Reports of the Academy of Sciences of the USSR
1960. Vol. 133, No. 5

PHYSICS

F. I. Vilesov and Academician A. N. Terenin

PHOTOELECTRIC EMISSION FROM SOLID LAYERS OF ORGANIC DYES

Photoemission from solid layers of organic dyes was studied in works (¹) in the spectral region down to 1900 Å. For various classes of dyes, a sharp increase in the quantum yield of photoemission was established beginning at 2000–2200 Å, which corresponds to a quantum energy of 6–5.5 eV. Recently one of the authors (²) showed that the photoionization thresholds of various classes of dyes in the gas phase also differ little from one another and lie in the region of quantum energies 7.0–7.3 eV.

In the present work, measurement of the energy distribution of photoelectrons was carried out by the retarding-field method in a spherical condenser made in the form of a platinized glass sphere 15 cm in diameter. The spectral dependence of photoemission was studied by the usual method in the same condenser. In some cases, a Geiger counter, previously used in work (³), was also employed as the photoelectron detector; in this case the current sensitivity of the apparatus was \(10^{-19}\) A. To decompose the ultraviolet radiation of a powerful hydrogen lamp, a vacuum monochromator described earlier (⁴) was used. The intensity of the monochromatic light flux incident on the photocathode was measured with an FEU-19m electron multiplier coated with sodium salicylate, having a constant quantum yield over a broad spectral region (⁵). The dye layers were thin polycrystalline films, obtained by sublimation in vacuum or by deposition from the corresponding solutions onto nickel disks 12–15 mm in diameter. Since most dyes decompose or sublime on heating, preliminary heat treatment of the layers in vacuum for degassing was not carried out. As we showed earlier (³), some adsorbed gases can significantly change the work function; therefore the work-function values given in Table 1 may differ from the true values. However, adsorbed gases cannot significantly change the distribution of photoelectrons over energies.

In the present work, layers of rhodamine 6G, erythrosin, β-carotene, metal-free phthalocyanine, and zinc phthalocyanine were investigated. The spectral distribution of the photoemission yield in relative units for some of the dye layers studied is presented on Fig. 1 on a semilogarithmic scale. The distribution of photoelectrons by kinetic energies for layers of rhodamine 6G and β-carotene at various photon energies is given in Figs. 2 and 3. The values of the photoelectric work function, determined from the Einstein equation from the distribution of photoelectrons over energies, and the position of the Fermi level, determined from its position for the substrate metal, are given in Table 1. The measurement errors were of the order of ±0.1 eV.

For rhodamine 6G, in the distribution of photoelectrons over energies (Fig. 2), throughout the investigated spectral region (6.5–10 eV), two groups were observed—

electrons for specimens prepared by different methods. It was established that the character of the electron energy distribution (the positions of the maxima) and the photoelectric work function do not change either with the thickness of the dye layer (0.01–1.0 μ), or with the method of preparing the layer (sublimation in vacuum at a pressure of \(10^{-5}\) mm Hg or deposition from solutions), or with the nature of the solvent (water, alcohols). The group of slow electrons (\(s\)) is characterized on the distribution curves by a maximum which, when the energy of the light quanta is increased to 10 eV, remains in place within the experimental error, broadening somewhat toward higher energies. The maximum of the group of fast electrons (\(f\)), as the energy of the quanta is increased, is regularly shifted toward higher energies by an amount equal to the increase in the quantum energy. In the case of β-carotene (Fig. 3), in the photoelectron energy distribution in the broad spectral region 6.0–9.5 eV only the group of slow electrons (\(s\)) is observed, and only for light-quantum energies greater than 10 eV does the group of fast electrons (\(f\)) appear.

Table 1

The energy distribution of photoelectrons for the other dyes investigated is qualitatively very similar to the two cases described. Thus, for erythrosin and zinc phthalocyanine, as for rhodamine 6G, we have two groups of electrons, but the group of fast electrons in these dyes is less clearly expressed, since its intensity is considerably smaller. For metal-free phthalocyanine, the energy distribution is in general features analogous to that of β-carotene, i.e., there is only one group of slow electrons.

Fig. 1. Spectral distribution of the quantum yield of the external photoeffect from layers: 1 — erythrosin, 2 — rhodamine 6G, 3 — β-carotene.

Analogous energy distributions of photoelectrons have been observed from the surface of inorganic semiconductors by a number of authors \((^{6})\). To explain the maximum of the group of slow electrons in these works, two mechanisms were proposed: (a) an exciton mechanism and (b) photoemission of electrons from a broad filled band. Since in the case of dyes we do not have broad filled bands, as is evidenced by the small difference between their absorption spectra in different aggregate states (gas phase, dilute solutions, solid phase), and at the same time the position of the (\(s\)) maximum practically does not change in the region of quantum energies 6–10 eV, both proposed mechanisms in the present case cannot be applied without invoking additional hypotheses.

The principal conclusions that follow from the present work are as follows: 1) the photoelectric work function of dye layers has a value of 5.5–6.0 eV; 2) for different dyes a close coincidence has been established of the spectral photoeffect curves in the region of quantum energies 6–10 eV; 3) a group of slow electrons is observed for all the layers investigated in this work; this makes it possible to suppose that the slow photoelectrons for all the dyes investigated have the same origin.

In organic dyes the least firmly bound electrons are the delocalized $\pi$-electrons. In this connection one should note the great similarity of the spectral curves of photoemission to the curves of photoionization, accompanied by the detachment of one of the $\pi$-electrons, of complex organic molecules in the gas phase. In addition, we have already noted that the photoionization thresholds of dye molecules of different classes lie very close to one another, within $7.0$–$7.3$ eV. The similarity of the data on the photoeffect and photoionization cannot be regarded as accidental, since dye molecules in a crystal interact comparatively weakly with one another and, to a certain degree, may be considered as having retained their electronic structure.

Fig. 2. Distribution of photoelectrons by energies for a layer of rhodamine 6G at different light-quantum energies

Proceeding from these data, the origin of the group of slow electrons may be explained as follows. The absorbed energy of a light quantum is spent not only on detaching a $\pi$-electron and imparting kinetic energy to it, but also on excitation of the positive ion formed in the process.

Fig. 3. Distribution of photoelectrons by energies for a layer of $\beta$-carotene at different values of the light-quantum energy

Part of the energy of the absorbed quantum is thus dissipated into intramolecular vibrations and electronic transitions in the positive ion, which leads to failure of the basic law of pho-

photoeffect (the increase in the kinetic energy of the electrons with increasing quantum energy). These viewpoints are also in complete agreement with the results on the photoionization of molecules. For example, it has been shown that upon photoionization of NO, simultaneously with the removal of one of the π-electrons, excitation of high vibrational levels of the ion occurs \((^7)\).

Part of the photon energy dissipated in exciting the positive ion depends on the structure of the molecule, in particular on how much the interatomic distances of the molecule and of the corresponding positive ion differ, and on the type of electrons being removed. If a light quantum is absorbed by an electron participating in a chemical bond, then considerable dissipation of energy into intramolecular vibrations should be expected. Conversely, if excitation and removal of an electron not participating in a bond occur, such dissipation of energy should be insignificant, and a group of fast electrons may be observed in the energy distribution of the photoelectrons. For different types of electrons participating in chemical bonds, the dissipation of energy should also be different.

In the case of the conjugated bond system of β-carotene, there are no valence electrons that do not participate in a chemical bond; therefore, the appearance of a group of fast electrons at quantum energies above 10 eV and the sharp increase in the photoemission yield in this region can be explained by the removal of σ-electrons.

The assumption discussed above—that the energy of the absorbed quantum is dissipated into intramolecular vibrations or electronic transitions—does not exclude the mechanism of dissipation of the kinetic energy of photoelectrons as they pass through layers of a solid, considered in a number of works by other authors \((^8)\).

Physical Institute
of Leningrad State University
named after A. A. Zhdanov

Received
9 V 1960

REFERENCES

1. F. Hlučka, Zs. Phys., 81, 66 (1933); 95, 486 (1935); 103, 246 (1936); 104, 653 (1937); A. Predvoditelew, N. Nechajewa, Z. Physik., 29, 332 (1924).
2. Ф. И. Вилесов, ДАН, 132, No. 3 (1960).
3. Ф. И. Вилесов, А. Н. Теренин, ДАН, 125, 1053 (1959).
4. Ф. И. Вилесов, Приборы и техн. эксперимента, 4, 89 (1958).
5. K. Watanabe, J. Opt. Soc. Am., 43, 32 (1953).
6. А. Н. Арсеньева-Гайль, Внешний фотоэффект с полупроводников и изоляторов, Moscow, 1957; О. М. Сорокин, ЖТФ, 28, 1413 (1958).
7. K. Watanabe, J. Chem. Phys., 22, 1564 (1954).
8. L. Apker, E. Taft, J. Dickey, J. Opt. Soc. Am., 43, 78 (1953); E. Taft, L. Apker, J. Opt. Soc. Am., 43, 81 (1953); L. Apker, E. Taft, Phys. Rev., 82, 814 (1951).