Luminescence in Chemical Reactions

A. Rabinovich.

The emission of light in a chemical reaction, not caused by heating, is called chemiluminescence. A whole series of chemical reactions is accompanied by bright light phenomena. Such, for example, are the flames that appear during the rapid oxidation of very many substances, when chlorine combines with antimony, with hydrogen, etc. However, the appearance of a flame is not analogous to chemiluminescence, since the light emitted by a flame is a secondary phenomenon caused by the heating of the reacting substances or of the reaction products, for the most part incompletely burned particles of carbon. Here we encounter so-called thermal or caloric radiation.

In thermal radiation, between the temperature of heating and the wavelength of the brightest emitted rays there exists a quite definite dependence, expressed by Wien’s law. More precisely, this law, like all other exact dependences concerning thermal radiation, applies not to real solid bodies, but to an absolutely black body, which has the property of absorbing all rays of the spectrum falling upon it at any temperature. The various real bodies with which the physicist deals in his laboratory differ, to a greater or lesser degree, in their properties from an absolutely black body, thereby giving quantitative deviations from the exact laws that determine radiation; but the general character of these laws, at least qualitatively, is preserved for other, real bodies as well.

Further, for thermal or caloric radiation Kirchhoff’s law is valid, according to which the ratio of emissive power (the quantity of emitted rays) to absorptive power is the same for all bodies and is equal to the same ratio for an absolutely black body. This law, repeatedly verified theoretically and experimentally, is of fundamental importance in theoretical physics.

It can serve, among other things, as a criterion for distinguishing purely thermal (caloric) radiation from any other. In the former case it is justified; to the others it is inapplicable. Thus, if it turns out that for a given radiation Kirchhoff’s law is inapplicable,

one may say with confidence that we have here a case not of thermal radiation, but of so-called luminescence, i.e. the transition of some other form of energy into light, bypassing the form of thermal energy. Thus, in triboluminescence light energy is formed from mechanical energy; in electro- and cathodoluminescence—from electrical energy; in photoluminescence (phosphorescence and fluorescence)—from light energy itself, but of another wavelength; in chemiluminescence—from chemical energy. The characteristic feature of chemiluminescence is, therefore, luminescence without heating, or with heating so slight that by itself it is incapable of producing luminescence. Of course, both kinds of luminescence may also occur together; in that case chemiluminescence will intensify the luminescence caused by heating from the heat of the reaction.

Chemiluminescence was first observed in 1669 by Brand, who discovered phosphorus. If this substance is kept in air (it is ordinarily stored under water), a cloud of luminous vapors is emitted from it, which gave grounds for calling the element “phosphorus,” i.e., in Greek, “light-bearing.” The cause of this luminescence was the subject of prolonged controversy, until Schrötter (B. A. Schrötter) proved definitively in 1852 that the cause of the luminescence of phosphorus is its oxidation in air. Subsequently it turned out that the oxidation of phosphorus under certain conditions is a periodic phenomenon[^1], i.e. it proceeds now with increasing, now with decreasing speed, depending on the pressure of oxygen. If the pressure is raised, the luminescence weakens, but then gradually intensifies, weakens again, and so on. This was explained[^2] by the fact that phosphorus gives different degrees of oxidation, which can oxidize further with greater or lesser ease, depending on the pressure of oxygen.

Since Brand’s time the number of examples of chemiluminescence has increased extraordinarily. Some reactions are accompanied by the liberation of a considerably greater quantity of light energy than the oxidation of phosphorus. Most of them are oxidation reactions; thus, fresh surfaces of sodium and potassium glow in air; this is observed especially in organic substances. Thus, with the emission of light, certain alcohols and polyhydric phenols, naphthols, etc., are oxidized. Among the brightest reactions are the oxidation of amarine, lophine, hydrobenzamide, various aldehydes and their derivatives[^3], and in particular the reaction of Wedekind (a compound of chloropicrin with phenylmagnesium iodide) and the oxidation, by hydrogen peroxide, of a mixture of pyrogallol with formaldehyde in alkaline solution (red luminescence)[^4]. In recent times two new groups of luminescent reactions have been discovered, namely—the oxidation of magnesium-organic compounds[^5] and, studied by Delepine[^6], sulfur-containing organic compounds.

The field of chemiluminescence also includes the glow of certain organisms: bacteria, noctilucae, fireflies, etc. These phenomena are described in detail in Harvey’s book⁷).

One of the first works containing a scientific investigation of chemiluminescence and of its dependence on the nature of the reaction, the concentration of the components, and external conditions was Trautz’s work⁸). Its principal results amount to the following: the amount of light in a chemical reaction increases with an increase in the rate of the reaction. All factors that increase the latter intensify chemiluminescence. Such factors are the concentration of the components, the rate of mixing, temperature (of course, not reaching the onset of thermal incandescence); a considerable influence is exerted by the order in which the components are mixed, the reaction of the medium (acidic or alkaline), and its “dissociating power,” i.e., its effect on the degree of electrolytic dissociation of the reaction components. Trautz also found that the addition of various substances indifferent in the chemical and physical sense does not affect the quantity and quality of the light emitted.

The spectrum of chemiluminescence has thus far been studied very little. The reason for this is the extreme weakness of the phenomenon itself, i.e., the small amount of light emitted, which does not permit the use of instruments of ordinary sensitivity. Trautz⁸) made a number of qualitative observations, and he was able to observe the spectrum only subjectively. In the brightest reaction he studied—the oxidation of a mixture of pyrogallol with formaldehyde—the spectrum proved to be continuous; it extended from red to blue, with a maximum in the orange region. The reactions of chlorine and bromine with ammonia, chlorine and bromine with acetylene, palmitic acid, cetyl alcohol with oxygen, and amarine with bromine water in alkaline solution also gave continuous spectra with a maximum in the yellow region.

Centnerszwer and Petrikaln⁹) were the first to study the spectrum of chemiluminescence by an objective method. As their object they used the classical reaction of phosphorus oxidation. They constructed a special device that allowed the phosphorus to oxidize slowly over many hours. If measures are not taken to remove the heat liberated during the reaction, then this heat warms the phosphorus, which accelerates the course of the reaction, and this, in turn, increases the liberation of heat. The reaction thus accelerates itself until it ends in a flash, i.e., in the rapid combustion of all the available phosphorus. To prevent this, the authors placed sticks of phosphorus in frames made of thick sheet lead, which cooled the reacting phosphorus owing to the high thermal conductivity of lead. These frames were placed in a large glass cylinder with an outlet tube through which the reaction products were removed, and after an exposure lasting several hours, on the photographic plate in the spectrograph of large

luminosity, chemiluminescence spectra of phosphorus vapors were obtained. These first photographs gave very interesting results. The spectrum was continuous in the visible region, while with an exposure of about 100 hours in a quartz spectrograph a photograph was obtained also of the ultraviolet part of the spectrum with separate lines; of these the brightest are the lines 2474 and 2458 Å. This emission spectrum does not coincide with the absorption spectra of phosphorus and its compounds in various solvents.

The intensity of chemiluminescence, i.e., the amount of light energy released during the reaction, until very recently had been noted only qualitatively. It is difficult to measure because of the weakness of the phenomenon. The first quantitative investigation of this question was carried out by Grinberg\(^{10}\), who compared the durations of “exposure” required for equal blackening of photographic plates on which small glasses were placed with luminescent reactions proceeding in them*).

As regards theoretical conceptions of the essence of chemiluminescence, in general all authors who wrote about it agreed on the idea that part of the energy released in a chemical reaction escapes conversion into heat; the atoms of the reacting substances thereby pass into an “excited” (angeregt) state, manifested in luminescence. No one doubted that it is precisely those atoms that take direct part in the reaction which emit light.

In recent years a number of interesting works have appeared which force one to doubt this assumption and, on the contrary, to ascribe the ability to emit light to those atoms of the reacting substances which are not directly affected by the reaction taking place—to atoms of the reaction products and even to extraneous substances mixed into the reacting mixture.

Strutt\(^{11}\) in 1913 observed that activated nitrogen, reacting with some substance, causes the appearance of spectral lines or bands characteristic of other substances present in the same space. From this he concluded that these substances, capable of emitting light rays, draw the energy necessary for this from the store released during the reaction, although they themselves take no direct part in it.

Stuchtey\(^{12}\) in 1914 showed that the light emitted during the decomposition of ozone contains spectral bands characteristic of undecomposed ozone. Here too, therefore, the energy released

*) At the meeting of the American Physical Society held in April 1924 in Washington, a paper was presented by Elliot Q. Adams, who determined the fraction of the reaction energy released in the form of light energy. The latter he determined photometrically, and the total energy of the reaction from calorimetric data. It turned out that, for example, in the case of the oxidation of phosphorus, the “yield” of light constitutes an extremely negligible fraction of the total amount of energy released in the reaction.

when ozone molecules decomposed, was transferred to particles that had not yet undergone decomposition and brought them into an optically excited state.

Important observations were made by Haber and Zisch¹³). They noticed that the light emitted by an ordinary gas flame is not wholly of purely caloric origin. Part of it draws its energy directly from the chemical reactions occurring in the flame, without the intermediate conversion of that energy into heat. Of the three parts of a gas flame—the inner cone, the outer cone, and the region of water gas—the inner cone is the brightest. Yet its temperature (1500°–1600°) is undoubtedly lower than the temperature of the outer cone (1800°). In the nature of the reactions occurring in them these parts differ from one another: in the outer cone the substances produced by incomplete oxidation of fresh gas in the inner cone are oxidized by atmospheric oxygen to carbonic acid and water. This led to the supposition that part of the light emitted by the relatively cold inner cone has a chemical, and not a caloric, origin. This difference in brightness can be shown especially clearly if the burner opening is given the form of a long, narrow slit and one looks along it. In order to make the pale gas flame brighter, sodium chloride is introduced into it, and the flame at the edges of the slit is decolorized by introducing a little hydrogen chloride into it. If one then looks along the slit, the “inner cone” stands out against the background of the darker water-gas region in the form of a luminous tunnel. The energy of the reaction imparts great brightness not only to the reacting substances of the cold inner cone, but also to the foreign sodium vapors admixed with them. In order to prove the “chemical” origin of the luminosity in such cases, it was necessary to try to separate this luminosity from thermal luminosity and obtain it in pure form. For this purpose the authors began to seek a cold flame, i.e. one whose temperature would be below the ignition temperature. In that case the emitted light could not be ascribed to heating, and the spectrum obtained would be the pure spectrum of chemiluminescence.

Such a flame was successfully produced by passing nitrogen, almost saturated with sodium vapor, into gaseous chlorine, into oxygen, into bromine vapor, or into iodine vapor. On entering there, the sodium produced a small flame 4 mm in diameter and 5–10 mm high, the spectrum of which was investigated. The nitrogen, which diluted the sodium vapor, was added in order to lower the temperature of the flame: the heat of reaction was spent on heating, besides the small mass of sodium, a large quantity of nitrogen, and the temperature could not rise above 525°. Despite the reaction that was taking place, the spectrum proved to be the ordinary spectrum of sodium, from which the authors concluded that the centers of emission were sodium atoms that did not take a direct part in the reaction but were excited in some way by the energy released in the reaction.

Similar experiments were carried out with mercury vapors instead of sodium. Mercury combined with chlorine in a vessel, \(HgCl_2\). The emitted light, resolved by Hilger’s quartz spectrograph, gave the spectrum of the reaction product—the sublimate. Thus, in one case the chemiluminescence spectrum coincided with the spectrum of one of the reacting substances, in the other—with the spectrum of the reaction product. Not once was there obtained a spectrum of the reaction, distinct from them, more precisely, a spectrum of the reacting molecules, as might have been expected if one imagined, as was done formerly, the mechanism of chemiluminescence to be essentially different from the mechanism of radiation of an incandescent solid body, a heated gas, or a fluorescent liquid.

Recently, similar investigations of reactions in solid and liquid systems were carried out by Kautsky and Zocher\(^{14}\). These authors worked on the unsaturated silicon compounds obtained by Kautsky\(^{15}\), which readily pass into one another under the action of light and exhibit in a very vivid form the phenomena of chemi-, tribo-, cathodo-, and photoluminescence (phosphorescence and fluorescence). Solid calcium silicide \(CaSi_2\), when treated with hydrochloric acid in an alcoholic solution, gives various unsaturated silicon compounds:

1) colorless oxydisilane \(Si_2H_2O\), which oxidizes into

2) deeply colored (red) compounds of silical \((Si_2OH)X\), among which the hydrate of silical oxide \((Si_2OH)OH = Si_2H_2O_2\) is especially interesting. On oxidation they pass into

3) colorless leucone of unknown composition, which in turn, upon further oxidation, gives

4) silicic acid \(SiO_2\).

The transition of compounds of the 1st stage into compounds of the 2nd, and of the 2nd into the 3rd, is accompanied by very bright light effects, such as hitherto have not been observed in any case of chemiluminescence. The light emitted upon the oxidation of oxydisilane is visible even in daylight. Mixtures of compounds of the 1st and 2nd stages, of indeterminate composition, are called silicon; their color varies from light yellow to red, depending on the content of silical. All these substances, from oxydisilane to leucone, have the same physical structure, being pseudomorphs after oxydisilane: they consist of aggregates of the finest plates, permeated by microscopic pores. Therefore they possess an enormous surface, readily adsorb from solutions, for example, dyes, and quickly undergo chemical changes occurring at the surface.

In view of the extraordinary variety of transformations of light energy observed in these substances, they provide a whole series of data for theoretical considerations on the nature of chemiluminescence, fluorescence, and similar phenomena.

Dry silicalic-acid hydrate, or its mixture with a small amount of alcohol, when illuminated (under cooling) with blue or violet light, begins to glow with green or yellow light. This is, obviously, an instance of fluorescence, obeying Stokes’s rule, which is justified in many cases of fluorescence. It says that when substances possessing this property are illuminated by rays of short wavelength, e.g. blue, violet, or ultraviolet rays, they begin to emit light of longer wavelength.

Interesting considerations on the nature of fluorescence were recently expressed by Perrin¹⁶). He proposed that fluorescence is a consequence of chemiluminescence: light of short wavelength, possessing a large energy in each quantum, causes a chemical (photochemical) reaction, proceeding between minimal quantities of substances. As a result of this reaction, in turn, light is emitted, which we perceive as fluorescence, but which is in essence induced chemiluminescence.

This opinion of Perrin’s, however, was not confirmed. The facts on which he based his hypothesis were interpreted differently. Thus, it turned out that the bleaching and decomposition of a solution of eosin in alcohol under the influence of light depends to a high degree on the presence of oxygen in a definite concentration¹⁷), and that the addition of alkali, which does not affect the fluorescence, completely changes the character of the reaction products¹⁸). There is also a known example in which cooling to −180° does not weaken the fluorescence, while very strongly slowing the course of the reaction¹⁸).

The most striking experiment refuting Perrin’s theory was carried out by Kautsky and Zocher with silicalic-acid hydrate¹⁹). Under the action of light, this substance is bleached, passing to a further degree of oxidation, leucone. Thus, here there really is the intermediate photochemical reaction which Perrin regards as the source of chemiluminescence, usually taken for fluorescence. If Perrin’s view were correct, slowing this reaction should have weakened the fluorescence. The authors performed this experiment by lowering the test tube with the preparation into liquid air. Cooling to −180° almost stopped the reaction and, consequently, should also have almost destroyed the fluorescence. However, it turned out that if the preparation cooled by liquid air is illuminated with blue light, it begins to fluoresce extremely brightly, incomparably more brightly than at ordinary temperature; thus the slowing of the reaction is accompanied not by a weakening, but by an enhancement of the fluorescence. This observation forced rejection of Perrin’s view of an intermediate photochemical reaction as the cause of fluorescence. On the other hand, it led Kautsky and Zocher to the idea that the sources of chemiluminescence may be the same material objects that cause fluorescence, i.e. atoms

substances that do not take direct part in the chemical reaction, but receive the energy exciting their luminescence from the reacting particles. Further reasoning led them to the idea that the energy of the reaction may be transferred either to one of the reacting substances, as in the present case or in the experiment of Haber and Zisch with sodium vapors, or to a reaction product, as in the work of the same authors with mercury vapors, or, finally, to entirely foreign molecules taking no part whatever in the reaction itself. They decided to carry out this artificial, or “synthetic,” chemiluminescence experimentally. The problem presented considerable difficulties, since the reaction had to be accompanied by the liberation of a significant amount of energy: the energy released by the reacting molecule must be greater than the quantum of the emitted light energy. In addition, the added substance to which the energy of the reaction is to be transferred must satisfy a number of conditions. Thus, it must itself readily pass into an optically excited state (in the visible part of the spectrum); furthermore, it must be in very close contact with the substance participating in the reaction, so that the transfer to it of the reaction energy may be facilitated as much as possible.

Certain fluorescent dyes satisfy the first requirement. Their molecules, absorbing light energy of short wavelength, enter an optically excited state, manifested in luminescence. These same substances also possess the second necessary property, being readily adsorbed from solutions by bodies with a large surface. In the adsorbed state they come into the closest contact with the reacting substances, which facilitates the transfer to them of the reaction energy. In choosing a reaction that would transfer its energy to fluorescent dyes, it is therefore necessary to seek components with a well-developed surface. This property is found in unsaturated silicon compounds, with which Kautsky and Zocher worked. For experiments on artificial chemiluminescence they chose the oxidation reaction of isiline oxide, proceeding almost without any chemiluminescence. Before the start of the reaction, oxysiline was placed in a solution of a fluorescent dye, which it vigorously adsorbed from the solution, becoming colored by it throughout its mass. When oxidizing substances, e.g. acid permanganate, were added to such colored oxysiline, it passed into a hydrate of silicon oxide, emitting a bright luminescence of the fluorescence color of the dye adsorbed by it. One might have supposed that the dyes affect the color of chemiluminescence only by absorbing the rays of “chemical” origin passing through them. It turned out, however, that this supposition was incorrect. The color of the luminescence coincided not with the color of the dye itself in transmitted light, but precisely with the color of fluorescence. Thus, rhodamine B gave a bright red luminescence, red

rhodamine \(CG\)—yellow, acid eosin (Echtsäureeosin) red—yellow-green. The authors regard these experiments as convincing proof that the energy released in the oxidation reaction was transferred to extraneous particles of fluorescent dyes, optically excited them, and caused a glow characteristic precisely of the given dye. There is, it is true, one known case that does not fall under this explanation, where the dye glows with a color that does not correspond to its ordinary fluorescence spectrum: if, instead of oxidiciline, one takes silicone oxidized by permanganate to such a degree that by itself it no longer fluoresces, and adds to it isochinoline red, fluorescing with a yellow-red (orange) color, then the reaction produces green luminescence. The authors explain this by the fact that isochinoline red itself is oxidized by permanganate and hydrochloric acid into a yellow substance possessing green fluorescence. In the presence of reducing oxidiciline this reaction is delayed; if, however, it has all been oxidized by permanganate, this reaction proceeds freely, which is manifested in the distinct coloration of the glow.

This anomaly was not observed with other, more stable dyes.

The experiments of Kautsky and Pocher cannot be considered entirely complete and decisive. A large part of their observations is qualitative in character and must be confirmed by quantitative measurements. If their conclusions are justified, then the former theory of chemiluminescence will be changed in the sense that the emission of light, previously ascribed only to those molecules that participate in the chemical process, can also be ascribed to chemically unexcited molecules of the reaction components that have not yet entered into interaction, to particles of the reaction products, and even to certain extraneous substances that take no part in it. In all these cases, a transfer of energy is assumed from the reacting particles to particles not affected by the process and excited only optically. If earlier the discussion was of chemiluminescent reactions, then now one will be able to speak only of chemiluminescent substances that absorb the energy of the reaction and give it off in the form of light energy.

This conception will fill the deep gulf that exists between our ideas about the mechanism of chemiluminescence, on the one hand, and the mechanism of the emission of rays by solids and gases and fluorescence, on the other.

The new view should exert a considerable influence on contemporary views in the field of chemiluminescence and fluorescence, photochemical reactions, optical sensitization, and other phenomena that are still not fully elucidated.

In general, the study of chemiluminescence phenomena has only just begun, and further work in this field will undoubtedly bring a number of brilliant successes.

References

1. Joubert. Thèse sur la phosphorescence du phosphore. Paris, 1874.
2. Scharff. Zeits. f. phys. Chem. 62, 179 (1908).
3. Radziszewski. Ber. d. d. chem. Ges. 10, 70 (1877).
4. Trautz und Schorigin. Zeits. f. wiss. Phot. 3, 121 (1905).
5. Lifschitz und Kalberer. Zeits. f. phys. Chem. 102, 393 (1922).
6. Délepine. C. R. 150, 876 (1910); 154, 1171 (1912); 174, 1291 (1922).
7. Harvey. The Nature of Animal Light. Philadelphia & London, 1919.
8. Trautz. Zeits. f. phys. Chem. 53 (1905).
9. Centnerszwer und Petrikaln. Zeits. f. phys. Chem. 80, 235 (1912).
10. Grinberg. Journal of the Russian Physico-Chemical Society 52, 151 (1920).
11. Strutt. Proc. Roy. Soc. A 88, 547 (1913).
12. Stuchtey. Zeits. f. wiss. Phot. 19, 161 (1920).
13. Haber und Zisch. Zeits. f. Physik 9, 302 (1922).
14. Kautsky und Zocher. Zeits. f. Physik 9, 267 (1922).
15. Kautsky. Zeits. unorg. allg. Chem. 117, 209 (1921).
16. Perrin. Ann. de Phys. (9), 10, 133 (1918).
    ” ” ” ” 11, 5 (1919).
17. Weigert. Nernstfestschrift (1912), 465.
18. Pringsheim. Zeits. f. Physik 1, 176 (1922).
    ” Ergebnisse der exakten Naturwissensch. 1, 335 (1922).
19. (See 14). Zocher und Kautsky. Die Naturwissenschaften 11, 194 (1923).
    Kautsky und Zocher. Zeits. f. Elektrochemie 29, 208 (1923).