Separation and Determination of Microgram Quantities of Tellurium in

of Microgram Quantities of Tellurium in Biologic Materials. Richard E. Kinser , Robert G. Keenan. American Industrial Hygiene Association Journal ...
1 downloads 0 Views 429KB Size
hand helps to judge the correct position). Then the balance is closed, and after 1 minute the beam is released. Thus, the correct rest point can be found. This performance must also be followed when any interruption delays the reading of swings much beyond the standard interval. ACKNOWLEDGMENT

The author is indebted t o John R. Dunning, under whose inspiring supervision this work was undertaken, and

t o Frank W. Hurd and Bernice Davis McAuliffe, who assisted with the experimental work. LITERATURE CITED

Benedetti-Pichler, A,, “Microtechnique of Inorganic Analysis,” pp. 180-3, Wiley, New York, 1942. Corner, M., Hunter, H., Analyst 66, 149 (1941). Corwin, A. H., IKD. ESG. CHEW, ANAL.ED. 16, 258 (1944). Hull. D. E.. “A Raoid Precise Microbalance TechniqGe,” U. s. Atomic Energy Commission, AECD-1816 (hIarrh 1948); Vortez (California

Section, ACS), p. 342 (preliminary report), September 1948. ( 5 ) Lindner, J., Mikrochemie 34, 67 (1948). (6) Pregl, F., Grant, J., “Quantitative Organic Microanalysis,” pp. 10, 13, Blakiston, Philadelphia, 1946. ( 7 ) Rodden, C. J., Kuck, J. A., BenedettiHuffman, Pichler, A,, Corwin, &4., E. W. D., I N D . EKG.CHEM.,A N A L . ED. 15, 415 (1943). (8) Waber, J. T., Sturdy, G. E., .IK.~L. CHEM.26, 1177 (1954). RECEIVEDfor review May 7 , 1954. (Resubmitted January 30, 1957). .-lecepted April 1, 1957.

Separation and Determination of Microgram Quantities of Tellurium in Urine C. K. HANSON Department o f Metallurgy, University of Ufah, Salt lake City, Utah

bMinute amounts of tellurium may be separated from all but a few elements and recovered quantitatively by extraction in n-amyl alcohol from solutions made approximately 1N in hydrogen ion and 0.6N in iodide. The extraction liquid after being washed is removed by evaporation. The residue, an essentially pure tellurium salt, is redissolved in hydrochloric acid, from which the element is precipitated in finely divided form and determined photometrically. Known, added amounts of tellurium were recovered satisfactorily by the process from wetashed urine samples.

0

methods for determining small amounts of tellurium involve photometric or visual comparison of suspensions of the element obtained in hydrochloric acid solution with stannous chloride (2, 3, 9). I n a more recent procedure, hypophosphite is used as precipitant under conditions carefully regulated for increased precision (6, 8). The more sensitive methods reported to date are those in which colored complexes of tellurium, notably the iodotellurite, are measured spectrophotometrically (5, 7 ) . With any of the methods preliminary separations are essential in all but the most simple solutions. Usual techniques involve precipitation of elemental tellurium by one of several available reagents, although stannous chloride is perhaps most often used. I n a series of analyses of human urines an attempt was made to determine the effectiveness of this precipitant. Quantities of element concerned were small. LDER

1204

ANALYTICAL CHEMISTRY

for the proposed allowable concentration of the toxic agent, tellurium, in urine is 20 y per liter (4). Known amounts of 5 to 20 y were accordingly added to 200-ml. samples prior to decomposition by wet ashing. Recoveries from the final hydrochloric acid solution were very unsatisfactory, being less than quantitative in about 10 to 15% of the samples. Failure of the precipitation method in the complex solutions derived from urine decomposition prompted attempts to isolate tellurium by solvent extraction. Of a number of solvents tried. n-amyl alcohol proved most satisfactory. From solutions made approximately 1N in hydrogen ion and 0.6N in iodide recoveries were complete. REAGENTS AND EQUIPMEN1

Standard solution containing 10 y of uurified tellurium uer ml. in 0.5N hvdrokhloric acid. Perchloric acid, 7Oy0. Hvdrobromic acid. 48%. containing 10 ml. of liquid bromine peyliter. Ammonium hydroxide, 1 to 1. Reagent grade sodium iodide, crystal or powder. Extraction liquid, prepared by adding 1 volume of ethyl ether to 2 volumes of n-amyl alcohol. Hydriodic acid, l N , freshly prepared by diluting 4iy0 reagent grade acid. Stannous chloride, 10% in 20% hydrochloric acid. Suitable photoelectric colorimeter with 440-mp filter. All-glass distilling apparatus for selenium distillation.

-

ANALYTICAL PROCEDURE

For application of the method to

urine, 200-ml. samples in 400-ml. beakers are used. Fifty milliliters of concentrated nitric acid are added, followed by evaporation, by gentle boiling, to 15 ml. After cooling, 15 ml. of 70% perchloric acid are added and slowly taken to fumes. Samples are again cooled. Cover glasses and the sides of the beakers are then washed down with distilled water. The fuming-washing down is repeated two or three times to make sure digestion is complete. A sample is then transferred to a 250ml. all-glass distilling apparatus by the use of 25 ml. of 4801, hydrobromic acid containing 10 ml. of liquid bromine per liter, and 50 ml. of water. The selenium is distilled off and may be recovered for determination. Distillation is permitted to proceed until the perchloric acid fumes strongly and the solution is water-clear. After cooling, the solution is transferred to a 125-ml. separatory funnel. I n making the transfer enough 1 to 1 ammonium hydroxide is used just to neutralize the sample. At the neutral point a voluminous white precipitate appears and is redissolved by 1 or 2 drops of dilute hydrochloric acid. With care, the total volume of solution in the separatory funnel may be kept within 55 to 60 ml. Six milliliters of concentrated hydrochloric acid are now used to rinse the distilling flask and added to the solution in the separatory funnel, followed by enough water, used also as a rinse, to bring the final volume to 70 ml. Usually a mark may be etched on the separatory funnel a t the 70-ml. level before use. Finally 6.6 grams of sodium iodide are added. The resulting solution becomes approximately 1N in hydrogen ion and 0.6N in iodide. The tellurium is now extracted with 2 volumes of n-amyl alcohol, to which has been added 1 volume of ether. The ether contributes nothing to the extrac-

tion but produces cleaner, faster separations of t'he water-solvent layers. Twenty milliliters of the mixture are poured into the separatory funnel and shaken for 30 seconds. After separation, the bottom Jyater layer is drained into a second separatory funnel. Here the extraction is repeated with a new 20-ml. portion of solvent. The water layer is extracted a third time after removal to a new separatory funnel and then discarded. Each portion of extraction solvent is non- shaken with a separate 15-ml. portion of I S hydriodic acid solution t o remove salts. The wash layers are permitted t o separate thoroughly and are combined in a fourth separatory funnel, \\-here they are extracted with a 20-nil. portion of the amyl alcohol-ether mixture. The water layer after complete separation is discarded. The four portions of alcohol-ether used in t.he extractions are now brought in order into a 250-ml. beaker. The portion in the first separatory funnel is drained first int'o the beaker. The portion in each succeeding funnel is then transferred t o the funnel ahead and used for rinsing. The second portion of extraction liquid is added to the first in the beaker, after n-hich each of the remaining two is shifted t o one separatory funnel ahead in the line and used again as a rinse. Finally, all four portions of extraction mixture are in the 250-nil. beaker. The combined extraction liquids are carefully evaporated to 15 to 20 ml. This niay be done satisfactorily on a steam bath with covers removed from the beakers or more rapidly from a lowheat hot plate covered with an asbestos pad, if solutions are kept below boiling. Thirty to 40 ml. of nat'er and 5 ml. of 30% hydrogen peroxide are now added. Cover glasses are placed on the beakers and evaporation is continued until all the alcohol is gone. More water is added, if necessary, to keep the bottom layer at 10 t o 12 nil. The sample is then cooled and 10 ml. of concentrated nitric acid containing 2 ml. of 707, perchloric acid are added in small portions. When violent action has ceased. the sample is again evaporated until white fumes appear. The sample is partially cooled, the cover and sides of the beaker are washed down, and evaporation is continued without a cover until fuming again begins. Sou- a t as low heat as possible fuming is continued to dryness. Solutions approaching dryness should be water-clear. Practically no material will ever be visible a' residue in the beakers. To make t,he final deterniination of the tellurium, 3 ml. of concentrated hydrochloric acid and 2 nil. of water are poured into the cooled beakers, and 1 ml. of stannous chloride (10% in 20% hydrochloric acid) is added. A purplish colored suspension of fine tellurium will appear if any is present. I n some samples a slightly yellon- tinge may be present after addition of the hydrochloric acid. If it disappears upon addition of 1 drop of stannous chloride, is completed. If it does not, and a greenish yellow cast persists

or deepens, the sample must be discarded. Fortunately this occurs but seldom, if the procedure is carefully applied. If the sample seems uncontaminated, it is transferred to a 10-ml. volumetric flask, water for diluting to volume being first used in 1-ml. portions to rinse the beaker. After mixing, the sample is transferred to a suitable colorimeter, where the transmittance is measured n-ith a 440-mp filter. Values are compared with those obtained from standard samples. ANALYTICAL RESULTS

The extraction process was applied

to a number of standards. Five, 10, or 20 y of tellurium were measured directly by microburet into a 125-nil. separatory funnel. Hydrochloric acid, mater, and sodium iodide were added to give 1S hydrogen and 0 . 6 s iodide in 70-ml. volume. The extractions, evaporations, and measurements were then carried through as described. In Table I are some typical additions and recoveries.

Table I. Standard Samples Carried through Extraction Procedure

Sample No. 1

Te .\dded,

Te Recovered,

Y

Y

2

5 5

4.9

7 8

20

18.9 20.0

5.0

20

Sixty-two urine samples were analyzed by the procedure. Kone mere known to contain tellurium, but some donors were believed to have suffered mild exposure. To each sample in one lot of 44 (Group I, Table 11), 10 y of tellurium mere added just after the samples had been measured into 400-ml. beakers. Thus added tellurium was carried through the entire procedure along with any originally present. To each sample in another lot of 18 (Group 11, Table 11). 20 y were added.

Table 11.

The mean value of assays on the first 44 samples is 10.03; the standard deviation is 10.8. The mean value of assays on the group of 18 samples is 20.12; the standard deviation is i1.38. DISCUSSION

Some difficulty was experienced in evaporating the n-amyl alcohol for recovery of the tellurium. It was necessary t o exclude water as far as possible from the beakers. A fern milliliters under the alcohol layer sometimes produced violent bumping and loss of sample. This could be avoided by using dry beakers t o catch the extraction liquid, by allowing solvent and xater layers to separate completely in the funnels, and by evaporating a t a temperature just below the boiling point of the alcohol. If the covers were removed from the beakers, evaporation proceeded rapidly without boiling. h number of attempts to wash the extracted tellurium from the alcohol layer rather than to separate by evaporation were unsuccessful. I n one instance dilute ammonia m s used. If sufficient was present to neutralize any acid carried over by the alcohol and to give a weakly alkaline solution, very troublesome emulsions formed, Thich persisted for so long as to make the method impractical. The use of fixed alkalies was avoided, for it \vas considered advisable to make the final precipitation IT-ith stannous chloride in solutions containing little else than tellurium and hydrochloric acid. Pure n-ater or dilute acid-Le., 0.05X hydrochloric-would not wash the tellurium quantitatively from the alcohol layer. S o satisfactory explanation was found for the appearance of the persistent greenish yellow color in some samples a t the final precipitation. Products likely t o form in mixtures of n-amyl alcohol, ether, iodine, and hydriodic acid under conditions prevailing during the extraction or later during evaporation should be easily volatilized. The possibility of preventing appearance of the undesirable color by removing impurities in the extraction liquid before use was investigated. The amyl

Micrograms of Tellurium Recovered from Urine Samples Containing Added Tellurium

Group I 9.6 11.0 9.8 11 0 11.2 9.0

10.6 9.6 11.4 10.6 10.0

9.6 10.0 9.7 10.0 9.5 9.6 9.9 9.5 9.1 10.3 10.0

Group I1 11.6 10.0 10.0 8.9 10.3 11.6 10.6 9.5 10 6 9.9 11.4

9.5 10.5 8.9 9.3 9.2 9.7 11.5 8.8 9.; 10.0 9.0 VOL. 2 9 ,

19.4 21.6 19.2

20.4 19.4 20.0 20.0 19.8 21.4

NO. 8, AUGUST 1957

19.0 22.8 19.3 20.6 22.2 18.8 19.2 18.9 20.3

1205

alcohol-ether mixture was shaken for 30 minutes with a solution 1N in hydrochloric acid and 0 . 6 s in sodium iodide. Considerable iodine formed during the process. The solution layer ivas drained off, after which the solvent mixture was shaken with dilute sodium thiosulfate to reduce the iodine. This in turn was drained off. The amyl alcohol-ether was then washed several times with water and redistilled. This reagent s h o m d no improvement over untreated solvent. Samples for a standard curve were prepared by measuring 5, 10, and 20 y of tellurium from a solution containing 10 y per ml. directly into 10-ml. volumetric flasks. Hydrochloric acid (3 ml.),water ( 2 ml.), and stannous chloride (1 ml.) were then added exactly as in the urine analyses. After dilution to volume and mixing, per cent transmittance was read a t 440 mp: Values w r e > respectively, 92.2, 83.7. and 68.5. The gentle slope of the linear plot of these results is indicative of a rather Ion- sensitivity for the tellurium sol method. Duplicate standards prepared with less than 5 y of tellurium per 10 ml. gave considerable variation in transmittance readings. The portion of the curve above the 5-7 level was, on the other hand, more precisely reproducible. Rather than to attempt to use the lower reaches of the curve, it was decided that adding a t least 10 y of trllurium to each sample might

give better precision. Measurements could then be made in the reproducible range and the difference between the tellurium added and that recovered would be considered the quantity originally present. It would appear logical, considering these undesirable features of the tellurium sol method, to measure the iodotellurite complex spectrophotometrically. This was considered, but results in previous attempts to use this method on very small amounts of tellurium were disappointing, presumably because of intereference from the oxygen-iodide side reaction (1, 7 ) . I n the application specifically to urine analysis, interfering elements would be more likely to be encountered if the iodotellurite method were used. Most elements which form complex iodides under the experimental conditions are extracted to some extent by the n-amyl alcohol: bismuth, lead, mercury, gold, platinum, palladium, rhodium, and iridium. Bismuth is sometimes present in urine and lead more often. Bismuth would definitely interfere in the iodotellurite method. I n this investigation there was some evidence that lead would also interfere, and its presence is considered harmful ( 7 ) . Only the elrments which precipitate or girc colored solutions with stannous chloride interfere in the tellurium determination as presented. Thus the method is inadequate only when mer-

cury, gold, palladium, platinum, rhodium, or iridium is present. Because of their comparative rarity, however, it may be very frequently applied. ACKNOWLEDGMENT

The m i t e r wishes to express appreciation for suggestions offered by John R. Lewis, Department of lletallurgy, by Randall E. Hamm, Department of Chemistry, University of Utah, and by D . E. Rushing, chief chemist, U. S. Public Health Service, Occupational Health Field Service, Salt Lake City, 'C'tah. LITERATURE CITED

(1) Brown, E. G., Analyst

(1954).

>LA.,

79, 50-4

I""".

ger, K,, Durst, A. M.,2. anal. Chem. 135, 11-14 (1952) Johnson, R. A,, Andersen, B. R., ~ ~ K B L CHEU. . 27, 120 (1955). Johnson, R. A , , Kwan, F. P., Ibzd , 23, 651 (1951). Johnson, R. A,, Kwan, F P , \Testlake, D. IT.,Ibid., 25, 1017 (1953). Steinberg. H. H.. 1Iassari. S. C.. MineryA. C., Fink, R., J . Ind. Hyy: Tozicol. 24, 183 (1942). RECEIVED for review April 6, 1956. .kccepted April 10, 1957.

Sensitive Quartz Beam Microbalance A. W. CZANDERNA and J. M. HONIG Department of Chtmisfry, Purdue University, lafayette, Ind.

,A sensitive quartz beam microbalance is described which operates on the principle of a normal gas density balance. The instrument is capable of detecting mass changes to a precision of 5 X l o p 8 gram and to an accuracy of gram; with proper temperature control, it is stable over long intervals of time. The device is very rugged in relation to its sensitivity; it can be constructed, maintained, and operated without requiring unusual experimental skill.

&lop7

microbalances, capable of detecting mass changes of the order of lo-' gram or less, have been described in the literature (1-7, 9, 10, 12-15. 17, 18). These balances can be classified as belonging to one of N u m m OF

1205 *

ANALYTICAL CHEMISTRY

three categories--.piral types (helices), torsional suspensions, and adaptations of gas density balances (pivotal types). Instruments belonging to the first two categories can be made very sensitive, but their installation, operation, and proper maintenance often require a high degree of evperiniental skill and patience. I n particular, quartz helices are quite susceptible to breakage; their operation in high yacuum is complicated by oscillations which result from incomplete shock mounting. Torsion balances of conventional design likewise are sensitive to shocks and require frequent recalibrations. However, a new model has recently appeared, in which these undesirable characteristics are eliminated (16). The design for the balance described below represents an adaptation of the

pivotal model described previously (9, 17). I n both cases the authors were primarily interested in measuring gas densities; however, no clear-cut indication was provided concerning the optimum sensitivity, reproducibility, and long range stability which could be achieved. It thus seemed desirable to report the results of extensive tests on the balance described below-. This instrument is capable of yielding a sensitivity comparable to that of other models; moreover, it is extremely rugged in relation to its sensitivity because of the rigidity of all parts of the assembly. The balance also exhibits satisfactory long term stability. Reproducible results may be obtained much more easily with a balance of this type than with other models. Finally, the instrument offers the ad-