Analyst, October, 1967, Vol. 92, $$. 645-649 645 Spectrographic Determination of Beryllium in its Minerals with a Gas-stabilised Arc BY M. D. MARINKOVIC (The Boris Kidric' Institute of Nuclear Sciences, Belgrade, Yugoslavia) AND A. M. ANTIC-JOVANOVIC (Imtitute of Physical Chemistry, Faculty of Sciences, University of Belgrade, Yugoslavia) A gas-stabilised arc in argon is described for the determination of beryllium in its minerals. The analytical procedure used involves the decom- position of the sample by fusing it with sodium fluoride, dissolving the melt, atomising the solution in an argon stream and excitation of the spray. Experimental variables that may influence the results, and the behaviour of some elements as internal standards, are investigated. When molybdenum was used as the internal standard, the coefficient of variation achieved was about 2.5 per cent.The method was checked on some natural and synthetic beryls. MANY spectrographic techniques for determining beryllium traces in various materials have been reported. However, the number of papers published on the spectrochemical determina- tion of beryllium in its minerals (as beryls, etc.) is rather Conventional chemical methods generally used for this purpose suffer because of the presence of elements with similar chemical properties, e.g., aluminium. This is a source of many difficulties. Emission- spectrographic methods depend much less on chemical properties but give far less accurate results at higher concentrations when, as a rule, they are not satisfactory.Unsatisfactory accuracy is mainly caused by the unsteady operation of the light source and unreproducible volatilisation from the solid sample, which are especially important when powdered ores and minerals are analysed by the d.c. arc technique. By applying a gas-stabilised arc, which has recently been used in different ~ a y s , ~ , ~ ~ ~ 3' 3 8 9 9 and introducing the sample into the analytical gap by atomising the solution in a stabilising stream, conditions have been established for the determination of elements at higher concentrations with satisfactory accuracy. In the present paper a method for the determination of beryllium in beryls is described that has fairly good accuracy and reproducibility. EXPERIMENTAL DESCRIPTION OF THE SOURCE- The first experiments were carried out with the stabilised arc described by MarinkoviC.1° However, this type of source was shown to be unsatisfactory because of the condensation of aerosols inside the chamber and the possibility of contamination with solutions used previously.In our further work we used an improved modification of the stabilised arc (Fig. 1). The source consists of electrically insulated segments, D, E, F, G and H. The arc burns between two graphite electrodes, B and C, the lower one, B, being situated in the chamber. Segments D, E and F are water-cooled and comprise the lower part of the chamber, into which a weak argon stream is introduced to prevent the lower electrodes from being consumed. Segments G and H comprise the upper part of the chamber into which the aerosol of the sample is tangentially introduced.This part of the chamber is not water-cooled646 MARINKOVIC AND ANTIC- JOVANOVIC SPECTROGRAPHIC DETERMINATION [klndyst, VOl. 92 and its temperature while the arc is burning is maintained above 100" C, whereby condensation of aerosol drops is prevented. Between segments D and E, and segments F and G, gaps are left for the outflow of argon and aerosol. The atomising pressure was 2 atmospheres, while the consumption of gas was at the rate of 4.5 litres per minute. The atomising rate of sample solutions was 4 ml per minute, with an efficiency of about 3.5 per cent. The arc was ignited by a high frequency spark. The holder of the lower electrode, A, is shaped so that it can be set in a de Grammon's holder, thus allowing the gas-stabilised arc assembly to be easily mounted on the standard optical bar of the spectrograph. CHOICE OF THE INTERNAL STANDARD AND WORKING CONDITIOXS- The most intense beryllium lines in our source are the spark lines beryllium(I1) 3131.07 and 3130-43 A.The former is used as the analytical line, while the latter interferes with one line of the OH band. It has been ascertained from table+ that only two elements, vanadium and molybdenum, have sufficiently intense lines in this spectral range that could be used as an internal standard. These are the molybdenum(1) 3132.59 and vanadium(I1) 3110.71 A lines. C H, .- ? ;I E' D / cm f Scale Aerosol in - - A - B and C = D, E and F = G and H = Argon in -c- Brass holder Graphite electrodes Water-cooled electrically insulated segments Upper electrically insulated segments into which aerosol is introduced I nsu lators Fig. 1.Diagram of stabilised arc To choose the internal standard and determine the optimum working conditions, a series of experiments has been carried out, and almost all of the experimental conditions that could influence the intensities of the spectral lines were varied. After considering the results of previous work,l0 which showed that the emission of lines with this type of source takes place in relatively narrow zones of the arc column, radial distribution of the emission of the above beryllium, molybdenum and vanadium lines was investigated. Fig. 2 shows a radial distribution of emission coefficients for all three lines, calculated according to Bockasten's method.12 The curves show that the beryllium and vanadium lines emit in close regions of the arc column.It could be expected from this that accidental changes of excitation conditions could be better compensated for if a vanadium line is used as the internal standard instead of a molybdenum line. However, investigationOctober, 19671 OF BERYLLIUM IN ITS MINERALS WITH A GAS-STABILISED ARC 647 I I Radius, mm Fig. 2. Radial distribution of emission coefficients: curve A, for beryllium (11) 3131.07 A line; curve B, for vanadium (11) 3110.71 ii line; and curve C, for molybdenum (I) 3132.59 4 line of the influence of elements with a low ionisation potential showed that this assumption is not correct. Fig. 3 shows the influence of sodium on the intensity ratio of the line pairs beryllium(I1) - molybdenum(1) and beryllium(I1) - vanadium(I1).The influence of aluminium has also been investigated. In the range of concentrations normally found in the samples (up to 15 per cent.) no effect on the intensity ratio of the line pair, beryllium 3131.07 A - molybdenum 3132.59 A, was detected. O 6ol 01 0.2 03 04 05 Sodium, mg per ml Fig. 3. Influence of sodium on the intensity ratio of line pairs: curve A, beryllium (11) - vanadium (11) ; curve B, beryllium (11) - molyb- denum (I) The effect of changes in the arc current is rather small and almost the same on the intensity ratio of both line pairs. When the arc current changes from 5 to 16 amperes, the intensity ratio increases by about 10 per cent. On the other hand, at low current, the in- stability of the burning arc increases. Bearing in mind the above, an arc current of 8 amperes has been chosen as a compromise, with molybdenum as the internal standard.In a further investigation the absence of self-absorption for the chosen lines was established. ANALYTICAL PROCEDURE- Preparation of samples-Beryl samples were decomposed by fusion with sodium fluoride.13 The sample was finely ground in a synthetic corundum mortar, and 100 mg were fused with 500 mg of sodium fluoride in a platinum crucible for 5 minutes. After cooling, concentrated sulphuric acid (2 ml) was added and the contents heated to a gentle melt. The heating was continued until fumes of sulphur trioxide were evolved and then for a further 5 minutes.648 MARINKOVIC AND ANTIC- JOVANOVIC : SPECTROGRAPHIC DETERMINATION [APzaZyst, Vol.92 After cooling, the contents were extracted with water (about 100ml), and the solution was boiled until it was clear and then transferred into a 500-ml calibrated flask. Fifty millilitres of standardised ammonium molybdate solution (5 g of molybdenum per litre) were added to the flask and the solution was diluted to 500 ml with distilled water. This final solution was then atomised into an argon stream. Preparation of standard solutions-Five standard solutions were prepared by diluting a beryllium sulphate solution so that the concentrations of beryllium were 2, 5, 10, 20 and 40mg per litre. Each of these standard solutions also contained ammonium molybdate (500 mg of molybdenum per litre) as an internal standard, and sodium sulphate (550 mg of sodium per litre) and aluminium sulphate (20 mg of aluminium per litre) to match the sample solutions. Aluminium is always present in beryls and sodium is introduced in large amount during the decomposition with sodium fluoride, SPECTROGRAPHIC PROCEDURE- The samples were excited and photographed under the conditions listed below- Spectrograph : Wavelength range : Slit width : Excitation source : Electrode gap : Polarity : Open circuit voltage : Current : Exposure period : External optics : Hilger quartz Lit trow.2450 to 3500A. 0.02 mm. Gas-stabilised arc (Fig. 1). 30 mm. Lower electrode anode. 240 volts. 8 amperes. 15 seconds. The 10-mm section of the arc column, just above segment H (Fig. l), is focused on the spectrograph collimator with lens F1025.The upper electrode is screened off. The densities of the spectral lines were measured on a non-recording Zeiss densitometer and converted to relative intensities by using a Respectra calculator. The analytical graph was constructed by plotting the logarithm of intensity ratio of analytical pair lines against the logarithm of concentration of beryllium (Fig. 4). The background intensity was negligible and no correction was made. I 10 I00 Log concentration of beryllium (pg per ml) Fig. 4. Analytical graph. Beryllium (11) 3131.072 A line and molybdenum (I) 3132-594 A line RESULTS AND DISCUSSION Beryllium was determined, by the method described, in several natural beryls from different Yugoslav bearings.14 The results are given in Table I.The accuracy of the method was checked by determining beryllium in three synthetic beryls prepared by a laboratory procedure developed by RistiC.15 Conditions of synthesis were chosen so as to ensure that the composition accorded with the stoicheiometric formula of the mineral. The results obtained are given in Table 11, and, as can be seen, they show good agreement between the amounts of beryllium found and calculated.October, 19671 OF BERYLLIUM IN ITS MINERALS WITH A GAS-STABILISED ARC 649 TABLE I DETERMINATION OF BERYLLIUM IN SEVERAL BERYLS FROM YUGOSLAV BEARINGS Mean value, Sample Beryllium content, pcr cent. per cent. Cer . . . . .. 4.88, 4.97, 4.90, 4-95 4-92 Prokuplje . . . . 4.72, 4.85, 4.98, 4.74, 4-69, 4.78, 4.73, 4.72 4.78 Kukavica .. . . 4.52, 4.71, 4.78, 4-79, 4.51, 4.74, 4.70, 4.86 4.70 Zheljin . . . . . . 4.87, 4-73, 4.72, 4.67, 4-55, 4.67, 4.58, 4.67 4.68 Bukulja .. .. 4.64, 4.75, 4.56, 4.66 4-65 Juhor . . . . .. 4.80, 4.72, 4.71, 4.69, 5.00, 4.82 4.79 TABLE I1 DETERMINATION OF BERYLLIUM IN SYNTHETIC SAMPLES Number of Mean beryllium content found, Theoretical beryllium content, Sample determinations per cent. per cent. I 4 5-10 5.03 I1 4 5.06 5.03 111 6 5.05 5.03 which 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. The reproducibility of this method is characterised by the coefficient of variation calcu- lated for samples photographed in duplicate on separate photographic plates and for different days. For the concentration of 10mg of beryllium per litre of solution the coefficient of variation is 2.6 per cent.It is noted that by evaluating different plates the analytical graph shows a parallel dis- placement from one determination to another. This displacement can be considerable and it is recommended that at least one standard be photographed on each plate to fix the position of the analytical graph. The method described can also be successfully used for the analysis of other beryllium minerals and ores. It is only necessary to prepare standard solutions, the composition of correspond to those of the samples analysed. REFERENCES Alekseeva, V. M., and Rusanov, A. K., Zh. Analit. Khim., 1957, 12, 23. Greitz, E. B., U.S. Bureau of Mines Report No. 5407, Washington, D.C., 1958. Kehres, P. W., and Poehlman, W. J., Appl. Spectrosc., 1954, 8, 38. Margoshes, M., and Scribner, B. F., Spectrochim. Acta, 1959, 15, 138. Korolev, V. V., and Vainshtein, E. E., Zh. Analit. Khim., Owen, L. E., Appl. Spectrosc., 1961, 15, 150. Collins, A. G., and Pearson, C. A., Analyt. Chem., 1964, 36, 787. Rieman, M., in “Proceedings of the 12th International Colloquium on Spectroscopy,” Hilger, Doerffel, K., and Lichtner, J., Spectrochim. Acta, 1966, 22, 1245. Marinkovid, M., Bull. Boris Kidric‘ Inst. Nucl. Sci., 1965, 16, 65. Meggers, W. F., Corliss, C. H., and Scribner, B. F., “Tables of Spectral-Line Intensities,” Parts I Bockasten, K., J . Opt. SOC. Amer., 1961, 51, 349. Patkor, A. J., and Varde, M. S., Indian J . Chem., 1964, 2, 123. RistiC, S., AntiC-JovanoviC, A., and JeremiC, M., 1st. Symp. Geochem., Beograd, 1965, pp. 409 RistiC, S., “Physikalisch-Chemische Untersuchungen an Beryllen in Zusammenhang mit ihren Received January 17th, 1967 1959, 14, 658. Exeter, 1965, pp. 199 to 204. and 11, Natn. Bur. Stand. Monogr. 32, Washington, D.C., 1961. to 431. Helium,” Doktordissertation, University of Mainz, 1956.