tions, the two sets of compositions obtained for an unknown would not be significantly inaccurate. If better precision were required, a large number of runs could be accumulated to evaluate statistically the best factor for each component. When empirical response factors are determined, in addition to standard column and sample handling techniques, some of the more important precautions to watch for are temperature changes, flow rate variation, and base line overlap. It is not always possible to transfer a precise table directly to a different Chromatographic instrument.
We have also observed the effect of aging of the thermal conductivity wires in both of the instruments over a period longer than 6 months. In general, the magnitude of response diminishes gradually while the relative proportionality of the various components remains the same. The relative proportionality of various components also remains the same when a new Triton x-305 on Teflon column replaces an old one. LITERATURE CITED
(1) Bennett, 0. F., ANAL. CHEM. 36, 684 (1964). ( 2 ) Casazza, W. T., Steltenkamp, R. S., J. Gas. Chromatog. 3, 253 (1965).
(3) CriPPen, R.
c.,
Smith,
c. E., Ibid.,
*?! 37. (4) Foster, J. S.,Murfin, J. W., Analyst 90, 1;s (1965).
(5) Kaiser, R., “Gas Phase Chroma-
tography,” Vol. 3, chap. 8, Butter1963*
(6) Kirkland, J. J., ANAL. CHEM.35, 2003 (1963). (7) Messner, A. L., Rosie, D. AI., Argabright, P. A., Ibid., 31, 230 (1959). (8ksSom!;$$5gy*~ Acta Chem. Stand* 1 3 1 (9) Zarembo, J. E., Lysyg, I., ASAL. CHEW31, 1833 (1959). RECEIVEDfor review May 31, 1966. Accepted September 19, 1966. Presented
at the First Middle Atlantic Regional Meeting, American Chemical Society, Philadelphia, Pa., Feb. 3, 1966.
Extraction of Metal Ions with IsooctyI Thioglycolate JAMES S. FRITZ, ROBERT K. GILLETTE, and H. E. MISHMASH‘ lnstitufe for Atomic Research and Deparfment of Chemistry, Iowa State University, Ames, lowa lsooctyl thioglycolate (IOTG) is a water-immiscible organic complexing agent that quantitatively extracts bismuth(lll), copper(ll), gold(lll), mercury (II), and silver(1) from aqueous nitric acid solutions. Antimony(ll1) and tin(lV) are partially extracted, but other metal ions studied are not extracted from 0.1M acid. The extraction of bismuth(ll1) and silver(1) from 3M and 7M nitric acid, respectively, was accomplished. Bismuth, copper, mercury, and silver are easily back-extracted into aqueous hydrochloric acid and can be determined by standard analytical methods. Several additional metal ions are extracted from solutions of higher pH. Overlapping curves were obtained for the extraction of lead(l1) and zinc(l1) plotted as a function of pH. However, these metal ions are separated quantitatively at pH 4 on a column packed with a solid Columns support containing IOTG. can also be used in more acidic solutions to isolate small amounts of extractable elements from larger amounts of nonextracted substances.
and it does not have an objectionable odor. It is a complexing or chelating agent, reacting with the metal ions that form insoluble sulfides when reacted with hydrogen sulfide. It can be made selective by making adjustments in the acid concentration. IOTG is immiscible with water and only slightly soluble in water. It has a fairly high boiling point and is not highly viscous. These properties make it possible to use IOTG alone for solvent extractions, or it can be diluted with an organic solvent such as cyclohexane, chloroform, or ethyl acetate. It is one of the few complexing organic liquids that can be used in solvent extraction without a diluent. Of the metal ions studied, bismuth (111), copper(II), gold(III), mercury (11), and silver(1) are extracted quantitatively from aqueous solutions 0.1M or greater in nitric acid. Antimony(II1) and tin(1V) are partially extracted from acidic aqueous solutions. Other common metal ions studied are not extracted from 0 , l M acid, although some are extracted from less acidic solutions.
A
Apparatus. Absorption spectra were measured with a Bausch and Lomb 600 spectrophotometer. A Nuclear - Chicago scintillation well counter, Model DS 5, was used as the detector for experiments involving radioactive tracers. A Nuclear-Chicago spectrometer, Model 1820, isolated the gamma emission from the tracers used. A decade scaler counted the pulses received from the spectrometer. Reagents and Solutions. Commercial isooctyl thioglycolate (IOTG) , b.p., 125’ C. a t 17 mm., sp. gr. 0.9736 a t 25’ C. and thioglycerol (both from Evans Chemetics, Inc.)
EXPERIMENTAL
a number of sulfur-containing analytical reagents such as dithieone, dialkyldithiocarbamates, and thiophosphorous compounds (3) have been used for separation of metal ions b y precipitation or solvent extraction, most of these are broad-spectrum reagents that react with a rather large number of metal ions. This paper describes the use of isooctyl thioglycolate, HSCH2COOCSHU,as a selective reagent for the solvent extraction of certain metal ions. Isooctyl thioglycolate (IOTG) is available commercially as a pure liquid LTHOUGH
were used without further purification. A11 acids and organic solvents were analytical reagent grade. Solutions of metal nitrates were 0.05M and contained enough nitric acid to prevent hydrolysis. Solutions of tin(1V) and antimony(II1) chlorides were 0.05M and contained 2 X hydrochloric acid. An aqueous solution of O.05M chloroauric acid was prepared. A 0.05M aqueous solution of EDTA was prepared from reagent grade disodium (ethylenedinitri1o)tetraacetate and standardized by titration with ainc nitrate (pure zinc metal as primary standard) a t pH 6 using NAS indicator [0.5% aqueous solution of 7-(6-sulfo-2naphthylazo) - 8 - hydroxyquinoline - 5sulfonic acid, disodium salt (@I. The Teflon4 was a special type, 70 to 80 mesh, obtained from Analytical Engineering Laboratories, Inc. Extraction Procedure. Add exactly 10 ml. of 0.05M metal ion in 0 , l N or 1 . O X nitric acid to a 125ml. separatory funnel. Add 1 ml. of isooctyl thioglycolate and shake for 2 minutes. Add 10 ml. of chloroform, cyclohexane, or ethyl acetate and shake for 1 minute. I n some cases more chloroform vas necessary to keep the complex in solution. Use cyclohexane when a solvent lighter than water is desirable; it is satisfactory for all metals except copper. After the final shaking allow a few minutes for the phases to separate. Run off the lower phase, add a little chloroform (or water if cyclohexane or ethyl acetate is used), swirl gently and again run off the lower phase. To recover the extracted elements, back-extract the organic phase with an equal volume of aqueous hydrochloric acid. Use a t 1 Present address, Department of Chemistry, Kansas State University, Manhattan, Kan.
VOL. 38, NO. 13, DECEMBER 1966
1869
,100
.
E2
60
E 63 X
u
S
% I
40
rn
a ae
20
0
-
ACID MOLARITY
Figure 1 , Effect of acidity on extraetion of bismuth with IOTG
least 4-71hydrochloric acid for bismuth and copper, at least 8 V for mercury, and at least 1021 hydrochloric acid to backextract silrer. The extraction was made with pure isooctyl thioglycolate because the extraction could be made quicker. Analytical Methods. BISMUTH. Add 1 t o 5 ml. each of nitric and perchloric acids and evaporate t o fumes of perchloric acid in order t o remove all chloride. Dilute with water, adjust the pH t o 2.2 with ammonium acetate, and titrate with 0.05X E D T A using xylenol orange indicator. Determine smaller quantities spectrophotometrically (4). COPPER. Add approximately 5 ml. of nitric acid and evaporate nearly to dryness in % tall-form beaker. Adjust the pH to 3.8 with ammonium acetate and ammonium hydroxide and titrate photonietrically a t 425 nip with 0.05U EDTA using N A S indicator ( 2 ) . XERCURY. The analysis is made without evaporating the strong solution of hydrochloric acid because of the volatility of mercuric chloride. Adjust the pH to approxjniately 2 with concentrated ammonium hydroxide and then to pH 6 with pyridine. Titrate ITith 0 . U thioglycerol using a freshly prepared solution of Thiomichler's ketone in acetone as the indicator ( 1 ) . SILVER.Precipitate the silver from homogeneous solution by slowly evaporating off the excess hydrochloric acid, then determine the silver gravimetrically as silver chloride. c ~ 4 D M X J M , COBALT, ZINC. Titrate these
LESD,
NICKEL,
metal ions with O.05M EDTA using NXS indicator (2)* Column Separation of Lead (11) and Zinc(I1). Teflon-6 was used as the column support for IOTG for the reversed-phase chromatographic separation of lead and zinc. Dilute isooctyl thioglycolate with chloroform and add to the support, then evaporate the chloroform off a t room temperature. Slurry the stationary phase in mater and put into a 1.2 X 8 cm. column; then run water through until no more organic is removed from the column. Pass about 30 ml. of ammonium acetate-acetic acid buffer (pH 4) through the co!umn, add the sample and continue elution with p H 4 buffer a t about 1 ml./min. The zinc 1870
m
ANALYTICAL CHEMISTRY
passes through the column while the lead remains behind. When the zinc is completely removed, change the eluent to 0.1~11nitric acid to elute the lead. Column Separation of Traces of Bismuth(II1) from zinc(I1). Impregnate Chroniosorb W, 80 to 100 mesh, with IOTG and remove the excess reagent by blotting repeatedly with absorbent paper. Add water to the Chromosorb-IOTG several times and each time remove the organic material that floats to the top. Add the IOTG impregnated support to the column (1.2 X 4 cm.) in water and rinse quickly with about 20 ml. of 1-W nitric acid. Add a 5-ml. sample to the column; the samples used contained minor amounts of bismuth and about 5 mmole of zinc in 1M nitric acid. Keep the flow rate slow (about 0.5 1O0lq
,
1
I
I
I
1 -
60
n
-
E60 4 K
c
w" 40 ap
20
A-TIN WITH ETHYL ACETATE 8 - T I N WITH CYCLOHEXANE C-ANTIMONY WITH ETHYL 0-ANTIMONY AND TARTRATE WITH ETHYL ACETATE E-ANTIMONY WITH CYCLOHEXANE
HYDROCHLORICACID MOLARITY
Figure 2. Effect of acid concentration on extraction of tin(lV) and antimony(1ll) with IOTG
ml./min.) during the sorption step. After the sample is completely sorbed, pass about 20 ml. of 1X nitric acid through the column a t about 2 ml./min. to remove the last traces of zinc. Rinse the column quickly with about 25 ml. of water to remove the nitric acid so that it will not interfere with the ultraviolet spectrophotometric determination of bismuth, Monitor the effluent with a spectrophotometer a t 327 mp until no absorbance is observed from nitrate. Elute the bismuth with 6 X hydrochloric acid and determine it spectrophotonietrically a t 327 mp. Regenerate the column by rinsing with water and then with 12' nitric acid a t very fast flow rates.
ping extracted bismuth from the organic layer. Silver(1) is quantitatively extracted into IOTG-cyclohexane from 0.1 to 7 M aqueous nitric acid using only a single extraction. The following distribution ratios were obtained for gold(I1I) and mercury(I1) using radioactive tracers: gold(III), D s 150 using 0.1M hydrochloric acid and IOTG in ethyl acetate; mercury(II), D s 1000 using 0.1M nitric acid and IOTG in ethyl acetate. Copper(I1) is best extracted by IOTG in chloroform because the copper complex is very sparingly soluble in ethyl acetate or cyclohexane. Extraction of copper(I1) from 0 , l M nitric acid into IOTG-chloroform was 9970 complete in a single extraction. From O.lM, 1 . O N nitric acid and probably from somewhat higher acid concentrations, quantitative separation of copper is obtained using two extractions with IBTG in chloroform. Studies on the extraction of tin(1V) and antimony(II1) are shown in Figure 2 . The extraction of tin is nearly complete from dilute hydrochloric acid into IOTG-ethyl acetate, but falls off rapidly with increasing hydrochloric acid concentrations. Results for antimony(II1) were difficult t o reproduce; they show significant but not quantitative extraction. Thus, tin and antimony appear to interfere in the batch type extractive separation of bismuth, copper, gold, mercury, and silver from other metal ions. However, data obtained using paper impregnated with IOTG (see below) indicate that tin and antimony are tightly held in the IOTG phase in this type of separation. Back-extraction with fairly concentrated aqueous solutions of hydrochloric acid permits easy recovery of extracted elements from the organic-IOTG phase. Data in Figure 3 show that bismuth(II1) and copper(I1) are quantitatively backextracted by approxiniately 4 X or greater solutions of hydrochloric acid, and mercury(I1) is back-extracted by 8 to 12M hydrochloric acid. Xilver(1) is returned to aqueous solution using 10 to 1 2 X hydrochloric acid. At this high 100
RESULTS A N D DISCUSSION
The extraction of bismuth(II1) from different concentrations of hydrochloric, nitric, and perchloric acid into IOTGcyclohexane is shown in Figure 1. The extraction is essentially quantitative in 0.1 to 3&! nitric or perchloric acid. Formation of water-soluble bismuth chloride complexes in hydrochloric acid and bismuth nitrate complexes in nitric acid apparently compete with isooctyl thioglycolate complexing. However, this effect makes 4 t o 12M solutions of hydrochloric acid quite usef d for strip-
I I
o
c
K HYDROCHLORIC - ACID 2 V LdA R I T Y4
z
Figure 3. Effect of HCI concentration on back-extraction of bismuth, copperr and mercury
'oom Table 1.
Quantitative Separations
Aqueous-organic Lead-bismuth Zinc-rnercurv Lead-silver Uranium-bismuth Cadmium-silver Zinc-tin Bismuth-copper a
Figure 4. Effect of pH on the extraction of lead and zinc
acid concentration, silver is back-extracted as a soluble, anionic chloro complex. Evaporation of niost of the acid gives beautiful crystals of silver chloride. Prior to the discovery of this backextraction method, silver was recovered by evaporating most of the organic solvent and oxidizing the residue with nitric and perchloric acids. Although this procedure ordinarily gives a metal perchlorate, most of the silver was obtained as a precipitate of silver chloride. Since the original extraction was carried out from aqueous nitric acid, the only source of the chloride in the precipitate was from decomposition of perchloric acid. The following elements are not extracted from 0.1M nitric acid by isooctyl thioglycolate using the extraction procedure given: alkaline earths(II), aluminuni(III), cadmiuni(II), chromium (111), cobalt(II), iron(III), lead(II), magnesium(II), manganese(II), molybdenum(VI), nickel(II), rare earths(III), uranium(VI), and zinc(I1). Some of these elements were extracted when the pH of the solution mas increased before the extraction. From qualitative chromatographic experiments using paper impregnated with IOTG, it appears reasonably certain that the following elements also are not extracted from 0 . W acid: scandium(III), thoriuni(IV), titanium(IV), vanadium(V), and yttrium(II1). Data for several quantitative separations are given in Table I. In the metal ion pairs listed in Table I, the first remains in the aqueous layer after extraction and the second is in the organic phase. The silver in the organic phase was determined by evaporating and oxidizing the solvent as described earlier. The other metal ions in the organic layer were back-extracted prior to analysis. The sample containing bismuth and copper waB extracted with IOTG in chloroform, then both metals were backextracted with aqueous 4Ll[ hydrochloric acid. After removal of the excess hydroc hloric acid, bismuth and copper
by !blvent Extraction.
Aqueous, mmole Taken Found
Organic, mmole Taken Found
0,4561 0.4802 0.4556
0.4561 0.4797 0.4581
0.4881
0.4836 0,4802
0.4841 0.4802
...
...
0.4886
...
0.5010 0.4881 0.4982
0 .io03 0.4881 0.4989
Bi, 0.2520 Cu, 0.2469
0.2535 0,2448
Results are averages of 3-4 t,rials.
were determined by a differentiating photometric titration with E D T S a t 745 m/.l(@. Several additional elements form extractable complexes with IOTG in less
Table 11. Quantitative Separations on u Chrorrtosorb W-IOTG Column
Bi taken, mole
Zn interference, mmole
Bi found, mmole
0.00494 0.1482 0.1976
5 15 20
0.00497 0,1486 0.1970
t5
5
E
! i 8 -I
a Y
0
40 60 80 EFFLUENT VOLUME, ML.
20
100
Figure 5. Elution curve for the ZnPb separation column: 1.2 X 8 cm. Zn: 0.025 rnrnole Pb: 0.05 rnrnole
acidic aqueous solutions. The extraction of lead and zinc into IOTG in ethyl acetate from aqueous solutions was studied as a function of pH, and curves were plotted of the percentage extraction against p H (Figure 4). These curves overlap, indicating that the separation of lead and zinc by batch extraction is not possible. A quantitative separation of lead and zinc was obtained with a column of Teflon-6 impregnated with IOTG. The column was conditioned with a p H 4 acetate buffer, then the sample was added and the separation accomplished by further elution with pH 4 buffer. Finally the lead was stripped from the column with 0.1Xnitric acid. The elution curve is shown in Figure 5. Columns impregnated with IOTG are also useful for extraction of elements from more acidic solutions. This technique appears to be particularly valuable for the isolation and concentration of traces of extractable elements. A column of Chromosorb W impregnated with IOTG effected a quantitative separation of small amounts of bismuth (111) from large amounts of zinc(I1). The metal ions retained by the column may be readily eluted with aqueous hy-
drochloric acid. Rata for several separations using these columns are given in Table 11. The columns could be used several times before they had t o be prepared freshly. PBper chromatography with Whatman-1 chromatographic paper impregnated with a solution of 570 10TG-95y0 cyclohexane was used to check for interfering elements and for small-scale separations. The chamber consisted of a large round glass jar lined with 3-mm. paper, which had been soaked in the developing solution and covered with a glass cover. At least one hour was allowed for the chamber to become saturated. When the developing solution was 0 , l N nitric acid, Hg(II), Au (111), As(III), Sb(III), Ag(I), Cu(II), Sn(IT'), and Bi(II1) remained a t the origin, Th(IV), U(VI), Ti(IV), Al(III), Zn(II), Ce(III), Pb(II), Ni(II), Fe(III), Y(III), Sc(III), Er(III), V(T), and Cd(I1) traveled with the solvent front, and Pt(IV), La(III), and Co(I1) formed
A IOTG
J
C Bi -IOTG
4
D
CU-IOTG E Au-IOTG
200
250
300
Figure 6. Spectra of metal-IOTG complexes
1
450
350
400
WAVELENGTH, mp
IOTG
VOL. 38, NO. 13, DECEMBER 1966
and
* 1871
streaks. When the developing solution was Q.li11 hydrochloric acid, Ag(I), Cu(II), Sn(IV), and Bi(II1) remained at the origin, Th(IV), U(VI), Ti(IV), A1 (111), Zn(II), Y ( V , Pt(IV), Cd(II), Y(IXI), Sc(III), Er(III), Ni(II), and Fe(II1) traveled with the solvent front, and Sb(III),Pb(II), and Co(I1) formed streaks. When the pH of the developing solution was increased, more elements formed streaks on the chromatograms. Ultraviolet and visible spectra of IOTG and its complexes with bismuth (111) copper (11), gold(II1) mercury (11) and silver(1) were recorded (Figure 6). The spectra were measured in chloroform. All of these metal-IOTG complexes except niercury(I1) are yellOW.
The combining ratio of IOTG to mercury(I1) was found to be 2 : l by a titration in methanol-water using Thiomichler’s ketone as indicator. This would be the expected combining ratio
for a divalent metal ion if a neutral complex is assumed. The logarithnis of the distribution ratios for extraction of lead (11) and zinc(I1) into IOTGethyl acetate were plotted against pH. I n each case a straight line of slope approximately one was obtained, indicating that only one proton per metal ion is ejected during complex formation. Quantitative results were obtained for the extraction of bismuth(II1) with a dilute solution of isooctyl thioglycolate instead of first extracting with pure IOTG. However, longer shaking was required with the diluted reagent. Experiments were undertaken to find the relationship between the distribution coefficient and the concentration of the extracting reagent. The concentration of bismuth in 1M nitric acid was kept constant a t 5.0 x 1Q-?11bismuth nitrate and the reagent concentration was varied from 2.9 X 10-3JI to 4.8 X 10-2M in cyclohexane. Equal volumes of aqueous arid organic mere shaken for
quid Extra I uoroaceto
two hours. The organic phase was then back-extracted with 6M hydrochloric acid and analyzed spectrophotometrically (4). When the log of the distribution coefficient was plotted against the log of the IOTG molar concentration, a slope of 2.07 was obtained. This implies that the complex extracted consists of two moles of IOTG to one of bismuth. LITERATURE CITED
(1) Frits, J. S., unpublished work, 1965. (2) Fritz, J. S., .4bbink, J. E., Payne, M. A,, ANAL.CHEY.33, 1381 (1961).
(3) HandJey, T. H., Talanta 12, 893 (1965). (4)Merritt, C., Hershenson, H. M., Rogers, L. B., AXAL. CHEM.25, 572 (1953). (5) Underwood, A. L., Ibid., 26, 1322 (1954). RECEIVEDfor review July 22, 1966. Accepted October 11, 1966. Presented before the Division of Analytical Chemistry, Fq’inter Meeting, ACS, Phoenix, Ariz., 1966.
e Iiu m(IV)
ication to the Purification and mination of Berkelium FLETCHER L. MOORE Analytical Chemistry Division, O a k Ridge National Laboratory, Oak Ridge, Tenn.
A new, simple, rapid method for the radiochemical purification and determination of berkelium i s based on liquid-liquid extraction of berkelium (iV) with QSM 2-thenoyltrifluoroacetone-xylene. The high stability of the berkeliurn(lV) chelate provides marked selectivity from aqueous sobtions of nitric, sulfuric, or hydrochloric acids. Sodium dichromate is an efficient oxidant for berkelium(lll) tracer under the conditions required for optimum chelation and extraction of berkelium(1V). Ten-normal nitric acid solution readily strips the berkelium from h e organic phase. Excellent sepcaration of berkelium is effected from many elements, including the alkalies, alkaline earths, trivalent lanthanides, ruthenium, zirconium, niobium, uranium, neptunium, plutonium, americium, curium, californium, iron, nickel, aluminum, and silver. Several useful analytical and process applications of this purification method are discussed.
associated with the radiochemical purification and determination of berkelium-249 were discussed recently (8). A two cycle IFFICULTIES
11
e
ANALYTlCAL CHEMISTRY
liquid-liquid extraction system was proposed a t the time. That method is of the ion association type, utilizing the solvents, di(2-ethylhexy1)orthophosphoric acid-heptane and tricaprylamine-xylene. Separation methods based on chelation with 2-thenoyltrifluoroacetone (TTA) possess much higher selectivity than those which function by the ion association mechanism. Indeed, TTA exhibits unparalleled selectivity for the chelation and extraction of zirconium(IV), neptunium(IV), plutoniuni(IV), cerium (IV), tin(IV), and iron(II1) (6, 7 , 1019).
The method described in this paper is based on the recent observation by the author that berkelium(1V) forms a highly stable chelate with TTA. It can be extracted essentially quantitatively from aqueous solutions containing nitric acid (0.5-3.5N), sulfuric acid (0.5l . O X ) i or hydrochloric acid (0.1N). Under these conditions very few metal ions form extractable chelates with TTA ( 6 , IO). The chelation and extraction of berkelium(1V) with TTA have not been reported previously. The extraction of berkeliurn(II1) with TTA at pH > 2 has been noted (3). Al-
though berkelium(II1) extracts readily under these conditions, the selectivity is very poor. Because of the considerable promise of a Bk(1V)-TTA system in separations chemistry, work was directed in this area. EXPERIMENTAL
Apparatus. An internal sample methane proportional counter with voltage settings of 2100, 2900, and 4300 was used for fission-fragment, alpha, and beta counting, respectively. A NaI (Tl) well-type scintillation counter, 13/4 x 2 inches, was used for gamma counting. Reagents. 2 - Thenoyltrifluoroacetone (TTA, M.W. = 222) is available from Columbia Organic Chemicals Co., Columbia, S. C. Procedure. Pipet a suitable fluoride-free aliquot, preferably containing 2 pg. or less of cerium, into a 50ml. borosilicate glass centrifuge tube. Adjust the aqueous solution to about 4.5-rnl. volunie containing about 1N nitric acid and 0.2X sodium dichromate. (As much as 100 pg. of cerium in the sample aliquot can be tolerated. I n this case, it is necessary to adjust the aqueous phase to about 4.5-ml. volume containing about 1N nitric acid-0.ZM
