EFFECT OF Cu++ AND Ag+ ADDITIVES ON THE THERMAL DECOMPOSITION, ELECTRICAL CONDUCTIVITY, AND PHOTOCONDUCTIVITY OF LEAD AZIDE

G. G. Savel’ev, Yu. A. Zakharov, G. T. Shchechkov, R. A. Vasyutkova

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

Earlier, some of us studied the effect of homophase impurities on the thermal decomposition and electrical conductivity of oxalate [1, 2] and silver azide [3]. It turned out that the homophase impurities Pb++, Hg++ and Cd++ reduce the rate of thermal decomposition of oxalate, while additions of Pb++ and CO₃-- do not affect the thermolysis of silver azide. The character of the change in electrical conductivity caused by additives is the same in both cases. The reasons for such behavior of AgN₃ were explained in [3] by localization of the thermolysis process in the region of disordered “precenters.” In investigating the reasons for such behavior of these preparations, we carried out studies with lead azide, which in its physicochemical properties is similar to silver azide. Thus, both azides belong to initiating explosives of approximately the same power, hardly differ in thermal stability, have low dark conductivity of ionic character, possess appreciable photoconductivity, and the edge of the absorption band of PbN₆ does not differ greatly from that of AgN₃ [4].

The rate of thermal decomposition of lead azide increases in the presence of graphite additives [5], powders of Pb, Zn, and Al [6, 7], and also when metallic silver is deposited on the surface of PbN₆ crystals [8].

We have also found the influence of mechanical impurities on the thermolysis of AgN₃ [9]. All these data indicate a great similarity of these substances. However, decomposition of lead azide begins at the crystal surface [9], whereas in AgN₃ nuclei arise throughout the volume in strongly disordered regions.

Experimental Part

Preparations of lead azide were obtained by precipitation when 0.2 N solutions of chemically pure sodium azide and lead nitrate were mixed. In preparing azide samples containing Cu++ and Ag+ additives, a calculated amount of 0.2 N solutions of Cu++ nitrate and Ag+ nitrate was added to the initial substances.

The operations for obtaining pure PbN₆ and azide with additives were carried out in diffuse daylight. The preparation of lead azide with an Ag+ additive, in view of the high photosensitivity of the latter, was carried out in red light. The azides obtained were dried and

were stored in a desiccator. We obtained preparations of pure lead azide, as well as preparations with additions of 1, 0.5, and 5 mol.% \(\mathrm{Cu}^{++}\); and 5 and 10 mol.% \(\mathrm{Ag}^{+}\).

In addition, we obtained single crystals of pure lead azide and with a copper additive. The single crystals were grown by the method of [11], from a 10% solution of ammonium acetate saturated at \(80^\circ\mathrm{C}\) with lead azide. The solution was cooled to \(20^\circ\mathrm{C}\) over 20 hours. To obtain single crystals with a copper additive, the solution was saturated with lead azide containing 5 mol.% \(\mathrm{Cu}^{++}\). The crystals obtained had a good shape and belonged to the rhombic system. The dimensions of the best crystals of pure \(\mathrm{PbN}_6\) were \(1\ \mathrm{mm} \times 0.5 \times 0.5\ \mathrm{mm}\), while the crystals with a copper additive were somewhat smaller; in addition, they were colored reddish brown (the crystals of pure \(\mathrm{PbN}_6\) were slightly yellowish).

We studied the thermal decomposition of the preparations on vacuum quartz microbalances having a sensitivity of \(1.2 \times 10^{-5}\ \mathrm{g}\) and described in [1, 2, 3].

The data on the thermal decomposition of powders are presented in Fig. 1 in coordinates of degree of decomposition \((\alpha)\)—time. Fig. 2 shows the kinetic curves of the thermal decomposition of single crystals.

Fig. 1. Thermal decomposition of finely crystalline preparations at \(245^\circ\mathrm{C}\). 1—preparation with an addition of 5 mol.% \(\mathrm{Ag}^{+}\); 2—preparation without additives; 3—preparation with an addition of 5 mol.% Cu.

Fig. 2. Thermal decomposition of single crystals at \(265^\circ\mathrm{C}\). 1—pure lead azide; 2—lead azide with an addition of 5 mol.% \(\mathrm{Cu}^{++}\).

It is seen that the addition of silver accelerates, while copper slows, the thermal decomposition of finely crystalline preparations. Single crystals with a copper additive, on the contrary, decompose at a higher rate than pure ones.

Kinetic analysis of the thermolysis curves of powder preparations showed that the data are well described by the Erofeev equation:

\[ 1-\alpha=e^{-k_1 t^n}, \tag{1} \]

and by the equation

\[ \alpha=k_2 t^n, \tag{2} \]

where in both cases

\[ n=\frac{1}{2}. \]

The activation energy, calculated by us from the temperature dependence of \(k_2\), proved to be \(11 \pm 2\ \mathrm{kcal/mol}\). The activation energy of the doped preparations does not differ appreciably from that of the pure preparation.

To study the conductivity of the preparations obtained in vacuum and in air, in the temperature interval \(25\text{–}190^\circ\mathrm{C}\), we used

a detachable measuring cell made of molybdenum glass, with four molybdenum wires sealed into it.

Two of them served for introducing the chromel–alumel thermocouple, while removable copper electrodes, made in the form of bent rectangular plates whose radius of curvature corresponded to the radius of the pellet, were placed on the others. The cell was closed with a ground-glass stopper.

To measure photoconductivity in the indicated temperature range, we used the same apparatus.

As the light source we used a 500-watt incandescent lamp with a broad spectrum.

To measure conductivity, we pressed pellets at pressures of 6 t/cm²; graphite electrodes were applied to their edges. The dimensions of the pellets were standard: thickness 0.55–0.65 mm, diameter 17 mm, weight 650–700 mg. To evacuate the apparatus we used a combination of an RVN-20 pump and a glass diffusion oil pump. Heating of the pellet was carried out in a tubular electric furnace placed over the projecting part of the cell stopper. After the required temperature had been reached, we measured the resistance of the pellet with an E6-3 teraohmmeter.

To eliminate pickup and ensure stable operation of the teraohmmeter, we used shielded lead wires and shielded the instrument inputs.

Our apparatus had the following parameters:

1. Pressure in the measuring cell during the experiment: \(10^{-4}\) mm Hg.
2. Upper limit of measured resistances: \(1 \times 10^{13}\) ohm.
3. Voltage applied to the pellet: 105 V.
4. Measurement error: \(\pm 5\%\).

The experimental results presented in Figs. 3 and 4 show that introducing additions of \(\mathrm{Cu}^{++}\) into lead azide decreases its dark and photoconductivity, whereas \(\mathrm{Ag}\) increases them. From the slopes of the straight lines in Fig. 3 we determined the activation energy of the dark conductivity, which proved to be \(22.4 \pm 2\) kcal/mol. Special experiments showed that air does not affect the electrical conductivity and photoconductivity of the preparations. Measurements of the electrical conductivity of single crystals were carried out in air at a temperature of 23°C. For the measurements, special microelectrodes were made, between which the single crystal was clamped under a microscope.

Table 1

Measurement of the electrical conductivity of single crystals*

* Here \(I_{\mathrm{t}}\), \(I_{\mathrm{uv}}\), and \(I_{\mathrm{i.l.}}\) are the currents: dark, under irradiation with ultraviolet light, and under an incandescent lamp.

Fig. 3. Dark conductivity of lead azide. 1—pure lead azide; 2—lead azide with an additive of 10 mol.% Ag+; 3—lead azide with an additive of 1 mol.% Cu++; 4—lead azide with an additive of 0.5 mol.% Cu++.

Fig. 4. Photoconductivity of lead azide. 1—pure lead azide; 2—lead azide with an additive of 10 mol.% Ag+; 3—lead azide with an additive of 1 mol.% Cu++; 4—lead azide with an additive of 0.5 mol.% Cu++; 5—lead azide with an additive of 5 mol.% Cu++.

The measurements were carried out using a U1-2 dc amplifier. To measure photoconductivity, in addition to an incandescent lamp, a source of short ultraviolet radiation of the BUF-30 type (a low-pressure mercury lamp) was used.

The data for a single crystal are summarized in Table 1, from which it is seen that the data obtained on single crystals agree well with the data obtained on pressed tablets.

Discussion of the Results

If, by analogy with AgN\(_3\), it is assumed that the conductivity of PbN\(_6\) in the dark is ionic, then the results obtained on the effect of additives are in qualitative agreement with the Frenkel and Schottky theory of disordering in ionic crystals [12]. However, the effect of a Cu\(^{++}\) additive on the electrical conductivity of PbN\(_6\) requires a special explanation, since this impurity has a valence identical to that of Pb\(^{++}\), and its incorporation into lattice sites should not be accompanied by a change in the concentration of ionic defects. If, however, the Cu\(^{++}\) additive is incorporated into interstices, then, to compensate its charge, some of the positively charged defects (anion vacancies or interstitial cations) must disappear, or the number of negatively charged ionic defects (cation vacancies) must increase.

In this case the observed effect can be explained if it is assumed that the positive defects are predominantly mobile.

The introduction of Ag\(^+\) ions into the lattice should lead to an increase in the concentration of positive ionic defects in order to compensate the missing charge of the impurity in comparison with Pb\(^{++}\), if the impurity is incorporated into the lattice in place of lead. In this case the substitution mechanism seems more probable, since the radius of the silver ion does not differ greatly from that of lead and is 1.13 Å. The data we obtained on the dark conductivity of preparations with a silver additive confirm this point of view.

Concomitantly with the change in ionic conductivity, the photoconductivity also changes under the influence of additives (although one would rather have expected an opposite dependence), which may be connected with the acceptor properties of the Cu\(^{++}\) ion and, possibly, the donor properties of the Ag\(^+\) ion.

It is of interest to compare these data with the effect of additives on thermal decomposition. In [14], on the basis of general considerations, the assumption was made that the stage limiting the process of thermal decomposition of PbN\(_6\) is the bimolecular reaction of holes: \(2\mathrm{N}_3 \xrightarrow{K_3} 3\mathrm{N}_2\). The stage of hole formation in this case may be considered close to equilibrium:

\[ \mathrm{N}_3^- \underset{K_2}{\stackrel{K_1}{\rightleftarrows}} \mathrm{N}_3^0 + e. \tag{1,2} \]

In this case, a considerable decrease in the concentration of electron traps should lead to an increase in the probability of reaction 2, which should lead to a decrease in the rate of the entire process. On the other hand, an increase in the concentration of electron traps and the accompanying increase in the concentration of cation vacancies may lead to an increase in the constant \(K_3\), owing to facilitation of reaction 3 in the vicinity of this defect. This mechanism of acceleration should apparently be accompanied by a decrease in the activation energy of the process.

The data on the effect of Cu\(^{++}\) on the decomposition of single crystals can be explained on the assumption that the process is limited by (stage 3).

Indeed, the photoconductivity data show that \(\mathrm{Cu}^{++}\) is an effective trap for electrons. This should reduce the probability of recombination by reaction 2, which will lead to an increase in the lifetime of the \(\mathrm{N}_3\) radicals and in the probability of reaction 3.

Unfortunately, we have not yet succeeded in growing single crystals with an \(\mathrm{Ag}^+\) additive because of the large difference in the solubilities of Ag and Pb azides. A study of such single crystals could clarify this question.

As kinetic analysis has shown, the decomposition of microcrystalline powders is well described by diffusion equations, since the exponents \(n\) in both equations used are equal to \(1/2\) [15, 16].

This agrees with the results of [17], in which diffusion kinetics was also observed; moreover, the activation energy calculated by us for these data also proved to be equal to 11 kcal/mol. In this case, ionic processes may play a larger role than electronic ones. The parallelism of the effect of impurities on ionic conductivity and thermal decomposition may indicate a substantial role of ionic defects in the rate-limiting thermolysis-diffusion process.

REFERENCES

1. V. V. Boldyrev, Yu. A. Zakharov et al. Kinetics and Catalysis, 4, no. 5, 1963.
2. Yu. A. Zakharov. Candidate’s dissertation, Tomsk, 1963.
3. Yu. A. Zakharov, G. G. Savel’ev et al. Kinetics and Catalysis, 5, no. 5, 1964.
4. F. Bowden, A. Ioffe. Fast Reactions in Solids, IL, Moscow, 1962.
5. Deb, B. Evans, A. Ioffe, VIII Symp. on comb. Baltimore, 829, 1962.
6. A. Ya. Apin. Collection of Articles on the Theory of Explosives, Oborongiz, Kiev, 1940.
7. S. M. Moskovich, V. A. Aleksandrovich. ZhFKh, 6, no. 9, 1935.
8. B. Reitzner, I. Kaufmann, E. Bartell, J. Phys. Chem., 421, 1962.
9. Yu. A. Zakharov, E. S. Kurochkina, G. G. Savel’ev. Kinetics and Catalysis (in press).
10. V. Garner, A. Gomm. Collection of Articles on the Theory of Explosives, Oborongiz, Kiev, 1940.
11. H. Miles. Collection of Articles on the Theory of Explosives, Oborongiz, Kiev, 1940.
12. K. Hauffe. Reactions in Solids and on Their Surfaces, 1, IL, Moscow, 1962.
13. Chemist’s Handbook, 1, Moscow–Leningrad, 1963.
14. Griffiths, Groocock, J. Chem. Soc., 3380, 1955.
15. G. V. Sakovich. Candidate’s dissertation, Tomsk, 1956.
16. V. V. Boldyrev. Methods for Studying the Kinetics of Thermal Decomposition of Solids, Tomsk, 1958.
17. B. Reitzner, Abel, I. Kaufmann, Nature, 192, 626, 1961.