Photolysis of Lead and Silver Azides and of Some Systems Based on Them

Yu. A. Zakharov, E. P. Abakumov, E. P. Surovoi

(Presented by the scientific seminar of the Department of Radiation Chemistry)

A number of investigators have reported data on the effect of additives coprecipitated with heavy-metal azides during their preparation on the photochemical stability of azides [1,2]. The observed effects are discussed, as a rule, within the framework of assumptions about the formation of solid solutions in the coprecipitated systems and of a model for the influence of introduced ions on the course of the elementary ionic and electronic stages of photolysis. A more careful study of the photolysis of coprecipitated azide systems shows, however, the complex nature of the influence of additives and the presence of a number of effects not previously detected or discussed.

In the present work—one of a series of studies we are carrying out on the photolysis of azide systems—new results are presented on the effect of additives on the photochemical decomposition of \(PbN_6\) and \(AgN_3\). The kinetics of decomposition was studied from the evolution of gaseous nitrogen—the product of photolysis. A schematic of the apparatus is given in Fig. 1. Light from the source (1), after passing through water (2) and glass band-pass filters (3), falls on the substrate bearing the deposited preparation. The azides were deposited on a flat glass plate in a layer about 1 mm thick, sufficient for complete absorption of the incident light. The pressure sensor was an LM-2 ionization manometer (4). The presence of a capillary (5), which is connected by a mercury seal (6), makes it possible to measure a quantity proportional to the rate of decomposition [3]. The system is pumped down to \(1 \div 2 \cdot 10^{-7}\) torr by a zeolite pump (7) cooled with liquid nitrogen. The sensitivity of the apparatus makes it possible to record a pressure change of \(1 \cdot 10^{-7}\) torr in the pressure range \(1 \cdot 10^{-7} - 9 \cdot 10^{-6}\) torr. Incandescent lamps and high-pressure mercury lamps served as the light source. The intensity

light, whose spectral range was selected by bandpass filters, was determined with the aid of a radiation thermoelement with an F-18 photomultiplier. The width of the spectral regions selected by the filters (at half-height of the intensities) was 30–40 nm. The sensitivity of the actinometric system was \(1 \cdot 10^{12}\) quanta/cm\(^2\)·s. The preparations were prepared by the exchange reaction between sodium azide and the nitrates of the corresponding metals.

I. Photolysis of silver and lead azides

The photochemical decomposition of pure preparations of lead and silver azides has common kinetic regularities, although the rate of photolysis of lead azide is approximately an order of magnitude lower than that of silver azide (when the photolysis is carried out under strictly identical conditions).

What is common in the photolysis of the azides studied by light in the band of intrinsic absorption \((\lambda_{\mathrm{AgN_3}} \le 365 \text{ nm};\ \lambda_{\mathrm{PbN_6}} \le 405 \text{ nm})\) is that, depending on the irradiation intensity, two forms of the kinetic decomposition curve are realized (two types of kinetics). The decomposition rate in fields of intense illumination \((J \ge 10^{14}\) quanta/cm\(^2)\), after rapidly reaching maximum values, then decreases (region A in Fig. 2, curve I), assumes a constant value (region B), and is then followed by region C with an increasing photolysis rate. The values of the photolysis rate in kinetic region A are very sensitive to the prehistory of the preparations and to the conditions under which photolysis is carried out. Preliminary heating of the azides in vacuum at temperatures \((60–140^\circ\text{C})\) below the temperatures at which their thermal decomposition begins, exposure to diffused white light, and aging of the samples smooth out the initial portion of the \(V-\tau\) curve, i.e., they reduce or even completely eliminate the rate maximum. The latter is observed after brief exposure or heating near the temperatures at which slow thermal decomposition begins. In this case the magnitude of the constant rate in region B practically does not change. For silver azide, studied in this respect, the photolysis rate in kinetic region A proved to be proportional to the square of the intensity of the incident light \((J^2)\), whereas at later stages of decomposition, for lead and silver azides, the photolysis rate is proportional to the intensity. The time required for the rate to reach maximum values for both azides does not change noticeably with

changes in intensity in the range \(10^{14}\)–\(10^{15}\) quanta/cm\(^2\)·s. When the light intensity in the intrinsic-absorption region is decreased, the value of the maximum photolysis rate gradually decreases, so that at \(J \leq 5\cdot 10^{13}\) quanta/cm\(^2\)·s the \(V\)–\(\tau\) curve changes to form II (Fig. 2). In the photolysis of lead and silver azides by light from the impurity-absorption region \((\lambda_{\mathrm{AgN_3}} > 366\ \mathrm{nm};\ \lambda_{\mathrm{PbN_6}} > 405\ \mathrm{nm})\), it is not possible to realize a kinetic curve of type I, even when the preparations are irradiated with the maximum light flux \((10^{15}\) quanta/cm\(^2\)·s). At incident-light intensities below \(5\cdot 10^{13}\) quanta/cm\(^2\)·s, decomposition proceeds according to a type-II curve. Increasing the light intensity eliminates the section of constant rate B, and the process proceeds with constant acceleration, according to curve type III in Figure 2.

The nonstationary portion of the photolysis kinetics of azides studied by us, observed at short existing irradiation times (and doses), had not been found by the authors in the available works on the photolysis of lead and silver azides and therefore had not previously been discussed. The presented data on the photolysis kinetics in the initial region can be discussed within the framework of two possible models. First, it is reasonable to assume that region A describes the intrinsic photolysis of azides, that is, its kinetics is not distorted by processes of the type of photodesorption or photolysis of impurities captured during crystallization (basic azides, carbonates); this is confirmed by data from mass-spectrometric analysis of azide photolysis products [4], showing that during illumination almost exclusively nitrogen is released. However, in view of the fact that in [4] the mass-spectrometric analysis of the products was carried out at the beginning of region C, the applicability of this conclusion to the region A under consideration should be additionally verified. According to the available concepts of the mechanism of photolysis of lead and silver azides,

\(V_{\mathrm{phot}} \sim k [N_3^{0}]^2\)

[5,6], and the kinetics of the change in the instantaneous concentration \(N_3^{0}\) must correspond to the profile of the \(V\)–\(\tau\) curve. In works [5,6] it was indeed shown that the photoconductivity in azides (p-type) changes upon illumination according to a curve with a maximum and a subsequent stationary value of the photocurrent. The rate of photodecomposition, taking into account recombination processes and trapping at deep traps (Tr), is determined by

by the expression:

\[ \frac{dV_{\varphi}}{dt} = k'\frac{dN_3^{0}}{dt} = kJN_3^{-}-k_2 eN_3^{0}-k_3N_3^{0}T-k_4\left(N_3^{0}\right)^2 . \tag{1} \]

The condition for a maximum has the form:

\[ \frac{dV_{\varphi}}{dt}=0 \quad \text{and} \quad -k_2\frac{de}{dt}N_3^{0}-k_3\frac{dT}{dt}N_3^{0}<0 . \]

For \(k_2\) and \(k_3\) constant over the course of photolysis, this condition is realized if \(dT/dt>0\), \(de/dt<0\), that is, when, during photolysis, the concentration \(N_3^{0}\) decreases and the concentration of free electrons increases owing to, respectively, an increase and a decrease in the number of deep traps for them. The feasibility of \(dT_e/dt<0\) is confirmed by the results for the dependence \(V=f(J)\). In region A, the dependence \(V-J^2\) can be realized under the condition of irreversibility of the first stage of photolysis:

\[ N_3^{-}+h\nu \longrightarrow N_3^{0}+e, \]

which, in turn, may occur under conditions of rapid capture of photoelectrons:

\[ e+T_e \longrightarrow T_e e . \tag{2} \]

Saturation of traps with electrons will lead to an increase in the probability of their recombination with \(N_3^{0}\), and to reversibility of stage (2), which explains the subsequently observed linear dependence of the photolysis rate on intensity. The photoconductivity in this region has the dependence \(i_{\varphi}\sim J^{0.5}\).

Thus, everything proposed confirms the probability that the inequality \(de/dt>0\) (i.e., \(dT_e/dt<0\)) exists.

Alongside this variant, an alternative explanation of the kinetic dependences obtained may be constructed from the assumption that the initial gas evolution is caused by processes of desorption or decomposition of captured impurities; moreover, the dependence of the decomposition rate of these products on intensity is sharper than the dependence \(V_{\varphi}=f(J)\). If it is further assumed that the consumption of \(T_e\) (or, in the second variant,

processes determining the initial gas evolution?) proceed via a monomolecular reaction, it can be shown that, in the region of high light intensities, the time required to reach \(V_{\max}\) depends only weakly on \(J\).

The acceleration of photolysis on section C, observed in all cases, is evidently associated with the accumulation of photolytic metal, which accelerates decomposition. As a probable scheme, one may propose the following photolysis scheme for this section, constructed on the assumption that electrons are transferred to metallic nuclei from the boundary regions of the azide:

\[ \begin{aligned} \mathrm{MeN_3^-} &\longrightarrow \mathrm{Me^- + N_3^0} \\ \mathrm{N_3^- + h\nu} &\longrightarrow \mathrm{N_3^0 + e} \\ \mathrm{N_3^0 + N_3^0} &\longrightarrow \mathrm{3N_2}. \end{aligned} \]

Thus, the process of photolysis of lead and silver azides may be represented as a superposition of two (in the first variant) or three (in the second variant) photochemical processes differing in mechanism (Fig. 3).

The distinctive kinetics of photolysis of azides by long-wavelength light must be associated with the specificity of impurity absorption. In this case, excitation of electrons apparently occurs to local impurity levels sufficiently deep that excitation of electrons from them into the conduction band at the photolysis temperature does not take place. The first stage of photolysis

\[ \mathrm{N_3^- + h\nu \xrightarrow{Te} N_3^0 + Te\,e} \]

is therefore irreversible, which corresponds to the dependence observed for silver azide,

\[ V_p \sim J^2. \]

The absence of electrons in the conduction band, and consequently of a stage of trap filling, leads to the absence of an initial maximum even at high light intensities.

In order to obtain additional information on the character of photolysis at individual kinetic stages, we carried out a series of experiments with interruptions. Interruption of decomposition in section B and po-

the second performance of photolysis, with a short interruption time (tens of minutes), converts the photolysis kinetics from type I to type II (Fig. 2), with a decrease in the decomposition rate. Exposure of the irradiated lead azide in air, or prolonged storage of it in vacuum (\(1 \cdot 10^{-6}\) torr), leads to the fact that, upon repeated photolysis, kinetics of type I is again realized. Illumination of lead azide with light of \(\lambda > 650\) nm causes photosorption; repeated decomposition by light from the intrinsic-absorption band proceeds along a curve approaching type I. These data show that, during the interruption period, restoration is possible of the special sites or defects responsible for the initial maximum of the photolysis rate.

If photolysis is interrupted in section C, the subsequent decomposition proceeds with the constant acceleration characteristic of section C (type III, Fig. 2).

II. Effect of additives on the photolysis of lead and silver azides.

The introduction of additives into azides, by coprecipitating them during preparation of the samples, has a very appreciable and complex effect on the photodecomposition of lead and silver azides. The results obtained in this direction are presented in Figs. 4, 5.

Acceleration by the additive \(Ag^+\) of the photolysis of lead azide on all kinetic curves, in the region of both intrinsic and impurity absorption, corresponds to the increase in photoconductivity in coprecipitated systems [6] and is naturally explained within the framework of concepts concerning the formation of a solid solution in the system, and taking into account the change in the equilibrium concentration of intrinsic microdefects in lead azide. The increase in such a solid solution of the concentration of anion vacancies accepting photoelectrons explains the increase in the initial maximum of the rate (for the reasons set forth above). The increase in the photolysis rate in sections A and B causes a natural decrease in the time until the appearance of section C (Fig. 4). The introduction of the additive \(Ag^+\) does not change the dependences, characteristic of lead azide, of the photolysis rate on the intensity of light and the dependence of the rate on wavelength. These results agree with the proposed causes of the influence of silver cations on the photolysis of lead azide. The effect-

Fig. 1.

Schematic of the apparatus for studying the rate of photodecomposition.

Fig. 2.

Observed types of kinetic curves during photodecomposition of lead and silver azides.

Fig. 3.

Contribution of individual processes at different decomposition depths:

1 — decomposition at defects;
2 — photolysis proper;
3 — effect of the metal.

Fig. 4.

Dependence of the photolysis rate of lead azide on the addition of the cations \(Cu^{++}\) and \(Ag^{+}\) (intensity \(10^{14}\) quanta/\(\mathrm{cm}^{2}\cdot\mathrm{s}\)).

Visible graph labels: \(V\), rel. units; \(\tau\), min; \(436\ \mathrm{nm}\); \(405\ \mathrm{nm}\); pure; \(0.5\%\,Ag^{+}\); \(+0.1\%\,Cu^{++}\); \(0.5\%\,Cu\).

Fig. 5.

Dependence of the photoconductivity (left) and the photolysis rate of \(AgN_{3}\) on the addition of the cations \(Cu^{++}\) and \(Pb^{++}\) (intensity \(10^{14}\) quanta/\(\mathrm{cm}^{2}\cdot\mathrm{s}\)).

Visible graph labels: \(\lg\); \(V\), rel. units; \(365\); \(436\); \(365\ \mathrm{nm}\); \(436\ \mathrm{nm}\); pure; \(0.5\%\,Pb\); \(0.1\%\,Cu^{++}\); \(0.5\%\,Pb^{++}\); \(\tau\), min.

Consequently, the influence of \(Ag^+\) entering the \(PbN_6\) lattice on the equilibrium concentration of intrinsic defects leads only to changes in the rates of the individual elementary ionic and electronic stages of photolysis; no change in the decomposition mechanism should be expected. Along with the absence of any effect on the dependence \(V_\varphi = f(J)\) and on the dispersion curve, we previously showed that \(Ag^+\) has no effect on the activation energy of lead azide photolysis.

We have already indicated earlier that copper(II) cations in lead and silver azides apparently play the role both of electron acceptors and of centers for recombination of free pairs. In addition, it turned out that copper azide is predominantly captured in the near-surface layers of azides. All this makes it difficult to explain the complex character of the influence of copper azide on photolysis. One can only note that the absence on the kinetic curves of photolysis of preparations containing copper azide of an initial maximum in the rate is naturally explained by the influence of \(Cu^{++}\) as a recombination center. The addition to the left-hand side of expression (I) of an additional term \(\left(-K_5 Cu^{++} N_3^0\right)\) decreases the probability that a maximum rate will be realized on the kinetic curves. The influence of a copper azide additive on the photoconductivity and the rate of photolysis of silver azide is qualitatively similar: the additive leads to an increase in both quantities upon irradiation with light in the intrinsic absorption band and decreases the photoconductivity and the photolysis rate in the long-wavelength region.

The character of the influence of \(Pb^{++}\) on the photolysis of silver azide (acceleration of the process under short-wavelength irradiation and retardation in the region of impurity absorption) qualitatively correlates with the influence of impurity ions \(Pb^{++}\) on the magnitude and spectral distribution of the photoconductivity in \(AgN_3\): the additive increases the photoconductivity in the intrinsic band and decreases it in the long-wavelength region.

Thus, the existence, in the cases studied, of a correlation between the influence of additives on the photoconductivity and on the photolysis rate of azides gives grounds to suppose that the principal physical (“electrophysical”) causes of their influence on the rate of photodecomposition are involved. A detailed clarification of the nature of the decomposition requires further experiments, in particular—on the laws-

regularities in the formation of solid solutions and the electrophysical properties of the systems.

Literature

1. H. D. Fair, Jr. and Q. C. Forsyth, “Reactivity of Solids,” edited by J. W. Mitchell et al., published by J. Wiley and Sons, Inc., 1969, 83–91.

2. D. A. Zakharov. Dissertation, Tomsk, 1963.

3. S. M. Ryabykh. Dissertation, Tomsk, 1968.

4. G. A. Kislin, D. A. Zakharov. Izv. TPI, 199, 44, 1969.

5. Yu. V. Gavrishenko. This collection.

6. G. G. Savel’ev, D. A. Zakharov, G. T. Pechkov. Journal of Physical Chemistry, 41, 1691, 1967.