716
Vol. 14, No. 9
INDUSTRIAL AND ENGINEERING CHEMISTRY
is shown, distillagain a t 140" m already described until a negative result is obtained. DISTILLATION OF GERMANIUM TETRACHLORIDE. Add 2 ml. of 1 to 1 hydrochloric acid and 2 ml. of water to the cool contents of the distilling flask, and distill while bubiling air through the solution until the temperature reaches 120 . Collect the distillate in the drawn-out test tube containing 2.00 ml. of 25 per cent sodium hydroxide solution. At the end of the distillation, neutralize the solution with l to l hydrochloric acid, transfer t o a 25-ml. volumetric flask, and make up to volume with water. Allow the distilling flask to cool, add 2 ml. of 1 to 1 hydrochloric acid and 2 ml. of water, and distill as before. Xeutralize t,he solution, transfer to a 25-ml. volumetric flask, and make up to the mark with water. With small amounts of germanium practically all is found in the first distillate, but with quantities in the neighborhood of 25 micrograms a small amount is found in.the sacond distillate. COLORIMETRIC DETERMINATIOK OF GERMANIUM. Prepare two standard germanium solutions as follows. Measure out two portions of germanium oxide solution containing, for example, 5 and 10 micrograms of germanium, and add 10 ml. of water, 1.5 ml. of 1 to 1 hydrochloric acid, and 2.00 ml. of 25 per cent sodium hydroxide solution. A-eutralize the solutions carefully with 1 to 1 hydrochloric acid (phenolphthalein as indicator) and dilute to 25 ml. In a similar manner prepare a blank solution for the color comparison by neutralizing 2.00 ml. of sodium hydroxide and diluting to 25 ml. Take a 10-ml. aliquot of each standard, of the two germanium distillates, and of the blank. Make each solution basic with a drop of 25 per cent sodium hydroxide and then acidify with 0.10 ml. of 1 to 1 acetic acid. Add 10.0 ml. of ammonium molybdateferrous ammonium sulfate reagent, mix, and dilute to 25 ml. with water. After 15 minutes compare the germanium distillates and the standards against the blank solution in a filt'er photometer, using a red filter. The standard series method of comparison may also be used. The color intensity of the solutions increases slowly on standing and the readings should, therefore, be made after the period of standing specified. Although germanium in appreciable amounts is unlikely to occur in the reagents, it is nevertheless advisable to run a blank through the procedure, especially since any contamination by silica will then be revealed.
TABLEI. DETERMINATION O F GERMANIUM I N SILICATE ROCKS Sample
Germanium Present0
Germanium Found
Error
%
'%
%
1 Graniteb 0.0005 0.0005 0.0000 2 Granite 0.0010 0.0009 -0.0001 -0.ooo1 3 Granite 0.0030 0.0029 4 Granite 0.0058 0.0050 -0.0008 5 Granitec 0.00026 0.00028 +o, 00002 6 Granited 0.00026 0.00030 - 0.00004 0.0001 7 Diabase' 0.0008 0.0007 8 Diabase 0.0014 0.0013 -0.0001 9 Diabase 0.0056 0.0052 - 0.0004 Sum of germanium originally present in samples as determined by method described (0.00026% in, granite. 0.000147, in diabase) and t h a t added. b Percentage composition: SiOz, 76.8; AlaOa! 13.2; FezOa, 0.3; FeO, 0.4; Mg-0, 0.2: CaO, 0.7; NalO, 3.8; Kz0, 4.5; TiOz. 0.08; PzOs,0.02; MnO,
+
Acknowledgments The authors are indebted t o Theodore C. Blegen, Dean of the Graduate School, and S. C. Lind, Dean of the Institute of Technology, University of Minnesota, for making laboratory facilities available for this work.
Literature Cited (1) Aitkenhead, W. C., and Middleton, A. R., IXD. ENG.CHEM., ANAL.ED.,10, 633 (1938). (2) Goldschmidt, V. M., and Peters, C., Nachr. Ges. Wiss. Gottingen, Math.-physik. Klasse, Fachgruppen, 1933, 141; Goldschmidt, V. M., Hauptmann, H., and Peters, C., Naturuiasenschaften, 21, 363 (1933). (3) Poluektov, N. S., Mikrochemie, 18, 48 (1935). (4) Schemer, J. A , , J . Research Nail. Bur. Standards, 16, 255 (1936).
A Volumetric Method for Determining Tin Based on the Formation of a Dioxalatothiometastannate HOBART H. WILLARD AND T A F T Y. TORIBARA University of Michigan, A n n Arbor, Mich.
W
HEELER (1) recognized that stannic tin in oxalic acid solution forms a fairly stable compound with
sulfur when hydrogen sulfide is passed into such a solution. H e made use of this fact in devising a volumetric method of determining tin in bronze by titrating the sulfur with iodine. The tin was separated from other metals by addition of phosphoric acid, the precipitate was dissolved in concentrated sulfuric acid, and the solution was neutralized with ammonia, using methyl orange indicator. Six or 7 grams of oxalic acid were added per 0.2 gram of tin and the solution was boiled until clear. The hot solution, 100 ml. in volume, was saturated with hydrogen sulfide, diluted to 260 ml., and cooled to room temperature, and a stream of air was bubbled through it 15 t o 20 minutes to remove the excess of hydrogen sulfide. A slight excess of standard iodine was added and back-titrated with thiosulfate.
Difficulties i n the Use of the Existing Method I n attempting t o verify Wheeler's work, it was found that conditions had t o be adjusted so carefully in order t o obtain t h e theoretical relationship between tin and sulfur t h a t the method was not practical. It was necessary t o pass hydrogen sulfide into the solution for at least 20 minutes a t a rate of 1 liter per minute to ensure complete absorption. Using air,
nitrogen, and carbon dioxide as agents t o sweep the excess hydrogen sulfide out of the solution, the results obtained were dependent upon the gas used and the time of its passage. The passage of air at a rate of 1.5 to 2 liters per minute for more than 15 minutes always gave low results, whereas nitrogen and carbon dioxide at the same rate gave low results when passed for much over 20 minutes. At t h e end of the passage of the gas, there was still a n odor of hydrogen sulfide. Since it was difficult to regulate the passage of gas in the same manner for each determination, it was likewise difficult to compare t h e times of passage on a quantitative basis.
ON TABLEI. EFFECTOF PASSISGNITROGEN RATIO
Time Min.
0 10 30 50 80 110
THE
S/Sn Ratio
SULFUR-TIS
September 15, 1942
ANALYTICAL EDITION
For this reason, nitrogen gas at room temperature was passed continuously through a solution, and samples for analysis were removed at regular intervals, Table I shows the results obtained. If the nitrogen was replaced b y air, lower results were obtained, owing to partial oxidation of the sulfur. T h e data show t h a t the dioxalatothiometastannic acid formed according to Wheeler's directions is relatively unstable. Solutions of oxalatostannic acid which were allowed to stand for some time in the cold were found t o become somewhat opalescent because of hydrolysis of the tin. These solutions always gave low sulfur-tin ratios, indicating that t h e absorption of hydrogen sulfide might be a measure of the unhydrolyzed tin present. To verify this t h e following experiments were performed: Some of the oxalatostannic acid solution was refluxed a t boiling temperature for 3 hours, and its opalescent appearance indicated that rather extensive hydrolysis had taken place. The sulfurtin ratio after the hydrogen sulfide treatment was found to be 0.104. Some of this same hydrolyzed solution was evaporated to fumes of sulfuric acid, and then the procedure as outlined by Wheeler (1) was followed. The sulfur-tin ratio of this solution was 1.000, indicating that the hydrolyzed tin had been rendered soluble. A second sample was treated in the same manner up through the addition of the oxalic acid. Bfter the solution had been heated to dissolve the stannic hydroxide, it v a s alloTYed to cool down to room temperature and stand overnight before treatment with hydrogen sulfide. Although the solution appeared perfectly clear, the sulfur-tin ratio was 0.640, indicating that heating even for this short time caused appreciable hydrolys1s.
TABLE 11. DETERMIN.4TIOZr pH after Addition of KzCz04
Final pH after GO* Passage
2.2 2.5 2.5 2.6 3.3
3.0
...
3.0 3.3 3.4
OF
TINUSING
POTASSIUM
SULFIDE
Tin Found
Tin Taken Gram
Gram
0.1301 0.1416 0.1328 0.1372 0.1312
0.1301 0.1414 0.1330 0.1370 0.1315
71'1
Sn taken, 0.1630 gram
pH after addition of K2C20r, 3.37
Passage of HzS a t 2 liters per minute, 25 minutes Passage of COZa t 2 liters per minute, 1 hour Final pH, 3.23 Sn found, 0.1632 gram
The iodine was standardized against arsenious oxide. Several other improvements i n t h e process were made. An appreciable length of time was required for the passage of hydrogen sulfide because the manner of introducing the gas through a tube did not afford very good contact with the solution. Furthermore, the use of the gas caused high localized concentrations a t the gas-liquid interface, such that a very small amount of stannic sulfide was always deposited on the tube. In order t o eliminate these difficulties, hydrogen sulfide was liberated internally by the addition of a solution of potassium sulfide. Carbon dioxide was bubbled through the solution to remove the exceSs hydrogen sulfide until a piece of lead acetate paper held in the escaping gases showed no blackening. The time of 1 hour used above was much greater than necessary, but it proved that dipotassium dioxalatothiometastannate in a large excess of potassium oxalate is very stable. Since the potassium sulfide solution is strongly alkaline, the pH resulting from the addition of potassium oxalate was set a t a lower value, such that the final pH after the addition of the sulfide solution was close t o the desired value of 3.3. Addition of potassium sulfide in the cold caused too high local concentrations of sulfide ion, resulting in the precipitation of some stannic sulfide. This eventually went back into solution, but in some cases only after rather extensive boiling. The difficulty was eliminated by adding the potassium sulfide solution dropwise a t a temperature of 60" with constant stirring. This temperature was found to be sufficiently high to give a satisfactory rate of reaction in forming the dioxalatothiometastannate.
Special Solutions POTASSICM SCLFIDE.Dissolve 20 grams of potassium hydroxide in 100 ml. of water and saturate with hydrogen sulfide, keeping the solution cold. Keep this solution in an inert atmosphere, because thiosulfates and sulfites formed by oxidation are not removed in the procedure and will be oxidized by iodine, An inert atmosphere can be maintained by bubbling some carbon dioxide through the solution before stoppering the flask each time after use. WASH LIQUID. Dissolve 5 grams of oxalic acid in 1 liter of water and add potassium oxalate until a pH of 3 is attained.
The New Method
Procedure
In t h e first place, the tin present should be converted to a form in which i t is less readily hydrolyzed than t h e oxalatostannic acid. It was shown in studies on the complex oxalatostannates (2) t h a t t h e potassium salt was very stable and could be recrystallized from hot solutions. Moreover, dipotassium dioxalatothiometastannate is very much more stable (3) than t h e dioxalatothiometastannic acid formed according t o Wheeler's directions. Instability of the potassium salt in solution is caused by hydrolysis and can be prevented b y employing a large excess of potassium oxalate. Keither sodium oxalate nor ammonium oxalate is suitable for this purpose because they are not sufficiently soluble. Accordingly, it was decided t o convert the oxalatostannic acid to t h e potassium salt b y the use of neutral potassium oxalate t o increase t h e pH. The proper p H was obtained from a study of the oxalatostannates and the dioxalatothiometastannates. T h e upper limit is a p H of 5, since a titration of potassium oxalatostannate (2)showed t h a t the hydrolysis of t h e tin takes place at this pH. T h e stable pH for dipotassium dioxalatothiometastannate is i n t h e region 3.3 to 3.5. The reactions involved may be represented by the following equations:
Dissolve about 0.15 gram of tin in 2 ml. of 18 -V sulfuric acid and several milliliters of concentrated nitric acid in a 200-ml. electrolytic beaker. Evaporate the solution to fumes of sulfuric acid to expel the nitric acid. Too large a quantity of sulfuric mid cannot be used because too much potassium oxalate is required to raise the pH to the desired point. Cool, and add 2 to 3 grams of solid potassium oxalate to the concentrated acid solution. M'ash the sides of the beaker and the cover glass with about 20 ml. of the wash liquid. At this point and during the subsequent addition of potassium oxalate, all the solid will not dissolve because much potassium bioxalate is formed. Continue adding solid potassium oxalate until a pH of 2.5 is attained (a range of 2.2 to 2.8 is permissible). A glass electrode was used in this work, but in later work the following indicators were found to be satisfactory: thymol blue, add potassium oxalate until all the red tint disappears; xylenol blue, add potassium oxalate until the pink tint disappears; metacresol purple, add potassium oxalate until the red tint disappears. After a little practice, no trouble was encountered in adjusting the pH to the proper range. Heat the solution, diluted to about 60 ml., t o 60", and, while stirring mechanically, add dropwise 3.5 ml. of the potassium sulfide solution. The beaker should be covered with a split watch glass because considerable gas is evolved. Continue heating a t this temperature for 5 minutes to ensure decomposition of the excess potassium sulfide and to ensure complete reaction with the tin. Add a few drops more of the potassium sulfide solution, continue heating for 1 minute, and cool the solution to room temperature under a stream of cold water. Neither the temperature nor the time of heating was found to be critical (temperatures from 55" to 80" C. and times ranging from 3 t o 10 minutes were satisfactory). At this point the volume should be sufficiently great (about 120 ml.) to prevent the precipitation of any solid material which would occlude some of the sulfur complex. Using mechanical stirring, pass in carbon dioxide until the escaping gas gives no test for hydrogen sulfide with lead acetate paper. (The
+ +
K6Sn~(C~04)7 2H2S 2K2SnS(C204)2 212
-
-
2K2SnS(GO& + 2H~C204+ K2C2Or + 2H~C204+ KK~Snz(C~04h G01 + 4HI + 25
With these modifications, a sample of pure tin gave the following results:
INDUSTRIAL AND ENGINEERING CHEMISTRY
718
dry tank carbon dioxide causes deposition of solid material by evaporation in the entry tube and may clog it. This difficulty is eliminated by bubbling the carbon dioxide through n.ater before passing it into the solution.) This usually requires shout 30 minutes but much longer if stirring is omitted. Titrate the solution with 0.1 .\-iodine solution to a faint yellow color (about 1 nil. past the end point), add starch solution, and bark-titrate with standard thiosulfate. One milliliter of 0.1 iodine is equivalent to 0.005935 gram of tin.
A moderate variat'ioii in pH is permissible. Because of the increased solubility of hydrogen sulfide, the addition of potassium sulfide a t higher p H values gives local concentrations sufficient to precipitate stannic sulfide. This can be dissolved by long boiling, but such a procedure is troublesome. High pH values also increase the difficulty of removing the excess of hydrogen sulfide.
Application to .4lloys Containing Tin In the case of pure tin, it has been demonstrated that this method of analysis gives accurate results. I n alloys the problem becomes one of a separation from those substances which would interfere with the titration. Table I11 shows the results obtained on Sational Bureau of Standards samples.
of Standards Saninlc I o .
13111..
Tin Piesent
pH after Addition of X2C201
Tin Found
C. /C
cc
.jAa, Tin-bane b e a i iiix inrtal
88.61
2.50 2.57 2.61
88.68 88,73 88.48
53. lead-babe b e a s i n bi metal
10 91
2.43 2.38 2.43
10.87 10.90 10.93
.;2, c a 8 t bronze
i.88
2.80 2.61 2 65
7.87 2.88
37b, sheet hraza
1 .oo
2.45 2.45
1 .UIJ 1 00
,
88
STASDKD SallrPLE 54a. This is a high-tin bearing metal with small quantities of antimony and copper. A sample of about 0.15 gram \vas weighed into a 150-ml. beaker and dissolved with 2 ml. of 18 N sulfuric acid and 1 ml. of concentrated nitric acid. The solution was evaporated to fumes of sulfuric acid and treated according to the procedure given above. The addition of the pohssium sulfide caused precipitation of the antimony and copper, and the solution was filtered hot into a 250-ml. wide-mouthed Erlenmeyer flask after the &minute heating period. I t was found best to use suction and a small funnel fitted with a Witte plate on which was placed a filter paper covered n-ith asbestos. The beaker was washed out with about 50 ml. of hot Jvash solut,ion. With constant stirring, the filtered solution was heated to 60", 0.5 mi. of potassium sulfide solution added, and the heating continued for another minute. The solution was cooled to room temperature, carbon dioxide vias passed through, and it was titrated as already described. For passing t,he carbon dioxide through the solution, it was found convenient to attach a 1-mm. capillary tube on the side (near the bottom) of the Erlenmeyer flask. Small quantities of antimony or copper sulfides which may have passed through the filt.erdid not interfere with t>hetitration. STANDARD S.mPLE 53. This sample contains about 79 per cent lead, 10 per cent antimony, and 10.91 per cent tin. Attempts to make all the separations in one operation or to separate the lead as sulfate, chloride, or oxalate gave low results. The alloy was dissolved in nitric acid, and the tin and antimony n-ere precipitated as insoluble acids. The precipitate was filtered off, and washed with hot water, and the whole filter paper was decomposed by sulfuric and nitric acids to recover the oxides. When the antimony was allowed to remain in the pentavalent form, the subsequent treatment always left a somewhat turbid solution, and the addition of the potassium sulfide gave a very gelatinous antimony precipitate which retained appreciable quantities of tin. The antimony was, therefore, reduced to the trivalent state
Vol. 14, No. 9
hy adding some hydroxylamine sulfate to the concentrated sulfuric acid solution and boiling until the excess hydroxylamine sulfate was decomposed. Potassium chloride added with the potassium oxalate aided in keeping the antimony in solution. A 1-gram sample was dissolved in 5 ml. of concentrated nitric acid in a 200-ml. electrolytic beaker. After the reaction had ceased, 30 nil. of water and some filter paper pulp were added and the solution mas allowed to digest for 30 minutes, filtered hot through a small Whatman S o . 12 filter, and washed several times with hot water. The filter paper was transferred back to the beaker, and 2 ml. of concentrated sulfuric acid and 10 mi. of concentrated nitric acid were added. The beaker was covered with a watch glass and heated to destroy the filter paper. More nitric acid was added as needed, but no more sulfuric acid. The solution was then evaporated to sulfuric fumes, 0.5 gram of hydroxylamine sulfate or hydrazine sulfate was added, and the solution was heated strongly until all the reducing agent was decomposed. This point can be recognized as that when all the evolution of gas ceases. After the solution had cooled, a few grams of solid potassium oxalate were added, then 20 ml. of water and 5 grams of pot,assium chloride. The procedure from here was exactly the same as for the tin-base bearing metal. STAXDARD SAMPLES 52 ASD 37b. Both samples contain large amounts of copper, and the same procedure applied to both. Attempts to remove copper as a sulfide were unsuccessful. The most satisfactory procedure was a preliminary separation of the tin as metastannic acid. (In samples which contain large quantities of iron, the precipitation of metastannic acid is incomplete. Preliminary experiments indicate that this difficulty may be eliminated by adding about 2 grams of ammonium sulfate to the diluted solution of the alloy followed by boiling and digesting, or by reducing all the iron to the ferrous stat'e by adding some sodium bisulfite.) The metastannic acid vias filtered off and treated as in the lead-base bearing metal except that no hydroxylamine sulfate or potassium chloride was added. For the bronze :I 1.5 gram sample was used, and for the sheet brass a 5-grain wmple was taken.
Discussion The method described is rapid, and the interferences are fern. It has an advantage over the volumetric method of reducing tin to stannous chloride and oxidizing i t with iodine in that special precautions to maintain an inert atmosphere are unnecessary. I n the method described here, the tin is a t all times in the stannic state, its most stable condition. Hydrogen sulfide gas may, of course, be used instead of potassium sulfide, but the time required for the reaction will he greater and more potassium oxalate will he needed.
Summary A solution of stannic tin is converted to potassium oxalatostannate, KeSn2(C2O4)7,by adding potassium oxalate to a definite p H (a range of 2.2 to 2.8 is satisfactory if potassium sulfide is subsequently used, or approximately 3.3 in case hydrogen sulfide gas is used). The salt is then converted t o dipotassium dioxalatothiometastannate, K2SnS (C204)2,by the addition of potassium sulfide or hydrogen sulfide. The excess of hydrogen sulfide is removed by a current of carbon dioxide, and the sulfur in the complex is titrated b y standard iodine. I n the presence of considerable amounts of other iiietals, the tin is first separated as metastannic acid. An inert atmosphere is required only t o remove the excess of hydrogen sulfide. Literature Cited (1) Wheeler, W-. C. G . , A n a l ~ s t63, , 883-4 ( 1 9 3 8 ) . (2) . . Willard, H. H., and Toribara, T. Y., J . Am. Chem. Soc., 64, 175961 (1942). (8) Ibid., p. 1762-5. PRESENTED before t h e Division of Analytical and Micro Chemistry a t the 103rd Meeting of t h e AMERICANCHEMICALSOCIETY,Memphis, Tenn. From a dissertation submitted b y T. Y. Toribara (Florence Fenwick Memorial Fellow, 1939-42) in partial fulfillment of t h e requirements for t h e degree of doctor of philosophy a t t h e University of Michigan, February, 1912.