate from 0.5, even in the absence of other competing processes. Relatively little is known about the E-Z isomerization of N-alkylimines (7a). The reversible photoisomerization of anils (7b), however, has been studied in some detail. Since the quantum yield of intersystem crossing a,,,
I
376
7.1
CIS-TRANS ISOMERIZATION OF DOUBLE BONDS
ORGANIC PHOTOCHEMISTRY
is relatively large, it is assumed to be a triplet reaction. (Cf. Paetzold et al., 1981.)
7.1.7 Azo Compounds Finally, azoalkanes (8) have lone pairs of electrons on both nitrogen atoms, and additional w n * transitions and additional kinds of motion have to be considered in discussing cis-trans isomerization. The effect of the n orbitals is apparent from the orbital correlation diagram shown in Figure 7.9. In constructing this diagram use has been made of the fact that the orbitals n, and n, of the lone pairs of electrons on the two nitrogen atoms split due to an appreciable interaction, and that the orbital ordering is the natural one in the cis isomer, with the combination n = (n, + n,)/fl below the combination n- = (n, - n,)/*, while in the trans isomer orbital interaction produces the opposite ordering. +
From PE spectroscopy results for azomethane (Haselbach and Heilbronner, 1970) the energy of the n MO is known to lie between those of the n + and n- orbitals. Thus, a hom*o-LUMO crossing results and the trans-cis
Figure 7.9.
HN=NH.
Orbital correlation diagram for the cis-trans isomerization of diimide
Figure 7.10. Computed potential energy surfaces of the ground state So and the (n,n*) excited states TIand S, for the cis-trans isomerization of diimide as a function of the twist angle 6 and the valence angle a at one of the nitrogen atoms.
isomerization is forbidden in the ground state and allowed in the first excited singlet and triplet states, as in the case of olefins. In contrast to the olefins, however, the lowest excited states of azoalkanes are the (n,llr) excited states. On the (n,n*) excited singlet and triplet surfaces the reaction encounters a correlation-imposed barrier. For instance, the configuration (n-)2(n)1(n+),(+)' of the trans isomer correlates with the doubly excited configuration (n +)'(n)2(n..)'(n*)2of the cis isomer. The computed potenti;ll energy surfaces of Figure 7.10 confirm these concepts. From these calculations it is also seen that isomerization by motion of the substituent in the molecular plane (variation of a) is energetically preferable to twisting the N=N bond (variation of 8) in the ground state, while the opposite is true in the 'v3(n,+) excited states. In agreement with the theoretical results, the photoisomerization of simple azoalkanes is found to be rather effective. For azomethane in benzene ,, = 0.45 have been observed at 25°C quantum yields of @,+,. = 0.42 and ,@ (Thompson et al., 1979). Cis-azo compounds are moderately stable. Only tertiary cis-azoalkanes are thermally unstable and decompose to nitrogen and radicals. (See Section 7.2.2.) Example 7.3: In azobenzene the cis-trans isomerization in the '(n.9)state apparently proceeds along a twisting path whereas in the '(n,n*) state it proceeds along the inversion path. This has been suggested by the fact that for azobenzenes such
OKC,r\l\llC I'HO'I'OCHEMISI KY
Schematic state correlation diagram for the cis-trans isomerization of azobenzene for two reaction paths that correspond to a twist mechanism and an inversion mechanism, respectively (adapted from Rau, 1984).
Figure 7.11.
as the bridged crown ether 9, for which twisting is inhibited for steric reasons, the quantum yield @,-, of trans-cis isomerization is independent of the exciting wavelength. In azobenzene itself, however, n + 9 excitation (A = 436 nm) and n-n* excitation ( A = 3 13 nm) give different quantum yields a,,,..According to the schematic state correlation diagram of Figure 7.1 1, inversion, which is preferred in the '(n,9) state, is forbidden in the ( n , ~state. ) However, from this state a funnel for a 90" twisted geometry may be reached, followed by a rapid transition to the ground-state surface (Rau, 1984).
7.2 Photodissociations These reactions are initiated by the cleavage of a single or a double bond. A double-bond dissociation is normally possible only if the bond is unusually
weak, that is, if one or both of the fragments are unusually stable. The C=N, and N=N, bonds in diazo compounds and azides are the best-known examples. Single-bond dissociation is the simplest when it takes place between atoms that carry no multiple bonds, lone pairs, or empty orbitals. In that case it is bitopic (two active orbitals). If the single bond is covalent and uncharged, the '(a@) state (which often represents S,) is weakly bound and often fluorescent, and the Yo,@) state (which often represents T,) is purely dissociative, as shown schematically in Figure 4.5 for the dissociation of H,. Another example is the photodissociation of the Si-Si bond, mentioned in Section 7.2.3. Bitopic photodissociation normally takes place in the triplet state, which correlates with a radical pair, and not in the singlet state, which correlates with an ion pair. If one of the termini of the single bond is charged (onium cation or ate anion), both the singlet and the triplet ( u , d ) configurations are dissociative (cf. Figure 4.24a), and the single bond can be cleaved in the singlet as well as the triplet state. Ammonium and sulfonium salts are the best-known substrates for such dissociations. If the single bond is of the dative type, which is uncommon in organic chemistry, neither the singlet nor the triplet (a,@) state is dissociative (both correlate with an ion pair). Single-bond dissociation reactions of higher topicity are more common. In these, at least one of the atoms originally connected by the cleaved bond carries a lone pair, an empty orbital, or a multiple bond. The number of active orbitals on the two reaction centers then determines the topicity of the reaction (Section 6.3.3). For instance, the cleavage of the carbon-heteroatom bond in amines and alcohols is tritopic, the cleavage of the saturated carbon-halogen and of the oxygen-oxygen bond is tetratopic, the cleavage of a vinylic carbon-halogen bond is pentatopic, etc. Already for tritopic bond-dissociation reactions, and more so for those of higher topicity, there is at least one configuration of each multiplicity that is dissociative (Figure 6.28), and hom*olytic photochemical bond cleavage is not restricted to the triplet state. Correlation diagrams of the kind discussed in Section 6.3.1 are useful for the construction of qualitative potential energy diagrams for these reactions. Often, the singlet or triplet ( u . d ) configuration does not enter the S, or TI state with significant weight at the initial geometry, and these are represented by other lower-energy configurations, such as (n,n*). This situation normally leads to a barrier in the reaction surface that needs to be overcome before dissociation can take place (Figure 6.20, cf. "intended" correlations in Section 4.2.2). An example of tritopic reactions of this kind are the singletstate dissociations of the benzylic C-X bonds in compounds containing the A r X - C moiety (photo-Fries, photo-Claisen, and others), in which the (n,n*)excited state of the arene chromophore lies below the ( u , d ) configuration, and for which the principles embodied in Figures 6.20 and 6.28 are relevant (Grimme and Dreeskamp, 1992).
ORGANIC PHOTOCHEMISTRY
380
Examples of correlation diagrams for a series of reactions of higher topicity, up to hexatopic, have been discussed in some detail elsewhere (Michl and BonaCiC-Koutecky, 1990). Here, we shall only outline the results for two important cases: Norrish type I C--C bond cleavage in carbonyl compounds (tritopic), and C-N bond cleavage in azo compounds (tetratopic).
7.2.1 a Cleavage of Carbonyl Compounds (Norrish Type I Reaction) a Cleavage of a carbonyl compound is known as a Norrish type I reaction. It gives an acyl and an alkyl radical in the initial step. According to Scheme 2 the acyl radical can lose CO (path a); the resulting radicals can undergo recombination or disproportionation. Alternatively, the acyl radical can react via hydrogen abstraction (path b) or by hydrogen loss and formation of a ketene (path c). R~CH-&=O
+
CR;-CHR,'
(a) + CO + R,CH* +
CR~-CHR,'
+ secondary products
Scheme 2
The correlation diagrams for the a-cleavage reaction have been derived in Section 6.3.1 for the path to a bent acyl radical as well as for that to a linear acyl radical and have been discussed with the aid of potential energy surfaces drawn as a function of the distance R of the dissociating bond and of the OCR angle of the acyl radical. In Figure 7.12 the potential energy surfaces of the photochemical a cleavage of the hydrogen atom of acetaldehyde and benzaldehyde are displayed in a similar way. The symmetry with respect to the molecular plane is indicated by a subscript s (symmetric) or a (antisymmetric), respectively. Detailed quantum chemical ab initio calculations for acetaldehyde have shown that reaction profiles for CH, cleavage and H cleavage are very similar in the 3(n,n*)state (Yadav and Goddard, 1986), and the surfaces in Figures 7.12a and b may be taken as characteristic for the a cleavage of saturated and unsaturated carbonyl compounds, respectively. In agreement with the results from Section 6.3.1, Figure 7.12a shows that in saturated ketones the lowest singlet and triplet states correlate directly with the biradicaloid states of the acyl radical at its linear geometry. Reaching this geometry, although allowed, is usually endothermic to such an extent that it is practically negligible. On the other hand, formation of the bent acyl radical is markedly less endothermic and occurs via a correlation-induced barrier. This barrier stems from a crossing of the Ti,and T, surfaces.
Figure 7.12. Potential energy surfaces for the a-cleavage reaction a) of acetaldehyde and b) of benzaldehyde. Formation of a bent acyl radical corresponds approximately to a reaction path along the front face (QOCC = 130") of the diagrams, whereas formation of the linear acyl radical proceeds on a path leading from the left front corner to the right rear corner (by permission from Reinsch and Klessinger, 1990).
For the initial geometry, these correspond to wn* and n-lt+ excitation, respectively. This crossing can also be discerned in the correlation diagram for the path to a bent acyl radical in Figure 6.24 and will be avoided when the arrangement of the reacting centers is no longer coplnnar; it therefore represents a conic;~linlt.rscclion.
Ulic 1, \1i1C I'liO I'OCHEMIS1 KY
382
In the absence of intersystem crossing (ISC) the '(n,n*) state can react only to give the linear acyl radical. Intersystem crossing followed by conversion into the bent acyl radical is favored, however, since it is associated with a transition of an electron between two mutually orthogonal p orbitals. (Cf. Example 1.8.) The barrier for a cleavage from the '(n,n*) state is considerably smaller than that for a cleavage from the '(n,n*) state. This is in agreement with experimental activation energies of 17 kcallmol and 6 kcall mol observed for the singlet and triplet states of acetone, respectively (Turro, 1978). The singlet ground state of the resulting radical pair is of much higher energy than that of the initial ketone and can therefore easily be deactivated back to the reactant. The corresponding reaction of the triplet radical pair with conservation of spin, however, is symmetry forbidden and in general also endothermic. Dissociation into free radicals is thus preferred. This is another reason why a cleavage is much more efficient from the '(n,n*) state than from the '(n,n*) state. From Figure 7.12b it is seen that in conjugated ketones the '(n,n*) state can be stabilized to such an extent that it will be energetically below the '.'(n,n*) states. Then, all the way along the reaction path leading to the bent acyl radical, the T, surface remains below the T, and S, surfaces, which for the initial geometry correspond to an n + f l excitation. In this case, the reaction rate is determined by the T,barrier. According to the discussion in Section 6.3.1, a barrier is to be expected from the fact that excitation does
I
not reside primarily in the a bond to be cleaved but mostly in the carbonyl group. (Cf. the natural correlations shown in Figure 6.24.) These results also follow from the cross sections through the potential energy surfaces calculated in the semiempirical all-valence electron approximation by the MNDOC-CI method for acetaldehyde, acrolein, and benzaldehyde, which are displayed in Figure 7.13. Stabilization of the '(n,n*) state by conjugation is seen to increase slightly the barrier, and this should reduce the reactivity. The a cleavage of t-butyl phenyl ketone is in fact slower by a factor of lo3 than the same reaction of t-butyl methyl ketone (Yang et al., 1970). If the '(n,n*) state is below the 3(n,n*) state, a considerably higher barrier is to be expected from Figure 7.13. Biphenyl t-butyl ketone (lo), whose lowest triplet state is (n,llr), is in fact photostable (Lewis and Magyar, 1972).
In Figure 7.14a the calculated barrier heights of the lowest triplet state are displayed for a number of alkyl methyl ketones with different alkyl groups R. For cleavage of the methyl group the barrier is independent of R, whereas for cleavage of the R group it is determined by the degree of branching at the a carbon. For primary alkyl groups --CH,R1 the calculated activation energy E, = 30 kcallmol is independent of R', for secondary alkyl
R= HpCH-
"
cH3-
E
AE= LL kcollmol
A€= L5 kcollmol
Q --AR = SO pm
Figure 7.13. Cross sections through the T, (n.n*) and T, (an*) triplet surfaces for the a cleavage of a) acetaldehyde, b) acrolein, and c) benzaldehyde, leading to the
bent acyl radicals. In a) and b), the surface crossing that is avoided for nonplanar geometries determines the barrier height and the location of the transition state (by permission from Reinsch et al., 1988).
Figure 7.14. a cleavage of saturated methyl ketones CH,COR; a) activation energy E, and b) location of the transition state for different groups R. ARC, is the increase in the length of the CC bond in the transition state (by permission from Reinsch and Klessinger. 1990).
O K W I C PHUI'OCHEMIS'I'KY
384
groups --€HRfR, E, ;= 25 kcal/mol, and for --CH(CH,),, E, = 20 kcallmol. This is in good agreement with experimental data that indicate that the rate and efficiency of the a-cleavage reaction increase with increasing stability of the resulting radicals. In unsymmetrical ketones the weakest CC bond is cleaved; methyl ethyl ketone thus preferentially yields ethyl and acetyl radicals, but the selectivity of the a cleavage decreases with increasing energy of the exciting light (Turro et at., 1972a). Finally, from Figure 7.14b it is seen that similar regularities are found for the transition-state geometries: the smaller the activation energy E,, the smaller the elongation ARcc of the bond to be cleaved. This is exactly what is to be expected from the correlation diagram in Figure 6.24, if for a given acyl radical the energy of the 3Ba,astate is determined essentially by the nature of the radical center at the cleaved group R. Further information about the mechanism of the a-cleavage reaction is available from the study of photochemical reactions in micelles (cf. Figure 7.15), formed in aqueous solutions of detergents (Turro et al., 1985). According to the Wigner-Witmer spin-conservation rules (Section 5.4. I ) a triplet radical pair is formed at first. In hom*ogeneous solutions intersystem crossing (ISC) is much slower than diffusion (kTs < kdi,) and free radicals will form. In micelles, however, k,, is appreciably smaller and of the same order as the escape rate from the micelle (106-10' s-I), and singlet radical pairs and their recombination products may well form.
Thus, upon photolysis of dibenzyl ketone (11) in micelles, 4-methylphenyl benzyl ketone (12) is formed as a side product according to Scheme 3.
Due to the cage effect in micelles, unsymmetrically substituted dibenzyl ketones such as 13 yield predominantly the unsymmetrical diphenylethanes on photodecarbonylation, whereas in hom*ogeneous solution all three possible products are formed in the statistical ratio 1:2:1 (Turro and Kraeutler, 1978).
hom*ogeneous
Solution
1
:
2
1
Miille
1
:
6
1
Photolysis of ketones in micelles with simultaneous application of an external magnetic field permits a I3Cisotope enrichment. (Cf. Section 6.1 S.5.) This is the case because I3Cnuclei have a magnetic moment and thus accelerate the spin inversion by the hyperfine interaction mechanism. (Cf. Example 4.9.) Due to the more efficient recombination of radicals containing I3C, the initial product formed after photolysis in a back reaction is 13Cenriched (Turro et at., 1980b). In spite of the occasionally rather low efficiency, decarbonylation has been used to generate strained ring systems. An important example is the synthesis of tetra-t-butyltetrahedrane (14) (Maier et al., 1981 ). Monomeric surfactant
Micelle
+
free surfoct~nf
Figure 7.15. Schematic representation of a micelle (by permission from Bunton and Savelli, 1986).
Example 7.4:
The photochemical behavior of cyclobutanone (15) contrasts sharply with that of other ketones. Cyclobutanone undergoes a cleavage also from the '(n,n*) state, with subsequent fragmentation to ketene and olefin, decarbonylation to cyclopropane or cyclization to oxacarbene (16), whose concerted formation has also been proposed on the basis of stereochemical observations (Stohrer et al.. 1974). In contrast, cyclohexanone cleaves exclusively from the triplet state and undergoes disproportionation reactions. The photochemical activity of cyclobutanone persists even at low temperatures (77 K ) where cyclohexanone is photostable.
radical is formed in an excited state and can dissociate into CO and CH,-
R s 0Q xn +
a x - c - R CY
X
Rco
O
xH
T
cleavage
R
+ COR
He.
RCO
Scheme 4
These results may be rationalized by means of the correlation diagram in Figure 7.16, which was constructed using thermochemical and spectroscopic data to estimate state energies. The (n,n*) states of the initial ketone are higher in energy than the B states of the linear acyl radical that is an unstrained openchain species, and the reaction corresponding to the direct correlation will be exothermic. Furthermore, it has been proposed that the resulting linear acyl
The rearrangement is almost exclusively intramolecular and proceeds within a solvent cage (Adam et al., 1973). In the case of phenyl acetate (X = 0,R = CH,), the o:p ratio increases from 1 in the absence of Pcyclodextrine to 6.2 in its presence; the macrocycle provides a cage for the radical pair (Ohara and Watanabe, 1975). However, these reactions are singlet reactions and should be classified a s dissociations of the benzylic C-X bond rather than a-cleavage reactions (Grimme and Dreeskamp, 1992), as mentioned above.
7.2.2 N, Elimination from Azo Compounds The lowest excited state of many azo compounds, like that of ketones, is an (n,n*) state. Photolytic cleavage of a C N bond analogous to the a cleavage of ketones is therefore to be expected:
Another conceivable route would be the concerted elimination of nitrogen:
7.16. Correlation diagram for the a cleavi~ge of cyclobutanone (adapted from Turro et al., 1976).
Figure
Experimental data and theoretical arguments indicate that the concerted path is energetically unfavorable, s o in general the two-step mechanism is involved (Engel, 1980). As a model for this reaction the orbital correlation diagram for the cleavage of one N H bond of cis-diimide is shown in Figure 7.17a, and the state correlation diagram derived therefrom is displayed in Figure 7.17b. The sim-
7.2
PHC)'f( )L)lSSOCIAI IONS
Cleavage of one cis-diimide NH bond; a) orbital correlation diagram, b) state correlation diagram (adapted from Bigot et al., 1978). Figure 7.17.
ilarity to the corresponding diagram for the a cleavage of ketones is apparent. The singlet state of the a,obiradical correlates with the ground state, and the triplet state with one of the higher excited (n,dc) states of the azo compound. The'conversion of the '.'(n,llc) states of the azo compound to the 'v3B,, biradical states is electronically allowed but is expected to be weakly endothermic. The results of quantum chemical calculations shown in Figure 7.18 confirm the expectations based on the correlation diagram. In comparing these results with experimental data it has to be remembered that in contrast to ketones, azo compounds can also undergo photochemical trans-cis isomerizations. (Cf. Section 7.1.7.) In the gas phase n+n? excitation results in photodissociation with nearly unit quantum efficiency. At higher pressures, however, and especially in solution, this reaction almost completely disappears and photoisomerization dominates. The latter is observed even at liquid nitrogen temperatures. This is understandable if it is accepted that photodissociation proceeds in the gas phase as a hot ground-state reaction. According to Figure 7.18, it has to overcome a barrier in the excited state and is therefore not observed in solution. For the trans-cis isomerization, on the other hand, no excited-state barrier is to be expected from the results in Section 7.1.7. For most acyclic azo compounds photodissociation in solution is an indirect process, that is, the photochemical reaction proper is a trans-cis isomerization, and the cis-azoalkane formed undergoes thermal decomposition to nitrogen and radicals. This occurs especially if the cis isomer is suffi-
Calculated potential energy curves of the ground and low-lying excited states of cis-diimide in the one-bond cleavage (by permission from Bigot et al., 1978).
Figure 7.18.
ciently unstable due to the presence of sterically demanding alkyl groups. The use of triplet sensitizers has shown that the major part of direct photolysis does not involve the triplet state; the extrusion of nitrogen from the triplet state requires an activation energy Ea (Engel, 1980). Cis-trans isomerization of cyclic azo compounds is only possible for sixmembered or larger rings. Otherwise only N, loss is observed, and direct irradiation and triplet sensitization can yield different products, as is the case for the cyclic azo compound 17 (Bartlett and Porter, 1968). This is due to the fact that the biradical R f .T R formed from the triplet-excited reactant cannot recombine to form a cyclobutane until spin inversion has occurred; the triplet biradical can live long enough to be diverted to the rotamer R t f R'. According to Scheme 5 cis and trans products are formed; direct photolysis produces 18a and 18b in a ratio 10: 1, whereas triplet sensitization yields a product ratio of 1.4: 1, which is nearly equal to what would be expected for an equilibrium distribution between the different rotameric biradicals. A biradical such as this, which gives different products depending upon its multiplicity, is said to exhibit a spin-correlarion effect. Photochemical elimination of Nzfrom bicyclic azo compounds produces cyclic cis-1,n-biradicals followed by stereospecific ring cleavage or cycliza-
€1
6 = 6 kcallrnol
@>,-
00 75k[r\
SY isomers
q; 0.2.kSy lo7 S-' 56 kcallrnd
.
I
3.
0.06
Scheme 5
tion as exemplified in Scheme 6, whereas disproportionation would require a change in conformation and is therefore in general not observed (Cohen and Zand, 1962).
Figure 7.19. Jablonski diagram and photochemical parameters of 7,8-diazatetracycl0[3.3.0.0~4.0'.~]oct-7-ene(by permission from Turro, 1978).
Sl+Tl intersystem crossing and therefore enhances the formation of diazacyclooctatetraene on direct irradiation. (Mlproport.1
Scheme 6
-
Example 7.5: Compound 19, which yields the valence isomers of benzene on direct irradiation, and diazacyclooctatraene (20) as the major product upon triplet sensitization, is an interesting example of the different reactivity of the S, and TI states of azo compounds (Turro et al.. 1977a):
The photochemical parameters for 19 are summarized in Figure 7.19. From these data it is apparent that at low temperatures diazacyclooctatetraene becomes the exclusive product, since the loss of N,from the '(n,lr*)state requires an activation energy of 5-6 kcal/mol. Oxygen has a catalytic effect on the
-
The photoextrusion of N, from cyclic azo compounds is a very useful way of producing strained ring systems such as 21 (Snyder and Dougherty, 1985) or 22 (Liittke and Schabacker, 1966). Unstable species such as the o-quinodimethanes 23 (Flynn and Michl, 1974) and 24 (Gisin and Wirz, 1976), and biradicals such as 25 (Gisin and Wirz, 1976), 26 (Watson et al., 1976), 27 (Platz and Berson, 1977), 28 (Dowd, 1966), and 29 (Roth and Erker, 1973), can also be generated in a matrix by this route and spectroscopically identified.
392
ORGANIC PHOTOCHEMISTRY
7.2
PHOTODISSOCIAI'IONS
Some azo' compounds undergo the usual photolysis (A > 300 nm) only to a minor degree or not at all and are therefore dubbed "reluctant azoalkanes." These are cyclic azo compounds such as 30,31, and 32.
Photolysis of such compounds can be accelerated by employing elevated temperatures or by introducing substituents that stabilize the radicals formed. (Cf. Engel et al., 1985.) Short-wavelength irradiation (A = 185 nm) also enhances photodissociation. Bridgehead azoalkanes such as 33 are also reluctant compounds and undergo photochemical trans-cis isomerization (Chae et al., 1981). Loss of nitrogen and formation of bridgehead radicals are observed upon excitation to the second singlet state (S,) of the trans or the cis isomer, with quantum yields of cD, = 0.3 and @, = 0.16, respectively (Adam et al., 1983).
tion reactions, but not as dissociation reactions, since more than one single or double bond is broken in these processes. They are closely related to the pericyclic processes discussed in Sections 7.4 and 7.5 and are formally isoelectronic with other excited-state-allowed four-electron pericyclic reactions, such as cheletropic elimination of CO from cyclopropanone and disrotatory electrocyclic ring closure in butadiene. The analogy of the chainabridgement reaction to the latter is illustrated in Figure 7.20, which shows the orbitals involved in the two reactions.
7.2.3 Photofragmentation of Oligosilanes and Polysilanes Alkylated and arylated oligosilanes and polysilanes, the silicon analogues of alkanes and of polyethylene, have recently attracted considerable attention (Miller and Michl, 1989). Unlike saturated hydrocarbons, these materials absorb in the near UV region. The reasons for this are related to the electropositive nature of silicon and can be understood in simple terms (Michl, 1990). Their excited states bear considerable similarities to those of polyenes, but also exhibit significant differences (Balaji and Michl, 1991). Upon irradiation, oligosilanes (Ishikawa and Kumada, 1986) and particularly polysilanes (Trefonas et al., 1985), readily fragment to lower-molecular-weight species, and polysilanes show promise as photoresists. Three distinct photochemical processes have been identified as shown in Scheme 7: (1) chain abridgement by silylene extrusion, (2) chain cleavage by silylene elimination, and (3) chain cleavage by hom*olytic scission (Miller and Michl, 1989). The two silylene-generating processes are believed to occur in the singlet excited state in pericyclic fashion, while the radical-pair-forming hom*olytic cleavage process is believed to occur in the triplet state (Michl and Balaji, 1991), as would be expected from Figure 4.5. According to the definition given in the beginning of this section, the singlet I , l-elimination (reductive elimination) processes qualify as fragmenta-
Figure 7.20. Comparison of the A 0 interactions in the photochemical chain abridgement in a polysilane (top) and in the disrotatory electrocyclic ring closure of butadiene (bottom) (by permission from Michl and Balaji, 1991).
394
()I<( ANIC I'HOTOCHEMISTRY
7.3
HYDROGEN ABSTRACTION REACTIONS
7.3 Hydrogen Abstraction Reactions One of the earliest photoreactions to be studied was the photoreduction of benzophenone (Ciamician and Silber, 1900bthat is, the conversion of a carbonyl compound into an alcohol by an intermolecular hydrogen abstraction reaction. intramolecular hydrogen abstraction by the carbonyl group, usually from the y site, is referred to as a Norrish type I1 reaction. Hydrogen abstraction by olefins and heterocycles has also been observed.
7.3.1 Photoreductions The correlation diagram for hydrogen abstraction by a ketone has been derived in Example 4.5 (Section 4.2.3). From this diagram it can be concluded that no reaction is to be expected from the '.'(n,jlr)states for the in-plane
I 0
I
I
200
1
I
I
I
LOO
600
800
R
'z*
3z*
fpml
Potential energy curves for the reductive elimination of SiH, from Si,H,, as calculated by an ab initio method (by permission from Michl and Balaji, 1991). Figure 7.21.
Figure 7.21 shows the behavior of the calculated singlet state energies along an idealized reaction path from trisilane to disilane and silylene (Michl and Balaji, 1991). quite similar to the classical behavior calculated for disrotatory ring closure of butadiene (van der Lugt and Oosterhoff, 1969). The So state of trisilane attempts to correlate with a doubly excited state of the products, in which both electrons of the silylene lone pair are in the Si 3p orbital. The So state of the products originates in a doubly HOMWLUMO excited state of the starting trisilane, and the would-be crossing is weakly avoided (along a less symmetrical path it may not be avoided at all), resulting in a barrier in the Soand a minimum or funnel in the S, state ("pericyclic funnel"). The optically strongly allowed HOMWLUMO excited ( a , @ )state correlates with a singly excited state of the products, with silylene in its first excited singlet, and does not impose a barrier in the excited state surface. The reaction presumably proceeds by relaxation into the ground state through the pericyclic funnel, followed by return to the starting geometry for some of the molecules, and dissociation into the products for the remainder.
'z
Ab initio results for hydrogen abstraction in CH-,= + CH,, as a function of the distance R,, between the carbon in methane and the oxygen in formaldehyde (by permission from Bigot, 1983). Figure 7.22.
ORGANIC PHOTOCHEMISTRY
396
7.3
HYDROGEN ABSTKACTION REACTIONS
attack; the reaction from the 'v3(n,n*)states will have to overcome a barrier that results from a strongly avoided crossing of states if natural orbital correlations are used. (Cf. Figure 6.21 .) The results of ab initio calculations for the formaldehyde-methane system depicted in Figure 7.22 show that the barrier decreases with decreasing distance Rco between the methane carbon and carbonyl oxygen. As a consequence, the hydrogen abstraction requires the separation between the reactants to be very small (Bigot, 1983). The crossing between the correlation lines starting at the ground state and the '(n,+) state of the reactant shown in Figure 7.22 becomes avoided if the system deviates from a coplanar arrangement. The crossing corresponds to a funnel, which mediates the return both to the product and reactant ground state in the usual way. This reduces the efficiency of the singlet reaction. Since in aryl ketones intersystem crossing is very fast, it is obvious why the reaction is observed only from the 3(n,n*)state. 0........ H-)(
In-plane attack
Perpendicular attack
In addition to the reaction between the no orbital of the carbonyl group and the a,, orbital, another conceivable reaction path for hydrogen abstraction is based on an interaction between the nc, and the a,, orbital. This is referred to as a perpendicular attack. In the correlation diagram for the perpendicular attack derived in Example 7.6, the 3 ( ~ , n *state ) correlates with the 3Bu,u ground state of the primary product. However, the corresponding reaction is expected to show a considerable barrier. The reaction from the '.3(n,n*)states is forbidden in the perpendicular case. These results are also confirmed by formaldehyde-methane calculations.
Perpendicular hydrogen abstraction reactions: a) the natural orbital correlation diagram and b) the resulting state correlation diagram (adapted from Bigot, 1980).
Figure 7.23.
should therefore be conceivable for molecules such as biphenylyl r-butyl ketone (lo), whose Yn,n*) state lies below the '(n,n*) state. No perpendicular hydrogen abstraction is observed even with such compounds because of the barrier expected from the natural orbital correlation. The reaction of benzophenone (34) with benzhydrol (35) is a representative example of hydrogen abstraction:
Example 7.6:
The natural orbital correlation for the perpendicular hydrogen abstraction reaction is shown in Figure 7.23a. The n,orbital of the ketone is seen to correlate with the n, orbital of the product, whereas the a,, orbital turns into the n;l orbital of the alkyl radical center and the n,, orbital turns into the a,, orbital. The product ground state is a biradicaloid 1.3B,, state with singly occupied I& and n, orbitals, which are both in the plane of the reacting centers and therefore have a symmetry. In addition, there is a '.'Bun state with a singly occupied no orbital with n symmetry. The resulting configuration correlation diagram is indicated in Figure 7.23b by dotted lines, whereas the state correlation diagram, which takes into account avoided crossings, is shown by full and broken lines for singlets and triplets, respectively. In contrast to the in-plane attack, the perpendicular attack should be most favorable from the '(n,n*) state and
In the first step the spectroscopically detectable ketyl radical 36 is formed, which then recombines to form benzopinacol (37) (Weiner, 1971). The same pinacol is obtained by reacting benzophenone with 2-propanol, since the dimethylketyl radical (38) produced in the hydrogen abstraction step is a strong reductant and transfers a hydrogen atom to the excess benzophenone to form another molecule of the diphenylketyl radical (36).
398
OK(;ANIC PHOTOCHEMISTRY
Thus the quantum yield for the disappearance of benzophenone has a limiting value of a, = 2, since only one photon is needed to convert two molecules of the reactant into one product molecule:
Some ketones such as 2-acetylnaphthalene (39) show no photochemical reactivity. In these cases the lowest triplet state is a (n,n*) state, the reaction of which is inhibited by a barrier which from the correlation diagram in Figure 4.18 is seen to result from an avoided crossing of the levels starting at the vertical (n,n*) and (n,x*) states. The efficiency of the hydrogen abstraction by 4-hydroxybenzophenone (40) is strongly solvent dependent. In cyclohexane, the quantum yield is high, a, = 0.9, but in the polar Zpropanol solvent, the (n, n*) state is destabilized to such an extent that it no longer is the lowest triplet state, and the quantum yield drops to a, = 0.02 (Porter and Suppan, 1964). These examples demonstrate the influence of the energy ordering of states: the wrong order results in correlation-induced barriers. (Cf. Section 6.2.4.)
7.3
399
HYDROGEN ABSTRACTION REACTIONS
Table 7.1 Quantum Yields of Norrish Type I1 Reactions (Adapted from Horspool, 1976) Ketone
Solvent
@S
@,
7.3.2 The Norrish Type I1 Reaction A common reaction of aliphatic ketones is intramolecular hydrogen abstraction from the y position (in rare instances from the 6 or even the p position). In addition to regenerating the reactant, the resulting biradical can cleave to give an olefin and an enol, or form a cycloalkanol. Scheme 9 illustrates the most important case of y hydrogen abstraction.
Scheme 9
Carbonyl compounds can also be reduced by electron transfer instead of hydrogen abstraction (Cohen et al., 1973). For the reduction of benzophenone with diisopropylamine the mechanism can be formulated as indicated in Scheme 8. The intermediate of this reaction again is the ketyl radical 36, which recombines to form the pinacol37:
This reaction is known as Norrish type I1 reaction. The correlation diagram for the intermolecular hydrogen abstraction discussed in the last section applies equally well to the intramolecular reaction. Accordingly, the inplane hydrogen abstraction yielding a ketyl radical is allowed from the '(n,n*) state as well as from the )(n,n*) state and forbidden from the (n,n*) states irrespective of their multiplicity. Since singlet states prefer tight biradicaloid geometries, and triplet states prefer loose ones that favor product formation, quantum yields asof the singlet reaction are generally smaller than those of the triplet reaction (a,), as is exemplified by the data in Table 7.1. In alkyl aryl ketones spin inversion is so fast that no singlet state reaction is observed (Wagner and Hammond, 1965).
Scheme 8
Naphthyl ketones, which are generally unreactive toward photoreduction by Zpropanol, are efficiently reduced by amines. Electron transfer (cf. Section 5.4.4) is considerably faster than hydrogen abstraction. Thus, the reaction cannot be quenched using common triplet quenchers although it proceeds from the triplet state.
Example 7.7: MIND013 calculations including configuration interaction for the type 11 reaction of butanal (Dewar and Doubleday, 1978) confirm the conclusions from the correlation diagram. The results are summarized in Figure 7.24: Reactions from the '(n,n*) state as well as from the >(n,n*)state proceed over a barrier. The triplet reaction then reaches the 'B,,, state of the biradical, which is just
ORGANIC PHOTOCHEMISTRY
7.3
HYDROGEN ABSTRACTION REACTIONS
401
tones with a chiral y-C atom such as 41, which competes with the hydrogen abstraction (Yang and Elliott 1969):
The reaction is not concerted and does not yield a triplet olefin, even when this process would be exothermic, as in the case of 42. Triplet stilbene decays to a 60:40 mixture of cis- and trans-stilbene, but in the reaction of 42,98.6% trans-stilbene was observed (Wagner and Kelso, 1969).
Schematic representation of energies of stationary points for the Norrish type 11 reaction of butanal. The diagram corresponds to a projection of multidimensional potential energy surfaces into a plane. The two energies given for the biradical on the Sosurface correspond to a geometry optimized for So(front bottom) and optimized for S, (middle rear), respectively. A broken line (---) is used for the TI surface and a broken-dotted line (-.-.-)for the S, surface. The relative energies of T,, Soand S, for the geometry of the funnel are not known (by permission from Dewar and Doubleday, 1978). Figure 7.24.
barely above the ground state. In the correlation diagram, this is apparently the point that corresponds to the crossing of the levels connected to the ground state and to the I(n,n*) state of the ketone. From this funnel the system can return to the initial geometry or proceed either to the singlet biradical or to the products (cyclobutanol or enol + olefin). The reaction path toward the elimination products passes through a geometry that is effectively the same as that of the transition state for cleavage of the singlet biradical and is stereospecific; that is, the stereochemistry of the initial ketone is preserved in the products. The calculated activation energies of 9 and 12 kcal/mol for the singlet and the triplet reaction, respectively, and of 10 kcallmol for cleavage of the singlet biradical are presumably too high by 3-5 kcallmol, but the relative values appear to be correct. Evidence that the triplet reaction is not concerted, but rather proceeds via the 1,Cbiradical, has been obtained from the photoracemization of ke-
Intersystem crossing to the So surface is believed to occur at the 1,4-biradical stage, and to yield one of the three possible types of singlet product (reactant, olefin, cycloalkanol) depending on the geometry at which it occurs, as discussed in Section 4.3.4. The singlet reaction also proceeds at least partially via a biradical, as was shown indirectly. If the reaction of 43 were concerted, the transfer of H should yield the deuterated cis isomer and the transfer of D the nondeuterated trans isomer:
With piperylene as the triplet quencher, 10% deuterated trans olefin is found, which must result from rotation about the /3,y bond of the singlet biradical (Casey and Boggs, 1972). If a molecule has two y hydrogens available, in the Norrish type I1 reaction the transfer proceeds over the lower of the two barriers, and according to Scheme 10 a preference for cleavage of the weaker secondary CH bond results (Coxon and Halton, 1974).
\
OK(;ANIC PHOTOCHEMISTRY (Hatransf.) CHI-CH=CH-CHI
+ CH3-CH=O (Major route)
c~,cH:,
(libtransf.)
(mrwr r v l
An increase in temperature or in photon energy reduces the selectivity. For alkyl aryl ketones electron-releasing substituents in the p position decrease the rate constant and quantum yields for type I1 cleavage. p-OH, p-NH,, and p-phenyl substituents inhibit the reaction completely. Similarly as in the: case of intermolecular hydrogen abstraction, this effect is thought to be a consequence of the 3(n,llc) state no longer being the lowest triplet state, resulting in a larger barrier. The rario of olefin to cyclobutanol product yield often depends on substitution. In order to understand this in detail, more would need to be known about the conformational dependence of the spin-orbit coupling matrix element. The present qualitative understanding (Section 4.3.4, Figure 4.26) of the critical intersystem crossing step suggests that it occurs at geometries at which the 2p orbitals of the two radical ends interact through a nonzero resonance integral while their axes lie approximately orthogonal to each other, such that after a 90-degree rotation of one of the orbitals about its center there still is a nonzero resonance integral with the other. According to this analysis a gauche conformation, in which the orbitals interact primarily through space (Figure 7.25a), and an anti conformation, in which they interact primarily through bonds (Figure 7.25b), can both be favorable, provided that the end groups are twisted properly. After intersystem crossing, the former is expected to yield the cyclobutanol, and the latter the fragmentation products, essentially instantaneously. The relative energies of the two types of conformation should be sensitive to the steric demands of substituents.
Stereoelectronic effects on the Norrish type 11 reaction. Presumed o p timal orbital alignments a) for cyclization and b) for elimination. Figure 7.25.
7.3
403
HYDROGEN ABSTRACTION REACTIONS
The cyclobutanol-forming path is diastereoselective, for example, a-methylbutyrophenone (44) and valerophenone (45) prefer to place the methyl and the phenyl groups on opposite sides of the four-membered ring (Scheme 11). Such steric discrimination on the T surface may already be present in the open-chain triplet biradical or become felt gradually as the radical ends approach each other and develop the covalent perturbation that leads to intersystem crossing (see Section 4.3.4), and both cases are known and exemplified by the two reactions in Scheme 1I. The high stereoselectivity of 44 is believed to be due to a repulsive interaction of the phenyl and the a-methyl groups in the 1,4-biradical (Figure 7.25a). The interaction between the phenyl and the y-methyl group in the biradical from 45 should be small until after the intersystem crossing has taken place and the 1,4-bond is almost completely formed (Lewis and Hilliard, 1972).
3
:
1
Scheme 11
Abstraction of a 6 hydrogen normally competes only when a y hydrogen is not available, and produces a 1,5-biradical. There are only two choices: return to the starting materials, or cyclization to a cyclopentanol. For instance, a-(0-ethylpheny1)acetophenone (46) yields 1-methyl-2-phenyl-2-indanol (47):
Diastereoselection is again observed and can be understood in terms of the relative energies of the two conformations that are ideally set up for intersystem crossing by spin-orbit coupling (Section 4.3.4). As is seen in Scheme 12, in the favored conformation, a hydroxyl, and in the disfavored conformation, a phenyl, have to be accommodated close to a benzene ring (Wagner et al., 1991).
ORGANIC PHOTOCHEMISTRY
7.4 Cycloadditions 7.4.1 Photodimerization of Olefins According to the Woodward-Hoffmann rules, the concerted cycloaddition + 2J of two olefins to afford a cyclobutane is allowed photochemically as reaction and thermally as the + J,] reaction. The different modes of addition give rise to products with different stereochemical structures as indicated in Figure 7.26. If the reaction does not follow a concerted pathway
us
us
Fipre 7.26. Stereochemical consequences of thermal and photochemical [2 cycloaddition.
+
21
but rather proceeds by a multistep process, and if ring closure of the intermediate is not very rapid compared with bond rotation, rotations about the CC bonds may occur, with subsequent loss of stereospecificity. Recent calculations for the [2 + 21 photoaddition of two ethylene molecules (Bernardi et al., 1990a,b) demonstrated that the bottom of the pericyclic funnel on the S, surface does not lie at the often assumed highly symmetrical rectangular geometry; instead, it is distorted to permit stabilization by diagonal interaction (cf. Section 6.2. l), as suggested by model calculations on H,(Gerhartz et al., 1977). Moreover, the calculations show that the S , S otouching is not even weakly avoided but actually is a conical intersection. These features, a rhomboidal distortion and a conical intersection, are likely to be general for photocycloadditions. Photodimerization often involves an excimer that can be treated as a supermolecule. (Cf. Section 6.2.3.) Then, the state correlation diagram for the singlet process (Figure 7.27a) ordinarily calls for a two-step return from S, to So along the concerted reaction path. First, an excimer intermediate E* is formed. Second, a thermally activated step takes the system to the diagonally distorted pericyclic funnel P* (cf. Section 4.4. l), and the return to So that follows is essentially immediate. The reaction will be stereospecific and concerted in the sense that the new bonds form in concert. However, it will not be concerted in the other sense of the word, in that it involves an intermediate E*. There may well be systems in which the excimer minimum occurs in the S, rather than the S, surface (Figure 7.27b). The approach to the pericyclic funnel P* on S, may then be barrierless, and an excimer intermediate will
Fipre 7.27. Schematic representation of the state correlation diagram for a groundstate-forbidden pericyclic reaction with an excimer minimum E* a) at geometries well before the pericyclic funnel P* is reached, and b) at geometries similar to those of P*.
not be detectable. (For a possible example of the latter case, see Peters et al., 1993.) These concepts are in very good agreement with experimental findings. There are relatively few examples of photodimerization of simple nonconjugated acyclic olefins because these compounds absorb at very short wavelengths. Irradiation of neat but-Zene, however, yields tetramethylcyclobutane with a quantum yield Q, = 0.04. For very low conversion, the observed stereochemistry of the adducts is the stereospecific one expected from Scheme 13 for a concerted [J, + J,] cycloaddition. However, since the major pathway is cis-trans isomerization with a quantum yield @ = 0.5 (cf. Section 7.1.2), it has been concluded that the molecules that undergo cistrans isomerization are not involved in photodimerization (Yamazaki et al., 1976).
Small-ring cyclic alkenes cannot deactivate by cis-trans isomerization. For instance, in contrast to acyclic alkenes, cyclopentene (48) therefore undergoes photosensitized [2 + 21 cycloaddition. For cycloalkenes with a six-membered or larger ring, a trans form becomes possible; for molecules such as cyclooctene (2), photosensitized cis-trans isomerization is the more efficient reaction path. (Cf. Section 7.1.2.)
Since most simple alkenes have a high triplet energy (ET = 75-78 kcall mol), triplet sensitizers have to be chosen accordingly to prevent oxetane formation (see Section 7.4.4), as shown in Scheme 14 for norbornene with acetophenone (ET = 75 kcallmol) and benzophenone (ET = 69 kcal/mol), respectively, as sensitizers (Arnold et al., 1965):
Scheme 13
In simple olefins, direct excitation of the triplet state and intersystem crossing from an excited singlet state to a triplet state do not play an important role. A sensitized reaction of the triplet state is possible and could in principle also be concerted. However, in the triplet state loose geometries with two separate radical centers are energetically more favorable than tight pericyclic geometries with cyclic interaction (cf. Section 6.2.1); it is most probable that one of these favorable minima will be reached prior to return to So. This tendency then favors a nonconcerted mechanism with the two new bonds formed in separate reaction steps. Formation of the first bond takes place on the TI surface, while the second one will close after the molecule has reached the So surface. The nature and stereochemistry of the product that results from the triplet species upon return to the Sosurface, in particular cyclobutane formation or back reaction to two olefins, is believed to be dictated by the geometry at which the conversion to Sotook place. The relative efficiencies are determined by the rates of the various processes, which in turn depend on such factors as the populations of the various conformers in TI and the size of the T , S o spin-orbit coupling matrix element, the geometrical dependence of which was discussed in detail in Section 4.3.4.
Scheme 14
phAph
Finally, it should be remembered that excimer minima E* and pericyclic funnels P* of different energies can result if different mutual arrangements of the reactants can lead to cycloaddition, as indicated in Scheme 3 in Section 6.2.3. The related issues of regio- and stereoselectivity of singlet photocycloaddition are dealt with in Section 7.4.2. Example 7.8: Sensitized irradiation of cyclohexenes and cycloheptenes in protic media results in protonation. This phenomenon, which is not shared by other acyclic or cyclic olefins, has been attributed to ground-state protonation of a highly strained trans-cycloalkene intermediate. In aprotic media, either direct or triplet-sensitized irradiation of cyclohexene produces a stereoisomeric mixture of [2 + 21 dimers 49-51 as the primary products, with 50 predominating. The reaction apparently involves an initial cis-trans photoisomerization of cyclohexene followed by a nonstereospecific nonconcerted ground-state cycloaddition, promoted by the high degree of strain involved. In contrast, cycloheptene undergoes only a slow addition to the p-xylene used as sensitizer,
408
ORGANIC PHOTOCHEMISTRY
presumably because the trans isomer is not sufficiently strained to undergo the nonconcerted cycloaddition. Copper(1)-catalyzed photodimerization of cyclohexene and cycloheptene affordsthe product expected for [2, + 2,] cycloaddition of the trans isomer to the cis isomer within a cis,trans complex 52. Stereospecificity is high in the former case and complete in the latter. (Kropp et al., 1980.)
Irradiation of butadiene in isooctane yields the isomeric 1,2-divinylcyclobutanes 53a and 53b and various other products, dependent on reaction conditions. The triplet-sensitized photoreaction yields 4-vinylcyclohexene (54) in addition to the divinylcyclobutanes fcf. Example 6.1 I), and the product ratio depends on the triplet energy of the sensitizer:
agonal bonds, to yield the x[2 + 21 product in one case; in the other case, there is preservation of perimeter bonding and eventual production of the [2 + 21 product. (Cf. Figure 6.1.5). Which product will be formed depends on the number of methylene groups between the double bonds. If either one or three CH, groups are present, the common [2 + 21 cycloaddition is observed; with two CH, groups x[2 + 21 cycloaddition predominates. For cyclic dienes the dependence of the reaction product on the number of CH, groups is even more pronounced; 1,5-cyclooctadiene (56) yields the product 57 exclusively (Srinivasan, 1963).
These findings have been rationalized by the rule offive (Srinivasan and Carlough, 1967), according to which five-membered cyclic biradicals are preferentially formed, as shown in Scheme 15. (Cf. also the Baldwin rules for radical cyclizations, Baldwin, 1976.)
Scheme 15
When ET is in excess of 60 kcal/mol, s-trans-butadiene (ET = 60 kcall @),which strongly predominates in the thermal equilibrium, is excited and produces mainly divinylcyclobutanes. When the E, of the sensitizer is not high enough to excite s-trans-butadiene, energy transfer to s-cis-butadiene (E, = 54 kcallmol) occurs instead, yielding vinylcyclohexene. If ET < 50 kcavmol, nonvertical energy transfer to a twisted diene triplet is believed to occur (Liu et al., 1x5). Intramolecular photoadditions can also occur as so-called x[2 + 21 cycloadditions, as demonstrated for 1,5-hexadiene (55): the terminal carbon atom of each double bond adds to the internal carbon atom of the other double bond in such a way the resulting a bonds cross each other.
It is likely that [2 + 21 and x[2 + 21 cycloadditions proceed through the same type of diagonally distorted pericyclic funnel (Section 4.4.1) with a preservation of the diagonal interaction, and eventual production of two di-
Gleiter and Sander (1985) proposed that the reaction is not concerted and that the different reaction course is due to differences in through-bond interactions. (Cf. Gleiter and Schafer, 1990.) Depending on the number of methylene groups between the double bonds, these through-bond interactions may change the orbital ordering. If such an interaction reverses, the "natural" ordering of n orbitals imposed by through-space interaction, the excited-state barrier for the x[2 + 21 cycloaddition will be lower than that for the [2 + 21 cycloaddition. Evidence for the deleterious effect of the reversal of orbital order on the [2 + 21 process is provided by cases in which the x[2 + 21 reaction is sterically impossible and the [2 + 21 reaction fails to proceed (Example 7.9). However, it is also possible that the reaction is concerted in its initial stages, and that the difference in the relaxation paths after return to So through a distorted pericyclic funnel is dictated by the steric requirements imposed by the alkane chains. Example 7.9: Tricyclo[4.2.0.0~~510ctadiene (58) does not undergo (2 cubanr,
although its ec~ornetr\rappc>:rr.; to he
+ 21 photocyclization to
ideally Yet tip for it. In thi.; mol-
sensitized mixed cycloadditions are summarized in Scheme 16 (Scharf, 1974). D l .
Scheme 16
The addition of trans-stilbene to 2,3-dimethylbutene-2 appears to be a singlet reaction: Figure 7.28. Correlation of the n orbitals of two cyclobutene molecules (left) with those of 58 (center) and 59 (right) (by permission from Gleiter, 1992).
ecule, the interaction of the normally more stable in-phase combination of the two n orbitals, a+ = (n,+ n,)lfl, with the two doubly allylic a bond orbitals is very strong and pushes its energy above that of the out-of-phase combination, n - = (n, - n,)lfl, inverting the natural orbital order (Figure 7.28). This converts the normal orbital crossing characteristic of ordinary [2 + 21 cycloadditions into an abnormal orbital crossing, so that the characteristic configuration no longer is the lowest in energy, and a barrier in the potential energy surface results for this reaction path. (See Section 6.2.4.)
Furthermore, [2 + 21 cycloadditions to N=N and C=N double bonds have been described (Prinzbach et al., 1982; Albert et al., 1984). The regioselectivity of the addition of electron-rich olefins to cyclic ketoiminoethers such as 60 may be rationalized using perturbation theory (Fabian, 1985):
7.4.2 Regiochemistry of Cycloaddition Reactions
In the quadruply bridged derivative 59, the [2 + 21 photocycloaddition proceeds. This can be understood as due to the restoration of the natural order of the orbital energies, n+ below n-,by the effect of the propano bridges (Figure 7.28). These can be expected to bring the two double bonds closer in space, increasing the through-space interaction, and to introduce additional throughbond interaction opposed in sign to the interaction through the doubly allylic bonds (Gleiter and Karcher, 1988).
Mixed cycloadditions between different olefins are also observed. The regiochemistry follows a pattern that has been rationalized by consideration of the HMO coefficients (Herndon, 1974). Some typical examples of triplet-
An important aspect of photocycloaddition consists of its regio- and stereoselectivity. Thus, dimerization of substituted olefins generally yields head-to-head adducts 61a and 61b in a regiospecific reaction, while head-totail dimen 63 are obtained from 9-substituted anthracenes (62) in polar solutions (Applequist et al., 1959; cf. Kaupp and Teufel, 1980).
412
ORGANIC PHOTOCHEMISTKY
The portion of the head-to-head product increases in nonpolar solvents and in micelles (Wolff et at., 1983). The regioselectivity of 9-methylanthracene is temperature dependent, and a photochemical equilibrium exists between the two isomers (Wolff, 1985). Singlet acenaphthylene (64) stereospecifically gives the syn dimer 65a, while in the TI reaction the anti dimer 65b predominates. The syn-anti ratio can be influenced by solvents with heavy-atom effect (Cowan and Drisco, 1970b), as well as by micellar solvents (Ramesh and Ramamurthy, 1984). Secondary orbital interactions (broken arrows) in the photochemical dimerization of acenaphthylene. The primary orbital interactions are indicated by full double arrows. Figure 7.29.
Triplet quenchers such as 0, or ferrocene inhibit formation of the anti dimer, while the formation of the syn dimer is hardly affected. The syn-anti ratio in the triplet reaction was determined by comparing the outcome of the reaction with and without use of a quencher (Cowan and Drisco, 1970a). Many factors must be considered to explain these facts, not least the relative stabilities of the various possible excimers or exciplexes and the accessibility of the pericyclic funnels. (Cf. Scheme 3 in Section 6.2.3.) In addition, the geometrical structure at the funnel and the ground-state relaxation pathways originating in the decay region determine which products will be formed. Excimer and exciplex formation will mainly influence the excited-state reaction path. Simple perturbation theory suggests that in concerted singlet photocycloadditions, electronic factors always favor the most highly symmetric excimer affording head-to-head regiochemistry and syn or cis stereochemistry, since hom*o-hom*o and LUMO-LUMO interactions are decisive in excimer formation according to Section 5.4.2. (Cf. Figure 5.21.) Interaction between centers whose LCAO coefficients are either both large or both small, that is, the head-to-head addition, is then favored:
head-to-head addition
head-to-tail addition
Similarly, secondary orbital interactions favor the syn arrangement of two acenaphthylene molecules, as illustrated in Figure 7.29.
The formation of head-to-tail dimers such as those of 9-substituted anthracenes and similar species can be rationalized by the conjecture that in this case it is not the most favorable excimer that determines the stereochemical outcome of the reaction but rather, the most readily reached pericyclic funnel. This can be understood if the effect of substituents on the pericyclic funnel is first considered at the simple two-electron-two-orbital model level (3 x 3 CI, Section 4.3.1). The discussion of cyclodimerization of an olefin is restricted to the four orbitals directly involved in the reaction. To simplify the presentation, we shall initially ignore the possible effects of a diagonal distortion. Then, the localized nonbonding orbitals at the rectangular pericyclic geometry have equal energies. They correspond to the degenerate n MOs $9 and $, of cyclobutadiene.
Since they are occupied with a total of two electrons, the system is a perfect biradical (6 = 0) and the energy splitting A E between Soand S, is relatively large. (Cf. Figure 4.19.) Substitution in positions 1,3 or 2,4 removes the degeneracy (Section 6.2.1). and the biradical is converted into a heterosymmetric biradicaloid (6 # 0). According to the results in Section 4.3.3, this brings the S, and So states closer and is likely to reduce the energy of the S, state. For instance, ab initio calculations on perturbed cyclobutadienes indicate that the splitting vanishes almost completely when two CH groups on diagonally opposite corners of the cyclobutadiene (positions 1 and 3) are replaced by isoelectronic NH@groups, yielding a critically heterosymmetric biradicaloid (BonaOiC-Kouteckyet at., 1989). As a result of the rectangular geometry, the pericyclic "minimum" at the head-to-tail dimerization path, which corresponds to a 1,3-disubstituted heterosymmetric biradicnloid, is likely to be deeper than that at the head-
to-head reaction path. This in turn should reduce the barriers around the minimum and thus make it easier to reach the head-to-tail rectangular "minimum" from the excimer. These arguments would suggest that head-to-tail regiochemistry is favored when access to the pericyclic "minimum" determines the product formation, while the head-to-head product is expected if the energy of the excimer intermediate is decisive for the reaction path (BonaCiC-Koutecky et al., 1987). Next, we consider also the effects of rhomboidal distortions, which permit a diagonal interaction. These may either reinforce or counteract the effect of the substituents, depending on which of the two diagonals has been shortened. For a head-to-tail approach of two substituted ethylenes, there will be two funnels corresponding to 1,3-disubstituted critically heterosymmetric biradicaloids, one at less and one at more diagonally distorted geometry than for the unsubstituted ethylene. This is confirmed by the results of calculations shown in Figure 7.30. In agreement with expectations from the two-electron-two-orbital model (lower part of Figure 7.30), the effect of donor substituents reinforces the orbital splitting produced by a 1.3-diagonal interaction, which brings the
a=24.5 ' E,I = 0.0
a=-15.S0
a =11.7'
Ere,= 22.0
Ercl = 15.2
a=-221' Ere, = 0.0 kcallmol
I
2.4 interaction
1.3 interaction
2.4 interact~on
1.3 interaction
Figure 7.30. Geometries of the conical intersections for the syn head-to-tail [2 + 21 cycloaddition of two aminoethylene molecules (left) ant1 of two acrylonitrile molecules (right), and a schematic representation of the energies of the cyclobutadienelike "nonbonding" biradicaloid orbitals. The effect of the donor (D) and acceptor (A) substituents on orbital energies is incorporated in the levels shown in the center of each half of the diagram. The effect of the diagonal interaction required to reach the critical value of the orbital energy difference (6,. cf. Section 4.3.3) is indicated by broken arrows.
substituted atoms closer to each other, while the effect of acceptor substituents opposes that of the 1,3 interaction. The opposite is true for the 2,4 interaction, which places the substituted atoms apart. As a result, a small rhomboidal distortion is sufficient to reach the "critically biradicaloid" geometry at which S, and Soare degenerate when the l ,3 diagonal is short for donor substituents and when the 2,4 diagonal is short for acceptor substituents. A large distortion is needed for donor substituents if the short diagonal is 2,4 and for acceptor substituents if it is 1,3. Numerical calculations yield a lower energy for the conical intersection when the substituent effect is opposed to that of the rhomboidal distortion, that is, at the more strongly distorted geometries (Klessinger and Hoinka, 1995). For small rhomboidal distortions, the peripheral bonding will dominate at the funnel geometry after return to So, and the formation of the head-to-tail product is likely to be favored. The diagonal bonding will dominate for large distortions, however, and head-to-head product could possibly be formed by x[2 + 21 cycloaddition-that is, by formation of one diagonal bond and subsequent closure of the other. Thus, the result of the photoaddition will depend on the distorted rhomboidal geometries, that is, on the relative heights of the barriers that lead to the various funnels, typically via an excimer or exciplex minimum. That is where steric effects as well as abnormal orbital crossing effects (cf. Example 7.9) may come into play. Finally, as we pointed out in Example 6.2, a number of different groundstate trajectories leading to one or the other of the possible photoproducts may emanate from the apex of the cone. This is illustrated schematically in Figure 7.31, which shows energy contour maps for the S, and So states of the syn head-to-tail approach of two acrylonitriles. Three valleys starting at the conical intersection are seen on the ground-state surface (Figure 7.31b), one leading to the cyclobutane minimum (B), the other one to the diagonally bonded product (D), and a broad slope back to the reactants (R). The steric outcome of the reaction will depend on the way in which the system enters the funnel region (Figure 7.3la) and on the dynamics of the passage through the cone. A model for estimating the barrier height between the excimer minimum and the pericyclic minimum for a given geometrical arrangement of the adducts, which ignores the effects just mentioned, was introduced in Section 6.2.3. The resonance integral fir,) for the bonds formed in the cycloaddition reaction may be estimated according to Equation (6.3) from the singlet and triplet excitation energies and the HMO coefficients of the hom*o and LUMO. A small absolute value will correspond to early surface crossings (cf. Figure 6.17) and thus to low barriers, and high reactivity. Table 7.2 summarizes some results obtained from this model. I t is seen that aromatic compounds and olefins undergo facile dimerizations when 1 P(rc)I < 20 kcaltmol, while for 1 f i r ) 1 = 20-24 kcaltmol the reactivity is moderate, and for 1 firc) 1 2 24 kcaltmol dimerizations have so far not been reported.
Table 7.2
Dimerization of Olefins and Aromatic Compounds: Singlet and Triplet Excitation Energies Es and E,, and Interaction Integral I Kr,) I (kcaVmol) ET
ICs
Anthracene Stilbene 1.3-Cyclohexadiene Simple alkenes Acenaphthylene Naphthalene Pyrene Phenanthrene
75.6 85.3 91.3
42.6 49 53
121 61.1 90.9
78 43--47 60.9
77.0 82.4
48.3 61.7
Caba
I
I
+ +
1.55* 1.55 2.88* 1.98 4.04 1.45 1.41* 0.98 0.72 1.41
I++
C a b = C (awob, en
+ aw,b,,)
for [2
+ 21 or (4
other reactions
2
-
-
a
Exp.
-
--
+ 41 cycloadditions; asterisks indicate result~for
the latter (adapted from Caldwell, 1980).
+
In discussing the state correlation diagram of the H, H,reaction, which can serve as a model for ethylene dimerization, it has been pointed out that the doubly excited state corresponds in the limit to an overall singlet coupling of two initial molecules 'M*in their triplet-excited states. As a consequence, the pericyclic minimum should be accessible not only from the excimer minimum but also directly through triplet-triplet annihilation. This was demonstrated for anthracene dimerization by determining the quantum yield as a function of the light intensity. The probability of excimer formation by triplet-triplet annihilation was determined experimentally as p l = 0.115, which is in good agreement with the spin statistical factor 119 and is accounted for by noting that apart from the singlet complex, triplet and quintet complexes are also formed upon triplet-triplet encounter. (Cf. Section 5.4.5.5.) Only the singlet species can yield the pericyclic intermediate, whereas the other complexes decay into monomers in the ground state or in a singlet- or triplet-excited state (Saltiel et al., 1981). It is likely that triplet-triplet annihilation also plays a role in other photochemical reactions. This is especially true at high concentrations and high intensities of the exciting light. The sensitized photodimerization of anthracene presumably proceeds entirely through triplet-triplet annihilation.
7.4.3 Cycloaddition Reaction of Aromatic Compounds Energy contour maps for the face-to-face [2 + 21 cycloaddition of two acrylonitrile molecules in the syn head-to-tail arrangement, showing (a) the lowest excited state S, and (b) the ground state So. Possible reaction paths via the low-energy conical intersection are indicated by arrows. Figure 7.31.
\
416
Photodimerizations are observed not only for olefins, but also for aromatic compounds, allenes, and acetylenes. The photodimerization of anthracene,
which may be considered to be a ground-state-forbidden [,4, + ,4,1 cycloaddition, was in fact described as early as in 1867 (Fritsche, 1867). Example 7.10: Much information about the detailed mechanism of anthracene dimerization was gained in the study of intramolecular photoreactions of linked anthracenes such as a,w-bis(9-anthryl)alkanes (66). I t was shown that luminescence and cycloaddition are competing pathways for the deactivation of excimers. In compounds with sterically demanding substituents R and R' that impair the cycloaddition reaction, the radiative deactivation is enhanced (H.-D. Becker, 1982).
The [2 + 21 cyclodimerization of benzene has been studied theoretically (Engelke et al., 1984) and used practically in the synthesis of polycyclic hydrocarbons such as 69 (Fessner et al., 1983). This represents a first step in the synthesis of the undecacyclic hydrocarbon pagodane (70), which in turn can be isomerized to give dodecahedrane (Fessner et al., 1987).
Mixed photocycloadditions of anthracene and conjugated polyenes yield products that correspond to a concerted reaction path, as well as others that are Woodward-Hoffmann-forbidden and presumably result from nonconcerted reactions. For example, the reaction of singlet-excited anthracene with 1,3-cyclohexadiene yields small quantities of the [,4, + ,2,] product 72 ,4,] product 71. in addition to the allowed [,4, Photocyclization of bis(9-anthryl)methane (67) and the corresponding photocycloreversion were shown by picosecond laser spectroscopy to have a common intermediate whose electronic structure is different in polar or nonpolar solvents as judged by different absorption spectra in various solvents (Manring et al.. 1985). a,af-Disubstituted bis-9-anthrylmethyl ethers (68) are characterized in their meso form by mirror-plane symmetry (a)and perfectly overlapping anthracene moieties (68a). The corresponding racemic diastereomers (68b) assume a conformation having a twofold axis of symmetry (CJ. Both the meso and racemic diastereomers cyclomerize. However, whereas the meso compounds are virtuall y nonfluorescent, the racemic diastereomers deactivate radiatively both from the locally excited state and from the excimer state. Thus, in the emitting excimer state of linked anthracenes the aromatic moieties overlap only partially. Apparently, the formation of luminescent excimers from bichromophoric aromatic compounds is associated with perfectly overlapping n systems only when intramolecular cycloaddition is an inefficient process (H.-D. Becker, 1982).
+
The yield of 71 increases with increasing polarity of either the solvent or the substituent X in the 9-position of anthracene. This result has been explained by invoking a stabilization of the exciplex, which should reduce the barrier between the exciplex minimum and the pericyclic minimum. The observation that in the presence of methyl iodide 72 becomes the major product for X = H, due to the heavy-atom effect, is compatible with the obvious assumption that the multistep process is a triplet reaction. However, if X = CN, no methyl iodide heavy-atom effect is observed (N. C. Yang et al., 1979, 1981).
420
ORGANIC PHOTOCHEMIS'I'KY
Addition of an alkene to benzene can occur in three distinct ways indicated in Scheme 17: ortho, meta, or para.
In general, dienophiles produce the ortho product in addition to small amounts of the para product. Maleic anhydride gives the 1:2 adduct 73 (Gilbert, 1980). Maleimides undergo analogous reactions. Both classes of compounds form CT complexes with benzene and its derivatives, and the reaction can be initiated by irradiation into the CT band (Bryce-Smith, 1973). Alkylethylenes, on the other hand, produce the meta adducts.
Figure 7.32. Orbital energy diagram for the photocycloaddition of excited benzene to an olefin a) with an electron-donating substituent X and b) with an electron-withdrawing substituent Z (adapted from Houk, 1982).
The selectivity may be explained by a consideration of the possible interactions of the degenerate benzene hom*o, &, and LUMO, $,., &,. (cf. Section 2.2.5 and Figure 2.33), with the ethylene n and n* MOs (Houk, 1982). In the ortho approach, the benzene MOs @a and &,. can interact with the ethylene MOs n and n*, while in the meta approach, and @a. can interact with nand d.Therefore, the configuration. , @ , is predicted to be energetically favored along the ortho cycloaddition path, while the configuration. , @ , is stabilized for the ortho and particularly the meta cycloaddition path. The magnitude of the stabilization depends on orbital overlap and relative orbital energies. The former factor favors the ortho approach and the latter the meta approach of the reactants. Altogether, one arrives at the prediction that for the addition of ethylene to the S, ('B,,) state of benzene, whose wave function can be written as (a,,. - @,,,.)/fl, comparable amounts of ortho and meta adducts should be formed, while the meta addition should dominate whenever' the alkene MOs n and n* are close in energy to the benzene MOs &, and @,., &..
+,,
+,,
The presence of substituents alters the situation in that charge transfer between the reaction partners may become significant, and ortho addition should then be preferred over meta addition. For electron-rich alkenes or electron-poor arenes, the ethylene IT MO lies at much higher energy than the singly occupied bonding MO of the excited arene ($a for the ortho path), and thus charge transfer from the alkene to the arene will take place. For an electron-poor alkene or an electron-rich arene, charge transfer from the arene to the alkene will occur (Figure 7.32). This analysis does not require a specification of the timing of the bondforming events. This could correspond (1) to a fully concerted, synchronous pathway, (2) to the cyclization of benzene to a biradical referred to as prefulvene followed by addition of the olefin, or (3) to the bonding of the olefin to meta positions and subsequent cyclopropane formation, as indicated for the case of meta cycloaddition in Scheme 18. Mechanism 2 of Scheme 18, which was first proposed in 1966 (BryceSmith et al., 1966), has been discarded because recent experimental evidence excludes the intermediacy of prefulvene (Bryce-Smith et al., 1986). Mechanism 3 is presently favored. Experimental and theoretical results sup-
( ) I \ ,rU\1IC I'HO'I'OCHEMISI'KY
dition is expected. Ortho cycloaddition is favored for AGE, values up to about 1.4-1.6 eV, and meta cycloaddition for more strongly positive values (Scheme 20).
port the notion that the a bonds between the reactants are formed during the initial stage of the reaction, while the cyclopropane ring closure is not expected until crossing to the ground-state energy surface (De Vaal et al., 1986; van der Hart et al., 1987). A fully concerted process (mechanism 1 in Scheme 18) is considered unlikely. An argument against it, and in favor of mechanism 3, is provided by independent photochemical generation of the proposed biradical intermediate (Scheme 19), which yields the same product ratio as is obtained from the corresponding arene-alkene photocycloaddition at low conversions (Reedich and Sheridan, 1985).
Scheme 20
Regioselectivity in the meta cycloaddition to substituted benzenes has been assumed to depend on charge polarization in the biradical intermediate shown in Scheme 21 (van der Hart et al., 1987). The theoretical calculations mentioned earlier, however, point out that in the early stage of the reaction the excited state has appreciable polar character that disappears as the reaction proceeds; the biradical itself does not have any particular polarity.
Irradiation of mixtures of various acetylenes with benzene gives cyclooctatetraenes, presumably via an intermediate ortho adduct 74 (Bryce-Smith et al., 1970). For acetylenes with bulky substituents, the bicyclooctatriene intermediate is sufficiently stable for a subsequent intramolecular photocycloaddition to a tetracyclooctene (75) (Tinnemans and Neckers, 1977):
Furthermore, the occurrence of an exciplex intermediate is assumed; an empirical correlation based on the AGE, values calculated from the Weller relation, Equation (5.28), has been established which allows the prediction of mode selectivity for a wide range of arene-alkene photocycloadditions (Mattay, 1987). For negative or very small values of AGE,, addition is preferred over cycloaddition. This is the case for the reactions of certain electron-rich alkenes with excited arenes. For positive values of AGE,, cycload-
ORGANIC PHOTOCHEMISTRY
424
I
7.4.4 Photocycloadditions of the Carbonyl Group Another photocycloaddition reaction that has been known for a long time is the Paterno-Buchi reaction, which involves the formation of oxetanes through the addition of an excited carbonyl compound to olefins:
As far as the addition of aromatic carbonyl compounds is concerned, only the triplet state is reactive, and hydrogen abstraction occurs as a side reaction. Another competing reaction path involves energy transfer (cf. Section 7.1.2); efficient oxetane formation is therefore observed only when the triplet energy of the carbonyl compound is not high enough for sensitized triplet excitation of the olefin. As is to be expected from the loose geometry of a triplet biradical intermediate, the reaction is not stereospecific; the same mixture of oxetanes 76 and 77 is produced at low conversion by irradiating cis- or trans-2-butene with benzophenone (Turro et al., 1972a; Carless, 1973): Figure 7.33. Mechanism of the photochemical oxetane formation: k,, k,,, and kc are rate constants of triplet biradical formation, intersystem crossing, and cyclization; kipand k, correspond to the formation of a contact ion pair and solvent-separated ion pairs; and kS and k,, are rate constants of the back reaction of the singlet biradical and back-electron transfer, respectively (by permission from Buschmann et al., 1991).
That the reaction is partly regiospecific, as indicated in Scheme 22, has been attributed to the differing stabilities of the biradical intermediates (Yang et al., 1964).
pair may also dissociate to a solvent-separated ion pair (k,) or return to the reactants by back-electron transfer (k,,). The triplet biradical intermediate was identified by picosecond spectroscopy (Freilich and Peters, 1985). A mechanism involving the formation of polar exciplexes in the first step, as well as electron-transfer processes, has been invoked in the interpretation of biacetyl emission quenching by electron-rich alkenes (Mattay et al., 1984b; Cersdorf et al., 1987). Diastereomeric oxetanes are formed from chiral carbonyl compounds such as menthyl phenylglyoxylate (78) (Buschmann et al., 1989).
Scheme 22
The mechanism of the photochemical oxetane formation is summarized in Figure 7.33. The 3(n,n*) excited state of the ketone produced by light absorption and subsequent intersystem crossing (ISC) attacks the olefin to form a triplet 1,Cbiradical (k,,). Intersystem crossing (kist) to the singlet, either directly or via a contact ion pair (kip),leads to ring closure (kc)to form the oxetane, or to /3 cleavage to re-form the reagents (k8). The contact ion I
The general kinetic scheme (Fig. 7.34) displays two stages of diastereoselection: (1) a preferred formation of that of the two diastereomeric 1,4-
0I:CANIC PHOTOCHEMISTRY
rn
e ndo
Figure 7.35. Preferred conformations for intersystem crossing during the diastereoselective oxetane formation (adapted from Griesbeck and Stadtmiiller, 1991).
Figure 7.34. A simplified kinetic scheme for the diastereoselective oxetane formation in the Paterno-Bilchi reaction (k, and k; as well as k, and k; include intersystem crossing steps) (by permission from Buschmann et al., 1989). biradicals (k, > k;) which leads to the majority oxetane product, and (2) a preferred /3 cleavage of the biradical that leads to the minority oxetane product (k; > &,), or a preferred ring closure of the biradical, which leads to the majority oxetane product (k, > k;). Upon continued irradiation, the Pcleavage products reenter the reaction cycle, establishing a "photon-driven selection pump." The Arrhenius plot for the diastereoselectivity (cf. Section 6.1.5.2) contains two linear regions, one with a positive and one with a negative slope, changing into each other at the "inversion temperature," Tin,. Each region selection has a different dominant selection step. At temperatures above Tin", is driven primarily by enthalpy, as expected for an early transition state in the bond cleavage in the energy-rich biradical intermediate. Below Ti,,,selection is driven primarily by entropy, as expected for an early transition state in the bond-forming step in the formation of the biradical from the energy-rich reactants (Buschmann et al., 1989). Example 7.11: An interesting example of diastereoselectivity is provided by the photocycloaddition of aromatic aldehydes to electron-rich cyclic olefins such as 2.3dihydrofuran (79):
The stereochemistry of the reaction can be accounted for by the conformational dependence of .;pin-orhit cor~plinpelements dihc~~rred in Section 4.3.4.
Figure 7.35a shows the postulated geometry of favored intersystem crossing in the biradical derived from an unsubstituted cycloalkene; bond formation following spin inversion is faster than conforrnational changes, and the endo adduct is formed from the less sterically hindered conformation. The exoselectivity observed in I-substituted cycloalkenes is explained by increasing gauche interactions with the Balkoxy group that favor the biradical conformations shown in Figure 7.35b (Griesbeck and Stadtmiiller, 1991). The photoaddition of a furan and an aldehyde can serve as a photochemical version of a stereoselective aldol reaction, since the photoadduct can be viewed as a protected aldol, as indicated in Scheme 23 (Schreiber et al., 1983).
Scheme 23
For aliphatic ketones, the situation is complicated by less efficient intersystem crossing, thus permitting reaction of the '(n,n*) as well as the '(n,llr) state of the carbonyl compound, as revealed by the use of triplet quenchers. The S, reaction is more stereospecific, presumably because of the tight geometry of the singlet biradical, and yields less cis-trans isomerization by a competing path. Thus, the stereochemistry of cis-I-methoxy-I-butene (80) is partially retained when acetone singlets attack the olefin, but it is almost completely scrambled in the reaction of triplet acetone (Turro and Wriede, 1970):
ORGANIC PHOTOCHEMISTRY
428
The situation is different with electron-poor olefins such as 1,2-dicyanoethylene. Only the S, state of acetone forms oxetanes, and the reaction is highly stereospecific, as indicated in Scheme 24. The competing cis-trans isomerization of the olefin arises exclusively from the triplet state (Dalton et al., 1970).
Scheme 24
+ NCAcN
CN
NC
The specificity of this reaction has been used to chemically rirrate both the excited-singlet acetone and the triplet acetone produced through thermal decomposition of tetramethyl-1,2-dioxetane.(Cf. Section 7.6.4.) For this purpose the thermolysis was carried out in the presence of trans-1 ,Zdicyanoethylene, and the quantities of singlet and triplet acetone formed were obtained from the yields of dioxetane and cis-1,2-dicyanoethylene, respectively (Turro and Lechtken, 1972). The various reactions of excited carbonyl compounds with olefins may be rationalized on the basis of correlation diagrams. In principle, four different pathways have to be discussed: the perpendicular and the parallel approaches (Figure 7.36) and the initial formation of a CO and a CC bond, yielding a C,C-biradical and C,O-biradical, respectively:
However, only three out of these four possibilities are realistic, since there is no carbonyl group orbital available for a perpendicular approach a ~ ford mation of a CC bond. Formation of a C,O-biradical is therefore possible only through a parallel approach.
Perpendicular approach
Parallel approach
The perpendicular and the parallel approach for the interaction of an mn*excited ketone with a ground-state olefin.
Figure 7.36.
Correlation diagram a) for the perpendicular approach of ketone and olefin, and for the parallel approach resulting in b) a C,C-biradical and c) a C,Obiradical. Figure 7.37.
From Figure 7.37a, the perpendicular approach leading to a C,C-biradical is seen to be electronically allowed. Since, however, the two p AOs of the unpaired electrons are oriented perpendicular to each other, a rotation of the CH, group of the ketone is required prior to cyclization both for the singlet biradical and for the triplet biradical, which assume similar geometries and therefore can each give an oxetane. Small rotations in the less sterically hindered direction to optimal geometries for ISC and subsequent reaction may lead to stereospecific oxetane formation from the triplet state. (Cf. Example 7.11 .) Correlation diagrams for the two modes of parallel approach are shown in Figures 7.37b and 7.37~.If the CO bond is formed first, the (n,n*) excited reactant states correlate with highly excited (n,aGo) states of the products and a correlation-induced barrier results. Hence this reaction is electronically forbidden (Figure 7.37b). If, however, the CC bond is formed first, the unpaired electron at the oxygen can be localized in a p A 0 with either a or n symmetry, that is, either in the fco MO or in the no orbital. Since the '.3(n,n*)reactant states correlate with the 'v3B,,., product states, no correlation-induced barrier is to be expected, and the reaction is likely to be exothermic (Figure 7.37~).In contrast to the perpendicular approach, the triplet biradical of the parallel approach will have a loose geometry and should result in cis-trans isomerization of the olefin. The conclusions from the correlation diagrams have been nicely confirmed by early ab initio calculations for the carbon-xygen attack of formaldehyde on ethylene (Salem, 1974)and by more recent calculations on the same model reaction considering both modes of attack. (Palmer et al., 1994).
The results for the carbon-oxygen attack are summarized schematically in Figure 7.38. The excited-state branch of the reaction path terminates in a conical intersection point at a CO distance of 177 pm before the biradical is fully formed (cf. Figure 7.37a). Thus the system can evolve back to the reactants or produce a transient C,C-biradical intermediate that is isolated by small barriers (< 3 kcallmol) to fragmentation (TS,) or to rotation and ring closure to oxetane (TS,). The singlet and triplet biradical minima are essentially coincident. A schematic representation of the surfaces for the carbon-carbon attack is shown in Figure 7.39. The very flat region of the S,, surface (barriers of the order of I kcallmol) corresponds to the '.3B,., C.0-biradical. The '.3B,, biradical has a CC bond length of 156 pm and corresponds to a conical intersection geometry in the case of the singlet, and to a minimum in the case of the triplet. Thus for the singlet photochemistry the decay to So occurs close to the products, and the reaction appears to be concerted. Since, however, the formation of the singlet biradical is also possible from the same funnel, a certain fraction of photoexcited reactant can evolve via a nonconcerted route. Figure 7.39. Photoaddition of formaldehyde and ethylene. Schematic representation of the carbondxygen attack as a function of the C,C distance R,, and the dihedral angle rp between the formaldehyde and ethylene fragments (by permission from Palmer et al., 1994).
Photocycloaddition of formaldehyde and ethylene. Schematic representation of the surfaces involved in the carbon-oxygen attack as a function of the C,O distance R, and the dihedral angle rp between the formaldehyde and ethylene fragments. )( and denote the parallel and perpendicular approach of the reactants, respectively; CI marks the conical intersection (by permission from Palmer et al., 1994). Figure 7.38.
The overall nature of the singlet reaction path is determined by whether the system returns to S,, through an early funnel along the path of carbonoxygen attack to produce a C,C-biradical intermediate (Figure 7.38) or through a late funnel along the path of carbon-carbon attack to produce oxetane in a direct process (Figure 7.39). The stereochemistry and efficiency of the reaction will depend on which of these two excited-state reaction paths is followed. Conclusions similar to those obtained from the correlation diagrams can be reached by means of either PMO theory (Herndon, 1974) or simple frontier orbital theory, as indicated in Figure 7.40. Orbital energies may be estimated from the assumption that the olefinic n MOs will be stabilized by electron-withdrawing substituents and destabilized by electron-releasing ones. The hom*o energy of an electron-rich olefin will then be comparable to that of the carbonyl n orbital, whereas the LUMO of electron-poor olefins will be similar in energy to the n* MO of the carbonyl group. Interactions between orbitals of comparable energy require a parallel approach in the case of electron-poor olefins and result in CC bond making and formation of a C,O-biradical. Perpendicular attack is required for electron-rich olefins and yields a C,C-biradical by CO bond formation.
ORGANIC PHOTOCHEMISTRY
7.4.5 Photocycloaddition Reactions of a,&Unsaturated Carbonyl Compounds Intermolecular photoadditions of a,@unsaturated carbonyl compounds can take place either at the CC or at the CO double bond. Photodimerizations with formation of a cyclobutane ring are quite common. In cases such as cyclopentenone (811, head-to-head as well as head-to-tail dimers are produced.
Electron-pwr olef ins
Electron - rich olef ins
Figure 7.40. Frontier orbitals for oxetane formation. Interaction of one of the halffilled carbonyl orbitals with the LUMO of electron-poor olefins (left) and with the hom*o of electron-rich olefins (right).
The magnitude of the LCAO MO coefficients of the interacting orbitals permits a prediction of the expected oxetane regiochemistry. Both in the hom*o of a donor-substituted olefin and in the LUMO of an acceptor-substituted olefin, the coefficient of the unsubstituted carbon atom is the larger one in absolute value. Therefore, electron-poor olefins regioselectively afford the oxetane with the substituted carbon next to the oxygen. (Cf. Barltrop and Carless, 1972.) In contrast, an electron-rich olefin predominantly yields the oxetane with the unsubstituted carbon next to the oxygen. (Cf. Scheme 22.) Finally, Figure 7.40 permits the conclusion that oxetane formation can be considered to be either a nucleophilic attack by a ketone on an electronpoor olefin or an electrophilic attack by a ketone on an electron-rich olefin, corresponding to the predominant interaction of a half-filled carbonyl orbital either with the empty n* MO or with the doubly filled rc MO of the olefin.
While double-bond cis-trans isomerization in Esubstituted a,&enones occurs at both 313 and 254 nm, oxetene forms only when short wavelengths are used (Friedrich and Schuster, 1969, 1972). This differentiation has been explained on the basis of calculations that revealed a S,-So conical intersection lying 15 and 10 kcallmol above the '(n,n*) minimum of s-trans- and scis-acrolein, respectively (Reguero et at., 1994). Oxabicyclobutane, however, has not been detected. This behavior of a,@enones is in contrast with that of butadiene, where cyclobutene and bicyclobutane are formed simultaneously. (Cf. Sections 6.2.1 and 7.5.1.) Mixed cycloadditions, for example, between cyclopentenone and ethyl vinyl ether, have also been observed frequently (Schuster, 1989). As indicated in Scheme 25, attack of the 3(3t.,n*)excited state of the enone on the olefin will give a triplet 1,4-biradical that ultimately yields cyclization and disproportionation products (Corey et at., 1964; De Mayo, 1971). The regio-
I
7.5
selectivity of the reaction, which was explained earlier in terms of an initial interaction of the enone triplet with an alkene to give an exciplex, can alternatively be accounted for by the relative efficiencies with which each of the isomeric biradical intermediates proceeds to the annelation products (as opposed to reverting to the ground-state enone). While the structural isomers 84 and 85 are produced in a ratio 1.0 : 3.1, trapping with hydrogen selenide H,Se indicates that the biradicals 82 and 83 are formed in the ratio 1 .O: I .O (Hastings and Weedon, 1991). Kinetic studies indicating that triplet 3-methylcyclohex-2-en-1-one (86) reacts with maleic or fumaric dinitrile directly to yield triplet 1,4-biradicals argue against the intermediacy of an exciplex (Schuster et al., 1991a). Dynamical properties of enone '(n,n*) states have been determined and discussed; torsion around the olefinic bond is the principal structural parameter that affects their energies and lifetimes (Schuster et a]., 1991b). Twisting around the C==C bond causes an increase in energy on the Sosurface and a concomitant decrease in energy on the '(n,n*) surface (cf. Figure 4.6). This reduces the TI-S,, gap and facilitates T,-S,, radiationless decay in a process dominated by spin-orbit coupling (Section 4.3.4). Average energies and lifetimes of the mixture of head-to-head and head-to-tail 1,Cbiradicals have been determined by photoacoustic calorimetry (Kaprinidis et al., 1993).
REARRANGEMENTS
435
characterized by a number of thermal as well as photochemical electrocyclic reactions and sigmatropic shifts, some of which are as follows:
Ergosterol
Precalciferol
Pyrocalciferol
The number of photoreactions that proceed in agreement with the Woodward-Hoffmann rules (Table 7.3) is very large. Some examples are collected in Scheme 26.
Synthetic applications of [2 + 21 cycloadditions of a#-unsaturated carbonyl compounds are numerous. The synthesis of cubane (Eaton and Cole, 1964), in which cage formation is achieved by the following photochemical reaction step, is an example: Scheme 26
7.5 Rearrangements 7.5.1 Electrocyclic Reactions Electrocyclic ring-opening and -closure reactions represent another important field in which the Woodward-Hoffmann rules apply. These rules were in fact derived as a rationalization of the chemistry of vitamin D, which is
Table 7.3 Orbital Symmetry Rules for Thermal (A) and Photochemical (hv) Electrocyclic Reactions Number of n Electrons in Open-Chain Polyene Reaction
4n Conrotatory Disrotatory
4n
+2
Disrotation Conrotation
436
ORGANIC PHOTOCHEMISTRY
The important role of avoided crossings and the resulting pericyclic minima for the mechanisms of photochemical reactions was first pointed out on the example of the butadiene-cyclobutene conversion (van der Lugt and Oosterhoff, 1969).
While the Oosterhoff model that follows from the state correlation diagrams discussed in Section 4.2.3 describes the stereochemistry of electrocyclic reactions correctly and in agreement with the Woodward-Hoffmann rules, it is oversimplified in that it does not attempt to actually locate the bottom of the pericyclic minimum and simply assumes a planar carbon framework. It therefore predicts a nonzero S,-So gap at perfect biradicaloid geometry. As discussed earlier (cf. Sections 4.4.1, 6.2.1, etc.) symmetry-lowering distortions that remove the exact degeneracy of the nonbonding orbitals at the perfect biradical geometry by introducing a heterosymmetric perturbation 6 lower the energy of the S, state and reduce the S,-So gap. Recent calculations on butadiene (Olivucci et al., 1993) have demonstrated that the gap is actually reduced to zero and that the funnel corresponds to a true S,-So conical intersection at geometries with pericyclic and diagonal interactions. According to these calculations, the ring closure of butadiene to form cyclobutene proceeds through the same funnel as the cis-trans isomerization, and the stereochemical decision is taken mainly on the excitedstate branch of the reaction pathway, where the low-energy pathway corresponds to disrotatory motion, as discussed in detail in Section 6.2.1. In fact, an excited-state cis-butadiene enters the conical intersection region in a conformation already consistent with the production of only one ground-state cyclobutene isomer (Figure 7.41). However, not all reactions that follow a Woodward-Hoffmann allowed path are necessarily stereospecific. (Cf. Clark and Leigh, 1987). Irradiation
Figure 7.41. Schematic representation of the geometry changes of excited state scis-butadiene a) entering the conical intersection region b) and producing groundstate cyclobutene c). Light and dark arrows indicate excited-state and ground-state pathways, respectively (adapted from Olivucci et al., 1993).
of bicyclo[4.2.0]oct-7-enes (87) yields the products shown in Scheme 27 (Leigh et al., 1991). Recent calculations suggest that the motion on the S, surface is disrotatory, as demanded by the Woodward-Hoffmann rules, and that a loss of steric information results from motion on the ground-state surface after return through the pericyclic funnel (Bernardi et a]., 1992~).The mechanism of the fragmentation step (also shown in Scheme 27) has not yet been elucidated in a definitive fashion.
Irradiation of s-cis-2,3-dimethylbutadieneyields both electrocyclic ring closure and double-bond isomerization products (Scheme 28). While the ring closure in s-cis-butadiene, isoprene, 2-isopropylbutadiene, and 1,3-pentadiene is considerably less efficient than s-cis-s-trans isomerization, in s-cis2,3-dimethylbutadiene it is about 50 times faster (Squillacote and Semple, 1990). This effect is counterintuitive, since the two methyl groups appear to hinder the rotation about the central C--C bond in butadiene.
Example 7.12: According to ab initio calculations of Olivucci et al., (1994b), the presence of the methyl substituents in s-cis-2,3-dimethylbutadienedoes not alter the nature of the pathway on S, to the pericyclic funnel. In both the parent and the dimethyl derivative, the atoms attached in positions 2 and 3 remain eclipsed. The different behavior could therefore be due to differences in the region of the Sosurface explored after return through the funnel, or to differences in the kinematics caused by the larger inertia of the methyl groups compared to hydrogen. The latter explanation is favored by Olivucci et al., but dynamical calculations are needed to settle the issue.
438
1
ORGANIC PHOTOCHEMISTRY
At the conical intersection the C,-C2-C3valence angle o is close to 90°, suggesting that in strained systems, where this value is hard to reach, the excited-state barrier on the pathway to the conical intersection o r the conical intersection itself may be s o high in energy that the photoreaction cannot take place. The quantum yields for cyclobutene formation in dimethylenecycloalkanes summarized in Scheme 29 (Aue and Reynolds, 1973; Leigh and Zheng, 1991) are in agreement with this expectation.
7.5
REARRANGEMENTS
439
Example 7.13:
The photochemical reaction network and photoequilibrium involved in the commercial synthesis of vitamin D using the photochemical electrocyclic ring opening of ergosterol to give precalciferol, which subsequently undergoes a thermal [1,7]-H shift to produce vitamin D itself (cf. Jacobs and Havinga, 1979). has been studied using a combination of molecular mechanics and simple 2 x 2 VB methods (Bernardi et al., 1992b). The results are summarized in Figure 7.42. The focal feature of the reaction network centered on precalciferol is the existence of three funnels analogous to those discussed for the cis-trans isomerization of cis-hexatriene in Section 7.1.3, except that there are two conical intersections CIZr and CIS for the s-cis-s-trans isomerization (c-t)since the
Bicyclo[l .l .O]butane is usually a side product of the photocyclization of butadiene to cyclobutene (Srinivasan, 1963); in isooctane, the quantum yield ratio is I : 16 (Sonntag and Srinivasan, 1971). It becomes the major product in systems in which the butadiene moiety is constrained near an s-trans conformation and bond formation between the two terminal methylene groups that leads to cyclobutene is disfavored. An example is the substituted diene 88 in Scheme 30, for which the bicyclobutane is the major product; a nearly orthogonal conformation should result from the presence of the 2,3-di-t-butyl substituents (Hopf et al., 1994).
HO Ergosterol
Me Toxisterol C2
Mechanistic scheme for the precalciferol reaction network including the three conical intersection regions CIA, CI; and CI,.,. Dark arrows indicate ground-state pathways, while light arrows indicate pathways on the excited-state surface (by permission from Bernardi et al., 1992b). Figure 7.42.
ORGANIC PHOTOCHEMISTRY
The electrocyclic ring-closure reaction proceeds exclusively in the S, state and yields solely the trans product by a conrotatory mode of reaction, as is to be expected from Table 7.3 for a 6n-electron system. Competing reactions are cis-trans isomerization (cf. Section 7.1.4) and intersystem crossing to T,. From the T, state, generally only cis-trans isomerization is observed. Stilbenes with substituents that enhance spin inversion, such as Br, RCO, and NO,, do not undergo the cyclization reaction efficiently. This reaction is not limited to stilbene itself. Whether a diarylethylene may be cyclized, and if so, in which way, can be predicted from the sum F* = 6' of the free valence numbers $*= fl - CQ; calculated for tre reacting positions Q and o of the hom*o-LUMO excited reactant. Here, C a is the sum of excited-state n-bond orders for all bonds p-g' originating from atom e. Laarhoven (1983) derived the following rules:
c*+
Figure 7.43. Illustration of the NEER principle: interconversion of the excited-state conformers of precalciferol is prohibited by the high probability of decay to the ground state through the conical intersection CI,., (by permission from Bernardi et al., 1992b). s-cis-Z-s-cis isomer of precalciferol has two diastereomers, which may be denoted as cZc + and cZc - . The possibility of bifurcation of the reaction path on the So surface corresponding to the various possible bond-forming schemes (indicated in Figure 7.42 by dark arrows) and the existence of several pathways in the S , state (light arrows) are the reasons for the great diversity of photoproducts from precalciferol. In agreement with the existence of three ground-state conformers cZc+, cZc- and cZt of precalciferol, there are three excited-state minima, cZc+*, cZc-*, and cZt*. Search for transition structures between these excited-state minima led toward the funnels Cl;, and C I , . This work explains the short singlet lifetime of an excited-state conformer, as well as the inability of excited conformers to interconvert, as stated by the NEER (nonequilibration of excited rotamers) principle. (Jacobs and Havinga, 1979; see also Example 6.9.) The interconverting conformer (e.g., cZc*) will have high probability of decaying to the ground state through the conical intersection CI,., before the barrier for interconverting to cZt* could be completely overcome. This is illustrated qualitatively in Figure 7.43.
The photocyclization of cis-stilbene gives dihydrophenanthrene (5), which may be thermally or photochemically converted back to cis-stilbene or oxidized to phenanthrene (Moore et al., 1963; Muszkat, 1980).
d=i
I . Photocyclizations do not occur when F; < 1 .O.
2. When two or more cyclizations are possible in a particular compound, only one product arises if one value is larger by at least 0.1 than all the others. 3. When planar as well as nonplanar products (pentahelicene or higher helicenes) can be formed, the planar aromatic will in general be the main product regardless of rule 2, provided rule 1 is fulfilled for its formation. The application of these rules is illustrated in the following example.
Example 7.14:
The excited-state n-bond orders of an alternant hydrocarbon are given by
where pw, is the n-bond order of the bond between atom e and its neighbor g', while cflo and c~.,, are the corresponding LCAO coefficients of the hom*o. Using q'= fl - 1.282 = 0.450, 4*= 0.620,and 6,.= 0.491, F ; , = 0.941, and 4;. = 1.1 1 1 are found for compound 89. From rule I only ring closure between positions 4 and 2' to yield the hydrocarbon 91 is expected, and no formation of 92 by ring closure between positions 2 and 2' should occur. ,. = 1.095 and = 1.188; from rule I, formation Similarly, for 90 one has & of both 91 and 92 could be expected, but since 91 is nonplanar, 92 should be the main product according to rule 3. Experimental results show that only products allowed by rules 1-3 are formed: 89 gives 91 in 60% yield and 90 yields only the planar compound 92, in 75% yield (Laarhoven, 1983).
e,;,.
7.5
REARRANGEMENTS
443
A situation similar to that of stilbene is found for dimethyldihydropyrene (%), which is converted into the dimethyl-substituted metacyclophanediene
97 by irradiation with visible light (A = 465 nm); complete back reaction to the reactants is achieved either with lower wavelength light (A = 313 nm) or thermally.
The analogous photocyclization of N-methyldiphenylamine has been studied in detail (Forster et al., 1973; Grellmann et al., 1981) and utilized synthetically (Schultz, 1983). In contrast to stilbene, the reaction proceeds from the triplet state of the amine by an adiabatic conrotatory ring closure to give a dihydrocarbazole (93), in accordance with the Woodward-Hoffmann rules. After return to the ground state, the initial product is oxidized to a carbazole (Scheme 31):
The two systems differ insofar as for stilbene the thermodynamic stability of the "open" form is greater, while for dihydropyrene the "ring-closed" form is more stable. The thermal back reaction 97 -, 96, which requires an activation energy of 22 kcal/mol, is apparently a symmetry-forbidden concerted reaction. An estimate of the ground-state barrier based on EHT calculations yields values that are in good agreement with experimental activation energies (Schmidt, 1971). In general, electrocyclic ring-closure and ring-opening reactions are singlet processes, believed to proceed via tight biradicaloid geometries. In contrast, triplet excitation frequently yields five-membered rings (see the rule of five, Section 7.4.1) via loose biradicaloid intermediates. For example, singlet cyclohexadiene gives bicyclo[2.2.0]hexene (98) and hexatriene, while triplet hexatriene yields bicyclo[3.1.O]hexene (99) (Jacobs and Havinga, 1979):
I Scheme 31
CH3
This reaction is of special interest because it is a stereospecific triplet reaction. Another known example is the ring opening by CC bond cleavage of the heterocyclic three-membered rings aziridine (94), oxirane (95). etc. (Huisgen, 1977; Padwa and Griffin, 1976). These reactions are believed to proceed through a cyclic antiaromatic triplet minimum which is isoelectronic with C,H,@; this in turn is an axial biradical with E(T,)< E(S,). (Cf. Section 4.3.2.) It has been suggested that the unusual degree of triplet stabilization at the tight pericyclic geometry in axial biradicals is responsible for the stereospecific course of the reaction (Michl and BonaCiC-Kouteckg, 1990).
The reactions of butadiene are very typical. As indicated in Scheme 32, direct irradiation yields predominantly cyclobutene; in the presence of Cu(1) or Hg, however, bicyclobutane formed by an x[2 + 21 process is the major product, with minor products formed through [2 + 21 cycloaddition. (Cf. Section 7.4.2.)
m--> Scheme 32
=%
- la*1-hv
1 '*I
444
ORGANIC PHOTOCHEMISTRY
Example 7.15: Direct irradiation of myrcene (100) gives the cyclobutane derivative 101 by an electrocyclic ring-closure reaction, together with Ppinene (102) formed in a [2 21 cycloaddition process. The sensitized reaction, however, yields the bicyclic compound 103.
biradical and products of its further transformation. These are indeed not observed.
+
7.5.2 Sigmatropic Shifts
The triplet reaction is assumed to proceed as a two-step process with the most stable five-ring biradical as intermediate according to the "rule of five" (cf. Section 7.4.1) (Liu and Hammond, 1%7).
The kinetics of the photochemical ring opening of cyclic dienes and trienes such as 1,3,5-cyclooctatriene (104) were determined by picosecond time-resolved UV resonance Raman spectroscopy (Ried et al., 1990) and provide excellent direct support for the Woodward-Hoffmann rules. The photochemical ring opening of cyclohexa-1,3-diene involves a rapid (10 fs) radiationless decay of the initially excited 1B, spectroscopic state to the lower 2A,, presumably through an S,-S, conical intersection reached by a conrotatory motion, in agreement with Woodward-Hoffmann rules (Trulson et al., 1989). According to the latest calculations (Celani et al., 1994), this is followed by further conrotatory motion on the 2A, surface along a pericyclic path distorted in the usual diagonal way from twofold symmetry, to a shallow minimum located most of the way toward the ringopen geometry of s-cis-Z-s-cis-hexatriene. As judged by the experimentally observed 6 ps appearance time of the ground-state 1,3,5-hexatriene product (Reid et al., 1993), the molecule has the time to equilibrate vibrationally in this shallow minimum before it decays to the ground state, presumably mostly via thermal activation to a more strongly diagonally distorted pericyclic conical intersection area located only about 1 kcallmol higher in energy. As is common for pericyclic funnels, this return to So is about equally likely to be followed by relaxation to the starting 1,3-cyclohexadiene and to the product, s-cis-Z-s-cis-hexatriene. The calculations assign little if any importance to an alternative ground-state relaxation path that would preserve the diagonal 1,5 interaction and lead to the 5-methylenecyclopent-2-en-I-yl
Sigmatropic shifts represent another important class of pericyclic reactions to which the Woodward-Hoffmann rules apply. The selection rules for these reactions are best discussed by means of the Dewar-Evans-Zimmerman rules. It is then easy to see that a suprafacial [1,3]-hydrogenshift is forbidden in the ground state but allowed in the excited state, since the transition state is isoelectronic with an antiaromatic 4N-Hiickel system (with n = I), in which the signs of the 4N AOs can be chosen such that all overlaps are positive. The antarafacial reaction, on the other hand, is thermally allowed, inasmuch as the transition state may be considered as a Mobius system with just one change in phase.
suprafacial
antarafacial
[I .3]hydrogen shift
For sigmatropic shifts of organic groups other than hydrogen, inversion of configuration at the migrating site is possible and introduces another change in phase. No phase change occurs if the configuration is retained. The resulting rules are summarized in Table 7.4.
Table 7.4 Orbital Symmetry Rules for Thermal (A) and Photochemical (hv) Sigmatropic [i, j1 Shifts Reaction
I + j = 4 n
A
supra Inv." antara Ret. supra Ret . antara Inv.,'
hv
+ + + +
i + j = 4 n supra-antara antara-supra supra-supra antara-anlara
" Hydrogen shifts with inversion are not possible.
I + j = 4n 2
+ srrprcc + Ret . untara + Inv." supra + lnv." onrara + Ret.
i+j= 4n + 2 sicpru-srrpra antara-untura supra-antarc1 anrura-supra
Interesting examples of sigmatropic shift reactions are provided by the rearrangements of 1,5-dienes, which, as indicated in Scheme 33 for geranonitrile (105), undergo a [3,3] shift (Cope rearrangement) in the ground state and a [1,3] shift in the singlet-excited state, while the triplet excited state undergoes an x[2 + 21 addition via a biradical, in agreement with the rule of five (Section 7.4.1). Stereospecificity of the singlet reaction, as well as results of deuterium labeling experiments, verify the expectations from the selection rules for concerted photochemical reactions (Cookson, 1968).
Scheme 33
(1 9%)
(8 1%)
However, mixtures of [I ,2]- and [I .3]-shift products may be obtained by irradiating the same reactant, as shown in Scheme 34 for the direct irradiation of 1,5-hexadienes (Manning and Kropp, 1981). Configuration and state correlation diagrams for the [I ,3] shift show that a pericyclic geometry in S, is indeed present, making the excited-singlet reaction quite analogous to the disrotatory interconversion of butadiene and cyclobutene discussed in the preceding section. Appropriate diagonal distortions of the pericyclic geometry may again lower the excited-state energy and possibly produce a conical intersection. The [1,2] or pseudosigmatropic shift, on the other hand, involves a cyclic array of an odd number of interacting orbitals that is isoelectronic with a generalized fulvene, and the reaction is always ground-stateforbidden because the product is a biradical.
[1,2] and [1,3] pathways is consistent with a suprafacial process and retention of configuration at the migrating carbon (Bernardi et al., 1992a).
The irradiation of diisopropylidenecyclobutane (106) causes an allowed antarafacial [ I ,5] shift (Kiefer and Tanna, 1969). For cycloheptatriene (107), a suprafacial [1,7] sigmatropic shift (Roth, 1963) and formation of bicyclo[3.2.0]hepta-2,6-diene (108) by intramolecular cyclization (Dauben and Cargill, 1961) are observed. Both these reactions originate from the 2A' excited state, which is reached from the optically prepared IA" state via an S,S, conical intersection, as has been suggested from lifetime measurements (Borell et al., 1987; Reid et al., 1992, 1993) and confirmed by theoretical calculations. The latter also showed that no such conical intersection exists for dibenzosuberene (109), and the relaxed S, minimum is still of A" symmetry (Steuhl and Klessinger, 1994). This state exhibits excited-state carbon acid behavior, showing that in fact the HOM-LUMO excited-singlet state of ground-state antiaromatic carbanions is stabilized compared to the excited state of ground-state aromatic carbanions (Wan and Shukla, 1993).
A formal [I ,5] hydrogen shift occurs in photoenolization of compounds such as o-methylacetophenone (110). The reaction proceeds in the triplet state by way of hydrogen abstraction and involves biradical intermediates and an equilibrium of the triplet states of Z- and E-en01 (Haag et al., 1977; Das et al ., 1979). These expectations have been confirmed by recent theoretical investigations on methyl shifts in but-I-ene (Scheme 35). The results demonstrate the existence of a funnel on the S, surface at a geometry from which travel on the ground-state surface can produce either a [1,2] or a [1,3] sigmatropic shift. The geometry and orientation of the migrating CH, radical along the
448
ORGANIC PHOTOCHEMISI'UY
cis-Crotonaldehyde (111) has been used as a model for the theoretical investigation of photoenolization reactions (Sevin et al., 1979; Dannenberg and Rayez, 1983).
Compound 112 represents an example of a photochromic material in which the colored enol is stabilized through hydrogen bonding and by the phenyl groups (Henderson and Ullmann, 1965).
7.5.3 Photoisomerization of Benzene The photochemical valence isomerizations of benzene summarized in Scheme 36 are wavelength-dependent singlet reactions. The first excited singlet state gives benzvalene (115) and fulvene (116); prefulvene (117), which had been postulated to be involved in the addition of olefins to benzene (cf. Section 7.4.3), was suggested to be an intermediate for both isomers (BryceSmith, 1968). The second excited singlet state leads to the formation of dewarbenzene (113) in addition to benzvalene (115) (Bryce-Smith et al., 1971) and fulvene (116); prismane (114) is not produced directly from benzene but rather results from a secondary reaction of dewarbenzene. (Cf. Bryce-Smith and Gilbert, 1976, 1980.) The conversion of dewarbenzene into prismane may be treated as a L2, + ,2,] cycloaddition.
The formation of benzvalene is formally an x[2 + 21 cyclo-addition. The S, (B,,) reaction path from benzene toward prefulvene starts at an excitedstate minimum with D,, symmetry and proceeds over a transition state to the geometry of prefulvene, where it enters a funnel in S, due to an S,-So conical intersection and continues on the So surface, mostly back to benzene, but in part on to benzvalene (Palmer et al., 1993; Sobolewski et al., 1993). At prefulvene geometries, So has a flat biradicaloid region of high energy with very shallow minima whose exact location depends on calculational details (Kato, 1988; Palmer, et al., 1993, Sobolewski et al., 1993). Fulvene has been proposed to be formed directly from prefulvene or via secondary isomerization of benzvalene (Bryce-Smith and Gilbert, 1976). Calculations support the former pathway with a carbene intermediate (Dreyer and Klessinger, 1995). The conversion of' benzene into dewarbenzene can be formally considered to be a disrotatory 4n ring-closure reaction; thus it is not surprising that it is forbidden in the ground state and allowed in the excited state. The correlation diagrams in Figure 7.44 indicate that the reaction proceeds from the S, state (at vertical geometries this is 'B,,, but it soon acquires E,, character as motion along the reaction path occurs). (Cf. Example 7.16.) The best currently available calculations (Palmer et al., 1993) suggest that the molecule moves downhill on the S, surface to an S,-S, conical intersection and
Figure 7.44. Benzene-dewarbenzene valence isomerization; a) orbital correlation diagram (C,,symmetry) and b) state correlation diagram, devised with the help of Table 7.5 and experimental state energies.
reaches the S, surface at a geometry where S,, too, has a funnel. This is caused by an S , S oconical intersection and provides further immediate conversion to the So state. Thus, there is no opportunity for vibrational equilibration in the S, state. The biradicaloid geometry at which these stacked funnels are located is calculated to be such that relaxation in Socan produce dewarbenzene, benzvalene, or benzene. In low-temperature argon matrices, dewarbenzene has been shown to be a primary product of benzene photolysis at 253.7 nm (Johnstone and Sodeau, 1991). Hexafluorobenzene yields hexafluorodewarbenzene upon irradiation in the liquid or vapor phase (Haller, 1967). Isomerization reactions of benzene that result in "scrambling" of hydrogen and carbon atoms presumably involve a benzvalene intermediate and thermal or photochemical rearomatization; depending on which of the CC bonds indicated in Scheme 37 by a heavy or a broken line is cleaved, either the rearranged product is formed or the reactant is recovered (Wilzbach et al., 1968).
D
Scheme 37
The major product of the direct irradiation of dewarbenzene is benzene, while some prismane is produced via an intramolecular [2 + 21 cyclization, which involves a funnel that lies between the geometries of dewarbenzene and prismane (Palmer et al., 1992). The electronic factors that control the existence of this funnel are the same as those for the intermolecular [2 + 21 cycloaddition of two ethylenes. (Cf. Sections 6.2.1 and 7.4.1 .) An important difference is due to the strong steric strain that arises in the cagelike a-bond framework as a diagonal distortion is introduced, producing a high-energy ridge along the rhomboidal distortion coordinate. This moves the conical intersection to higher energies, but it apparently remains accessible. Furthy strain may locate the bottom of the pericyclic funnel at a less diagonally distorted structure, where the S o S , touching is still weakly avoided. Triplet sensitization (ET> 66 kcallmol) exclusively yields triplet excited benzene in an adiabatic reaction. According to Scheme 38, this catalyzes the conversion of dewarbenzene into benzene. The reaction thus proceeds as a chain reaction with quantum yields of up to aIi,= 10 (Turro et al., 1977b).
Scheme 38
3[o[a -- 0 '[or +
+
The major reaction to result from direct excitation of benzvalene is a degenerate [1,3] sigmatropic shift as indicated in Scheme 39. The same reaction is observed upon sensitized triplet excitation if the triplet energy of the sen-
sitizer ET is below 65 kcal/mol. If, however, ET exceeds 65 kcal/mol, reversion to benzene takes place. The latter is an adiabatic hot triplet reaction, which proceeds from one of the higher excited triplet states (T,) of benzvalene and produces triplet excited benzene. It results in a chain reaction with QIi, 2: 4, similar to the reaction of dewarbenzene shown in Scheme 38 (Renner et al., 1975).
Scheme 39
Direct irradiation as well as sensitized triplet excitation of prismane results in rearomatization to produce benzene and in isomerization to form dewarbenzene, which predominates even in the triplet reaction. An adiabatic rearrangement to excited benzene is not observed (Turro et al., 1 977b). Benzene valence isomers undergo thermal rearomatization to form benzene rather easily. Half-times of dewarbenzene and benzvalene at room temperature are approximately 2 and 10 days, respectively. Ythough all these reactions are strongly exothermic, chemiluminescence (cf. Section 7.6.4) is observed only for the thermal ring opening of dewarbenzene. This reaction produces triplet excited benzene in only very small amounts. While benzene phosphorescence is undetectable in fluid solution near room temperature, triplet-to-singlet energy transfer with 9,lO-dibromoanthracene as acceptor, followed by fluorescence, leads to a readily detectable indirect chemiluminescence (Lechtken et al., 1973; Turro et al., 1974). --
Example 7.16: The experimental results on valence isomerization of benzene can be rationalized by means of correlation diagrams that were first discussed for hexafluorobenzene by Haller (1967). The orbital correlation diagram for the conversion of benzene to dewarbenzene (Figure 7.44a) has been constructed on the basis of a classification of the benzene n MOs according to C,, symmetry. The full symmetry of benzene is D,. The symmetry elements common to both point groups C,, and D, are E, C,, and a,(x,z) = al'),which is the plane through
Table 7.5 Interrelations between Irreducible Representations of Point Croups C , and D ,
ORGANIC PHOTOCHEMISfKY While the electrocyclic ring opening to o-quinodimethanes is the major reaction pathway in the irradiation of substituted benzocyclobutenes (cf. Example 6.14), the irradiation of unsubstituted benzocyclobutene yields 1,2dihydropentalene (119) and 13-dihydropentalene (120) a s major products. The mechanism shown with "prebenzvalene" (118) as primary photochemical intermediate has been proposed to explain the formation of the isomeric dihydropentalenes (Turro et al., 1988). Supporting calculations that yield the same mechanism for the benzene-to-fulvene transformation have been published (Dreyer and Klessinger, 1995).
Figure 7.45. Benzene-prismane valence isomerization; a) orbital correlation diagram (C,, symmetry) and b) state correlation diagram, devised with the help of Table 7.5 and experimental state energies.
atoms C-l and C-4, and d,(yz) or a:1', perpendicular to the former plane. From symmetry behavior with respect to these elements, interrelations between irreducible representations of the two point groups may be derived as indicated in Table 7.5. Making use of experimental singlet and triplet excitation energies and of the relative ground-state energies (cf. Turro et al., 1977b). one can obtain the state correlation diagram of the benzene-dewarbenzene interconversion shown in Figure 7.44b; according to Table 7.5 the 'B,state of dewarbenzene correlates with the 'B,,state of benzene, located 5.% eV above the ground state. This explains why dewarbenzene is formed only by short-wavelength excitation (Scheme 36). In agreement with the ab initio results discussed earlier, both an S 2 4 ,conical intersection and an S , 4 , funnel are easily recognized, although the calculated intersections occur at geometries of lower symmetry, as is to be expected from the discussion in Section 6.2.1. The rearomatization observed on singlet excitation and the adiabatic formation of triplet excited benzene observed for sensitized triplet excitation are apparent from this diagram. It is also seen that correlations can be drawn such that the So and TI lines cross, permitting chemiluminescence to occur. (Cf. Section 7.6.4.) Corresponding correlation diagrams for the benzene-prismane interconversion in Figure 7.45 explain why the direct irradiation of benzene does not produce prismane, while the reverse reaction is quite efficient. Correlation diagrams for the other benzene valence isomerization reactions can be derived in a similar way. (See also Halevi, 1977.)
7.5.4 Di-n-methane Rearrangement 1,4-Dienes and related compounds undergo a photochemical rearrangement reaction known as the di-n-methane or Zimmerman rearrangement (Zimmerman et al., 1966, 1980; Hixson et al., 1973). The reaction occurs also with B,y-unsaturated carbonyl compounds and is then called the oxa-di-nmethane rearrangement. (See Section 7.5.5.) According to Scheme 40, the reaction formally involves a [I ,2] shift, but the second double bond between carbons C-1 and C-2 is apparently also involved in the reaction process. The most favorable structural feature for di-n-methane rearrangement is an interaction of the C-2 and C-4 centers of the I ,4-diene in a way that permits a stepwise reaction involving a I ,4-biradical and a 1,3-biradical, as formulated in Scheme 40. In contrast to experimental results (Paquette and Bay, 1982), a b initio calculations suggested that the 1,4-biradical may be a true intermediate (Quenemoen et al., 1985).
Scheme 40
1.4-Biradlcal
1.3-Blradicai
~
7.5
However, more recent calculations for the excited-singlet reaction of 1,4pentadiene (Reguero et al., 1993) suggest that the preferred relaxation from S, to So occurs via a funnel that corresponds to a 1,2 shift of a vinyl group (see Section 7.5.2) and produces the 1,3-biradical, avoiding the formation of the 1,4-biradical intermediate altogether. The 1,3-biradical is unstable and undergoes a ground-state barrierless ring-closure process to yield the final vinylcyclopropane product. In particular, three funnels have been found: one structure with both double bonds twisted, which would lead only to cistrans isomerization or back to the starting material (cf. Section 7.1.3); one rhomboidal structure similar to the conical intersection for the ethylene + ethylene [2, + 2,] cycloaddition, which would lead to cycloaddition products or to the I ,4-biradical (which arises from an asynchronous [2 + 21 addition, cf. Section 7.4.2); and the lowest-energy conical intersection (23 kcallmol lower in energy than the first and 31 kcallmol lower than the second), which corresponds to the sigmatropic [1,2] shift and would produce the 1,3-biradical. A ground state reaction path emanating from the geometry of this funnel leads to vinylcyclopropane. From the stereochemistry at C-1 and C-5, it has been concluded that the ring closure occurs in a disrotatory way and anti with respect to the migrating vinyl group, as shown in Scheme 41. Only when the anti-disrotatory ring closure is not possible for steric reasons is a stereochemistry that corresponds to a syn-disrotatory mode of reaction observed.
8, hv
Scheme 41
Both reaction paths that involve breaking of one a and two n bonds and the making of three new bonds are in agreement with the Woodward-Hoffmann rules. According to Figure 7.46, the disrotatory ring closure between C-3 and C-5 requires the orbitals D and F to overlap anti or syn with respect to the bond between C-2 and C-3 which is being cleaved; overlap of orbitals C and B produces the new bond between C-2 and C-4, and orbitals A and E form the new n bond between C-l and C-2. The rearrangement may therefore be formulated as a Qa+ 2,+ 2,1, or a [J, + 3, + ,2,] reaction, for the anti- or syn-disrotatory mode of reaction, respectively. Both cases correspond to ground-state-forbidden, photochemically allowed reactions. The
REARRANGEMENTS
Orbital interactions at the di-n-methane rearrangement a) anti-disrotatory mode (G [J, + 2, + 3 [,2, + 2,]), b) syn-disrotatory mode (e tn2, + 2,+ 2.1, e 3 [J, + 2 3
Figure 7.46.
s,],+
same conclusion may be reached by the observation from Figure 7.46 that the relevant orbitals form a (4N + 2)-Mobius system, since signs of the six orbitals can be chosen in such a way that just one overlap is negative. Inversion of configuration at C-3, which is expected for an anti-disrotatory reaction path, has been observed, for instance, for the acyclic chiral I ,4-diene 121.
The stereochemistry of the di-n-methane rearrangement has thus been shown to be in complete agreement with a concerted reaction course (Zimmerman et al., 1974). The Qa + s , ] or L2, + J,] mode has the same stereochemical consequences. However, this mechanism, which does not involve the n bond between C-1 and C-2, was excluded because no di-n-methane rearrangement is observed upon irradiation of compound 122, which contains only one double bond.
For acyclic and monocyclic 1,4-dienes, the di-n-methane rearrangement occurs in general from the singlet excited state, since loose geometries are favored in the triplet state and cis-trans isomerization is the preferred reaction, as shown in Scheme 42.
ORGANIC PHOTOCHEMISTRY
Scheme 42
For singlet reactions, the direction of ring closure is the one that involves the most stable 1,3-biradical, as shown in Scheme 43. This is equivalent to the assumption that in a concerted mode of reaction the bond between centers C-3 and C-5 is only weakly formed during the initial stages of the reaction.
7.5
REARRANGEMENTS
457
(cf. Scheme 40) occurs during the evolution of the conical intersecction structure toward the 1,3-biradical structure, as was shown by reaction path calculations: the a bond between the two terminal methylene radical centers is formed via a barrierless process on the side of the --CH, fragment opposite to the initial position of attachment of the migrating vinyl radical. The triplet reaction probably proceeds by a different mechanism, likely to involve a I ,Cbiradical intermediate. In an elegant study of the rearrangement of deuterated m-cyanodibenzobarrelene (124) to the corresponding semibullvalene 125, it was in fact shown that the 1 ,Cbiradical is the productdetermining intermediate (Zimmerman et al., 1993).
A large number of rearrangements such as 126 -+ 127 can be classified as belonging to the di-n-methane type even though the molecules do not formally contain a 1 ,4-diene unit, since in these cases an aryl ring can take the place of one of the olefinic bonds. The di-n-methane rearrangement is not confined to acyclic and monocyclic systems. Bicyclic 1,Cdienes such as barrelene (123) also rearrange, but they do so upon triplet sensitization (Zimmerman and Grunewald, 1966). In all probability, the built-in geometry constraint of a bicyclic molecule does not allow for the loose geometries otherwise characteristic of the triplet state, while in the singlet state other reactions such as [2 + 21 cycloaddition predominate.
Substituent effects have been studied in detail and have been rationalized on the basis of the proposed mechanism (Zimmerman, 1980). Example 7.17:
Donor-substituted benzonorbornadienes 128 generally yield the product of a di-n-methanerearrangement corresponding to m-bridging, while acceptor-substituted compounds 129 give the p-bridged product: All these experimental features are consistent with a concerted pathway that passes through a funnel for a [1,2] vinyl shift and leads to the 1,3biradical region (Reguero et al., 1993). First, the regiospecifity is readily rationalized via the different accessibility of the two conical intersections leading to the 1,3-biradicals (Scheme 43): the one with the most stable substituted radical moiety will be lower in energy, and the vinyl radical without radical stabilizing substituents will migrate. Second, the double bond in the migrating vinyl radical retains its original configuration along the excitedstate-ground-state relaxation path. Third, inversion of configuration on C-3
01-:tiANIC PHOTOCHEMISTRY
Figure 7.47.
Orbital ordering in donor- and acceptor-substituted benzenes.
If the triplet states of 128 and 129 involve mainly single configurations with a half-occupied hom*o and a half-occupied LUMO. the pronounced regioselectivity may be rationalized on the basis of an interaction of the half-occupied LUMO of the aromatic moiety with the vacant LUMO of the ground-state ethylene moiety. Substitution removes the degeneracy of the benzene LUMO, and the product formed (meta or para) depends on the magnitude of the LCAO coefficients of the lower of these MOs. From Example 3.10 the @., MO is known to be the LUMO in acceptor-substituted benzenes, while @,,. is the LUMO in donor-substituted benzenes. (Cf. Figure 3.18.) This is shown again in Figure 7.47, together with the labeling of the MOs corresponding to C,, symmetry; in the b, MO the LCAO coefficient of the para position is largest in absolute value, while in the a, MO the meta coefficient is larger than the para coefficient (Santiago and Houk, 1976).
Example 7.18: The complex mechanism of the rearrangement of metacyclophanediene 130 into dihydrocyclopropapyrene 136 was elucidated by Wirz et al. (1984) and is 131 are symshown in Figure 7.48. Valence isomerization of the type 130 metry-allowed ground-state reactions and are extraordinarily fast. The adiabatic formation of 131* was mentioned in Example 6.7.
-
7.5
REARRANGEMENTS
459
Figure 7.48. Mechanism of the transformation of metacyclophane 130 into dihydrocyclopropapyrene 136 with intermediates 1321133 and 1341135. The suggested position of the original methano bridge is marked by a bullet. The hatched arrows indicate the two-quantum process that occurs at room temperature and does not involve the ground-state intermediate 134 (adapted from Wirz et al., 1984).
Below - 30°C the reaction path consists of a three-quantum process involving two thermally stable, light-sensitive isomers 132 and 134. The first two steps from 130 to 133 may formally be viewed as di-n-methane rearrangements (cf. Scheme 40, the carbon marked with a bullet in Figure 7.48 corresponds to C3 in Scheme 40). while the last step from 134 to 136 represents a [1,7] hydrogen shift. At room temperature the reaction proceeds as a two-quantum process, bypassing the ground-state intermediate 134.
A related reaction, the "bicycle rearrangement," is schematically shown in Scheme 44: A carbon with substituents R'and RZconnected to a n system by two hybrid AOs moves along the molecule as if the hybrid AOs were the wheels and the substituents the handlebars of a bicycle.
An example is the rearrangement of 2-methyleneJ,6-diphenylbicyclo[3.l .O]hexene (137). which yields 1,5-diphenylspiro[2.4]-4,6-heptadiene
I
460
ORGANIC PHOTOCHEMISTRY
7.5
REARRANCEMENI'S
(138). The shift does not proceed over two bonds as depicted in Scheme 44, but over three bonds (Zimmerman et al., 1971):
The reaction proceeds in the singlet state and is stereospecific in that the configuration of the "handlebars" is retained for migrations of up to three bonds. Both the mechanism of the di-a-methane rearrangement and that of the 2,5-cyclohexadienone rearrangement (dealt with in Section 7.5.9, involve a step that may be formulated as a bicycle rearrangement (Zimmerman, 1982).
7.5.5 Rearrangements of Unsaturated Carbonyl Compounds In addition to the normal photochemical reactions of saturated ketones, P,yunsaturated carbonyl compounds undergo carbonyl migration via a [1,3] shift. Compound 139 in Scheme 45 represents a typical example. Compounds with an alkyl substituent in the /? position such as 140 may also undergo a Norrish type I1 reaction (Kiefer and Carlson, 1%7), while for ketones with electron-rich double bonds such as 141, oxetane formation is also observed (Schexnayder and Engel, 1975).
a Cleavage is much faster for F,y-unsaturated ketones than for saturated ketones. This can be rationalized by the relative stability of the acyl-ally1 radical pair, which can experience an additional stabilization by simultaneous bonding interaction of the acyl radical with the y carbon (Houk, 1976). Whether this interaction is large enough to render the [I ,3] shift a concerted
Figure 7.49. Possible mechanisms of the [I ,3] shift of P,y-unsaturated ketones: a) concerted, b) concerted with inclusion of carbonyl orbitals, c) biradical intermediate, and d) radical pair intermediate (adapted from Houk, 1976).
reaction, or whether a mechanism involving a cleavage and subsequent radical recombination prevails, cannot be generally decided. Four different mechanisms can be envisioned as indicated in Figure 7.49: the [I ,3] shift may be concerted and can be classified either as a [,2, + ,2,] reaction (Figure 7.49a) or with inclusion of the acoand no orbitals in the orbital array (Figure 7.49b). Two stepwise mechanisms, one involving the formation of a biradical, the other of a radical pair, can also be formulated (Figure 7.49c,d). Since the [I ,3] shift is a ground-state forbidden by orbital symmetry, the ground-state surface will have a saddle point at a geometry approximating the transition state of this reaction, while the S, surface will have a minimum or a funnel at a nearby geometry. Thus, whether a biradicaloid intermediate or a radical pair will be involved, or whether the reaction will be concerted, depends on the magnitude of a,y interaction or on the degree to which the crossing is avoided. For the '(n,n*) state of a B,y-unsaturated ketone to undergo radical a cleavage or a [I ,31 shift, the a bond must be approximately parallel to the n orbital, since only then can ally1 resonance stabilize the biradical and the carbonyl orbital overlap with the a system at the y terminus. These geometrical requirements are not fulfilled for the aldehyde 142, and intersystem crossing dominates a cleavage and [I ,3] shift (Baggiolini et al., 1970). The triplet photochemistry of P,y-unsaturated carbonyl compounds depends very much on tl-cir electronic configuration. The '(n,n*) state under-
OIIl A N I C PHOTOCHEMISTRY
7.5
REARRANGEMENTS
463
[I ,3] shift (Barber et al., 1969) has been observed for cyclopentenones (143). Radical stabilizing substituents in the 5-position, as in 144, favor formation of cyclopropyl ketenes (Agosta et al., 1969). goes a [1,3] shift, presumably via a radical pair. In most p,y-unsaturated carbonyl compounds, however, the lowest triplet state is the 3(n,n*) state, for which a [1,2] shift termed the oxa-di-n-methane rearrangement is characteristic (Schaffner and Demuth, 1986). Concerted mechanisms as well as biradical intermediates may be formulated for this reaction, in a way analogous to the di-n-methane rearrangement. Although the problem has not yet been solved generally, evidence in favor of a stepwise reaction path has been obtained-for instance, for the reaction shown in Scheme 46 (Dauben et al., 1976). This contrasts with the corresponding results for the di-n-methane rearrangement of the hydrocarbon 121, which occurs in the singlet state.
Scheme 49
p(
144
' ~ h
For cyclohexenones, only [1,2] shifts are typical. According to Scheme 50, two types of [1,2] shift with ring contraction can occur: one through rearrangement of the ring atoms (type A), the other through migration of a ring substituent (type B) (Zimmerman, 1964): Scheme 48
Because the intersystem crossing probability is low, the triplet reaction can in general be initiated only by sensitization. The difference between singlet and triplet reactions can be utilized synthetically, as indicated in Scheme 47 (Baggiolini et al., 1970; Sadler et al., 1984):
Scheme 47
@gib-sh
R
~ y pe
R' 'R'
~yp A
Scheme 50
Photochemical ring opening of linearly conjugated cyclohexadienones affords dienylketenes (145), which react in one of the following ways: recyclization to the original or to a stereoisomeric cyclohexadienone, formation of bicyclo[3.l.O]hexenones (146), or addition of a protic nucleophile to yield substituted hexadienecarboxylic acids (147) (Quinkert et al., 1979).
In addition to the reactions discussed in Section 7.4.5, a,&unsaturated ketones undergo isomerization to give the deconjugated p,y-unsaturated compound. According to Scheme 48, this reaction may be analyzed as a Norrish type I1 reaction with cleavage of a n bond instead of a a bond. However, it may also be interpreted as a photoenolization. (Cf. Section 7.5.2.)
Scheme 48
JyeDeyJeJy
Cyclic enones and dienones undergo a number of photorearrangements. As indicated in Scheme 49, both ring contraction via a [1,2] shift (Zimmerman and Little, 1974) and formation of cyclopropanone derivatives via a
In addition to these reactions typical of the '(n,n*) state, 6-acetyloxycyclohexadienones (0-quinol acetates) form phenols from the 3(n,n*)state. These results are summarized in Figure 7.50 (Quinkert et al., 1986). Dienylketenes are produced irrespective of whether irradiation occurs into the n-n* or n-n* band, and phenols can be formed not only through triplet
464
ORGANIC PHOTOCHEMISTRY
OAc
+
.XYCH3
Figure 7.50. Schematic state correlation diagram for o-quinol acetate photoreactions. Wavy arrows designate physical as well as chemical radiationless processes (by permission from Quinkert et al., 1986).
sensitization (ET > 42 kcal/mol, corresponding to the Yrc,n*)state), but also by extended direct irradiation in aprotic solvents if the quinol acetate is regenerated from the kinetically unstable dienylketene.
7.6 Miscellaneous Photoreactions 7.6.1 Electron-Transfer Reactions
7.6
MISCELLANEOUS PHOTOREACTIONS
465
From Scheme 51 it is seen that an exciplex with a certain degree of charge-transfer character can be formed either from an excited donor molecule ID* and an acceptor molecule 'A or from an excited acceptor molecule 'A* and a donor molecule ID. Both alternatives have been observed experimentally. Thus, the use of diethylaniline as donor and either biphenyl or anthracene as acceptor yields exciplexes that can be identified by the typical structureless fluorescence at long wavelengths. This corresponds to the two possibilities depicted in Figure 5.25a and b, excitation of the donor in the first case and of the acceptor in the second case. The charge-transfer character of the complex causes the wavelength of its fluorescence to be highly sdvent dependent, and the value of p = 10 D has been derived for the dipole moment (Beens et at., 1967). Electron transfer then gives a radical ion pair ( D e A e ) , which for k, > k, decays to the reactants by back electron transfer; this corresponds overall to an electron-transfer quenching. (Cf. Section 5.4.4.) However, if fast secondary reactions lead to product formation, k, %k,, the process becomes a photoinduced electron-transfer reaction, and this can be utilized synthetically. Example 7.19: As an example, photochemical excitation of donor-acceptor complexes may be considered. Irradiation into the CT band of the anthracene-tetracyanoethylene complex leads directly to the radical ion pair, the components of which are identifiable from their UV-visible spectra. The transient absorptions decay in -60 ps after excitation, as the radical ion pairs undergo rapid back electron transfer to afford the original donor-acceptor complex (Hilinski et al., 1984). With tetranitromethane as acceptor, however, an addition product is obtained in both high quantum and chemical yield. This is due to the fact that the tetranitromethane radical anion undergoes spontaneous fragmentation to a NO2 radical and a trinitromethyl anion, which is not able to reduce the anthracene radical cation (Masnovi et al., 1985):
It has become evident only fairly recently that photochemical electron-transfer processes (PET) play an important role in many reactions. In this connection it is of great importance that in an excited state a molecule can be a better oxidant as well as a better reductant than in the ground state. For instance, after H O M b L U M O excitation the half-occupied hom*o can readily accept another electron, while the single electron from the LUMO can readily be transferred to an acceptor. (Cf. Figure 5.25.)
A detailed study showed that after dissociation of the radical anion a contact ion pair [Dm C(NO,)f)] in a solvent cage is initially formed. It transforms into
OK(;ANIC PHOTOCHEMISTRY a solvent-separated ion pair within a few ps, which in turn converts into the free ions within a few ns. Both processes follow first-order kinetics. The free ions then form the adduct in a second-order process (Masnovi et al., 1985).
The photoreduction of carbonyl compounds or aromatic hydrocarbons by amines was one of the early electron-transfer reactions to be studied. Observation of products from primary electron transfer depends on the facility of a deprotonation of the amine, which must be fast compared to back electron transfer. For amines without a hydrogens, quenching by back electron transfer is observed exclusively (Cohen et al., 1973). The solvent plays a quite important role since it determines the yield of radical ion pairs formed from the exciplex (Hirata and Mataga, 1984). As a typical example the photoreduction of naphthalene by triethylamine (Barltrop, 1973) is shown in Scheme 52. The radicals generated by a deprotonation couple to the products 148 and 149, and disproportionate to the reduction product 150.
I!
i
7.6
more readily from the conformationally favored transition state 152 rather than from 153 (Lewis et al., 1981):
While electron-transfer reactions of aromatic hydrocarbons are in general reactions of the (n,n*) state, electron transfer in carbonyl compounds can be into the (n,n*) as well as into the (n,n*) state. Correlation diagrams may be constructed for these reactions and are fully equivalent to the corresponding diagrams for hydrogen abstraction. (Cf. Section 7.3.1 .) From such diagrams it is, for instance, evident that coplanar electron transfer (attack of the amine within the plane of the carbonyl group) is symmetry allowed in the (n,n*) state but symmetry forbidden in the (n,rt*) state. Photoreduction of benzophenones and acetophenones by amines has been studied in detail. A mechanism has been derived for the reaction of benzophenone with N,N-dimethylaniline that involves a triplet exciplex in the formation of the radical ion pair, as indicated in Scheme 49.
t PhN(CH& 140
Scheme 52
With singlet excited trans-stilbene (151) and tertiary alkyl amines only products characteristic for radical coupling are observed (Lewis et al., 1982).
With unsymmetrical trialkylamines selective formation of the least-substituted a-amino radical is observed. The, stereoselectivity is thought to be stereoelectronic in origin, as can be most easily seen in highly substituted amines such as diisopropylmethylamine, where a deprotonation occurs
467
MISCELLANEOUS PHOTOREACTIONS
+ Ph2C=0
.O P~N(CHJ),
+ P~,CO 0
Scheme 53
Subsequent proton transfer yields the ketyl radical and the usual products. (Cf. Section 7.3.1.) This mechanism was confirmed by CIDNP measurements. At the same time it was proven that on excitation of the amine, electron transfer occurs from the singlet excited amine to the ground-state ketone (Hendricks et al., 1979). The exact time scale for those processes that follow the electron transfer was studied by picosecond spectroscopy (Peters et al., 1982). Photochemical addition reactions may also occur as electron-transfer reactions involving a radical ion pair. An illustrative example is the photochemical reaction of 9-cyanophenanthrene(154) with 2,3-dimethyl-2-butene, which, in nonpolar solvents, gives good yields of a [2 + 21 cycloadduct via a singlet exciplex, while in polar solvents radical ions are formed in the primary photochemical process. The olefin radical cation then undergoes deprotonation to yield an allyl radical or suffers nucleophilic attack by the solvent to produce a methoxy alkyl radical. Coupling of these radicals with
468
ORGANIC PHOTOCHEMISTRY
7.6
MISCELLANEOUS PHOTOREACTIONS
the aromatic radical anion produces acyclic adducts such as 156 in addition to the cycloaddition and reduction products (Lewis and Devoe, 1982).
Formation of cycloadducts can be completely quenched by conducting the experiment in a nucleophilic solvent. This intercepts radical cations so rapidly that they cannot react with the olefins to yield adducts. In Scheme 54 the regiochemistry of solvent addition to 1-phenylcyclohexene is seen to depend on the oxidizability or reducibility of the electron-transfer sensitizer. With I-cyanonaphthalene the radical cation of the olefin is generated, and nucleophilic capture then occurs at position 2 to afford the more stable radical. Electron transfer from excited 1,4-dimethoxynaphthalene, however, generates a radical anion. Its protonation in position 2 gives a radical that is oxidized by back electron transfer to the sensitizer radical before being attacked by the nucleophilic solvent in position 1. Thus, by judicious choice of the electron-transfer sensitizer, it is possible to direct the photochemical addition in either a Markovnikov (157) or anti-Markovnikov (158) fashion (Maroulis and Arnold, 1979).
Electron transfer can also induce valence isomerizations such as the transformation of hexamethyldewarbenzene (159) to hexamethylbenzene. The quantum yield of this reaction is larger than unity; a chain reaction mechanism is therefore assumed (Jones and Chiang, 1981).
Spectroscopic studies with photo-CIDNP techniques revealed the existence of two distinct radical cations generated from hexamethyldewarbenzene, presumably rapidly interconverting. In one of these, the central carbon-carbon bond is significantly stretched and bears the unpaired spin density. In the second. the spin density is confined to one of the olefinic bonds. This example is the first to show conclusively that two different radical ion structures can correspond to a single minimum on the ground-state surface of the neutral (Roth et al., 1984). Quadricyclanes (160) also undergo a valence isomerization to norbornadienes if irradiated in the presence of electron acceptors such as fumaronitrile (Jones and Becker, 1982). Two distinct radical cation structures are observed for the hydrocarbon, corresponding roughly to the bonding patterns of norbornadiene and quadricylane, respectively (Roth et al., 1981).
Quadricyclane in Csl or KBr matrices, prepared by deposition in the salt under conditions that yield single-molecule isolation, is rapidly converted into norbornadiene under conditions that induce color center formation in the alkali halide: rapid-growth vapor deposition, or U V or X-ray irradiation. The reaction proceeds only at temperatures at which color centers of the "missing electron" type (H center) are mobile. At lower temperatures (T < 90 K), UV irradiation of norbornadiene converts it into quadricyclane in the usual fashion (Kirkor et al., 1990). The electron-transfer sensitized interconversion of trans- (161) and cis1.2-diphenylcyclopropane (162) (Wong and Arnold, 1979) is thought to proceed via the radical ion pair from which back electron transfer generates a triplet biradical that undergoes geometric isomerization; the corresponding ring-opened radical cations 163are conformationally stable (Roth and Schilling, 1981). Cycloadditions of carbonyl compounds to olefins generally involve exciplexes and biradicals. (Cf. Section 7.4.4.) While normal olefins frequently yield a number of products, photoinduced electron transfer may be utilized in the case of electron-rich olefins to influence the regioselectivity. Thus, irradiation of the ketene acetal 165 and biacetyl (164) yields exclusively the oxetane 167. Since the radical cation 166 could be trapped, electron transfer
OKtiANIC PHOTOCHEMISTRY
470
7.6
MISCELLANEOUS PHOTOREACTIONS
is assumed to be the photochemical primary step. As the regioisomeric oxetane 168 is generated by a thermal process involving a dipolar intermediate, this reaction constitutes a case of reversal ("Umpolung") of the carbonyl reactivity by photoinduced electron transfer (Mattay et al., 1984a).
6~ Ph Scheme 56
In the photoinduced singlet dimerization of indene shown in Scheme 55, the endo head-to-head dimer expected from the more favorable excimer geometry is the preferred product. With electron-transfer sensitization, the less sterically hindered exo head-to-head dimer is formed (Farid and Shealer, 1973).
Scheme 55
Ph
Ph
170
The [4 + 21 cycloaddition of electron-rich dienes to electron-rich dienophiles in nonpolar solvents can be catalyzed by electron-poor arene sensitizers (Calhoun and Schuster, 1984). The proposed mechanism involves a triplex (ternary complex) formed by the reaction of the diene with an exciplex composed of the sensitizer and the dienophile. Using a chiral sensitizer, ( - )-I, 1 '-bis(2,4-dicyanonaphthyl), an enantioselective cycloaddition was observed as shown in Scheme 57. In the case of 1,3-cyclohexadiene and trans-/?-methylstyrene, the enantiomeric excess was 15 2 3% (Kim and Schuster, 1990). (Cf. the enantioselective cis-trans isomerization via a triplex discussed in Section 7.1.2.)
6~
Besides cyclobutane formation, alternative ring closures are sometimes observed. One example is the 9,lO-dicyanoanthracene sensitized dimerization of 1,l-diphenylethylene (169). The six-membered ring is formed via the 1 ,4-radical cation, which results from the addition of the free radical cation to diphenylethylene as indicated in Scheme 56, while the l,4-biradical generated by back electron transfer from the radical ion pair yields tetraphenylcyclobutane (170) (Mattes and Farid, 1983).
Products resulting from photolysis of alkyl iodides indicate that according to Scheme 58 hom*olysis of the C-X bond is followed by electron transfer, which results in an ion pair in a solvent cage (Kropp, 1984). In the case of norbornyl iodide (171) and other bridgehead iodides, bridgehead cations re-
ORGANIC PHOTOCHEMISTRY
472
sult; these are difficult to obtain from solvolytic reactions. In other cases Wagner-Meerwein rearrangements of the ionic intermediates have been observed.
\
OCH,
Scheme 58
A particularly important photoinduced electron-transfer process occurs in photosynthesis in green plants. The overall process amounts to the splitting of water by sunlight into oxygen and metabolically bound hydrogen, and this forms the basis for the existence of higher organized living systems on earth (Kirmaier and Holten, 1987; Feher et al., 1989; Boxer, 1990). In contrast to simpler photosynthetic bacteria that have only one photosystem and are not able to oxidize water, two quanta of light are used by plants to split water by means of two photosystems (PS I and PS 11). This proceeds via a sequence of redox processes, indicated schematically in Figure 7.5 1. Photosynthesis may be formally described as charge separation induced by electron transfer with the positive charge (electron defect) being used for
I
I
PS 1
("P700") @--
-----r
7.6
MISCELLANEOUS PHOTOREACTIONS
473
oxidation of H,O (or H,S) and the negative charge for reduction of CO,, according to the overall equation
However, the saccharides (CH20), are not produced by the photoreaction but by a subsequent dark reaction of the photochemically generated hydrogenated nicotinamide adenine dinucleotide phosphate (172) (NADPB-H,). The energy absorbed by pigment-protein complexes in the light-gathering antennae, which also contain carotenoids as triplet quenchers, is transferred to the photochemically active reaction center and produces the excited singlet state of pigment P680 absorbing at 680 nm, which constitutes a special dimer complex of chlorophyll a (173) referred to as the special pair. The oxidation potential of excited P680 is sufficient to remove one electron from water; an electron is transferred from the excited special pair, via a primary electron acceptor, possibly a pheophytin a molecule (an Mg-free chlorophyll a), to a plastoquinone (174). From Figure 7.52 it is seen that after these two electron-transfer processes the ground state of P680 is regenerated; P680 thus acts as a photocatalyst for the transfer of an electron from water to plastoquinone.
"P 870"
H20 oxidation
PS2 ("P680") Figure 7.51. Schematic representation of charge separation in the photosynthetic cycle a) in green plants involving photosystems PS I1 and PS I and b) in photosynthetic active bacteria (by permission from Rettig, 1986).
(In bacteriochlorophyll C3 carries a COCHI group and the C7-C8 double bond is hydrogenated)
Plastoquinone in turn is a reductant for excited WOO of photosystem PS I, which operates similarly to the system PS I1 and has a reduction potential sufficient for an electron transfer to the iron-sulfur complex of ferredoxin and finally to N A D P . producing [emailprotected].
ORGANIC PHOTOCHEMISTRY
- -
I
+
Chl a*
PQ
I +'- - - -+3' H20
+
+ S t + H~O?
Chl a
1
7.6
MISCELLANEOUS PHOTOREACTIONS
475
in meta positions, and strong electron-releasingsubstituents such as 4 C H 3 behave as activating substituents and direct incoming nucleophiles to the ortho and para positions. The photoreactions of the three nitroanisoles with @CN are collected in Scheme 59 to illustrate this behavior (Letsinger and McCain, 1969).
PQ?
Redox process of photosystem PS I1 during electron transfer from water to plastoquinone.
Figure 7.52.
The protein matrix essentially determines the properties and the n-electron redox chemistry of the reaction center and has to be considered an apoenzyme of this functional unit. The geometrical arrangement of the chromophores in the reaction center of the purple bacterium Rhodopseudomonas viridis has been elucidated using X-ray crystallographic analysis (Deisenhofer et al., 1985). On the basis of these structures, several mechanisms have been proposed to explain the primary electron-transfer event (Bixon et al., 1988; Bixon and Jortner, 1989; Marcus, 1988). The special pair in this system consists of a bacteriochlorophyll dimer that transfers an electron from the excited singlet state via a bacteriopheophytin to an ubiquinone (175). The high symmetry of this arrangement and the small overlap of the chromophores in the reaction center are apparently essential for the high efficiency of the system, as has been concluded from a comparison with TICT states, or "sudden polarization" states (cf. Section 4.3.3) (Rettig, 1986). In solution, electron transfer from a metal porphyrin to benzoquinone is, in fact, efficient only in the triplet state, and in the singlet state back electron transfer predominates (Huppert et al., 1976). INDOIS calculations demonstrate that the observed charge separation can be reproduced only by including the effects of the protein; utilizing the selfconsistent reaction field solvent model dramatically lowers the energy of all the charge-transfer states (Thompson and Zerner, 1991).
7.6.2 Photosubstitutions Many aromatic compounds undergo heterolytic photosubstitution on irradiation. Nucleophilic aromatic substitutions are particularly frequent. The orientation rules are reversed compared to those known for ground-state reactivity; that is, electron-withdrawing substituents orient incoming groups
Aromatic photosubstitution can proceed by various mechanisms. Electron-withdrawing substituents presumably include a S,2 3Ar*mechanism involving a a complex of the triplet excited aromatic moiety with the nucleophile, as indicated in Scheme 60.
Scheme 60
This mechanism has been proven conclusively for the CH30@exchange reaction of nitroanisoles and similar compounds. Attack by the nucleophile occurs at the position with the highest calculated positive charge in the triplet state. The directing effect of the nitro group can also be described by hom*o and LUMO charge densities but the activating influence is difficult to rationalize on the basis of charge densities. This has been explained by the fact that the excited triplet and the ground-state potential energy surfaces are particularly close to each other at the geometry of the a complex for meta substitution, which is very unfavorable in the ground state (Van Riel et al., 1981). The activating effect of methoxy groups and similar substituents has been explained by an SRoNljAr* mechanism with the formation of a radical cation through electron transfer as the principal step as shown in Scheme 61 (Havinga and Cornelisse, 1976):
ORGANIC PHOTOCHEMISTRY
7.6
MISCELLANEOUS IJHOI'OKEACI'IONS
Route I:
Scheme 61
CN
Route II:
For halobenzenes an SRoNlAr* mechanism with the formation of a radical anion by electron transfer as shown in Scheme 62 has been discussed (Bunnett, 1978):
Aliphatic and alicyclic molecules such as cyclohexane undergo photosubstitution with nitrosyl chloride (Pape, 1967). The reaction is of considerable industrial importance in the synthesis of E-caprolactam, an intermediate in the manufacture of polyamides (nylon 6). (Cf. Fischer, 1978.) At long wavelengths a cage four-center transition state between alkane and an excited nitrosyl chloride molecule is involved, as indicated in Scheme 63. In contrast to light-induced halogenation, photonitrosation has a quantum yield smaller than unity, and is not a chain reaction.
Scheme 83
7.6.3 Photooxidations with Singlet Oxygen Under photochemical conditions oxygen can insert itself into a substrate with the formation of hydroxyhydroperoxides (176), hydroperoxides (177), peroxides (178), and dioxetanes (179). There are essentially two different reaction courses, involving either a photochemically generated radical reacting with ground-state oxygen or a ground-state substrate molecule reacting with singlet oxygen. (Cf. Example 7.20.) Some typical examples are summarized in Scheme 64.
Formation of endoperoxides of type 178 can be interpreted as a + ,?,I cycloaddition that involves singlet oxygen and therefore constitutes a photochemical reaction. Similarly, formation of hydroperoxides from nonconjugated olefins with an allylic hydrogen generally appears to be a concerted ene reaction as indicated in Scheme 65. Other mechanisms have
k4,
1
(
478
) I . II\IVIC I7 I 0 10CHCILIIb I KL
been proposed, and according to Scheme 65 involve a biradical or a perepoxide 180. Ab initio calculations favor the biradical mechanism, at least in the gas phase (Harding and Goddard 111, 1980), while semiempirical calculations suggest the perepoxide to be a genuine intermediate (Dewar and Thiel, 1977). Reactions of singlet oxygen are characterized by low activation energies and very fast reaction times. Therefore, a detailed mechanism is in general difficult to establish. Many experimental findings suggest, however, that an interaction with charge-transfer character occurs at the initial stages of reaction, with the stereochemistry given by the hom*o of the olefin as the electron donor and the n* LUMO of the oxygen (Stephenson et al., 1980). This is exemplified in Scheme 66 for the hom*o-LUMO interaction of 2-butene with oxygen.
7.6
MISCELLANEOUS PHOTOREACTIONS
St ate
Complex Orbitals
Real Orbitals
Figure 7.53. Spin-orbital diagrams for the lowest molecular oxygen states (by permission from Kasha and Brabham, 1979).
Scheme 6c
Some substituted alkenes react with singlet oxygen to form a dioxetane in a [ J , + ,2,] cycloaddition reaction. Most dioxetanes readily decompose to carbonyl compounds in an exothermic reaction that is accompanied by a bluish luminescence. The chemiluminescence will be dealt with in more detail in Section 7.6.4.
Example 7.20: The ground state of molecular oxygen is a 'I:; state according to Hund's rule, and the MOs n, and n,,, which are degenerate by symmetry, are singly occupied by electrons with parallel spin. The corresponding singlet state 'X: is higher in energy by 35 kcallmol. Between these two I: states there is a degenerate 'A8 state, 22 kcallmol above the 'I:; ground state. This 'Agstate is generally referred to as singlet oxygen. The description of these states becomes particularly intelligible if it is recognized that the 0, molecule is a perfect axial biradical. especially if the symmetry-adapted complex n MOs n , = (JZ,+ in,)/ and n = (n,- in,)lfi are used. Figure 7.53 gives a schematic representation of the wave functions of the lowest states of molecular oxygen expressed in terms of complex and real orbitals, with MOs doubly occupied in all states not shown. In an axial biradical, all possible real combinations of the degenerate orbitals are localized or delocalized to the same degree. (Cf. Section 4.3.2.) Potential energy curves of molecular oxygen are shown in Figure 7.54.
Figure 7.54. Potential energy curves for the lowest molecular oxygen states (adapted from Herzberg, 1950).
ORGANIC PHOTOCHEMlSTRY
480
7.6
MISCELLANEOUS I'HUI'OKEAC'I'IONS
Singlet oxygen can be generated either by thermal or by photochemical methods. The most general and synthetically useful method is photosensitization with a strongly absorbing dye such as Rose Bengal or methylene blue, which can be used advantageously as a polymer-bound sensitizer (Schaap et al., 1975). Singlet oxygen is generated by triplet-triplet annihilation according to Sens + hv + 'Sens* -,3Sens* 'Sens* + '0, + Sens + '0, The thermal generation of singlet oxygen from hydrogen peroxide and hypochlorite presumably involves the chloroperoxy anion; other synthetically useful examples involve decomposition of phosphite ozonides or endoperoxides, as indicated in Scheme 67 (cf. Murray, 1979).
--
--
Figure 7.55.
Orbital correlation diagram for endoperoxide formation.
Figure 7.56.
Schematic state correlation diagram a) for a singlet photoreaction and
-
Example 7.21: The orbital correlation diagram for the [4 is shown in Figure 7.55 for the reaction
+ 21 cycloaddition of singlet oxygen
The ordering of the reactant orbitals is obtained from ionization potentials of molecular oxygen and butadiene. The reactant configuration with singly occupied MOs $ and ('0,)correlates with a highly excited product configuration, while the configuration with doubly occupied or MOs correlates with the product peroxide in'the ground state. Due to perturbation by the reaction partner, the configuration develops as the reaction proceeds.
< ~
7.6.4 Chemiluminescence From the schematic representation in Figure 7.56 it is seen that chemiluminescence can be described as a reverse photochemical reaction. Chemiluminescence is afforded by a transition from the ground-state potential energy surface to an isoenergetic vibrational level of an excited-state surface and
b) for a singlet chemiluminescent process.
7.6
by escape over a small barrier into a deeper well. If the separation of the ground-state and excited-state surfaces has increased in the process as shown in Figure 7.46b, the resulting excited molecule can reveal its presence by emission. Oxidation of luminol (see Example 7.22) and thermolysis of endoperoxides or I ,Zdioxetanes provide important examples of chemiluminescent reactions. Tetramethyl-1,Zdioxetane(181) has been studied in great detail; the thermolysis is clearly first order and the activation enthalpy in butyl phthalate is AHS= 27 kcallmol. The enthalpy difference between the reactants and ground-state products is AH,, = -63 kcallmol.
The sum -AH,, + AHt is greater than the excitation energies of the TI (n,n*) state (A& = 78 kcallmol) as well as the S, (n,n*) state (AEs = 84 kcallmol) of acetone. Both states can be formed exothermally from the transition state for thermolysis (Lechtken and Hohne, 1973). It could be shown independently that triplet excited acetone ()A*) is produced directly with a quantum yield Q, = 0.5. Singlet excited acetone ('A*) is formed with a quantum yield = 0.005.
MISCELLANEOUS PHOTOREACTIONS
483
These results have been rationalized by means of the correlation diagram shown in Figure 7.57 (Turro and Devaquet, 1975). The ground state of tetramethyl-1,tdioxetane correlates with a doubly n+n* excited state of the supermolecule consisting of two acetones, while the ground state of the latter correlates with a doubly v d r excited state of the dioxetane. The (n,n*) states of acetone will correlate directly with excited dioxetane states of appropriate symmetry, probably with the (n,dr) states, which are antisymmetric with respect to the reaction plane. The crossing between So and T, will be weakly avoided due to spin-orbit coupling, and a "surface jump" to the TI product surface ('A* + A,) can occur. The origin of spin-orbit coupling that makes the crossing avoided may be visualized by means of the schematic diagram given in Figure 7.57, which represent the electronic structure of the reaction complex for different stages of the reaction: To the left of the transition state, the 0-0 bond has lengthened with the "unpaired" electrons in orbitals of a symmetry; at the product side, however, one of the unpaired electrons is in a n* MO of one of the carbonyl groups. That a reaction of the type So+ TI + Sois not observed may be due to the fact that the first crossing can be reached many times until spin inversion finally takes place, while the second crossing is passed only once. Probabilities are therefore much lower for the TI -* Sotransition than for the So+ T, transition. The situation is somewhat different in the case of dewarbenzene, which undergoes an electrocyclic ring opening to give triplet excited benzene. (Cf. Example 7.16.) In contrast to I ,2-dioxetanes, this reaction possesses a very low chemiluminescence efficiency. The reason is thought to be the low intersystem crossing probability, which is due to the very weak spin-orbit coupling inherent in hydrocarbon systems. Thus, although the available energy is very favorable for chemiluminescence, the rearrangement proceeds as a ground-state reaction. Example 7.22: It is generally agreed that the excitation-producing step in the oxidation of luminol(182) is decomposition with loss of nitrogen of the dianion of an azoendoperoxide produced by the action of a base and oxygen:
Figure 7.57. State correlation diagram for the chemiluminescence reaction of tetramethyl-] ,Zdioxetane (by permission from Turro, 1978).
This can be viewed as an allowed [2, + 2, + 2,J pericyclic reaction, and it is not obvious that there should be an avoided or unavoided touching of Sowith S, that could provide an easy "surface jump" to the S, surface along the way.
484
ORGANIC PHOTOCHEMISTRY
The following alternative mechanism may well provide an explanation of luminol chemiluminescence:
The first step is analogous to the easy retro-Diels-Alder reaction of 1.4-dihydrophthalazine (183). whose A I F is 15 kcavmol (Flynn and Michl, 1974). The ground state of the resulting energy-rich peroxide dianion has 18 n electrons. Stretching of the 0-0 bond converts it into a much more stable aminophthalate dianion product, whose ground state, however, has only 16 n electrons. It therefore does not correlate with the ground state, but with a doubly n-dc excited state of the peroxide. The 18 n-electron ground state of the peroxide correlates with a doubly n-n* excited state of the aminophthalate dianion product. At high symmetries, the crossing of the two states will be avoided. In the first approximation, the splitting will be given by twice the exchange integral between the two orbitals that cross (cf. Section 4.2.3). One of these is a
415
SUPPLEMENTAL READING
n-type orbital, the other an n-type orbital, and the overlap density and the exchange integral will therefore be only very small. It is likely that an S , S , conical intersection can be reached when the symmetry is lowered, and this ought to provide a facile radiationless crossing between the two surfaces as indicated in Figure 7.58 (Michl, 1977).
Several additional chemiluminescence mechanisms have been described, which are based on excited-state generation by electron transfer in a radical ion pair according to where D@ + A- constitutes an excited state of the system D + A. The radical ion pair may be produced either electronically or chemically by electron transfer to a peroxide that subsequently rearranges and loses a neutral molecule. An example of this CIEEL (chemically induced electron-exchange luminescence) mechanism is provided by the thermal reaction of diphenoyl peroxide (184) (Koo and Schuster, 1977):
184
ACT = activator, for example, aromatic hydrocarbon.
Supplemental Reading Barltrop, J.A., Coyle, J.D. (1975),Excited States in Organic Chemistry; Wiley: London. Coyle, J.D., Ed. (1986), Pholochemistry in Organic Synthesis, Special Publ. 57; The Royal Soc. Chem.: London. Cowan, D.O., Drisko, R.L. (1976). Elerrrents New York.
c?f
Organic Photochemistry; Plenum Press:
Coxon, J.M., Halton, B. (1974), Organic Photochemistry; Cambridge University Press: Cambridge. Gilbert, A., Baggott, J. (1991). Essentials of Molecular Photochemistry; CRC Press: Boca Raton. Horspool, W.M., Ed. (1984). Synthetic Organic Photochemistry; Plenum Press: New York. Reactant
Product
Figure 7.58. Schematic state correlation diagram for the chemiluminescence of a strongly exothermic reaction forbidden in the ground state (adapted from Michl, 1977).
Kagan. J. (1993), Organic Photochemistry: Principles and Applications; Academic Press: London. Kopecky, J. (1992), Organic Photochemistry: A Visual Approach; VCH: New York Ramamurthy, V.,Turro, N.J.,Eds. (1993). "Photochemistry," thematic issue of Chem. Rev. 93, 1-724.
486
UI, I;IWIC PHOTOCHEMISTRY
Scaiano, J.C., Ed. (1989), Handbook of Organic Photochentistry; CRC Press: Boca Raton, Vols. I and 11.
Houk, K.N. (1976). "The Photochemistry and Spectroscopy of B,y-Unsaturated Carbonyl Compounds," Chem. Rev. 76. I .
Turro, N.J. (1978), Modern Molecular Photochemistry; BenjaminICummings Publ.: Menlo Park.
Kramer, H.E.A. (1990). "Salicylates, Triazoles, Oxazoles," in Photochromism, Molecules and Systems; Durr, H., Bouas-Laurent, H., Eds.; Elsevier: Amsterdam.
Wayne, R.P. (19881, Principles and Applications of Photoclremistry; Oxford University Press.
Sammes, P.G. (1986), "Photoenolization." Acc. Chem. Res. 4, 41.
Collin, G.J. (1987). "Photochemistry of Simple Olefins: Chemistry of Electronic Excited States or Hot Ground State?", Adv. Photochem. 14, 135.
Schuster. D.I. (1980). "Photochemical Rearrangements of Enones," in Rearrangements in Ground and Excited States, 3; de Mayo, P., Ed.; Academic Press: New York.
Dauben, W.G., McInnis, E.L., Michno, D.M. (1980). "Photochemical Rearrangement in Trienes," in Rearrangements in Ground and Excited state.^. 3; de Mayo. P., Ed.; Academic Press: New York.
Turro. N.J., Dalton, J.C., Dawes, K., Farrington, G., Hautala, R., Morton. D., Niemczyk, M., Schore, N. (1972)- "Molecular Photochemistry of Alkanones in Solution: a-Cleavage, Hydrogen Abstraction, Cycloaddition, and Sensitization Reactions," Acc. Chem. Res. 5, 92.
Gorner, H., Kuhn. H.J. (1995). "Cis-trans Photoisomerization of Stilbene and Stilbene-Like Molecules," Adv. Photochem. (Neckers, D.C., Volman, I1.H.. von Bunau, G., Eds.) 19, 1. Jacobs, H.J.C.. Havinga, E. (1979). "Photochemistry of Vitaniin D and Its Isomers and of Simple Trienes," Adv. Photochem. 11,305. Saltiel, J.. Sun. Y.-P. (1990). "Cis-trans Isomerization of C=C Double Bonds," in Photochromism, Molecules and Systems; Dilrr. H.. Bouas-Laurcnt, H., Eds. ; Elsevier: Amsterdam. Saltiel, J.. Sun, Y.-P. (1989). "Application of the Kramers Equation to Stilbene Photoisomerization in n-Alkanes Using Translational Diffusion Coefficients to Define Microviscosity," J. Phys. Chem. 93, 8310.
Azoalkanes and Azomethines Adam, W., De Lucchi, 0. (1980). "The Synthesis of Unusual Organic Molecules from Azoalkanes." Angew. Chem. Int. Ed. Engl. 19, 762.
Schaffner, K., Demuth, M. (1980). "Photochemical Rearrangements of Conjugated Cyclic Dienones," in Rearrangements in Ground and Excited States. 3; de Mayo, P., Ed.; Academic Press: New York.
Wagner, P.J. (1980). "Photorearrangements via Biradicals of Simple Carbonyl Compounds," in Rearrangements in Ground and Excited States. 3; de Mayo, P., Ed.; Academic Press: New York. Wagner. P.J., Park. B.-S. (1991). "Photoinduced Hydrogen Atom Abstraction by Carbonyl Compounds." Org. Photochem. (Padwa, A., Ed.) 11, I I!.
Photocycloadditions Arnold, D.R. (1968). "The Photocyloaddition of Carbonyl Compounds to Unsaturated Systems: The Syntheses of Oxetanes," Adv. Photochem. 6, 301. Becker, H.-D. (1990). "Excited State Reactivity and Molecular Topology Relationships in Chromophorically Substituted Antracenes," Adv. Photochem. 15, 139. Bernardi, F., Olivucci, M.. Robb, M. A. ( 1990). "Predicting Forbidden and Allowed Cycloaddition Reactions: Potential Surface Topology and Its Rationalization," Acc. Chem. Res. 23,405.
Durr, H., Ruge, B. (1976). "Triplet States from Azo Compounds," Topics Curr. Chem. 66, 53. Engel. P.S. (1980), "Mechanism of the Thermal and Photochemical Decomposition of Azoalkanes," Chem. Rev. 80.99.
Bouas-Laurent. H.. Desvergne, J.-P. (1990). "Cycloaddition Reactions Involving 4n.Electrons: [4 41 Cycloaddition Reactions Between Unsaturated Conjugated Systems," in Photochromism. Molecrtlrs and Systems; Diirr. H., Bouas-Laurent, H., Eds.; Elsevier: Amsterdam.
Meier, H.. Zeller, K.-P. (1977). "Thermal and Photochemical Elimination of Nitrogen," Angent. Chem. Int. Ed. Engl. 16,835.
Caldwell, R.A., Creed, D. (1980). "Exciplex Intermediates in [2 Acc. Chem. Res. 13,45.
Paetzold, R., Reichenbticher, M. Appenroth, K. (1981). "Die Kohlenstoff-Stickstoff-Dop pelbindung: Spektren, Struktur, thermische und photochemische EIZ-lsomerisierung," Z. Chem. 12.42 1.
Cornelisse, J. (19931, "The Meta Photocycloaddition of Arenes to Alkenes," Chem. Rev. 93,615.
Rau, H. (1990). "Azo Compounds" in Photochromism. Molec.rrles and Systems; DOrr. H.. Bouas-Laurent, H., Eds.; Elsevier: Amsterdam. Wentrup. C. (1984). Reactive Molecules; Wiley: New York.
Carbonyl Compounds Cohen, S.G., Parola, A., Parsons Jr., G.H. (1973). "Photoreduction by Amines." Chem. Rev. 73, 141.
+
+ 21 Photocyloadditions,"
Crimmins, M.T. (1988). "Synthetic Applications of Intramolecular Enone-Olefin Photocycloadditions," Chem. Rev. 88. 1453. Desvergne, J.-P.. Bouas-Laurent. H. (1990). "Cycloaddition Reactions Involving 4n Electrons: [2 + 21 Cycloadditions; Molecules with Multiple Bonds Incorporated in or Linked to Aromatic Systems," in Photochromism, Molecrtles and Systems; Durr, H., BouasLaurent, H., Eds.; Elsevier: Amsterdam. Jones, 11. G. (1990). "Cycloaddition Reactions Involving 4n Electrons: [2 + 21 Cycloadditions; Photochemical Energy Storage Systems Based on Reversible Valence Photoiso-
ORGANIC PHO'I'OCHEMISTRY SUPPLEMENTAL READING merization" in Photochromism, Molecules und Systems; Diirr, H., Bouas-Laurent, H., Eds.; Elsevier: Amsterdam.
489
Photosubstitution
Laarhoven. W.H. (1987), "Photocyclizations and intramolecular Cycloadditions of Conjugated Olefins," Org. Photochem. (Padwa, A., Ed.) 9, 129.
Cornelisse, J., Havinga, E. (1975). "Photosubstitution Reactions of Aromatic Compounds," Chem. Rev. 75, 353.
McCoullough, J.J. (1987). "Photoadditions of Aromatic Compounds," Chem. Rev. 87, 81 1.
Davidson, R.S., Gooddin, J.W., Kemp, G. (1984). "The Photochemistry of Aryl Halides and Related Compounds," Adv. Phys. Org. Chem. 20, 191.
Schuster, D.I. (1993). "New Mechanism for Old Reactions: Enone Photocycloaddition," Chem. Rev. 93, 3. Wagner-Jauregg, T. (1980). "Thermische und photochemische Additionen von Dienophilen an Arene und deren Vinyloge und Hetero-Analoge." Synthesis 165.769. Wender, P.A. (1986). "Alkenes: Cycloaddition," in Photochemistry in Orgunic Synthesis, Special Publ. 57; Coyle, J.D., Ed.; The Royal Soc. Chem.: London. Wender, P.A., Siggel, L.. Nuss, J.M. (1989). "Arene-Alkene Photocycloaddition Reactions." Org. Photochem. (Padwa, A., Ed.) 10, 356.
Rearrangement Reactions Arai, T., Tokumaru, K. ( 1993). "Photochemical One-way Isomerization of Aromatic Olefins," Chem. Rev. 93. 23. Bryce-Smith. D., Gilbert, A. (1980). "Rearrangements of the Benzene Ring," in Reurrungements in Ground and Excited Stutes. 3; de Mayo, P., Ed.; Academic Press: New York: Demuth, M. (1991), "Synthetic Aspects of the Oxadi-n-Methane Rearrangement," Org. Photochem. (Padwa, A., Ed.) 11.37.
Havinga, E., Cornelisse, J . (1976), "Aromatic Photosubstitution Reactions." Pure Appl. Chem. 47, 1. Kropp, P.J. (1984). "Photobehavior of Alkyl Halides in Solution: Radical, Carbocation, and Carbene Intermediates," Arc. Chem. Res. 17, 131.
Singlet Oxygen Foote, C.S. (1968). "Photosensitized Oxygenations and the Role of Singlet Oxygen," Acc. Chem. Res. 1 , 104. Lissi, E., Encinas, M.V., Kubio, M.A. (1993). "Singlet Oxygen Bimolecular Photoprocesses," Chem. Rev. 93, 698. Stephenson, L.M., Grdina, M.J., Orfanopoulos, M. (1980), "Mechanism of the Ene Reaction between Singlet Oxygen and Olefins," Acc. Chem. Res. 13,419. Wasserman, H.H., Murray, K.W. (1979). Singlet Oxygen; Academic Press: New York.
Chemiluminescence
Leigh, W.J. (1993). "Techniques and Applications of Far UV Photochemistry. The Photochemistry of the C,H, and C,H, Hydrocarbons," Chem. Rev. 93,487.
Gundermann, K.-D., McCapra, F. (1987). Chemiluminescence in Orgunic Chemistry; Springer: Berlin.
Zimmerman, H.E. (1991). "The Di-v-Methane Rearrangement," Org. Photochem. (Padwa, A., Ed.) 11, I .
Schuster, G.B., Schmidt, S.1'. (1982). "Chemiluminescence of Organic Compounds," Adv. Phys. Org. Chem. 18, 187.
Electron-Transfer Reactions Davidson, R.S. (1983). "The Chemistry of Excited Complexes: a Survey of Reactions," Adv. Phys. Org. Chem. 19. 1. Eberson, L. (1987). Electron Transfer Reactions in Orgunic Chemistry; Springer: Berlin.
Turro, N.J., Ramamurthy, V. (1980). "Chemical Generation of Excited States"; in Rrcrrrungements in Ground ctnd Excited Stutes, 3; de Mayo, P., Ed.; Academic Press: New York.
Various
FOX,M.A., Ed. (1992). "Electron-Transfer Reactions," Chem. Rev. 92,365.
Balzani, V., Scandola, F. (1991). Suprumolecular Photochemistry; Ellis Horwood: New York.
Fox, M.A. (1986). "Photoinduced Electron Transfer in Organic Systems: Control of the Back Electron Transfer," Adv. Photochem. 13,237.
BUnau, G. von, Wolff, T. (1988). "Photochemistry in Surfactant Solutions," Adv. Pl~otochem. 14,273.
Julliard, M., Chanon, M. (1983). "Photoelectron-Transfer Catalysis: Its Connections with Thermal and Electrochemical Analogues." Chem. Rev. 83,425.
Dbrr, H., Bouas-Laurent, H.. Eds. 0990). Photochromism, Molecrtles ctnd Systems; Elsevier: Amsterdam.
Kavarnos, G.J., Turro, N.J. (1986). "Photosensitation by Reversible Electron Transfer: Theories, Experimental Evidence, and Examples." Chem. Rev. 86,401.
Guillet, J. (19851, Polymer Photophysics and Photochemistry; Cambridge University Press: Cambridge.
Mariano, P.S., Stavinoka, J.L. (1984). "Synthetic Aspects of Photochemical Electron Transfer Reactions," in Synthetic Organic Photochemistry; Horspool. W.M., Ed.; Plenum Press: New York.
Platz, M.S., Leyva, E., Haider, K. (1991). "Selected Topics in the Matrix Photochemistry of Nitrenes, Carbenes, and Excited Triplet States," Org. Photochem. (Padwa, A., Ed.) 11,367.
Mattay, J. (1987). "Charge Transfer and Radical Ions in Photochemistry,"Angew. Chem. Int. Ed. Engl. 26, 825.
Turro, N.J., Cox. G.S., Paczkowski, M.A. (19851, "Photochemistry in Micelles," Top. Curr. Chem. 129.57.
Yoon, U.C., Mariano, P.S. (1992). "Mechanistic and Synthetic Aspects of Amine-Enone Single Electron Transfer Photochemistry," Acc. Chem. Res. 25, 233.
Epilogue
In the preceding seven chapters, we have gradually developed the framework necessary for a qualitative understanding of the photophysical and photochemical behavior of organic molecules in terms of potential energy surfaces. After introducing the basics of electronic spectroscopy, ordinary and chiral, and the fundamental concepts of photophysics in Chapters 1 4 , we described the basic notions of organic photophysics and photochemistry in Chapter 5 and 6, and illustrated their utility on a fair number of specific examples in Chapter 7. Throughout, we have attempted to concentrate on the key concepts provided by the quantum theory of molecular structure, and to relate these to experimental observations. In a sense, this text aspires to being a textbook of both theoretical and mechanistic photochemistry but it makes no pretext of providing practical experimental information on light sources and the like. More than anything else, our goal has been to introduce the reader to a way of thinking about problems in photophysics and photochemistry. Although many additional organic photochemical processes could be added to Chapter 7, we have chosen not to do so. Instead, we hope that the reader will be able to apply the understanding of the material that we have chosen to present as he or she approaches the study of additional reactions.
References
Abrash, S., Repinec, S., Hochstrasser, R.M. (1990), J. Chem. Phys. 93, 1041. Adam, W., d e Sanabia, J.A., Fischer, H. (1973), J. Org. Chem. 38,2571. Adam, W., Mazenod, F., Nishizawa. Y., Engel, P.S., Baughman, S.A.. Chae, W.-K.. Horsey, D.W., Quast, H., Seiferling, B. (1983), J. Am. Cltem. Soc. 105, 6141. Adam, W., Grabowski, S., Wilson, R.M. (1990), Ace. Chem. Res. 23, 165. Agosta, W.C., Smith, 111, A.B., Kende, A.S., Eilermann, R.G., Benham, J. (1969). Tetruhedron Lett. 8, 4517. Albert, B., Berning, W., Burschka, C., Hiinig, S., Prokschy, F. (1984). Chern. Ber. 117, 1465.
Allinger, N.L. (1977). J. Am. Chern. Soc. 99, 8127. Almgren, M. (1972), M o l . Photochem. 4,213. Alves, A.C.P., Christofferson, J., Hollas, J.M. (1971 ), M o l . Phys. 20, 625. Angliker, H., Rommel, E., Wirz, J. (1982). Chem. Phys. Lett. 87, 208. Aoyagi, M.,Osamura, Y., I v ~ a t a S , . (1985), J. Chem. Phys. 83, 1140. Applequist, D.E., Litle, R.L.., Friedrich, E.C., Wall, R.E. (1959). J. Am. Chem. Soc. 81, 452. Arnett, J.F., Newkome, G., Mattice, W.L., McGlynn, S.P. (1974), J. Am. Chem. Soc. 96, 4385. Arnold, B.R., Balaji, V., Michl, J. (19901, J. Am. Chem. Soc. 112, 1808. Arnold, B.R., Balaji, V., Downing, J.W., Radziszewski, J.G., Fisher, J.J., Michl, J. (1991). J. Am. Chem. Soc. 113, 2910.
REFERENCES Arnold, D.R., Trecker, D.J., Whipple, E.B. (1965). J. Am. Cl1c.m. Soc. 87, 25%. Atchity, G.J., Xantheas S.S., Ruedenberg, K. (1991) J. Chem. Phys. 95, 1862. Aue, D.H., Reynolds, R.N. (1973). J. Am. Chem. Soc. 95, 2027. Bachler, V., Polanski, O.E. (1988). J. Am. Chem. Soc. 110, 5972,5977.
Bischof, H., Baumann, W., Detzer, N., Rotkiewicz, K. (1983, Chcm. Phys. Lett. 116, 180. Bixon, M., Jortner, J. (1968), J. Chcm. Phys. 48, 715. Bixon, M., Jortner, J. (1989). Chem. Phys. Lett. 159, 17. Bixon, M., Michel-Beyerle. M.E., Jortner, J. (1988). Israel J. Chem. 28, 155.
Badger, G. M. ( 1954). The Strut-tirre und Rc~actionsof Aromcrtic Compounds; Cambridge University Press: Cambridge.
Blair, J.T., Krogh-Jesperen, K., Levy, R.M. (1989). J. Am. Chem. Soc. 111,6948.
Baggiolini, E., Hamlow, H.P., Schaffner, K. (1970). J. Am. C h m . Soc. 92, 4906.
Blais, N.C., Truhlar, D.G., Mead, C.A. (1988). J. Ckem. Phys. 89, 6204.
Balaji, V., Michl, J. (19911, Polyhedron 10, 1265.
Blout, E.R., Fields, M. (1948). J. Am. Chem. Soc. 70, 189.
Baldwin, J.E. (1976). J.C.S. Chem. Commun. 734.
BonaCiC-Kouteckq, V., Ishimaru, S. (1977). J. Am. Chem. Soc. 99, 8134.
Barber, L.L., Chapman, O.L., Lassila, J.D. (1969), J. Am. Chclm. Soc. 91, 3664.
BonaCiC-Kouteckq, V., Michl, J. (1985a). Tlteor. Chim. Actu 68, 45.
Barltrop, J.A. (1973). Pure Appl. Chem. 33, 179.
BonaCiC-Koutecky, V., Michl, J. (1985b). J. Am. Chem. Soc. 107, 1765.
Barltrop, J.A., Carless, H.A.J. (1972), J. Am. Chem. Soc. 94, 1951.
BonaCiC-Kouteckg, V., Bruckmann. P., Hiberty. P.. Koutecky, J., Leforestier, C., Salem, L. ( 1975). Angew. Cheri~.Int. Ed. Engl. 14, 575.
Bartlett, P.D., Porter, N.A. (1968). J. Am. Chem. Soc. 90, 5317. Bauschlicher, Jr., C. W., Langhoff, S.R. (199 I), Theor. Chim. Arta 79.93. Bearpark, M.J., Robb, M.A., Schlegel, H.B. (1994). Cltem. Phys. Lett. 223,269. Beck, S.M.. Liverman, M.G., Monts, D.L., Smalley, R.E. (1979). J. Chem. Phys. 70,232.
BonaCiC-Koutecky, V., Michl, J., Kohler, J. (1984). Chem. Phys. Lett. 104, 440. BonaCiC-Kouteckq. V., Koutecky, J., Michl, J. (1987). Angew. Cltem. Int. Ed. Engl. 26, 170. BonaCiC-Kouteckq, V., Schoffel, K., Michl, J. (1989). J. Am. Chem. Soc.. 111, 6140.
Becker, H.-D. (1982). Pirre AppI. Chem. 54, 1589.
Bonneau, R., Joussot-Dubien, J . , Salem, L., Yarwood, A.J. (1976). J. Am. Chem. Soc. 98, 4329.
Becker, R.S., Freedman, K. (1985). J. Am. Cltem. So(.. 107, 1477.
Borell, P.M., Lohmannsroben, H.-G., Luther, K. (1987). Cliem. Phys. Lett. 136, 371.
Beckmann, S., Wessel, T., Franck, B.. Honle. W., Borrmann. H.. von Schnering, H.G. (1990). Angew. Cltem. Int. Ed. Engl. 29, 1395.
Boumann, T.D., Lightner, D.A. (1976). J. Am. Cltern. Soc. 98, 3145.
Beens, H., Weller, A. (1968). Chem. Phys. Lett. 2, 140. Beens, H.. Knibbe, H., Weller, A. (1967). J. Cltem. Pltys. 47, 1183. Beer, M., Longuet-Higgins, H.C. (1955), J. Cltem. Phys. 23, 1390. Bentzien, J., Klessinger, M. (1994). J. Org. Chem. 59, 4887. Bernardi, F., Olivucci, M., McDonall, J.J.W., Robb, M.A. (1988). J. Chem. Pltys. 89,6365. Bernardi, F.. De, S., Olivucci, M., Robb, M.A. (1990a) J. Am. Ckem. Soc. 112, 1737. Bernardi, F., Olivucci, M., Robb, M.A. (1990b). Ace. Chem. Res. 23,405. Bernardi, F., Olivucci, M. Robb, M.A., Tonachini, G. (1992a), J. Am. Chem. Soc. 114,5805.
Boxer, S.G. (1990). Annrr. Rev. Biopliys. Cltem. 19, 267. Briegleb, G., Czekalla, J. ( 1960), Ange\v. Cltem. 72, 401. Brogli, F., Heilbronner, E., Kobayashi, T. (1972). Helv. Cltirn. Actcr 55, 274. Brooker, L.G.S., Keyes, G.H., Heseltine, D.W. (1951), J. Am. Cliern. Soc. 73, 5350. Bruckmann. P., Salem, L. (1976). J. Am. Cliern. Soc.. 98. 5037. Bryce-Smith. D. (1968). Prrre Appl. Chem. 16, 47. Bryce-Smith, D. (1973). Pure Appl. Chem. 34, 193. Bryce-Smith, D., Gilbert, A. (1976). Tetruhedron 32, 1309. Bryce-Smith, D., Gilbert, A. (1980). in Rearrangements in Ground and Excited States, Vol. 3; de Mayo. P., Ed.; Academic Press: New York; p. 349.
Bernardi, F., Olivucci, M., Ragazos, I.N., Robb, M.A. (1992b). J . Am. Chem. Soc. 114, 821 1. Bernardi, F., Olivucci, M., Ragazos, I.N., Robb, M.A. (1992~).J. Am. Chem. Soc. 114, 2752.
Bryce-Smith, D., Gilbert, A., Orger, B.H. (1966). J.C.S., Chem. Commun. 512.
Bigot. B. (1980). in Qucrntrrm Theory of Chemical Reactions. Vol. 11; Daudel, R.. Pullman. A., Salem, L.. Veillard, A., Eds.; Reidel: Dordrecht.
Bryce-Smith, D., Gilbert, A., Robinson, A.D. (1971), Angen~.Cltem. Int. Ed. Engl. 10,745.
Bigot, B. (1983). Israel J. Chent. 23, 1 16. Bigot, B., Sevin, A., Devaquet, A. (1978), J. Am. Chem. Soc. 100,2639. Bigwood, M., Bout, S. (1974). J.C.S. Chem. Commun. 529. Birks, J.B. (1970). Photophysics of Aromatic Molecules; Wiley: London. New York.
Bryce-Smith, D., Gilbert, A. Grzonka, J. (1970). J.C.S. Chem. Commrrn. 498.
Bryce-Smith, D., Gilbert, A., Mattay, J. (1986). Tetrahedron 42, 601 1. Buckingham, A.D., Ramsy, D.A., Tyrell, J. (1970). Can. J. Phys. 48, 1242. Buenker, R.J., Peyerimhoff, S.D. (1974). Chem. Rev. 74, 127. Buenker, R.J., Shih, S., Peyerimhoff, S.D. (1976), Chem. Phys. Lett. 44, 385. Buenker, R.J., BonaCiC-Kouteckv, V., Poglinai, L. (1980). J. Cltem. Phys. 73, 1836.
REFERENCES
REFERENCES
Buncel, E., Rajagopal, S. (1990), Ace. Chum. Res. 23, 226.
Cohen, S.G., Zand, R. (1962). J. Am. Chem. Soc. 84, 586.
Bunnett, J.F, (1978), Acc, Chem. Res. 11,413.
Cohen, S.G., Parola, A., Parsons, Jr., G.H. (1973), Chem. Rev. 73, 141.
Bunton, C.A., Savelli, G. (1986). Adv. Phys. Org. Chem. 22,213.
Colle, R., Montagnani, R., Kiani, P., Salvetti, 0. (1978). Theor. Chim. Actu 49, 37.
Buschmann, H., Scharf, H.-D., Hoffmann, N., Plath, M.W.. Runsink, J. (1989), J. Am. Chem. Soc. 111,5367.
Condon, E.U. (1928), Phys. Rev. 32,858.
Buschmann, H., Scharf, H.-D., Hoffmann, N., Esser, P. (1991). Angew. Chem. Int. Ed. Engl. 30,477. Caldwell, R.A. (1980), J. Am. Chem. Soc. 102, 4004. Caldwell, R.A., Carlacci, L, Doubleday, Jr., C.E., Furlani, T.R., King, H.F., Mclver, Jr., J.W. (1988). J. Am. Chem. Soc. 110,6901. Calhoun, G.C., Schuster, G.B. (1984). J. Am. Chem. Soc. 106, 6870. Callomon, J.H., Dunn, T.M., Mills, I.M. (1966). Phil. Truns. R. Soc. London 259A, 499.
Constanciel, R. (1972), Theor. Chim. Actu 26, 249. Cookson, R.C. (19681, Q. Rev. 22,423. Corey, E.J., Bass, J.D., LeMahieu, R., Mitra, R.B. (1%4), J. Am. Chem. Soc. 86, 5570. Cotton, F.A. (197 1), C h e m i c ~Applicutions l of Group Theory; Wiley: Chichester Cowan, D.O., Drisko, R.L.E. (1970a,b), J. Am. Chem. Soc. 92, 6281,6286. Cowan, D.O.. Gleiter, R., Hashmall, J.A., Heilbronner, E., Hornung, V. (1971), Angew. Chem. Int. Ed. Engl. 10, 401.
Calzaferri, G., Gugger, H., Leutwyler, S. (1976). Helv. Chint. Actu 59, 1969.
Coxon, J.M., Halton, B. (1974). Organic Photochemistry; Cambridge University Press: Cambridge.
Carlacci, L., Doubleday, Jr., C., Furlani, T.R., King, H.F., Mclver, Jr., J.W. (1987). J. Am. Chem. Soc. 109,5323.
Craig, D.P. (1950). J. Chem. Sue. (London), 2 146.
Carless, H.A.J. (1973). Tutrcrhedrort Lett. 14, 3 173. Casey, C.P., Boggs, R.A. (1972). J. Am. Chem. Soc. 94,6457. Castellan, A., Michl, J. (1978). J. Am. Chem. Soc. 100, 6824.
Dalton, J.C., Wriede, P.A., l'urro, N.J. (1970), J. Am. Chem. Soc. 92, 1318. Dannenberg, J.J., Rayez, J.C. (1983), J. Org. Chem. 48,4723. Das, P.K., Becker, R.S. (1978). J. Phys. Chum. 82, 2081.
Castellan, A., Kolc, J., Michl, J. (1978), J. Am. Chem. Soc. 100, 6687.
Das, P.K., Encinas, M.V., Small, Jr., R.D., Scaiano, J.C. (1979). J. Am. Chem. Soc. 101, 6%5.
Cave, R.J., Davidson, E.R. (1988), Chem. Phys. Lett. 148, 190.
Datta, P., Goldfarb, R.D., Boikess, R.S. (1971). J. Am. Chum. Soc. 93, 5189.
Celani, P., Ottani, S., Olivucci, M.. Bernardi, F., Robb, M.A. (1994). J. Am. Chem. Soc. 116, 10141.
Dauben, W.G., Cargill, R.L. f 1961). Tetruhedron 12, 186.
~. 102. 5733 Celani. P.. Be!-nardi. F.. Olivucci. M.. Robb. M.A. (IWS). J. C l r c ~ nP11y.s.
Chae, W.-K., Baughman, S.A., Engel, P.S., ~ r u c h M., , Ozmeral, C., Szilagyi, S., Timberlake, J.W. (1981), J. Am. Chem. Soc. 103,4824. Charlton, J.L., Dabestani, R., Saltiel, J. (1983)-J. Am. Clrem. Soc. 105, 3473. Charney, E. (1979). The Molrculur Basis of Opticul Acitivity; Wiley: New York. Chu, N.Y.C., Kearns, D.R. (1970), J. Phys. Chem. 74, 1255. Ciamician, G., Silber, P. (1900), Berchte 33. 291 1. Clar, E. (1964), Polycyclic Hydrocarbons; Academic Press: London, New York. Clark, K.B., Leigh, W.J. (1987), J. Am. Chem. Soc. 109, 6086. Closs, G.L., Miller, J.R. (1981). J. Am. Chem. Soc. 103, 3586.
Dauben, W.G., Kellogg, M.S. (19711, J. Am. Chem. Soc. 93, 3805. Dauben, W.G., Salem, L., Turro, N.J. (1975), Ace. Chem. Res. 8 , 41. Dauben, W.G., Lodder, G., Robbins, J.D. (1976), J. Am. Chem. Soc. 98, 3030. Day, A.C., Wright, T.R. (1969), Tetruhedron Lett., 1067. De Mayo,P.(1971),Acc. Chew. Res.4.41. De Mayo, P., Shizuka, H. (1973). J. Am. Chem. Soc. 95, 3942. De Vaal, P., Lodder, G., Cornelisse, J. (1986), Tetrahedron 42, 601 1. DeBoer, C.D., Herkstroeter. W.G., Marchetti, A.P., Schultz. A.G., Schlessinger, R.H. (1973). J. Am. Chem. Soc. 95, 3963. Dehareng, D., Chapuisat, X., Lorquet, J.-C., Galloy, C., Raseev. G. (1983). J. Chem. Phys. 78, 1246.
Closs, G.L., Miller, J.R. (1988). Science 240,440.
Deisenhofer. I. Epp, O., Miki. K., Huber, R., Michel, H. (1985). Nature 318. 618.
Closs, G.L., Calcaterra, L.T., Green, N.J., Penfield, K.W., Miller, J.R. (1986). J. Phys. Chem. 90, 3673.
Del Bene, J., Jaffe, H. H. ( 1x8). J. Chem. Phys. 48, 1807.
Closs, G.L., Johnson, M.D., Miller, J.R., Piotrowiak, P. (1989). J. Am. Chern. Soc. 111, 375 1. Closs, G.L., Forbes, M.D.E., Piotrowiak, P. (1992). J. Am. Chum. Soc., 114, 3285.
Devaquet. A.. Sevin. A.. Bigot. B. (1978), J. Am. Chem. Sue. 100. 2009. Dewar, M.J.S. (19501, J. Chenr. Sac. (London), 2329. Dewar, M.J.S. (1969), The Moleculur Orbitul Theory of Organic Chemistry; McGraw-Hill: New York.
I
Dewar, M.J.S., Doubleday, C. (1978). J. Am. Chem. Soc. 100,4935.
Eriksson. M.. Norddn, B.. Lycksell. P.-0.. Grislund. A.. Jernstrom. B. (1985). J.C.S. Chem. Comnir~rt.1300.
Dewar, M.J.S., Dougherty, R.C. (1975), The P M O Tlieory of Organic Chemistry; Plenum Press: New York, London.
Evleth. E.M.. Kassab. E. ( 1978). J. Ant. Chent. Soc.. 110. 7859.
Dewar, M.J.S., Thiel, W. (1977), J. Am. Chem. Soc. 99, 2338.
Eyring. H.. Walter. J.. Kimball. G. (1944). Qr~trrttrrr~i Ckcjntistry; Wiley: New York.
Dewar, M.J.S., Thompson, Jr., C.C. (1966), Tetrahedron Suppl. 7.97.
Fabian. W. ( 1985). Z. Ntrtr~r:fbrsc.lt.40a. 279.
Dexter, D.L. (1953). J. Chem. Phys. 21, 836.
Farid. S.. Shealer. S.E. ( 1973). J.C.S. C ~ P I I IContntrrrt. . 677.
Dick, B., Nickel, B. (1983). Chem. Phys. 78, 1.
Feher. G.. Allen. J.P.. Okamura. M.Y.. Rees. D.C. ( 1989). N t ~ t r r 339. r~ l 1I.
Dick, B., Gonska, H., Hohlneicher, G. (1981), Ber. Bunsenges. Phys. Chem. 85,746.
Fessner. W.-D.. Prinzbach. H.. Rihs. G. (1983). Etrtrhe~tlro~t Lc.tr. 24. 5857.
DMS-UV-Atlas organisclrer Verbindungen ( 1966-1971 ); But terworths: London: Verlag
Chemie: Weinheim.
Fessner. W.-D.. Sedelmeier. G.. Spurr. P.R.. Rihs. G.. Prinzbach. H. (1987). J . Ant. Chem. Soc.. 109. 4626.
Doany, F.E., Hochstrasser, R.M., Greene, B.I., Millard, R.R. (1985). Chem. Phys. Lett. 118, 1.
Fieser. L.F.. Fieser. M.. Rajagopalan. S. (1948). J. Org. Cltertt. 13. 800.
Dohnert, D., Kouteckq, J. (1980), J. Am. Chem. Soc. 102, 1789.
Fischer. S.F.. van Duyne. R.P. (1977). Cheni. Pliys. 26. 9.
Dorr, F. ( 1966). in Optische Anregung organischer Systeme; Foerst , W., Ed. ; Verlag Chemie: Weinheim.
Flynn. C.R.. Michl. J. (1974). J. Ant. Cllern. Soc.. 96. 3280.
Dorr, F. Held, M. (1960). Angent. Chem. 72, 287. Doubleday, Jr., C., Mclver, Jr., J.W., Page, M. (1982). J. Am. Chem. Soc.. 104,6533.
Chc~ni.1111.Etl. Otgl. 17. 16. Fischer. M. ( 1978). A~tgc.~c*.
Forster. E.W.. Grellmann. K.H.. Linschitz. H. (1973). J. A ~ t t .Chc~nt.Soc.. 95. 3108. Forster. T. ( 1950). Ber. Brrn.\c~rtges.Phys. C'lic~nt.54. 53 1.
Doubleday, Jr., C., Mclver, Jr., J.W., Page, M. (1985). J . Am. Chem. Soc. 107,7904.
Forster. T. (1951). FIrrorus,-.enz orgonisc.lit~r Vc.rhi~telrr~igc~rt: Vandenhoek und Ruprecht: Gottingen.
Doubleday, Jr., C., Turro, N.J., Wang, J.-F. (1989), Acc. Chent. Res. 22, 199.
Forster, T. ( 1959). Disc.. firrtrtlery Soc.. 27. 7.
Dougherty. T.J. (1992). Adv. Photoc-hemistry 17, 275.
Forster. T. (1970). Prrrc. Appl. Cltc~ni.24. 443.
Dowd, P. (1966), J. Am. Chem. Soc. 88,2587.
Forster. T.. Kasper. K. (1955). Z. EleXtroc.hent. 59. 976.
Downing, J.W., Michl, J., JBrgensen, P., Thulstrup, E.W. (1974). Tlteor. Chim. Acta 32,203.
Forster. T.. Seidel. H.-P. ( 1965). Z. Phys. Cltc.111.N . F. 45. 58.
Dreeskamp, H., Koch, E., Zander, M. (1973, Chem. Pliys. Lert. 31, 251.
Forbes. M.D.E.. Schulz. G.R. (1994). J. Ant. Cht.111.Sot.. 116. 10174.
Dreyer. J.. Klessinger. M. (1994). J. Cltc>rit.Phys. 101. 10655.
Franck. J. ( 1926). Trtrns. Ftrrtrtlery Soc.. 21, 536.
Dreyer. J.. Klessinger. M. (1995). Cltentistry Errr. J. (in press).
Freilich. S.C.. Peters. K.S. (1985). J. Ant. Cltc>r~t. Soc.. 107. 3819.
Dubois. J.T.. van Hemert. R.L. (1964). J. Cltcarli. Phys. 40. 923
Freund. L.. Klessinger, M. (1995). to be published.
Dubois. J.T.. Wilkinson. F. (1963). J. Chc.111. Pltys. 39. 899.
Friedrich. L.E.. Schuster. G.B. ( 1969). J. Ant. Cht~nt.Sot.. 91. 7204.
Dunning, Jr., T.H.. Hosteny, R.P., Shavitt. I. (1973). J. AIU. C'lte~ni.Soc.. 95, 5067.
t . 94. 1193. Friedrich. L.E.. Schuster. G.B. (1972). J. Am. C h e ~ ~Sot..
Dvorak. V.. Michl. J. (1976). J. Ant. Cliant. Soc.. 98. 1080.
Friedrich. J.. Metz. F.. Dorr. F. (1974). Bcv. B~r~i.\enge~s. Pltys. Cltc>nt.78. 1214.
Eaton, P.E., Cole. Jr., T.W. (1964). J. Ant. Chc.111. Soc.. 86, 3158.
Fritsche. J. (1867). J. Prtrkt. Ckern. [I] 101. 333.
Eberson, L. (1982). Ad,.. Pltys. Org. Clicm. 18. 79.
Frolich. W.. Dewey. H.J.. Deger. H.. Dick. B.. Klingensmith. K.A.. Piittmann. W., Vogel, E.. Hohlneicher. G.. Michl. J. ( 1983). J. Ant. Cheln. Soc.. 105. 62 1 1.
64. 356. Eckert. R.. Kuhn H. (1960). Z. Elektroc.lte~~t.
. Rettschnick. R.P.H.. Hoytink. G.J. (1969). Cltem. Pltys. Lett. 4. 59.
El-Sayed, M.A. (1963). J . Clte~n.Pltys. 38. 2834.
Geldof. P.A.
Engel. P.S. (1980). Chem. Rel,. 80. 99.
Gerhartz. W.. Poshusta, R.D., Michl. J. (1976). J. Am. Chern. Soc.. 98. 6427.
Engel. P.S., Keys. D.E., Kitamura. A. (1985). J. Ant. Cltem. Soc.. 107, 4%4.
Gerhartz. W.. Poshusta. R.D.. Michl. J. (1977). J. Am. Chem. Soc.. 99. 4263.
Engelke. R., Hay, P.J., Kleier, D.A., Wadt, W.R. (19841, J. Ant. Cltc~nt.Soc. 106, 5439.
Gersdorf. J.. Mattay. J.. Giirner. H. (1987). J. Am. Cltem. Soc.. 109. 1203.
Englman. R.. Jortner. J . (1970). Mol. Phvs. 18. 145.
Gilbert. A. (1980). Prrre Appl. Cltem. 52. 2669.
REFERENCES
REFERENCES
Gisin. M.. Wirz. J. (1976). Hell*. Cltirn. Ac.tcr 59. 2273.
Havinga. E. ( 1973). Experic,iiticr 29, 1 181.
Gleiter. R. (1992). Angew. Cltem. Int. Eel. EngI. 31. 27.
Havinga, E., Cornelisse, J. (19761, Pure Appl. Chem. 47, I.
Gleiter. R.. Karcher, M. (1988). Angot.. Cltern. lnt. Ed. Eirgl. 27. 840.
Hayes. J.M. (1987). Chem. Rev. 87. 745.
Gleiter. R., Sander. W. (1985). Ange~v.Cheiti. Int. Ed. G i g l . 24. 566.
Heilbronner. E. ( 1963). 7i~tt~trliedrori 19. Suppl. 2.. 289.
Gleiter. R.. Schafer, W. (1990). Acac.. Clrem. Rrs. 23. 369.
Heilbronner. E. (1966). in 0pti.sc.ltc~Aitrcyrrrt~orgcriii.sc.lter Systenic.; Foerst. W., Ed.; Verlag Chemie: Weinheim.
Gleiter. R.. Schang, P.. Bloch. M.. Heilbronner. E.. Bunzli. J.-C.. Frost. D.C.. Weiler. L. (1985). Client. Bc~r.118. 2127. Godfrey. M.. Murrell. J.N. (1964). Proc.. R. Soc.. Loncloii A278. 57.
Heilbronner. E.. Bock. H. ( 1968). Dcrs HMO-Modell rrrid seirtc. An~~~ertclrrrtg; Verlag Chemie: Weinheim.
Goldbeck. R.A. ( 1988). Acec.. Cltern. Res. 21,95.
Heilbronner. E.. Murrell. J.N. ( 1963). Mol. Pliys. 6. 1.
Gould. I.R., Moser. J.E.. Ege. D.. Farid. S. ( 1988). J. Ant. Cherti. Soc.. 110. 1991.
Henderson. Jr.. W. A.. Ullnian. E. F. ( 1965). J. Am. Clic~m.Soc.. 87. 5424.
Gouterman. M. ( 1961). J. Mol. Spec.trosc.. 6. 138.
Hendriks. B.M.P.. Walter. K. I..Fischer. H. ( 1979). J. Ani. Chc>rti.Soc.. 101. 2378.
Grabowski. Z.R.. Dobkowski. J. (1983). Prrro Appl. Clteni. 55. 245.
Herman. M.F. ( 1984). J. Cl!c.~ti.Phys. 81. 754.
Granville. M. F.. Holtom. G.R.. Kohler. B.E. ( 1980). J. Cher~i.Phys. 72. 4671.
Herndon. W.C.
Greene. B.I.. Farrow. R.C. (1983). J. Clteitt. Pliys. 78. 3336.
Herndon. W.C. ( 1974). For~sc.ltr.Cheni. fir.sc.lt. 46. 14 1 .
Grellmann. K.-H.. Kuhnle. W.. Weller. H.. Wolff. T. (1981). J. Ani. Clieni. Soc.. 103. 6889.
Herzberg. G. ( 1950). Molc~c~rrlcrr Spec.trtr crrrcl Molc~c~rrlcrr Strrrc.trrrc~:Van Nostrand: Princeton.
( 19711.
Tc.trrrlieclron Lett. 12. 125.
Griesbeck. A.G.. Stadtmuller. S. (I9911. J. Ani. Cheni. Soc.. 113. 6923.
Herzberg. G.. Longuet-Higgins. H.C. ( 1963). Trccris. Fcrrtrcl(ry Soc.. 35. 77.
Griffiths. J. (1972). Chem. Soc.. Rer*. 1. 481.
Herzberg. G.. Teller. E. (1947). Rev. Mod. Phys. 13. 75.
Academic Press: LonGriffiths. J. (1976). Colorrr t r r i e l Con.stitrrtion (!/'Orgernic. Molc~c~rr1e.s: don. New York.
Hilinski. E.F.. Masnovi. J.M.. Kochi. J.K.. Rentzepis. P.M. ( 1984). J. Am. Ckem. Soc.. 106. 807 1.
Grimbert, D.. Segal, G.. Devaquet, A. (1975). J. Ant. Cltc~iti.So(.. 97. 6629.
Hirata. Y.. Mataga. N. (198-1). J. Phys. Chent. 88. 309.1.
Grimme. S.. Dreeskamp. H . ( 1992). J. Pliotoc.hent. Pliotohiol. A: Cltc>rn.65. 37 1.
Hixson. S.S.. Mariano. P.S.. Zimmerman. H.E. (1973). Cheni. Re\.. 73. 531
Grinter. R.. Heilbronner. E. (1962). Hel~..Cliint. Ac.tcr 45. 2496.
Hochstrasser. R.M.. Noe. I..J. (1971). J. Mol. Spec.trosc.. 38. 175.
Gustav. K.. Suhnel. J. (1980). Z. Cltc~ni.8. 283.
Hohlneicher. G.. Dick. B. (1984). J. Photoc.hem. 27. 215.
Gustav. K.. Kempka. U.. Suhnel. J. (1980). Chc.nr. Pltys. Lett. 71. 280.
Hopf. H.. Lipka. H.. Traetleberg. M. ( 1994). Aiige~c-.Circ~nt.Int. Eel. Engl. 33. 204.
Haag, R.. Wirz. J.. Wagner. P.J. (1977). He/\,. Cliini. Ac.ttr 60. 2595.
Horspool. W.H. ( 1976). A.sl)clc.rs c!fOr,qer~iicPhotoc.henii.stry; Academic Press: London.
Halevi. E.A. (1977). Norr~..J. Chint. 1. 229.
Hosteny. R.P.. Dunning. JI-.. T.H.. Gilman. R.R.. Pipano. A.. Shavitt. I. (1975). J. Chc~iti Pltys. 62.4764.
Haller. 1. (1967). J. Chem. Pkys. 47. 1 1 17. Ham. J.S. (1953). J. Cltent. Phys. 21. 756. Hammond. G.S.. Saltiel. J. (1963). J. Am. Client. Soc.. 85. 2516. Hammond, G.S.. Saltiel, J., Lamola. A.A.. Turro. N.J., Bradshaw. J.S.. Cowan. D.O.. Counsell. R.C.. Vogt. V.. Dalton. C. ( 1%4). J. Ant. Cltcm. Soc.. 86. 3197. Hansen. A.E. ( 1967). Mol. Plivs. 13,425. Harada. N.. Nakanishi. K. (1972). Acc. Cltern. Reu. 5. 257. Harding. L.B.. Goddard 111. W.A. (1980). J. Am. Cliem. Soc.. 102.439.
Houk. K.N. (1976). Chc.~tr.Hcl*. 76. I. Houk. K.N. ( 1982). Prrrc. Al)pI. Cltern. 54. 1633. Howeler.
U.. Michl. J. (19951, to be published.
Howeler. U.. Chatterjee. P.S.. Klingensmith. K.A.. Waluk. J.. Michl. J. (1989). Prrre Appl. Chcnt. 61,21 17. Hudson. B.S.. Kohler. B.E. (1972). Chenr. Phys. Lett. 14. 299. Hudson. B.S.. Kohler. B.E. (1973). J. Client. P1ry.s. 59. 4984.
Haselbach. E.. Heilbronner. E. (1970). HCl1.. Chirn. Actcr 53.684.
Hudson. B.S.. Kohler, B.E.. Schulten. K . (1982). in E-rcitcd Stcrtes. Vol. 6; Lim, E.C.. Ed.: Academic Press: New York.
Hassoon. S.. Lustig, H.. Rubin. M.B., Speiser. S. (1984). J. Pltys. Cliern. 88,6367-
Huisgen. R. (1977). Angc#11-. Chern. Int. Ed. E ~ t g l .16.572.
Hastings. D.J., Weedon, A.C. (1991). J. Am. Cliem. Soc.. 113.8525.
Huppert. D., Rentzepis. P. M.. Tollin. G. ( 1976). Bioc.hi~rt.Biopliys. Ac.ter 440. 356.
Hautala. R.R.. Dawes. K.. Turro. N.J. (1972). 72trerhedroii Lett. 13. 1229.
Imamura. A.. Hoffmiinn. K II9hX). .IAIII. . Clrc,ltr. So(..90. 5379.
REFERENCES
REFERENCES
Innes, K.K. (1975) in Excited States, Vol. 2; Lim, E.C., Ed.; Academic Press: New York.
Kearvell, A., Wilkinson, F. (1969). 20"" Ro'union Soc. Chim. Phys., 125.
Inoue, Y., Takamuku, S., Sakurai, H. (1977). J. Phys. Chem. 81. 7.
Keller, J.S., Kash, P.W., Jensen, E., Butler, L.J. (1992),J. Chem. Phys. 96, 4324.
Inoue, Y., Yamasaki, N., Yokoyama, T., Tai, A. (1993), J. Org. Chem. 58, 1011.
Kestner, N.R., Logan, J., Jortner, J. (1974), J. Phys. Chem. 78, 2148.
Ireland, J.F., Wyatt, P.A.H. (1973), J.C.S. Faraday Trans, 1 69, 161.
Khudyakov, I.V., Serebrennikov, Y.A., Turro, N.J. (1993). Chem. Rev. 93, 537.
Ishikawa, M., Kumada, M. (1986), Adv. Organomet. Cltem. 19, 51.
Kiefer, E.F., Carlson, D.A. (1967). Tetrahedron Lett. 8, 1617.
Jacobs, H.J.C., Havinga, E. (1979), Adv. Photochem, 11, 305
Kiefer, E.F., Tanna, C.H. (1969). J. Am. Chem. Soc. 91,4478.
Jaffd, H. H., Orchin, M. ( 1962), Theory and Applications of Ultrcrviolet Spectroscopy; Wiley: New York, London.
Kim, J.-I., Schuster, G.B. (1990). J. Am. Chem. Soc. 112, 9635.
Jensen, E., Keller, J.S., Waschewsky, G.C.G., Stevens, J.E.. Gvaham. R.L., Freed, K.F., Butler. L.J. (1993). J. Chern. Phys. 98, 2882.
Kiprianov, A.I., Mikhailenko, F.A. (1961), Zh. Obshch. Khim, 31, 1334. Kirkor, E.S., Maloney, V.M., Michl, J. (1990). J. Am. Chem. Soc. 112, 148.
Job, V.A., Sethuraman, V., Innes, K.K. (1969), J. Mol. Spectrosc.. 30, 365.
Kirmaier, C., Holten, D. (1987). Photosynth. Res. 13, 225.
Johnson, D.R., Kirchhoff, W.H., Lovas, F.J. (1972). J. Phys. Cltc.m. Ref. Data 1, 1011.
Klessinger, M. (1968). Fortschr. Chem. Forsch. 9, 354.
Johnson, P.M., Albrecht, A.C. (1968). J. Chem. Phys. 48,851.
Klessinger, M. (1978). Chem. rrrts. Zeit 12, 1.
Johnston, L.S., Scaiano, J.C. (1989). Chem. Rev. 89,521.
Klessinger, M. (1995). Angew. Chem. Int. Ed. Engl. 34,549.
Johnstone, D.E., Sodeau, J.R. (1991). J. Phys. Chem. 95, 165.
Klessinger, M., Hoinka, C. (1995)-to be published.
Jones, 11, G., Becker, W.G. (1982). Chem. Phys. Lett. 85, 271.
Klessinger, M., Potter, T.. van Wullen, C. (19911, Theor. Chirn. Acvn 80, 1.
Jones, 11, G., Chiang, S.-H. (1981), Tetruhedron 37, 3397.
Klingensmith, K.A., Puttmann, W., Vogel, E., Michl, J. (1983), J. Am. Chem. Soc. 105, 3375.
Jones, L.B., Jones, V.K. (1969). Fortschr. Chem. Forsch. 13, 307. Joran, A.D., Leland, B.A., Geller, G.G., Hopfield. J.J., Dervan. P.B. (1984). J. Am. Chem. Soc. 106,6090.
Kobayashi, T., Nagakura. S. (1972). Bull Chern. Soc. Jpn. 47, 2563. Koga, N., Sameshima, K., Morokuma, K. ( 1993),J. Phys. Chem. 97, 131 17.
JBrgensen, N.H., Pedersen, P.B., Thulstrup, E.W., Michl, J. (1978). Int. J. Quuntirm Chem. Sl2, 419.
Kohler, B.E. (19931, Chent. Rev. 93, 41.
Jug, K., Bredow, T. (1991). J. Phys. Chem. 95,9242.
Kollmar, H., Staemmler, V. (1978), Theor. Chim. Acta 48, 223.
De Kanter, F.J.J., Kaptein, R. (1982). J. Am. Chem. Soc. 104.4759.
Koo, J., Schuster, G.B. (1977). J. Am. Chem. Soc. 99, 6107.
De Kanter, F.J.J., den Hollander, J.A., Huizer, A.H., Kaptein. R. (1977). Mol. Phys. 34, 857.
Kosower, E.M. (1958), J. Am. Clrem. Soc. 80, 3253, 3261, 3267.
Kaplan, L., Wilzbach, K.E. (1968, J. Am. Chem. Soc. 90, 3291.
Kouteckq, J. (1965). in Modern Quantrrm Chemistry, Part 1; Sinanoglu, 0.. Ed.; Academic Press: New York, London.
Kaprinidis, N.A.. Lem. G., Courtney, S.H., Schuster, D.I. (1993)-J. Am. Chem. Soc. 115, 3324.
Kouteckq, J. (1966), J. Clrc~n.Phys. 44, 3702.
Karelson, M., Zerner, M.C. (19901, J. Am. Chem. Soc. 112,9405. Karelson, M., Zerner, M.C. (1992), J. Phys. Chem., 96,6949. Karwowski, J. (1973). Chem. Phys. Lett. 18, 47. Kasha, M. ( 1950). Disc.. Faraday Soc. 9, 14. Kasha, M., Brabham. D.E. (1979). in Singlet Oxygen; Wasserman, H.H., Murray, R.W., Eds.; Academic Press: New York. Kato, S. (1988). J. Chem. Plrys. 88, 3045. Kaupp, G., Teufel, E. (1980), Cltem. Ber. 113, 3669. Kearns, D.R., Case, W.A. (1966). J. Am. Chem. Soc. 88,5087.
Kolc, J., Michl, J. (1976). J. Am. Chem. Soc. 98, 4540.
Kouteckq, J. (1967). J. Chem. Phys. 47, 1501. Kouteckq. J.. Citek. J.. Dubskq. J.. Hlavatq, K. (1964). Theor. Chim. Actu 2. 462. Kroon. J., Verhoeven, J.W.. Paddon-Row. M.N.. Oliver, A.M. (1991), Angew. Chem. Int. Ed. Engl. 30, 1358. Kropp, P.J. (1984). Acc. Chem. Res. 17, 131. Kropp, P.J.. Reardon. Jr.. E.J., Gaibel, Z.L.F.. Williard. K.F.. Hattaway. Jr.. J.H. (1973). J. Am. Clrem. Soc. 95, 7058. Kropp, P.J., Snyder. J.J.. Rawlings. P.C., Fravel, Jr., H.G. (1980). J. Org. Chem. 45,4471 Kubota, Y.. Motoda. Y., Shigemune, Y.. Fujisaki, Y. (1979). Photochem. Photobiol. 19. 1099.
REFERENCES Kuhn, H. (1949), J. Chem. Phys. 17, 1198.
REFERENCES Lippert, E. (19691, Ace. Clrem. Res. 3, 74.
Kuppermann, A., Flicker, W.M., Mosher, O.A. (1979), Chem. Rev. 79.77.
Lippert. E., Rettig, W., BonaCiC-Kouteckq, V., Heisel, F., Miehk, J.A. (1987), Adv. Chem. Phys. 68, 1.
Laarhoven, W.H. (1 983), Reel. Trav. Chim. Pays-Bas 102, 185.
Liptay, W. (1963). 2.Natrr~:forsch.18a, 705.
Labhart, H. ( 1%6), Experientia 22, 65.
Liptay, W. (19661, in Optische Anregung organischer Systeme; Foerst, W., Ed.; Verlag Chemie: Weinheim.
Labhart, H., Wagniere, G. (1959). Helv. Chim. Acta 62, 2219. Lakowicz, J.R. (1983), Principles of Fluroescence Spectroscopy; Plenum Press; New York, London.
Liptay, W. (1969). Angew. Chem. Int. Ed. Engl. 8, 177.
Lamola, A.A. ( 1968), Photochem. Photobiol. 8, 126.
Liptay, W., Wortmann, R.. Schaffrin, H., Burkhard, O., Reitinger, W., Detzer, N. (1988a), Chem. Phys. 120,429.
Landau, L. (1932), Phys. 2. Sowjet. 2-46.
Liptay, W., Wortmann, R.. Bohm, R., Detzer, N. (1988b), Chem. Phys. 120,439.
Langkilde, F.W., Thulstrup, E.W., Michl, J. (1983a), J. Chem. Phys. 78, 3372.
Liu, R.S.H., Hammond, G.S. (1967), J. Am. Chem. Soc. 89,4936.
Langkilde, F.W., Gisin, M., Thulstrup, E.W., Michl, J. (1983b). J. Phys. Chetn. 87, 2901.
Liu, R.S.H., Turro, N.J., klammond, G.S. (1%5), J. Am. Chem. Soc. 87, 3406.
Laposa, J.D., Lim, E.C., Kellogg, R.E. (1969, J. Chem. Phys. 42, 3025.
Longuet-Higgins, J.C., Abrahamson, E.W. (1965)-J. Am. Chem. Soc. 87,2045.
Lauer, G., Schafer, W., Schweig, A. (1975). Chem. Phys. Lett. 33, 312.
Longuet-Higgins, H.C., Murrell, J.N. (1955). Proc. Phys. Soc. A 68, 601.
Lechtken, P., Hohne, G. (1973). Angew. Chem. Int. Ed. Engl. 12, 772. Lechtken, P., Breslow, R., Schmidt, A.H., Turro, N.J. (1973). J. Am. Chem. Soc. 95,3025.
Lorquet. J.C., Lorquet, A.J., Desouter-Lecomte, M. (1981), in Quantum Theory of Chemical Reactions. Vol. 11; Daudel, R., Pullman, A., Salem, L., Veillard, A., Eds; Reidel: Dordrecht; p. 241.
Lee, Y.S., Freed, K.F., Sun, H., Yeager, D.L. (1983), J. Chem. Phys. 79, 3862.
LOttke, W., Schabacker, V. ( 1966), Liebigs Ann. Chem. 698.86.
Leigh, W.J. (1993). Can. J. Chem. 71, 147.
Luzhkov, V., Warshel, A. ( 19911, J. Am. Chem. Soc. 113, 4491.
Leigh, W.J., Zheng, K. (1991),J. Am. Chem. Soc. 113,4019; Errata, ibid. 114 (1992) 7%.
Maciejewski, A., Steer, R.1). (1993). Chem. Rev. 93, 67.
Leigh, W.J., Zheng, K., Nguyen, N., Werstink, N.H., Ma, J. (1991)- J. Am. Chem. Soc. 113,4993.
Maier, G. (1986)- Pure Appl. Chem. 58,95.
Letsinger, R.L., McCain, J.H. (1969), J. Am. Chem. Soc.. 91, 6425.
Maier, G., Pfriem, S., Sch2fer, U., Malsch, K.-D., Matusch, R. (1981). Chem. Ber. 114, 3%5.
Levy, R.M., Kitchen, D.B., Blair, J.T., Krogh-Jespersen, K. (1990). J. Phys. Chem. 94, 4470.
Maier, J.P., Seilmeier, A., Laubereau, A., Kaiser, W. (1977). Chem. Phys. Lett. 46, 527.
Lewis, F.D., DeVoe, R.J. (1982), Tetrahedron 38, 1069. Lewis, F.D., Hilliard, T.A. (1972). J. Am. Chem. Soc. 94, 3852. Lewis, F.D., Magyar, J.G. (1972), J. Org. Chem. 37,2102. Lewis, ED., Ho, T.-I., Simpson, J.T. (1981), J. Org. Chem. 46, 1077.
Majewski, W.A., Plusquellic, D.F., Pratt, D.W. (1989). J. Chem, Phys. 90, 1362. Malhotra, S.S., Whiting, M.C. (19591, J . Chem. Soc. (London), 3812. Malmqvist, P.-A., Roos, B.O. (1992). Theor. Chim. Actu 83, 191. Manning, T.D.R., Kropp, P.J. (1981), J. Am. Chem. Soc. 103, 889.
Lewis, ED., Ho, T.-I., Simpson, J.T. (1982). J. Am. Chem. Soc. 104, 1924.
Manring, L.E., Peters, K.S., Jones, 11, G., Bergmark, W.R. (1985),J. Am. Chem. Soc. 107, 1485.
Li, R., Lim, E.C. (1972), J. Chem. Phys. 57,605.
Manthe, U., Ktbppel, H. (1990), J. Chem. Phys. 93, 1658.
Libman, J. (1975). J. Am. Chem. Soc. 97,4139.
Marchetti, A.P., Kearns, D.R. (1967), J. Am. Chem. Soc. 89, 768.
Lin, S.H., Fujimura, Y., Neusser, H.J., Schlag, E. W. (1984). Mitltiphoton Spectroscopy of Molecules; Academic Press: New York.
Marcus, R.A. (1964), Annu. Rev. Phys. Chem. 15, 155. Marcus, R. A. ( 1988), Israel J. Chem. 28, 205.
Linder, R.E., Morrill, K., Dixon, J.S., Barth, G., Bunnenberg, E., Djerassi, C., Seamans, L., Moscowitz, A. (1977), J. Am. Chem. Soc. 99,727.
Maroulis, A.J., Arnold, D.R. (19791, Synthesis, 819.
Linderberg, J. (1967). Chem. Phys. Lett. 1, 39.
Martin, H.-D., Schiwek, H.-J., Spanget-Larsen, J., Gleiter, R. (1978), Chem. Ber. 111,2557.
Linderberg, J., Michl, J. (1970), J. Am. Chem. Soc. 92, 2619.
Martin, R.L., Wadt, W.R. (19821, J. Phys. Chem. 86, 2382.
Lippert, E. (1966), in Optische Anregung organischer Systeme; Foerst, W., Ed.; Verlag Chemie: Weinheim.
Masnovi, J.M., Kochi, J.K.. Hilinski, E.F., Rentzepis, P.M. (1986),J. Am. Chem. Soc. 108, 1126.
REFERENCES
1 I
507 .
Mason, S.F. (1962). J. Chem. Soc. (London), 493.
Michl, .I. (19921, J. Mol. Stritct. (Theochem)260, 299.
Mason, S.F. (1981), Advances in IR and Raman Spectroscopy. Vol. 8, Chapter 5; Clark, R.J .H., Hester, R.E., Eds. ;Heyden: London.
Michl, J., Balaji, V. (1991). in Computational Advances in Organic Chemistry: Molecular Structure and Reactivity; ogretir, C., Csizmadia, I.G., Eds. ; Kluwer: Dordrecht.
Mason, S.F., Philp, J., Smith, B.E. (1968), J. Chem. Soc. (London), A 3051. Mataga, N. (1984). Pure Appl. Chem. 56, 1255.
Michl, J.. BonaCiC-Kouteckq, V. (1990). Electronic Aspects of Organic Photochemistry; Wiley: New York.
Mataga. N.. Nishimoto. K. (1957). Z. Phys. Chem. N.F. 13, 140.
Michl, J., Kolc, J. (1970)- J. Am. Chem. Soc. 92, 4148.
Mataga, N.. Karen, A.. Tadashi. 0.. Nish*tani, S., Kurata. N.. Sakata. Y.. Misumi. S. (1984). J. Phys. Chem. 88, 5138.
Michl, J., Thulstrup, E.W. (1976). Tetrahedron 32, 205.
Mattay. J. (1987). J. Photochem. 37. 167.
Michl, J. Thulstrup, E.W. (1986), Spectroscopy with Polarized Light; VCH Publishers. : Deerfield Beach.
Mattay, J., Gersdorf, J., Freudenberg, U. (1984a). Tetrahedron Lett. 25, 817.
Michl, J., West, R. (1980)- in Oxocarbons; West, R., Ed.; Academic Press: New York.
Mattay. J.. Gersdorf. 1.. Leismann, H., Steenken, S. (1984b). Anpew. Chem. lnt. Ed. Engl. 23, 249.
Miller, R.D., Michl, J. (1989), Chem. Rev. 89, 1359. Moffitt, W. (1954a). J. Chem. Phys. 22, 320.
Mattes, S.L., Farid. S. (1983), J. Am. Chem. Soc. 105, 1386.
Moffitt, W. (1954b). J. Chem. Phys. 22, 1820.
McDiarmid, R. (1976). J. Chem. Phys. 64,514.
Moffitt, W., Moscowitz, A. (1959). J. Chem. Phys. 30, 648.
McDiarmid, R., Doering, J.P. (1980). J. Chem. Phys. 73,4192.
Moffitt, W., Woodward, R.B., Moscowitz, A., Klyne, W., Djerassi, C. (1961), J. Am. Chem. SOC.83,4013.
McGl ynn. S.P., Azumi, T., Kinosh*ta, M. (1%9). Molecular Spectroscopy of the Triplet State: Prentice-Hall: Englewood Cliffs.
Momicchioli, F., Baraldi. I., Berthier, G. (19881, Chem. Phys. 123. 103.
McLachlan, A.D. (1959). Mol. Plrys. 2, 271.
Moore, W.M.. Morgan, D.D., Stermitz, F.R. (1963). J. Am. Chem. Soc. 85, 829.
McWeeny, R. (1989). Methods of Molecular Quantum Mechattic.~:Academic Press: London.
Mosher, O.A.. Flicker, W.M., Kuppermann, A. (1973). J. Chem. Phys. 59, 6502.
Meinwald, J., Samuelson, G.E., Ikeda, M. (1970), J. Am. Chent. Soc. 92, 7604. Melander, L.C.S., Saunders, Jr., W.H. (1980). Reaction Rates oj.lsotopic Molecules; Wiley: New York.
Moule, D.C., Walsh, A.D. (1975), Chem. Rev. 75.67. Mulder, J.J.C. ( 1980). Noirv. J. Chim. 4, 283. Mulliken, R.S. (1939). J. Chem. Phys. 7, 20.
Merer, A.J., Mulliken, R.S. (1969). Chem. Rev. 69. 639.
Mulliken, R.S. (1952), J. Am. Chem. Soc. 74, 81 1.
Meth-Cohn, O., Moore, C., van Rooyen, P.H. (1985). J.C.S. Perkin Trans. 1. 1793.
Mulliken, R.S. (1977), J. Chem. Phys. 66, 2448.
Michl. J. (1972). Mol. Photochem. 4, 243,257,287.
Murray, R.W. (1979). in Singlet Oxygen. Chapter 3; Wassermann, H.H.. Murray. R.W., Eds.; Academic Press: New York.
Michl, J. (1973). in Physical Chemistry. Vol. VII; Eyring. H.. Henderson. D.. Jost. W., Eds.; Academic Press: New York. Michl, J. (1974a). Top. Curr. Chem. 46, 1.
Murrell, J.N. (1963), The Theory of the Electronic Spectra of Organic Molecules; Methuen: London. Murrell, J.N., Tanaka, J. (1%3), Mol. Phys. 7, 363.
Michl. J. (1974b), in Chemical Reactivity and Reaction Paths. Chapter 8; Klopman, G.,Ed.; Wiley: New York.
Muszkat, K.A. (1980). Top. Curr. Chem. 88, 91.
Michl, J. (1974~).J. Chem. Phys. 61,4270.
Nayler, P., Whiting, M.C. (1955). J. Chem. Soc. (London), 3037.
Michl, J. (1977). Photochem. Plt~tobiol.25, 141.
Neurnann, J. von, Wigner, E. (1929), Z. Phys. 30,467.
Michl, J. (1978), J. Am. Chem. Soc. 100,6801,6812,6819.
Neusser, H.J., Schlag, E.W. (1992), Angew. Chem. Int. Ed. Engl. 31. 263.
Michl, J. (1984). Tetrahedron 40,3845.
Nickel, B., Roden, G. (19821, Chem. Phys. 66, 365.
Michl, J. (1988). Tetrahedron 44, 7559.
Niephaus, H., Schleker, W., Fleischhauer, J. (19851,Z. Naturforsch. 40a, 1304.
Michl, J. (1990), Acc. Chern. Res. 23, 127.
Ohara, M., Watanabe, K. (1975). Angew. Chem. Int. Ed. Engl. 14, 820.
Michl, J. (1991). in Theoretical and Computational Models for Organic Chemistry; Formosinho, S.J. Csizmadia, I.G., Arnaut, L.G., Eds.; Kluwer: Dordrecht.
Ohmine, I. (1985). J. Chent. Phys. 83, 2348. Olivucci, M.. Ragazos. I.N.. Bernardi. F., Robb. M.A. (1993). 1.Am. Chem. Soc. LI5.3710.
508
REFERENCES
REFERENCES
Olivucci, M., Bernardi, F., Celani, P., Ragazos, I.N., Robb, M.A. (1994a1, J. Am. Chem. SOC. 116, 1077.
Prinzbach, H., Fischer, G., Rihs, G., Sedelmeier, G., Heilbronner, E., Yang, Z. (1982), Tetrahedron Lett. 23, 125 1 .
Olivucci, M., Bernardi, F., Ottani, S., Robb, M.A. (1994b), J. Am. Chrm. Soc. 116, 2034.
Quenemoen, K., Borden, W.T., Davidson, E.R., Feller, D. (1985). J. Am. Chem. Soc. 107, 5054.
Onsager, L. (1936), J. Am. Chem. Soc. 58, 1486. Orlandi, G., Siebrand, W. (1975), Chem. Phys. Lett. 30, 352. Padwa, A., Griffin, G. (1976), in Photochemistry of Heterocyclic Compounds; Burchardt, 0.; Ed.; Wiley: New York.
Quinkert, G., Finke, M., Palmowski, J., Wiersdorff, W.-W. (1969). Mol. Photochem. 1,433. Quinkert, G., Cech, F., Kleiner, E., Rehm, D. (1979), Angew. Chem. Int. Ed. Engl. 18,557.
Paetzold, R., Reichenbacher, M., Appenroth, K. (1981), Z . Chem. 21,421.
Quinkert, G., Kleiner, E., Freitag, B.-J., Glenneberg, J., Billhardt, U.-M., Cech, F., Schmieder, K.R., Schudok, C., Steinmetzer, H.-C., Bats, J.W., Zimmermann, G., Diirner, G., Rehm, D., Paulus, E. F. ( 1986), Helv. Chim. Acta 69,469.
Paldus, J. (1976), in Theoreticul Chemistry, Advances and Perspectives. Vol. 2; Eyring, H., Henderson, D., Eds.; Academic Press: New York.
Rabek, J.F. (19821, Experimc.nta1 Methods in Photochemistry and Photophysics, Parts I and 2; Wiley: Chichester.
Palmer, I.J., Olivucci, M., Bernardi, F., Robb, M.A. (1992), J. Org. Chem. 57, 5081.
Rademacher, P. (19871, Strilkturen organischer Molekiile; Verlag Chemie: Weinheim.
Palmer, I.J., Ragazos, I.N., Bernardi, F., Olivucci, M., Robb, M.A. (1993), J . Am. Chem. SOC. 115, 673.
Radziszewski, J.G., Burkhalter, F.A., Michl, J. (1987), J. Am. Chem. Soc. 109,61. Radziszewski, J.G., Waluk, J . , Michl, J. (1989)- Chem. Phys. 136, 165.
Palmer, I.J., Ragazos, I.N., Bernardi, F., Olivucci, M., Robb, M.A. (1994). J. Am. Chem. SOC. 116, 2121.
Radziszewski, J.G., Waluk, J., Nepras, M., Michl, J. (1991), J. Phys. Chem. 95, 1963.
Pancir, J., Zahradnik, R. (1973). J. Phys. Chem. 77, 107.
Ragazos, I.N., Robb, M.A., Bernardi, F., Olivucci, M. (1992), Chem. Phys. Lett. 197, 217.
Pape, M. (1967). Fortschr. Chem. Forsch. 7, 559.
Ramesh, V., Ramamurthy, V. (1984), J. Photochem. 24, 395.
Paquette, L.A., Bay, E. (1982), J. Org. Chem. 47,4597.
Ramunni, G., Salem, L. (1976). Z. Phys. Chem. N . F . 101, 123.
Pariser, R. (1956). J. Chem. Phys. 24, 250.
Ransom, B.D., Innes, K.K., McDiarmid, R. (1978), J. Chem. Phys. 68,2007.
Pariser, R., Parr, R.G. (1953), J. Chem. Phys. 21,466.
Rau, H. (1984), J. Photochetn. 26, 221.
Parker, C.A. (1964), Adv. Photochem. 2, 306.
Reedich, D.E., Sheridan, R.S. (1985), J. Am. Chem. Soc. 107, 3360.
Pasman, P., Mes, G.F., Koper, N.W., Verhoeven, J.W. (198% J. Am. Chem. Soc. 107,5839.
Reguero, M., Bernardi, F., Jones, H., Olivucci, M., Ragazos, I.N., Robb, M.A. (1993). J. Am. Chem. Soc. 115, 2073.
Petek, H., Bell, A.J., Christensen, R.L., Yoshihara, K. (1992), J. Chem. Phys. 96, 2412. Peters, K.S., Pang, E., Rudzki, J. (1982). J. Am. Chem. Soc. 104, 5535. Peters, K.S., Li, B., Lee, J. (1993), J. Am. Chem. Soc. 115, 11 119. Peyerimhoff, S.D., Buenker, F.J. (1973). NATO Adv. Study Inst. Ser. C 8.257. Peyerimhoff, S.D., Buenker, F.J. (1975), Adv. Quantum Chem. 9.69. Philips, D., Salisbury, K. (1976), in Spectroscopy, Vol. 3; Straughan, B.P., Walker, S., Eds.; Chapman and Hall: London. Pichko, V.A., Simkin, B.Ya., Minkin, V.1. (1991). J. Mol. Strucf. (Theochem) 235, 107.
~ e ~ u e rM., o , Olivucci, M., I)ernardi, F., Robb, M.A. (1994). J. Am. Chem. Soc. 116,2 103. Rehm, D., Weller, A. (1970a). Z. Phys. Chem. 69.83. Rehm, D., Weller, A. (1970b). Isrue1 J. Chem. 8 , 259. Reichardt, C., Dimroth, K. (1968), Fortschr. Chem. Forsch. 11, 1. Reichardt, C., Harbusch-Gornert, E. (1983), Liebigs Ann. Chem., 721. Reid, P.J., Doig, S.J., Mathies, R.A. (1990), J. Phys. Chem. 94, 8396. Reid, P.J., Wickham, S.D., Mathies, R.A. (1992). J. Phys. Chem. 96, 5720.
Pickard, S.T., Smith, H.E. (1990), J. Am. Chem. Soc. 112,5741.
Reid, P.J., Doig, S.J., Wickham, S.D., Mathies, R.A. (1993),J. Am. Chem. Soc. 115,4754.
Platt, J.R. (1949), J. Chem. Phys. 17,484.
Reinsch, M., Klessinger, M. (1990). Phys. Org. Chem. 3,81.
Platt, J.R. (1%2), J. Mol. Spectrosc. 9, 288.
Reinsch, M., Howeler, U., Klessinger, M. (1987), Angew. Chem. Int. Ed. Engl. 26, 238.
Platz, M.S., Berson, J.A. (1977), J. Am. Chem. Soc. 99, 5178.
Reinsch, M., Howeler, U., Klessinger, M. (1988), J. M o l . Struct. (Theochem) 167, 301.
Pople, J.A. (1953), Trans. Faraday Soc. 49, 1375. Pople, J.A. (1955). Proc. Phys. Soc. A 68,8 1.
Renner, C.A., Katz, T.J., Pouliquen, J., Turro, N.J., Waddell, W.H. (1975), J. Am. Chem. SOC. 97, 2568.
Porter, G., Strachan, W. (1958), Spectrochim. Acta 12, 299.
Rentzepis, P.M. (1970), Scietlce 169, 239.
Porter, G., Suppan, P. (1964), Pure Appl. Chem. 9,499.
Rettig, W. (1986). Angew. Clrem. Int. Ed. Engl. 25, 971.
Rettig, W., Majenz, W., Lapouyade. R., Haucke. G. (1992). J. Photoclrem. Photobiol. A : Chem. 62,415. Ridley, J., Zerner, M. (1973). Theor. Chim. Acta 32, 111. Ridley, J., Zerner, M. (1976), Theor. Chim. Acta 42,223. Riedle, E., Weber, T., Schubert, E., Neusser, H.J.. Schlag, E.W. (1990). J. Chem. P h F . 93, 967. Roos, B.O., Taylor, R.P., Siegbahn, P.E.M. (1980). Chem. Phvs. 48, 157. Rosenfeld (1928). Z. Phys. 52, 161. Roth, H.D., Schilling, M.L.M. (1980). J. Am. Chem. Soc. 102. 4303. Roth. H.D., Schilling, M.L.M. (1981). J. Am. Chem. Soc. 103, 7210. Roth, H.D., Schilling, M.L.M.,Jones, 11, G. (1981). J. Am. Ch1.m. Soc. 103, 1246. Roth, H.D., Schilling, M.L.M., Raghavachari, K. (1984). J. Anl. Chem. Soc. 106,253. Roth, W.R. (1963). Angew. Chem. 75.921.
Schaffner, K., Demuth, M. (1986), in Modern Synthetic Methods, Vol. 4; Scheffold, R., Ed.; Springer: Berlin. Heidelberg. Scharf. H.-D. (1974). Angew. Chem. Int. Ed. Engl. 13, 520. Schenck, G.O., Steinmetz, R. (1962). Brrll. Soc. Chim. Belg. 71, 781. Schexnayder, M.A., Engel, P.S. (1975), J. Am. Chem. Soc. 97,4825. Schmidt, W. (1971). Helv. Ckitn. Acta 54, 862. Scholz, M.. Kohler, H .-J. ( 1981), Quanrenchemische Naherungsverfahre~tund ihre Anwendrtng in der orgcrnischen Chentie; Hiithig: Heidelberg. Schreiber, S.L., Hoveyda, A.H., Wu, H.-J. (1983). J. Am. Chem. Sou. 105, 660. Schulten, K., Ohmine, I., Karplus, M. (1976a). J. Chem. Pltys. 64, 4422. Schulten, K., Staerk, H., Weller, A., Werner, H.-J., Nickel, B. (1976b), Z. Phys. Chem. N.F. 101, 371. Schultz, A.G. (1983), Acc. Cltem. Res. 16, 210.
Roth, W.R.. Erker, G. (1973). Angew. Chem. Int. Ed. EngI. 12, 503.
Schuster, D.I. (1989). in The Chemistry of Enottes; Patai. S., Rappoport. Z.. Eds.; Wiley: New York.
Sadler, D.E., Wendler, J., Olbrich, G., Schaffner, K. (1984). J. Am. Cltem. Soc. 106, 2064.
~. In?. Ed. EngI. 30, 1345. Schuster, D.I., Heibel, G.E., Woning, J. (199la). A n g e ~ Chem.
Salem, L. (1974). J. Am. Cltem. Soc. 96, 3486.
Schuster. D.I. Dunn, D.A.. Heibel. G.E., Brown, P.B.. Rao, J.M., Woning, J., Bonneau, R. (1991b). J. Am. Chem. Soc. 113. 6245.
Salem, L. (1982). Electrons in Chentical Reactions: First Princ.iples: Wiley: New York. Salem, L., Rowland, C. (1972), Angew. Chem. I n t . Ed. Engl. 11.92. Salem, L., Leforestier, C., Segal, G., Wetmore, R. (1975). J. Am. Chem. Soc. 97,479. Salikhov, K.M., Molin, Y.N., Sagdeev, R.Z., Buchachenko. A.L. (1984). Magnetic and Spin Effects in Chemical Reactions; Elsevier: Amsterdam. Saltiel, J., Charlton. J.L. (1980). in Rearrangements in Ground crnd Excited States, Vol. 3; de Mayo, P., Ed.; Academic Press: New York.
Schwartz, S.E. (1973). J. Chem. Educ. 50, 608. Scott. A.I. (1964). Interprettrtion ofthe Ultruviolet Spectra r?fNaturirl Products; Pergamon Press: Oxford. Seamans, L., Moscowitz, A.. Linder. R.E., Morill. K., Dixon, J.S., Barth, G., Bunnenberg, E., Djerassi, C. (1977), J. Am. Chem. Soc. 99, 724. Sevin, A., Bigot, B.. Pfau, M. (1979), Helv. Chim. Acta 62, 699.
Saltiel, J., Zafiriou, O.C., Megarity, E.D.. Lamola, A.A. (1968). J. Am. Chem. Soc. 90, 4759.
Serrano-AndrCs, L.. Merchan. M.. Nebot-Gil, I., Lindh. R., Roos, B.0. (1993), J . Chem. Phys. 98, 3151.
Saltiel, J., Metts, S., Wrighton, M. (1969). J. Am. Cltem. Soc.. 01, 5684.
Sharafy, S., Muszkat, K.A. (19711, J. Ant. Chem. Soc. 93, 41 19.
Saltiel, J., Metts, L., Wrighton, M. (1970). J. Am. Chem. Soc. 92, 3227.
Siebrand, W. ( 1966). J. Chern. Pltys. 44, 4055.
Saltiel, J., D'Agostino, J., Megarity, E.D., Metts, L., Neubcrger, K.R.. Wrighton, M., Zafiriou, O.C. ( 1973), Org. Photochent. 3, 1.
Siebrand, W. (19671, J. Ckent. Pltys. 46. 440, 241 1.
Saltiel, J., Marchand, G.R., Smothers, W.K., Stout. S.A.. Charlton. J.L. (1981). J. Am. Ckem. Soc. 103, 7159.
Snatzke, G. (1982). Chem. Zeit 16, 160.
Sklar. A.L. (1942), J. Chern. Phys. 10. 135.
Snyder, G.J., Dougherty, D.A. (1985), J. Am. Chem. Soc. 107, 1774.
Saltiel, J., Sun, Y.-R. (1990), in Photochromism. Molecules an11 Systems; Dorr, H., BouasLaurent, H., Eds.; Elsevier: Amsterdam.
Sobolewski, A.L., Woywod, C., Domcke, W. (1993). J. Chem. Phys. 98,5627.
Sandros, K., Biickstrom, H.L.J. (1%2), Acta Chem. Scand. 16. 958.
Sonntag, F.J., Srinivasan, R. (19711, Org. Photochem. Synth. 1, 39.
Santiago, C., Houk, K.N. (1976). J. Am. Chem. Soc. 98, 3380.
Souto, M.A., Wallace, S.L.. Michl, J. (1980), Tetrahedron 36, 1521.
Scaiano, J.C. (1982), Tetrahedron 38,8 19.
Squillacote. M., Semple, T.C. (19901, J. Ant. Chem. Soc. 112, 5546.
Schaap, A.P.. Thayer. A.L.. Blossey, E.C., Neckers. D.C. (1975). J. Am. Chem. Soc. 97, 3741.
Squillacote. M.E.. Sheridan. R.S., Chapman, O.L.. Anet, F.A.L. (1979). J. Atn. Chem. Soc. 101. 3657.
Schafer, 0..Allan. M., Szeimies, G., Sanktjohanser, M. (1992). J. Am. Chem. Soc. 114, 8 180.
Squillacote. M.. BeWnan. A.. De Felippis. J. (1989), Tetruhedrrpn Lett. 30, 6805.
REFERENCES Srinivasan, R. (1%3), J. Am. Chem. Soc. 85, 819.
Trulson, M.O., Mathies, R.A. ( 1990). J. Phys. Chem. 94, 5741.
Srinivasan, R. (1968). J. Am. Chem. Soc. 90,4498.
Trulson, M.O., Dollinger, G.D., Mathies, R.A. (1989), J. Chem. Phys. 90,4274.
Srinivasan, R., Carlough, K.H. (1%7), J. Am. Chem. Soc. 89,4932.
Tseng, K.L., Michl, J. (1976). J. Am. Chem. Soc. 98,6138.
Staab, H.A., Herz, C.P., Krieger, C., Rentea, M. (1983), Chem. Ber. 116, 3813.
Turner, D.W., Baker, C., Baker, A.D., Brundle, C.R. (1970), Molecular Photoelectron Spectroscopy; Wiley-Interscience: London, New York.
Steiner, R.P., Michl, J. (197% J. Am. Chem. Soc. 100,6861. Steiner, U.E., Ulrich, T. (1989), Chem. Rev. 89, 51. Stephenson, L.M., Grdina, M.J., Orfanopoulos, M. (1980), Acc. Chem. Res. 13,419. Stern, O., Volmer, M. (1919), 2.Phys. 20, 183. Steuhl, H.-M., Klessinger, M. (1994), Angew. Chem. Int. Ed. Engl. 33, 2431. Stevens, B., Algar, B.E. (1967), Chem. Phys. Lett. 1, 219. Stevens, B., Ban, M. I. ( 1964). Trans. Furaday Soc. 60, 1515. Stohrer, W.-D., Jacobs, P., Kaiser, K.H., Wiech, G., Quinkert, G. (1974). Top. Curr. Chem. 46, 181. Strickler, S.J., Berg, R.A. (1962). J. Chem. Phys. 37, 814. Struve, W.S., Rentzepis, P.M., Jortner, J. (1973), J. Chem. Phys. 59, 5014. Sugihara, Y., Wakabayashi, S., Murata, I., Jinguji, M., Nakazawa, T., Persy, G.. Wirz, J. (1985). J. Am. Chem. Soc. 107,5894. Sun, Y.-P., Wallraff, G.M., Miller, R.D., Michl, J. (1992a). J. Photuchem. Photobiol. 62, 333.
Turro, N.J. (1978). Modern Alolecular Photochemistry; Benjamin: Menlo Park.
lbrro, N.J., Devaquet, A. (1975), J. Am. Chem. Soc. 97,3859. Turro, N.J., Kraeutler, B. (IY78), J. Am. Chem. Soc. 100, 7432. Turro, N.J., Lechtken, P. (1972). J. Am. Chem. Soc. 94, 2886. Turro, N.J., Wriede, P.A. (1970). J. Am. Chem. Soc. 92, 320. Turro, N.J., Dalton, J.C., Dawes, K., Farrington, G., Hautala, R., Morton, D., Niemczyk, M., Schore, N. (1972a), Ac-c. Chem. Res. 5,92. Turro, N.J., Kavarnos, G., Fung, V., Lyons, Jr., A.L., Cole, Jr., T. (1972b),J. Am. Chem. SOC. 94, 1392. Turro, N.J., Schuster, G., Pouliquen, J., Pettit, R.. Mauldin, C. (1974). J. Am. Chem. Soc. 96,6797. Turro, N.J., Farneth, W.E., Devaquet, A. (1976). J. Am. Chem. Soc. 98, 7425. Turro, N.J., Cherry, W.R., Mirbach, M.F., Mirbach, M.J. (1977a), J. Am. Chem. Soc. 99, 7388. Turro, N.J., Ramamurthy, V., Katz, T.J. (1977b). Nouv. J. Chim. 1, 363.
Sun, Y.-P., Hamada, Y., Huang, L.-M., Maxka, J., Hsiao, J.-S., West, R., Michl, J. (1992b), J. Am. Chem. Soc. 114,6301.
Turro, N.J., Liu, K.-C., Show, M.-F., Lee, P. (1978). Photochem. Photobiol. 27, 523.
Sun, S., Saigusa, H., Lim, E.C. (1993). J. Phys. Chem. 97, 1 1635.
Turro, N.J., Aikawa, M., Butcher, Jr., J.A., Griffin, G.W. (1980a). J. Am. Chem. Soc. 102, 5127.
Tapia, 0. (1982),in Molecular Interactions, Vol. 3; Orville-Thomas, J.W., Ed.; Wiley: New York. Taube, H . ( 1970), Electron Transfer Reacrions of Complex Ions in Solution; Academic Press: New York, London. Taylor, G.N. (197I), Chem. Phys. Lett. 10, 355. Taylor, P.R. (1982), J. Am. Chem. Soc. 104,5248. Teller, E. (1937). J. Phys. Chem. 41, 109.
Turro, N.J., Anderson, D.R., Kraeutler, B. (1980b), Tetruhedron Lett. 21, 3. Turro, N.J., Cox, G.S., Paczkowski, M.A. (1985). Top. Curr. Chem. 129, 57. Turro, N.J., Zhang, Z., Trahanovsky, W.S., Chou, C.-H. ( 19881, Tetruhedron Lett. 29,2543. Vala, Jr., M.T., Haebig, J., Rice, S.A. (1965), J. Chem. Phys. 43, 886. van der Auweraer, M., Grabowski, Z.R., Rettig, W. (1991) J. Phys. Chem. 95,2083.
Terenin, A., Ermolaev, V. (1956), Truns. Furuduy Soc. 52, 1042.
van der Hart, J.A., Mulder, J.J.C., Cornelisse, J. (IY87), J. Mol. Struct. (Theochem) 151, 1.
Thiel, W. (1981). J. Am. Chem. Soc. 103, 1413.
van der Lugt, W.T.A.M., Oosterhoff, L.J. (1969), J. Am. Chem. Soc. 91, 6042.
Thompson, A.M., Goswami, P.C., Zimmerman, G.L. (1979). J. Phys. Chem. 83,314.
van Riel, H.C.H.A., Lodder. G., Havinga, E. (1981), J. Am. Chem. Soc. 103,7257.
Thompson, M.A., Zerner, M.C. (1991). J. Am. Chem. Soc. 113,8210. Thulstrup, E.W., Michl, J. (1976), J. Am. Chem. Soc. 98,4533.
Vogel, E., lux, N., Rodriguez-Val, E., Lex. J.. Schmickler, H. (1990), Angew. Chem. Int. Ed. Engl. 29, 1387.
Thulstrup, E.W., Michl, J., Eggers, J.H. (1970). J. Phys. Chem. 74, 3868.
Wagner, P.J. (1989), Ace. Chcm. Res 22,83.
Tinnemans, A.H.A., Neckers, D.C. (1977), J. Am. Chem. Soc. 99,6459.
Wagner, P.J., Hammond, G.S. (1965). J. Am. Chem. Soc. 87, 1245.
Trefonas, Ill, P., West, R., Miller, R.D. (1985), J. Am. Chem. Soc. 107, 2737.
Wagner, P.J., Kelso, P.A. ( 1969), Tetrahedron Lett., 4151.
Troe, J., Weitzel, K.-M.(1988), J. Chem. Phys. 88, 7030.
Wagner, P.J., Kochevar, 1. (1968). J. Am. Chem. Soc. 90. 2232.
REFERENCES Wagner, P.J., Meador, M.A., Zhou, B., Park, B.-S. (1991). J. AIM. Chem. Soc. 113,9630.
Yamazaki, H., CvetanoviC, R.J. (19691, J. Am. Chem. Soc. 91, 520.
Wagniere, G. (1966). J. Am. Chem. Soc. 88, 3937.
Yamazaki, H., CvetanoviC, R.J., Irwin, R.S. (1976). J. Am. Chem. Soc. 98, 2198.
Wallace, S.L., Michl, J. (1983). Photochent. Photobiol. ; Proc. Int. Conf. Alexandria; Zewail,A.H..Ed.;Vol.2.p. 1191.
Yang, K. H. (1976). Ann. Pltys. N Y 101, 62, 97.
Wallraff, G.M., Michl, J. (1986). J. Org. Chem. 51, 1794. Waluk, J., Michl, J. (1991). J. Org. Chem. 56, 2729.
1 1
Yang, K.H. (1982), J. Phys. A 15,437. Yang, N.C., Elliott, S.P. (1969). J. Am. Chem. Soc. 91, 7550.
Waluk. J.. Michl. J. (1991). J. Org. Chem. 56, 2729.
Yang, N.C., Feit, E.D., Hui, M.H., Turro, N.J., Dalton, J.C. (1970). J. Am. Chem. Soc. 92, 6974.
Waluk, J., Mueller, M., Swiderek, P., Koecker, M.. Vogel, E., Hohlneicher, G., Michl, J. (1991). J. Am. Chem. Soc. 113, 5511.
Yang, N.C., Nussim, M., Jorgenson, M.J., Murov, S. (1964), Tetrahedron Lett., 3657
Wan, P.,Shukla, D. (1993),Chern. Rev.93.571.
I
Wasielewski, M.R. (1992). Ckem. Rev. 92, 435. Watkins, A.R. (1972), J.C.S. Fcrrudrry Truns, I 68, 28. Watson, Jr.,C.R., Pagni, R.M., Dodd, J.R., Bloor, J.E. (1976). J. Am. Chem. Soc. 98,2551.
Yang, N.C., Loeschen, R., Mitchell, D. (19671, J. Am. Chem. Soc. 89, 5465. Yang, N.C., Yates. R.L., Masnovi, J., Shold, D.M., Chiang, W. (1979). Pure Appl. Chem. 51. 173. Yang, N.C., Masnovi, J . . Chiang, W., Wang, T., Shou, H., Yang, D.H. (1981). Tetrahedron 37. 3285.
Weber, T., von Bargen, A., Riedle, E., Neusser, H.J. (1990). J. C'hem. P l ~ v s92.90. .
Yang, N.-C., Noh, T., Gan, H., Halfon, S., Hrnjez, B.J. (1988). J. Am. Chem. Soc. 110, 5919.
Weeks. G.H., Adco*ck, W., Klingensmith, K.A., Waluk, J.W., West, R.. VaSak, M., Downing. J., Michl, J. (1986). Pure Appl. Chem. 58, 39.
Yarkony, D.R. (1990). J. Chem. Phys. 92, 2457.
Weiner, S.A. (1971). J. Am. Chem. Soc. 93,425.
Yarkony, D.R. (1994). J. Chem. Pltys. 100, 3639.
Weller. A. (1958). Z. Phys. Chem. N.F. 15.438.
Zechmeister, L. (1960). Fortschr. Chem. Org. Naturst. 18, 223.
Weller, A. (1968). Pure Appl. Cltem. 16, 1 15.
Zener, C. (1932). Proc. R . Soc. London A137.696.
Weller, A. (1982a), Z. Phys. Chem. N F 133.93-
Zimmerman, H.E. (1964). Pure Appl. Chem. 9,493.
Weller, A. (1982b). Pure Appl. Chem. 54, 1885.
Zimmerman, H.E. (19661, J. Am. Chem. Soc. 88, 1564, 1566.
West, R., Downing, J.W., Inagaki, S., Michl, J. (1981). J. Am. C'ltem. Soc. 103, 5073.
Zimmerman, H.E. (1969). Angew. Chem. Int. Ed. Engl. 8, 1.
Wigner, E., Wittmer, E.E. (1928), 2. Phys. 51, 859.
Zimmerman, H.E. (1980). in Rearrungements in Ground and Excited States, Vol. 3; de Mayo, P., Ed.; Academic Press: New York; p. 131.
Wilkinson, F. (1964). Adv. Photochem. 3, 241.
Zimmerman, H.E. ( 1982). Chimia 36,423.
Wilkinson, F. (1968). in Luminescence, Chapter 8; Bowen, E.J.. Ed.; Van Nostrand, London.
Zimmerman, H.E., Grunewald, G.L. (1966). J. Am. Chem. Soc. 88, 183.
Wilzbach, K.E., Harkness, A.L., Kaplan, L. (I%@, J. Am. Chclm. Soc. 90, 11 16.
Zimmerman, H.E., Little, R.D. (1974), J. Am. Chem. Soc. 96.4623.
Winkelhofer, G., Janoschek, R., Fratev, F., v. R. Schleyer, P. (1983). Croat. Chim. Acta 56,509.
Zimmerman, H.E., Hackett, P., Juers, D.F., McCall, J.M., Schroder, B. (1971), J. Am. Chem. Soc. 93,3653.
Wirz, J., Persy, G., Rommel, E., Murata, I., Nakasuji, K. (1984). Helv. Chim. Acta 67, 305.
Zimmerman. H.E., Robbins, J.D., McKelvey, R.D., Samuel. C.J.. Sousa, L.R. (1974)J. Am. Chem. Soc. 96, 1974.
Wolff, T. (1 985). Z. Natrrrforsch. 40a. 1105. Wolff, T., Miiller, N., von Bilnau, G. (1983). J. Phorochem. 22, 61.
Zimmermann, H.E., Penn, J.H., Johnson, M.R. (1981), Proc. Nutl. Acud. Sci. U.S.A. 78, 202 1.
Wong, P.C., Arnold, D.R. (1979), Tetrahedron Lett. 23, 2101.
Zimmerman, H.E., Sulzbach, H.M., Tollefson, M.B. (1993), J. Am. Chem. Soc. 115,6548.
Woodward, R.B. (1942)- J. Am. Chem. Soc. 64, 72. Woodward, R.B., Hoffmann, R. (1969). Angew. Chem. Int. Ed. Engl. 8,781. Xantheas, S.S., Atchity, G.J., Elbert, S.T., Ruedenberg, K. (1991). J. Chem. Phys. 94,8054. Yadav, J.S., Goddard, J.D. (1986). J. Chem. Phys. 84, 2682.
Index
A term, 155-56, 158, 163 quantum mechanical expression, 160 Ab initio calculations, absorption spectra, 58-60 acetaldehyde, 380 benzene valence isomeriziltion, 449 butadiene, 60, 338-39,436-37 but-I-ene, methyl shift, 446-47 di-n-methane-rearrangement, 453, 45657 2.3-dimethylbutadiene, 437 electron transfer, 292 ethylene, cis-trans isomerization, 363 dimerization, 405 formaldehyde-methane. 395.429-3 1 H,. 235. 332-33 hexatriene. 36748 hydroperoxide formation. 478 methyl mercaptane, 358 Paterno-BUchi reaction, 429-3 1 perturbed cyclobutadienes, 4 13 Absorbance, 7,8, 265 Absorption coefficient, 7 Absorption spectrum. See ulso Spectrum; Polarization spectrum ab initio calculations, 58-60 anthracene, 19,72, 263 aromatic hydrocarbons, 7 1-76 4- and 5-azaazulene, 104, 106
azulene, 33-34, 106, 273-74 benzene, 37-38,86, 107 benzyl radical, 102 biphenyl. 128 biphenylene, 98-99 @carotene, 66 croconate dianion, 159 cyclooctetraene dianion, 86 3,8-dibromoheptalene, 169 1.3-di-t-butylpentalene-4.5-dicarboxylate, 99, 169 diphenylmethyl anion and cation, 170 I,4-disilabenzene, 105, 107 ethylene, 64-65 isoquinoline, 104, 106 1- and 2-methylpyrene, 166 naphthalene, 33.42, 104, 106 N-nitrosodimethylaniline, 133-34 octahydrobenzoquinoxaline, 144-45 [3,3]paracyclophane-quinhydrone,124 pentalene, 99 perylene, 261-62 phenanthrene, 8, 19, 273, 275 polyene aldehyde, 120 polyenes, 65-7 1 pyrene, 40,280-81 quinoline, 104, 106 rotational fine structure, 9
518
I
Absorption spectrum (conr.) silabenzene, 105, 107 substituted benzenes, 1 15-17 tetracene, 72, 103 tetracene radical anion and cation, 103-4 triphenylene, 274, 276 tropylium ion, 86 vibrational structure, 9 Acceptor, 1 19, 123-25, 173,464-65. See also Electron transfer; Exciplex Acenaphthylene, 165, 168, 346 dimerization, 412-13 MDC spectrum and polarized absorption, 157-58 perimeter model, 87 Acenes, 7 1-73 Acetaldehyde, 1 19-20, 380-82 Acetone, chemical titration, 428 oxetane formation, 3 18,427 singlet and triplet states, 382. 428 Acetophenone. 407. 467 Acetylene. 203. 348 cycloaddition. 4 16.423 excited state geometry. 4 5 4 6 2-Acetylnaphthalene, 398 6-Acetyloxycyclohexadienones, 463 Acidity, excited states, 48-52 Acrolein, 34, 38243,433 Acrylonitrile, 328,414-15,417 Activation energy. 38243,400 Acyl radical, 352-55, 380432,460 Adiabatic. See Potential energy surface; Reaction; Wave function Alkene, photocycloaddition, 366,420-23 addition to benzene, 420 substituted, 432 Alkyl amines, tertiary, 466 Alkyl aryl ketones, 399. 402 Alkylethylenes, 420 Alkylidenecyclopropene, 57-58 N-Alkylimines, 375 Alkyl iodide, 471 Alkyl methyl ketones, 383-84 Alkyl radical, 380 Allene, 416 Allyl radical. 102,460 All yl resonance, 46 1 Alternant hydrocarbons, 33.86.97. 1 12, 127, 167.441 n-bond order, 441 excited states, 17 first order Ci, 17, 54, 70 longest-wavelength transition, 74-75
INDEX mirror image theorem, 170-7 1 pairing theorem, 17, 90, 103 plus and minus states, 17, 18, 33, 54 radicals and radical ions, 101-104 topology and geometry, 70 Aminoborane, twisted, 207,214,218,226 Aminoethylene. 4 14 Aminophthalate dianion, 484 p-Amino-p'-nitrobiphenyl, 26142 Ammonia-borane adduct, 216, 218 Ammonium and sulfonium salts. 379 Angular momentum, 76.81, 161, 164 z component. 77.81. 164 operator, 223. 229 orbital, 28, 29, 76-78, 81 quantum number, 77.81 spin, 28, 29 Aniline. 52, 1 15-16, 264 Anils, 375 [nIAnnulene, 78.85, 161, 164. See also 4Nand (4N + 2)-electron perimeter [ IO]Annulene, 83 nodal properties of frontier orbitals, 92, 175 perimeter, 175 [8]Annulene dianion. frontier orbitals, 90 [I4]Annulene perimeter, 87, 174 Annulene, perimeter model, 78, 85 antiaromatic. 167, 205,445 bridged. 70 2n-electron and k-hole, 162-63 substituent effects on frontier orbitals, 172-73 [I I] and [13]Annulenyl ion, 87, 167-68 Anthracene. absorption spectrum, 19.72, 263 delayed fluorescence, 296-97 dimerisation. 3 19-20,416,418 electron transfer, 465 fluorescence and phosphorescence, 263, 266, 282 internal conversion, 253 intersystem crossing rate, 256,266 hom*o+LUMO transition, 19, 20 'La and 'L, band, 94, 263 linked, 418 methyl substituted, 350 orbital energy levels, 18 oscillator strength, 253 perimeter model, 93-94 photocycloaddition, 419 substituted, 302.41 1,413,419 triplet-triplet annihilation, 29697,320
INDEX
Anthracene-dimethylaniline exciplex, 28182 Anthracene-tetracyanoethylene complex, 465 Anthrylmethyl radical, 350 Aromatic hydrocarbons. 92, 151 absorption spectrum, 71-76 barrier in the S, state, 34546 condensed. 73-76.259 electron transfer reactions. 467 energy gap, 254-56 excimer formation, 281 intersystem crossing, 255-56 radiationless transition, 254, 259 Aromatic molecules, B term, 164 cycloaddition, 416-23 derived from (4N + 2)-electron perimeter, 87 dimerization, 415-19 electron-poor and electron-rich, 421 photosubstitution, 474 Arrhenius plot. 426 Atomic orbital (AO), 1 1 Atomic vector contributions. See Spinorbit coupling Azaazulene, absorption spectra, 104, 106 Aziridine. 442 Azoalkanes, 37677,392 Azobenzene. 121, 377-78 Azo compounds, 1 19, 121,358 cis-trans isomerization, 376-78 cyclic and bicyclic. 389-92 N2elimination, 387-92 reluctant, 392 Azo dye, 133 Azoendoperoxide dianion, 483 Azomethane, 121,376 Azomethines, 374-76 Azulene, absorption and emission spectra, 33-34, 104-106, 1 14,273-74 anomalous fluorescence, 253-54.27374 hom*o-LUMO transition, 33, 91 'L, band. 91-92 perimeter model, 83.88, 91-92 substituent effect. 1 14 triplet quencher. 37 1 0 term. 155-58. 163-67 11 and ~c - contribution. 164-67 quantum mechanical expression. 160 substituted benzenes, 172-73 sign. 156, 164, 167, 169 +
Back electron transfer, 2 16, 284, 363,425, 46546,46869,474 Bacteriochlorophyll, 474 Bacteriopheophytin, 474 Baldwin rules, 409 Band shape, 2 1-44 CD band, 143 MDC band, 155-56 ORD band, 143 'B, and 'B, band, 71-76 'B, and 'B, state, 79-81, 83,92 Barrelene, 456 Barrier, 180,200,232, 31 1, 318, 321, 34144,415 abnormal orbital crossing, 345 a cleavage, 380 correlation induced, 197, 324, 345, 380, 398.429 excited state, 328, 370. 388 natural correlation, 35 1 photodimerization, 342-44 Basicity, excited states, 48-52 Bathochromic shift, 104-105, 1 12, 123 solvent effect, 132 by steric hindrance, 127-28 Benzaldehyde, 299, 381-83 Benzene, absorption spectrum, 37-38.69. 73.86. 94-96 aza derivatives, 122-23 cyclodimerization, 4 19 density of states, 257 dimers, 324,419 excited state geometry, 43-44 fluorescence, 264-65 frontier orbitals, 32. 80. 90, 420 Ham effect, 134 highest resolution spectra, 43 'L, transition, 172-73 phosphorescence, 45 1 photoc ycloaddition, 420-23 rate constants of unimolecular photophysical processes, 250 rotational constants, 44 selection rules, 32 substituent effects on the intensity, 109 substituted. 115-18.458 transition densities, 80 transition moments, 95 triplet excited, 45 1, 483 two-photon spectrum, 43 valence isomerization, 26465, 302, 448-53 vibronic coupling, 32, 37-38, 96
INDEX Benzene oxide-oxepin equilibrium, 326-27 Benzhydrol, 397 Benzocyclobutene, 350,453 Benzonorbonadienes, substituted, 457 Benzophenone, 26748,407,424,467 Jablonski diagram, 252 oxetane formation, 407,424 photoreduction, 397-98,467 as sensitizer, 294, 367, 407 substituted, 52 Benzopinacol, 397 Benzoylox y chromophore, 154 Benzvalene, 264-65, 302,448-5 1 Benzyl anion and cation, 171 Benzyl radical, 102 Biacetyl, 266,291,425, 469 Jablonski energy diagram, 25 1 9,9'-Bianthryl, 48 Bichromophoric system, 305-6,418 Bicycle rearrangement, 459 Bicyclobutane, 23 1, 333, 336, 339, 34 1, 36647,433.438.443 Bicyclo[3.2.0]hepta-2.6-diene, 447 Bicyclo[2.2.O]hexene, 443 Bicyclo[3. l .O]hexene. 368.443 Bicyclo[3. I .O]hexenones, 463 Bicyclononadiene, 329 Bicyclooctatriene, 423 Bicyclo[4.2.0]oct-7-enes,437 Bimolecular process, 244,276-301, 313, 34 1 Biphenyl, 45, 203, 348 absorption spectrum, 128 electron transfer, 465 fluorescence spectrum, 263 hom*o and LUMO energy, 128 methyl derivatives, 128 Biphenylene, 98 Biphenylquinodimethane, 128 Biphenylyl t-butyl ketone, 383, 397 1 $3-Biradical,453-54, 456-57 1,4-Biradical, 236, 400,424,433-34,45354,457,470 l ,n-Biradical, 389 C,O- and C,C-Biradical, 428-3 1 Biradical, 187, 195, 197,205-10,2 19-230, 342.35 1, 391,430,434,461. See also Biradicaloid; ~o-electron-twoorbital model axial, 1 10, 210,442,478 cyclobutadiene like, 334 intermediates, 423-24.447, 461 magnetic field effect, 33 1
pair, 98, 210, 212, 231-32 perfect, 208-10,212, 223, 225,231-32, 234,236, 334,413,436 spin-orbit coupling, 219-29 triplet, 229, 323,403,424,469 wavefunctions and energy levels, 206, 208-2 12 Biradicaloid, 187.205.2 10-30,461. See also Biradical; Geometry; Minimum; Structure; Two-electron-two-orbital model critically heterosymmetric. 195, 214. 217. 228-29. 334. 338. 363 heterosymmetric. 2 10-19. 213. 236. 303. 334.413-15 hom*osymmetric. 2 10-12. 224. 234. 362 nonsymmetric. 210. 212.224 ( - 1- I . I '-Bis(2.4-dicyanonaphthyl). 47 1 ct.co-Bis(9-anthryl)alkanes. 4 18 Bis(9-anthry1)methane.4 18 Bis-9-anthrylmethyl ethers, a,a'disubstituted, 418 Bitopic, 190. 357. 379. See ulsn Topicity Blue shift, 133 Boltzmann's Law, 6 Bond dissociation. See Dissociation Bond order-bond distance relation, 45 Bond order, excited state, 441 Born-Oppenheimer approximation, 10, 34, 179,328 Hamiltonian, 180 states. 185-86 surface, 180-81. 316 Bracket notation, 5 Branching space, 183, 217,31617,339 Bridging, 87. 175. See also Perturbation, hierarchy of Brillouin's theorem, 54 I -Bromonaphthalene, 255,264 Brooker dyes, 135 Butadiene, 230-3 1, 336-39, 341, 366,408, 433,443 ab initio calculations, 43637 ab initio state energies, 60 HMOs, 26,3 1 isomerization, 333, 336-39, 366-67 ring closure, 332-33,43637 s-trans- and s-cis-, 2627, 31-32, 70, 276, 338,366,409,436 twisted, 341 Butanal, Norrish type I1 reaction, 399-400 l -Butene, 446 2-Butene, 364,406,424,478 r-Butyl ketones, 383
IZCisotope enrichment, 385 C term, 155-56, 158 quantum mechanical expression, 160 Cage effect, 385 Caldwell model, 343-44.415. 417 E-Caprolactam, 476 Carbazole. 442 Carbonyl compounds, 1 19-2 1, 365 aromatic, 92, 151, 3% chiral, 425 a cleavage, 380-87 dipole moment, 47-48 intersystem crossing, 29 optical activity, 147-49 photocycloaddition, 424-32 photoreduction, 466 rearrangement reactions, 460-64 solvatochromism, 133 spin-orbit coupling, 29-30 a,/?-unsaturated, 433-34.46263 #3.y-unsaturated, 120.453.460-62 singlet and triplet reactions, 462 /?-Carotene, absorption spectrum, 66 Carotenoids, 473 CASSCF, 58.363 Cationic dye, 217 CD spectrum, 147-50 cholest-5-ene-3#3,4/?-bis(p-
chlorobenzoate), 153-54 octahydrobenzoquinoxaline, 144
single chromophore systems. 147-52 two-chromophore systems, 152-54 CH and CD vibrations, 259-60 Characteristic configuration. 344-46. 349, 410 Charge-transfer (CT), 239, 421. See ulso Charge translocation; Electron transfer; Excimer; Exciplex band, 1 16, 125,420 character, 1 14, 282,465,478 complex, 34, 186,215, 34 1,420,465 configuration, 126 interaction, 238 intramolecular, 115, 126, 303 state, 1 16,238,474 transitions, 32, 115, 123-26 Charge translocation, 214-15 Chemical titration, 428 Chemically induced electron-exchange luminescence (CIEEL), 485 Chemiluminescence, 451-52,478,480-85 Chirality, 143-46, 150 Chiroptical measurements. 141-43
Chloranil, 123 I -Chloronaphthalene. 252 Chlorophyll a, 286,473 Cholest-5-ene-3#3,4/?-bis( p-chlorobenzoate), 153 CI. See also Configuration interaction 3 x 3, 230,232-33, 236, 334,336 20 x 20, 23633,235, 237, 336 . CIDNP, 220,469 CI matrix, perimeter model, 78-79,82,84, 93,97-99 Circular birefringence, 139, 141 Circular dichroism, 139, 141-42 magnetic, 154-77 natural. 143-54 Circularly polarized light, 139-4 1 Cis-band, 67 Cis-trans isomerization, 319, 329, 333, 336, 362-78, 388,406-7,427-29,433, 436-37.441.454-55. See also Synanti isomerization; State correlation diagrams azo compounds, 376-78 azomethines, 374-76 benzophenone sensitized, 367 butadiene, 367,437 cycloal kene. 364-65 dienes and trienes, 366-69 diimide, 376-77 double bonds, 362-78 enantioselective, 364 ethylene. 36243 heteroatom, substituent and solvent effects, 372-73 cis-hexatriene, 439 mechanisms, 36264 olefin, 364-66 Schiff base, 191, 373 stilbene, 369-72 triplet-sensitized, 363, 367 Clar's nomenclature, 20 Classification of photoreactions, 361 a Cleavage, carbonyl compounds, 352-55, 380-87.460-6I c yclobutanone. 386 ester, amide, 387 formaldehyde, 353-55 potential energy surfaces, 353, 355 state correlation diagram, 354 /? Cleavage, 425-26 CNDOIS method, 55 Collisional frequency, 247 Color center, 469
522
INDEX
Cone. 183,217. 316-17.366.415 Configuration, 12, 23 1-32 closed-shell. 205 electronic, 16-20, 193 excited, 12, 54, 58 ground, 12, 16, 54, 77 orbital, II Configuration correlatien diagram. 193. See also Correlation diagram Configuration interaction. 13. 16-20. 24. oI 52, 56, 72, 192. See ~ I s C complete. 55-57, 69 doubly excited (DCI, SDCI), 55, 57, 69 first-order, 16-17, 54, 70, 78, 102, 195 H4,234 second-order, 16 singly excited (SCI). 55-57. 70 Configurational functions, 11, 12, 17, 53 Conical intersection. 195, 2 17. 229. 23 1. 236-37.254, 315-18, 333-38, 363, 366-68. 375, 381.405.414-15.417, 430,433,43640,444,446-47.44950,452,454,457 true and weakly avoided, 182-86 Conjugated n systems. See cilso Polyenes; Annulenes; Aromatic hydrocarbons cyclic. 71-101, 171 linear, 63-70 Conrotatory. See Electrocyclic reactions; Reaction pathway Contact ion pair, 283,424-25,465 Continental divide, 3 12-1 3 Cope rearrangement, 446 Correlation diagram, 179, 184, 193-205, 23 1-32. 332, 334. 366,405,428-3 1, 449. 451-52. See ulso Orbitalcorrelation diagram: State correlation diagram Correlation, dynamic. 233 intended, 197-200, 35 1, 379 natural, 197-200, 204-5, 351, 383, 39697 Correlation effects, 56 Coulomb energy, 336 Coulomb gauge, 22 Coulomb integral. 14-15. 88. 97. 104. 208. 237 Couplet. 154 Covalent. See Perturbations; Structure Croconate dianion. 158-59 Crossing, Set. trlso Surface touching allowed. 3 16. avoided. 184-85. 200. 205. 23 1-32. 3 15-16. 777. 746. 10F 700 3 0 % -175 76 JX7
(.is-Crotonaldel~yde.448 Cross-link. 87. Y I. 101. 109. 169. Set' (IISO Perturbation. hierarchy o f CT. See Charge transfer Cubane. 409.434 Cyanethylene + NH,.304 Cyanine dye. 16. 69, 129. 218, 373 i n stretched poly(viny1 alcohol), 39 m-Cyanodibeniobarrelene. 457 I~cyanoheptalene,329-30 I-cyanonaphth;~lcne, 325, 468 9-cyanophenani hrene, 467 Cyclic n Systcnis. See ulso Annulene; Conjugated n Systems. Perimeter with a 4N-electron perimeter. 96-101. 167-70 with a ( 4 N + 2)-electron perimeter, 81-96. 16447 Cyclization. Sc,c Photocyclization Cycloaddition. 404-35. See also photocycloaddition aromatic compounds, 4 16-23 carbonyl group. 424-33 crossed. 23 1. 340. 342 enantioselect ive. 47 1 formaldehyde + ethylene, 430 ground-state forbidden, 230. 341 mixed. 410.4 19,433 photosensitized. 470-71 regiochemistry. 4 17-23 two ethylene molecules. 202-3. 237. 33336. 339. 417 c1.l-i-unsaturitrcd carbonyl compounds. 433-34 Cycloaddition. 12 + 21. 23 1. 237. 334-36. 364. 404-1 1. 414.417.443-44.448. 450.454-56. 461.467.478 N = N and C = N double bonds. 4 1 1 [4 + 2j.419, 471.477. 480 [4 + 41. 418-19 (2 + 2 21.455 .r[2 + 21. 23 1. 237. 333. 335.408-9.415. 443.446. 449 Cycloalkene. 30445. 407 I-Cycloalkenes. substituted. 427 Cyclobutadiene. 233. 236. 334.4 13 Cyclobutadienc dianion and dication. 16243 Cyclobutane. 404. 406. 415. 433. 444 fragmentatioii. 202-3 Cyclobutanol. JOO, 402-3 Cyclobutanone. 386 Cyclobutene, 230, 323-33, 336, 339, 36667. 410. .133. 436. 447
+
INDEX conrotatory ring opening. 196, 199 disrotatory ring opening, 194-201, 340 Cyclobutenophenanthrene, 203,34648 Cyclodecapentaene, even and odd perturbation, 9 1-92 Cycloheptatriene. 447 Cycloheptene, 407 Cyclohexadiene, 4 19, 443-44.47 1 Cyclohexadienone. 460,463 Cyclohexane. 476 1.2-Cyclohexanediol. 152 Cyclohexanone. 386 Cyclohexene. 407 Cyclohexenones, 463 Cyclononatriene. 329 1.5-Cyclooctadiene, 409 Cyclooctatetraene. 236. 423 dianion, 90 1.3.5-Cyclooctatriene. kinetics o f ring opening. 444 Cyclooctene, 364. 407 Cyclopentene. 407 Cyclopentenone, 433. 463 C yclopropane, 42 1-22 Cyclopropanone, 463 Cyclopropyl ketenes. 463 Cycloreversions. 347
1.4-Dibromonaphthalene, 269 Dicarbonyl compounds, PE and U V data, 121 9.10-Dicyanoanthracene, 470 I,2-Dicyanoethylene, 428 Dielectric constant, 13 1 Diene, 112, 444, 446. 453, 455 acyclic chiral. 455 bicyclic, 456 cis-trans isomerization. 366-369 Dienones. 462 Dienophil, 420 Dienylketenes, 4 6 3 4 4 Diethylaniline. 465 Differential overlap. Sot) Zero differential overlap hb. IOb-Dihydrobenzol3,4jcyclobutI I-2-trji~ccnaphthylene.3 12 Dihydrocarbi~zole.442 Dihydrocyclopropapyrene. 458-59 2.3-Dihydrofuran. 426 Dihydropentalene. 453 Dihydrophenanthrene. 37 1,440 1.4-Dihydrophthalazine, 484 Dihydropyrene. 443 Diimide. 376-77, 388-89 Diisopropylamine. 398 Diisopropylidenecyclobutane. 447 Davydov splitting. 152-53 Diisopropylmethylamine, 466 De Broglie relationship. 15 Dimer, syn and anti. 412 rrcrtts-Decalin. 292 head-to-head and head-to-tail. 41 1-14, Decarbonylation, 385 470 Degree o f anisotropy, 272 Dimerization, acenaphthylene, 4 12-1 3 Degree o f polarization. 272 ethylene. 202-3. 339. 416 Density o f states. 244. 257-58. 290 indene. 470 Deuterium labeling. 446 olefins and aromatic compounds, 415 Dewarbenzene. 448-52.483 I-4-Dimethoxynaphthalene, 468 Dewar-Evans-Zimmerman rules, 445 p-I)imethylaminohenzonitrilc, 303 I,4-Dewarnaphthalene. 32 1-23, 328. 346 Dimethylaniline. 115. 467 Di-n-methane rearrangements. 453-60. 462 9.10-Dimethylant hracene. 247 Diabatic. See Reaction; States Dimethylbutadiene, 367, 437 Diarylethylene. 44 1 2.3-Dimethyl-2-butene. 299, 41 1, 467 Diastereoselecti v i t y, 403 Dimethylcyclopropene. 367 oxetane formation. 326-27. 425-27 Dimethyldihydropyrene, 443 Diazabic yclooc tene, 292 Dimethylenecycloalkane, 438 Diazacyclooctatetraene, 390 2.5-Dimethyl-2.4-hexadiene. 3 0 - 1 3.5- and 3.6-Diazaindoles, 174 N.N-Dimethylindigo. 127 7.8-Diazatetracyclo[3.3.0.0~~4.0'~hJoct-7-ene,Dimethylketyl radical. 397 39 1 Dioxetane. 428. 476. 478. 482-83 Dibenzosuberene. 447 Diphenoyl peroxide. 485 Dibenzyl ketone. 385 Diphenyl amine. 340 9, !0-Dibromoanthracene, 256.45 1 9.10-Diphenylanthracene, 330 1.2-Dibromoethane. 270 Diphenylcarbene, 330 3.8-Dibrc~mnheptalcr~n. I69 1.2-Diphcnylcyclopropane. 469-70
INDEX Diphenyldiazomethane, 330 l-Diphenylethylene, 470-7 1 Diphenylketyl radical, 397 Diphenylmethyl anion and cation, 170 9, I0-Diphenylphenanthrene,330 1.5-Diphenylspiro[2.4]-4.6-heptadiene, 459 Dipole-dipole interactions, 290 Dipole field, 131 Dipole length formula. 23, 56-58 Dipole moment, 465. See also Electric dipole moment; Magnetic dipole moment exciplex, 282 excited-state, 47-48, 132 induced, 130 permanent. 130-3 1 Dipole strength, 158 Dipole velocity formula, 2, 56-57 Direct reaction. See Reaction 1,4-Disilabenzene, 105-7 Disproportionation, 228-29, 380, 390,433 Disrotatory. See Electrocyclic reaction; Reaction pathway Dissociation benzylic C-X bond, 379, 387 C-H bond, 348-49 a bond, 188-90, 210, 214-15, 356-57 n bond, 190-91, 210, 214, 216-17 B-N bond, 2 19 C-C bond, 190,350 C-N bond, 358, 380. 387-92 C-Ne bond, 215. 219 C - 0 bond, 190, 357-58 C-S@bond, 190 double bond, 378 H-H bond, 188-89, 356 polar bond, 2 17 single bond, 188-90, 378-79 Si-Si bond, 190, 216, 356, 379 toluene, 348-49 Distortion, diagonal, 333, 339-42, 413, 444, 446,450 rhomboidal, 334-35.405.4 14-1 5,450, 454 Disulfides. 150-5 1 1.3.-Di-~-butyl-pentaIcnt'-4.5-di~i1rboxylic ester. 169 2.3-Dithia-cr-steroids. I5 I 1.2-Divinylcyclobutanes. 408 Dodecahedrane, 4 19 Donor, 110-11, 114, 123-24, 135, 173,421, 465 Donor-acceptor chromophores, 134, 2 18 Donor-acqeptor complex, 28 1,465
Donor-acceptor pairs, 123-24, 2 14, 2 16, 218, 286. See also Charge transfer; Exciplex rigidly fixed, 286, 305 Doppler broadening, 42-43 Double bond. twisted. 188,205, 218. See crlso Ethylene; Propene Dynamic correlation, 232 Dynamic spin polarization, 206 Dynamical memory. 373 Dynamics of nuclear motion, 3 15.415.437 Efficiency, 247-49, 32 1, 406, 43 1 EHT calculation, 443 Einstein probability, 245 Electric dipole moment operator, 5, 13, 15, 23, 25 Electric quadrupole moment operator. 5, 13, 25 Electrocyclic reaction. 321,434-44 Electrocyclic ring-closure, 430,434,43637.44 1,443-44.456 conrotatory. 442 disrotatory. 436.449.454-55 Electrocyclic ring opening. 434. 443-44. 453.483 conrotatory, 196, 199, 338,439,444 cyclobutenoacenaphthylene, 347 disrotatory, 194-201, 340 kinetics, 444 Electromagnetic spectrum, 1-2 Electron affinity, 53, 282 interaction, 14, 52 Electronegativity, 191, 212, 218, 365, 372 Electron energy loss spectroscopy, 28 Electronic energy transfer. See Energy transfer Electronic excitation, 3 10 MO models, 9-2 1 quantum chemical calculations, 52-60 Electronic transitions, intensity, 21-27 notation schemes, 20 polarization, 38-40 selection rules, 27-34 Electron repulsion. 53. 55, 74.78. 234 Electron transfer, 123, 28347,292,398. 425,475-76,485 dependence on solvation, 304-6 free enthalpy. 285 light induced, 286, 304, 325,464-75 reactions, 464-75 sensitization, 468-70 Ellipticity, 140, 142, 154 El Sayed's rules, 255,257,266
INDEX Emission, 244-45, 260-76, 3 18, 323. See also Fluorescence; Luminescence; Phosphorescence Encounter complex, 278. 285. 29 1, 34 1 Endoperoxide. 477,480-8 1 Ene reaction, 477 Energy gap, 218, 255-56.258-59. 266, 316, 328.365, 372 law, 247, 254 Energy transfer, 277-78, 283. 287-97, 3 18, 363, 365.424.45 1 Coulomb and exchange mechanism, 290-95 nonvertical, 408 radiative and nonradiative. 287-88 Z- and E-enol, 447 Enthalpy and entropy control. 426 Environmental effects, 3 0 1 4 Ergosterol, 439 Ether-pentane-alkohol mixture (EPA). 250. 266 Ethylbenzene, 350 Ethylene. 64-65. 218. 420-21. 429-31.450. 454 cis-trans isomerization. 362-63 correlation diagram. I90 dimerization. 202-3. 236. 335. 339. 404. 416.454 electronic states. 64 [1,2] hydrogen shift, 363 MO diagram, 64-65 N t V transition, 24, 64 Rydberg orbitals, 5 9 4 , 64 spectrum, 64 substituted, 334 twisted, 190, 193, 206-14, 325-26, 36243 Ethyleneiminium ion, 2 14 a-(o-Ethy1phenyl)acetophenone.403 Ethyl vinyl ether, 433 Exchange energy, 287 Exchange integral. 14-15, 121, 208-10, 237, 289, 335-36.484 Excimer, 238,278-81, 335. 341-43.405, 412-13.418 fluorescence, 27940,418 intermediate, 405,414 minimum, 186,232,238-39. 279,320, 341-42,405,407,415 MO scheme, 279 wave function, 280 Exciplex, 48, 238, 278, 281-83, 285, 335, 342, 363, 366,412,419, 425,434, 46547,469 emission. 282. 3054. 326
intermediate, 422,434,444 minimum, 186, 341-43,405,415,419 triplet, 444, 467 wave function, 282 Excitation, electronic, 13, 310-13 Excitation energy, 13-14. 56. 93 HMO model. 13 quantum chemical calculation. 5 2 4 0 semiempirical calculations. 53-56 Excitation polarization spectra, 273-75 Excited State, 44-52, 263. See also Acidity; Basicity; Dipole Moment; Geometry n-bond order, 44 1 carbon acid. 447 degenerate, 156, 161-62 magnetic moment, 158-59 potential energy surfaces, 179-82 vertically. 3 18 Exciton-chirality model, 147, 152 Exciton state. 238, 280 Exo-selectivit y. 427 Exponential decay. 246 Extinction coefficient, 8, 21, 40, 139, 143, 265, 327. 365 Face-to-face approach, 186, 333.4 17 Faraday effect, 154 Far-UV region. 9 FEMO model, 15-16.67-68. 76-77 Fermi golden rule, 223, 255, 257, 290 Ferredoxin. 473 Fluorene, 238. 330 Fluorescence, 2 17, 244-45, 260-65, 282, 3 1 1, 320. 465 anomalous, 254, 273 benzene, 264-65 benzenoid aromatics, higher excited states, 253 delayed, e-type and p-type, 245, 295 donor-acceptor pair, 304-5 excimer, 238,279-8 1, 320 exciplex, 282, 304-5, 326 intensity, 32 1 lifetime. 47. 322 polarization, 39, 272-76 quantum yield, 248-49, 263-64 quenching, 277. 283 rate constant, 322 stilbene, 264, 370 Fluorescence excitation spectrum, 265, 27 1, 367 Fluorescence polarization spectrum, 272-73
INDEX
Fluorescence quenching. 277-78, 283 benzene. 265 diazabicyclooctene. 292 10-methylacridinium ion. 300-1 Fluorescence spectrum. anthracene. 263 anthracene + dimethylaniline. 282 azulene. 273-74 perylene. 2 6 1 4 2 phenanthrene. 273-75 porphine. 2 6 8 4 9 pyrene. 280-8 I triphenylene. 274. 276 Formaldehyde. excited state geometry. 45. 120. 186 barrier t o inversion. 120 energy level diagram. 119 solvent effect. 132-33 ethylene. 429-31 Formaldehyde Formaldehyde + methane. 351. 395-96 Formaldimine. 374-75 Formaldiminii~mion. 373 Formyl radical, 252 Forster cycle, 49. 51 Fiirster mechanism. 280. 290-91 Fragmentation. 228-29 Franck-Condon envelope. 189 Franck-Condon factor. 34-35. 223. 25840. 263 Franck-Condon principle. 34-36. 13 1. 186. 288. 310 Franck-Condon region. 254 Free-electron model, 15-16. 76 Free-valence number. 44 1 Frontier orbital. 432 Fulvene. 448-49 Fumaronitrile, 434 Funnel. 18246. 195. 197. 314-15. 317. 3202 1. 325-27. 332-33.336. 338-39. 342. 363. 36547.414.436.439-40.446. 449-50.452.454.456.461 bottom. 184. 230. 236. 317. 333. 366.436 pericyclic. 229-39. 332-39. 363. 36647. 407. 412. 437. 450 diagonally di5tortcd. 342. 366. 405. 408-409. 4 13. 444. 450 region. 335. 338-39. 3 6 6 4 7 . 4 15 }:our-clectrc)n-four-orhit;il model. 230. 232. 333 Furiin. 427
+
g Value. 220 0-0 Gap. 260 S,-S,, gap. 2 18
Gaussian linesh;~pe. 143. 155 Geometry. antiiiromatic, 205 biradicaloid. 187-88, 191, 193, 198,205, 208. 228. 13 1. 323. 339-4 1. 362-369. 449-50 excited states. 44-47. 56. 186 loose and tight. 191. 339-41. 362. 399. 406.424. 429.442-43 pericyclic. 33.1. 334. 337.442.446 rhomboidal. 2 36. 335-36 Gcrade and ung[.ri~de.30-3 1. 4 1. S ~ itl.~c) C Parity Geranonitrile. 4 I 0 Glyoxal, 121 Gradient. 180. 133-84. 317 Guanosine-5'-moncjphosphate.300-1 Hl molecule. INS-93 correlation di;~gritm.232 dissociated. 207. 216 electronic cibr~figuratians.64 M O and VR t tci~tment.191-93 H, H?. 231-32. 232-36. 238.279.295. 333-34.4 17 H,. 230-38.333. 336. 342.405 Hairpin polyenes. 70 Half-times. den :irhenzene and benzvalene. 45 1 Half-wave potc~~ti;ils. 285 Hiim effect. 134 Hiimiltonian. 14. 2 1. 78-79. 180 electronic. 10. IHO. 184 H M O model. I 3 spin-orbit c o ~ ~ p l i n g28. vibr;.rtioni~l. I I Hiird and soft cl~roniophore. 1 6 4 4 8 Harmonic oscill,rtor eigenfunction. 36 Hartree-Fock approximation. 52 Heavy atom effect. 28. 223-24.255.264. 269-70. 306. 412. 419 external and internal. 270 Heitler-London. 237 Helicene. I 5 I.44 I Hematoporphy rin. 295 Heptacyclene. 3-46 Heptalene. 169 Hcrzberg-Tcllc~vibronic coupling. 37 flcteroatom replacement. 8748,91, 109. See ctlso I'erturhat ion. hierarchy of Hetero-TTA. 2'h-97 Hexacene. 254 1.5-Hexadiene. -108. 446 2.4-Hexadicne. 16%
+
INDEX
Hexadienecarboxylic acids, 463 Hexafluorobenzcne. 450-5 1 Hexafluorodewarhenzene. 450 Hexahelicene. 15 1 Hexamethylbcnzcne. 126. 468 Hexamethyldewarbenzene. 468 Hexatriene. 46. 212-13. 329. 36748.439. 444 Highest resolution spectrii. 42. 44 AHI.. AHSI.. 1 0 7 4 9 H M O approuiniiit ion. 208 H M O model. 13-14. 17. 09. 74. 77. 8%. 332 coefficients. 410. 415 excitation energy. 13.73-75 polyenes. 67-68 radical ions. 102- 103 Hole transfer. 292 hom*o. 13. 43 1-32. 441. 428 Ahom*o. 82. 84. 10%. 110-I I. 164-67. 170. 172-75 l1OMO-hom*o ititcr;rction. 343. 411 hom*o-L-UMO crossing. 194. 198.376 H O M O - L U M O excitation. 75. 77. 345.441. 447 hom*o-1,UMO intcraction. 478 hom*o-1,UMO splitting. 91. 105 hom*o-.I,UMO tr;~nsition. 13-14. 16-17. 26. 31. 33-34. 65. 72. 91. 114. 127. 34% polyencs. 62. 69-70 s ~ ~ h s t i t u c cfl'cc't. nt 105 IIOMO-SOMO tran\ilion. I 0 1 Hot molcculcs. 252. 3 10-1 2. 373. 45 1 Hot I-c;~cticbn. 330-22. Scco crl.\ct Reactions cxcitcd stiitc. 301. 210 g r o t ~ n dsti~te.204. 310-1 1. 38% Hund's rule. 37% Hydrocarbon. Sclc~trlso Alternant hydrocarbon cata-condensed. 7 1-72.95 nonalternant. 104 Hydrogen abstraction. 203-4. 380. 395403,424.447 carbonyl compound. 35 1-52, 395-99.424 intramolecular, 399-403 in-plane iind perpcndiceliir attack. 39697 natural orhitill correlittion dii~griini.204 Hydrogen sclcnide, 434 [ I .2j-Hydrogen shift. 363 11.31-Hydrogen shift. 445 II.5 ]-Hydrogen shift. 447 [1.7]-Hydrogen shift. 439. 447. 459 Hydroperoxidcs. 476-77
p-Hydroxyacetophenone. 27 1 o- and p-Hydroxyaryl ketones. 387 4-H ydroxybenzophenone. 398 Hydroxyhydroperoxides. 476 3-Hydroxyquinoline. 50-5 1 Hyperconjugation. 88. 114. 173.227 Hyperfine interaction. 220-21. 331, 385 Hypsochromic shift, I04 solvent effect. 132 hy steric hindrance. 127-28 Increment rules. diene absorption. 112 enone absorption. 1 13 Indene, 470 Indicator equilibrinm, 48 I N DOlS method. 55.474 Indole. perimeter niodel. 88. 91 Indolizine. 174 Inductive effect. 104-9 In-plane inversion. 374 Intensity. 21-44 acene, 95 'B,. 'B?. 'L,. 'L?transitions. 85 benzene. 95-96, 108-9 borrowing. 37.96 Cl calculation. 57-58 CT band. 126 electronic transitions. 2 1-27 emission. 245-46 exciting light. 327-30 integriited. 246 I~itcriiction.t c . y. 461. Sc#t*crlso Throt~ghhond: -1'hrough-spircc . tliagoniil. 239. 334-39. 30647. 405. 4 1415.444 Intcrniediirtc. hir;~dical. 423-24. 447. 46 1 excimer. 405. 4 14 exciplex. 422. 434. 444 excited state. 3 14 pericyclic. 4 16 triplet. 368 Internal conversion (IC). 2 16. 244-45. 25254, 264. 287. 310. 320. 366 intersection coordinate suhspace, 183-84 Intersystem crossing (ISC). 30. 244-45. 254-56. 264. 266. 283. 286-87. 295. 302. 3 10- II 320. 328. 340. 349, 363. 382. 384. 396.401-3.406.424.42629.441.461-62 aromatic hydrocarbons, 255 biradicals and biradicaloids. 219-29 hyperfine coupling mechanism, 220-21 spin orbit coupling mechanism, 221-27
.
Inversion temperature, 426 Ionization potential, 53, 282 I o n pair, 424-25.47 1. See ulso Contact ion pair; Radical ion pair states, 238 Isoconjugate hydrocarbon model, 116 Isoquinoline, 104, 106, 109 Isoprene, 437 2-Isopropylbutadiene, 437 Isotope effects, 33 1, 385 Jablonski diagram, 243-45, 25 1-52 Jahn-Teller distortion. 96, 98 Kasha's rule. 253, 3 10 Ketene, 380, 386 acetal, 326-27, 469 Ketoiminoether, 4 1 1 Ketone. See ulso Carbonyl compounds aldol reaction, 427 excited state basicity. 52 K e t y l radical, 397-98. 467 Kinetics o f photophysical processes, 250. 297-30 1 Koopmans' theorem. 14, 12 1
'L, and 'L, band. 52. 7 1-76, 167, 173 IL, and 'L2band, 169 'L,, IL,, 'B,, 'B, states, 83, 86, 92-93, 167 Laarhoven rules, 441 Lactam-lactim isomerism. 174 Lambert-Beer law, 7-8, 366 Landau-Zener relation. 3 16 Langevin dipoles solvent model, 133 Laser spectroscopy. 4 18, 425, 444 Lifetime, 369, 434 excited singlet and triplet states. 245 higher excited states, 253 natural and observed, 245-47. 249-50. 253,266 triplet biradicals,, 229 Light absorption, 5-9 M O models. 11-13 Light, circularly polaridzed, 41, 139-44, 154, 158, 162-63 elliptically polarized, 139-43 linearly polarized, 1-3. 5.. 38-4 1, 139-4 1 Light-gathering antennae, 473 Linear momentum operator, 22,24, 56, 145 Line-shape function, 156 Liquid crystals, 272 Localized orbital model, 115-16
Locally excited states. 116 London formula, 237 Luminescence, 244, 260,418,478. See cilso Emission; Fluorescence; Phosphorescence Luminescence polarization. 272-76 Luminescence quenching, 293 by oxygen, 286 Luminescence spectrum, 1.4dibromonaphthalene, 269 free-base porphine. 268 naphthalene and triphenylene, 270 1.uminol. 482-84 I.UMO. 13. 43 1-32. 458 A12UM0. 82. 84. 110-1 1. 16447. 170. 17275 l . U M 0 - 1 - U M O interirctions. 343. 412 Magnetic circular dichroism (MCD), 142, 154-77 Magnetic dipole moment operator, 5, 13, 25, 145, 160 Magnetic field effects, 33 1 Magnetic moment, 77, 331-32, 385 z component. 1 6 1 6 3 p' and y - contributions, 164-65, 170 excited state. 158-59 Magnetic optical rotary dispersion (MORD), 142 Maleic anhydride, 420 Maleic dinitrile, 434 Maleimide, 420 Marcus inverted regicin, 284. 286 Marcus theory. 284-86 Markovnikov and anti-Markovnikov, 468 Mataga-Nishimoto formula, 53, 55 M C D spectrum, acenaphthylene, 157-58, 169 anthracene. 263 applications, 17 1-77 croconate dianion, 159 3.8-dibromheptalene, 169
1,3-Di-t-butylpentalene-4.5-dicarboxylic ester, 169 diphenylmethyl anion and cation, 170 4N-electron perimeter, 167-70 ( 4 N 2)-electron perimeter, 164-67 I- and 2-methylpyrene, 166-67 monosubstituted benzenes, 114, 172 mutually paired alternant systems, 169 orbital ordering. 175 pleiadiene. 157-58, 169
+
mirror image law, 170-7 1 temperature dependence, 155 vibrational structure, IS9 Mechanism cis-trans isomerization, 302-64 energy transfer, 287-91 oxetane formation. 425 photocatalytic, 364 photosubstitution, 474-76 [I,3] shift 6, y-unsaturated ketones, 461 transformation o f metacyclophane, 459 Menthyl phenylglyoxylate. 326-27.425 Merocyanine, 129, 135 Mesomeric effect, 109-18. 171 Meta c ycloaddition, 420-23 Metacyclophanediene, 324.443.458-59 I,6-Methano[lO]annulene. 175-77 Methano-cis-dihydropyrene, 324 I-Methoxy-l -butene, 427 p-Methoxyphenylacetic acid. 325 o-Methylacetophenone, 447 10-Methylacridinium chloritle, 300-1 Methylamine, protonated. 2 18 9-Methylanthracene, 403 Methylbicyclo[3. I.O]hexenc , 329 a-Methylbutyrophenone, 403 ( )-3-Methylcyclohexanon~.,148-49 Methylcyclohexene, 365 3-Methylcyclohex-?-en- l-one. 434 I-Methylcyclopropene, 336 N-Methyldiphenylumine. 4-12 Methylene blue. 295. 480 Methylenecyclohexene, I12
+
2-Methylene-5.6-diphenylbicyclo(3.1.0]hexene, 459 Methyl ethyl ketone, 384 2-Methylhexadiene, 3 18 Methylhexatriene- l,2,4, 329 Methyl iodide, 264, 419 4-Methylphenyl benzyl ketone, 385 I-Methyl-2-phenyl-2-indanol, 403 I-and 2-Methylpyrene, 166-67. 1?4-75 Methyl shift, 446 PMethylstyrene, 47 1 Methylvinylcyclobutene, 329 Micellar solvents, 412 Micelle, 384435.4 12 MIM (Molecules-in-Molecules), 117, 126 MIND013 calculations, 399 Minimum, antiaromatic, 442 biradicaloid, 187-91, 195-96, 205, 325, 36263, 365, 370-7 1
excimer, 186, 232, 238-39, 279, 320, 34142,405,407,415 exciplex, 186, 341-44,405,415, 419 excited state, 186-93 pericyclic, 195-96, 229-39, 320, 332-33, 338-44, 349-50, 370,419,436 reactive, 186-93, 349-50 spectroscopic, 186-93, 3 11, 320, 323, 349-50, 370 Minus states, 18-19, 27, 33, 54, 70 Mirror-image, absorption-emission, 260-63, 266 Mirror-image theorem (MCD), 170-7 1 M N D O C method. 56.383 M O (Molecular Orbital). II.See also Orbital M O configuration. 187-90. 192. 101. 131 M O model. electronic excitation, 9-21 light absorption. 1 1-13 Mobius array. 205. 329.445. 455 Molecular dynamics simulations. 132-33. 339 Molecular mechanics, 439 MRD-CI method, 58 Multiplicity, 12, 28-29, 180-81. 244. 247. 253. 257, 379 Mutually paired systems, 104, 170 Myrcene. 444 N and P transitions. 99-100, 167-69 N, Elimination from azo compounds. 38792 Naphthalene, 256, 321-24, 328, 346 absorption spectrum, 33. 42, 72. 104. 106 aza derivatives. 122 electron affinity and ionization potential. 14 excited state geometry, 44-45 HOM-LUMO transition, 14-55. 33 luminescence spectrum, 270 perimeter model, 83, 88, 91-92 photoreduction, 466 sensitization, 294-95 transition density. 33 triplet lifetime, 260 two-photon absorption spectrum, 41-42 Naphthalene- I-carboc ylic acid, 50 Naphthalene, 2-substituted. 84 I-Naphthol, 50 Naphthyl ketones, 351-52, 398 N E E R (non-equilibrium o f excited rotamers), 440
Nicotinamide adenine dinrtcleotide phosphirte (NAIII'). 473 Nitroanisoles. 475 Nitrogen heterocycles, 122-23 N-Nitrosodimethylamine, 133-34 Nitrosyl chloride. 476 Noncrossing rule. 182-83. 193 Norbornadiene. 230. 469 Norbornene. 407 Norbornyl iodide. 47 1 Norrish type Ireaction. 357. 380-97 Norrish typc II reaction. 323, 395. 399-404. 460,462 Nuclear kinetic energy. 217 N ~ ~ c l c o p h i laromatic ic substittrtions. 47476 Number o f active orbitals. 357. 379 Numbering systeni o f orbiti~ls.17 Nylon 6. 476 Octirhydrobenzoquinoxirline. 144 Octant rule. 148-49 Octatetraene. 70 Olefin. 363-66. 4 1 1. 4 12. Sc*cncilso Alkene cis-trans isomcrizirtion. 3 6 4 4 6 electron-poor. 42 1.428.43 1-32 elect ron-rich, 42 1, 423, 432 photodimerization. 404-1 1 Oligosilanes. 217. 302. 357. 392-94 One-electron model. 13-16 Oosterhoff modcl. 332. 335. 436 Optical activity, 140-54 exciton-chirality model. 147. I52 one-electron model. 147 Optical density, 7 Optical rotatory dispersion (ORD). 142-43 Orbital. See c11so AO: MO: Spin orbital active. 357, 379 canonical. 207 complex. 77. 79-81. 207. 478-79 c yclobutadiene-like, 334.4 14 degenerate, 187,413.420.458 frontier. 17. 90. 1 1 1. 17 1. 174-77. 292. 43 1 lone-pair. I 2 I. 150 most delocalized. 207-9. 478 nlost localized. 207-9. 225. 232. 334, 36243.413-14.478 nonbonding. 206. 232, 236. 334. 362-63. 413-14.436 nonort hogonal. 208 orthogonal, 206. 208 perimeter. 76. 89. 97
Orhitirl correlillion diagram. 189. 333 alternant hydl ocarbon. 345-46 benzene valc~rceisomerization. 449. 452 o bond dissociation. 188-89 n bond dissociirtion. 190-91 cubane. 410 cyclobutene. conrotatory r i n g opening. 196-97. 1'19 cyclobutenc. tlisrotatory ring opening. 194-98. ZcH) cyclohutenopl~cnirnthrene. 203 diimide. u-clc;~virge.3 8 7 4 9 cis-trans-i\c~merization. 377 endoperoxid lormation. 480-8 1 ethylene dinicrizirtion. 202-3. 333 hydrogen ab\t rirction. 204-5. 397 Iwo-step procedure. 197-W. 346 0rbit;rl crossiny. normal and abnormal. 344-48. 4o9-10.415 Orbital energy. 13-15. 104 perimeter niotlcl. 77 Orbital energy diagram. anthracene and phenantht-cne. 18 benzene pholocycloaddition. 42 1 two orbital s!,\tcni. 187 Orbital interacliun. 197-99,201-3.342.455 secondary. 4 1 2- 13 Orbital labeling system. 17 Orbital magnetic moment. 77. 161 Orbital ordering. 376.409-10.458 Orbital symmet I y. 193-97. 202.435 Orientation field. 131 Oriented molec~tles,38-40 Orlandi-Siebratid diagram. 369 Ortho cycloaddit ion. 420-23 Oscillator strength. 21-24. 38. 56-57. 67. 116. 246. 253. 289-Yo Outer-sphere clcctrcm-transfer reactions. 284 Overlap charge density. 26. 33 Overlap densit). 14. 79. 125. 150,209.485 Overlap integrirl. 237. 362 Overlap selection rule, 32-33 Oxabicyclobutatic. 433 Oxacarbene. 380 Oxa-di-n-met h;111crearrirngement 453. 462 Oxepin. 326 Oxctanc. 299. 3 19. 3M. 407.424-32. 460. 470 kinetic schcnit.. 426 Oxelene. 433 Oxirirnc. 340. 7 F X . 442
.
Oxygen effect. 256 Oxygen wave functions. 478-79. See crlso Singlet oxygen Pagodane$ 419 Pairing theorem. 17. 33. 90. 170. See ulso Alternant hydrocarbon Para cycloaddition. 420 Paracyclophanc, 124. 281 Parity. 3 1 Paterno-Riichi reaction. 424-32 carbon-carbon attitck. 43 1 carbon-ox ygen attack. 429-3 1 parallel and perpendicular approach. 428-30 PE (Photoelectron) spectra. IIS. 121. 376 Pentacene isomers, 94-95 Pentadiene. 367.437. 454 Pentahelicene, 44 1 Pentalene, 169. 236 Perepoxide. 478 Pericyclic reactions. 205. 230. 235. 238-39. 332-48, 393. Sene cilso Cycloaddition: Elect roc yclic reaction ground-state allowed. 197 ground-state forbidden. 194-95. 202. 229. 332. 344.454 spectroscopic nature o f states. 238-39 Perimeter modcl, 76-101. 161-70. 236 applications. 87-92. 17 1-77 C I matrix. 79. 82. 84. 97 complex MOs. 76 generalization. 81-101 Perimeter. 4N-electron. 96-101. 167-70, 445 charged. 98. 340 perturbation-induced orbital splitting. I00 substituent-induced perturbation. 108 uncharged, 97-98. 340 Perimeter ( 4 N + 2)-electron. 77.81. 85-87. 90. 92. 109-10. 161. 164-67. 171 charged, 78.8247 uncharged. 78.90.92 Peripheral bonding. 337. 367. 415 Peroxides. 338. 480 Perturbation. chiral. 147 covalent ( y ) . 21 1-13. 220-21. 224-25. 228-29, 332. 339. 366 hierarchy of. 87 polarizing ( 0 ) .212-13. 218. 224. 236. 332. 334. 339. 36243.436 structural. 82
theory, 2 1. 37. 104. 122. 127. Set. erlso PMO method time-dependent. 2 1 Perylene. 262 Phantom state, 365. 371.413 Phase angle, perimeter model, 82-83, 9192.97 Phase difference. 140 Phase polygon. 79-80 Phenanthrene, 203. 348. 440 absorption spectrum. 19 H M O orbital energy levels. 18 perimeter model, 93-94 polarization spectra. 275 Phenes. 71 Phenol. 52. 463 Phenylalanine. 152 I-l'henylcyclohexene. 468 I - and 2-Phenylnaphthirlene. 46-47 I'heophytin a. 473 I'hosphoresccnce. 244-45. 266-71. 2W. 31 1 I'hosphot-escence cxcitirtion spectrum. 27071 p-hydroxybenzophenone. 27 1 Phosphorescence polirriz:rtion spectrum. 273-76 Phosphorescence spectrum. Scp crlso Luminescence Spectrum anthracene. 263. 266 porphine. 268 Photoacoustic calorimetry. 434 Photocatalyst. 364. 473 Photochemical electron transfer (PET). See Electron transfer Photochemical nomenclature. 18 1 Photochemical reaction models. 309-32 Photochemical variables. 324-3 1 pitrirmcters. 3%) I'hotochromic material. 448 Photo-Claisen. photo-Fries. 358. 379, 387 Photocyclization. N-methyldiphenylamine. 340,442 ( Z ) -I.3.5-hexatriene. 229 cis-stilbene. 238, 440 Photocycloadditions. 341-43. 404-5. See also Cycloaddition arene-irl kene. 422 carbonyl group. 424-32 excited benzene to olefin. 42 1 ground-state forbidden, 341 n./l-unsaturated cirrhonyl compounds. 433-34 Photocycloreversion. 4 18
INDEX Photodimerization, 34 1-44, 405. See ulso Dimerization acenaphthylene, 413 anthracene, 341, 416 Copper(1) catalyzed, 408 olefins, 344, 404.41 1 schematic potential energy curves, 343 Photodissociation, 378-94. See also Dissociation Photodissociation spectroscopy, 238 Photodynamic tumor therapy, 295 Photoenolization, 447-48, 462 Photofragmentation o f oligosilanes and polysilanes, 392-94 Photoinduced electron transfer, 286, 466, 470, 472. See crlso Electron transfer Photoisomerization, 32 1, 368. 388 o f benzene, 448-53 Photon-driven selection pump, 426 Photonitrosation. 476 Photoorientation, 276 Photooxidations with singlet oxygen, 47680 Photophysical parameters, 249 Photophysical process, 179, 186, 196, 243306, 310-13 i n gases and condensed phases, 301-2 lifetime, 2 4 8 4 9 quantum yield, 248-50 rate constant, 249-50 temperature dependence, 302-3 Photoracemization o f ketones, 400 Photorearrangements. See Rearrangement reactions Photoreduction, 395-99 benzophenone, 395,397-98 benzophenones and acetophenones, 467 carbonyl compounds, 466 naphthalene by triethylamine, 466 Photoselection, 272, 275 Photosensitization, 292, 407. See ~ l s o Sensitization Photostationary state. 329. 365. 371 Photosubstitutions, 444-76 Photosynthesis, 286,472-74 Photosystem Iand 11,472-74 Pigment P680 and P700.472-74 Pinacol, 397-98 ppinene, 444 Piperylene, 299, 367, 401 pK value, 49-52 Plastoquinone, 473 Platt's nomenclature, 21, 79, 91, 167, 169
Pleiadene, 165, 168, 3 1 1 M C D spectra, 157-58 polarized absorption, 157-58 Plus and minus states, 18-19, 27, 33, 54, 70 PMO method, 75, 88-89, 343. 431 Polarization degree, 41 Polarization direction, 2, 3 8 4 0 , 57, 272, 274 absolute and relative, 6 'L,, 'L,, 'B, , 'B2 band, 86-92-93. 109 N a n d P bands. 100 substituent effect, 109 Polarization. electronic transition, 38-40 Polarization spectrum, 273-76 acenaphthylene, 157 azulene, 274 cyanine dye. 39 o f fluorescence and phosphorescence, 273-76 phenanthrene. 273 pleiadicnc. 157 pyrcne. 40 triphenylene. 276 Polyacene, 92-96 Polyamides, 476 Polyene aldehydes, 120 Polyenes, 48. 54. 65-71, 370 'A, state, 70 alternating double and single bonds, 67-68 dimethyl and diphenyl, 67 Polysilane, 392-94 Porphine, free base, 268, 276 Porphyrin-quinone systems, 286 Porphyrins, 171, 295,474 expanded, 296 Potential energy curves. ethylene, 66, 373 diatomic molecule, 36, 259 formaldimine, 374-75 formaldiminium ion, 373 molecular oxygen, 478-79 SiH, elimination, 394 Potential energy surfaces, 179-93, 309, 3 14 acetaldehyde, a cleavage, 38 1-82 acrolein, a cleavage. 382 adiabatic and diabatic; 179, 185-86, 315-16 anthracene dimerization. 3 19 benzaldehyde, a cleavage, 38 1-82 chemiluminescence, 48 1 1.4-dewarnaphthalene, photoisomerization, 322
INDEX diimide, cis-trans isorneriration, 377 excimer formation, 279 excited states, 179-8 1 , 186-93 formaldehyde, a cleavage, 353-55 nonconcerted reactions. -349-55 12 + 2) and .r[2 + 2) processes. 230-38 PPP met hod. 17-1 8. 53-55. 70-7 1. 102-3. 273. 332 Prebenzvalene, 453 Precalciferol, 4 3 9 4 0 Prefulvene. 42 1-22.44849 Prismane. 448.450.452 2-Propanol, 398 [ I.I.ljpropellane. 189 Propene. twisted, 210, 214 Proton transfer, 49 Protonation site, 174 Pseudosigmatropic shift, 446 Purple bacterium Rhodopseudomonas viridis, 472 Pyramidalization. 2 18, 362 Pyrene. 87. 134. 256. 280-8 1. 295 absorption and fluoresce~icespectra. 280 2-Pyridone, 174 Quadricyclane, 230,469 Quantum yield, 247-50, 3 17-18, 365, 367. 378,406,416,438 benzvalene isomerization. 45 1 chemiluminescence, 483 cyclobutene formation. 438 dewarbenzene isomerization, 450 differential, 248 emission, 253 fluorescence, 248-50, 263, 266, 297, 30 1-2 hexamethyldewarbenzenc valence isomerization, 468 hydrogen abstraction, 398 internal conversion, 248-49 intersystem crossing. 248 -49 N?elmination, 392 Norrish type 11 reaction, 399 oxetane formation, 299 phosphorescence. 248-50, 266, 269 photodimerization, 406 photonitrosation, 476 radiationless deactivation, 249 stilbene isomerization. 370-71 temperature dependence, 373 total, 248 triplet formation, 249 Quartet state, 102
Quencher, 298 triplet, 371, 398. 412, 473 Quenching. 3 1 1,425 concentrat ion, 277-78 diffusion-controlled, 299 dynamic and static, 299-301 electron-transfer, 283-87, 465 excited states, 277-78 heavy atom, 283-87 impurity, 277 oxygen, 286-87 rate constant, 298 Quinhydrones o f the [3.3]paracyclophane series. 124 o-quinodimethanes, 391, 453 o-quinol acetate, 463-64 Quinoline. 104. 106, 26748, 303 Quinone, 100. 286 Radiation, electromagnetic, I intensity. 4 interaction with molecule. 21 Radiationless deactivation, 245, 249, 25260. 31 1 processes. 244, 253. 263 transition, 253-54, 257-60 Radical anion, 476 Radical cation. 467. 469-70. 475 Radical. doublet and quartet configuration, 102 odd alternant, 101-3 Radical ion pair, 285-86, 363, 465-67, 46970.485 Radical ions. 284 alternant hydrocarbons. 102 pairing theorem, 103 Radical pair, 33 1, 460-62 Rate constant, 248-49 electron transfer processes, 285 fluorescence, 246, 248 internal conversion, 247, 249 intersystem crossing, 247, 249, 266 isomerization, 370 nonradiative energy transfer, 290 phosphorescence. 246. 249 spontaneous emission. 245 triplet-triplet energy transfer. 292, 294 vibrational relaxation, 247 Rate. internal conversion, 247 intersystem crossing, 256 radiationless transitions, 259 unimolecular processes, 245-47 Rate law. exponential. 246. 248
INDEX
534
I
Reaction, adiabatic and diabatic, 31 I, 32224,450-5 1 antarafacial. 445, 447 complex, 3 13 concerted, 340 direct, 310-1 1. 327 electron-transfer. 464-75 ground-state allowed, 197, 324 ground-state forbidden, 194-97, 205, 229, 405 HI + H,, 230-36, 238, 279. 295, 342, 417 hot, 301, 310-1 1, 320-22, 388 intermolecular, 3 18 nonconcerted, 349-59.406 pericyclic, 205. 320 stereospecific. 340 symmetry forbidden, 443 with and without intermediates, 313-20 Reaction coordinate. 180 Reaction dynamics, 3 18 Reaction field. 130-32. 474 Reaction medium. 324-26 Reaction pathway, 193, 3 18 bifurcated. 3 18, 440 conrotatory. 339. 366 disrotatory, 339, 366-67.436-37 ground-state, 436 rectangular. 235. 335 tetrahedral. 235 Rearrangements. 434-64 unsaturated carbonyl compounds, 460-64 Recombination, 380 Reductive elimination of SiH,, 392-94 Refractive index, 2, 140 Regioselectivity, 326, 407, 410-17, 420, 422-24.432. 434,456,458,469 Rehm-Weller relationship, 285-86 Relaxation. 6 Reorganization energy, 284 Resonance integral, 53, 56, 1 1 I, 2 11, 224, 344. 402,415 Retinal Schiff base, 373 Retro-Diels-Alder reaction, 484 Rhodamine dyes. 250. 373 Rhodopsin. 373 Right-hand rule. 150-51 King opening and closure. S c p IIISO Correlation diagram in hiradicals. 228-29. 430 Ring strain. 350. 364. 438 Rose Bengal, 295,480 Rosenfeld formula. 145-46 Rotation, specific and molar. 142
Rotational barriers. 329 Rotational fine structure, 9, 42-44 Rotational strength. 145-46 Rule of five. 409. 443-44.446 Rydberg orbital. 59-60 Rydberg transition. 2 1 Saccharides. 473 Saddle point. 180. 333 Salem diagram. 205, 355-56 Schcnck mcch;~ni\m.364 Schiff base. 191. 217. 273 Schrodinger equittion. 180 Scrambling of hydrogen. 450 Selection rules. 27-34 Self-consistent field (SCF) methods. 52 Self-consistent reaction field, 131-32. 474 Self-quenching. 277-78 Self-repulsion energy. 33 Semibullvalene. 457 Sensitization. 277. 292-95. 31 1. 364-67. 388-90.407-8, 416.450.456.462-63 Sensitizer. 292. 364. 372. 389.407 chiral. 364, 47 1 Sigmatropic shift, 435. 445-48. See crlso Hydrogen shift [ 1.21, 363, 446. 453-54.456.46243 type A and type B. 463 [ 1,3]. 445-46. 450,46043 (3.31.446 [ij]. 445 antarafacial. 445. 447 degenerate, 450 suprafacial, 23 1. 445, 447 vinyl. 454. 456 Silabenzene, 105-7 Silylene, 392-94 Single bond. dissociation, 188-90 stretching. 205 Singlet, 13. 180 impure. 28 Singlet oxygen. 295,476-80 Singlet-singlet energy transfer. 278. 291 Singlet-triplet splitting. 33-34. 55. 76. 123. 220-2 1. 227, 229. 266, 289 Sing!ct-triplet ~ t i ~disposition. te 267-68 Singlet-triplet tr;lnsitic>n. 29 12 Slater determini~~it. Slater rules. 15, 234 Solvation. 284. 305. 324 Solvatochromisn~.positive or negative, 132-33 Solvent cage, 326, 387, 465, 47 1
INDEX Solvent effect. 48, 129-35, 303-6, 325 continuum and discrete theories, 131 Solvent parameters, 1 3 1 Solvent reorganization, 303. 306 Solvent-separated ion pair, 425 SOMO, 102, 167 SOMO+LUMO excitation. 10? Special pair. 473-74 Spectra, atomic and molecular. 9 Spectral distributions. 288 Spectral ovcrlap, 288-90 Spectroscopist's convention, 181 Spin correlation effect. 389 Spin density. 469 Spin functions, 54, 224, 227. 274 singlet and triplet. 209. 222 Spin inversion, 255, 324, 389. 399,483 Spin orbital, I I, 78, 289,479 Spin-orbit coupling, 28-29. 180, 220-29. 255, 257, 264, 266, 268, 273, 363, 402-3,406,426,434,483 atomic vector contributions. 224-26 operator, 28. 222. 274 parameter, 223-24 strength, 222. 224. 227-28 through-bond. 222. 225-27 through-space. 225-27 vector, 222-23 Spin relaxation. 2 19-20, 228 Spin selection rule, 28. 32, 291 Spin-statistical factor. 320. 416 Spin. z component. 1:' State correlation diagram. 189. 194. 200-5. 446 alternant hydrocarbon, 345 anthracene dimerization. 3 19-20 azobenzene. cis-trans-isomerization. 377-78 benzene valence isomerization. 449. 452 o bond dissociation. 188-89 n bond dissociation. 190-91 butanal. Norrish typ 11. 400 chemiluminescence. 48 1-82 u cleavage, 354-55. 381-82. 388 cyclobutanone. 386 cyclobutene ring opening. 195-99, 201. 436 cyclob~~tenophenanthrenc fragmentation. 347-48 diimide, 376-77, 388-89 ethylene, cis-trans isomerization. 191. 362, 365 dimerization. 334
formaldehyde + methane. 35 1, 395 H4,232-35.416 hydrogen abstraction, 203-4, 35 1, 397 methanol, C - 0 bond dissociation, 358 a-quinol acetate reactions. 464 photodimerization. 403 stilbene. 369 tetraniethyl-1,2-dioxetane,482 State diagram. Sue Jablonski diagram State. electronic. 16-20 degenerate, 159. 161 diabatic. 185-86, 3 17 267disposition of (n,n*) and (n,n*), 68 G . S and D. 98-100. 192 N and P, 99-1 00 Stationary point, 184 Stereochemistry, 404,412,431,436,455 head-to-head and head-to-tail, 334-35, 342.41 1-15.417, 433-34 syn and anti, 342,412,414,417 Stereoelectronic effects. 466 Stereoselectivity, 403. 407, 41 1, 427. 466 Stereospecificity. 4044,428, 442,446 Steric discrimination. 403 Steric effects. 126-29, 450 Stern-Volmer equation. 298 Stern-Volmer plot, 299-301 Steroids, 1 13 Stilbene, 218,401. 41 1,440-43, 466 cis-trans isomerization, 369-72 fluorescence quantum yields. 264 triplet, 401 Stokes shift, 260-62 Stretched polymer films, 39-40, 272 Structure, biradicaloid, 352, 355-56 charge-separated, 362 covalent, 54. 207, 233 dot-dot, 207-8, 21 1-14, 217, 233, 356 hole-pair, 207-8. 2 12-1 8, 223-24, 233, 356 zwitterionic. 189, 191-92. 207. 223. 233-34 Substituent effect, 104-1 18, 172-73.41315.42 1. 438. 453. 457. See crlso Steric effects aniline. 1 15-16 hyperconjugative. 173 MCD signs, 171 methyl group, 1 14. 173-75 Substituent parameter. I I 1 Substituents. acceptor. 414-15, 432, 457-58 &ma-. 414-15.458
INDEX electron-donating and withdrawing, 17172,236, 335,402,421, 431-32 inductive, 104-9 +Mand - M , 110 mesomeric, 109- 1 18 strength, 115 Substitution, 87-88. See also Perturbation, hierarchy of Sudden polarization, 212-13, 218, 317, 362, 474 Supermolecule, 132, 23 1, 3 13, 341, 405 Supersonic jet laser spectroscopy, 44, 367 Surface jump, 2 17, 3 15-17.483 Surface touching, 181-83, 217, 236, 3 15, 333. See also Crossing Symmetry selection rules, 30-32 Syn-anti isomerization, 374-75 Temperature dependence, photophysical processes, 302-3 stereoselectivity, 326-27 Tetra-t-butyltetrahedrane, 385 Tetracene, 254 absorption spectrum, 72, 103 Tetracene radical anion and cation, 103 Tetracyclooctene, 423 Tetramethylcyclobutane, 406 Tetramethyl-l,2-dioxetane,428, 482-83 Tetramethylethylene, 326-27 Tetranitromethane, 465 a,a,a1a'-Tetraphenylbenzocyclobutene. 350 Tetraphenylcyclobutiine, 470 Tetraphenylethylene, 33 1 Tetraradical, 230 Thermal equilibration, 260, 310, 321 Thiobenzophenone, 328 Thiocarbonyl compounds. 254 Third row elements, 105, 107 Three-quantum process, 459 Through-bond interaction, 121, 292, 402, 409- 10 Through-space interaction, 12 1, 177, 402, 409- 10 TICT (twisted internal charge transfer). 191, 214, 218, 303,474 Time-resolved spectroscopy, 250, 444 Toluene, 348-49 Topicity, 190, 352, 356-58, 379 Trajectory, 180, 317,415 Transannular interactions, 175-77 Transient spectrum, 369,465
Transition. See also Electronic transition 0-0, 37. 26162 allowed and forbidden, 27 charge-transfer, 123-26 degenerate, 158 electric dipole, 5 quadrupole. 5 intensity, 21-27 longest-wavelength, alternant hydrocarbons, 74-75 magnetic dipole, 5 n-m*, 1 18-23 carbonyl compounds, 34, 119-22 nitrogen heterocycles, 122-23, 133 solvent effect. 133 radiationless, 323 spin-forbidden, 266 two-photon, 40-42 vertical, 35-36 vibrationally induced. 36-38 Transition density, 25, 34, 79-8 1, I5 1 Transition dipole moment. 23. 32. 37. 56. 85 electronic, 36. 144-46. 150-5 1, 160. 164 magnetic, 144-46, 150-51, 153, 160, 164 Transition moment, 5, 13. 15. 18, 25-27. 29, 126, 246, 263, 266, 272-74, 289, 303 computation. 56-58 direction, 25, 272-73 perimeter model, 79-8 1, 85, 99 solvent effect, 133 tensor, two-photon. 4 1 Transition state, 3 17-18, 334-35 geometry. 384 Translocation of formal charge, 214 Tricyclo[4.2.0.0?']octadiene. 409 Trienes. 1 12. 366.444 Trimethinecyanine, FEMO model, 16 Trimethylene biradical, 225, 227 Trinitrobenzene, DA complex, 125 Triphenylene, 270, 276 Triphenylmethane dyes, 373. , Triple bond. bending, 205 Triplet, 12, 14, 180 impure, 28, 220-21, 227 Triplet energy of sensitizer, 372, 407 Triplet function, 222 Triplet state, aromatics, 76 calculation, 55 ethylene, 64 Triplet-triplet absorption, 256 Triplet-triplet annihilation, 238, 278, 287, 295-97, 31 1, 318, 320, 41617,480
Triplet-triplet energy transfer. 278, 291-94 Triplex, 364, 471 Trisilane, 394 Tritopic. See Topicity Tropylium ion, 86 . Tunneling. 321, 33 1 Two-chromophore system, 152-54 Two-electron-two-orbital model, 191-93, 205-18, 230, 336.413-14 Two-photon excitation, 3 1 I . 370 Two-photon process, 248, 330 Two-photon spectroscopy, 28.40-44. 54, 70 Two-photon spectrum. benzene, 43 naphthalene, 4 1-42 Two-quantum process, 459 Two-step procedure, 198-9. 203, 346 Ubiquinone, 474 Umpolung, 470 Ungerade, 30-3 1 , 4 1. See ulso Parity Unimolecular processes, 243-52, 3 13 Unitary group approach (UGA), 55 Valence isomerization, 324, 329,458,46869 . of benzene, 264-65. 302,448-52 Valence isomers of benzene. 390-91.448 Valence tautomers, 326 Valerophenone, 403 triplet quenching, 300-1 VB (Valence Bond), 208, 230, 232 VB correlation diagram. Src. Salem diagram VB exchange integral, 237 2 x 2 VB model, 230, 232-33, 235-37 VB structure, 189-92, 202, 205-208, 232, 236, 355-56, 362. Set. trlso Structure Vector potential, 22 Vibrational equilibration, 185, 444, 450 Vibrational relaxation, 245, 252, 264-65, 288, 301, 303, 310. 375 Vibrational structure, 9-1 1, 13, 35, 43, 159 Vibronic coupling, 31, 33, 37, 44, 273 benzene, 32,37-38, % Vibronic progression, 38
Vibronically induced transitions, 36-38 2-Vinylbicyclo[l . 1 .O]butane, 368 4-Vinylcyclohexene, 408 Vinylcyclopropane, 454 Vinyl radical, 45657 [I ,21 vinyl shift, 456 Vinyl substituent, l I0 Viscosity, 299-300, 306, 325, 370 Vision, primary steps, 373 Vitamin D. 434-35, 439 Wagner-Meerwein rearrangements, 472 Walsh rules, 44 Wave function, electronic, 10. adiabatic and nonadiabatic, 184 excited state, 53-54 total, 10 vibrational, 10. 23. 35-36 Wave number, 2-3 Wave packet, nuclear, 185, 3 16 Wave vector, 3. 22 Wavelength. 2- 4 Wavelength dependent photoreaction, 302, 327-30,448 Weakly coupled chromophores, 126 Weller equation, 285, 422 Wigner-Witmer rule, 277-78, 291. 295, 384 Woodward-Hoffmann rules, 324, 404, 419, 434-37.442-444-45.454 x , and xz vector, 183-84, 217, 316, 339
Xanthone, 52, 296-97 p-Xylene, 407 o-Xylylene, 350 Zeeman splitting. 160-65 Zero differential overlap (ZDO) approximation, 26, 53, 56, 79, 96, 124, 126, 234 Zero field splitting, 22 1 tensor, 22 1, 225, 227 Zimmerman rearrangement, 453 Zwitterionic character, 54, 207, 214, 23536, 356
