Determination of Parts-per-MiIIion Qua’ntities of TelIurium in Various Alloys by X-Ray Spectrometry Keith E. Burke, M. M. Yanak, and C. H. Albright Paul D . Merica Research Laboratory, The International Nickel Co., Inc., Sterling Forest, Suffern, N . Y . 10901 I n a rapid and accurate X-ray spectrographic determination of 1 to 200 ppm of tellurium in nickel-, iron-, aluminum-, and copper-base alloys, tellurium is separated from its matrix, after dissolution of the sample and the removal of silica and selenium, by the reduction of tellurium(1V) with tin(ll) chloride in solutions 3 to 5M in hydrochloric acid. The resulting suspension of colloidal tellurium is filtered on a micropore disk. The final operation involves a direct x-ray spectrographic measurement of the tellurium on the filter disk.
SPECTROGRAPHIC METHODS for the determination of tellurium in a variety of alloys are either lacking in sensitivity or subject to interferences (])-for instance, the direct x-ray determination of tellurium in the concentration range of interest is not practical. A concentration technique must be employed to increase spectral line intensity (2). Optical emission spectrographic methods also lack sensitivity and are subject to chromium interference. The problems associated with the determination of tellurium in metallurgical systems have been discussed and a cathoderay polarographic method proposed for the determination of tellurium at the 20- to 800-ppm level (3). Tellurium is separated from copper-base alloys with ammonium hydroxide, using iron as a carrier, and from iron-base systems by precipitation with sulfur dioxide using selenium as carrier, prior to the polarographic determination. Neutron activation has also been successfully used to determine trace concentrations of the element (4). This method is highly accurate; but it requires three precipitations and three scavenging steps plus an anion exchange separation of selenium and tellurium prior to the final precipitation of tellurium (IV) oxide. We have developed a rapid method for concentration and subsequent x-ray determination of less than 100 ppm of tellurium. It consists of four steps: dissolution of the sample; prefiltration of the solution to remove any precipitate such as silica, or selenium which is selectively reduced with hydroxylamine; separation of the colloidal tellurium which is precipitated with tin (11); and counting the dried micropore filter containing elemental tellurium with the vacuum x-ray spectrograph. EXPERIMENTAL Reagents. STANDARDTELLURIUM SOLUTIONS,500 and 5 pg per ml. Dissolve 0.5000 gram of high purity tellurium (Spex Industry, Metuchen, N. J.), 99.999%) in 10 ml of aqua regia and dilute to 1 liter with water. This solution contains 500 pg of tellurium per ml. (1) I. M. Kolthoff, P. J. Elving, and E. B. Sandell, “Treatise on Analytical Chemistry,” Vol. 7, Part 11, p. 192, Interscience, New York, 1961. (2) W. J. Campbell, E. F. Spano, and T. E. Green, ANAL.CHEM., 38, 987 (1966). (3) E. J. Maienthal and J. K. Taylor, Ibid., 37, 1516 (1965). (4) D. E. F. Morris and N. Hill, Metallurgia, 71,99 (1965).
14
0
ANALYTICAL CHEMISTRY
TIN(II) CHLORIDE, 1 gram per ml. Weigh 250 grams of stannous chloride, SnC12.2H20, into a 400-ml beaker, add 100 ml of concentrated hydrochloric acid, and place on a warm hot plate until the solution becomes clear. Stir occasionally. Transfer the solution to a 250-ml bottle and dilute to 250 ml with concentrated hydrochloric acid. Prepare this solution daily, as tin(I1) is readily air-oxidized. Use reagent grade chemicals and class A volumetric glassware for the preparation of all solutions. Procedure. DISSOLUTION. Accurately weigh the sample into a 100-ml beaker. Select a sample weight so that 5 to 100 pg of tellurium will be present. Dissolve the sample in 35 ml of concentrated hydrochloric acid with dropwise additions of nitric acid. If it is necessary to add more than 2 to 3 ml of nitric acid, either take the sample to near-dryness and remove the nitrate by repeated evaporations with hydrochloric acid or fume with 5 ml of perchloric acid. Add 5 ml of sulfuric acid when the sample is known to contain 100 mg or more of molybdenum, or to volatilize selenium partially when it is present as a major constituent. Adjust the volume to about 15 ml if selenium is known to be absent and proceed with the prefiltration. PREFILTRATION. Use a microfiltration apparatus (XX1002500, Millipore Filter Corp., Bedford, Mass.) fitted with glass fiber filters (934 AH, Reeve Angel, Clifton, N. J.) for rapid filtration of any silica, graphite, etc., which might be present. For samples containing less than 50 mg of selenium, make the concentration of hydrochloric acid 6 to 9N, add 8 grams of hydroxylamine hydrochloride, and boil the solution for 5 min to precipitate elemental selenium (5). Separate the selenium along with any other precipitate on an 8-micron micropore filter. Repeat this step if the selenium level is greater than 50 mg. Rinse the filtrate back into the original beaker and dilute to 40 ml. If the volume is too large at this point, transfer the sample from the filtration flask to a 150-ml beaker and reduce the volume by evaporation. Adjust the hydrochloric acid concentration to at least 3N for the tellurium precipitation. TELLURIUM PRECIPITATION. Add a minimum of 10 ml of the tin(1I) chloride solution to the sample. Add 10 ml more of reductant for a 2-gram sample, or a total of 30 ml for a 3-gram sample. Stir the solution and allow it to stand for 5 to 10 min before filtration onto a 2.5-cm, 8-micron micropore filter. Wash the precipitate with 50 ml of 3N hydrochloric acid followed by 50 ml of water to remove any trace of the reductant, any other metal ion, or acid. Increase the air flow through the filter for about a half minute. place the filter on a spot plate and dry for 10 min in an oven set at 90” C. Mount the dry filters with doubleback adhesive tape on squares or circles of inch thick phenol fabric. Any material may be used which does not have detectable x-ray lines in the spectral region of interest. Place the dry filter containing the preMEASUREMENT. cipitate of tellurium in the vacuum x-ray spectrometer quipped with a chromium target, lithium fluoride crystal, 0.01- x 4-inch collimator, gas flow proportional counter, and (5) I. M. Kolthoff, P. J. Elving, and E. B. Sandell, “Treatise on Analytical Chemistry,” Vol. 7, Part 11, p. 166, Interscience, New
York, 1961. (6) W. R. Schoeller, Analysr, 64, 318 (1939).
1
1200-
c I
900 8-
I
Sn
-
0
z 0
B cn ln
6000 V
300
-
28, DEGREES
Figure 1. Comparison of spectrum tracings from 2.85 to 3.48 A 0
Micropore filter;
-
50 pg of
Te on micropore filter;
phenol fabric masks W Ith openings 1 inch in diameter. Measure the tellurium Lal radiation for 40 seconds at a 28 of 109.54', at 50 ma and 65 kv. Focus the crystal with pure tellurium. Determine the background by reference to a blank. Calculate the concentration of tellurium by comparison with standards after subtracting the background correction. Count a few calibration points along with the samples. CALIBRATION CURVI:. Place aliquots containing 0, 5, 10, 20, 50, and 100 pg of tellurium in 100-ml beakers with 8.5 ml of concentrated hydrochloric acid, dilute to 40 ml, and add 10 ml of tin(I1) solution. Filter the precipitates on 8-micron filters. Wash, dry, and mount the precipitates. Measure the tellurium Lal radiation, subtract the background, and draw a calibration curve. The standards may be re-used for subsequent analyses. DISCUSSION
Of the several separation schemes considered as potential methods for the concentration of tellurium prior to its determination by x-ray spectrometry, the most selective separation procedure was based on the reduction of tellurium(1V) to tellurium(0). In aqueisus media, only a few elements are reduced to their elemental states by tin(I1) : tellurium, selenium, mercury, silver, bismuth, arsenic, and gold. The rapid reduction of tellurium(1V) by tin(I1) chloride makes it an ideal choice as a precipitant (5). During the developnient of the concentration technique for the x-ray determination of tellurium several conditions were evaluated. The tellurium level was kept at 100 pg and its Loll radiation measured for 40 sec. Tin(I1) was used as the reductant. Initially 5000 pg of selenium were used as a carrier for tellurium and were: collected on a glass fiber filter. Tellurium appeared to be evenly distributed on the surface be-
- - - - potential interferences
cause the Loll radiation was essentially the same when the counting was done on stationary or rotating sample holders. Under these conditions the ratio of the peak to background was 1.5 and the limit of detection about 3 pg of tellurium. When the selenium concentration was increased by 500 p g , the ratio of peak to background remained at 1.5; however, the background increased with the increasing selenium concentration. A IO-fold decrease in the selenium concentration produced a decrease in the background, while the ratio of peak to background increased 3.5 times. The ratio of peak to background increased to about 8 when the selenium concentration was decreased to 100 pg. The selenium concentration could not be decreased to a lower level because the colloided precipitate was not retained by the glass fiber filter. The filter also contributed to the background. Glass fiber filters are made of borosilicate glass and contain calcium among other elements. Calcium radiation occurring at 28 values of 100" and 110' causes intensification of the background. On the other hand, micropore filters contain only traces of calcium and no appreciable radiation was observed at the Loll wavelength of tellurium. Selenium was removed prior to the addition of tin(I1) to avoid interference (see Figure 1). Selenium was separated by boiling a solution of the sample, which is at least 6N in hydrochloric acid, with hydroxylamine (@. By this method 50 pg of tellurium were quantitatively separated from 0.02, 0.1, 0.5, 1.0, 5.0, and 10 mg of selenium by a single selenium precipitation. At the 50- and 100-mg levels, the precipitation was incomplete and a second selenium precipitation was required to prevent low tellurium results. There was no evidence of any tellurium coprecipitation ; low results were apparently due to dilution of the tellurium precipitate with selenium. VOL. 39, NO. 1, JANUARY 1967
15
TabIe I.
Determination of Tellurium with Diethyldithiocarbamate at 340 mp Tellurium, pg A340 Added Found
0.038
5
5
0.082 0.198 0,402 0.790
10 25
10 25
50
50
100
99
Figure 1 shows a tracing of the x-ray spectrum for the micropore filter (unbroken dotted line), for 50 pg of tellurium on a filter (unbroken line), and for several potential interferences such as selenium, tin, or nickel (broken lines). One hundred micrograms of selenium on a micropore filter produce Kal (n = 3) radiation which occurs at 110.78” and contributes to the background for tellurium Kal radiation. No interference occurs from the micropore filter; however, care must be taken to wash the filter and precipitate free of contamination, especially nickel. No radiation due to iron occurs in this area. When colloidal tellurium agglomerates, it is known to occlude tin and copper (7). Appreciable amounts of these or other elements were not retained by the colloidal tellurium, possibly because the digestion step was eliminated. Although trace amounts of tin were certainly present with the tellurium precipitate, Figure 1 shows tin Lpl and L/%radiation, which indicates the presence of tin with the tellurium precipitate. No interference was observed when tellurium was precipitated in the presence of nickel. The initial procedure used only 1.5 grams of stannous chloride dihydrate for the precipitation; however, if the iron concentration in the original sample was greater than 0.3 gram, it was necessary to increase stannous chloride concentrations from 1.5 to 10 grams to avoid low results. An excess of tin(I1) is essential to ensure reduction of both tellurium(1V) and iron(III), as well as any other reducible ion. There is some disagreement in the literature regarding the ability of tin(I1) to precipitate small amounts of tellurium. Goto and Kakita (8) report that 5 grams of tin(I1) chloride will separate tellurium from iron at the 0.01 level. They found complete precipitation when the concentration of hydrochloric acid was greater than 3N and no interference from perchloric acid up to lN(9). Our data show the precipitation to be complete when the hydrochloric acid concentration is between 1.6 and 8N. Subsequently it was reported (IO)that tin(I1) cannot be used to separate microgram amounts of tellurium. At the 66-pg level they reported that only 3 9 z recovery could be obtained when 10 grams of tin(I1) chloride were used to reduce tellurium(1V) in 3N hydrochloric acid. In order to establish the degree of tellurium precipitation with tin(II), the reduction was evaluated with three of the samples described in Table I11 at the 5-, lo-, and 50-pg levels. After filtration with micropore filters, the samples were dis-
x
(7) R. Stiennon-Bovy and G. Haegeman-Geladi, Centre Etude Energia Nuclear (Brussels), BLG-97 (1962); C. A., 63, 10655b (1965). (8) H. Goto and Y. Kakito, Sei. Repts. Res. Inst., Tohaku Unic. Ser. A, 7, 365 (1955); C. A . , SO, 6252a (1956). (9) H. Goto and Y. Kakito, Japan Analysr, 3, 299 (1954); Anal. Abstr., 3, 394 (1956). (IO) Y.Uzumasa, K. Hayashi, and S. Ito, Bull. Chem. SOC.Japan, 36, 301 (1963); Anal. Abstr., 11, 1742(1964). 16
ANALYTICAL CHEMISTRY
i
1
I
I
I
1
I
I
I
I
I
5
IO
20
50
100
zoo
I
.
eooa
,
I
i
‘i IO
I
~ELLW],
Figure 2. X-Ray calibration curve for tellurium Vacuum x-ray spectrometer LiF crystal, Cr tube 65 kv, 50 ma Soller slit 0.01 X 4 inches
Micropore filter matrix solved in 2 ml of 5 0 x sulfuric acid; the filter was destroyed by the addition of 30 hydrogen peroxide to the hot solution and then fumed. Tellurium was then determined spectrophotometrically using diethyldithiocarbamate. The recoveries were 4.8, 10.3, and 50 pg or very near the theoretical values. Perhaps Uzumasa and coworkers (IO) found incomplete precipitation of tellurium because they used an incorrect separation procedure or failed to add a sufficient quantity of tin(I1). Certainly they could not have lost tellurium in the dissolution step because volatilization does not occur during dissolution in a nonoxidizing acid (IO). Tellurium is volatilized only to from hydrochloric-perchloric acids at the extent of 0.5 200” c (11). The spectrophotometric determination of tellurium with diethyldithiocarbamate has been described (9). Tellurium is precipitated with diethyldithiocarbamate from a 5 % sulfuric acid solution and extracted with chloroform, and the absorbance is measured at 428 mp. The sensitivity of this method at 428 mp was not satisfactory; however, the spectra of the telluriumdiethyldithiocarbamate complex (12) show a region of greater sensitivity at wavelengths of less than 400 mp. Accordingly, we verified Beer’s law at 340, 345, and 350 m p for 5 to 100 pg of tellurium in 10 ml of chloroform with 1-cm cells, e349 = 10,200 liters per mole-cm. The blank at 340 mp was less than 0.05 absorbance unit. Table I shows the recovery obtained at this wavelength. Diethyldithiocarbamate is reported unstable in acid solutions (12-13); however, no difficulties were encountered in this work, per-
z
z
( 1 1 ) G. F. Smith, “Mixed Perchloric, Sulfuric, and Phosphoric Acids and their Applications in Analysis,” p. 61, G. F. Smith
Chemical Co., Columbus, Ohio, 1942. (12) H. Bode, Z . Anal. Chem., 142, 414(1954); C.A., 48, 13522h (1954). 25, 1260(1953). (13) A. E. Martin, ANAL.CHEM.,
haps because the extraction was performed as soon as the reagent was added. In addition to the spectrophotometric verification, x-ray data proved that microgram quantities of colloidal tellurium were completely retained on micropore filters. When a micropore filter with iin 8-micron pore size was used in tandem with a filter having a pore size of 1.2 microns, all of the precipitate was found on the upper 8-micron filter while the 1.2-micron filtt; did not contain a countable amount of tellurium. Figure 2 also verifies the reliability of the separation scheme for tellurium. The points on the curve are the actual micropore filters. The gradation in gray to black makes it possible to estimate the telluriuni concentration visually. The numbers given on the right y-axis are the actual counts for each point. The background correction does not vary with tellurium concentration. Replacement of the glass fiber filter with micropore filters and the elimination of selenium as a carrier reduced the background (see Figure 1) and decreased the limit of detection to about 0.2 pg. The background was decreased to lower levels by narrowing the Soller slit collimator from 0.02 x 11iz to 0.01 X 4 inches. This did not change the limit of detection. The peak to background ratio was increased to about 50. Table I1 shows the x-ray results obtained in the determination of tellurium added to solutions of various alloys. There is no serious interference from the following elements at the concentration levels given on the standard certifications : aluminum, antimony, arsenic, boron, calcium, carbon, chromium, cobalt, copper, lead, magnesium, manganese, molybdenum, niobium, nitrogen, phosphorus, silicon, sulfur, tin, titanium, tungsten, uranium, zinc, and zirconium. Linear calibration curves are not obtained when more than 200 pg of tellurium is held on a filter 25 mm in diameter, When a filter 47 mm in diameter is used, the precipitate is spread over a greater area and calibration curves are linear up to 1000 pg. RESULTS
The accuracy of the x-ray method is given in Table 111, which points out that .the average error is 4z at the 5-pg level and 1 or less at the upper levels. A statistical evaluation of the precision of i.he method is also given. For the 50pg sample, the precision of an isolated result is 49.6 i 2 pg. The 50-pg samples also contained 500 pg of selenium which was separated prior to the tellurium precipitation. There are very few standards available for the determination
z
Table 11. X-Ray Determination of Tellurium Added to Various Alloys
Maraging steel Maraging steel Nickel-chromium alloy 600 Copper-nickel alloy Copper-nickel alloy 400 Cupro-nickel 4 Copper-aluminum alloy Nimonic 90 Permanent magnet alloy Bronze 6 x Zn-aluminum alloy Mild steel Basic open hearth steel Ingot iron Ingot iron Cr 18-Ni 9 steel Cr 18-Ni 10-Ti 0.4 steel
INCO 73 INCO 75 INCO NX 1476 INCO M 300 B Wiggin B 6998 BCS 180/1 BCS 181/1 BCS 310 BCS 233 BCS 183/1 BCS 300 BCS 322 NBS 15f NBS 55ea
50 50 50 50
53 47 53 47 47 53 47 48 47
50
50
x
NBS 55e5 NBS lOle NBS 121e a
Tellurium, pg Added Found
Material
Standard No.
50 50
50 50 50 50 50
55
51 50 52 26 10 48 49
25 10 50 50
Sample weight 1 gram, but 0.2 gram with 55e cases.
Table 111. Statistical Evaluation of the X-Ray Method
No. of detn 10 8 8 24 10
Tellurium, c(g Found Added (av)
Std dev
Re1 std dev
Mean error
5.2 9.9 19.6 49.6 100
0.7 0.8 1.7 2.0 3.3
12.5 8 9 4 3.3
0.2 0.3 0.6 0.4 1.o
5 10 20
50 100
Table IV. X-Ray Determination of Tellurium in Standards Described by Morris and Hill
Composition range. C 0.1-0.2, Cr 10-14, Co 0-15, Mo 3-4.5, V 0.8-2, Ti 1-5, A1 5.5-6, Zr 0.07-0.1, B 0.01, Ni balance Standard No.
Tellurium, ppm Reptd (12) Found
N432 N433 N690/4 N434 N435 N689/4
0.52 1.2 1.1 2.0 4.8 4.9
0.2 0.6 0.5 1.6 5.1 4.0
Tabk V. Determination of Tellurium N B S White Cast Irons and Copper-Base Standards
NBS No.
NBS 730 90 (140) (240) (90) 35 15 3
X-Ray 1174 620 1175 80 1177 110 1180 180 1182 70 c 1100 27 c1101 12 c1102 1 1118 ... 2 .*. 0.5 1119 1120 *.. 1 Values in parenthesis not certified, given for information.
Tellurium, ppm Polarographic ( 3 ) 710 87
...
... ...
30,40 18, 19
... ...
...
...
Activation ( 3 ) 750 94
...
... ... 32, 32 11
... ... ... ...
Extraction (3) 720 f 110 80 20
*
... .,. ... ...
...
... ... ... ...
VOL. 39, NO. 1, JANUARY 1967
17
of tellurium in the range of 5 to 100 ppm. A series of nickel-base alloys has been analyzed by activation analysis in the range of 0.5 to 5 ppm (4). Table IV presents a comparison of the values. Five milliliters of sulfuric acid were added in the dissolution process to prevent the separation of molybdenum ; however, some precipitation occurred. The prefiltration for the removal of molybdenum may also have removed some tellurium and caused the tendency toward low results. Table V shows the values obtained in the analysis of two certified NBS white cast irons, and certified copper-base alloys. The values for the copper-base standards agree very well with the NBS values as well as with the polarographic and activation methods (3). The x-ray values for the white cast irons (NBS 1174 and 1175) fall within the limits measured by the butyl acetate-hexone extraction method (3); however, they are low compared with the polarographic and activation methods (3). Any loss of tellurium in the preflltration of silica and graphite was small; thus the lack of agreement is probably due to the nature of these spectrographic standardsLe., only the top 6 / ~ einch is certified. The x-ray sample was not taken from this portion of the standard.
The method has been applied to the analysis of highpurity material such as 99.8x nickel and 99.999xselenium. This type of nickel had been analyzed by the solid mass spectrograph and found to contain 0.3 ppm of tellurium. Using the x-ray method 0.6 ppm was found. A semiquantitative determination of tellurium in selenium showed the presence of between 1 and 3 ppm; the x-ray method found 1 ppm. Most of the selenium was volatilized before the method was applied (14). ACKNOWLEDGMENT
The authors are indebted to Ute Moehring and Thomas Ruppert for obtaining many of the analytical data. RECEIVED for review May 26, 1966. Accepted October 17, 1966. Seventeenth Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, February 1966.
(14) N. Etten and J . Muschaweck, 2.Anal. Chem., 206, 17 (1964); Anal. Abstr., 13, 37 (1966).
Arsenazo 111 as a Sensitive and Selective Reagent for the Spectrophotometric Determination of Palladium in Iron and Stony Meteorites J. G.Sen Gupta Geological Survey of Canada, Ottawa, Ontario, Canada Arsenazo Ill is a very sensitive and selective reagent for the spectrophotometric determination of palladium in mixtures of palladium, platinum, rhodium, and iridium isolated from iron and stony meteorites and copper-nickel matte by perchloric acid decomposition and ion exchange separation. Palladium forms a 1:l complex with Arsenazo 111, with a dissociation constant of 5.04 X 10-6 at 24O C. The optimum concentration range for the determination of palladium is from 1.16 to 3.00 ppm. Also, a combination of ion-exchange separation and the sensitive thorium-arsenazo 111 reaction has been used in the determination of microgram amounts of thorium in some stony and iron meteorites. Palladium concentrations were found to be in the range of 2.4-13.4 ppm in three iron meteorites, 0.4-0.8 ppm in five stony meteorites; the thorium concentrations were in the range of 0.10-0.15 ppm in the iron meteorites and 0.05-0.09 ppm in the stony meteorites.
THEREAGENT ARSENAZO I11 (1,8-dihydroxynaphthalene-3,6disulfonic acid-2,7-bis[(azo-2)-phenylarsonic acid)], as represented in Figure 1, has been used for the colorimetric determination of a number of elements as summarized in recent reviews by Savvin (1-3). A search of the literature indicated that this reagent had never been proposed nor used as a colorimet(1) S. B. Savvin, Tuluntu, 8,673 (1961). (2) Zbid., 11,l (1964). (3) Ibid., p. 7.
18
ANALYTICAL CHEMISTRY
ric reagent for the platinum metals, nor had it been used for the determination of traces of thorium in meteorites, The complex formed between Arsenazo I11 and palladium(I1) in a buffered aqueous medium of pH 3.4to 5.9 has an intense purple color and is suitable for the spectrophotometric determination of palladium. The reaction is highly selective because no other noble metals react with the reagent under similar conditions. Osmium and ruthenium are usually separated from a mineral, ore, or meteorite containing the platinum metals by distillation with perchloric acid; the base metals are then separated from palladium, platinum, rhodium, and iridium by first adjusting the pH of the solution to 1.5 and then passing it through a Dowex 50W-X8cation exchange resin which retains the base metals, including the thorium, but not the platinum metals (4, 5). It is then possible to determine the palladium directly with Arsenazo I11 in the effluentin the presence of the other platinum metals. The reagent Arsenazo I11 has, therefore, an advantage over other reagents which can be used only after partial or complete separation of palladium from the associated noble metals (6, 7). (4) J. G. Sen Gupta and F. E. Beamish, Am. Mineralogist, 48, 379 (1963). (5) J. G. Sen Gupta and F. E. Beamish, ANAL.CHEM.,34, 1761 (1962). (6) J. G. Sen Gupta, Tuluntu, 8, 729 (1961). (7) J. H. Yoe and J. J. Kirkland, ANAL.CHEM., 26,1335 (1954).
