J. Phys. Chem. 1980, 84, 365-369
(23) (24) (25) (26) (27)
(29) I. H. Kolthoff, "Treatise on Analytical Chemistry", Interscience, New York, 1959. (30) C. A. Bunton, N. Camsco, S. K. Huang, C. H. Paik, and L. S.Romsted, J. Am. Chem. Sm., 100, 5420 (1978), and references cited therein. (31) M. Almgren, F. Grieser, and J. K. Thomas, J. Am. Chem. Soc.,101, 279 (1979). (32) For an "ideal" buffer, the limit of [OH,], the analytical concentration of "bound" hydroxide ion at "infinite" CT, is given bya
neutral cyanohydrin intermediate into the micellar phase, following ratedeterminina' OH- attack, rather than a variation of OH, itself. J. M. Corkill, J. p. Goodman, and T. Walker, Trans. Faraday SOC., 63,768 (1967). L. R. Romsled, PhD. Thesis, Indiana University, Bloomington, IN, 1975. N. Funasaki, J . Colloid Interface Sci., 64, 461 (1978). D. H. Smith, J . Colloid Interface Sci., 68, 70 (1979). Our experimental values of the cmc of CTAB in water (9.4 X lo-' M) and 0.020 M borate buffer pH 10.10 (5.0 X lO-'M) may be used to estimate K, on the basis of a charpd-phase separation treatment of micellization.28 Thus, for the equilibrium: mm,
+ m(1 - a' - P)[Br,'] + mP[B,]
KOH/Br(l-
clsLoHb] =
-I-P(log [Br;]
a) [OH,]
1'
it can be readily shown that
- (1 - a') log [Br,']
a
Thus, with, ,K , = 0.08 and a = 0.2, an intermicellar pH of 11.5 ([OH, = 3.2 X M at 25 OC) corresponds to a limiting value of M, assuming that the buffer system is capable [OH, = 1.0 X of maintaining the intermicellar pH at 11.5. I n this regard, it is interesting to note that the CTAB micelle is in certain respects a weakly acidic species; however, unlike a conventional weak acid, its "acidky" (binding of OH') varies with the pH and the ionic content of the intermicellar aqueous phase. (33) Evidence for possible additional kinetic complications arising from the interaction of bnlc buffer species with the micellar phase of CTAB has been reported by Funasaki." (34) An interesting case in point is SDS in the presence of sodium phosphate buffer at alkallne pH, where the influence of added buffer is treatable as an added commn salt effect (M. PolRi, I.M.Cuccovla, F. H. Quina, and H. Chaimovich, unpublished results). While this manuscript was in preparation, Funasaki*' reported that, among several systems, a buffer system compatible with our criteria [2amim2methyC1,3-propanedii (AmptHBr] appeared to be the nwst suitable for studies of alkaline hydrolysis In CTAB. In fact, the data of Funasaki2' for the alkaline hydrolysis of NPA in the presence of 0.01 M Amp-HBr buffer are quite nicely reproduced by eq 3-7 (including the additional terms for the added common salt In eq 5 and 6") using the same parameter values empbyed In our calculation (Le., Ks = 54 W', kI:, = 0.37, a = 0.20, V = 0.37 Llmol, and KOHp= 0.08) and his value of k:[OH,] = 0.049 min-' at pH 9.59 (25 C).
F=
micelle (separate pseudophase)
log m , = C ,
365
- log [B,])
where C, is nominally equal to -(log K,)IN. At the cmc in the cmc; thus log cmc = presence of borate buffer, m , [BrF] C, -(1 - a') log cmc &log cmc - log [B,]) and
+
Since the literature values" of the constants C, = -4.84 and (1 - a')= 0.60 agree with our cmc in pure water and [By] can be calculated from eq 13, these equations permit one to estimate that P 0.28 and KBI, 0.025 at the cmc. Obviously, the above equations are at best approximate since C , will certainly vary as 6 N (1 - a');nonetheless, they cleariy indicate that an unfavorable selectivity coefficient for the added ionic species implies a small effect on the cmc and vice versa. (28) T. Sasaki, M. Hattori, J. Sasaki, and K. Nukina, Bull. Chem. SOC. Jpn., 48, 1397 (1975), and references cited therein.
Miscibility of Fluorocarbon and Hydrocarbon Surfactants in Micelles and Liquid Mixtures. Basic Studies of Oil Repellent and Fire Extinguishing Agents KBzO Shinoda" and Toshio Nomure Department of Applied Chemistry, Faculty of Engineering, Yokohama National University, Hodogaya-ku, Yokohama 240, Japan (Recelved May 16, 1979; Revised Manuscript Received October 11, 1979) Publication costs assisted by Yokohama National University
The mutual solubility of C,F2,+lCOOH (n = 7-12) with C,H2,+lCOOH (rn = 7-17) and of C,F2,+1CH2CH20H (n = 8-10) with C,Hh+lOH (rn = 11-18) has been studied. From the solubility and critical solution temperatures, it is concluded that a carbon chain longer than 8 carbons is necessary to cause the phase separation in liquid-liquid mixtures of these fluorocarbon and hydrocarbon surface-activesubstances. Theoretical equations for the partial and total critical micelle concentrations (cmc) of surfactant mixtures have been derived in the case when the energy of mixing between surfactants has to be taken into account. Miscibilities of fluorocarbon and hydrocarbon surfactants at the interface or in micelles have been examined by the comparison of the theoretical and the cmc values of surfactant mixtures. It is concluded that C,FI5COONa will mix with CloH2,S04Nain all compositions but that C8F17COO"4 will only partially mix with C12H25S04NH4 at 25 "C. The importance of the immiscibility of fluorocarbon and hydrocarbon surfactants for oil repellent and fire extinguishing properties is discussed.
Introduction The critical solution temperature of heptane and perfluoroheptane is 50 "C, above which two liquids mix over all compositions.' If one end of the respective molecules is converted into the same group, such as COOH or OH, both components will readily mix a t 25 "C because the enthalpy of mixing is considerably depressed. Actually, a 1:3 volume mixture of C7F15COOHand C7H15COOHand a n equivolume mixture of C8FI7COOHand CllH2J!OOH 0022-3654l8Ol2084-0365$0 1.OO/O
are one liquid phase above 19 and 47 OC, respectively, below which crystals of respective perfluoroalkanoic acids precipitate. This direct observation implies the miscibility of two surfactants, C7F15COONaand C7H15COONa,in micelles, because a micelle is a pseudophase in a liquid state. When the chain lengths of both surfactants are sufficiently long, however, two types of micelles, one rich in fluorocarbon surfactant and the other rich in hydrocarbon surfactant, may be formed. If this happens, two 0 1980 American Chemical Society
366
The Journal of Physical Chemistry, Vol. 84, No. 4, 1980
Shinoda and Nomura
types of micelles, i.e., two liquid phases, are in equilibrium, so that the chemical potentials of both species in micelles or monolayers of different compositions in equilibrium are equal, respectively. This means that an aqueous surfactant solution containing long hydrocarbon and fluorocarbon surfactants which will spread over an air-oil interface may have superior surface properties, because a fluorinated surfactant whose surface tension is small may preferentially adsorb at an air-water interface, whereas a hydrocarbon surfactant will preferentially adsorb at an oil-water interface. Hence, the fluorocarbon chain length of surfactant necessary to form two types of micelles is an important and interesting problem to study.
O N
0
E
YE
0
+
0 -
0
E
01
I
\
0
E
0
cl
Theory Miscibility or immiscibility of two types of surfactants in micelles is a phenomenon similar to the mixing of a binary solution. If the energy of mixing per molecule exceeds 2kT, two liquids will only partially mix. But we cannot visually observe phase separation of two surfactants in micellar solution. Whether two types of surfactants are miscible or not, however, may be revealed from the cmc values of surfactant mixtures, because the cmc in solution corresponds to the vapor pressure of a binary liquid mixture in the gas phase. Just like the relation between vapor pressures and the composition of a binary solution, the partial concentrations of singly dispersed surfactants vs. the composition in mixed micelles may be expressed statistical thermodynamically2 as cmcl = cmcloxl e x p ( B ' ~ ~ 4 , ~ / R T )
(1)
cmc2 = cmc2x2 exp(B'u2412/RT)
(2)
where cmco is the cmc of pure surfactant, x2 the mole fraction of the second component (surfactant) in the micelle, 4 the volume fraction, u the molal volume, and B' the energy of mixing of two surfactants per cubic centimeter. If we assume u1 = u2, we obtain cmcl = cmc:xl
e x p ( ~ ~ ~ x ~ ~ / k T ) (3)
cmc2 = cmc20x2exp(w12x12/kT)
(4)
where u12is the interchange energy per molecule. Equation 3 and 4 are identical with equations of partial pressures in strictly regular solutiona3Since u12 = 2kT is the condition for critical mixing, the partial and total concentrations relative to pure components of singly dispersed surfactant as a function of mole fraction in the micelle were calculated and are shown by dashed curves in Figure 1. In binary solution, the amount of respective components in the vapor phase is very small and the vapor is in equilibrium with a large amount of solution. In surfactant solution, however, the cmc is a concentration at which micelles are formed. At this point, a few micelles (solution) are in equilibrium with the relatively large amount of singly dispersed surfactants, Le., the experimental cmc curves have to be plotted against the mole fraction of singly dispersed surfactants, cmcz/(cmcl + crnc,). The solid curve in Figure 1is the cmc (total saturation concentration) vs. the composition of singly dispersed surfactant at the cmc calculated from dashed curves. If the energy of mixing is more than 2kT per molecule, two types of micelles will be formed. The maximum cmc value is higher, and the shape of the dashed curve is different but that of the solid curve is similar. When we are dealing with the ionic surfactant, the effect of the concentration of gegenions on the cmc has to be taken into account. The cmc is inversely
O N
\
0
\
E 0 k 0 E
\ \
\ \ \
0
\ \
0 -
0
E
0
\
0
E
0
- _ _ _CmC v s = CmC v s .
X, in micelle X,of dlspersd species
Flgure 1. The cmc of surfactant mixture as functions of mole fraction in micelle (dashed curve) and mole fraction of singly dispersed surfactant (solid curve) assuming w l P= 2kT, and cmc = 1. The mole fractlon of singly dispersed surfactant is defined as x 2 = cmc2/(cmc, cmc,). The chain curve Is that of ionic surfactant assumlng K, = 0.8 and a l p = 2kT.
+
proportional to the K,th power of the concentration of gegenions4 mu In cmc = -- - K, In C, + const (5) kT where m is the number of carbon atoms in the chain, w the energy to transfer a CF2 (or CH2) group from the micelle to the bulk of solution, C, the concentration of gegenions in mol/L, and Kgan experimental constant obtained from the linear relation between log cmc and log C,. If no salt is added, C, = cmc, and we obtain from eq 54 mu (1 K,) In cmco = --k T + const (6)
+
The partial cmc of respective surfactants in the mixed micelle may be derived from eq 3, 4, and 5 . mlwl In cmcl = -- K,, In C, kT
w12x2 + In x1 + + constl kT
(7) mzw2 In cmcz = -- Kg2In C, kT
12xl2 + In x 2 + W+ constz kT
(8)
Substituting eq 6 into 7 and 8 and assuming K,, = K,, = Kg
+
In cmcl = (1 K,) In cmclo - K, In C,
+ In x1 + W12X22 kT (9)
+
In cmcz = (1 K,) In cmcZ0- K, In C,
+ In x 2 + W12XI2 kT (10)
Miscibility of Fluorocarbon and Hydrocarbon Surfactants
The Journal of Physical Chemistry, Vol. 84,
1980
TABLE I: Miscibility of Fluorinated and Paraffin Chain Compounds
or W12XZ2
cmcl CgKg= cmclO(l+Kg)xl expcmcz CgKg= cmc20(1+Kg)x2 exp-
kT
compound 1
W12Xt2
kT Since C, = cmcl + cmcz in the case when no salt is added, we obtain (cmcl
No. 4,
+ cmc2)1+Kg= crnclO(l+Kg)xlexp-
w12xz2
kT
+
or w12xz2
cmclo(l+Kg)xlexpL
kT
+
C,F, ,COOH
compound 2
T,,“Cb
C,H,,COOH
19
Tc, “Cc
(1:3v/v) C,H,,COOH 38 C,,H,,COOH 45 C.F, .COOH C,H,.COOH 42.5 C; ,Hi,COOH 47 C,,H,,COOH 49 C, F ,,C OOH C,,F,,COOH C, ,H,,COOH 76 C, ,H,,COOH 83 C,,H,,COOH 100 c,,Fz P O O H C, ,H3 ,COOH 114 30.5 C,F, ,CH,H,OH C12H250H C, ,F,,CHzCH,OH C,,H,,OH 77 83 1‘ ‘lHZSoH 95.2 C113H370H (gola 81 C,,F, ,CH,CH,OH C, ,H,,COOH 101 C,,H,,COOH
367
-
nfd
>7.5 >7.4
>8.3 >8.4 >8.2
>9.4 >9.0
>7.9 >8.4 >8.2
=7.9 >8.3 =7.8
Two liquid phases were observed in these systems. T , means a temperature above which solid melts and one liquid phase was obtained for equivolume mixture. T, n’ is the calculated is the critical solution temperature. number of carbon atoms in fluorocarbon chain to cause phase separation a t 25 “C. a
The cmc curve of ionic surfactant vs. the composition of singly dispersed surfactant was calculated assuming K , = 0.6 and plotted by chain curve in Figure 1. Cmc values when w12 = 3kT and w12 = WkT at x1 = 0.5 are also indicated.
Experimental Section Materials. Perfluoroalkanecarboxylic acids, C6F13COOH, C8F17COOH,CloF21COOH,and C12F2&OOH,were furnished by the Asahi Glass C O . ~Perfluorooctanoic acid is the same material used in the preceding article^.^^^ Perfluorodecanoic acid was supplied by Dr. R. A. Guenthner of 3M. Fluorinated alcohols, C8Fl7CH2CH,OH and CloF21CH2CH20H,were also donated by the Asahi Glass Co. Fatty acids were extra pure grade materials donated by Nihon Yushi C O . ~ The respective salts were prepared by neutralizing the acids with aqueous alkaline hydroxide, ammonia, and amine solutions, respectively. Ammonium dodecyl sulfate was prepared by double decomposition of silver dodecyl sulfate and an exact amount of ammonium chloride. Silver dodecyl sulfate was prepared by double decomposition of sodium dodecyl sulfate with silver nitrate in water.g Procedures. Consolute temperatures were observed by repeated heating and cooling while shaking the ampules containing known amounts of fluorinated and paraffin chain alcohols (or acids). Cmc was determined from the electrical conductivity-concentration curve at constant temperature. Results and Discussion 1. Liquid-Liquid Miscibility of Paraffin Chain and Fluorocarbon Chain Compounds. Since a direct observation of the miscibility of two types of surface-active materials is straightforward and reliable, liquid-liquid miscibility in mixtures of perfluoroalkanoic acid or perfluoroalkyl ethanol (C,Fk+lCH2CH20H) with alkanoic acid or alkanol was studied and is summarized in Table I. T, means an upper critical solution temperature, and T, means the temperature below which solid compound separates from one liquid phase, Le., only one liquid phase exists above these temperatures. In most systems studied separation into two liquid phases was not observed. Critical solution temperatures were observed in CloF21CH2CH20H-C18H370Hand C ~ O F ~ ~ C H ~ C H ~ O H - C , H ~ , +(~nC=O13, O H15, 17) mix-
tures. Critical solution temperatures of fluorinated alcohols and paraffin chain acids were gradually depressed with time, probably because of a reaction. If we assume that the energy of mixing of a fluorocarbon chain with a hydrocarbon chain per -CF, corresponds to two -CF2groups, considering the surface area of these groups we can roughly estimate the fluorocarbon chain length, n’, necessary to cause separation into two liquid phases at 25 “C from T,, 298 ( n + 1)- 1 = n’ TAK)
where n is the number of carbon atoms in the fluorocarbon compound studied. In the case when T , is not observed, we can only estimate the fluorocarbon chain length, which is not long enough to cause phase separation by T,. It is evident from Table I that a fluorocarbon chain length of a t least 8 is necessary to cause the separation into two liquid phases at 25 “C. Since perfluoroalkanoic acid is a strong acid and fatty acid is a weak acid, mutual solubility would have been enhanced. 2. Mutual Solubility of Fluorocarbon and Hydrocarbon Surfactants in Micelles. Mukerjee and Yang’O have measured the cmc of mixtures of sodium perfluorooctanoate (C7F15COONa)with sodium decyl sulfate at 25 “C. They said, “The experimental cmc data are not too far from the calculated case of no mixing of micelles (in which infinitely large energy of mixing is assumed), indicating that the nonideality of mixing is indeed severe and suggest the coexistence of two kinds of micelles in this mixture (C7F15COONaand C1oHz1SO4Na)at 25 “C. The cmc data by themselves do not provide conclusive evidence on this point.” From the measurements of the differential conductance plotted against the mean concentration of surfactants, they concluded, “Two kinds of micelles exist together in aqueous solution of mixture of C7F15COONa with Cl2HZ5SO4Na.” The theoretical cmc curve of the mixture of these surfactants was calculated by substituting w12 = 1.8kT and K , = 0.59 (average of Kg = 0.645 for CloHz1SO4Naand K = 0.53 for C7F15COONa)in eq 13 or 14 as shown by solid curve in Figure 2. The maximum cmc value when w12 =
The Journal of Physical Chemistry, Vol. 84, No. 4, 1980
368
I,
Shinoda and Nomura
8 K T , Kg = 0 59
I
I
I
T w o Kinds of Coexisting
Micelles Region
I I
-----I
I
1
I
1
I
02
04
06
08
10
CBFi7C OONHe
"
0
02
04
06
00
IO
C ioHniSO4Na
C~FISCOON~
Figure 2. The comparison of the experimental and calculated cmc values in the mixture of C7F,5COONa and Ci0H2,SO4Na at 25 "C. Experimental data are obtained from ref 10. The cmc curve was calculated by substituting w12 = 1.8kTand K, = 0.59 in eq 14.
1.7kT is also indicated. From the good agreement between the observed cmc and the calculated cmc assuming w12 = 1.8-1.7kT, we conclude that these surfactants mix in micelles over all compositions at 25 "C. It is clear from Figure 1 that the difference between the maximum cmc value calculated by assuming w12 = kT and that of complete mixing (w12 I2 k T ) is fairly large. Therefore, we have to compare the data with the theoretical curve of critical mixing. Since the area per molecule at the air-solution interface does not differ much for C7Fl,COOK, CllH25COOK, and C12H25S04Na,6 the effect of the difference in cross-sectional areas to the energy of critical mixing is small. Since we are interested in knowing the necessary chain length to cause the "phase separation" in pseudophase (micelle), we have studied the cmc of mixtures of longer chain compounds, i.e., C8F17COONH4and C12Hz5S04NH4. The reason we have selected ammonium salts is that the Krafft points of these surfactants are well below 25 "C and that the cmc of mixture and the effect of the concentration of gegenions are observable at 25 0C.5 The results are plotted in Figure 3. The change of the cmc of these surfactants and their mixture with the total concentration of gegenions has been determined by electrical conductivity measurements and is given by eq 16. The K , values obtained are listed in log cmc = -Kg log C,
+ const
(16)
Table 11. The Kgvalue was nearly constant over all mole fraction ranges. Theoretical cmc values of the mixtures were calculated by substituting w12 = 2.2kT, Kg= 0.72, and the respective cmc values of surfactants in eq 14, as shown
CizHzt.S04"4
Figure 3. The comparison of the experimental and calculated cmc values in the mixture of C8F,,COONH4 and C12H25S04NH4 at 25 OC. The cmc curve was calculated by substituting wi2 = 2.2kTand K, = 0.72 in eq 14.
TABLE 11: Cmc with No Added Salts and the Experimental Slope, K,, of t h e Logarithm of the Cmc vs. the Logarithm of the Concentration of Univalent Gegenions (",NO,) in the Equation, log Cmc = - K , log Ce -1 const a t 2 5 "C surfactants C,F, ,COONH, C,F,,COONH,
+ C,,H,,SO,NH,
cmc, mmol
K,
9.1 9.8
0.71 0.71
10.9
0.72
7.9
0.73
7.2
0.73
(x, = 0.146)
C,F,,COONH, t C,,H,,SO,NH, (x, = 0 . 4 5 ) C,F,,COONH, + C,,H,,SO,NH, (x, = 0.85) Cl,H,,SO,",
by the solid curve in Figure 3. The maximum cmc value when w12 = 2.lkT is also indicated. It is clear from Figure 3 that the observed maximum cmc value closely agrees with the calculated cmc, assuming w12 = 2.2-2.lkT. The fact implies the "pseudo phase separation" in micelles. Deviations of the observed cmc values from the calculated curve are larger at the medium composition between that of maximum cmc and x = 0 or 1. Since a few percent of surfactant is in the micellar form at the cmc, the observed cmc should deviate from the solid curve toward the dotted curve. The estimated energies of mixing w12 = 1.8--1.7kT for the C7Flj compound and w12 = 2.2-2.lkT for the C8F17 compound mutually agree rather nicely. From the NMR absorption of fluorine atoms (CF,), we have obtained the other independent evidence that C7F16COOLimixes with C12H25S04Liover all mole fraction ranges, but c8F17coOLi, CgFl9C0OLi,and ClJ?21COOLido not mix completely with CI2Hz5SO4Li. 3. T h e Importance of Immiscibility of Fluorocarbon and Hydrocarbon Surfactants as Oil Repellent and Fire Extinguishing Agents. The importance of the immiscibility of two surfactants is the coexistence of two kinds of
J. Phys. Chem. 1980, 84, 369-376
air
f
,
i uo r o cy;;;
water
'hydrocarbon oi i
chain
Figure 4. Illustration of the coexistence of two mixed monolayers of different compositions at the air-water and at the ail-water interfaces when fluorocarbon and hydrocarbon surfactants are only partially miscible.
monolayers of different compositions. Aqueous solutions (or foams) containing such mixtures of fluorocarbon and hydrocarbon surfactants should exhibit superior properties spreading over the oil surface, because the air-water surface will be preferentially covered with the fluorocarbon surfactant to depress the surface tension effectively (15-20 dyn/cm) whereas the hydrocarbon surfactant will preferentially adsorb a t the oil-water interface to depress the interfacial tension (0-2 dyn/cm), as illustrated in Figure 4.
369
The longer the chain lengths of fluorocarbon and hydrocarbon surfactants, the more distinct the preferential adsorption. Since the Krafft point (melting point of hydrated solid surfactant) is raised with the chain length just as the melting point of paraffi chain compounds," various devices to depress the Krafft point will be necessary for practical applications.12 Acknowledgment. The financial support of the Asahi Glass Foundation for Industrial Technology is gratefully acknowledged. We thank Mr. M. Hanlin for some miscibility measurements, and Asahi Glass Co. and 3M Co. for fluorochemicals.
References and Notes (1) J. H. Hikiebrand, B. 8. Fisher, and H. A. Benesi, J . Am. Chem. SOC., 72, 7348 (1949). (2) K. Shinoda, "Principles of Solution and Solubility", Marcel Dekker, New York, 1978; concept in Chapter 9 and eq (4.10), (4.11), and (11.1). (3) R. Fowler and E. A. Guggenheim, "Statistical Thermodynamics", Cambridge University Press, London and New York, 1956, pp 357-358,432. (4) K. Shinoda, T. Nakagawa, B. Tamamushi, and T. Isemura, "Colloklal Surfactants", Academic Press, New York, 1963, pp 41-42. (5) H. Kunieda and K. Shinoda, J . Phys. Chem., 80, 2468 (1976). (6) K. Shinoda and H. Nakayama, J. Colloid Sci., 18, 705 (1963). (7) K. Shinoda and K. Katsura, J. Phys. Chem., 68, 1568 (1964). (8) K. Shinoda, J. Phys. Chem., 58, 1136 (1954). (9) P. Mukerjee and K. J. Mysels, J . Phys. Chem., 82, 1400 (1958). (10) P. Mukerjee and A. Y. S. Yang, J. Phys. Chem., 80, 1388 (1976). (11) Reference 4, pp 5-8. (12) K. Shinoda "Solution and Solubility", Maruzen, Tokyo, 1974, p 191 (in Japanese).
Adsorption of Weak Organic Electrolytes from Aqueous Solution on Activated Carbon. Effect of pH G. Muller,+ C. J. Radke,' and J. M. Prausnitr Chemical Engineering Department, University of California, Berkeley, California 94720 (Received August 6, 1979)
A theoretical model is presented to predict the effect of pH on adsorption equilibria of weak organic acids or bases from dilute aqueous solution on activated carbon. The solid surface exhibits heterogeneity which is characterized by the model of patches and an exponential adsorption energy distribution function. Adsorption on each patch is described by a local competitive Langmuir isotherm with identical nonelectrostatic parameters for the undissociated and dissociated solute. The solid surface charges in response to solution pH and ionic strength; the resulting (smeared) surface electrostatic potential influences the adsorption affinity of the ionized solute. The response of the surface to solution pH is obtained from surface titration measurements using strong acids and bases. For certain combinations of solute pK, solute concentration, and surface charge, the model predicts an adsorption maximum with hydrogen ion concentration. When the surface is homogeneous and surface coverage is low, the adsorption change with pH parallels the solute dissociation curve. New experimental adsorption isotherms are reported for benzoic acid and for p-nitrophenol on activated carbon at varying pH. Quantitative agreement is found between predicted adsorption equilibria and experiments over a wide range of pH.
Introduction Adsorption on activated carbon is the basis of an effective and widely used process for removing organic pollutants from municipal and industrial wastewaters. Therefore, the literature provides several reports concerning adsorption kinetics and equilibria for both singleand multisolute aqueous systems a t neutral pH.1-3 By comparison, few studies are available for adsorption at pH On leave from Institut fur Thermische Verfahrenstechnik, Universitat Karlsruhe, 75 Karlsruhe, W. Germany. 0022-3654/80/2084-0369$01 .OO/O
values removed from 7.0 in spite of the qualitative observations of early workers who noted that pH has a dramatic effect on adsorption for ionizing Early quantitative measurements of the pH effect were given by Phelps and Peters who showed a pronounced decrease in organic acid adsorption with increasing ioniati ion.^,^ The adsorption decrease closely followed the dissociation curve, leading Phelps and Peters to postulate that only the molecular form of the solute adsorbs. Phelps and Peters later found similar results for organic bases, but they indicate deviations when the activated carbon is 0 1980 American Chemical Society