Separation of Metal Ions Using a Hexylthioglycolate Resin Elizabeth M. Moyers' and James S. Fritz" Ames Laboratory (ERDA) and Department of Chemistry, Iowa State University, Ames, Iowa 500 1 1
The synthesis, characterization, and some analytical applications of a new chelating resin are described. The new resin, HTG-4, is synthesized from a highly cross-linked, macroporous resin (XAD-4) and contains the hexylthioglycolate functional group. HTG-4 is highly selective for silver( I), mercury(11), bismuth( Ill), and gold( 111) In acidic-aqueous solution. The first three metal Ions are sequentially eluted with 0.5 to 6 M hydrochlorlc acid, and gold(ll1) is eluted with a thiourea solution. Several quantitative analytical separations are reported.
Studies have shown t h a t sulfur-containing ligands are more selective toward the noble and heavy metals than their oxygen a n d nitrogen analogues (1-5). Chelating resins employing these complexing groups have been widely used for concentrating a n d separating mercury(II), silver(I), a n d gold(II1) from each other (6, 7) and as a group from t h e base metals (8, 9 ) . Other chelating resins have been of valuable assistance in toxicity studies of mercury (10-13) and in recovering trace quantities of gold (14-18). Of these, Koster's a n d Schmuckler's benzyl isothiouronium resin ( 1 4 ) has received t h e most attention for its recovery of gold from sea water. By contrast, bismuth(II1) has received little attention in t h e realm of chelating resins (19)and t o date no actual separations have been performed. This paper describes the synthesis a n d analytical applications of a new chelating resin. This resin, HTG-4, has a hexylthioglycolate group attached t o a n XAD-4 polymer matrix through a n ester linkage.
m . 4 Synthesis of a resin of this type was suggested by the b e havior of isooctylthioglycolate (IOTG), a n organic liquid which can extract quantitatively silver(I), mercury(II), gold(III), bismuth(III), and copper(I1) from acidic aqueous solution (20). T h e new resin also retains t h e first four of t h e aforementioned metal ions from acidic solution. Separation of these metal ions from each other and from other metal ions is achieved by liquid chromatography using a short column of HTG-4 resin.
EXPERIMENTAL Instrumentation. The liquid chromatograph has already been described ( 2 1 , 2 2 ) .However, modifications have been made on the instrument. A system for monitoring the pH of the eluent after its passage through the resin column and an electronic noise filter which allows readings as low as 0.01 absorbance unit to be made are described in Ref. 23. A micrometering valve is used for fine adjustment of the helium pressure to give faster initial equilibration and diminished drift of the eluent flow rate. The eluent tanks, previously made from cast iron, are now light weight polycarbonate and Plexiglas to eliminate eluent contamination by iron and lead.
Present address, The Procter and Gamble Co., Cincinnati, Ohio.
Synthesis '-r
1
pet ether
-"-
I
anhydAIC13 (75 T I
_L
(XAD4)
I
T
+
HO(CH2)eOH
2 ml H 8 0 ,
(70
T)
-L
T
+ HOOCCH$H
2 m l Ha0 (70 "C)
m - 4
III The resin matrix, Rohm and Haas XAD-4, is air dried, Soxhlet extracted with methanol, ground, and sieved. The 250-325 mesh fraction is used for the synthesis after washing with 12 M hydrochloric acid, water, and acetone. To 5 g of XAD-4 resin are added 35 g of anhydrous aluminum trichloride and 20 ml of petroleum ether (60-110 OC). Then 20 ml of 5050 petroleum ether/acetic anhydride is added over a period of 30 min through a water cooled condenser, and the resulting mixture is refluxed at 75 "C for 3.5 h. Product I is cooled to room temperature and hydrolyzed in 12 M hydrochloric acid and ice. It is then collected by suction filtration and washed with 15 M ammonium hydroxide, 12 M hydrochloric acid, water, and acetone. This procedure is repeated to obtain a higher yield. I is light green when dry, dark green in acid, and brown in base. I is oxidized to the carboxylic acid by stirring for 1.0 h in 0.1 M potassium permanganate which is 2% in sodium hydroxide. 11, which is light yellow, is collected and washed in a like manner to I. I1 is mixed with 30 g of melted 1,6-hexanediol(mp 41 "C) containing 2 ml of 18 M sulfuric acid as catalyst. The mixture is reacted at 70 OC for 12 h. I11 is collected by a hot suction filtration using boiling methanol as the wash solution. I11 is added to 40 ml of thioglycolic acid containing 2 ml of 18 M sulfuric acid. The mixture is reacted at 70 "C for 12 h. The final product is collected by suction filtration and washed with 12 M hydrochloric acid, water, and acetone. The HTG-4 is tan in color and darkens somewhat with use. Characterization. To determine the hydrogen capacity, 0.2 g of the resin in the acidic form is weighed into 20 ml of distilled, deionized water. A 10-ml aliquot of 0.1 N sodium hydroxide is added and the mixture is equilibrated with stirring for 2.0 h. The excess base is collected by filtration and titrated to a methyl red end point with 0.1N hydrochloric acid. Sulfur contents are determined by Dumas' method (24).
For the potentiometric titration, 0.2 g of resin is slurried in 50 ml of distilled, deionized water and 0.1 N sodium hydroxide is added in 0.05 to 0.20-ml increments. Nitrogen gas is bubbled through the solution throughout the procedure. Between each increment, the mixture is equilibrated with stirring for 0.5 h and the pH of the solution is read. Batch capacities of the resin are determined by weighing 0.3 g of resin (batch 1) to 25 ml of 0.1 M perchloric acid-0.0002 M sodium hydrogen tartrate to which 10 ml of 0.1 M silver(I),mercury(II), or gold(II1) is added. After a 2.0-h equilibration, the mixture is filtered and the resin washed with four 10-ml portions of perchloric acidtartrate. The eluent from the filtration and the wash procedure is analyzed for metal content. Column capacities are determined for gold(II1) in 0.0 to 8.0 M hydrochloric acid media using 0.18 g of the resin in a gravity flow column. ANALYTICAL CHEMISTRY, VOL. 48, NO. 8, JULY 1976
1117
IO.00
7 ei
9.oot
6'ool i 5.00
-00.3 0.0
1.0
2.0
3.0
4.0
m l OF TITRANT
Figure 1. Titration of HTG-4 with 0.1 N sodium hydroxide
The resin is conditioned by passing 30 ml of the acid through the column at 2.0 ml/min. Similarly, 0.1 M gold(II1) in the acid solution is passed through the resin until saturation is reached. The column is then washed with 50 ml of acid to remove excess metal. The bound M thiourea. All solutions metal is eluted with 25 ml of pH 3.0, used in this study contain 2 drops of 0.01%bromine water per 100 ml to prevent gold reduction on the resin. Bismuth(II1) and mercury(I1) are determined by titration with 0.025 M EDTA to a Xylenol Orange end point, silver(1) by potentiometric titration with 0.1 M sodium chloride using a silver wire electrode, and gold(II1) by spectrophotometry as the thiourea complex. An ir spectrum of the HTG-4 resin shows bands for ester carbonyl and hydroxyl groups. Gold-Thiourea Spectrophotometric Determination. The absorbance of 1.2 X lo-* M gold(II1) in M thiourea vs. a M thiourea blank is measured on a Cary Model 14. Bromine and hy-
RESULTS AND DISCUSSION Several synthetic procedures were tried for steps 1-111, but those described in t h e Experimental section give the maxim u m yield with least degradation of the polymer matrix. Petroleum ether (60-110 "C) is used as solvent for the Friedel-Crafts acetylation because i t is inert, has a convenient boiling range, and, unlike nitrobenzene or dioxane, does n o t produce a tarry product. T h e yield of t h e acetylation can be increased b y recycling I a n d by using acetic anhydride rather than acetyl chloride. F o r the two esterifications, melted diol and liquid thioglycolate, respectively, are used as solvent and reagent. T h e sulfur content of t h e two batches of HTG-4 are 1.96 and 2.43 mmol/g, respectively. T h e total acidic hydrogen content of batch 1is 2.61 mmol/g, which is 1.84 mmol/g mercaptan and 0.54 mmol/g carboxylic acid b y potentiometric titration (Figure 1).The two titratable hydrogens have p K values of 4.40 a n d 7.90; these a r e assigned to the unreacted carboxylic acid and the mercaptan, respectively. These assignments seem valid because the pK of benzoic acid is 4.20 a n d that of ethyl mercaptoacetate is 7.95. Because of the mercapto group, t h e resin is not stable t o
T a b l e I. Capacity of Resin f o r Gold(III)a
M HCl pmol Au(III)/g resin 0.0 0.5 1.0 2.0 3.0 a
29.4 94.4 111 117 133
M HC1 pmol Au(III)/g resin 4.0 5.0 6.0 7.0 8.0
drochloric acid are omitted to show that the absorbance is indeed due to the gold-thiourea complex. Separation Procedure. Samples are made by weighing 10 to 400 mg of each of the metals, the metal salt, or the metal oxide into a digestion flask and wet ashing with aqua regia-hydrogen peroxide or with nitric acid alone when silver is present. The species used are reagent grade antimony, bismuth, copper, iron, silver, and tin metals, gold chloride, and mercuric oxide. The samples are diluted to 100 ml with distilled, deionized water and two 5- to 10-ml aliquots are taken. These are further diluted to 100 ml so that the samples injected are 1.0 to 20.0 ppm in analyte. All samples in which hydrochloric acid is used contain 0.3 to 0.4 g copper, and gold samples contain 2 drops of 0.01% bromine water per 100 ml of solution. The resin is conditioned with 20 ml of 0.1 M perchloric acid-0.0002 M tartrate at 2.0 ml/min and 51.4 pl of the sample is injected onto a 3.3 X 0.6 cm column of HTG-4 resin. After injection, the resin is further washed with perchloric acid-tartrate for 5 to 10 min. Bismuth(III), silver(I), and mercury(I1) are sequentially eluted with 0.5, 2.0, and 6.0 M hydrochloric acid using a stepwise gradient, mixed on line 1:l with 12 M hydrochloric acid and determined as the chloride complexes at 225 nm (25). Gold(II1) which is retained by the resin during this process is eluted by 25 ml of M thiourea-pH 3.0 in hydrochloric acid, and diluted by half with distilled, deionized water. After a 15-min wait, the gold(II1) is determined spectrophotometrically at 256 nm as the thiourea complex.
160 170 160 153 133
Data obtained on batch I1 of HTG-4.
T a b l e XI. Silver, Gold, Bismuth, a n d M e r c u r y Recovery in the Presence of O t h e r Metals a Molar ratio Metal detd Amount taken, ( p g ) Amount found, pg Recovery, % Ag Au Bi Hg a
0.861 0.409 0.412 1.092
0.856 0.409 0.412 1.093
Re1 dev,pph
99.4 100.0 100.0 100.0
Quadruplicate analyses using least squares plot.
0.6 0.1 0.5 0.3
Cu/M 18 44 73 15
Fe/M
Pb/M
Sn/M
Sb/M
3 3 12 2
1 2 1
...
...
2 5 1
2 2 2
1
1 bismuth and 2 mercury also present.
T a b l e 111.Bismuth Recovery in the Presence of O t h e r Metalsa Molar ratio Bismuth taken, pg
a
Bismuth found, pg
Recovery, %
0.292 0.283 97.0 0.394 0.387 98.4 0.310 101.4 0.306 0.403 0.404 100.2 Quadruplicate analyses using least squares plot.
1118
ANALYTICAL CHEMISTRY, VOL. 48, NO. 8, JULY 1976
Re1 dev, pph
Interf. ion
Inter/Bi
Cu/Bi
3.2 3.1 1.4 0.7
Fe(II1) SbW) Pb(I1) Sn(1V)
10 10 10 10
134 87
... 66
,
li 1
0.10-
w t-a
a
0.09-
vr
c_ z
3
W 0
z 4 m
w
m
I
I
I
5 IO MINUTES
0
d
IE
Figure 2. Separation of bismuth(lll), silver(l), and mercury(l1) in the presence of tin(lV), copper(ll), iron(lll), and lead(l1)
0
5
0
5 MINUTES
Figure 4. Effects of hydrochloric acid and copper(l1) on the recovery of bismuth(ll1) in 0.1 M perchloric acid-0,0002 M sodium hydrogen tartrate
Table IV. Gold Recovery in the Presence of Other Metalsa -
01
I I
,
I
pm Au(II1) Interfering found 2 M HC1 Bi(II1) Cd(11) Cu(I1) Fe(II1) Hg(I1) Pb(I1) Ru(II1) Sb( 111)
Sn(1V)
1.34 1.35 1.33 1.57 1.35 1.29 1.29 1.56 1.30
diff +0.02 +0.03 +0.01 +0.25 +0.03 -0.03 -0.03 +0.24
-0.02
pm Au(II1) found 6M HC1
diff
1.32 1.33 1.30 1.31 1.31 1.29 1.31 1.33 1.30
0.00 +0.01 -0.02 -0.01 -0.01 -0.03 -0.01 +0.01 -0.02
All samples, contained 1.32 pm of Au(II1) and 18.80 pm of interfering ion. Pt(II), Pd(II), and Rh(II1) interfered severely. Quadruplicate analyses were performed using least squares plot. a
ry(II), and 0.76 mmol/g for gold(II1). Since the sulfur content of this batch of resin is 1.96 mmol/g, a 1:1,2:1, and 3:l complex of the resin functional group with silver, mercury, and gold, respectively, is indicated. These capacity measurements were carried out a t p H 1,which is sufficiently acidic t o avoid any interaction of t h e metal ions with unreacted carboxyl groups in t h e HTG-4 resin. T h e column capacities of HTG-4 for gold(TI1) in 0 t o 8 M hydrochloric acid are given in Table I. Table I1 gives t h e results obtained in the separation of bismuth(III), silver(I), and mercury(I1) each from 5 foreign metal ions, and of gold(II1) from 7 other metal ions a t a 1:l molar ratio. These ions were chosen as those most likely t o interfere based on t h e strength of their metal t o sulfur complexes. Table I11 presents data for separation of bismuth(II1) from a tenfold excess of each of 5 different metal ions. In Table IV d a t a are given for quantitative separation of gold from a fifteenfold excess of various metal ions. These separations were performed both in 2 M a n d in 6 M hydrochloric acid. Bismuth(II1) is retained quantitatively by the HTG-4 resin up t o 0.2 M hydrochloric acid. However, when the bismuth(II1) sample solution contains 0.0002 M sodium hydroANALYTICAL CHEMISTRY, VOL. 48, NO. 8, JULY 1976
*
1119
gen tartrate, t h e bismuth is partially eluted even in 0.02 M hydrochloric acid (see Figure 4).This is evidenced by t h e reduced bismuth peak on elution with 0.5 M hydrochloric acid. It is likely t h a t this decrease results from t h e competition of t h e tartrate with t h e resin 1igand.for t h e bismuth(II1). As shown by t h e last chromatogram in Figure 4, this problem is solved by adding 0.3 t o 0.4 g of a copper salt t o all samples containing hydrochloric acid.
ACKNOWLEDGMENT T h e authors thank Rohm and Haas for a gift of the XAD-4 resin used in t h e synthesis, and John J. Richard who performed t h e sulfur analyses.
LITERATURE CITED (1) Yu Zolotov. "Oraanic Sulfur Reaaents", Seminar Iowa State University, Oct. 1975. (2) F. E. Beamish, Talanta, 14, 991 (1967). (3) E. E. Rakorskii and M. II Starozhitskaya, Zh Anal. Khim., 29, 2094 (1974). (4) R. F. Propistosova and S. B. Savvin, Zh. AnalKhim., 29, 2097 (1974). (5) L. A. Demina, 0. M. Petrukhim, and Yu. Zolotov, Zh. Anal. Khim., 25, 1463 119701 - -, (6) E. Bayer, Angew. Chem. 73, 659 (1961). (7) F. H. Pollard, G. Nickless, K. Burton, and J. Hubbard, Microchem. J., I O , 131 (1966). %
(8) A. Lewqndowski and W. Szezepaniak, Fresenius' Z.Anal. Chem., 202, 321 (19641. (9) E. Eayer, H: Fiedler. L. Hock, Dotterbach,G. Schenk, and V. Voelter, Angew. Chem., 76, 76 (1964). (IO) H. J. Kramer and B. Neidhart, Radiochem.-Rad. 22, 209 (1975). (11) D. E. Leyden and G. H. Luttrell, Anal. Chem., 47, 1612 (1975). (12) J. F. Dingman, K. M. Gloss, E. A. Milano. and S. Siggla, Anal. Chem., 46, 774 (1974). (13) P. Heizmann and K. Ballschmiter, Fresenius'Z. Anal. Chem., 266, 206 (1973). (14) G. Koster and G. Schmuckler, Anal. Chlm. Acta, 38, 179 (1967). (15) J. S. Fritz, and W. G. Millen, Talanta, 18, 323 (1971). (16) G. V. Myasoedova, 0. P. Eliseeva, S. B. Savvin, and N. I. Uryanskaya, Zh. Anal. Khlm.. 27, 2004 (1972). (17) G. V. Myasoedova, L. I. Bol'shakova, 0. P. Shroeva, and S. B. Savvin, J. Anal. Chem. USSR, 28, 1382 (1973). (18) Albert Zlatkis, W. Bruening, and E. Bayer, Anal. Chem., 41, 1692 (1969). (19) R. Hering, K. Trennl, and P. Neske, J. Pract. Chem., 32, 291 (1966). (20) J. S. Fritz, R. K. Giliette, and H. E. Mishmash, Anal. C h m . , 38, 1669 (1966). (21) M. D. Seymour and J. S. Fritz, Anal. Chem., 45, 1394 (1973). (22) J. S. Fritz and L. Goodkin. Anal. Chem., 46, 959 (1974). (23) M. D. Arguello and J. S. Fritz, future publlcation. (24) S. M. Ahrned and B. J. P. Whailey, Fuel, 51, 190 (1972). (25) L. Goodkin, M. D. Seymour, and J. S. Fritz, Talanta, 22, 245 (1975).
RECEIVEDfor review February 17, 1976. Accepted April 8, 1976. Work performed for t h e U S . Energy Research a n d Development Administration under Contract No. W-7405eng-82.
Exchange Rates and Water Content of a Cation Exchange Membrane in Aprotic Solvents Maria Lopez, Brian Kipling, and Howard L. Yeager* Department of Chemistry, The University of Calgary, Calgary, Alberta T2N 1N4, Canada
The properties of a perfluorinatedsulfonic acid ion exchange membrane (Nafion) in the solvents water, acetonitrile, and propylene carbonate are described. Equilibrium solvent compositions within the membrane phase for sodium and cesium ion forms have been determined using different membrane pretreatments. The membranes may be effectlvely dehydrated in nonaqueous media without prior heating. Rates of exchange of sodium and cesium ions for hydrogen ions decrease markedly in these solvents.
Several authors have investigated potential analytical a p plications of commercially available ion exchange membrane materials. Blaedel a n d co-workers (1-4) have demonstrated t h a t these materials are highly permselective and can function as effective preconcentrating devices; a corresponding analytical application has been reported ( 5 ) .Extensive use of ion exchange membranes in electrochemical systems has been made in both aqueous and nonaqueous media (6-10). Nafion-brand sulfonated fluorocarbon cation exchange membranes are chemically inert, highly permeable t o cations, permselective, a n d are available in a variety of forms (8, 11-13). T h e material is based on a perfluorinated ethylene backbone with pendant side chains of the form
where n < 20. A recent study suggests t h a t Nafion contains fewer impurity sites capable of binding metal ions than some other commonly used ion exchange membranes ( 4 ) .We are 1120
ANALYTICAL CHEMISTRY, VOL. 48, NO. 8, JULY 1976
interested in applying these membranes t o nonaqueous systems and have chosen t o study the properties of Nafion in the solvents propylene carbonate (PC) and acetonitrile (AN). These solvents were chosen because of their general suitability in electrochemical applications. We have performed tracer diffusion studies of alkali metal ions in Nafion for P C and water systems t o determine membrane diffusion coefficients ( 1 4 ) .Although satisfactory results were obtained in aqueous experiments, the nonaqueous results showed scatter. A recent study of the physical properties of Nafion suggests t h a t the water content of the membrane may be a significant factor in t h e morphology of the polymer (15). T o clarify significant factors controlling ion transport in Nafion, it is important t o characterize membrane composition under various experimental conditions. We report studies of solvent uptake and water content of Nafion as a function of counter ion a n d pretreatment procedure in P C and AN. I n addition, t h e rates of ion exchange in each solvent have been surveyed.
EXPERIMENTAL Materials. Ion exchange membranes were H-form Nafion-125 (Plastics Dept., Du Pont and Co.),with nominal capacity and thickness of 0.83 mequiv/g and 0.013 cm, respectively. Capacity measurements conducted on these membranes in water-equilibrated form yielded a value of 0.87 mequiv/g. Capacities of various samples of membrane used in this study were calculated from this value and the respective membrane densities. Densities were determined from the measured dimensions of weighed membrane portions. Sodium iodide (Fisher Certified) and cesium iodide (Alfa Ventron, 99.9%) were used without further purification. The purification and analysis of propylene carbonate has been described previously (16). Water content of PC used in this study was found to be 5 X M