SPLITTING OF MOLECULES BY THE ACTION OF LIGHT*

A. N. Terenin

1. In recent decades the mutual interpenetration of physics and chemistry has been steadily increasing, and the commonality of the aims of these two fundamental sciences of the structure and transformations of matter has become apparent. This striving for rapprochement is vividly attested by the fact that Mendeleev’s chemical law of the periodic properties of the elements has been placed, as a common banner, precisely here—in the physical auditorium of the University—and has been set here alongside the laws of Newton and Faraday.

My communication is devoted to a question whose development by physicists began about 25 years ago. This question is the following: by what path does the energy of a light quantum \(h\nu\), having been absorbed by a molecule \(AB\), lead to rupture of the valence bond \(A—B\) and to the decomposition of the molecule into free atoms or radicals \(A\) and \(B\)? Schematically this simplest photochemical reaction may be written as follows:

\[ h\nu + A—B \longrightarrow \overset{\leftarrow}{A}\cdot + \cdot\overset{\rightarrow}{B}, \tag{1} \]

where the dots denote the remaining valence electrons that had previously participated in the bond, and the arrows above the letters indicate the possibility of the particles flying apart with an excess of kinetic energy.

Foreign scientists, headed by James Franck, in solving this question took the path of theoretical interpretation of the already known absorption spectra of diatomic molecules. Namely, the structureless continuous part observed in the absorption spectra of many molecules even in the state of a rarefied gas was attributed to the process represented above in (1). As justification for this assertion, the consideration was advanced that the continuous, uninterrupted character of the absorption spectrum indicates that the molecule can take up energy of any magnitude, beginning from a certain limit. Such a capacity for continuous

* The content of an address delivered by the author at the Second D. I. Mendeleev Reading on February 21, 1947, at Leningrad University, at a joint meeting of the Leningrad branch of the All-Union Chemical Society named after D. I. Mendeleev and the Academic Council of the University.

changes in the internal energy of the molecule—especially a simple diatomic one—does not in principle possess. According to the quantum theory, which asserts the discreteness of energy states or energy levels, the energy of molecules can assume only a series of quite definite discrete values. Consequently, these scientists concluded, the energy of light quanta of any magnitude, absorbed from the continuous part of the spectrum, must lead to a violation of the integrity of the molecule, to the rupture of the valence bond between the parts of the molecule. Then, indeed, any excess of energy over the work of breaking the bond can be absorbed by the molecule, since it is converted into the energy of translational motion of the flying-apart parts $\vec A$ and $\vec B$; kinetic energy, however, is not subject to those quantum restrictions which are imposed on the energy states of an undisturbed molecule.

The development of the problem indicated above was begun by us at the Optical Institute, in the building of Leningrad University, almost simultaneously with foreign scientists. However, unlike them, we followed an independent path of directly experimentally detecting the decomposition products in the very act of photochemical dissociation or, so to speak, in statu nascendi. My report is devoted to this topic and will not adhere to the chronological sequence of the development of our investigations.

1. Of the simplest diatomic gases, oxygen is the most accessible to radiation action, since it exhibits the strongest absorption of light in the region of short ultraviolet wavelengths. This absorption makes air opaque to ultraviolet rays shorter than 1850 Å.

Calculations of the energy required to break the bond between the two atoms in an oxygen molecule, and analysis of the structure of the absorption spectrum of oxygen, lead to the conclusion that the opacity of atmospheric oxygen is due to the fact that, beginning with the wavelength 1750 Å, the incident light splits $O_2$ molecules into atoms, according to the process:

\[ h\nu + \cdot O_2 \cdot \longrightarrow \underset{{}^3P}{\cdot O \cdot} + \underset{{}^1D}{\ddot O}. \tag{2} \]

The separately written dots indicate that the electrons are unpaired. Indeed, the oxygen molecule is paramagnetic, since it has two unpaired electrons, while the oxygen atom in its normal state possesses two free valences. The pair of dots above the second oxygen atom indicates that in it the valence electrons are paired. Below are given the spectral symbols of the electronic states of the atoms.

The peculiarity of the action of light, in comparison with other agents, consists in the fact that the oxygen molecule is split by it not into two

ordinary atoms $\mathrm{O}+\mathrm{O}$, but one of the atoms $(\ddot{\mathrm O})$ is obtained at a higher energy level and carries an energy reserve of 46 kcal. Both oxygen atoms are reactive, although in the valence respect the second atom ought to have a lower ability to react owing to the pairing of the valence electrons, which is absent in the first (normal) oxygen atom.

The photoprocess (2) is of practical interest because it takes place in the upper layers of the atmosphere under the action of the Sun’s short ultraviolet radiation. Thus this radiation, which does not reach the lower layers, is absorbed. The oxygen atoms formed, attaching themselves to oxygen molecules, lead to the appearance in the upper parts of the atmosphere of a layer of ozone, which serves as a protective screen that does not allow longer ultraviolet waves, also harmful to organisms, to reach the surface of the Earth. Hence the importance of this simplest photochemical reaction for the preservation of life on Earth is clear.

Fig. 1. Quartz reaction vessel with a hydrogen discharge tube.

In our investigations we set ourselves the aim of directly detecting free oxygen atoms by their reaction with other gases, such as, for example, $\mathrm{H}_2$ and CO, under conditions of low pressures (of the order of 0.01 mm Hg). At such pressures the free path of an oxygen atom is so large that the interpretation of the reactions is greatly simplified. Fig. 1 shows the apparatus used. A discharge tube filled with hydrogen served as the source of the short ultraviolet radiation that dissociated oxygen. This radiation, through the thinnest bulb (wall thickness $10\,\mu$) made of quartz, transparent down to a wavelength of 1500 Å, entered the lower part of the apparatus, which was filled with oxygen[^1].

The pressure of the oxygen, amounting to several hundredths of a millimeter of mercury, was measured by a corresponding sensitive manometer.

If the vessel is carefully freed from contamination of organic origin, the pressure of the oxygen, when it is illuminated by short ultraviolet radiation, undergoes only a slight change, since the oxygen atoms formed rapidly recombine into molecules on the quartz walls, as a result of which the stationary concentration of free atoms in the vessel is very small. However, if...

if substances are introduced into the vessel that do not react with molecular oxygen but are capable of binding oxygen atoms—for example, if a layer of sprayed silver is applied to the side walls—then under illumination there occurs a continuous consumption of oxygen, detected by a fall in pressure. The entire vessel is immersed in liquid air, since the low temperature facilitates fixation of the volatile products on the walls, preventing them from spreading into the illuminated zone and undergoing secondary photochemical reactions, which would greatly complicate the interpretation of the experiments.

Similarly, a fall in pressure takes place if adsorbed CO molecules are present on the walls of the vessel (Fig. 2). If hydrogen is admixed with oxygen at the same low pressure, there is likewise a fall in pressure until the supply of molecular hydrogen is exhausted (Fig. 3). A detailed kinetic study carried out on this apparatus by my collaborators G. G. Neuymin, B. V. Popov, and M. M. Pavlyuchenko—an analysis of which cannot be dealt with here—has led to a certain, fairly clear picture of the individual stages through which the reaction of oxygen atoms develops. In particular, from analysis of the data one may arrive at the conclusion that an oxygen atom rich in energy, \(\dot O\), enters more readily into reaction with an \(\mathrm{H}_2\) molecule than does a normal atom. However, this conclusion, based on measuring the rate of pressure change, i.e. on the usual method for studying the kinetics of gas reactions, cannot be regarded as convincing, like most conclusions based only on the analysis of formulas of kinetic regularities.

Fig. 2. Oxidation of CO molecules adsorbed on quartz by atomic oxygen at the temperature of liquid air (the initial values of the pressures of \(\mathrm{O}_2\) and CO are marked by vertical arrows; the latter is oxidized to \(\mathrm{CO}_2\)).

The use, for splitting the molecule, of shorter wavelengths, i.e. of quanta of greater magnitude, exceeding the threshold value \(h\nu_0\) (needed only for rupture of the valence bond), makes it possible within certain limits to vary the kinetic energy with which the fragments of the molecule fly apart, i.e. the oxygen atoms in reaction (2). However, the limited range of transparency of quartz did not permit, in the present cycle of investigations, changing

the kinetic energy of the atoms flying apart over a sufficiently large range of values, as could be done in the investigations described below.

In addition to the manometric method for studying reactions of free atoms—which has the disadvantage of being cumulative, recording only the overall result of the reaction—we used, in the case of other objects, special substances that serve as selective indicators for free atoms or for organic radicals appearing in the illuminated zone. Such indicators are: 1) metal oxides that change or acquire color under the action of free atoms, as, for example, molybdenum trioxide, which turns blue under the action of atomic hydrogen²; 2) thin films of metals that disappear as a result of reaction with free aliphatic radicals³; and, finally, 3) the method developed by us of luminescent substances whose luminescence is quenched under the action of certain atoms and radicals⁴.

Fig. 3. Oxidation of hydrogen by atomic oxygen.

1. All the methods indicated above are, in essence, “blind” methods, in the sense that they make it possible only indirectly to judge the formation of free atoms and radicals in the illuminated zone. Therefore, the splitting of molecules into luminous particles, discovered by us at the Optical Institute approximately 22 years ago⁵, has acquired great importance.

In this phenomenon the flying-apart parts of the decomposing molecule are observed directly not only in statu nascendi but, more importantly, in statu luminescendi, i.e., they are made visible by the characteristic radiation they emit.

The essence of the phenomenon consists in the following photoprocess, which may be represented schematically as follows:

\[ h\nu^{*} + \mathrm{A}-\mathrm{B} \rightarrow \mathrm{A}^{*}\!\cdot + \cdot\mathrm{B} \tag{3} \]

\[ \mathrm{A}^{*}\!\cdot \rightarrow \mathrm{A}\!\cdot + h\nu_{A}, \]

whereupon

\[ h\nu^{*} - h\nu_{A} \gg D, \]

where \(D\) is the dissociation energy of the molecule \(\mathrm{AB}\). In other words, under the action of a light quantum \(h\nu^{*}\) of sufficiently large magnitude, the molecule \(\mathrm{AB}\) dissociates into two particles flying apart (atoms or radicals), one of which (\(\mathrm{A}^{*}\)) contains an excess of energy, denoted by an asterisk, and then emits it in the form of a photon \(h\nu_{A}\), characteristic of the spectrum of such a particle.

To observe this phenomenon it is necessary to act with the short ultraviolet radiation of a spark discharge (wavelengths shorter than 2500 Å). As a result, the atoms or radicals liberated in the decomposition become “labeled” in the sense that it becomes possible to trace directly not only their appearance, but also their subsequent fate.

The splitting of molecules into luminous particles marked the beginning of “emission” photochemistry and provided a new method for studying the photokinetics of reactions in gases. The first objects of study were halide salts in the vapor state. Volatile salts of metals of all groups of the periodic system were investigated: the halides of Na, Ag, Tl, Hg, Sb, and others. This stage of our work is set forth in detail in the brochure Photochemistry of Salt Vapors, published in 1934[^5].

As an illustration of the selectivity of the action of light in photodissociation, I shall cite a result obtained for AgJ vapors. The silver atom is split off from the AgJ molecule, according to scheme (3), in an energetically high state Ag* and emits a spectrum characteristic of the silver atom:

\[ h\nu^{*} + \mathrm{AgJ} \to \mathrm{Ag}^{*} + \mathrm{J}, \]
\[ \mathrm{Ag}^{*} \to \mathrm{Ag} + h\nu_{\mathrm{Ag}}. \tag{4} \]

The Ag* atom may be in either of two excited energy levels and, correspondingly, emit two spectral lines 3281 and 3383 Å (forming a doublet), the quanta of which differ from one another by an interval of 0.1 eV, i.e. by only 2 kcal. It turns out that, when the AgJ molecule is split, the silver atom can be separated at will in one or the other of these two states, depending on the wavelength of the ultraviolet light employed.

Fig. 4. Emission spectrum of Ag atoms arising during the photodissociation of AgJ vapors.

In the photographs of Fig. 4 it is seen that, when illuminated by the light of a Ca spark, silver atoms glow predominantly in the lower of the two excited states, which is manifested by the appearance of the emission of one spectral line, 3383 Å, associated with this level. Under the action of light of shorter wavelengths, contained in the radiation of a Pb spark, the second line of the silver atom, 3281 Å, belonging to the higher of the two levels, is emitted predominantly. When the light of a Zn spark acts, however, splitting of the AgJ molecule occurs into both excited Ag* atoms[^6].

The photograph presented testifies to the very fine selectivity of action accessible only to the optical method of excit—

action on the molecule. In an electric discharge or under thermal excitation we would obtain simultaneously Ag** atoms in both states.

For demonstrating luminescent photodecomposition, the most convenient objects are volatile mercury salts, such as, for example, mercuric chloride \(\mathrm{HgCl}_2\), whose vapors are produced in vacuum by slight heating (to \(60\text{–}70^\circ\)). Under the action of the short ultraviolet radiation of a spark, passing through a quartz vessel in the form of a thin beam, the path of this beam in the vapors of the mercuric chloride becomes visible by its bright green luminescence. Upon spectral resolution of this luminescence, the spectrum obtained is not that of the mercury atom, but of the diatomic molecule or radical \(\mathrm{HgCl}\) (Fig. 5).

Fig. 5. Emission spectrum of the \(\mathrm{HgCl}\) radical arising during photodecomposition of \(\mathrm{HgCl}_2\) vapors.

An analogous luminescence, both in color and in spectral composition, is observed if ultraviolet light is applied to the vapors of an organomercury compound \(\mathrm{H}_3\mathrm{C}\!-\!\mathrm{Hg}\!-\!\mathrm{Cl}\)*).

The identity of the emission spectra in these two cases leads to the following interpretation of the photodissociation process:

\[ h\nu^{*} + \mathrm{RHgHal} \longrightarrow \mathrm{R} + \mathrm{HgHal}^{*}, \]

\[ \mathrm{HgHal}^{*} \longrightarrow \mathrm{HgHal} + h\nu_{\mathrm{HgHal}}, \tag{5} \]

where \(\mathrm{HgHal}^{*}\) is the luminescent radical excited in the decomposition process, and \(h\nu_{\mathrm{HgHal}}\) is the quantum of the spectrum emitted by it.

Knowing the supplied energy \(h\nu^{*}\), absorbed by the molecule \(\mathrm{AB}\), on the one hand, and, on the other, the energy of the quantum \(h\nu_{\mathrm{A}}\) emitted by the atom or radical \(\mathrm{A}\), we obtain from the difference of these two quantities a direct estimate of the energy that was expended in breaking the valence bond in the molecule, according to the reaction scheme (3). Such a purely optical estimate of the bond strength does not give the exact value of the heat of dissociation, but only its upper limit. It turns out that, in order to rupture a molecule by the action of light, it is not sufficient to supply a quantum exactly corresponding to the bond strength, which can be determined from thermochemical data; it is always necessary

* In fact, the experiment was carried out with the compound \(\mathrm{H}_3\mathrm{CHgJ}\).

impart to the molecule a certain, sometimes even considerable, excess of energy, i.e. act with a quantum of greater magnitude. Of course, in order to obtain a luminous product, the magnitude of the acting quantum \(h\nu^*\) must be increased in comparison with the quantum \(h\nu\) of scheme (1) by the value of the energy of the emitted quantum, i.e. by the excitation energy of particle \(A\).

Fig. 6. A powerful hydrogen discharge tube for photodissociation of molecules in a vacuum ultraviolet region; mounted in the drawing is a photograph of the luminescence observed along the path of the beam.

Halide salts of metals and organometallic compounds can decompose with the liberation of a luminous radical under the action of accessible ultraviolet radiation from a spark discharge between metallic electrodes. To effect luminous photodecomposition of the firmly built molecules of ordinary compounds, still shorter ultraviolet radiation is required, strongly absorbed, as we mentioned above, by the oxygen of the air.

The advance into the region of such far ultraviolet radiation was accomplished by me in work with G. G. Neŭmin.^8

A peculiar “photon gun” was constructed, emitting large-caliber photons, each of which was equivalent to an energy of the order of 150–200 kg-cal. Fig. 6 schematically shows a hydrogen discharge tube with a power of 10 kW, which served as the light source. The radiation passed through a thin fluorite window and entered directly into the space filled with vapors of the compound under investigation. Incorporated into the diagram of the figure is a photograph of the luminescence observed along the path of the short ultraviolet radiation. This luminescence, as spectral investigation showed, belongs to free radicals OH, NH, CN, etc., flying out from the cleaved molecules of water, alcohols, organic acids, ammonia, and other gaseous compounds. A table containing the principal data obtained in this investigation is given below.

The structure of the spectrum of the radicals liberated upon the decomposition of molecules shows that these diatomic particles possess anomalously high vibrational and rotational energies, i.e. that they are anomalously “hot.” Thus, part of the energy of the quantum absorbed by the initial compound is transformed during decomposition not only into the excitation energy of the luminous radical, but passes

also into the vibrations of its atoms and the rotation of the diatomic radical as a whole. The “temperature” of such an anomalously “hot” particle, formed as a result of the decomposition of the initial compound, amounts to several thousand degrees, although the photochemical reaction is carried out at room temperature.

Conventionally, this process may be written by the following scheme:

\[ h\nu + \mathrm{ROH} \to \mathrm{R} + \mathrm{OH}^{*}, \]

where the signs \(\sim\) and \(\odot\) denote an excess of vibrational and rotational energy in the particle being split off.

The method of luminous photodissociation therefore makes it possible to trace the dynamics of the processes of transformation of the absorbed light energy inside the molecule.

1. In addition to the application of the method indicated above, it also gives the researcher the possibility of studying the kinetics of elementary intermolecular reactions by tracing the subsequent fate of the particle liberated during photodissociation. By adding to vapors undergoing photodissociation an extraneous gas, we observe quenching of the luminescence with increasing pressure of the admixed gas; moreover, the magnitude of the quenching depends both on the nature of the luminous particle and on the nature of the added gas.

The quenching of the luminescence is caused by the fact that the excited particle, ejected from the initial molecule during photodissociation, undergoes collisions with molecules of the added gas. The frequency of these collis—

\(^*\) \(D\) is the dissociation energy of the corresponding valence bond, and \(E\) is the excitation energy of the luminous radical.

collisions increases with increasing gas pressure. If the excited particle produced in photodissociation manages to emit light in the interval between two successive collisions with gas molecules, then no weakening of the luminescence will occur. However, as the pressure increases, collisions become more frequent, and the excited particle undergoes a collision before it is able to emit light.

Depending on the partner, the collision either is accompanied by deactivation of the particle, i.e., by loss of its excitation energy, or it is not.

In the first case, the excitation energy of the particle is distributed among various types of vibrations if the colliding particles are molecules, and also passes into the translational energy of the particles. In other words, in quenching there occurs a conversion of the store of excitation energy of the particle into thermal energy. However, if the added gas is capable of entering into a reaction with the particles separated out in photodissociation, then the quenching of the luminescence should undoubtedly be attributed to the circumstance that the luminous particle disappears as a result of a chemical reaction with a molecule of the added gas. Such is the nature of the quenching of the luminescence of the sodium atom emitted from the NaJ molecule in an atmosphere of gaseous iodine.^9 The reaction may be represented by the following scheme:

\[ \begin{aligned} h\nu+\mathrm{NaJ} &\to \mathrm{J}+\overset{\to}{\mathrm{Na}}^{*},\\ \overset{\to}{\mathrm{Na}}^{*}+\mathrm{J}_{2} &\to \mathrm{NaJ}+\mathrm{J}. \end{aligned} \tag{6} \]

Especially valuable in our method is the circumstance that, by carrying out the photodissociation of the NaJ molecule with different ultraviolet wavelengths, we impart to the sodium atom \(\mathrm{Na}^{*}\) different kinetic energy, conventionally denoted in (6) by the short arrow above the atom \(\mathrm{Na}^{*}\). Consequently, by measuring the quenching as the wavelength of the dissociating radiation is varied, we obtain the possibility of measuring the change in the “yield” of the reaction indicated above as a function of the translational energy of only one partner in the chemical interaction, namely the sodium atom. In other words, in this method we have the possibility of controlling—and, moreover, very finely—only one factor that is significant for the activation of a chemical transformation. Such selective control of the mechanism of a chemical reaction is impossible in the ordinary method of carrying out a reaction by heating the reacting mixture, since with such a method of supplying energy all degrees of freedom of intramolecular and intermolecular motion are activated simultaneously and chaotically.

The systematic pursuit of this direction of research has shown, for the systems studied by us, that an increasing excess of kinetic energy of the atom promotes its reaction when, for the rearrangement of valence bonds, it is necessary to overcome a certain

energy barrier. This was observed in the case of quenching the glow of Tl* (from TlJ) by addition of CO₂ (see Fig. 7). On the other hand, the independence of the reaction yield from the velocity of translational motion of the atom was established for those cases where, evidently, there is no chemical reaction between the colliding particles and the quenching of the atom’s glow must be attributed to the physical process of removal and dissipation of the excitation energy by the inert collision partner. Such a case occurred for Na* + J and Tl* + N₂ (or CO) (Figs. 7 and 8).

Fig. 7. Change in the yield of the reaction of quenching a luminous Tl* atom by foreign gases, as a function of its kinetic energy or velocity \((E_0\) is the kinetic energy imparted to the atom upon absorption of threshold-value dissociating quanta).

Fig. 8. Change in the yield of the reaction of quenching a luminous Na atom by iodine molecules and atoms, as a function of its kinetic energy.

However, for those reactions which do not require additional thermal activation energy, as, for example, in the case of the above-mentioned reaction of the sodium atom with an iodine molecule, we found that the reaction yield falls rapidly with increasing kinetic

energy and velocity of the translational motion of the sodium atom obtained as a result of the photodecomposition of $\mathrm{NaJ}$ (Fig. 8); the same occurs for the reactions $\mathrm{Tl}^{*} + \mathrm{J}_{2}$ and $\mathrm{Tl}^{*} + \mathrm{O}_{2}$ (Fig. 7).

The decrease in reaction yield with increasing velocity of the translational motion of one of the partners, established by us, is of great fundamental significance. This fact shows that, for the rearrangement of atoms in a chemical reaction to take place, certain optimal conditions of approach must be observed, as well as a certain duration of interaction. If, as a result of an increase in the energy of translational motion, the approach of the reacting particles becomes excessive and the time spent at the optimal distance becomes exceptionally short, then, as our experiments show, the reaction yield falls, i.e., the reaction has a lower probability. These results were obtained in a series of investigations with the active participation of N. A. Prilezhaeva[^10].

No less interesting results were obtained by the same method of sensitized photodecomposition in the case when a diatomic radical is formed, possessing an excess of vibrational energy, i.e., energy of vibration of the constituent atoms. The changes occurring in the spectrum of the radical upon the addition of a foreign gas give a clear idea of the role of the excess of vibrational energy in chemical interaction. For this purpose we tested the luminous radicals $\mathrm{HgHal}^{*}$, formed during the photodecomposition of mercury halides in the gaseous state. Photodissociation of the halide vapors was carried out in the presence of oxygen or gaseous ammonia. The work of S. P. Tibilov, who studied this object, established that radicals possessing an excess of vibrational energy obtained upon photodissociation are more capable of reaction than those radicals which do not possess such an excess[^11]. At the same time it was shown that the presence of a greater or lesser electronic excitation energy in the radical is by no means decisive for the course of the reaction. Indeed, the radical $\mathrm{HgJ}^{*}$ can be obtained not only in the lowest possible excited state for it, but also at a higher level, exceeding the preceding one by an energy of 23 kcal; nevertheless, radicals brought to such a high level proved to be considerably less capable of reaction than radicals at a lower excited level. It follows from this that the mere presence of an excess of electronic excitation energy is not the determining factor in the reactions we studied, and that our conclusions are applicable also to those conditions under which radicals are obtained thermally in the zone of an ordinary reaction and do not possess excitation energy.

The investigation carried out by us ultimately made it possible to separate and study individually the role, on the one hand, of translational energy and, on the other, of vibrational energy in the activation of a chemical reaction.

1. When we spoke above of the possibility of tracing, by means of our method, the details of intramolecular dynamics, we had in mind the simplest diatomic molecules undergoing dissociation under the action of light. The extension of our investigations to polyatomic molecules, which was partly touched upon in the preceding section, has led us to a new phenomenon of fundamental importance, which may be called internal compensation of an energy deficit. The light absorbed by a molecule acts as a kind of trigger mechanism, releasing stores of internal energy. As a result, small quanta of light energy can produce effects considerably greater than would be expected from their magnitude.

Such effects of compensation for an energy deficiency at the expense of internal resources are possible, for example, when the rupture of one valence bond in a molecule under the action of light induces a polyvalent atom participating in this bond to pass to a lower valence state. The transition from a higher valence to a lower one, accompanied by pairing of valence electrons, releases significant portions of energy which can be used in an intramolecular process. Thus, for example, the transition of a carbon atom from the tetravalent state \(^{IV}\mathrm{C}\) to the divalent \(^{II}\mathrm{C}\) liberates about 80 kcal, and in the transition \(^{IV}\mathrm{Sn} \to {}^{II}\mathrm{Sn}\) 30 kcal are released. We carried out experiments with halides of the latter metal in the gaseous state. It was found that, under the action of ultraviolet light, vapors of \(\mathrm{SnJ}_4\) give a luminescence belonging to \(\mathrm{J}_2\) molecules[^12]. A detailed experimental analysis of the origin of this luminescence showed that we are dealing here with photodecomposition of the \(\mathrm{SnJ}_4\) molecule, but that in the emitting state this time not an atom but a whole molecule, \(\mathrm{J}_2^*\), is separated, according to the following reaction scheme:

\[ h\nu + \mathrm{SnJ}_4 \to \mathrm{SnJ}_2 + \mathrm{J}_2^* . \tag{7} \]

This type of photodecomposition took place only when the rarefied vapors of tin iodide were additionally heated to a higher temperature. From the magnitude of the temperature coefficient of the luminescence it was established that the process depicted above required an additional thermal activation energy of about 10 kcal. The very fact that a ready-made molecule is separated during photodecomposition was not in itself new, since the possibility of decomposition of acetone into \(\mathrm{H}_3\mathrm{C}—\mathrm{CH}_3\) and \(\mathrm{CO}\) had already earlier been inferred on the basis of indirect conclusions. In our experiment such decomposition was demonstrated directly by a direct experiment. However, our experiment proved not only the possibility of a rearrangement of valence bonds in a complex molecule under the action of light. From the values of the strengths of the valence bonds \(\mathrm{Sn}—\mathrm{J}\) and \(\mathrm{J}—\mathrm{J}\) it followed that the energy of the quantum causing this process was clearly insufficient. The deficit amounts to approximately

23 kcal and could be covered only at the expense of the molecule’s internal store of electronic energy. The only source of internal energy could only be the transition of the central tin atom, in the process of photodecomposition, from the tetravalent to the divalent state; consequently, scheme (7) must be replaced by the following, more accurate interpretation of the observed photoreaction:

\[ h\nu + {}^{\mathrm{IV}}\mathrm{SnJ}_4 \longrightarrow \left( \begin{matrix} \mathrm{J} & & \mathrm{J}^{*}\\ & \mathrm{Sn} & \\ \mathrm{J} & & \mathrm{J} \end{matrix} \right) \longrightarrow {}^{\mathrm{II}}\mathrm{SnJ}_2 + \mathrm{J}_2^{*}. \tag{8} \]

1. The photodecomposition of the ionic molecule \(\mathrm{Na}^{+}\mathrm{J}^{-}\) into neutral atoms, described in the third section, indicates that, during photodecomposition, a redistribution of the electronic charge occurred, and an intramolecular transfer of an electron from the anion \(\mathrm{J}^{-}\) to the cation \(\mathrm{Na}^{+}\). The inverse process of rupture of a homopolar valence bond with the formation of ions was discovered by us in the photodecomposition of the iodides of Tl, Ag, and others, which in the gaseous state are molecules with a covalent bond.\(^{13}\) Under the action of light, along with the luminescent photodecomposition described in Section 3, the following photoreaction took place with a larger yield:

\[ h\nu + \mathrm{TlJ} \longrightarrow \mathrm{Tl}^{+} + \mathrm{J}^{-}. \tag{9} \]

The \(\mathrm{Tl}^{+}\) and \(\mathrm{J}^{-}\) ions liberated in this process were detected not only by the appearance of photoconductivity in the illuminated vapors, but were also recorded by a mass spectrograph. The phenomenon was observed in gallium halides, and also in the vapors \(\mathrm{AgJ}\) and \(\mathrm{BiJ}_3\).

This cycle of investigations led us, considerably later, to the question of ionization in photochemical reactions. It is of interest to establish the possibility of detaching the simplest ion—namely, the proton—from complex organic compounds possessing acid groups, \(-\mathrm{OH}\) or \(-\mathrm{COOH}\). For detaching protons from such acid compounds under conditions of the gaseous state, i.e., from free molecules, it would be necessary to expend considerable quanta of light belonging to the short ultraviolet region of the spectrum, inaccessible to experiment. Therefore we turned to the condensed phase and attempted to apply to complex organic systems an analogous method of fine analysis of the reaction mechanism using luminescent products. The experiments, conceived during the war years, were carried out with the help of A. V. Karyakin in 1946.\(^{14}\) It proved possible, in molecules of aromatic compounds bound by a hydrogen bond, to demonstrate the possibility of splitting off and transferring a proton under the action of near ultraviolet radiation.

Let \(\mathrm{A{-}COOH}\) be an organic acid capable of donating a proton \(\mathrm{H}^{+}\), and \(:\mathrm{N}<\mathrm{B}\) a fluorescent organic base capable of accepting a proton with a change in the color and spectrum of fluores-

cence. As the proton donor \((\mathrm{A—COOH})\), oxalic, succinic, and terephthalic acids were taken, transparent in the region where the absorption spectrum of the base—the proton acceptor—lies; acridine was chosen as this acceptor. Upon joint sublimation of both components in vacuum onto a surface cooled with liquid air, the normal violet fluorescence of acridine changes into the bright green fluorescence characteristic of an acridine molecule that has accepted a proton. This process of proton transfer proceeds in the dark and does not require photoactivation. It is represented in the following scheme by an arrow directed from top to bottom:

\[ \begin{array}{c} \mathrm{A—COO^{-}H^{+}\ldots :N \le B} \quad \begin{array}{c} \text{violet fluor.} \end{array} \\[-0.2em] \begin{array}{c} \text{dark} \end{array} \ \downarrow \quad \uparrow h\nu_{\mathrm{B}} \\[-0.2em] \mathrm{A—COO^{-}\ldots H^{+}:N \le B} \quad \begin{array}{c} \text{green fluor.} \end{array} \end{array} \tag{10} \]

If this system, fluorescing with green light, is acted upon by quanta of ultraviolet light \(h\nu_{\mathrm{B}}\) in the region of the absorption spectrum of the acridine cation \(\mathrm{H^{+}:N \le B}\), then a restoration is observed of the original violet fluorescence, belonging to the initial acridine molecule \(:\mathrm{N \le B}\). When the acting ultraviolet light is switched off, the violet fluorescence again becomes green, and this phenomenon can be repeated many times without any residual changes. The measured spectra established beyond doubt that the observed phenomenon is caused by a reversible displacement of the proton, as shown in scheme (10).

Remarkable is not only the fact that, under the action of light, an elementary particle more massive than the electron is mobile, but, in particular, the fact that this displacement is caused by small ultraviolet photons. As in the preceding photoprocess described above, here we have a new example of an internal compensation of the energy deficit, possible in this case owing to the conjugation of the participants into a single intermolecular compound with common energetics. The missing energy is supplied through the formation of a covalent \(\mathrm{O—H}\) bond upon the reunion of the proton with the original acid anion \(\mathrm{A—COO^{-}}\).

1. In summing up my exposition, I note that the photoprocesses considered here are of interest not only within the framework of laboratory kinetics of photochemical reactions. I have already mentioned the significance of the photodissociation of oxygen for the regime established in the upper layers of the atmosphere.

Not only for the present state of the Earth’s atmosphere does the action of the Sun’s ultraviolet light have primary importance. It may be supposed that the radiation effect of solar irradiation on the original atmosphere of the terrestrial globe, which consisted of water vapor and carbon dioxide and contained no oxygen, was also exceptionally strong. The possibility is not excluded that

the simplest carbohydrate—formaldehyde, which marked the beginning of the organic world—was first synthesized precisely by the direct action of the sun’s short ultraviolet radiation.

The latter phenomenon that we have described—the phototransfer of a proton under the action of much smaller quanta of ultraviolet light—is related to one of the links in that complex chain of processes which determine the splitting of water molecules and the addition of hydrogen to the $CO_2$ molecule, i.e., to the process of photosynthesis that takes place in the living leaf of a plant.

It seems to me that from what has been said above, the direct connection becomes evident between the branch of our investigations touched upon here and life, in the direct and figurative sense of the word. I must, moreover, express my deep gratitude for the high honor shown to me and to my pupils by the honorable invitation to deliver a report at the readings named for our great Russian scientist D. I. Mendeleev. I attribute this honor to that direction of photochemical science which has been successfully realized in our Soviet country, which has created exceptional opportunities for scientific work.

The attention shown to me by granting me the floor in this honorable and solemn setting inspires me and the collective that I head to new labors and new achievements, to the glory of Russian science and for the benefit of our great motherland!

REFERENCES CITED

1. G. Neumin and B. Popov, ZhETF 5, 87 (1935); B. Popov and G. Neumin, ibid. 5, 440 (1935); B. Popov, Diss. Inst. Chem. Phys. (1940); M. Pavlyuchenko, Zh. fiz. khim. 14, 605, 877 (1940).
2. G. Neumin, DAN 16, 453 (1937).
3. N. Prilezhaeva and A. Terenin, Zh. fiz. khim. 8, 111 (1936).
4. V. Gachkovsky and A. Terenin, Izv. AN OMEN, 805 (1936); Acta phys. chim. URSS 7, 521 (1937).
5. A. Terenin, Optical Dissociation of Salt Molecules, Tr. GOI, issue 40 (1928); A. Terenin, Photochemistry of Salt Vapors [Problems of Modern Physics, issue XII] GTTI, 1934.
6. A. Terenin, Physica 10, 209 (1930).
7. A. Terenin and N. Prilezhaeva, ZhETF 5, 599 (1935).
8. G. Neumin and A. Terenin, Izv. AN OMEN, 529 (1936).
9. A. Terenin and N. Prilezhaeva. Tr. GOI 7, issue 66 (1931); Zeits. f. phys. Chem. (B) 13, 72 (1931).
10. N. Prilezhaeva, Zh. fiz. khim. 5, 1239 (1934); Sov. Phys. 2, 351, 367 (1932); Acta phys. chim. URSS 2, 647 (1935).
11. S. Tibilov, Zh. fiz. khim. 10, 1 (1937); Acta phys. chim. 7, 171 (1937).
12. A. Terenin and R. Chubarov, Zh. fiz. khim. 10, 636 (1937); Acta phys. chim. URSS 7, 1 (1937).
13. A. Terenin and B. Popov, Sow. Phys. 2, 299 (1932); R. Kagan, ZhETF 5, 811 (1935).
14. A. Terenin and A. Karyakin, DAN 58, 425 (1947).