Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
THE NATURE OF THE PRIMARY PROCESS IN PHOTO- CHEMICAL REACTIONS G. K. ROLLEFSON Department of Chemistry, University of California, Berkeley, California Received May 25, 1958 In the previous reports the nature of the primary action of light in photo- chemical reactions has been discussed in some detail. The principal de- velopments since that time have been concerned with the determination of the products of the photodissociation processes and with the process known as predissociation. The term "predissociation" was introduced by Henri (4) in describing certain diffuse bands discovered by himself and his asso- ciates. Their choice of this name was based on their idea that these diffuse bands corresponded to excitation of the molecule to a loosely bound state in which it could dissociate readily. The present-day views are based on the picture, offered by BonhoefTer and Farkas (2) and by Kronig (11), that the molecule in the activated state produced by the absorption of these diffuse bands undergoes a radiationless transfer to another electronic state and dissociates in a time which is short compared to the period of rotation. Such changes have been expressed graphically by Herzberg (6) by means of potential energy diagrams (figure 1~. The transition from one state to another by this predissociation process must occur without any appreciable change in the separation of the atomic nuclei or in the energy of the sys- tem. From the standpoint of the diagrams, this means that we have radiationless transitions occurring only at the intersection of two curves. In our diagram (figure 1), if the excitation-by light takes the molecule from the normal state, n, to a state represented by a point on the curve a above the level of the intersection of that curve with a', then as the molecule vi- brates in the excited state we have the possibility of a transfer from a to a' occurring. If this intersection point is above the level corresponding to the dissociation of the state represented by a', the molecule will dissociate within the next vibration. The diagrams (figure 1) must be looked upon as schematic only, since, actually, we cannot specify the positions and energies of the atoms with the precision indicated by the curves. According to Heisenberg's Un- certainty Principle, the product of the uncertainties of position and mo- ~ Contribution No. 3 to the Third Report of the Committee on Photochemistry, National Research Council. 35
36 G. K. ROLLEFSON Momentum is h/2~. A more convenient form for us to use in the discussion of predissociation is that the uncertainty in the energy multiplied by the uncertainty of the time is hider. Our curves therefore should be looked upon as representing mean values only. Furthermore the intersections of two curves must be considered not as points but rather as regions within which the molecule may be thought of as being in an inde- terminate state. The probability of a transfer occurring will depend upon the time that the molecule is in this indeterminate condition. This time will depend on the kinetic energy of vibration and the range over which the transition may occur. Thus in figure la we should expect the transition to occur only if the condition of the molecule corresponds to a point rather close to the intersection. On the other hand, in figure lb, where the two curves cross at a small angle, when 1~ . t/ r la RE. FIG. 1. Potential energy diagrams ~ \~7/K r lb we take into consideration the- blurring called for by the Uncertainty Principle, we see that there will be a long region along the curve a which is in the indeterminate area. Hence we should expect, other things being equal, that molecules represented by diagrams similar to figure lb will show predissociation over a wider range and more frequently than those corresponding to the other diagrams. At the present time we do not know enough about these potential curves to make any quantitative calculations concerning predissociation. Furthermore, these curves must be considered as only schematic for polyatomic molecules, as they are far from simple vibrators. The Uncertainty Principle also accounts for the appearance of a diffuse spectrum. If the excited state dissociates within a time ~ the width, w, of the spectrum line which put the molecule into that state will be given by
PRIORY PROCESS IN PHOTOCHEMICAL REACTIONS 37 we = h/2,r or, if we express w in cm.-i, we = h/2?rc. In any ordinary spec- trum which shows a fine line structure, the broadening due to the Doppler eff ect is ten to one hundred times the natural width of the line; therefore the predissociation process must shorten the life of the excited state by a factor of this magnitude for the eject to be noticeable. It is to be expected there- fore that, whenever we find a diffuse absorption spectrum, fluorescence will either be very weak or absent. If we consider the competition between fluorescence and decomposition as represented by AB*klAB+hv AB* 2 A + B + kite. the fraction of the activated molecules which radiate will be given by 1 + k /k Hence, the greater k2 is relative to kl, i.e., the shorter the life period, the weaker will be the fluorescence. As k2 usually must be ten to one hundred times kit for a diffuseness to be observed, it follows that a marked weakening of fluorescence is a more sensitive indicator of predis- sociation than a disuse spectrum. It is by no means universally true that we have transitions from one state to another occurring whenever we have a crossing of the potential energy curves for those states. Kronig (11) has set up the following selection rules for diatomic molecules: (~) There shall be no change in the total angular momentum. (2) Transitions occur only between states of the same multiplicity. (Invalid for large multiplet separations.) (~) The quantum number ~ changes by O or +1. (~) Transitions occur from a positive to a positive state or from a negative to a negative. (5) If both atoms are the same, the states involved in a transition are either both symmetrical or both antisymmetrical. The interpretation of the spectra of polyatomic molecules has not progressed to the point where we can say whether an analogous set of rules applies or not. However, we may say that we can expect to find the probabilities of radiationless transfers oc- curring ranging from practically every time the molecule reaches the state represented by the intersection of the potential energy curves to practically complete prohibition of the change. The probability of a radiationless transfer occurring is modified consider- ably by a magnetic field or collisions with other molecules; This gives rise to the phenomenon known as induced predissociation. Experimental evidence for the occurrence of this process was obtained by Loomis and Fuller (12) and by Kondratiew and Polak (10) in the study of the absorp- tion spectrum of iodine. They found that the addition of inert gases in- creased the absorption coefficient of iodine in the vapor state for those bands involving values of v', the vibrational quantum number in the upper
38 G. E. ROLLEFSON state, greater than 12. Turner (15) found that, if iodine vapor were il- luminated with light absorbed in the band region in the presence of a magnetic field or inert gases, there was a marked absorption of the spec- trum lines characteristic of iodine atoms. More recently, Rabinowitch and Wood (13) have made a more quantitative study of the effect of inert gases. In their experiments, the dissociation of the iodine was determined by measuring the light absorbed by the remaining iodine molecules. They concluded that argon, nitrogen, and oxygen caused the dissociation of the activated molecule at every "gas-kinetic" collision. Helium was some- what less effective but, with a pressure of 500 mm. of that gas, all of the molecules absorbing light were dissociated. In the case of bromine we have some evidence that dissociation occurs in the band region even without the aid of collisions. Urmston and Badger (16) found in their experiments on the photochemical reaction of bromine with platinum that the rate of the reaction was independent of the distance between the platinum and the illuminated zone. Their ex- periments were performed at low pressures, s that induced predissociation did not need to be considered. Neither could their results be accounted for on the basis of active molecules, as they wer able to demonstrate that fluorescence was confined to the illuminated zone It is possible that the reaction was due to a continuum underlying the band absorption, a situa- tion analogous to that in the hydrogen-chlorine reaction which has been discussed recently by Bayliss (1~. However, no appreciable difference was noted in rates whether blue or yellow light was used. Other reactions in which no effect was noted on comparing the rates using blue and yellow light are the bromination of acetylene (3) and the formation of hydrogen bromide Ail. The efl ect of inert gases on the latter reaction has been stud- ied in considerable detail. Instead of finding an accelerating erect, which could be attributed to induced predissociation, a retarding effect was noted, due to an increased rate of recombination of bromine atoms (7, 8, 14~. This is understandable if we assume that with bromine, as has been found with iodine, every collision of an activated molecule results in dis- sociation. Under such conditions, no effect assignable to inducedpre- dissociation would be detectable above about 10 mm. pressure. A quite different result has been reported for the hydrogen-chlorine reaction by Hertel (5~. He found that, if the light absorption occurred in the banded region of the spectrum, this reaction was accelerated by the ad- dition of inert gases. This effect was believed to be due to the activated molecules being dissociated by collisions. The mechanism of such a dis- sociation is probably induced predissociation. The examples which we have just discussed show no sign of predissocia- tion in their absorption spectra, indicating that the life of the undisturbed activated molecule is at least of the order of 10-9 sec.; with iodine, studies
PRIMARY PROCESS IN PHOTOCHEMICAL REACTIONS 39 of the fluorescence indicate a life of 10-7 sec. It is apparent, therefore, that spontaneous or induced predissociation may occur with any kind of light absorption. Hence, the possibility of dissociation in the primary step of a photochemical reaction cannot be excluded on the basis of well- defined lines in the absorption bands nor even on the basis of fluorescence observations. Another interesting example is furnished by tellurium vapor, Tee. In this case the limits of the natural and induced predissociation do not coincide. The difference must be due to the induced effect involving lower lying level than the spontaneous process (9~. If the light is absorbed by a complex molecule it is much more difficult to demonstrate an effect due to induced predissociation on account of the complexities introduced by secondary reactions. Usually, the photochem- ical experiments are carried out at such high pressures that the induced predissociation is a maximum if it is anything like that with simple mole- cules. Any effects of this type must therefore be sought for at low pres- sures, particularly in systems for which the quantum yield of the primary process is less than 1. REFERENCES (1) BAYLISS: Trans. Faraday soc. 33, 1339 (1937~. (2) BONHOEFFER AND FARKAS: z. PhYSik. Chem. 134, 337 (1928). (3) BOOHER AND ROLLEFSON: J. Am. Chem. soc. 66, 2288 (1934). (4) HENRI AND TEVES: Nature 114, 894 (1924); Compt. rend. 179, 1156 (1924~. (5) HERTEL: z. physik. Chem. B15, 325 (1932~. (6) HERZBERG: z. Physik 61, 604 (1930~. (7) HTI.FERDING AND STEINER: z. physik. Chem. B30, 399 (1935~. (8) JOST AND JUNG: z. physik. Chem. B3, 83 (1929). (9) KONDRAT]EW AND LANDIS: z. Physik 92, 741 (1934~. (10) KONDRAT]EW AND POLAK: Physik. Z. Sowietunion 4, 764 (1933). (11) KSONIG: z. Physik 60, 347 (1928); 62, 300 (1930~. (12) LOOMIS AND FULLER: Phys. Rev. 39, 180 (1932~. (13) RABINOWITC~ AND WOOD: Trans. Faraday Soc. 32, 547 (1936~. (14) RITCHIE: Proc. Roy. soc. (London) A146, 828 (1934~. (15) TURNER: Phys. Rev. 41, 627 (1932) . (16) URMSTON AND BADGER: J. Am. Chem. Soc. 5S, 343 (1934~. /
