STUDIES ON THE DETERMINATION OF MICROQUANTITIES OF TIN AND LEAD IN GALLIUM ARSENIDE

A. A. KAPLIN, I. P. MAMONTOVA

(Presented by the scientific-methodological seminar of the Department of Physical Chemistry)

Recently, in the semiconductor-device industry, alloys of elements of Groups III and V of the Mendeleev periodic system have been finding ever broader application; in some of their properties they surpass the semiconductors germanium and silicon. It is also known what large changes, for example, in the properties of semiconductors are caused by microquantities of donor or acceptor electron impurities. Therefore, an important problem arising simultaneously with the intensive use of semiconductor materials is the analysis of these materials for the presence of traces of impurities. A comparatively large number of papers has been published in the literature [1–3] on the analysis, by various methods, of arsenic, gallium, and gallium arsenide; the method of amalgam polarography with accumulation has been applied to the analysis of semiconductor compounds comparatively recently.

The aim of the present work was to develop an amalgam-polarographic method for determining tin and lead in crystals of gallium arsenide.

Selection of the Supporting Electrolyte

In developing the method, it is very attractive to select a supporting medium for the determination of tin and lead on which anodic peaks of these elements can be obtained without separating the matrix. As such a medium, a solution of 0.1 N nitric acid was investigated. It was assumed that arsenic in nitric acid solution is in a polarographically inactive pentavalent state. Anodic peaks of tin and lead were obtained on this background (\(\varphi_{\mathrm{Sn+Pb}} = -0.5\) V), the heights of which depend linearly on concentration. As a result of studying the dependence of the anodic current of tin and lead on the time and potential of electrolysis, the conditions for the polarographic determination of tin and lead were selected: \(\varphi_{\mathrm{e}} = -1.0\) V, \(\tau_{\mathrm{e}} = 10\) min. The interfering effect of arsenic on the anodic peaks of tin and lead in 0.1 N nitric acid solution was studied. The tin peak decreases by half if \(9 \cdot 10^{-6}\) g/ml of arsenic is present in the solution, while the height of the anodic lead peak begins to decrease slightly if the arsenic concentration in the solution is \(2 \cdot 10^{-5}\) g/ml. In a gallium arsenide sample, it was not possible to obtain the anodic peaks of Sn and Pb because the electrode surface was coated with a brown film of elemental arsenic; consequently, 0.1 N nitric acid cannot be used as the supporting medium for the direct determination of Sn and Pb.

Tin is most often determined polarographically in supporting electrolytes containing a sufficiently high concentration of chloride ion; in this case a well-defined tin peak is obtained. In a chloride medium, the anodic lead peak is superimposed on the tin peak. To obtain separate peaks of tin and lead, we used a method based on preliminary recording of the combined Sn and Pb peak in a 3 N HCl medium and subsequent removal of tin in the form of its volatile bromides or chlorides¹. The dependence of the anodic peaks of Sn and Pb on the concentration in solution in a 3 N HCl medium is linear, which makes it possible to apply the method of standard additions. The dependences of the anodic peaks of Sn and Pb on the potential and electrolysis time in a 3 N HCl solution were studied. Electrolysis conditions: $\varphi_{\mathrm{e}} = -1.0\ \mathrm{V}$, $\tau_{\mathrm{e}} = 5$ min.

Effect of the matrix on the anodic peaks of tin and lead

Effect of gallium

According to literature data, in an HCl medium gallium forms the compound GaCl$_3$, which is soluble in HCl. In [4] it is noted that gallium, although it does not substantially interfere with the determination of Sn and Pb, somewhat lowers the sensitivity of their determination.

The half-wave potential of gallium in an HCl supporting electrolyte is about $-1.4\ \mathrm{V}$. Apparently, under our conditions ($\varphi_{\mathrm{e}} = -1.0\ \mathrm{V}$), gallium is practically not reduced at the electrode. Thus, in chloride supporting electrolytes, amalgam-polarographic determination of impurities is possible in the presence of an excess of gallium.

Effect of arsenic and the possibility of its removal

In studying the effect of arsenic (As$^{III}$) on the anodic tin peak in a 3 N HCl solution, we found that upon addition of $2 \cdot 10^{-6}\ \mathrm{g/ml}$ As, the anodic tin peak decreases already by 78% of its initial value. Arsenic also has a noticeable effect on the anodic lead peaks. The suppression of the anodic peaks is associated with the reduction of arsenic on the mercury electrode in the potential region under study.

A film of elemental arsenic forms on the mercury electrode; in this case, the overvoltage for hydrogen evolution decreases, the diffusion of metal atoms in the bulk of the electrode is hindered, and the current efficiency of the metal decreases.

Most authors, in order to remove arsenic, use its distillation in the form of volatile chlorides or bromides [3]. In our case, arsenic distillation cannot be used, since tin chloride is volatile ($t^\circ = 114^\circ$C). Therefore, extraction was used to separate arsenic. Tributyl phosphate was tested as an extractant for separating As and Ga, followed by determination of impurities in a 3 N HCl medium. A series of analyses of a GaAs sample was carried out after its decomposition in a mixture of hydrochloric and nitric acids (3:1); reproducible results could not be obtained, since arsenic is not completely extracted, is reduced at the electrode, and interferes with the polarographic determination of Sn and Pb. The studies described above showed that arsenic must be present in solution in an amount less than $2 \cdot 10^{-6}\ \mathrm{g/ml}$. When tributyl phosphate is used as an extractant for arsenic, a number of complicating circumstances arise that increase the error of determination.

1. The residues of the organic phase are ignited several times with HClO₄ + HNO₃ at \(t = 80^\circ\)C, which entails losses of tin and increases the duration of the analysis.

2. If the organic phase is not completely charred, then tributyl phosphate adsorbed on the electrode and the products of its incomplete combustion reduce the sensitivity and accuracy of the determination².

3. Residual amounts of HClO₄ are reduced on the electrode and increase the residual current. The above-listed factors diminish the reproducibility of the results, the sensitivity of the analysis, and increase its duration.

It was decided to select another extractant for separating arsenic from the elements under study. The most suitable extractant for our case proved to be benzene [5]. Arsenic is extracted almost 100%, while the impurity elements remain in the aqueous phase (Table 1).

Table 1

Results of the extraction of tin and lead with benzene from a hydrochloric acid medium

On the basis of the investigations carried out, a procedure was developed for the determination of Sn and Pb in GaAs (Table 2).

Table 2

Content of tin and lead in gallium arsenide

Conditions: \(\varphi_{\text{e}} = -1.0\ \text{V}\) (relative to sat. cal. el.); \(\tau_{\text{e}} = 5\) min.; \(S = 2 \cdot 10^{-8}\ a/\text{mm}\);
\(V = 4\ \text{ml}\); \(\varphi_{\mathrm{Sn+Pb}} = -0.48\ \text{V}\) (relative to sat. cal. el.); \(n = 6,\ \alpha = 0.95\)

Course of the Analysis

A 0.1 g sample of gallium arsenide is dissolved in two ml of a mixture of hydrochloric and nitric acids (3:1) at \(t^\circ = 70^\circ\)C. The solution is evaporated to dry salts in a quartz beaker at the same temperature. The dry residue is then dissolved in 10 ml of 12 N HCl, and arsenic is extracted with benzene (10 ml) three times for 5 min each. After settling, the aqueous phase is transferred to a quartz beaker and evaporated at 70°C to a volume of one milliliter. Then 3 ml of bidistilled H₂O is added to the same beaker; the amalgam-polarographic determination of tin and lead is carried out in the resulting background solution (3 N HCl). On the polarogram a combined peak of these elements is recorded. Tin is removed in the form of the readily volatile chloride (HCl + H₂O₂). The height of the anodic tin peak is calculated from the difference of the anodic peaks, and the contents of Sn and Pb are calculated by the method of additions. Several drops of a 0.05% potassium bromate solution are added to the solution for polarography in order to convert As³⁺, if it remained after extraction, into the form

² The influence, established in the work, of microquantities of surfactants on the anodic peaks of Sn requires additional investigations.

$\mathrm{As}^{V}$, which is polarographically inactive. On the basis of the developed procedure, an analysis of gallium arsenide samples was carried out.

Conclusions

1. An amalgam-polarographic procedure has been developed for the determination of tin and lead in gallium arsenide samples.

2. The influence of arsenic on the anodic peaks of tin and lead in a background of $0.1N\ \mathrm{HNO}_{3}$ and $3N\ \mathrm{HCl}$ has been studied.

3. An extractant has been selected for the complete separation of arsenic from the impurities being analyzed.

References

1. P. I. Artyukhin. “Zavod. labor.,” 33, 677, 1967.

2. “Methods for the Analysis of Gallium Arsenide.” Moscow, 1970.

3. Analytical Chemistry of Rare Metals and Semiconductor Materials. Seminar Proceedings. Moscow, 1970.

4. N. F. Zakharchuk, V. A. Trukhacheva, I. G. Yudelevich. Izv. SO AN SSSR, No. 2, 1970, 1, 73.

5. N. M. Kuzmin, G. D. Popova, I. A. Kuzovlev, V. S. Solomatin. ZhAKh, vol. 24, issue 6, 1969, 899.

1. Later we showed the possibility of separating Sn and Pb by introducing NaF into the solution being analyzed. ↩