Sorption of copper from dilute aqueous solutions onto poly(vinyl mercaptan)

Reactive Polymers, 4 (1986) 225-235

225

Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

SORPTION OF COPPER FROM DILUTE AQUEOUS SOLUTIONS ONTO POLY(VINYL MERCAPTAN) M. CHANDA *, K.F. O ' D R I S C O L L and G.L. REMPEL

Department of Chemical Engineering, University of Waterloo, Waterloo, Ontario (Canada) (Received August 2, 1985; accepted in revised form February 11, 1986)

A hydrophilic poly(vinyl mercaptan) (P VM) has been prepared from poly(vinvl alcohol) by partial conversion of the OH groups to S H groups. The resin with a free mercaptan content of 2.23 m e q / g has a copper sorption capacity of 70 m g / g . The polymercaptan used as a sodium salt, P VM(Na), has several-fold higher sorption capacity for copper than the same resin used in thiol from, P VM(H). From a solution containing Cu(II), copper is sorbed mostly as Cu(II) by PVM(Na) and as Cu(I) by PVM(H). Significantly, there is a large increase in the copper sorption capacity of both PVM(H) and PVM(Na) in the presence of sodium chloride in solution. In the absence of added NaCl, the copper sorption capacities of PVM(H) and PVM(Na) are influenced in different ways by the type of coion in solution. Thus, PVM(H) takes up more copper from copper(II) chloride solution than from copper(II) sulfate solution under comparable conditions, while the opposite is true for PVM(Na). These features can be explained by the formation of copper(I1) chloro complexes in the presence of added chloride, more efficient Donnan exclusion of multivalent coions, and greater degree of hydrolysis of copper(II) chloride than copper(ll) sulfate. The copper sorption on PVM(Na) is highly sensitive to pH, decreasing rapidly both at lower and higher p H from a maximum at p H ~ 5. In contrast, the sorption on P VM(H) is fairly constant in the mildly acidic p H range and increases rapidly above pH-~ 9, approaching the sorption capacity of P VM(Na).

INTRODUCTION

Conventional cation-exchange resins containing sulfonic or carboxylic acid functional groups have limited potential for removal and recovery of heavy metals from process solutions and waste streams because of their low selectivity, particularly for alkali and alkaline * On leave from Indian Institute of Science, Bangalore, India. 0167-6989/86/$03.50

earth metal ions. A number of chelating ionexchange resins have been produced by resin manufacturers to overcome this problem. Iminodiacetic acid resins, p o l y a m i n e (2aminomethyl pyridine) resins and aminophosphonic acid resins are some of the more widely used resins in this category [1-4]. A major aim of resin manufacturers has been to produce resins of higher selectivity and capacity for application in primary metal recovery. The ability of mercaptan-containing polymers

© 1986 Elsevier Science Publishers B.V.

226 to form stable mercaptides with heavy metals, such as Cu, Ag, Au, Hg, Pb, Cd, Co, Ni and Pd, has been used for the preparation of exchange resins as selective sequestering agents for removal of heavy metal ions. Poly(thiol styrene) [5,6] and poly(vinyl benzyl mercaptan) [7] are two of the earliest examples which proved effective for selective removal of mercuric ions. Because of the high chain transfer tendency of thiol groups, polymercaptans are synthesized by indirect routes, of which there are two general types, viz., polymerization of monomers containing a potential mercaptan group protected by a removable blocking group, and chemical modification of preformed polymers to introduce mercaptan groups. Lee and Daly [8] presented an exhaustive review of mercaptan-containing polymers synthesized by both of the above two general types of approach, highlighting both their problems and potential. An advantage of the latter approach is the possibility of employing suitable high molecular weight polymers in order to obtain desirable mechanical and physico-chemical properties in the derived polymercaptans. Since heavy metals recovery usually concerns aqueous process streams or waste water, it is advantageous to have p o l y m e r c a p t a n s with neighbouring hydroxyl groups to impart hydrophilicity and greater accessibility of the mercaptyl functional sites to aqueous species. Since mercaptans are easily synthesized from alcohols [9], hydrophilic polymercaptans could be obtained from poly(vinyl alcohol) by suitably limiting the conversion of hydroxyl to thiol to yield polymers which could be considered to be copolymers of vinyl alcohol and vinyl mercaptan. There are several reports of such poly(vinyl alcohol)-derived mercaptyl resins in the literature [10-12]. However, no data exist for such resins as regards their metal sorption behavior. This paper therefore presents such a study, initially selecting copper as the sorbate species.

EXPERIMENTAL Sorbents

Poly(vinyl alcohol) used for preparing polymercaptan was obtained from Polysciences, Inc., Warrington, Pennsylvania, U.S.A. It was a 99.7% hydrolyzed product, having molecular weight 78,000. The polymer was converted to polymercaptan in two stages. In the first stage, the polymer was converted to a thiouronium salt by reacting with thiourea and hydrochloric acid or hydrobromic acid (eqn. 1); in the second stage, the thiouronium salt was hydrolyzed with sodium hydroxide and then acidified to yield the mercaptan: NH2 (~)-on + S=C( NH2 +

NH2C1+HC1--> ( ~ - S - C (

(1) NH2

+

NH2C1- (i) NaOH (~)-S-C~ NH 2

> (~)-SH (ii) H2SO 4

(2)

In a typical preparation, 20 g poly(vinyl alcohol) and 40 g thiourea were added together to 500 ml of conc. HC1 (sp. gr. 1.19), magnetically stirred and heated to 60°C in a 3-necked 1-1itre flask fitted with a thermometer and a vertical condenser. After complete dissolution of the polymer, the temperature was raised to boiling with gentle reflux and maintained constant for 20 h. Using a water pump, 340 ml of the acid solution in the flask was distilled out. The remainder was then poured in a thin stream into 600 ml of 12% (w/v) NaOH solution with vigorous stirring, whereupon the polymer precipitated. The polymer slurry in alkali was heated under reflux in a nitrogen current for 5 h, filtered, and treated with an excess of 2 N H z S O 4 for 2 h. The slurry was finally filtered, washed with water and methanol, and dried at 50°C.

227 8O 60 40 2O

~- 8o
40 (b)

~- 80

3400

2340

1640

I100

630

WAVENUMBER ,cm-I Fig. 1. IR spectra of (a) poly(vinyl alcohol) film, and of (b) poly(vinyl mercaptan) (HC1 method) and (c) poly(vinyl mercaptan) (HBr method), derived from it. Spectra (b) and (c) were taken in KBr.

The resulting polymercaptan in thiol form is henceforth referred to as PVM(H). It was sieved to particle size less than 75 /~m. The product contained 48.76% C, 6.52% H, 7.16% N and 15.74% S. Another batch of PVM(H) was prepared in the same way as above, using, however, 400 ml aqueous HBr (sp. gr. 1.49) in place of HC1. The polymer contained 48.37% C, 6.07% H, 8.41% N and 21.45% S. The presence of nitrogen in the final product, indicating thiouronium groups, may be ascribed to the instantaneous precipitation of the polymer in alkaline medium and the consequent inaccessibility of some groups buried in the particles to alkali. The IR spectra of PVM(H) prepared by the HC1 method and the HBr method are compared with that of poly(vinyl alcohol) starting material in Fig. 1. The presence of residual nitrogen is confirmed by the strong C - N band at 1110 cm -1 and the N - H band at 1640 cm-1. Prolonged digestion of up to 30 h with

a large excess of alkali did not reduce the intensity of these bands, indicating that the residual thiouronium groups are inaccessible to the reagent, as explained above. Being inaccessible, these groups would be expected not to take part in the metal sorption process. This is corroborated by the fact that, as shown below, the measured copper sorption capacities of the resin in sodium form are in good agreement with the mercaptan content of the resin. The mercaptan content of PVM(H) was determined by equilibrating the resin with an excess of 0.1 N NaOH and estimating the amount of NaOH reacted with the resin. The value was 2.23 m e q / g for PVM(H) prepared by the HC1 method and 3.01 m e q / g for that prepared by the HBr method. These values are in good agreement with the copper sorption capacities--l.10 m m o l / g and 1.14 m m o l / g , respectively--determined by equilibrating the resins in sodium form with an excess of 0.3 M CuSO4 solution. The PVM(H) resins had redox capacities of 4.73 and 6.44 m e q / g with 12 in aqueous KI [5]. These values are considerably higher than what would be expected from the mercaptan content given above. This is ascribed to strong adsorption of I 2 by the resin itself, as was evident from the fact that the yellowish-white resin was stained brown by the iodine and the stain could not be removed completely. The bulk density and particle density of the resins are 0.33 and 1.95 g / c m 3 for PVM(H) prepared by the HC1 method and 0.36 and 1.20 g / c m 3 for PVM(H) prepared by the HBr method. The resins swelled in water to the extent of 52% and 45%, respectively. Though the PVM(H) resin prepared by the HBr method has higher mercaptan content (3.01 meq/g), as compared with the PVM(H) resin prepared by the HC1 method (2.23 meq/g), the latter resin, however, in addition to being cheaper, has much superior drainage characteristics, giving a significantly faster

228

drainage rate which constitutes a considerable advantage in handling the resin in sorption and regeneration cycles. In the present work, therefore, the PVM(H) resin prepared by the HC1 method was chosen for the sorption studies. Henceforth, PVM(H) will therefore refer only to the resin prepared by the HC1 method. A portion of PVM(H) was converted to the salt form by treatment with 2 N NaOH followed by washing with water and methanol and drying at 50°C. This polymer is henceforth referred to as PVM(Na). Sorption studies were carried out with both PVM(H) and PVM(Na) to afford a comparison of the polymercaptan in thiol and salt forms in sorption behavior.

(b) Cu(2P3/z) satellile peek7 •e~

• w"

Peak due to differential charging 7

•o

°w~

,.,,.. rj UJ I---

Cu (2P3/z) main peak ---....,, • ~,

°

k.t,'

¢

w r~

I

I

[

I

(o)

co Z 0 re"

I

I

I

I

.=

w _J w

I

•

Ie ~

Cu (2P3/z) p . ~

No satellite peak ~

945

• •

•

925

BINDING ENERGY (eV) Fig. 2. N a r r o w s c a n s (20 eV) of X-ray photoelectron

Sorption experiments

spectra of copper 2p region: (a) copper complex of PVM(H) and (b) copper complex of PVM(Na).

In all equilibrium studies, measured amounts of the resin were vigorously shaken with definite volumes of solution of known metal ion concentration for 20 h in tightly stoppered glass bottles at 25°C, using a gyrotory shaker with 2 cm eccentricity at 300 rpm. The residual metal ion concentration was determined by measuring the absorbance in ammoniacal solution using a Varian Model 219 spectrophotometer. A range of concentrations of the metal ion chosen for the study, Cu 2+, were employed. The sorption of Cu z+ was also measured as a function of time under same conditions as in equilibrium studies.

from PVM(H) and PVM(Na) are shown in Fig. 2(a) and Fig. 2(b), respectively. The absence or presence of a 2p satellite peak in this region is known [13] to characterize, respectively, cuprous and cupric compounds. It thus appears from Fig. 2 that copper is sorbed mostly as Cu(II) by PVM(Na) and as Cu(I) by PVM(H). The formation of Cu(I) on PVM(H) is explained by the reducing properties of mercaptans [14,15]:

R E S U L T S AND D I S C U S S I O N

Sorption mechanism Both PVM(H) and PVM(Na) were loaded to capacity with copper (0.47 and 1.10 mmol C u / g resin, respectively) by equilibrating them in an excess of 0.3 M CuSO 4 solution. Narrow scan (925-945 eV) X-ray photoelectron spectra of the copper complexes derived

2 RSH + C u 2 + ~ RS. + RSCu(I) + 2 H ÷ (3) The fact that the copper sorbed on PVM(Na) is in + 2 state suggests that the mechanism of copper sorption by the mercaptyl resin in sodium form is quite different from that in thiol form and that the sorption probably takes place predominantly by ion exchange (eqn. 4) instead of a redox reaction: 2 RS-Na + +

C u 2+ ~

( R S - ) 2 C u 2+ + 2 Na +

(4)

229

There is also a possibility that Cu(II) is precipitated as hydroxide by the mercaptyl resin used in sodium form. However, this is considered not to be significant in the present study because of the acidic pH of the copper(II) solutions which were, moreover, used in considerable excess of the stoichiometric amount based on the resin capacity. Moreover, the equilibrium sorption capacity of PVM(Na), as shown below, is significantly dependent on the copper(II) concentration in the substrate, especially in the lower concentration range, even though the Cu(II)-to-thiol ratio is greater than unity in all cases. In the event of chemical precipitation of copper(II) hydroxide by PVM(Na), its copper removal capacity in the low concentration range would be relatively insensitive to copper concentration in the substrate. In the following discussion, the removal of copper(II) by PVM(Na) was considered to take place primarily by ion exchange.

Sorption isotherm The equilibrium data for the sorption of copper from aqueous solutions by PVM(H)

and PVM(Na) are plotted against equilibrium solution concentration in Fig. 3. The results show that PVM(Na) has a 3 to 4 times higher sorption capacity than PVM(H), signifying that the value of the equilibrium constant of the copper sorption reaction given by eqn. (4) is considerably higher than that of reaction (3). Due to the release of protons from PVM(H) by redox reaction (3), the substrate became more acidic after equilibrium. However, no attempt was made to neutralize the acid during equilibration, since, as shown below, the sorption capacity is unaffected by acidic pH up to a sufficiently low value. The copper sorption behavior of PVM(H) and PVM(Na) in the presence of added salt (NaC1) is compared in Fig. 4. Unlike in conventional ion exchange, the salt is seen to promote the sorption of copper. This apparently abnormal behavior may be attributed to the tendency of chloride ion to form chloro complexes with copper(II). The equilibria can be represented by equations CuCln3 - n 1 + C1 - ~ CuCI~~ ) - n

't

K o = [CuCl~ "]/[CuCl,3,-~][C1 l j' n=1,2,3,4

2t 1.0

E ~- 08 c~ 0.6 u')

04 3

PVM(H)

0-

~o 0 2

0

0 2 4 6 8 I0 EQUILIBRIUM CONC. OF Cu2'" IN SOLN., rnmol/~

Fig. 3. Equilibrium sorption of copper from aqueous solutions of CuSO4 (pH 5.0-5.4) by PVM(H) (©) and PVM(Na) (zx). Resin concentration: 0.2-0.4% (w/v) slurry. Temperature 25°C.

(5)

By using spectrophotometric measurements McConnell and Davidson [16] concluded that only the chloro complexes for 1 _< n < 2 are present in detectable concentrations in aqueous solution; they determined the stability constants to be K ] = 1 . 3 0 + 0 . 0 3 and K z = 0.23 _+ 0.15 l/tool at 25.2°C. Therefore, even in the presence of excess chloride, only CuCI + and CuC1 z chloro complexes are formed in appreciable quantities. The formation of CuC1 + species with single cationic charge would be expected to result in a higher uptake of copper by PVM(Na) by the ion-exchange reaction RS Na + + C u C I + ~ R S

(CuC1 + ) + N a +

(6)

230 b) PVM(NoI

_ (0) PVM(H)

~ I:=

/-I.0M NoC/,

1.2

05M NoCI

1.0

~Z"

'0.8

-

n

j__,,o

LO.OM NaC~

~I

/~,0.5 M NoC~

M=,~,.,NoC~

=~ 0,4 a:~ 0.2

~OOM NaCe i

2

4

i

i

6

EQUILIBRIUM

J

8

I

I0

i I

i

2

0

i I r I i 4

r i

6

8

I

I0

CONC. OF Cu z+ IN SOLUTION, mrn01/t

Fig. 4. Equilibrium sorption of copper from aqueous solutions of CuSO 4 (pH 5.0-5.4) in the presence of NaCI by (a) PVM(H) and (b) PVM(Na). Resin concentration 0.2-0.4% ( w / v ) slurry. Temperature 25°C.

as compared with the simple C u 2+ exchange represented by eqn. (4). The promotional effect of excess NaC1 on copper sorption is seen to be more pronounced in the case of PVM(H). This may be attributed to N a + playing a useful role by virtue of its high concentration in the ion-exchange process, as given by the following equations: RSH + N a + ~ R S - N a + + H + R S - N a + + CuC1 + ~ R S - (CuC1 +) + Na +

(7)

-

(o)

Q8

IJ_l

PVM(Na)/CuoSOu

.

o

lz

PVM(H)/CuCf, z

,

0

(8)

The adverse mass action effect of excess N a ÷ on reaction (7) would not be significant if the equilibrium constant of the reaction is sufficiently high. In order to determine the effect of C1- in the absence of N a ÷, equilibrium sorption on both PVM(H) and PVM(Na) was also measured using cupric chloride solution in the

C

0 I-.,'~ 0.6 ¢r." 0 0'3 0.4 n.." 0.2 -----J 0 0

RSH + CuC1 ÷ ~ RS-(CuC1 +) + H ÷

(b)

1.0 E z

in addition to possibly some direct exchange with the thiol resin itself, represented by equation

I

,

I

,

I

,

I

0 2 4 6 8 2 4 6 8 I0 EQUILIBRIUM CON(:. OF Cu 2÷ IN SOLUTION, m m o l I [

,

I

I0

Fig. 5. Equilibrium sorption of copper from aqueous solutions of CuSO 4 and CuC12 by (a) PVM(H) and (b) PVM(Na). Resin concentration 0.2-0.4% ( w / v ) slurry. Temperature 25°C.

231

absence of NaC1. The results are presented in Fig. 5 and compared with the respective sorption from copper(II) sulfate solution. Interestingly, PVM(H) picks up considerably more copper from copper(II) chloride solution than from copper(II) sulfate solution, while the reverse is tue for PVM(Na). The effect on PVM(H) may be attributed in part to reaction (8) and in part to Donnan potential which excludes multivalent coions more efficiently than monovalent coions. In the case of PVM(Na) these effects are probably more than counterbalanced by the greater tendency for hydrolysis and consequent acidity of copper(II) chloride solution. The sorption capacity of PVM(Na), as shown later, is highly sensitive to solution pH. Representing the copper sorption by the thiol and sodium form resins (eqns. 3 and 4) by a general mass action equation of the form 2 R . A + B ~ (R)2B + 2 A

(9)

an apparent equilibrium constant, K a, can be defined by the equation Ka

(10)

--

where C represents molar concentration and the subscripts r and s denote resin phase and solution phase, respectively. The values of Ka were calculated from eqn. (10) using the data for copper sorption on PVM(H) and PVM(Na) in the absence of NaC1 (Fig. 4). Data on the sorption on

PVM(H) in the presence of NaCl were also used to calculate K a by the same eqn. (10) so as to afford a ready comparison. These values are presented in Table 1. It should be mentioned that the calculated values of K a were fairly constant over most of the range of sorption isotherms in each of the above cases. Average values are reported in Table 1.

Effect of pH on sorption capacity The effect of pH on the copper sorption capacity of PVM(H) and PVM(Na), both in the acidic and alkaline range, is shown in Fig. 6. The pH was adjusted by H2SO 4 in the acidic range and by N H a O H in the alkaline range. PVM(Na) is seen to be extremely sensitive to acidic pH, with the copper loading decreasing from a value representing 80% of the maximum resin capacity at pH = 5 to less than 10% of the capacity at pH below 2. In comparison, PVM(H) is less sensitive to acidic pH, with the sorption capacity decreasing to less than 10% of the maximum value at pH below 1.5. In ammoniacal solutions above pH 9, the copper loading of PVM(H) increases

,g F

p. ~.. r~ 0 o3

I0

0.8-

06 -

(Na)

""

0.4

TABLE 1 Apparent equilibrium constants, Ka, for copper sorption System

Ka

PVM(H)/CuSO4 PVM(H)/CuCI 2

0.013 0.036 0.16 0.33 14.7 1.45

g C~

0.2 f , . ~ 0

~ 0

I 2

~PvM(.) ,

I 4

i

I 6

,

I 8

~

I ~0

J

I 12

pH

PVM(H)/CuS04/O.05 M NaC1 PVM(H)/CuSO4/0.5 M NaC1 PVM(Na)/CuSO4 PVM(Na)/CuC12

Fig. 6. Effect of pH on equilibrium sorption of copper from CuSO 4 solution by PVM(H) and PVM(Na). Initial feed concentration: 4.0 mmol/l. Resin concentration: 0.2-0.4% (w/v) aq. slurry. Temperature 25°C. pH adjusted by H2SO 4 in acidic range and by NH4OH in alkaline range.

232

while that of PVM(Na) drops, both approaching the same value at strongly alkaline pH. These results are attributable to PVM(H) and PVM(Na) becoming nearly identical in chemical form both at low pH and high pH levels of the substrate.

where C is the volume swollen resin capacity (mol/1), C is the total concentration of counterions in solution (mol/1), D is the diffusion coefficient in solution (cm2/s), ? is the average radius of wet resin particles (cm), 6 is the fihn thickness (cm) and K A is a dimensionless selectivity coefficient defined by (CA)n(CB)s/(CB)n(CA)s, the value of I was calculated substituting appropriate values of C_ D, ~, K A (calculated from equilibrium sorption data), D (as calculated above from the sorption kinetic data), and assuming ~ -10 -3 cm (for well-stirred solution). For the lowest solution concentration used in kinetic runs I is approximately 0.24. Since this value is much smaller than unity, the sorption of copper on PVM(Na) under the conditions employed may be assumed to be particle-diffusion controlled. The results of the "interruption test" [25], presented in Fig. 7(b), which was carried out for the lowest solution concentration used in the kinetic runs, also indicate particle-diffusion control of the sorption kinetics.

Kinetic consideration

The copper sorption kinetics were measured only on PVM(Na), since it behaves more like an ion exchanger, whereas the sorption on PVM(H) is complicated by the redox reaction causing conversion of Cu(II) to Cu(I). The sorption of copper on PVM(Na) was measured as a function of time under conditions of vigorous agitation at ambient temperature using a gyrotory shaker with 2 cm eccentricity at 300 rpm. Further increase of shaker speed did not increase the sorption rate. The sorption data are plotted in Fig. 7(a): The rate of attainment of equilibrium sorption is seen to be nearly independent of substrate concentration. Using the simple criterion developed by Helfferich [24] I-

CD8 (5 + 2K A) >< 1 CD?

I--Z

Copper sorbed on both thiol- and sodiumform PVM resins can be stripped by either

(16)

1.0

I.L

z

Stripping behavior

Z

~-~~,~INTERRUPTION

~0.8

I--

I-- 0 <~ ~ 0 . 4

z ~ :m

0

(b)

0 i

I

i

I I

i

TIME, h

i

J

I 2

J

I

I

20

J

I

40

I

I

60

I

I

80

TIME, min

Fig. 7. Rate of sorption of copper from CuSO 4 solution by PVM(Na) used as 0.2% ( w / v ) aq. slurry under vigorous agitation at 25°C. (a) Initial concentration of Cu 2+ in solution (pH 5.2): O, 4.09; zx, 5.92; and × , 8.04 mmol/1. (b) Interruption test for particle-diffusion control with initial Cu 2+ concentration 4.09 m m o l / l . - - , without interruption; . . . . . . , with interruption.

233 80 hi

a.. 60 130 i,

o

4-0

¢.,.'.'3 Z a._

20 b) PVM (Ne)

F--

I

2

3

I

2

3

4

TIME ,h

Fig. 8. Rate of stripping of copper from (a) PVM(H) and (b) PVM(Na) loaded to capacity. Stripping with 10 bed volumes of solutions at 25°C under vigorous agitation.

dilute hydrochloric acid or dilute ammonia at ordinary temperature, the former giving a higher rate of stripping in both cases (Fig. 8). The copper sorbed by the sodium-form resin, which, as shown earlier, is predominantly in + 2 state, is seen to be stripped considerably more rapidly than the copper sorbed by the thiol-form resin that contained a large fraction of the sorbed copper in + 1 state.

Regeneration of the spent resin According to eqn. (3), the process of metal sorption by PVM(H) is accompanied by oxidation of 50% of the initial mercaptan groups, resulting in disulfide cross-links, which cannot take part in further sorption of the metal ion by the redox process. Thus stripping of the sorbed copper with dilute HC1 or NH4OH, as described above, regenerates only about 50% of the initial mercaptan groups taking part in the redox reaction. It is known that the redox reaction of mercaptan-containing polymer can be reversed with strong reducing agents, such as sodium bisulfite, sodium hydrosulfite, sodium sulfide or monomeric thiols [5]. In the present work, the effectiveness of sodium bisulfite for regeneration of the oxidized resin was studied. To obtain the oxidized resin, the thiol-form resin was initially subjected to a redox reaction

with Ag + as it produced a larger proportion of the oxidized form. The oxidized resin was agitated with an excess of 10% sodium bisulfite solution for different periods of time. The degree of conversion of the oxidized form to mercaptan was estimated by measuring the silver sorption capacity of the resin before and after bisulfite treatment. It is evident from the results presented in Fig. 9 that the initial rate of regeneration is rapid, with nearly 70% conversion taking place within 1 / 2 hour at 25°C, and the rate is very slow thereafter. Thus at 65°C, about 70°% regeneration takes place in 1 / 2 hour and full regeneration in 4 h.

100 W N

"O

o

.-e~

~

o25 C

60

~.40 la_l z

,,,

20

laJ re" L

o

I

2

i

]

4

i

L

i

I

6 8 TIME, h

J

I

to

I

12

Fig, 9. Rate of conversion of the oxidized form of PVM(H) to mercaptan by treatment with 10% (w/v) aq. sodium bisulfite solution.

234

CONCLUSIONS

The hydrophilic poly(vinyl mercaptan), derived from poly(vinyl alcohol) by partial conversion of OH groups to SH groups, can be used efficiently for the removal of Cu 2+ from dilute aqueous solutions. The resin used in sodium form exhibits several-fold higher copper sorption than the same resin in thiol form, which is attributed to the higher value of the equilibrium constant of the ion-exchange process compared with the redox process with polythiols. The resin with a mercaptan content of 2.23 m e q / g has a maxim u m copper sorption capacity (in sodium form) of 1.10 m m o l / g , in good agreement with the mercaptan content. Sodium chloride has a significant promotional effect on the copper loading of both PVM(H) and PVM(Na). This may be attributed to the formation of the chloro complex CuC1 +, resulting in lowering of the cationic charge of the counterion. Copper is predominantly sorbed in the cupric state by the sodium-form resin and in the cuprous state by the thiol-form resin. The sorbed copper can be removed by stripping with dilute HC1 or NH4OH. The resin in the oxidized (disulfide) form can be easily regenerated by treatment with an excess of 10% sodium bisulfite solution to complete the redox cycle. The copper sorption process is particle-diffusion controlled over the whole period.

ACKNOWLEDGEMENTS

The technical help of Ms. Anna Grinshpun, who carried out a part of the analyses, is gratefully acknowledged. We thank Dr. W.J. Chauvin of the Surface Science Laboratory, University of Western Ontario, London for carrying out XPS analysis of the c o p p e r / polymer complexes. Finally, we wish to thank the Natural Sciences and Engineering Council

of Canada for financial support and the award of an International Scientific Exchange Award to one of us (M.C.).

REFERENCES 1 J.J. Wolff, Ion exchange purification of feed brine for chloralkali electrolytic cells. The role of Duolite ES 467, Oslo Symp. 1982, Ion Exch. Solvent Extr., Society of Chemical Industry, London, IV/62IV/74, 1982. 2 B.R. Green and R.D. Hancock, Useful resins for the selective extraction of copper, nickel and cobalt, J.S. Afr. Inst. Min. Metall., 82 (1982) 303. 3 K.C. Jones and R.R. Grinstead, Properties and Hydrometallurgical applications of two new chelating ion exchange resins, Chem. Ind., (1977) 637. 4 R.R. Grinstead, Copper-selective ion-exchange resin with improved iron rejection, J. Met., 31 (1979) 13. 5 H.P. Gregor, D. Dolar and G.K. Hoeschele, Polythiolstyrene--a new oxidation-reduction ion exchange resin, J. Amer. Chem. Soc., 77 (1955) 3675. 6 C.G. Overberger and A. Lebovits, The synthesis of poly-p-thiolstyrene, an oxidation-reduction polymer, J. Amer. Chem. Soc., 77 (1955) 3175. 7 J.R. Parrish, Selective ion-exchange from polystyrene, Chem. Ind., (1956) 137. 8 C.S. Lee and W.H. Daly, Mercaptan Containing Polymers, Advances in Polymer Science, Vol. 15, Springer-Verlag, Berlin, 1979. 9 R.L. Frank and P.V. Smith, The preparation of mercaptans from alcohols, J. Amer. Chem. Soc., 68 (1946) 2103. 10 Y. Nakamura, Insoluble high polymer containing thiol groups, Kogyo Kagaku Zasshi, 58 (1955) 269; C.A., 49 (1955) 14376. 11 J. Cerny and O. Wichterle, Polythiouroniumverbindungen, J. Polym. Sci., 30 (1958) 501. 12 M. Okawara and Y. Sumitomo, Polymers containing sulfhydryl groups derived from poly(vinyl alcohol), Kogyo Kagaku Zasshi, 61 (1958) 1508. 13 D.C. Frost, A. Ishitani and C.A. McDowell, X-ray photoelectron spectroscopy of copper compounds, Mol. Phys., 24 (1972) 861. 14 H.G. Cassidy, Electron exchange polymers--I, J. Amer. Chem. Soc., 71 (1949) 402. 15 K.A. Kun and R. Kunin, Macromolecular redox polymers. II. J. Polym. Sci., A1, 4 (1966) 847. 16 H. McConnell and N. Davidson, Spectrophotometric investigation of the copper(II) chloro-complexes in aqueous solution of unit ionic strength, J. Amer. Chem. Soc., 72 (1950) 3164.

235 17 A. Ringbom, Complexation in Analytical Chemistry, Interscience Publishers, New York, NY, 1963, p. 261. 18 G. Adams, P.M. Jones and J,R. Millar, Kinetics of acid uptake by weak-base anion exchangers, J. Chem. Soc., (1969) 2543. 19 R.E. Warner, A.M. Kennedy and B.A. Boho, Kinetics of neutralization of weak electrolyte ion-exchange resins, J. Macromol. Sci.-Chem., A4 (1970) 1125. 20 G. Schmuckler and S. Dolstein, lnterphase mass transfer rates of chemical reactions with crosslinked polystyrene, in: J.A. Marinsky and Y. Marcus (Eds.), Ion Exchange and Solvent Extraction, Marcel Dekker, New York, NY, 1977, Chap. 1.

21 J. Babjak, A study of ion-exchange kinetics involving chelating resin, Ph.D. Dissertation, Univ. Waterloo, Canada, 1982; Diss. Abstr. Int., 43 (1982) 1555-B. 22 M. Chanda, K.F. O'Driscoll and G.L. Rempel, Ion exchange sorption of thiosulfate and tetrathionate on protonated poly(4-vinyl pyridine), Reactive Polymers, 2 (1984) 269. 23 O. Levenspiel, Chemical Reaction Engineering, John Wiley, New York, NY, 1972, p. 361. 24 F. Helfferich, Ion exchange kinetics, in: J.A. Marinsky (Ed.), Ion Exchange, Vol. 1, Marcel Dekker, New York, NY, 1966, Chap. 2. 25 T.R.E. Kressman and J.A. Kitchener, Cation exchange with a synthetic polysulfonate resin. V. Kinetics, Discuss. Faraday Soc., 7 (1949) 90.