Photochemistry Volume 2

A Review of the Literature Published between July 1969 and June 1970

By D. Bryce-Smith

The Royal Society of Chemistry

Copyright © 1971 The Chemical Society
All rights reserved.
ISBN: 978-0-85186-015-2

Contents

Introduction and Review of the Year By D. Bryce-Smith, xi,
Part I Physical Aspects of Photochemistry,
Chapter 1 Spectroscopic and Theoretical Aspects By D. Phillips,
Chapter 2 Photophysical Processes in Condensed Phases By D. Phillips,
Chapter 3 Gas-phase Photochemistry By D. Phillips,
Part II Inorganic Photochemistry By D. Phillips,
1 Photochemistry of Water, H2O2,and Aqueous Anions, 235,
2 Photochemistry and Photoluminescence of Transition-metal Co-ordination Complexes, 243,
3 Gas-phase Studies, 281,
4 Solid-phase Luminescence and Photoreactions, 287,
Part III Organic A spects of Photochemistry,
Chapter 1 Photolysis of Carbonyl Compounds By W. M. Horspool,
Chapter 2 Enone Rearrangements and Cycloadditions: Photoreactions of Cyclohexadienones, Tropones, Quinones, etc. By W. M. Horspool,
Chapter 3 Photochemistry of Ole(lns, Acetylenes, and Related Compounds By W. M. Horspool,
Chapter 4 Photochemistry of Aromatic Compounds By A. Gilbert,
Chapter 5 Photo-oxidation and -reduction Reactions By A. Gilbert,
Chapter 6 Photoreaclions of Compounds Containing Heteroatoms other than Oxygen By A. Gilbert,
Chapter 7 Photoelimination Reactions By A. Gilbert,
Part IV Polymer Photochemistry By D. Phillips,
1 Photopolymerization, 757,
2 Optical Properties of Photoexcited Polymers, 770,
3 Photo-cross-linking and Grafting, 776,
4 Photodegradation and Stabilization,
Errata, 800,
Author Index, 801,

CHAPTER 1

Part I

PHYSICAL ASPECTS OF PHOTOCHEMISTRY

1

Spectroscopic and Theoretical Aspects

As was stated in Volume 1 of this series, the experimental photochemist can learn much from a study of the absorption spectroscopy of molecules under his investigation. Thus, details of the nature of the excited states formed upon absorption, radiative lifetimes of these states, changes in dipole moment and polarizability (and consequently reactivity) may be obtained from a careful study of the electronic absorption spectra. Frequently information obtained in this way is augmented by ab initio, or more usually semi-empirical, calculations upon the energy levels and properties of excited states of molecules. Together, theoretical and spectroscopic considerations can often provide a rationale for the photochemists’ experimental observations. The volume of material of a spectroscopic and theoretical nature published annually is vast, however, and it would be impossible, and indeed misplaced, to attempt to include a comprehensive survey of this field in this volume. Instead a brief section is included on absorption spectra and energy level calculations which may be of direct interest to the photochemists for whom this volume is intended. References to other theoretical and spectroscopic work will often be found in sections dealing with the photochemistry of specific molecules.

As in Volume 1, a considerable section of this chapter will be devoted to theoretical considerations of radiationless transitions. This is an area in which experimental evidence is building up, and the phenomenon provides a challenge to theoreticians and experimentalists alike to gain a deeper understanding of the complex nature of these processes. It must be stressed that the authors of this volume are primarily experimentalists, and the account given here of the efforts of colleagues engaged upon the difficult task of providing an adequate theory which will quantitatively account for observed rates of internal conversions and intersystem crossings may be coloured by a lack of understanding of the true complexities of the problem.

The extensive data which are available on the rates of decomposition of molecules in the gas phase have been treated in the past in only a semi-quantitative manner, but it is encouraging to note that there has been an increasing tendency of late to apply the theories of unimolecular decomposition to excited states of molecules. A short section is devoted to this and other theoretical aspects of photodecomposition, and some further applications of the Rice–Rampsberger–Kassel–Marcus (RRKM) treatment to excited states of molecules in the gas phase will be found in Chapter 3.

Finally, we have included a section on developments of new experimental techniques in this Chapter. This section belongs properly with the experimental sections Chapters 2 and 3, but since the techniques described are principally of a spectroscopic nature, and often apply equally to condensed phase and gas-phase investigations, the section will be included here as a preface to both Chapter 2 and Chapter 3.

1 Absorption Spectra and Energy Level Calculations

As stated above, this section is not intended to be comprehensive, but a selection of published papers on molecules which may be of interest to photochemists is given. These molecules will be dealt with in order of increasing complexity. The calculation of Franck-Condon factors is of importance in that these determine the intensities of both radiative and non-radiative transitions. In small molecules, work has continued on investigations into hitherto neglected vibration-rotation interactions in the calculation of Franck-Condon factors. For the hydrogen molecule, the (B1Σu+<- X1Σg+), (I1Πg<- B1Σu+), (d3Πu<- a3Σg+), (C1Πu<- X1Σg+), (D1Πu<- X1Σg+), (h3Σu+<- c3Πu) systems, the E1Σg+<- B1Σu+), (G1Σg+<- B1Σu+), and (kΠu<- a3Σg+) systems, have been extensively studied. In all cases good agreement with experiment is possible if vibration–rotation interactions are considered. Previous computations, for e.g. the Lyman bands, have not included the effects of the centrifugal potential, but it has been tacitly assumed that the principal effect of the rotational energy is a shift of the potential curve by a constant energy displacement. This assumption is most severely tested in calculations of Franck–Condon factors for electronic transitions in light molecules, and especially if the transition involves states with appreciably different potential curves, as in the (BΣu+<- XΣg+) Lyman system. The rotational angular momentum necessarily changes in all transitions (since there is no Q branch, it being a Σ <- Σ transition), and this serves to emphasize the vibration–rotation interaction effects. By way of contrast, the (d3Πu<- a3Σg+) Fulcher bands arise from states which have similar potential curves, and smaller vibration-rotation interaction effects on intensities would be anticipated.

When the centrifugal potential is explicitly taken into account, the Franck–Condon factor appropriate to the electronic transition (v’, J’ ->v”, J”)] is given by:

q(v’, J’; v”, J”) = | [∫ φ’v’, J’ (R) φ”v”, J”(R) d R|2 (1)

where φv’, J’(R) and φv”, J”(R) are the eigenfunctions of the nuclear motion for the two states. For 1E states the eigenfunctions are solutions of the Schrodinger equation:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

Solutions of this equation can be obtained by replacing it with the equivalent finite-difference equation which is solved numerically by a computer procedure. Details will not be given here, but will be found in ref. 2. It is found that vibration–rotation interaction effects for e.g. the Lyman bands are substantial, but small for the Fulcher bands as expected. It should be noted that the calculations of distribution of intensity in these systems have been in the limit of a strictly valid Franck-Condon principle i.e. the molecular electronic dipole moment does not depend on internuclear separation. It is known, however, that this is not strictly true, and in fact there may be an appreciable variation with internuclear distance in some systems. Small variations of dipole moment with internuclear distance tend to be amplified if vibration–rotation effects are appreciable. Thus, if µ(R) represents the dipole moment function, then expansion of µ(R) around R = 0 leads to

∫ φ’v’ J’ (R) µ(R) φ”v”, J”(R) d R = µ0 ∫ φ’v’ J'(R) φ”v”, J”(R) d R + µ1 ∫ φ’v’, J'(R) Rφ”v”, J”(R) d R + … (3)

If the first term, the overlap integral, vanishes, then the other terms may have great significance. It has been shown in the papers discussed briefly above that the overlap integral may vanish because of vibration–rotation interaction, a fact which has not hitherto been appreciated, and clearly any complete theoretical treatment must necessarily involve inclusion of an explicit dipole moment function as well as the effect of perturbations mentioned previously. Franck–Condon factors for the v= 0 progression of the N2 fourth positive system have been derived.

High-resolution studies of the C2 Π ->X2 Π emission bands of NO, the electronic spectrum of cyanogen, and the (B3Σ – ->X3Σ-) and (A3Σ- [right arrow]X3Σ-) bands of SO have been recorded, and forbidden absorption bands of O2 in the argon continuum region of the spectrum observed.

Ethane is the only alkane which exhibits structure in its molecular electronic absorption spectrum. High-resolution photographs reveal that this is diffuse vibrational structure in the case of C2H6. However, the 0–0 of C2D6 near 1406 Å exhibits a distinctive rotational contour. Computer simulation of this contour shows that the transition moment for the 0–0 band lies perpendicular to the C — C bond. The assignment of the electronic transition is 1Ag ->1Eu. Vibrational analysis shows that the transition effects a large reduction in the C — H stretching frequency, consistent with a 731 cm-1 shift of the origin band towards higher frequencies on deuterium substitution, and with the promotion of an electron out of a C — H bond. These results are consistent with the observed photochemistry of this molecule (see Chapter 3). Although the C — C bond energy is lower than that of the C — H bond in the ground state of ethane, the photochemistry suggests that C — C cleavage is not the main primary process at ca. 1400 Å, indicating either that the 1Eu state is formed via promotion of an electron from the C — H bond, or that excitation of an electron from the C-C bond is followed by a rapid redistribution of energy. The present results favour the former alternative. The linewidths for C2D6 are ca. 3 cm-1 indicating a lifetime of the state of 10-11 S. The lifetime of the C2H6 excited state is at least ten times shorter.

The absorption spectrum of bicyclohexylidene vapour is shown in Figure 1 to consist of two bands, both associated with the C = C bond. The crystal absorption spectrum is similar, and polarization measurements reveal the very interesting fact that both transitions are polarized in the same direction and correspond to 1B1u excited states. The existence of two strong 1Ag ->1B1u transitions in the low-energy spectral region of simple olefins is incompatible with the predictions of π-electron theory. There can be only one strong π-electron transition with the observed transition moment direction, the π -> π* (N [right arrow] V)] transition. It is thus felt that the other band must be due to a σ -> σ* transition and this recognition of the existence of a strong σ -> σ* transition in the energy region previously assumed to be the sole domain of π-electrons is of obvious importance to those concerned with the photochemistry of such molecules. Should other simple olefins also show such bands, the impact of these observations upon photochemistry and indeed π-electron theory could be very great. However, a Rydberg transition should occur in this wavelength region, and although the assignment of the second observed band to this transition has been discounted, further evidence must be presented before the σ–-σ* assignation can be accepted unequivocally.

Singlet–triplet absorptions in mono-olefins are very weak, but the intensity of such bands can be increased by the presence of oxygen at 70 atmos pressure. An attempt has been made to seek the absorption spectra of methyl-substituted ethylenes corresponding to the ground state to triplet Rydberg state transition (TR<- N) under these conditions, but only bands corresponding to the (T<- N) transition and due to contact charge-transfer transitions from the ethylene to O2 could be observed.

Zero-differential-overlap (ZDO) molecular orbital calculations of the geometry for excited states of unsaturated molecules have been criticized on the grounds that such calculations completely discount one of the major factors influencing bond lengths and bond angles in excited molecules. As an example, ethylene can be considered. For the π-electron energy levels επ and επ*, according to one-electron LCAO–MO theories which retain the overlap integrals S

επ = (α + β)/(1 + S) (4)

επ* = (α – β)/(1 – S) (5)

Since the π* level is more antibonding than the π level is bonding, the total π-electron energy E for a ππ* electron configuration is net antibonding.

E = 2(α – βS)/(1 – S2) (6)

the excited molecule, the energy can be stabilized by decreasing the magnitudes of β and S; this is accomplished by lengthening the C — C bond and/or by twisting the pπ-orbitals away from each other. Any theoretical method which neglects S must predict that the one-electron π -energy of ππ* configurations is independent of geometry, and thus neglects a major driving force in the determination of the equilibrium geometry.

According to an analysis based on first-order perturbation theory, the net antibonding character of the lowest 3ππ* state of an acyclic polyene containing 2m unsaturated carbon atoms is given approximately by

-8S(β – αS)/(m + 1)2 (7)

The destabilization due to overlap effects decreases rapidly with increasing chain length, and the exothermicity associated with rotation by 90° about a C = C double bond becomes zero in longer polyenes. Hence, errors in calculated geometry due to the ZDO approximation should be significant only for the shorter chains. Semi-empirical SCF–LCAO–MO π-electron calculations predict that the 90° twisted conformations of both the excited singlet and triplet states of ethylene are more stable than the planar forms, in agreement with the conclusions reached spectroscopically.

Allene is a molecule which has excited the interest of photochemists of late, and Gaussian SCP calculations on the ground state of this molecule have generated values for the energy levels of the excited states. These are, however, only in poor agreement with experiment, although they agree qualitatively in that they predict the spectrum of allene to consist of three weak transitions followed by a strong transition at shorter wavelengths, as observed. The absorption and fluorescence spectra of crystalline trans– stilbene,14 and the electronic properties of polyenes and polyphenylacetylenes have been described, 15 the latter from a theoretical point of view.

Benzene is a molecule whose photochemistry has been much studied, and interest in this molecule by theoreticians and spectroscopists is also intense. It is known that the lowest state of benzene is the 3B1u state, but higher triplet levels have been predicted by calculations and have in some cases been observed experimentally. Recently, absorption in the 4300 Å region has been attributed to the first allowed T–T transition in benzene, 16 producing the T4 state which is of 3E20 symmetry. The absorption decayed with the same lifetime as other well characterized triplet absorptions, and also has been observed in another study. In the latter work, absorption by excited singlet state benzene molecules was also monitored, and the transition attributed to the S1 -> S4 states, i.e.1B2u ->1E2g. From the positions of the spectral bands, the 3E2g and 1E2g states of benzene can be placed at 52,800 and 58,400 cm-1 respectively above the ground state. The figure for the 3E2g level is ca. 2 eV below that calculated using the Pariser–Parr–Pople approach. The order of the excited singlet states has also been called into question by a thorough experimental study of higher π–π* transitions in solid argon, krypton, xenon, and nitrogen matrices in the spectral region 2800–1700 Å. On the basis of the observed vibrational structure the second excited singlet state of benzene has been confirmed as 1B1u rather than 1E2g. Moreover, theoretical calculations of the dynamic electronic vibrational coupling between the 1B1u and the 1E1u states also support the 1B1u assignment of the 2100 Å transition. A phenomenon closely related to electronic relaxation in large molecules is the occurrence of line broadening in the absorption spectra of the higher excited electronic states. In these excited states the Born–Oppenheimer separability conditions for electronic and nuclear motion break down because of vibronic coupling between isoenergetic zero-order Born–Oppenheimer vibronic states which correspond to different electronic configurations. In the so-called ‘statistical’ limit when the density ρ of vibronic levels is sufficiently high to exceed the reciprocal of the mean vibronic coupling term v between the zero states, inhomogeneous line broadening is expected to be observed. For an isolated resonance, the line shape is expected to be Lorentzian, and the line is given by Δi = 2πv2 ρ. In the solid phase medium effects may also cause line broadening by the following mechanisms:

(i) coupling of states with the lattice vibration of the host should lead to temperature-dependent broadening Δph of all the vibronic levels for a given electronic state;

(ii) vibrational relaxation effects will lead to the broadening of the higher vibrational components within a given electronic state. These effects should be negligible for the 0–0 band, and a contribution to the line broadening additional to that of the 0–0 band gauges the vibrational relaxation process. This additional broadening due to vibrational relaxation is in the range Δ vr = 1–10 cm-1;

(iii) in the statistical limit the radiationless relaxation times should be independent of the medium.

Thus the total linewidth of a large molecule in a solid can be expressed in the form

= Δ = Δρh + Δvr + Δi (8)

(Continues…)Excerpted from Photochemistry Volume 2 by D. Bryce-Smith. Copyright © 1971 The Chemical Society. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.