Indian Journal of ChemistryVol. 34A, May 1995, pp. 342-350

Removal of some anionic and cationic detergents using aninorganic gel adsorbent

S K Srivastava", V K Gupta, Anupam' & Dinesh Mohan

Department of Chemistry, University of Roorkee, Roorkee 247 667, India

Received 13 September 1994; revised and accepted 25 November f994

The uptake of two cationic detergents, cetyltrimethylammonium bromide (crAB), cetylpyridini-um bromide (CPB) and two anionic detergents, sodium dodecylsulphate (SDS) and sodium dodecyl-benzenesulphonate (ABS) has been observed on copper ferrocyanider which is prepared and charac-terised by well known methods. Sorption data have been correlated with both Langmuir and BETadsorption models. The presence of mono and divalent cations increase the uptake of ABS and SDSand the enhancement is much more in the presence of trivalent salt like AlCl3• Thermodynamic par-ameters indicate the feasibility of the process while kinetic measurements provide the necessary me-chanistic information. Desorption of anionic detergents from loaded copper' ferrocyanide has alsobeen achieved.

One of the most important series of compounds,from the view point of environmental pollution,has been synthetic detergents. Most commonlyemployed detergents are the anionic type, and be-ing non-biodegradable in nature, their removalhas been a subject of interest for quite some time.Out of all the methods tried for the purpose,adsorption at solid/liquid interface is extremelyimportant in view of its adaptability in wide rangeof applications. Adsorbents used, so far, are gran-ular carbon, carbon black, sulphur, y-alumina,silica and activated carbon I· 7.

Efforts made to use sand, flyash, digestedsludge, soil etc did not meet with much success.Recent investigations in this regard, still deal withcarbon, clay minerals etc''".

Inorganic gels. possessing high sorption or ex-change capacity are much better materials for theuptake of surfactants from aqueous solutions.These are mechanically and chemically morestable, exhibit specific selectivity and thus can beused even in drastic physical conditions. Theirstability, possibility of chemical regeneration andquantitative elution of adsorbate from these mate-rials, are some of the added advantages of thisclass of compounds and make the process econo-mically feasible.

Investigations presented in this paper includethe uptake and kinetics of the two cationic deter-

'Present address: National Environmental EngineeringResearch Institute, Nagpur, India.

gents viz., cetyltrimethylammonium bromide(CTAB) and cetylpyridinium bromide (CPB) andtwo anionic detergents viz. sodium dodecylsulph-ate (SDS) and sodium dodecylbenzenesulphonate(ABS) on copper ferrocyanide gel, under differentconditions of temperature, concentration, particlesize of the adsorbent etc.

Materials and MethodsCTAB, CPB, SDS and ABS were obtained

from BDH. All other reagents used were A.R.grade chemicals. All solutions, standards and dilu-tions were made using doubly distilled water. pHmeasurements were made with a pH-meter, mod-el pH 821, E.Cl., India. Absorbences were re-corded on Bausch and Lomb spectronic- 20 andUnicam-500 spectrophotometers. The mean par-ticle radius was evaluated using Agfa Diameter(Germany).

Preparation and characterization of the adsorbentmaterial

Preparation and characterization of copper fer-rocyanide were carried out as reported by Kou-rim et al." The product, brown in colour, wasground and sieved to different particle sizes. Themechanical and chemical stabilities of the gel areevident by the fact that no significant change inmean radius of the particles was observed whenkept in water or in detergent solution understudy.

The cation and anion exchange capacity of the

SRIVASTAVA et al.: REMOVAL OF ANIONIC & CATIONIC DETERGENTS 343

;;;0., !!

0 ..~ ~~ E.

II<~ ;./1 •• lOmg. 2.,.

compound as determined by standard method I~

are 2.60 meq/g and 0.21 meq/g respectively.

Adsorption experimentsSorption studies were conducted in a routine

manner by batch technique. Isotherms were runby taking different concentrations of surfactants(10-4 to 10-5 M in the case of cationic and 1O-sto 10-6 M for anionic detergents). at pH 6.2 and4.0 respectively. Stoppered glass tubes containingdetergent solution and a definite amount of gel(particles 170-200 B.S.S. mesh size) 'were stirred(10-15 min) and then kept for 12 h (in the case ofcationic detergents) and 48 h (in the case of an-ionic detergents) at 30°C with intermittent shak-ing. The supernatant liquid was centrifuged andana lysed for the detergent. The effect of NaCI,CaCl2 and AICl3 (varying concentrations+ on theuptake of anionic detergents has also been ob-served.

Desorption measurementsThe desorption of anionic detergents was tried

with OH- and Cl " ions and also with K2S04 anda mixture of C2HsOH and H2S04,

Kinetic studiesKinetic measurements were made by batch tech-

nique (finite bath system). A number of test tubes(50 ml capacity) containing a known volume(20 ml in each case) and concentration of solu-tions of the anionic and cationic. detergents wereplaced in thermostatic shaker. When the desiredtemperature was reached, a known amount of ad-sorbent (0.025 g) was added into each tube andthe solutions were agitated by mechanical shaking.After pre-decided interval of time, the solution of

t(hro)

the specified tube was separated from the sorbentmaterial and the uptake of the detergent was ob-tained as a function of time.

Estimation of detergentsAnionic detergents were estimated by the meth-

od of Lawrence" using methylene blue dye andthe method of Few and Ottawill!' was used forthe estimation of cationic detergents.

Results and DiscussionIn order to determine the equilibrium time and

quantity of adsorbent for optimum adsorption, ex-periments were conducted by varying the time ofcontact, using different amounts of copper ferro-cyanide at a fixed adsorbate concentration(5 x 10-5 M for cationic and 5 x 10-6 M for an-ionic detergents). Adsorbent concentration is vari-ed from 10 to 50 mg per 20 mJ of solution. Thesevalues of adsorbents were arrived at after a fewtrial runs. Results of these ~periments are shownin the form of plots of percent detergent removalversus time of contact in the case of SDS [Fig.1(A)]..Although...similar plots were obtained for allthe surfactants,. only the data for SDS has beengiven. Many preliminary runs indicate the equili-bration time as 48 hand 24 h for anionic and ca-tionic surfactants respectively. The logarithmicplot of the data shows a linear variation [Fig.1(B)] and the same may be mathematically repre-sented as

P, = k, 1m

or log P, = m log t+ log k, ... (1)

where P, is the percent solute removal and I thetime of mixing in hours. The slope m of the linear

logt(hro)

~ig. 1- Plots showing the percentage removal of SDS with time at varying copper ferrocyanide concentrations([SDSj= 5 x 1O-~ M, pH =4.0, particle size = 170-200 mesh, temp. = 30 ±0.2°C).

344 INDIAN J CHEM. SEe. A. MAY 1995

plot reflects the mechanism of adsorption whilethe constant k, is indicative of the rate factor. It isobserved that the amount of adsorbent affectsboth rate as well as the mechanism of detergentuptake. It is seen from the plots that the value ofm decreases in each case with increase. in theamounts of adsorbent, thus indicating a better ad-sorption mechanism at low adsorbent concentr-ations. Increasing values of the rate factor k, indi-cates an increase in the rate of removal at higheradsorbent concentration. Although percent deter-gent removal is more at high adsorbent concentr-ations, a large quantity of the same is comparat-ively less efficient if percent detergent removalper mg of adsorbent is considered. Moreover,very large quantities of adsorbent create handlingproblems and difficulties in the process of desorp-tion. Also taking into account the lesser removalof adsorbate at low concentration, 20 mg of ad-sorbent for cationic detergents and 30 mg of ad-sorbent for anionic detergents per 20 ml havebeen used. For running the sorption isotherms, amoderate concentration of adsorbate, 10 - 4 to10 - 5 M in the case of cationic and 10 - 5 to 10 - h

M in the case of anionic detergents, has beenchosen, keeping in view the surfactant levels likelyto be present in various industrial processes. Var-ious fractions of particle size have been tried andit was found that the uptake as well as desorptionis most convenient and maximum on particles of170-200 B.S.S. (0.082 mm) mesh size. Adsorptionisotherms of CPB, CTAB, SDS and ABS on cop-pcr ferrocyanide at pH 6.20 and 4.0 are given inFigs 2(A) and 2(B). All the isotherms [Figs 2(Aand B)] indicate a positive adsorption. These areregular and concave to the equilibrium concentra-tion axis. The extent of adsorption increases withincrease in detergent concentration, but the per-

2.0cpt

cent removal is high at low adsorbate concentra-tion. Further, the nature of isotherms, in eachcase, reveal a rapid initial uptake of detergent,followed by a slow approach to saturation condi-tion at high adsorbate concentration. Maximumuptake of detergents (both cationic and anionic) ismuch less than the cation or anion exchange ca-pacity of the adsorbent material.

The effect of pH on the adsorption of CPB,CTAB, SDS and ABS shows that the uptake ofCPB and CTAB increases with the increase in pHand that of SDS and ABS increases with acidityof the system. Variation of the uptake of CPB andSDS on copper ferrocyanide gel is given in Figs3(A and B). Figure 3(A) shows that the adsorp-tion of CPB increases with pH and the rise isquite sharp upto pH 5.0. Beyond this value amarginal increase is observed and thus the mea-surements of cationic detergents are made atpH 6.2. On the other hand the adsorption of SDSis almost constant upto pH 4.0 and beyond thisvalue, a sharp decrease in the adsorptionpheno-menon is observed as shown in Fig. 3(B). Thus allmeasurements of anionic detergents are made atpH 4.0. Adsorption data correlate well with theLangmuir and BET models. It is observed [Figs4(A,B) and 5(A,B)] that a good linear relationshipfor BET plots is only found at low concentrationof adsorbate while the entire adsorption data isbetter represented by Langmuir model.

BET as well as Langmuir parameters are givenin Tables 1 and 2. QO values as calculated fromBET expression for anionic and cationic deter-gents are much less as compared to those calcu-lated from Langmuir plots. Since BET plots haveonly been shown at low concentration of adsorb-ate, the same would not give a correct value ofthe amount of detergent necessary to form a

(A) (8)505

pH-4,O

Ad.OO'IMnt conc.=30m9 pe' 20ml

0.0(50S)

J.. l~O 1~' :z~(AU) •

( ••M conc.n".' •••• (Ce,,)·10

Fig. 2-Adsorption isotherms of (A) cationic, (8) anionic detergents on copper ferrocyanide (temp. = 30 ± 0.2°C, particl-size = 170-200 mesh).

SRIVASTAVA et al.: REMOVAL OF ANIONIC & CATIONIC DETERGENTS 345

tAl

pH

Fig. 3-Effect of pH on the adsorption of (A) cationic, (B) anionic detergents on copper ferrocyanide (temp. = 30 ± 0.2°C, par-ticle size-170-200 mesh) .

(A)

...••...

•••••c~u•~ 1.14.•2·~f""u

'~'~"~'---'I~A----~--~T----n~(••"'._.""., ••••(e.,.) I.'

:1.1c.:D••- Lit1!.-.!!....~ ••M

•• so (6)

G.21A8S

pH·4.0

AclMrbent co"cn.-30 ft'IG p«r 20",1•.H

••• '.1I

2.0I.'(A8S)

( ••••.concent,"ion Ce:...)~ 10'

Fig. 4-Langmuir plots of (A) cationic, (B) anionic detergents on copper ferrocyanide (temp. = 30 ± 0.2°C; particle size = 170-200 mesh).

Table I-Values of the constants of BET expression for anionicand cationic detergents

Ax 10-·( I mol"] Oil x Hlh (mol g-I)

36.20 44.20

Detergents

SDS

ABS

CPB

CTAB

11.53

54.95

45.66

0.909

monolayer on the adsorbent surface. Probablydue to limited validity of adsorption data for BETmodel and different' nature of sorption sites in-volved in the process of anionic and cationic ad-sorption, the BET parameter, A, shows great var-iation and does not reflect any consistency with

Table 2- Values of the constants of Langmuir expression foranionic and cationic detergents

h x 10-·(1 mol"] O"x 1(I"(molg-')

2.95 113.)2.50 H).6Y1.72 U\O

Detergents

SDSABS(,PB

('TAB 1.04

the adsorption behaviour of the detergents. Asmentioned earlier the Q(I values as calculatedfrom Langmuir model 'are higher for SDS andCPB as compared to ABS and CTAB. The valuesof the Langmuir parameter b which reflect thesteepness of the approach to saturation are higher

346 INDIAN J CHEM. SEC. A, MAY 1995

for SDS and CPB as compared to ABS andcrAB, thereby indicating greater tendency tow-ards adsorption. The values of the constant Awhich reflect the energy of interaction with thesurface are higher for SDS as compared to ABSand QO values calculated from BET expressionare higher for ABS as compared to SDS while itwas not possible to make the BET plots forCTAB.

Uptake of SDS and ABS has also been studiedin the presence of different concentrations ofNaCl, CaCI2 and AlCI3 in solution at fixed ad-sorbate and adsorbent concentrations. Mono aswell as divalent cations increase the adsorption ofSDS and ABS in the pH range 4-7 as shown inFigs 6(A and B). Trivalent salt like AlCI3 causesmuch increase in adsorption in this pH range. A25 to 40 per cent increase in adsorption with mo-no or divalent cations and 50 per cent increase inthe presence of Al3+ ions, is observed at the

•...~

(Al

"••~,.u"1.2J

'.1

Ceq/C. (CPI)

same pH. At higher salt concentration, adsorptionof these two compounds decreases.

The increased removal of SDS and ABS in thepresence of ionic salts .in solution, probably, can-not be accounted for by any single mechanism.Several probabilities seem consistent in terms ofinterpretation of these results such as:

(i) Possible lowering of cmc of detergents in thepresence of salts!". Micelles are known to have ahigher degree of surface activity.

(ii) Formation of smaller micelles below thenormalcmc.

(iii) Adsorption of ABS or SDS as Al-OH-ABS/SDS complex in acid solution.

A quantitative understanding of the sorptiondynamics and other mechanistic aspects help inproper optimization and modelling of the process.Kinetic data, as obtained by the Finite Bath meth-od has been treated by Eqs 2 and 3 which are va-

••sI _.oJ ••"':r -u

pH-U

•••••• ""t_n. _10•••• p.r zo ••••

(8)

u

••QeSOS

..•••

Cect!c. (SOS,AIS)

Fig. 5-B.E.T. plots of (A) cationic, (B) anionic detergents on copper ferrocyanidc (temp.=30±0.2°C; particle size-l70·200 mesh).

o pilaU (A)

~pH-7.0

Sail concentration (M)

(8)

sail concent_ (M)

Fig. 6-Plots showing the effect of salt solutions on the uptake of anionic (A) SDS, (B) ABS surfactants using copper ferro-cyanide (temp. e 30 ±0.2"C, particle size-l70-200 mesh).

SRIVASTAVA et al.: REMOVAL OF ANIONIC & CATIONIC DETERGENTS 347

lid for finite as well as infinite bath techniques,under the present experimental conditions em-ployed. This is in accordance to the observationsof Reichenberg." and have also been used in ear-lier studies 17 •

eo ( 2 2)6 1 - D, t Jl nF= 1--2 L 2: cxp 2

st n-I n ro... (2)

... (3)

Bt values as derived from Eq. (3), have been ob-tained from Reichenberg's Tablels. The linearitytest of Bt versus I plots are employed to distin-guish between the film and particle diffusion con-trolled rates of exchange 10.

The experimental conditions were set up forparticle diffusion as the sole rate determining stepto study the thermodynamics of the sorption pro-cess. The energy of activation E; has been deter-mined by applying the Arrhenius type expression(6)

where F is the fractional attainment of equilibri- o, = Do exp ( ~~a)urn at time t and is obtained by the expression

F= 0,00

... (6)

The pre-exponential constant Do (analogous to... (4) Arrhenius frequency factor) gives the entropy of

activation !:1S#where Of is amount taken up at time t and 00 isthe maximum equilibrium uptake and

2. Jl Dj •

B = --2 - = time constantro

. .. (5)

whereD, = effective diffusion coefficient of ions in the

adsorbent phase,ro = radius of the adsorbent particle assumed to

be spherical, andn = 1, 2, 3 ..... are the integers defining the infi-

nite series solution obtained for a fourier typeof analysis.

For every observed value of F, corresponding

Conc,(M)lIl1O'

Il.,'" 6.~

0.. ~.O

.," 4.0

2 [kt] (!:1S"i)Do=2.72d h exp R ... (7)

t.II.

0.7

o.siii

us

0.0

where d is average distance between the succes-sive exchange sites and is taken as 5 A (ref. 11).

The uptake of SDS, ABS, CAB and crAB oncopper ferrocyanide is particle diffusion con-trolled at and above 5.0 x 10-6 M for SDS andABS and 3 x 10-s M for CAB and CTAB asshown in Figs 7(A and B).

The increase in sorption rate of the four deter-gents with decreasing particle size of the adsor-bent material [Figs 8(A and B)] and increase intemperature [Figs 9(i)A; 9(i)B and 9(ii)A; 9(ii)B]further support the particle diffusion mechanism

(A) D.' (6)

0.1

0.5

0.15

0.5

0.••

0.0

Fig. 7-Plots of F vs t and Bt vs t for (A) anionic surfactants, (B) cationic surfactants, using different solution concentrationsat 30·C on copper ferrocyanide; [r= 9.4 x 10-3 em].

348 INDIAN J CHEM. SEe. A, MAY 1995

as the rate controlling process. An enhancedsorption rate at high temperature is probably dueto an increase in the mobility of sorbing species.It is evident that the concentration of anionic de-tergents necessary for particle-diffusion mechan-ism are less in comparison to cationic adsorbateson the adsorbent copper ferrocyanide. This is at-tributed to a higher rate of sorption of cationicdetergents on copper ferrocyanide as comparedto anionic detergents.

The non linear behaviour of log (I-F) versus tplots drawn for some typical sorption system[Figs lO(A and B)] is also an evidence in favour ofparticle diffusion mechanism as rate controllingstep. The initial portions of Mckay plots [FigslO(A and B)] are convex to log (I-F) axis whilethe later portion lies on a straight line in eachcase. It is inferred that the sorption of CP+ andCTA+ cation and SD- and AB - anions, mainlyinvolve two inter diffusion processes, a faster one

...

Fig. 8-Plots of F vs t and Bt vs t for (A) anionic surfactants (solution conc. =6.5 x 10-0 M), (B) cationic surfactants (solutioncone. = 7.5 x 10 - 5 M) using different particle sizes of copper ferrocyanide at 30°C: (1,2) r = 9.4 x iO -} em; (3,4) r= 5.13 x 10 - 3

ern.

(A) (6)

2.0 2.0

II_ 5O..1I1Ir'_-'e_37.3X..,..'MC-1

I.S 8_42.1..10"' __ ' U_2Q.7J:"'-'.cc-'

"- •• :n.,•.,.- •••..'CD 1.0 11-27.0""'~

iii 1.0 8-n; K1(f"i«<"'

8_16XI)-'tC'C-1

0.' 0.5

0.0 0.010 tChro)

Fig. 9(i)-Plots of F vs t and Br vs t for (A) SDS; 6.5 x 10-6 M, (B) ABS; 6.5 x 10-6 M at different temperatures using cop-per ferrocyanide (r= 9.4 x 10-3 ern),

'.0

t

(A)

Fig. 9(ii)-Plots of F vs t and Bt vs t for (A) CPB; 7.5 x 10-5 M, (B) CTAB; 7.5 x 10-5 M at different temperatures usingcopper ferrocyanide (r= 9.4 x 10-3 em).

SRIVASTAVA et al.: REMOVAL OF ANIONIC & CATIONIC DETERGENTS

corresponding to the residual curve and a slowerone corresponding to the latter portion of theplots lying on straight line, contributing to theoverall rate of exchange, As the system tends tow-ards equilibrium, the faster one becomes less sign-ificant and the slower one dominates. Srivastavaet a/17 have reported similar results for the sorp-tion of complex cations on chromium ferrocya-nide gel. Rawat et al.ls have also observed similarbehaviour for the exchange adsorption of metalions on different substrates.

It may be argued that the effective diffusion co-efficient, D, in these systems, is mainly comprisedof two components, which are due to the simul-taneous diffusion of the ingoing detergent ionsthrough the pores of different widths and differ-ent electric fields along the diffusion path. Thediffusion within the pores of wider-widths andweaker retarding forces of electrostatic interaction(particularly at the surface of the sorbent) ac-counts for the faster one and the one within thepores of narrower mesh widths and stronger re-tarding forces, accounts for the slower componentof Dj• ~s the temperature increases the contribu-tion of faster component of D. increases. It is be-cause of the increasing mobility of ingoing speciesat higher temperature, which to some extent over-comes the influence of retarding forces. The va-lues of effective diffusion coefficient D. are higher

349

for CPB and SDS as compared to CTAB andABS (Table 3). This is because of the geometryand bigger size of CTAB and ABS as comparedto that of CPB and SDS, causing, thereby, morehindrance to its mobility within the pores of sor-bent material. Further, the D, values are less foranionic detergents in comparison to cationic de-tergents which may be due to a low anion-ex-change value of copper ferrocyanide as comparedto its cation-exchange capacity.

The log D, versus lIT plots for the sorption ofCPB, CTAB, SDS and ABS on copper ferrocya-nide are linear in nature. This permits the use ofArrhenius equation for the determination of Doand E. from the intercepts and slopes of the line-ar plots. Do values are further used to calculatethe entropy of activation for the sorption of CPB,CTAB, SDS and ABS. The values of Dn, E" andI:lS" for all the four systems investigated are list-ed in Table 4.

The values of activation energy (E.) are higherfor CTAB and ABS as compared to CPB andSDS. This is due to the bigger size of CTAB andABS, which causes more steric hindrance to theirmobility through the pores of the sorbent materialand thus need more energy to undergo diffusion.The entropy of activation values are negative forall the systems and this indicates the lack oforientational effect upon the water environment

Table 3-Effective diffusion coefficient values for anionic and cationic detergents

D;(cm2s-')

Detergents 30° 35° 40° 45°CCPB 1.99 x 10-9 2.49 x 10-9 3.26 x 10-9 3.77 x 10-9

CTAB 1.57 x 10-9 1.91xlO-9 2.82 X 10-9 3.48 X 10-9

SDS 2.45 x 10-10 2.83 X 10-10 3.79 x 10- 10 4.55 X 10-10

ABS 1.43 x 10- J(I 2.06 X 10-10 2.65 x 10 -.n 3.33 x 10- III

...•

Fig. IO-McKay plots for the sorption of (A) crAB, (B) SDS on copper ferrocyanide.

350 INDIAN J CHEM. SEe. A, MAY 1995

Table 4-Do and thermodynamic parameters for the sorption of detergents

Parameters CPB CTAB SDSDo(cm2s-l) 3.18xlO-4 1.60 x 10-2 1.07 x 10-·

E.(kcalmol-') 7.17 9.69 7.81~S"(Caideg-'mol-') -9.54 -1.75 -11.70

ABS

3.02 x 10-3

10.10-5.07

o!

100

80!

60c2

"~~ 40

0.0 0.' 1.U t.lJ- ----r.Ir~ConClntr.tlon 0' dllOrbing reagent CK2SQr.)addeo

Fig. II-Desorption of ABS and SDS from copper ferrocya-nide detergent complexes.

of cations and anions and their faster diffusionthrough the sorbents. It may also be inferred thatthe uptake of detergent ions do not cause anystructural changes in the sorbent material.

Recovery of the adsorbed material as well asthe regeneration of the adsorbent is quite an im-portant aspect for an efficient scavenging process.As such desorption of adsorbate from anionic de-tergent loaded copper ferrocyanide was tried witha number of eluting agents. Data of these experi-ments have been shown in the form of plots ofpercent desorption of anionic detergents versusconcentration of eluting reagent added (Fig 11). Itis observed that SO~- ions are more effective ascompared to Cl" ions. An almost complete de-sorption of both SDS and ABS from copper fer-rocyanide detergent complex, could be achievedwith the help of two desorbing reagents K2S0;1

and a mixture of 1 M H2S04 and alcohol (1 : 1,v/v) (only the plots with K2S04 are shown in fi-gure). Both these eluting agents are equally effec-tive and the adsorbent copper ferrocyanide canbe regenerated for further use.

AcknowledgementOne of the authors, Dinesh Mohan is highly

thankful to the CSIR, New Delhi for providingthe financial assistance.

ReferencesI Zhu B-Y & Gu T, J chem Sac, Faraday Trans I, 85 (1989)

3813.2 Carberry J B & Geyer A T, J Water Polin Cant Fed, 49

(1977)452.3 Douillard J M, Pougnet S, Faucompre B & Partyka S, J

colloid interface Sci, 154 (1992) 113.4 Narkis N & David B B, Waler Res, 19 (1985) 815.5 Harwell J H, Roberts B L & Scamehorn S F, Colloids

Surf, 32 (1988) 1.6 Zhu. B-Y, Gu T & Zhao X, J chem Sac, Faraday Trans,

85 (1989) 3819.7 Hukao K & Takeda Y, Carbon, 9 (1991) 173.8 Douillard J M, Pougent S, Faucompre B & Portyka S, J

colloid interface Sci, 154 (1992) 113.9 Luisa Gonzaler Martin M & Rochester C· H, J chem Sac,

Faraday Trans i, 88 (1992) 973.10 Kourim V, Rais J & Million B, J inorg nucl Chern, 26

(1964) llll.11 Helfferich F, Ion exchange (McGraw Hill, New York)

1962.12 Lawrence K W, JAm Waler Works Assoc, 67 (1925) 19.13 Few A V & Ottewill R H, J Call Sci; 11 (1958) 34.14 Bezan M, Malavasia M & Vesnarver G, J chern Sac Fara-

day Trans, 89 (1993) 2445.15 Reichenberg D,J Am chem Sac, 75 (1953) 589.16 Boyd G E, Adamson A W & Myers L S, J Am chem Sac,

69 (1947) 2836.17 Srivastava S K, Saraswat 1 P & Sharma A K, Can J

Chern, 57 (1979) 1214.18 Rawat J P & Thind P S, J phys Chern, 80 (1976) 1384.