INTERMOLECULAR ENERGY TRANSFER IN THE PHENOMENON OF SENSITIZED LUMINESCENCE OF ORGANIC SYSTEMS. II*)

A. N. Terenin and V. L. Ermolaev

The nonradiative transfer of a quantum of excitation energy, imparted to a particle \(A\) in the act of absorption of light, to another nearby particle \(B\) is of great fundamental interest in connection with the possibility of energy migration in photoreactions of complex organic molecules[^1]. Such transfer is observed directly in the phenomenon of sensitized luminescence of particles \(B\), observed in a mixture of particles \(A\) and \(B\) upon excitation by light absorbed only by particles \(A\). In this case, along with the emission of quanta \(h\nu_A\) by excited particles \(A\), quanta \(h\nu_B\), characteristic of excited particles \(B\), are emitted, while the exciting light does not act directly on \(B\). Schematically the process is represented as follows: \(A^* + B \to A + B^*\), where the asterisk denotes the excited electronic state of the particles. The energy level of excitation of \(B\) must be lower than the level of \(A\), or exceed it by an amount accessible to additional thermal excitation at the temperature of the experiment. Such sensitized excitation of particle \(B\) can occur during the time \(\tau\) for which \(A\) remains in the excited state.

Experiment and theory show that two different types of transfer of electronic excitation energy between molecules are possible. According to one of them, energy transfer occurs when the particles approach to “contact,” i.e., to a distance between their centers equal to the sum of the radii of the colliding molecules. Averaged, conditional radii of molecules, determined by known statistical methods[^2], give a certain measure of the “impenetrability” of the volume occupied by the electron shell of the molecule in its normal unexcited state. It is assumed that upon excitation of a complex molecule its radius changes comparatively little. However, intermolecular transfer of energy, especially between molecules of the same kind, can take place—

*) For the first paper see Usp. Fiz. Nauk 43, 347 (1951).

…take place also at distances considerably exceeding the radii indicated above, but by a different mechanism. This mechanism, which has been given the name “inductive” (J. Perrin, S. I. Vavilov, T. Förster[^1]), consists in the fact that the electromagnetic field of an electronic oscillator, with which an excited molecule may be identified, as a result of resonant interaction with the unexcited oscillator of a second molecule induces in it forced oscillations of the same frequency. Such a process of energy “pumping” occurs at distances within the wavelength emitted by the oscillator, i.e., at distances certainly exceeding the linear dimensions of molecules[^1]. Whereas this latter, inductive type of transfer in sensitized luminescence does not require that particles approach to the point of contact and can occur for molecules fixed at distances large in comparison with molecular dimensions in viscous and solid media, the former, kinetic type of transfer occurs only under conditions in which particles have freedom of motion in a gas-like or dissolved state during their free or diffusive path. Under these latter conditions, there undoubtedly also remains the possibility of energy transfer by the inductive mechanism between molecules passing by one another without coming into direct contact. Such a phenomenon must account for the anomalously high efficiency of transfer in sensitized luminescence, leading to a sphere of “action,” or of mutual sensitivity of molecules, with a radius considerably exceeding their dimensions.

The greatest difficulties in experiments on sensitized luminescence are caused by two circumstances. First, when complex organic molecules have an extended absorption region in the spectrum, it is almost impossible to excite selectively only molecules \(A\), without simultaneously affecting molecules \(B\). A correction for direct excitation of the luminescence of component \(B\) is usually introduced by taking into account the ratio of light absorption, i.e., the optical densities of components \(A\) and \(B\) at the concentration used in the experiment, and by determining the quantum yields of the luminescence of the components. The second and most unpleasant complication in the interpretation of the observed phenomenon as intermolecular energy transfer is introduced, however, by absorption by particles \(B\) of the light emitted by the energy donor \(A\). Indeed, when it is necessary—especially for an inductive mechanism—for the maximum of the emission band of donor \(A\) to lie close to the maximum of the absorption band of acceptor \(B\), their overlap inevitably occurs, as does the trivial excitation of \(B\) by the luminescence light from \(A\). To take into account this internal reabsorption in the system and the secondary luminescence of \(B\) caused by it is exceptionally difficult.

We shall consider successively, by sections, the progress that has been achieved in the field of sensitized luminescence since approximately 1950. The state of the question up to that time was covered in an earlier review by one of the authors[^1].

1. SENSITIZED FLUORESCENCE IN VAPORS OF ORGANIC COMPOUNDS

Primezhaeva, in a mixture of aniline* vapors + indigo3, and later, together with Klimova, in a mixture of benzene* + aniline4, observed an enhancement of the fluorescence of the second component when the first was added under the action of undecomposed ultraviolet light exciting both components. The vapor pressure of the second component was insignificant \((10^{-1}—5 \cdot 10^{-3}\ \mathrm{mm\ Hg})\); the vapor pressure of the energy donor reached \(10\ \mathrm{mm\ Hg}\). Upon the sensitized luminescence, if it occurred here, there must have been superimposed an independent effect also leading to an intensification of fluorescence upon the addition of foreign polyatomic gases—namely, the effect investigated by Neporent5. It consists in the fact that, when aromatic compounds are excited by ultraviolet light falling in the region of diffuse bands of the absorption spectrum, the fluorescence yield is small in comparison with excitation in the long-wavelength absorption band, because the absorbed quanta of high energy create in the molecule a considerable excess of vibrational energy, leading to decomposition or internal deactivation of the molecule6. The presence of a foreign polyatomic gas transparent to the exciting light considerably increases the fluorescence yield, since collisions with its molecules remove the excess vibrational energy without affecting the electronic state of excitation, and thereby stabilize the excited molecule. Of special interest is Neporent’s study of the fluorescence of β-naphthylamine vapors in the presence of benzene vapor; in this work the enhancement caused by sensitized fluorescence was clearly separated from the enhancement due to stabilization7. The addition of benzene vapor to fluorescing vapors of naphthylamine (pressure \(0.1\ \mathrm{mm\ Hg}\)) upon excitation by monochromatic light of the lines \(280.4\ \mathrm{m\mu}\) and \(265.2\ \mathrm{m\mu}\), not absorbed by benzene, causes a considerable increase in the fluorescence yield of β-naphthylamine, as shown in Fig. 1 (solid curves). Pentane has the same effect (dashed curves). The phenomenon is due to stabilization. However, upon excitation by the line \(253.7\ \mathrm{m\mu}\), absorbed by benzene vapor, the parallelism in the action of benzene and pentane is sharply disrupted (Fig. 1): the fluorescence intensity of naphthylamine increases with the pressure of the added benzene three times faster than with the pressure of pentane*), and also in comparison with excitation by two lines that do not affect benzene. This spectral selectivity of the enhancement makes it plausible to assert that, along with the stabilizing action of benzene as a polyatomic molecule, sensitized fluorescence also occurs here for the naphthylamine molecule excited by the \(253.7\ \mathrm{m\mu}\) line,

*) Calculated for an equal number of gas-kinetic collisions.

i.e., the transfer of energy to naphthylamine in collisions with excited benzene molecules. However, in this case the possibility of internal reabsorption by naphthylamine of the fluorescence light of benzene was likewise not excluded.

To achieve preferential excitation of only one of the components of the vapor mixture, Terenin and Karyakin used naphthalene as the energy donor, which in the vapor has a discrete absorption spectrum. Excitation was produced by a group of lines of a magnesium

Fig. 1. Increase in the fluorescence yield of β-naphthylamine vapor upon addition of benzene vapor (points and solid curves) and pentane (dashed curves). \(Z\) is the number of collisions of an excited molecule with molecules of the foreign gas per second.

spark at \(280\,m\mu\) in the region of the spectrum where the intense absorption bands of naphthalene are located \(^{8}\). At the same time, as energy acceptors they used such compounds which in this spectral region have a minimum in the absorption spectrum, such as, for example, acridine, phthalimide derivatives, etc. The light of wavelength \(280\,m\mu\) used for excitation does not produce in the vapors of these compounds the visible fluorescence bands characteristic of them or, in other cases, it is very weak. The addition to the vapors of these compounds, having a pressure of the order of \(50\) mm Hg, of naphthalene even at the insignificant pressure of \(0.1\) mm Hg causes the appearance or sharp intensification of their fluorescence. Separation of the vapors of naphthalene and acridine by a thin quartz wall in a special vessel showed that such

intense glow of the latter could not have been caused by reabsorption of the fluorescence light of naphthalene. An explanation of the observed flare-up by the above-mentioned effect of stabilization of acridine molecules directly absorbing the exciting light is ruled out by the fact that the vapor pressure of naphthalene was insufficient and, moreover, the presence of considerably larger amounts of benzene, tetralin, and hexane does not, under the same conditions, produce a flare-up of the fluorescence of acridine vapor. A flare-up upon the addition of naphthalene vapor was also observed for vapors of 3-aminophthalimide, for which the stabilization effect is not observed at all, even upon the addition of 500 mm Hg of pentane. On this basis one may conclude that, for the binary vapor mixtures of various organic compounds with naphthalene that were studied, a very effective transfer of energy in collisions with excited naphthalene molecules does indeed take place. In the investigation described, in addition to acridine, naphthalene-sensitized fluorescence was also obtained for vapors of 3-oxyphthalimide, 3-monomethylaminophthalimide, acridonimine, indigo blue, aluminum 8-oxyquinolate, and magnesium phthalocyanine.

2. SENSITIZED FLUORESCENCE IN LIQUID SOLUTIONS

The transfer of excitation energy in an inductive resonance interaction between unlike molecules, occurring in solutions at distances clearly exceeding their dimensions, has been convincingly demonstrated in the works of S. I. Vavilov’s school \(^{9-13}\). This transfer manifests itself as quenching of the fluorescence of solutions by absorbing substances, observed in those cases when the fluorescence spectrum of the energy donor overlaps the absorption spectrum of the nonfluorescing quencher \(^{10}\), and the lifetime \(\tau\) of the donor in the excited state is shortened. However, measurements show that the fluorescence yield falls with increasing quencher concentration faster than \(\tau\). This question was investigated in detail, both experimentally and theoretically, in a series of works by M. D. Galanin \(^{9-13}\), who in his last paper \(^{13}\) came to the conclusion that the difference between the fall of the yield and the shortening of \(\tau\) in viscous solvents can be explained without resorting to the idea of the participation of “instantaneous” quenching caused by the presence of nearby molecules that remove the energy immediately after the act of excitation.

The most convincing proof of energy transfer in inductive interaction between unlike molecules nevertheless remains the observation of sensitized luminescence of the added second component—the quencher. Meanwhile, up to now, for liquid solutions we do not have sufficiently clear results that would make it possible unambiguously to separate nonradiative energy transfer from trivial absorption by the second participant of the light emitted by the donor, and secondary luminescence. Employing-

the large thickness of the fluorescing layer used hitherto in the experiments aggravates the spectrophotometric difficulties of such a separation. M. D. Galanin\({}^{12}\) proposed a subtle method for the nonspectrophotometric detection of sensitized fluorescence, based on measuring with a phase fluorometer the change not only in \(\tau_1\) of the donor, but also in \(\tau_2\) of the acceptor as the concentration of the latter is increased. From the kinetics of sensitized fluorescence it follows that, along with the shortening of the donor \(\tau_1\), intermolecular energy transfer causes an increase in the mean duration \(\tau_2\) of the acceptor emission in comparison with its normal value under direct excitation. The measurements confirmed the increase of \(\tau_2\); however, as the author himself noted, reabsorption of the donor fluorescence light and secondary luminescence of the acceptor lead to the same course of change of \(\tau_2\), and it is difficult to separate out the portion of the effect caused by transfer. It is necessary to pass to thin layers, for which the measurements become difficult because of the weakness of the luminescence. The binary systems studied by M. D. Galanin in aqueous solution were: anthracene\({}^{*}\) + benzoflavin, benzoflavin\({}^{*}\) + rhodamine B, anthracene\({}^{*}\) + acridine orange, where here and below the asterisk denotes the primarily excited energy donor. Qualitatively, he observed a slight strengthening of the emission maximum of benzoflavin when the emission maximum of sulfuric-acid quinine, which played the role of donor, was weakened. An analogous qualitative result had earlier been obtained by Förster\({}^{14}\) for the systems trypaflavin\({}^{*}\) + rhodamine B, fluorescein\({}^{*}\) + erythrosin, trypaflavin\({}^{*}\) + methylene blue.

A check on the existence of sensitized fluorescence in dye solutions due to inductive energy transfer was carried out recently by Schmillen\({}^{15}\), likewise by measuring the decay times of the emission of the components with the aid of a phase fluorometer. The following systems were investigated: fluorescein-Na\({}^{*}\) + erythrosin in aqueous solution, trypaflavin\({}^{*}\) + rhodamine 6G and trypaflavin\({}^{*}\) + rhodamine B; the two latter systems in methanol solution. As in Galanin’s experiment, \(\tau_1\) of the energy donor decreased, while \(\tau_2\) of the acceptor increased, beginning at concentrations of the order of \(10^{-4}\) mole/liter\({}^{*}\), reached a maximum at concentrations of about \(10^{-3}\) mole/liter, and then fell because of the onset of concentration quenching (Fig. 2). The increase of \(\tau_2\) reached, for trypaflavin\({}^{*}\) + rhodamine B, a factor of two (from \(2.6\) to \(4.8 \cdot 10^{-9}\) sec). The concentration of acceptor at which \(\tau_1\) of the donor (trypaflavin) is shortened by half, \(c_{1/2}\), was, in the case of rhodamine 6G, \(6 \cdot 10^{-3}\) mole/liter, and for rhodamine B about \(8 \cdot 10^{-3}\) mole/liter. For fluorescein the concentration \(c_{1/2}\) of erythrosin proved to be \(1.5 \cdot 10^{-3}\) mole/liter, whereas the concentration of the acceptor (erythrosin) producing a “half-quenching” of the luminescence intensity

\({}^{*}\) The measurements were carried out, as in Galanin’s work, under conditions of increasing the total concentration while keeping the ratio of the component concentrations constant. The layers were likewise taken to be thick (10 mm).

fluorescein; in Förster’s experiments^14 it was \(7 \cdot 10^{-4}\) mole/liter. Thus, here too the emission yield of the donor falls more rapidly than the lifetime \(\tau_3\), as in the case of quenching by nonfluorescent compounds in Galanin’s work.

Consequently, qualitatively here as well the results agree with the presence of sensitized fluorescence at mean intermolecular distances corresponding to the conditions of inductive transfer of excitation energy. However, in this case too it is difficult to decide what fraction of the observed increase in \(\tau_2\) is due to energy transfer, and what fraction to secondary fluorescence.

Fig. 2. Fluorescence decay time of fluorescein:
\(a\) — fluorescein in \(H_2O\), layer thickness 10 mm;
\(b\) — fluorescein in \(H_2O\) according to Shimanovskii;
\(c\) — fluorescein in \(H_2O\), thin layer;
\(d\) — fluorescein and erythrosin in \(H_2O\), \(C_{\phi} = C_{\mathrm{e}}\).

The absence of a change in the form of the spectra of the components, both in emission and in absorption, argues against the possibility that association of the components occurs in solution and against explaining the effects by a noninductive process of “near” transfer. The experiments of Zanker^16, as well as of Schmilen^18, indicate that, in the presence of dimerization and a clear change in the absorption spectrum, acridine orange also exhibits a change in the form of the fluorescence spectra and, at the same time, an increase of \(\tau\) by a factor of 4.5.

Bouin and Broklehurst^19 investigated by a spectrophotometric method the fluorescence yield of mixtures of polycyclic aromatic hydrocarbons in liquid benzene, proceeding from the assumption that under these conditions association does not occur*). Into the spectra was introduced

* Measurements were made with an instrument calibrated for the number of quanta emitted by the object. The ratio of the intensities of the donor and acceptor emission spectra was measured. The relative values of the quantum yields of their luminescence were determined from the areas of the spectra.

correction for reabsorption of the donor luminescence light. Excitation was carried out, as usual, by a group of lines of the mercury spectrum at 365 mμ.

The following systems were studied: 9-phenylanthracene* + 9,10-dichloroanthracene, 1-chloroanthracene* + perylene. Figure 3 shows the observed increase in the fluorescence yield of the acceptor—perylene—and the decrease in the yield of the donor with increasing concentrations of the components at their unchanged ratio.

Fig. 3. Fluorescence intensity of a mixed solution of 1-chloroanthracene and perylene in benzene at a constant ratio of the molar concentrations of the components, 5:1. 1 — total emission of the components; 2 — fluorescence intensity of perylene, separated from the total; 3 — fluorescence intensity of 1-chloroanthracene.

Proceeding from the data obtained and having calculated the fraction of reabsorption, the authors believe that energy transfer takes place at intermolecular distances of the order of 50 Å.

In the case of the second system investigated by them, 1-chloroanthracene* + perylene, the first compound has a small luminescence yield; nevertheless it effectively sensitizes the fluorescence of the second (see Fig. 3). From this the authors conclude that the energy of a quantum absorbed by the donor can be transferred to the acceptor before internal deactivation. In such a case the energy transfer should not belong to the category of inductive transfer, but rather to “instantaneous” transfer between molecules that have approached to “contact.”

In the last work on the same question, Bowen and Livingston²⁰, along with expanding the range of objects, carried out a more careful spectrophotometric quantitative study of all the details of the phenomenon. The objects used were: 1-chloroanthracene* + perylene, 1-chloroanthracene* + rubrene, 9-cyanoanthracene* + rubrene, 9-aldehydeanthracene* + rubrene. The solvents for the first pair of compounds were liquid paraffin, benzene, chloroform; for the others—benzene.

The values of the fluorescence intensity of each of the components were measured for various concentrations*). It was established that energy transfer is more effective than would be the case with reabsorption of the radiation by the acceptor. Comparison of the results for the first system in three solvents showed that the effect is not determined by diffusion and, consequently, has features characteristic of inductive transfer. From the concentration values for the donor quenching constant, whose magnitude is about \(6—7\cdot 10^{-3}\) mole/liter, the authors find, for the mean interatomic distance for transfer, \(41\) Å for the first system, \(38\) Å for the second, and \(84\) Å for the third.

By means of not especially convincing calculations the authors found that the probability of inductive transfer is approximately proportional to the concentration of the energy acceptor, as follows from the theory.

The absence of association of the components, in their opinion, is demonstrated by the invariance of the fluorescence spectra and by fulfillment of Beer's law in mixed solutions**).

3. MIGRATION OF EXCITATION ENERGY IN MOLECULAR CRYSTALS

The question of sensitized luminescence of an impurity of naphthacene in anthracene crystals and similar systems was the subject of a previous review\(^1\), which covered works up to 1950. In the intervening period there has appeared the monograph by A. S. Davydov, Theory of Light Absorption in Molecular Crystals, in which he examines in detail the processes of dissipation and migration of excitation energy in a crystal consisting of polyatomic molecules, refining and developing the earlier conclusions of Ya. I. Frenkel.

Essential for the possibility of energy migration in the form of a free exciton is the weak interaction of the corresponding electronic state with lattice vibrations, in contrast to a localized exciton, for which the accompanying deformation of the lattice inevitably leads to dissipation of the excitation energy into heat. To prevent the possibility of transition of a migrating exciton into a localized one and thereby create a large yield of luminescence, it is assumed that the energy level of the free exciton must lie below the level of the localized one.

Raitt’s work\(^ {21}\) is a refinement and development of already known results on the fluorescence yield of mixed crys-

*) The cuvette with the fluorescing objects was placed before the slit of a double monochromator, at whose exit there was a photomultiplier; measurements were made at wavelengths of spectral bands of the components in the region of minimal reabsorption.

**) Recently a new article by Bowen and Brocklehurst\(^ {17}\) appeared, in which they proved the presence of sensitized fluorescence in a mixed solution of 1-chloroanthracene and perylene at \(-180^\circ\)C in chloroethane.

naphthalene crystals with anthracene. Excitation by the lines 253.7 mµ and by α-particles was used.

Among recent experimental works on this question, of interest is the study by Simpson and Norforon^22 of the sensitized luminescence of a series of polynuclear aromatic hydrocarbons as impurities in anthracene. The specimens were prepared by depositing thin layers of thickness of the order of 5 microns from vapor in vacuum. In such layers the influence of reabsorption and secondary radiation was completely eliminated. For ovalene (6 rings) the luminescence and absorption spectra have the form of narrow maxima, practically separated from one another (Fig. 4).

By measuring the relative fluorescence yields of the matrix and the impurity, the authors showed that the efficiency of energy transfer decreases with increasing concentration of the ovalene acceptor from \(10^{-4}\) to \(10^{-3}\) mole/liter. For hydrocarbons with fewer than 6 rings, on the contrary, the efficiency increases with increasing concentration of the energy acceptor. The efficiency of transfer was taken to be the ratio

Fig. 4. Fluorescence spectra of ovalene in anthracene.

of the number of quanta of sensitized emission of the acceptor to the number of quenched quanta of emission of the donor. As the impurity concentration is increased, changes are observed in its spectrum.

Measurements of the lifetimes of components in mixed crystals of aromatic hydrocarbons by a phase fluorometer were carried out by Schmillen and coauthors^15,23. The decay time of the luminescence \(\tau\) of pure crystalline anthracene proved to be considerably greater than the decay time for anthracene in liquid solution, namely: \(1.3 \cdot 10^{-8}\) instead of \(0.38 \cdot 10^{-8}\) sec, in accordance with the greater luminescence yield, which confirms the concept of exciton migration. In accordance with this concept, for the systems chrysene—anthracene, anthracene—naphthalene, a shortening of \(\tau_1\) of the lumi-

cence of the host in the presence of an impurity, although not in all cases (Fig. 5). An interesting, but as yet unexplained, fact is the increase of $\tau_2$ of the energy acceptor—anthracene or naphthacene in naphthalene—with increasing acceptor concentration from $10^{-6}$ to $10^{-3}$ mole/liter upon excitation in the absorption band of the acceptor (anthracene, tetracene, Fig. 6). This increase cannot serve as an unambiguous criterion of energy migration, since reabsorption leads to the same result.

N. D. Zhevandrov24 investigated the polarization of the fluorescence of single crystals of anthracene, 9,10-dibromoanthracene, 3-dimethylamino-6-amino-phthalimide, and also of naphthacene as an impurity in anthracene. In all cases he established, for different wavelengths of the exciting light, a complete independence of the degree of polarization of the fluorescence light from the orientation of the polarization of the exciting light with respect to

Fig. 5. Decay time of the fluorescence of solid solutions:
$a$—tetracene in anthracene, emission of anthracene;
$b$—tetracene in anthracene, emission of tetracene;
$c$—pentacene in anthracene, emission of anthracene;
$d$—pentacene in anthracene, emission of pentacene;
$e$—tetracene in chrysene, emission of chrysene;
$f$—tetracene in chrysene, emission of tetracene;
$g$—tetracene in chrysene, emission of tetracene upon excitation at 4358 Å.

Fig. 6. Decay time of anthracene (exc. 365 mμ) and tetracene (exc. 436 mμ):
$a$—anthracene in naphthalene; $b$—tetracene in naphthalene.

the crystal. This is interpreted as the result of energy migration, since as a result of migration the excitation energy is trans—

is distributed among molecules of different orientation. Upon excitation by light with wavelength 436 mμ, absorbed only by naphthalene in an anthracene medium, the polarization of the fluorescence of naphthalene also proves to be independent of the polarization orientation of the exciting light. From this the author concludes that energy migration is possible between naphthalene molecules located in a foreign lattice and separated by a distance tens of times greater than their dimensions.

4. SENSITIZED LUMINESCENCE IN LIQUID SYSTEMS UNDER RADIATION ACTION

When particles and quanta with high energy values act on highly dilute solutions of intensely fluorescing organic compounds, luminescence of these dissolved molecules is observed with a large yield, although the exciting energy is known to be absorbed, for the most part, by the entire mass of the solvent. Kallmann and Furst \(^{25,26}\), who investigated this luminescence with the aid of a photomultiplier, came to the conclusion that the initially excited solvent molecules transfer their excitation energy to neighboring ones, until the migration of such an exciton in the solvent medium ends in excitation of a dissolved molecule and its emission. In the preceding review we expressed doubt as to the correctness of such an interpretation. Indeed, exciton migration in liquid media, analogous to the phenomenon established for crystals, is possible for solvents of the benzene series, which have sufficiently stable electronic excitation levels and their own fluorescence spectrum. Meanwhile, the luminescence phenomenon described is observed with equal success also in such solvents whose molecules do not possess a discrete absorption spectrum, such as, for example, \( \mathrm{CCl}_4 \), or do not exhibit fluorescence even in the short ultraviolet region and in the gaseous state, such as heptane and dioxane. This indicates that they lack the possibility of storing energy upon excitation of the electron shell, since strong interaction with the motion of atomic nuclei leads to a highly effective dissipation of the excitation energy into heat.

From our point of view, in agreement with other authors \(^{1}\), it is more plausible to interpret the luminescence considered here as a kind of electro- and chemiluminescence of dissolved fluorescing molecules due to processes of recombination and reaction of electrons, ions, and radicals, formed in liquids under the action of particles and quanta of high energy, occurring in their vicinity.

Furst and Kallmann \(^{24}\) expressed doubt as to the significance of the latter processes on the grounds that the effect of radiation luminescence depends selectively on the nature of the solvent and does not change in parallel with the efficiency of excitation of fluorescence of the same solutions by ultraviolet light. Some systems, comp—

especially those that do not luminesce under irradiation nevertheless fluoresce well under the action of light. This argument seems unconvincing to us, since the formation of ions and radicals depends precisely on the specific nature of the solvent and is not a universal mechanism.

Another point of view on the essence of radioluminescence belongs to Birks[^29][^30], who believes that the molecules of the solvent are initially excited to all high energy levels, up to ionization of the molecule. As a result of such primary excitation, in the author’s opinion, light is emitted from all excited levels in the form of short ultraviolet radiation in the region of 100–200 mµ, strongly absorbed by the solvent and therefore not observed outside it. The quanta of such hidden radiation, as a result of repeated acts of emission and reabsorption, migrate from molecule to molecule within the solution, reach the dissolved molecules of the fluorescent compound, and cause its characteristic emission. Birks’s hypothesis is in sharp contradiction with the existing generally accepted ideas concerning the internal deactivation of high electronic levels in complex molecules even in the gaseous state, and with the absence in them of the short-wave radiation required by the author.

We shall dwell here only on that aspect of the question which is of interest for sensitized fluorescence.

To prove the existence of intermolecular energy transfer and to exclude other interpretations of radioluminescence, Förster and Kallmann[^27], as well as Kogan and Weinreb[^31], applied excitation by ultraviolet light to the same liquid solutions that luminesced in their experiments under the action of γ-radiation. The fluorescence intensity of terphenyl in mixtures of xylene with cyclohexane was measured upon excitation by the 254 mµ line of the mercury spectrum, which is absorbed also by xylene but not by cyclohexane. As shown in Fig. 7, when the relative content of both solvents is changed toward enrichment with xylene at a constant concentration of terphenyl, the fluorescence intensity decreases upon excitation with 254 mµ light and does not change upon excitation with 313 mµ light. Unlike the first, the latter wavelength is not absorbed by xylene; therefore the fall in fluorescence is naturally

Fig. 7. Fluorescence of p-terphenyl in a mixture of cyclohexane and xylene.

can be explained by the inclusion of xylene in the path of the exciting light as a light filter. However, the final value of the fluorescence intensity of terphenyl, equal to 36% for pure xylene, is significantly greater than that expected on the basis of the ratio of the optical densities of the absorbing medium—xylene—and the terphenyl dissolved in it, in comparison with 100% in pure cyclohexane. From the optical calculation one should have expected, in pure xylene, a decrease in the fluorescence of terphenyl to 5%, and not to 36%. Meanwhile, under excitation by γ-rays, for which the optical densities of the absorbing substances are of no importance, the fluorescence intensity increases continuously in passing from cyclohexane to xylene (Fig. 7). On the basis of this result the authors believe that, under light excitation, as under excitation by γ-rays, there is a transfer of energy from the excited molecules of the solvent to the molecules of terphenyl.

An analogous comparison of the luminescence of terphenyl in hydrocarbon solvents (benzene, toluene, and xylene), on the one hand, and in chloroform, on the other, was carried out by Kogen and Weinreb^31. They established that, upon excitation by the 254 mμ line, absorbed only by the former solvents, the intensity of the luminescence of terphenyl in them is 20 times greater than in chloroform, which does not absorb this line. In doing so, the decrease in the fluorescence yield of terphenyl in chloroform by 35% in comparison with toluene, measured under excitation by the 313 mμ line, not absorbed by either solvent, was taken into account. As in the paper by Förster and Kallmann, the considerable intensity of the luminescence in hydrocarbon solvents exceeds that calculated on the basis of the ratio of the optical densities of the solvent and the dissolved substance. The possibility of reabsorption of the ultraviolet radiation of toluene, assumed by other authors, was excluded by the observation that a decrease in the thickness of the layer did not affect the ratio of the intensities.

In the work of Förster and Kallmann^28, in addition to the results presented above, experiments are also described for elucidating energy transfer between two fluorescing polycyclic aromatic hydrocarbons, such as naphthalene, anthracene, 9,10-diphenylanthracene and p- and m-terphenyl, α,α′-dinaphthyl, dissolved in a common solvent (xylene, n-butyl phosphate), which does not absorb light of the exciting wavelength (313 mμ).

With an increase in the concentration of the energy donor (for example, naphthalene) to large values, a decrease is observed in the fluorescence intensity of the energy acceptor (anthracene), which is present in a considerably smaller concentration. However, this decrease in the intensity of the acceptor emission reaches a constant value, clearly exceeding that expected from the ratio of the optical densities of the donor and acceptor. Hence the conclusion is drawn that energy transfer occurs in collisions of excited donor molecules

with the acceptor molecules. Unlike other cases, for the naphthalene* + terphenyl system, an increase in the concentration of naphthalene leads to the complete disappearance of the luminescence of terphenyl (under excitation at $\lambda 313\,m\mu$). This phenomenon is not caused by specific quenching of the terphenyl emission by naphthalene, as is shown by the experiment with excitation by light of $\lambda 334\,m\mu$, not absorbed by naphthalene. From a comparison of the long-wavelength limits of the absorption spectrum, the authors conclude that for effective energy transfer the electronic level of the acceptor must not lie too close to the donor level, and that there is a rather sharp threshold for the acceptor level, above which the probability of transfer falls to zero. Hence the conclusion is drawn that energy is transferred only from the very lowest electron-singlet level of the donor (naphthalene), and not from all higher levels, as Birks assumes in his hypothesis (see above). Furthermore, in view of the small value of the dipole moment for the long-wavelength electronic transition in the donor—naphthalene—the authors suppose that the energy transfer is not inductive in character, but is associated with the approach of the interacting molecules to contact.

In their qualitative approach, the experiments of Förster and Kallmann are unfavorably distinguished from the careful spectrophotometric studies of Bowen and coauthors, discussed above. For this reason their conclusions cannot be regarded as decisive.

5. ENERGY TRANSFER IN HETEROGENEOUS SYSTEMS

Two-component systems of aromatic hydrocarbons, of a type intermediate between a solid solution in a crystal and a joint molecular solution in a liquid neutral solvent, were studied in a series of works by Reid and coauthors32, 33, 34. They established conditions under which, in a solution of one component, for example anthracene, a second component, for example naphthacene, was suspended in the form of a suspension, and the system was brought to a low temperature ($-180^\circ$C). In such a heterogeneous or colloidal system, in sharp contrast to the conditions of molecular dissolution of both components, peculiar spectral effects are observed, interpreted by the authors as transfer of excitation energy from the dissolved molecules to the microcrystals of the suspension. In Fig. 8, under the letter (d), the emission spectrum is shown of a suspension of naphthacene alone in EPA*) at $-180^\circ$C; under (e) is the effect of a sharp enhancement of the luminescence of the naphthacene suspension, with a change in its spectrum, as a result of introducing anthracene into the solution in a concentration ensuring its solubility in molecular form. The dashed

*) A generally accepted abbreviation for a solvent consisting of ether + isopentane + ethanol in the proportion 5:5:2, possessing good dissolving ability with respect to organic compounds of various classes and freezing in the form of a glassy body.

A. N. TERENIN AND V. L. ERMOLAEV

the solid curve gives the emission spectrum of the added anthracene that appears; the dotted curve reproduces the form of the spectrum of the naphthacene suspension when the concentration of anthracene is increased threefold. The lower picture (f) corresponds to the same system (e) at 20°C. It is noteworthy that in the latter case the spectrum of naphthacene is practically identical with the spectrum of its molecular solution (see (c)), i.e.,

Fig. 8. Fluorescence spectra of the anthracene—naphthacene system. The emission of anthracene is indicated by a dashed curve; the spectra of naphthacene are indicated by solid and dotted curves.
First column — naphthacene in solution: a) naphthacene alone; b) the same concentration of naphthacene as in a) with added anthracene in solution at −180°C; c) the same as b) at 20°C.
Second column — naphthacene in suspension: d) naphthacene alone at −180°C; e) naphthacene at the same concentration as in d) with anthracene added in solution at −180°C; the dotted curve is the effect of a threefold increase in the concentration of anthracene; f) the same as e) at 20°C.
Third column — anthracene in suspension: g) anthracene alone at −180°C; h) the same concentration of anthracene as in g) with naphthacene in solution at −180°C, the ratio of the concentrations of naphthacene and anthracene being 1:10⁵; i) the same as h) at 20°C. EPA was used as the solvent in all cases.

obviously, the suspension of naphthacene is formed only under deep cooling. It is very interesting that the absorption spectrum of such a colloidal two-component system differs sharply from the sum of the spectra of the separated components under the same conditions (Fig. 9). A new intense band appears at 403 mμ, which, judging by its appearance, might be attributed to the strongly shifted (by ~40 mμ) first absorption maximum of dissolved anthracene molecules as a result of their adsorption and interaction with naphthacene microcrystals. However, such a considerable shift of the spectrum is contradicted by the negligibly small value estimated by the authors

of adsorption energy on microcrystals is on the order of 0.3 kcal/mole. In all cases, for excitation there was apparently used, chiefly, the 365 mµ line group of a mercury lamp, which affected both components. The appearance of a new significant absorption band in the spectral region of the exciting light, in our opinion, exhaustively explains the effect of the sharp enhancement of naphthacene emission. The decrease of the short-wavelength maximum at 520 mµ in the luminescence spectrum of the naphthacene suspension (Fig. 8) in comparison with its solution is naturally attributable to the superposition of the intense long-wavelength maximum of the absorption spectrum of the suspension (Fig. 9).

The characteristic and large changes described above in both the absorption spectrum of anthracene and the spectrum of naphthacene microcrystals, as well as of other analogous hydrocarbon systems, are attributed by the authors to adsorption of dissolved molecules of one of the components on the surface of microcrystals of the other. In their opinion, only under conditions in which molecular binary complexes of the participants are formed on the surface of microcrystals is effective intermolecular transfer of excitation energy possible. The observed reduction of the effects upon addition or upon complete transition to a polar solvent (alcohol) is explained by the authors by a decrease in the concentration of adsorbed hydrocarbon molecules removed by this solvent.

Fig. 9. Absorption of the anthracene–naphthacene system in a naphthacene solution and in suspension at −180° C: A) — combination of anthracene in solution and naphthacene in suspension in EPA, revealing a new absorption band at 403 mµ; B) — absorption of the individual components at the same concentration as in A).

Significant changes in the spectra are difficult to explain if one assumes, as the authors do, that only adsorption of the light-absorbing anthracene molecules on the naphthacene suspension takes place. We are evidently dealing here with a surface layer on the microcrystal containing naphthacene molecules dispersed in the layer of anthracene that has precipitated on the particles of the suspension. Indeed, the selectively enhanced maximum of the naphthacene suspension at 530 mµ (Fig. 8, e) is very close to the mean maximum in the emission spectrum when an anthracene suspension is present in the naphthacene solution (Fig. 8, h). There is every reason to suppose that, when the naphthacene suspension was formed, it also captured on the surface of the particles ...

and anthracene molecules, which separated the naphthalene molecules. As a result, the fluorescence yield increased in comparison with crystalline naphthalene, which has a low yield.

Thus, from our point of view, the effects observed by Ray belong to the category of energy-transfer phenomena observed in crystalline media. The differences in the spectral patterns may be explained by the incompleteness of the three-dimensional lattice and by the state of stress in the surface layer of the microcrystals. Confirmation of this point of view may be seen in the fact, established by Schmillen and Rode,^23 that the fluorescence spectrum, as well as the decay time, for anthracene changes only as a function of its state of dispersion. The authors come to the conclusion that on the surface of microcrystals the molecules are in a different state than inside the microcrystal.

More striking and unexplained is the strong attenuation of the intense long-wavelength maximum of absorption of the naphthalene suspension and the redistribution of intensity that occurs here upon adding anthracene solution (Fig. 9*). Purely optical effects are also possible, well known in the presence of turbid media that selectively absorb light at a definite maximum.

In the heterogeneous reverse system—anthracene in suspension, naphthalene in solution—there is also observed a considerable increase in the emission of naphthalene, the spectrum of which (Fig. 8, h) differs from that of a solution of naphthalene alone, but, according to the authors’ admission, coincides with the spectrum of crystalline anthracene contaminated with naphthalene. In the absorption spectrum there also appear new intense bands

Fig. 10. New absorption bands of the system anthracene in suspension and naphthalene in solution: A) anthracene in suspension and naphthalene in EPA solution at −180° C; B) naphthalene in solution at −180° C. The new bands are designated 3 and 4, whereas 1 and 2 are the usual absorption bands of a naphthalene solution.

* The two absorption maxima observed in the naphthalene suspension, 475 and 446 mµ (Fig. 9), actually belong to its dissolved molecules, as is easy to see by comparing Figs. 2 and 3 of the original.^33

(Fig. 10), shifted toward the long-wavelength side relative to the spectrum of the dissolved molecules. All these facts fit within the picture of the formation of surface inclusions of naphthacene in microcrystals of the anthracene suspension.

The diffuse appearance of the luminescence spectrum of the anthracene suspension, in comparison with the discrete spectrum of the dry crystal and the discrete spectrum of a solution of anthracene in a polar solvent (alcohol), due, apparently, purely to the optical effect of multiple scattering of the emitted light in the suspension, is regarded by the authors as a manifestation of the statistical distribution of molecular surface complexes of anthracene. In doing so, they arbitrarily assume that energy migration takes place between dissolved molecules up to the moment of emission at the surface of the microcrystal. This interpretation serves as the basis for the following assertions: 1) energy migration between identical molecules in solution is possible, and 2) since for anthracene molecules the absorption and emission spectra overlap only weakly, energy is transferred from the excited (by the act of light absorption) high vibrational levels of the donor even before the onset of the process of vibrational deactivation to the zero vibrational level.

The authors found no indications whatever of the existence of excitation-energy transfer in homogeneous solutions of both components under similar experimental conditions for 50 hydrocarbons in various combinations in a variety of solvents at \(+20^\circ\) and \(-180^\circ\)C. This negative result contradicts the above-described work of Bowen and Brocklehurst\(^{19}\) and Bowen and Livingston\(^{20}\), in which the presence of sensitized fluorescence was established for molecules of aromatic hydrocarbons in liquid benzene. Reid\(^{33}\) attributes his negative result to the low solubility of the hydrocarbons under his conditions, which did not ensure a sufficient concentration of molecules necessary for energy transfer by the inductive mechanism. Thus, the maximum concentrations attainable in EPA at \(-180^\circ\)C are, for anthracene, \(3.56 \cdot 10^{-4}\) mole/liter, and for naphthacene, \(1.54 \cdot 10^{-5}\) mole/liter. The author also used other solvents of similar composition. According to his data, the mean distance between dissolved molecules was about 200 Å, which obviously exceeded the distance for inductive transfer (approximately 60 Å).

Reid\(^{34}\) likewise did not observe sensitized fluorescence in the system terphenyl + diphenylhexatriene in liquid xylene under irradiation with the lines 253.7 and 313 mµ, in contrast to the result of Kallmann and Furst\(^{26}\), who established in this system, under the action of X-rays, enhancement of the luminescence of the second component upon addition of the first, which thereby lost its own luminescence. It should be pointed out that the authors overlooked the point that they sought to take a high concentration of the energy donor (terphenyl) and a low con-

concentration of the acceptor, whereas for the optimal conditions for observing transfer the presence of a high concentration of acceptor was required.

The absence of quantitative spectrophotometric measurements makes the authors’ far-reaching conclusions, in many cases, highly strained. Although the authors give a number of arguments against explaining the observed phenomena by simple reabsorption of the light emitted by one of the components, this possibility nevertheless is also not excluded in their experiments.

The study of energy migration in frozen alcoholic and glycerol solutions of a mixture of anthracene and naphthacene is also the subject of work by Feidish and co-workers35, 36. Even before the appearance of Reid’s work they established the dependence of the phenomenon on the rate of freezing of the solution, the spectrum of the slowly cooled solution proving independent of the solvent and analogous to the spectrum of a solid solution of naphthacene in anthracene. The authors believe that microcrystals are formed only by anthracene, while naphthacene dissolves in them, leading to the known sensitized fluorescence in the crystalline phase.

6. ENERGY TRANSFER IN MOLECULAR COMPLEXES

Of considerable interest are the luminescence spectra of molecular complexes of trinitrobenzene with aromatic hydrocarbons, discovered by Reid and a co-author37, 38. The formation and stability of such well-known intermolecular compounds had previously been ascribed to polarization of the electron shell of the hydrocarbon by the dipole moments of the nitro groups. Recently the view has come to prevail that the cause of association, and also the origin of the new long-wavelength maximum upon formation of the complex, is the electronic donor–acceptor interaction39.

The luminescence spectrum (upon excitation with wavelengths at 366 mµ) of solutions of crystallizing compounds of trinitrobenzene with a series of aromatic hydrocarbons (chrysene, pyrene, phenanthrene, benzanthracene, fluorene, acenaphthene, etc.), and also with acridine, appears when the solution is frozen to −180° C). It consists of a sequence of diffuse maxima lying in the region 500–700 mµ, close in their positions to the known discrete maxima of the triplet–singlet luminescence spectrum of the frozen solution of the same hydrocarbons observed in the absence of trinitrobenzene (Fig. 11)*). In contrast to the known great duration—

*) Complexes of the same hydrocarbons with picric and styphnic acids proved to be nonfluorescent.

**) In the case of anthracene, the authors attribute to an erroneous preceding interpretation of the phosphorescence spectrum of anthracene by Lewis and Kasha40, and assign to the triplet level of anthracene the value 19,000 cm\(^{-1}\) instead of 14,700 cm\(^{-1}\)37.

...of the triplet-singlet spectrum, the luminescence decay time of the complexes is, according to the authors’ qualitative estimate, less than \(10^{-4}\) sec. Consequently, if the emission is attributed to the triplet level, then in the complex there occurs a substantial weakening of the transition prohibition associated with reversal of the electron spin.

The authors further assume that the luminescence spectrum of the complex also occupies the spectral region in which one might expect the appearance of luminescence of the trinitrobenzene molecule itself, if it possessed the ability to emit. From this they conclude that intermolecular energy transfer takes place within the complex, with excitation of the triplet level of the hydrocarbon component. They see confirmation of their point of view in the fact that the complex of carbazole with trinitrobenzene does not fluoresce under the same conditions, while the luminescence yield of the complex with phenanthrene is very low. Indeed, the triplet level of carbazole (\(24\,480\ \text{cm}^{-1}\) at \(450\ \text{m}\mu\)) and of phenanthrene (\(21\,740\ \text{cm}^{-1}\) at \(460\ \text{m}\mu\)) lies higher than in the other hydrocarbons and, possibly, exceeds the assumed excitation level of the complex.

The mechanism of excitation of the complex emission received contradictory interpretations in two papers by the authors. On the basis of the continuous character of the long-wavelength edge of the absorption spectrum of the complex, they concluded\({}^{37}\) that photodissociation of the complex into its constituent molecules occurs, the hydrocarbon molecule being left at the triplet level. Having established that the complexes retain fluorescence also in the form of crystals, the authors soon rejected this explanation. In the second paper\({}^{38}\) they adopt the interpretation of the long-wavelength absorption maximum of the complex as the spectrum of intermolecular electron transfer from the hydrocarbon

Fig. 11. Emission spectra of hydrocarbons and their complexes with trinitrobenzene. Dashed line: long-wavelength triplet emission of hydrocarbons; solid line: emission spectra of complexes of hydrocarbons with trinitrobenzene.
\(a)\) 2-methyl-1,2-benzanthracene, \(b)\) 3-methyl-1,2-benzanthracene, \(c)\) 6-methyl-1,2-benzanthracene, \(d)\) 8-methyl-1,2-benzanthracene.

to trinitrobenzene. Obviously, in the reverse transition the electron ends up on the triplet level. The authors try to ascribe the shortening of the lifetime at the latter to distortion of the system of \(\pi\)-electrons of the aromatic polycycle by the nitro groups*).

Further uncertainty was introduced into the question by the careful spectrophotometric measurement by Bjur and Ketelaar\(^{41}\) of the long-wavelength maximum of the absorption spectrum of the same complexes. They showed that this maximum for the complexes of trinitrobenzene with anthracene and phenanthrene is a mirror image of the envelope curve of the luminescence spectra of the complex obtained by Reid (Fig. 12). The position of the center of mirror symmetry of the pattern agrees well with the value of the energy of the zero transition to the triplet level of the corresponding hydrocarbon, including naphthalene. Hence it follows that the observed spectra, both emission and absorption, belong to singlet–triplet transitions of the hydrocarbon. Then the perplexing question arises of the ability of trinitrobenzene to remove the prohibition of the singlet–triplet transition and even to stimulate this transition. On the other hand, it is known that trinitrobenzene gives similarly intensely colored complexes, similar in spectrum, with \(OH^{-}\) or \(OCH_3^{-}\) ions, which do not have triplet levels like those of hydrocarbons. Finally, complexes of the same hydrocarbons with iodine molecules or sulfur dioxide do not reveal the absorption maxima observed in the case of trinitrobenzene. Apparently, the mirror symmetry of the absorption spectrum of the complex with its luminescence spectrum is accidental in origin, and the long-wavelength absorption maximum nevertheless belongs to the electron-transfer spectrum, and not to singlet–triplet absorption**).

Fig. 12. Complex of sym-trinitrobenzene—anthracene, absorption spectrum in chloroform at \(25^\circ\text{C}\) (upper scale of wave numbers); — — — emission spectrum according to Reid at \(-180^\circ\text{C}\) in organic glass (lower scale of wave numbers); \(MP\) — reflection point; \(I\) — intensity in arbitrary units.

*) The authors also attempt to relate the height of the triplet level to the carcinogenicity of the hydrocarbon, but unsuccessfully.

**) See also \(^{39}\).

7. INTERMOLECULAR ENERGY TRANSFER WITH EXCITATION OF THE TRIPLET LEVEL

The question of the possibility of energy transfer from a molecule in a metastable triplet state to an identical molecule in the ground state has been the subject of discussion. Vavilov and Shishlovsky^42, as well as Levshin and Vinokurov^43, showed that when the concentration of a dye (rhoduline orange, fluorescein-Na) is increased in sugar or boric phosphors, the decay time is shortened. By contrast, Lewis, Lipkin, and Magel^44 found only a negligible shortening of $\tau$ for the neutral form of fluorescein in the mixed solvent glycerin + phosphoric acid at $-100^\circ\mathrm{C}$ when the concentration of fluorescein was varied from $5\cdot 10^{-5}$ to $5\cdot 10^{-2}$ mole/liter. From this the latter authors concluded that energy exchange between metastable and normal molecules is impossible. The same point of view is defended by Feofilov^45, who found no depolarization of phosphorescence in the course of decay.

However, Sveshnikov^46 confirmed the existence of a shortening of the phosphorescence of fluorescein-Na in boric acid with increasing dye concentration at $-100^\circ$ and $+18^\circ\mathrm{C}$. The shortening of the $\tau$ of the phosphorescence of benzene and its derivatives with increasing concentration in alcoholic solution at $-180^\circ$^47, observed by him as well, also convincingly speaks in favor of the existence of nonradiative removal of energy from the triplet level by neighboring molecules located on the singlet ground level. The question of the possibility of transferring them, in this process, to the triplet level remained open.

This question was unambiguously resolved by Terenin and Ermolaev^48,49, who discovered the new phenomenon of sensitized phosphorescence in frozen (at $-180^\circ\mathrm{C}$) solutions containing certain pairs of aromatic compounds. Carbonyl derivatives of benzene (benzaldehyde, benzophenone, acetophenone, ethylphenylketone, o-oxybenzaldehyde, benzoin), as well as carbazole and diphenylamine, were used as energy donors; naphthalene, diphenyl, $\alpha$-methylnaphthalene, and $\alpha$-chloronaphthalene were used as acceptors. The arrangement of the singlet and triplet levels of the partners was such that it allowed selective excitation by light of the energy donor without affecting the acceptor (Fig. 13). From the fact of the appreciable duration of the afterglow of the donor luminescence spectrum it may be concluded that a large part of its molecules, after primary excitation by light, are found on the triplet level $T_D$, situated very close to the singlet level $S_D^*$.

All combinations of donors with acceptors for which $T_D > T_A$ led to the appearance of sensitized phosphorescence upon excitation $h\nu = S_D^* < S_A^*$. In other words, the exciting wavelength was always outside the long-wavelength limit of the absorption of the acceptor, and direct excitation of the latter was excluded. The transfer

energy is also detected by the shortening of the decay time \(\tau\) and by the decrease in the donor yield \(^{50}\).

The phenomenon is not reproduced if the triplet level \(T_A\) of the acceptor lies above the level \(T_D\) (or \(S_D^*\)) of the donor. In the phenomenon of sensitized phosphorescence, the possibility of simulating energy transfer by trivial reabsorption by the acceptor of the donor’s luminescence light is completely excluded. The experimentally established \(^{48}\) identity of the triplet emission spectrum of the acceptor with its ordinary phosphorescence spectrum in the absence of the donor shows the absence of any association of the participants. In Ermolaev’s works \(^{50}\), the phenomenon was subjected to quantitative study, and it was established, in particular, for some systems, that the shortening of the donor decay time with increasing acceptor concentration occurs even in those cases where it was not possible to detect the spectra of sensitized phosphorescence (for example, in the systems carbazole\(^*\)—naphthalene, diphenylamine\(^*\)—naphthalene). The transfer of excitation energy thus has the character of “long-range action,” although the overlap of the spectral regions of donor emission and acceptor absorption required for inductive interaction is absent.

Fig. 13. Scheme of the electronic levels of molecules participating in radiationless energy transfer in sensitized phosphorescence. On the left are the unexcited \((S_D)\), fluorescent \((S_D^*)\), and phosphorescent (triplet \(T_D\)) levels of the energy donor; on the right, the same for the acceptor.

The high efficiency of transfer with excitation of an optically forbidden transition in the acceptor molecule must, evidently, be compared with Wigner’s well-known rule \(^{52}\) on the conservation of the total spin of the system in effective processes of deactivation of triplet states of atoms in collisions in a gas. In the present case, the change of spin as a result of the triplet–singlet transition in the donor is compensated by the reverse change of spin in the acceptor.

Analogous experiments have not yet been published by other investigators. Indeed, the effects observed by Reid \(^{37,38}\), as we have seen, have another origin and are caused by an obviously close interaction leading to distortion of the spectra. There is only one observation by Ferguson and Tinson \(^{51}\) of a change in the spectrum of a frozen solution of benzophenone in petroleum ether, which the authors mistakenly interpret as transfer of excitation energy to another benzophenone molecule with its transition to the triplet level.

With an increase in the concentration of benzophenone from \(10^{-4}\) to \(10^{-2}\) mole/liter, its normal low-temperature luminescence spectrum in solution is replaced by another, shifted toward the long-wave side. The authors suppose that the latter belongs to the triplet level of the compound. Meanwhile, they do not take into account the possibility of separation of benzophenone, upon freezing of its concentrated solution, in the form of microcrystals. Ermolaev showed that the altered luminescence spectrum of benzophenone is close to the spectrum of its crystals at low temperature.

8. MIGRATION OF EXCITATION ENERGY AND SENSITIZED FLUORESCENCE IN BIOLOGICAL OBJECTS

In the preceding review, attempts were described to detect transfer of excitation energy from accessory pigments to chlorophyll in marine algae \(^{1}\) by observing sensitized fluorescence; these attempts, however, did not yield sufficiently convincing results. In recent years carefully performed studies have appeared in which this phenomenon was reproduced more distinctly. French and Young \(^{53}\) carried out a quantitative investigation of the fluorescence spectrum of the red alga Porphyridium, using high-intensity monochromators with diffraction gratings for isolating the exciting lines of a mercury lamp, and photoelectric recording of the fluorescence spectrum. The apparatus was calibrated in terms of energy values of intensity and number of emitted quanta.

Fig. 14. Fluorescence spectrum of the alga Phorphyridium and its decomposition into the spectra of the pigments phycoerythrin, phycocyanin, and chlorophyll present in it.

Figure 14 shows the summary curve of the fluorescence spectrum of the alga and its decomposition into component maxima belonging to chlorophyll, phycocyanin, and phycoerythrin. Upon excitation in the region \(405\text{–}450\ \mathrm{m}\mu\), only chlorophyll fluoresces; upon excitation in the region \(450\text{–}550\ \mathrm{m}\mu\), along with the intense fluorescence of chlorophyll, maxima of fluorescence of the pigments—phycocyanin and phycoerythrin—also appear in the spectrum.

The authors measured, for eleven wavelengths, the efficiency of excitation of the luminescence of all three pigments. In Fig. 15 is shown

comparison of the spectral distribution of excitation of chlorophyll fluorescence (circles) with the absorption spectra of phycoerythrin in the alga, and of chlorophyll and phycocyanin in solutions. It follows from this that chlorophyll fluorescence is excited in the intrinsic

Fig. 15. Absorption spectra of the pigments phycoerythrin, phycocyanin, and chlorophyll (curves) and the excitation spectrum of chlorophyll fluorescence (circles) in Phorphyridium.

short-wavelength absorption maximum at 436 mµ with a small yield, and with a large yield within the absorption maximum of phycoerythrin, located at 550 mµ.

From the spectral distribution of excitation of phycocyanin fluorescence it also follows that this pigment receives excitation energy from phycoerythrin. The authors suppose that phycocyanin serves as an intermediate link, providing better resonance matching of the energies transferred by the inductive mechanism from phycoerythrin. According to calculations, the pigment concentration in algae is sufficient to ensure the necessary intermolecular distances (about 100 Å).

Similar results were obtained, with a less refined spectral method, by Duysens54 with the red algae Porphyra and Porphyridium. He also confirmed the transfer of energy from phycoerythrin to phycocyanin, and then to chlorophyll “a,” with a yield exceeding 80%, by exciting the two pigments separately with monochromatic light of 546 and 600 mµ. For other red and blue algae, the fact is also striking that there is a low yield of fluorescence of chlorophyll “a” when it is excited in its own absorption band at 420 mµ, and a high yield when excited in the absorption region of phycoerythrin. The author comes to the conclusion that there exists

of chlorophyll in two forms—strongly and weakly fluorescing. For Porphyra, a very bright fluorescence maximum of unknown origin is found at 725 mµ, dominating in this region over the luminescence maximum of chlorophyll and excited, together with it, by a wavelength of about 420 mµ. The author assumes that energy is transferred from chlorophyll to this unknown fluorescing pigment, which, in his opinion, explains the weak yield of chlorophyll fluorescence. Meanwhile, an analogous poor yield occurs for Porphyridium, where there is no fluorescence of such an unknown pigment.

Duysens extended analogous experiments also to photosynthesizing purple bacteria, which contain, instead of chlorophyll, the red pigment bacteriochlorophyll. The absorption spectrum of bacteriochlorophyll in living bacteria consists of three maxima, situated in the infrared region at 800, 850, and 890 mµ, which, as Krasnovsky has shown,^55 apparently belong to three different states of bacteriochlorophyll. The infrared fluorescence of bacteriochlorophyll is observed upon excitation at the absorption maximum 890 mµ, and not at the other two, and also at the shorter-wavelength maximum 590 mµ, corresponding to the chlorophyll maximum at 420 mµ. In addition, in the absorption spectrum of the bacteria there is present, in the region 550–450 mµ, a series of intense absorption maxima

Fig. 16. Absorption spectra (△—·—△), phototaxis action (○------○), and excitation of bacteriochlorophyll fluorescence—the band at 890 mµ (●—●)—for the purple bacteria Chromatium.

belonging to carotenoids. It turned out that the spectral distribution of excitation of the infrared fluorescence of bacteriochlorophyll exactly reproduces the absorption spectrum of the carotenoids (Fig. 16). Thus, in this case as well, there is found

transfer of energy to the photosynthesizing pigment from accompanying pigments of a different nature, which absorb shorter wavelengths. The author also carried out experiments on the transfer of energy from chlorophyll “b” to chlorophyll “a” both in acetone solution and in living cultures of Chlorella.

Despite the persuasiveness of the results described here, some doubt is nevertheless raised by the failure to take into account the influence of reabsorption of fluorescence light, especially in the experiments of Duysens, who apparently does not take this correction into consideration at all. It is known that the fluorescence of phycoerythrin is extremely intense even in ordinary aqueous solution.

In both papers the view is advanced that, of all the pigments listed, only chlorophyll is connected with the chemistry of the photosynthesis reaction. It should be noted in passing that the fluorescence of chlorophyll changes antibatically with the photosynthesis reaction.

In connection with the question of the transfer of excitation energy between pigments by the inductive mechanism, there arises the question of the possibility of migration of excitation energy in protein, which is the medium in which, or on the surface of which, essentially all biochemical processes take place.

In the preceding review1, the results of the work of Bücher and Kaspers were critically examined; these results were interpreted as migration of excitation energy, absorbed by tyrosine and tryptophan of the protein of myoglobin, to the iron-containing pigment—heme—localized on the surface of the protein molecule. As a result, there occurs the cleavage from the heme of a CO molecule weakly bound to it, with the same efficiency as under the direct action of light on the heme. We pointed out that direct proof of the possibility of exciton migration in protein, postulated by a number of investigators (see review1), would be the detection of sensitized fluorescence of a chromophore fixed on the surface of a protein molecule under the action of light absorbed by the aromatic amino acids—tryptophan and tyrosine—present in small concentration in the bulk of the protein.

Such an experiment was indeed carried out by Bannister56, who used the natural pigment, the chromoproteid phycocyanin, which includes 8–16 molecules of the fluorescent chromophore mesobiliviolin, firmly bound to a globulin-type protein molecule having a molecular weight of 280,000. The absorption and fluorescence spectrum of an aqueous solution of phycocyanin, extracted from algae and carefully purified, is shown in Fig. 17. The maxima at 615 and 400–350 mμ belong to the chromophore, while the maximum at 275 mμ is ascribed to the benzene ring of the tryptophan and tyrosine groups of the protein. The dotted line shows the maximum of the fluorescence spectrum of phycoerythrin according to Duysens. Figure 18 gives Bannister’s measured values of the relative quantum yield of fluorescence for four preparations of different purity for a series of exciting wavelengths.

waves. For the purest preparation (4, 7), despite the large scatter of points, one may conclude that the fluorescence yield is practically constant, independently of the spectral region of excitation. In other words, both in the region 310–400 mμ, where only the chromophore absorbs,

Fig. 17. Solid line — absorption coefficient of phycoxanthin extracted from algae; dashed line — fluorescence spectrum of phycoxanthin.

and in the region 260–310 mμ, where, according to the author’s determinations, about 50% of the light absorption is due to the protein, the yield is practically the same. From this the author concludes that there is a very

Fig. 18. Value of the relative quantum yield of fluorescence of phycoxanthin as a function of the wavelength of the exciting light for preparations of various degrees of purification.

efficient transfer of excitation energy from the amino acids composing the protein to the chromophore. It is assumed that the transfer mechanism is inductive in origin, i.e., that fluorescence must exist in the named aromatic amino acids,

which, by its spectral position, should coincide with the absorption band at 350–400 mμ of the chromophore and thereby produce excitation of its fluorescence. The ultraviolet emission postulated by the author has not been observed either in protein in general or in the named aromatic amino acids. The sufficiently high concentration of the energy acceptor required for inductive transfer is clearly not present here. Other modes of migration of excitation energy in protein are excluded because of the absence of the corresponding low electronic levels (see the review ^1).

It seems to us that Bannister’s interpretation of the results obtained is not convincing. His determination of protein absorption at a maximum of 275 mμ was based on measurement of the absorption spectrum of hydrolyzed, i.e., denatured, protein, which obviously cannot reproduce the absorption of the original, intact native protein in phycoerythrin. Moreover, the removal of residues of the chromophore structure was not controlled by him with sufficient care. The author does not give the spectrum of the solution of the isolated chromophore itself—mesobiliviolin—which, possibly, also has an absorption maximum in the region of 275 mμ*). Apparently, what occurs here is direct excitation of the chromophore in its own absorption spectrum.

The present review of numerous works devoted to radiationless transfer of excitation energy between molecules testifies to the great interest of researchers in this question both in the Soviet Union and abroad. A great influence on the development of this field of science in our country was exerted by the works of S. I. Vavilov devoted to the theory of energy migration in concentrated dye solutions.

What is new in the research of recent years is the extension of the study of energy-transfer processes to solutions of simpler organic molecules. Thus, alongside crystals of polycyclic aromatic compounds, their mixed solutions have been studied in detail. These compounds have the advantage, in comparison with dye solutions, that in them the effects of energy migration are not complicated by side phenomena (association, influence of pH, polymerization, etc.). In these works, energy transfer between molecules is demonstrated sufficiently convincingly by experiments on sensitized fluorescence. In the new phenomenon of sensitized phosphorescence, the presence of energy transfer between tri-

*) Havemann ^57, carrying out careful measurements of the spectrum of another chromophore, heme, in myoglobin, showed the doubtful nature of the above-mentioned explanation by Brocq and Kasperse of the transfer of energy from the protein to the heme chromophore on the grounds that the spectrum of the chromophore also has an absorption maximum in the region of 275 mμ and that the estimate of the protein’s share of the absorption was incorrect.

excited levels of organic molecules. For biological objects, such as algae, for example, the use of a refined experimental technique employing the phenomena of sensitized luminescence in a living object made it possible to demonstrate more convincingly the transfer of energy between pigments. It should be noted, however, that in many works the internal secondary absorption of luminescence light from the energy donor was insufficiently taken into account, and therefore the conclusions drawn in them concerning the presence of sensitized luminescence are not sufficiently substantiated.

REFERENCES

1. A. N. Terenin, UFN 43, 347 (1951).
2. H. A. Stuart, “Die Struktur des freien Moleküls,” Berlin, Göttingen, Springer, 1952.
3. N. A. Prilezhaeva, Acta phys. chim. URSS 1, 785 (1935).
4. N. A. Prilezhaeva, A. A. Klimova, ZhFKh 10, 353 (1937); Acta phys. chim. URSS 7, 163 (1937).
5. B. S. Neporent, Trans. Farad. Soc. 35, 39 (1939); ZhFKh 13, 965 (1939); 21, 1111 (1947).
6. B. S. Neporent, B. I. Stepanov, UFN 43, 380 (1951).
7. B. S. Neporent, ZhFKh 24, 1219 (1950).
8. A. N. Terenin, A. V. Karyakin, Izv. AN SSSR (ser. fiz.) 15, 550 (1951); DAN SSSR 96, 269 (1954).
9. M. D. Galanin, M. Frank, ZhETF 21, 114 (1951).
10. M. D. Galanin, L. V. Levshin, ZhETF 21, 121 (1951).
11. M. D. Galanin, ZhETF 21, 126 (1951).
12. M. D. Galanin, Izv. AN SSSR (ser. fiz.) 15, 543 (1951).
13. M. D. Galanin, ZhETF 28, 485 (1955).
14. Th. Förster, Zeits. f. Elektrochem. 53, 93 (1949); Zeits. Naturforsch. 4a, 321 (1949).
15. A. Schmillen, Zeits. f. Physik 135, 294 (1953).
16. V. Zanker, Zeits. phys. Chem. 199, 225 (1952); 210, 250 (1952).
17. E. J. Bowen, I. Brocklehurst, Trans. Farad. Soc. 51, 774 (1955).
18. A. Schmillen, Zeits. Naturforsch. 9a, 1036 (1954).
19. E. J. Bowen, H. Brocklehurst, Trans. Farad. Soc. 49, 1131 (1953).
20. E. J. Bowen, R. Livingston, J. Amer. Chem. Soc. 76, 6300 (1954).
21. G. T. Wright, Proc. Phys. Soc. 66A, 77 (1953).
22. O. Simpson, D. S. Northrop, Physika 20, 1122 (1954).
23. A. Schmillen, L. Schmillen, F. Rohde, Zeits. Naturforsch. 8a, 213 (1953).
24. N. D. Zhevandrov, DAN SSSR 83, 677 (1952); 100, 455 (1955).
25. H. Kallmann, M. Furst, Phys. Rev. 79, 857 (1950); 81, 853 (1951).
26. M. Furst, H. Kallmann, Phys. Rev. 85, 816 (1952); 89, 912 (1953).
27. M. Furst, H. Kallmann, Phys. Rev. 94, 503 (1954).
28. M. Furst, H. Kallmann, Phys. Rev. 96, 902 (1954).
29. J. B. Birks, Phys. Rev. 94, 1567 (1954); 95, 277 (1954).
30. J. B. Birks, “Scintillation Counters” (1953), p. 68.
31. S. G. Cohen, A. Weinreb, Phys. Rev. 93, 1117 (1954).
32. M. M. Moodie, C. Reid, J. Chem. Phys. 19, 986 (1951).
33. M. M. Moodie, C. Reid, J. Chem. Phys. 20, 1510 (1952).
34. M. M. Moodie, C. Reid, J. Chem. Phys. 22, 1126 (1954).
35. A. N. Faidysh, O. P. Kharitonova, DAN USSR, No. 5 (1951).
36. A. N. Faidysh, DAN USSR, p. 215 (1953).

1. C. Reid, J. Chem. Phys. 20, 1212 (1952).
2. M. M. Moodie, C. Reid, J. Chem. Phys. 22, 252 (1954).
3. A. N. Terenin, Uspekhi Khim. Nauk 24, 121 (1955).
4. G. Lewis, M. Kasha, J. Amer. Chem. Soc. 66, 2100 (1944).
5. A. Bier, J. A. A. Ketelaar, Recueil 73, 264 (1954).
6. S. I. Vavilov, A. A. Shishlovskii, Phys. Zeits. Sowjet 5, 379 (1934).
7. V. L. Levshin, L. A. Vinokurov, Phys. Zeits. Sowjet 10, 10 (1936).
8. G. Lewis, D. Lipkin, T. Magel, J. Amer. Chem. Soc. 63, 3005 (1941).
9. P. P. Feofilov, Dokl. Akad. Nauk SSSR 44, 159 (1944).
10. B. Ya. Sveshnikov, Dokl. Akad. Nauk SSSR 58, 49 (1947).
11. P. P. Dikun, A. A. Petrov, B. Ya. Sveshnikov, ZhETF 21, 150 (1951).
12. A. N. Terenin, V. L. Ermolaev, Dokl. Akad. Nauk SSSR 85, 547 (1952).
13. V. L. Ermolaev, A. N. Terenin, in Collection in Memory of S. I. Vavilov, Publishing House of the Academy of Sciences of the USSR, 1952, p. 137.
14. V. L. Ermolaev, Dokl. Akad. Nauk SSSR 102, 925 (1955).
15. J. Ferguson, H. Tinson, J. Chem. Soc. (L.), p. 3083 (1952).
16. E. Wigner, Göttingen Nachrichten, p. 375 (1927).
17. C. S. French, V. K. Young, J. Gen. Physiology 35, 873 (1952).
18. L. N. M. Duysens, Nature 168, 548 (1951).
19. K. K. Voinovskaya, A. A. Krasnovskii, Biokhimiya 18, 626 (1953).
20. T. T. Bannister, Archiv. Biochem. Biophys. 49, 222 (1954).
21. R. Havemann, P. Wolff, Zeits. phys. Chem. 201, 246 (1952).