It is clear, therefor€, t h a t the oxidation product of hydroxyproline intermediate in the present method is neither pyrrole nor pyrrole-o -carboxylic acid. The dyestuff formed by condensation with p-dimethylamir o benzaldehyde does, however, appear to be the same for these three compounds. It is likely, though by no means certain, that the variations in A,, and perhaps some of the variations in absoabance are due to a n association of the dyestuff which is dependent on the dielxtric constant of the medium and whici responds slowly to changes in the environment. ~i*-Pyrroline-4-hydri,sy-2-carbo x y l i c acid has also been suggested by Radhakrishnan and Jleister as a product of the oxidation of hydroxyproline ( 8 ) , but has not been directly tested in the present work. Thiay found that this compound, when trezted with dilute mineral acid, was jlot%Tly transformed into pyrrole-a-carbor ylic acid. This fact was used in the present work to inTestigate the nature of the hydrosyproline oxidation product. Four aliquots of hydroxyproline, each Kith its corresponding blank, were taken through the method r-ith the following variations. (The blanks all gave equally
low absorbances.) Aliquot KO.1 was heated for 25 minutes (the optimum) in stage 2 of the method, and aliquot No. 2 for 35 minutes. The 10-minute overheating resulted in a 3% decrease in color yield. Aliquots No. 3 and 4 were taken through the oxidation stage, then the perchloric acid and propanol were added and, after an hour at room temperature, the solid aldehyde. Aliquot KO.3 was then heated at 60" C. for 25 minutes, and S o . 4 for 35 minutes. The 10-minute overheating again produced a 3% decrease in absorbance, though the color yields of aliquots No. 3 and 4 were 10% lower than those of KO. 1 and 2, respectively. Most compounds related to pyrrole are somewhat unstable in acid solution, and therefore, i t is not surprihing that the color yield was decreased by alloning the intermediate oxidation product to stand in acid for a n hour. If, however, this intermediate mere hydrosypyrroline carboxylic acid, then the treatment with acid should have transformed it, at least in part, into pyrrole-a-carboxylic acid. This latter has an optimum heating time in the present method of about 80 minutes at 60' C., and the color yield would, therefore, have increased rather than
decreased, for an increase in heating time from 25 to 35 minutes. The inference is that the oxidation product of hydroxyproline in the present work is not AI-pyrroline-4-hydroxy-2-carboxylic acid. LITERATURE CITED
( 1 ) Bergman, I., Tuck, G. C., Safety in
Mines Research Establishment, Sheffield, England, unpublished work. (2) Bowes, J. H., J . SOC.Leather Trades' Chemists 4 3 , 2 0 3 (1959). (3) Grunbaum, B. W., Glick, D., Arch. Biochem. Biophys. 6 5 , 2 6 0 (1956). (4) Hutterer, F., Singer, E. J., ANAL. CHEM.3 2 , 5 5 6 (1960). ( 5 ) Lollar, J., J . Am. Leather Chemists' Assoc. 53, 2 (1958). (6) Neuman, R. E., Logan, bl. A4., J . Biol. C h e m 184, 299 (1950). ( 7 ) Prockop, D. J., Udenfriend, P., Anal. Biochem. 1 , 228 (1960). (8) Radhakrishnan, A. N., Illeister, A., J . Biol. Cheni. 226, 559 (1957). (9) Stegemann, H., Z. Physiol. Cheni. 3 1 1 , 4 1 (1958). (10) Woessner, J. F., Jr., Arch. Biochena. Biophys. 93,440 (1961).
RECEIVED for review January 31, 1963. Accepted July 15, 1963. The illustrations to this paper are British Crown Copyright and are reproduced by permission of the Controller of Her Britannic Majesty's Stationery Office.
Analysis of Mixtures of Antimony and Bismuth Tellurides Containing Selenium and Iodine K. L. CHENG and B. 1. GOYDISH RCA laboratories, Princeton, N . J .
b There i s a need fot, determining the composition of the tellurides o f antimony and bismuth. Simpler and more r a p i d methods for defermining bismuth, antimony, tellurium, selenium, and iodine in such alloys have been developed. Tellurium i s determined gravimetrically as tellurium dioxide. Bismuth and antimony are determined volumetrically b y step (ethylenedinitri1o)tetraacetic acid titrations in the presence a n d in the absence of tarirate, which can mask antimony. Selenium i s determined photometrically with 3,3'-diaminobenzidine, and iodine i s determined photometrically by measuring the turbidity o f the silver iodide precipitate in an ammoniacal medium after a simple distillation. If desired, each individual element may b e determined independently. All five elements may b e determined sequentially in a single sample when the sample supply i s limited.
M
such as those of bismuth and antimony, possess interesting thermoelectric properties. One of the important thermoANY TELLUHIUES,
electric materials is a bismuth telluride compound in which bismuth is partly replaced by antimony, tellurium is partly replaced by selenium, and the material is doped with small amounts of iodine which can affect the conductivity type ( n or p ) of the thermoelectric material. It is necessary to analyze various materials to ascertain their compositions and to obtain desired properties. Reed ( 7 ) reported methods of analyzing this type of material in which selenium and tellurium are determined gravimetrically by precipitation with sulfurous acid at different acidities. The antimony is titrated potentiometrically after its separation as antimony sulfide from the filtratp, from which selenium and tellurium have been previously removed, and bismuth is titrated with (ethylenedinitri1o)tetraacetic acid (EDTA) in the filtrate after selenium and tellurium have been removed. H e did not report the method of analyzing for iodine. Cluley and Proffitt (6) reported essentially similar methods of analyzing this type of material, and also reported a method of
iodometric determination of iodine after separation by steam distillation. This paper reports rapid methods of analyzing the material. Tellurium is determined gravimetrically as TeOz while bismuth and antimony are masked by E D T A in the presence of acetone. Selenium is directly determined photometrically with 3,3'-diaminobenzidine in the presence of tartrate (3). Both bismuth and antimony are titrated Kith E D T A in the presence of acetone. By masking antimony with tartrate, only bismuth is titrated by EDTA. Then the values for antimony and bismuth can be obtained by stepwise E D T A titrations with and without addition of tartrate ( 4 ) . The proposed methods are more rapid, simpler, and at least as accurate as those previously reported. Another significant feature of the proposed methods is that when the determination of only one element is desired, it can be done rapidly and independently without separation of other elements. If desired, iodine, tellurium, bismuth, antimony, and selenium may be determined in a single sample. VOL. 35, NO. 12, NOVEMBER 1963
1965
EXPERIMENTAL
Table 1. Reagents and Apparatus. E D T A solution, 0.025M, standardized against a standard lead solution. Standard lead solution, 0.025M, prepared from pure lead and nitric acid. Standard cobalt solution, O.O25M, prepared from cobalt nitrate. Potassium thiocyanate solution, 67% aqueous solution (w./v.). Dissolve 67 grams of KSCN in water and dilute to 100 ml. with water. Xylenol Orange indicator solution, 0.001M. 3,3'-Diaminobenzidine solution, freshly prepared 0.5% in water stored in a refrigerator. Formic acid , 2.5N. Acetate buffer solution, pH 3.5, PrePared bv misine 100 nil. of acetate solition cohaining-50 grams of sodium acetate trihydrate a i t h 350 ml. of acetic acid and diluting to 500 nil. a i t h water. .idju.t to p H 3.5 with either sodium hydrouide or acetic acid, if neceqaary. Other reagent. mere of analytical reagent grade. A Beckman spectrophotometer Nodel D U and Beckman p H meter Model 76 were used. The Ace Glass Co. (Vineland, S.J.) ar-pnic distillation apparatus No. 5319 was used for iodine distillation (8). Determination of Tellurium. Dissolve a suitable amount ( 0 2 to 0.5 gram) of sample in approuiniatelv 5 ml. of nitric acid on a hot plate and evaporate off most of the acid. Cool add 7 5 nil. of acetone and a known, slightly-eucess amount of EDTA%solution. (The slight excess is uqed to complev both Bi and Sb, but a large excess of E D T A has no effect on the tellurium determination.) Carefully adjust the pH to 4.0 t o 5.0 with ammoniiini hydroxide or hydrochloric acid. Cool to 20" C. or below, and filter through a medium-porosity sinteredgla- crucible K a s h 2 to 3 times n i t h a fen milliliters of 70% acetone solution, and dry the precipitate at 120' to 140' C , until a constant neight is obtained. Keicrh as TeO,. the eravimetric factor beiiG 0.7995. -' Titration of Bismuth and Antimony. Transfer the filtrate from the tellurium determination into a 400-ml. beaker and rime with acetone or water to adjust the acetone concentration to approuimately 607,. Add 3 nil. of p H 3.5 acetate buffer solution and adjust to pH 3.0 to 3.5, if necessary, with acetic acid or sodium acetate. After adding 4 nil. of 677, potasqium thiocyanate solution, back-titrate the esceqs EDTA%with a standard cobalt solution, the color change a t the end point being from pink to violet. A magnetic stirrer may be used during the titration. The result gives the total amount of biqmuth and antimony. Adjust the same solution to approximately p H 9 with ammonium hydroside or sodium hydroside to precipitate antimony, add a sufficient amount of tartaric acid to diqsolve the antimony oxide, and adjust to p H 5.5 to 6.0. Back-titrate the free E D T A releaqed from antimony-EDTA% complex with a standard lead solution, using 6 drops of Xylenol Orange solu-
-
u
1966
ANALYTICAL CHEMISTRY
Analysis of Synthetic Mixtures
Taken,
Found, mg.
Element mg. Bismuth 20 90 20 80 Antimony 34 20 34 10 Tellurium 51 04 51 14 Selenium 1 01 0 96 Iodine 0 050 0 048" Bismuth 26 12 26 22 Antimony 57 00 57 20 Tellurium 38 28 38 30 Selenium 2 02 2 00 Iodine 0 025 0 023" Bismuth 26 12 26 25 Antimony 45 60 45 60 Tellurium 63 80 63 80 Selenium 3 06 3 06 Iodine 0 500 0 495 Bismuth 15 67 15 70 Antimony 34 20 34 40 Tellurium 51 04 51 10 Selenium 0 50 0 a n Iodine 0 750 0 '744 a Made up t o 25 ml. instead of
Difference, nig. -0 -0 +0 -0 -0 +0 +0 +0 -0 -0 +0 -0 0 0 -0
+o
$0 +0 -0 -0
10 10 10 05 002 10 20 02 02 002 13 10
005 03 20 06 01 on6
50 ml.
tion as indicator. The end point change is from yellow to red. This second titration gives the result for antimony alone and the result for bismuth niay be obtained by difference. Determination of Selenium. Dissolve a suitable amount (approximately 0.1 gram) of a separate portion of the sample in nitric acid and evaporate off most of the acid. Add a sufficient amount of tartaric acid to keep antimony and bismuth in solution, 2 nil. of formic acid, and dilute to approsimately 50 ml. with water. Adjust the pH to 2 to 3. .idd 2 ml. of 0.57, diaminobenzidine solution and let stand for 30 to 50 minutes. (EDTA% is not necessary in such mixtures, but its presence is not harmful, either.) Adjust the pH to 6 to 7 with ammonium hydroxide. Transfer to a 125-mI. Feparatory funnel, add elactly 10 ml. of toluene, and shake vigorously for 30 seconds. Discard the aqueous layer and filter the estract through a glass aool placed a t the tip of the funnel. Measure the absorbance a t 420 mp using a reagent blank. Prepare a calibration curve with known amounts of selenium in a similar manner. Determination of Iodine. Place a suitable amount (approximately 1 gram) of t h e sample in a No. 5319 Ace Glass Co. arsenic distillation apparatus, and add 10 ml. of concentrated nitric acid from the modified separatory funnel. Nitric acid dissolves the sample, in t h e form of powder or chips, rather rapidly upon heating. Gently bubble nitrogen gas through the solution. Heat the distillation apparatus at 100' C. on a heating mantle. Distill for 10 minutes and collect the iodine in 10 ml. of 1Asodium hydroxide solution. To the sodium hydroside solution, add 5 ml. of concentrated ammonium hydroxide and 2 ml. of 0 . l N silver nitrate solution. Make up to a 50-ml. volume with water. Measure the turbidity at 400 mpagainat
a reagent blank. Prersare a calibration curvcwith 0.05 to 0.5'mg. of iodine in a similar manner. RESULTS AND DISCUSSION
Synthetic mixtures were analyzed and the results are shown in Table I. Tellurium. Tellurium has been commonly determined gravimetrically as its elemental form. I n Reed's method for tellurium, the filtered precipitate is dried in vacuo a t room temperature to avoid oxidation of tellurium, which is stated by Duval (6) to occur a t temperature above 40" C. Later, Cluley and Proffitt found that tellurium precipitated in a flocculent form was less susceptible to oxidation; even so, a small positive error of 0.6% n-as obtained on drying the precipitate a t 105" C. for 1 hour. Precipitation of tellurium as its elemental form requires a long time of standing before filtration t o settle the fine particles. This occasionally causes a negative error, especially if the precipitation conditions are not carefully controlled and the precipitate is filtered too soon after precipitation. Probably good results niay be obtained by coincidental compensation of the tlvo errors. Cluley and Proffitt used the dichromate method for the L oluinctric determination of tellurium after it is precipitated as elemental tellurium and redissolved with nitric acid. Cheng ( 1 ) reported a gravimetric method for determining tellurium as tellurous acid in the presence of EDTX with advantages of high selectivity, rapidity, and favorable gravimetric factor. Furthermore, T e 0 2 is very stable and can be dried at 120" to 140" C. n ithout danger of oxidation. Therefore we adopted the TeOz method for determining tellurium in bismuth telluride and the related materials. Bismuth is strongly complexed by EDTA% in a nide p H range but antimony is not. Howeyer, antimony can be kept in a 60% acetone solution by EDTA a t pH 4 to 6, which is also the optimum pH for precipitation of tellurous acid. I n the presence of acetone, the solubility of tellurous acid is further decreased. Tellurous acid, once separated, may, if desired, be determined volumetrically with permanganate or other osidizing agent3. Tellurous acid is easily dissolved with an acid or a base. Bismuth and Antimony. Reed determined bismuth b y E D T A titration when antimony was coniplexed by tartrate, and Cluley and Proffitt titrated bismuth in the presence of antimony hydroxide precipitate, which often occludes or adsorbs small amounts of bismuth, causing slight error or delayed end point change. I n the present work, a back-titration technique was used and bismuth and antimony were
stepwise titrated with E D T A for simplicity and accuracy. Takamoto (9) first reported the EDTA titration of antimony in the pre:ence of acetone. The antimony-EDTA complex is so weak that tartrate, citrate, or fluoride replaces E D T A to i'orm a tartrate, citrate, or fluoride complex of antimony, Thus, the amount of E D T A originally combined with antimony is released quantitatively. Bismuth can be titrated with E D T A in the presence of tartrate, which masks antimony, using lead as a back titrrtnt and Xylenol Orange or hIethylthyino1 Blue as indicator. I n the absmce of tartrate, citrate, or fluoride, antimony and other metal ions may be totally titrated with E D T A in the presence of acetone using cobalt as a back titrant and thiocyanate as indicator. The IZDTA titrations either in the same solution or in the separate aliquots give a highly-selective volumetric method for antimony. The detail? of the step$-ise E D T A titrations of antimony will be dxcribed in a separate report. When only the determination of bismuth is desired, this may be done by adding escess E D T A and enough tartrate or fluoride (masking antimony and tellurium) ; the excess of E D T A is then back-titrated with lead at p H 5 to 6. When only the determination of antimony is desired, the stepmise E D T A titrations of bismuth and antimony may be done in the presen1:e of acetone and hydrolyzed tellurous avid. I n the back-
titration with cobalt, take the reading for the first end point change and continue the titration after addition of tartrate or fluoride The titer from the first end point t o the second end point is for antimony. Selenium. Bismuth telluride as a thermoelectric material usually contains small amounts of selenium. T h e selenium is better determined photometrically n i t h 3,3'-diaminobenzidine which has been proved t o be a highly selective and sensitive reagent for selenium when the masking agents a n d solvent extraction technique are employed. Iodine. Bismuth telluride thermoelectric material often contains small amounts of iodine which can be conveniently determined with t h e silver iodide turbidity method. Silver iodide is one of the most stable silver precipitates known; its turbidity may be measured in a n ammoniacal medium (a). .4 simple apparatus can be used for distilling iodine; for example, an 125-m1. Erlenmeyer flask with ground glass neck was connected to a 25-ml. volumetric pipet whose tip was dipped into 10 ml. of 1A' sodium hydroxide solution. The use of the pipet was to prevent suction of the sodium hydroxide solution into the Erlenmeyer flask when the pressure was decreased after the sample was dissolved with heating. Since the iodine is easily carried over by the nitrogen oxides generated from the dissolution of the
sample with nitric acid, the sweeping with an inert gas may be omitted. When the sample supply is limited, the determinations of iodine, tellurium, bismuth, antimony, and selenium may be carried out in a single sample. The iodine is first distilled and determined as AgI; the tellurium is precipitated as tellurous acid in the remaining solution (or an aliquot) after addition of E D T A and acetone; after separation of the tellurous acid precipitate, the filtrate is back titrated with a cobalt solution a t p H 3.5 for both bismuth and antimony and a t p H 6 for antimony alone using lead as a back titrant in the presence of tartrate or Auoroboric acid; a few drops of EDTA4are added, the selenium is then determined b y diaminobenzidine after removing acetone by heating. LITERATURE CITED
(1) Cheng, K. L., A \ ~ L CHEM. . 28, 1738
(1936).
( 2 ) Ibzd., 33, 761 (1961). (3) Zbid., p. 783. (4)Cheng, I