Ionic

8790

Langmuir 2006, 22, 8790-8799

Water Behavior Revisited at the Air/Ionic Solution and Silica/Ionic Solution Interfaces. Extension to Coadsorption of Ions and Organic Molecules Abdelhaq Acharid, Mustapha Sadiki, Galal Elmanfe, Nawal Derkaoui, Rene´ Olier, and Mireille Privat* De´ partement de Chimie, UMR CNRS 6521, UniVersite´ de Bretagne Occidentale, 6 AV. Le Gorgeu, C.S. 93837, 29238 Brest Cedex 3, France ReceiVed April 28, 2006. In Final Form: July 7, 2006 This paper reports on investigations about the adsorption at the air-water surface, and for the sake of comparison at the silica-solution interface, of two 1-2 electrolytes, Pb(NO3)2 and PbCl2, at first alone and then from a mixture with carbofuran or with benzene; all of them were at concentrations below 10-2 M. The limited domain, where the Debye and Hu¨ckel formalism for solutions and the Wagner-Onsager-Samaras (WOS) model for surfaces are correct, is then respected. This study was aimed at trying to identify the part played in the surface by the different particles of the system components and in particular the role of water. When aqueous solutions of nonorganic salts are dilute enough, their surface tensions are known to be salt concentration-independent; however, the zero value of the resulting relative adsorption has never been the subject of analysis about water behavior. By combining experimental relative adsorptions and Gibbs excesses calculated from the WOS theory, we will show that, in well-known solutions such as KCl ones, where the negative excess in salt can be very precisely modeled by the WOS theory, the resulting water excess ΓW is negative. The same result can be obtained by taking into account the Ray-Jones effect. This observation drove us to wonder about the results of a similar analysis done on solutions of unsymmetrical electrolytes and on mixtures of salt and organic molecules. Experiments showed that, for all of the systems, ΓW was negative. For a given salt, ΓW was more negative in the presence of organic molecules, and carbofuran was a more efficient water repellent than benzene; water repulsion was greater with nitrates than with chlorides. From these data, it seems that water was repelled toward the solution bulk, whereas ions probably took place between the bulk and a layer of organic molecules. These observations called for a more detailed modeling.

1. Introduction Electrolyte adsorption at the water interface is involved in many crucial phenomena. In the current literature, three of them have brought a new surge of interest. Biological mechanisms either in vivo or implied in devices such as sensors are often driven by electrochemical phenomena where liquid interfaces are implicated.1-3 The knowledge of ionic organization at aerosols surface has thus become topical.4,5 The role of counterions in ionic surfactants-containing interfaces has been taken into account. Experiments as well as theoretical approaches have deeply changed the views about ion-containing surfaces. In most of these systems, the simultaneous presence of ions and organic molecules affects the adsorption behavior and may lead to a rich chemistry within the microsystems formed from surface breaking.6,7 These considerations explain the present development of studies about surfaces in such mixtures.8 This paper reports on investigations about the adsorption at the air-water surface of two 1-2 electrolytes, Pb(NO3)2 (partially studied in ref 9) and PbCl2, at first alone and then in mixture with * Corresponding author. E-mail: [email protected]. (1) Suci, P. A.; Klem, M. T.; Douglas, T.; Young, M. Langmuir 2005, 21, 8686. (2) Pasche, S.; Vo¨ro¨s, J.; Griesser, H. J.; Spencer, N. D.; Textor, M. J. Phys. Chem. B 2005, 109, 17545. (3) Lund, M.; A° chesson, T.; Jo¨nsson, B. Langmuir 2005, 21, 8385. (4) Garrett, B. C. Science 2004, 303, 1146. (5) Jungwirth, P.; Tobias, D. J. J. Phys. Chem. B 2005, 105, 10468. (6) Vione, D.; Maurino, V.; Minero, C.; Lucchiari, M.; Pelizzetti, E. Chemosphere 2004, 56, 1049. (7) Vione, D.; Maurino, V.; Minero, C.; Pelizzetti, E. EnViron. Sci. Technol. 2005, 39, 1101. (8) Karraker, K. A.; Radke, C. J. AdV. Colloid Interface Sci. 2002, 96, 231.

carbofuran or 2,2-dimethyl-2,3dihydro-benzofuran-7-yl methylcarbamate (Figure 1). The experiments and analyses discussed here were conducted with the hope of identifying the part played in the surface by the different particles of the system components, in particular the role of water. As water is responsible for numerous oriented interactions, its behavior should provide a lot of information about specific spatial arrangements, but till now adsorption data have been the subject of very little analysis of this kind. The solutions used in this study were all at concentrations below 10-2 M, i.e., within the domain limits where the Debye and Hu¨ckel formalism for solutions and the Wagner-OnsagerSamaras (WOS) model for surfaces are correct. We compare our experimental results first with previous data on mixtures, in water, of Pb(NO3)2 and benzene at the air/solution interface10 and second with preliminary data obtained at the silica/solution interface with carbofuran-Pb(NO3)2 mixtures. The experimental adsorption data discussed here were obtained by application of the Gibbs isotherm formula to surface tension values for investigations at the air/solution interface and by a depletion method for those at the silica/solution surface; it is worth underlining that they are clearly relative adsorptions of salts or organics with respect to water. To get information about individual components, the knowledge of their Gibbs excesses is of high interest but requires the use of a model to calculate, at least, one of them. Ionic theories in an easy-to-use analytical (9) Acharid, A.; Quentel, F.; Elle´ouet, C.; Olier, R.; Privat, M. Chemosphere 2006, 62, 989-997. (10) Sadiki, M.; Quentel, F.; Elle´ouet, C.; Huruguen, J. P.; Jestin, J.; Andrieux, D.; Olier, R.; Privat, M. Atm. EnViron. 2003, 37, 3551.

10.1021/la061160c CCC: $33.50 © 2006 American Chemical Society Published on Web 09/07/2006

Water BehaVior ReVisited

Langmuir, Vol. 22, No. 21, 2006 8791

Figure 1. Developed formula of carbofuran.

form are rather scarce. In fact, since the historical treatment by Wagner,11 popularized and extended by Onsager and Samaras,12 no decisive progress has been made in this domain despite several improvements.13-15 Till now, the relative layout of cations and anions in the surface layer has been gained through molecular dynamic simulations.5,15-17 They have shown that, according to its polarizability, the anion may be very displaced toward the air, contrary to most cations. This result is important because of the observation of negatively charged aerosols;16 moreover, such results have been confirmed by second-harmonic generation experiments.18 However, they have usually been conducted in more concentrated solutions than those studied here. In this report, we will also examine the Jones-Ray effect19,20 occurring at air/ solution interfaces at about 10-3 M. This effect will be discussed with respect to the WOS model in the light of our experimental data and Jones and Ray’s ones. Our first concern, here, will not be directly the layout of ions but rather the very specific behavior of water. Though it has been known for ages that, on condition that aqueous solutions of nonorganic salts be dilute enough, their surface tension is unaffected by the salt concentration, to our knowledge, the zero value of the resulting relative adsorption has never been the subject of analysis about water behavior. We will show later that the resulting water excess is negative, even in well-known solutions such as KCl ones, where the salt excess can be very precisely modeled by the WOS theory. The same result can be obtained by taking into account the RayJones effect. This observation drove us to wonder about the results of a similar analysis done on mixtures of salt and organic molecules. We, thus, continued to use the WOS theory, adapted to such systems despite its serious deficiencies for unsymmetrical electrolytes. As seen later, the order of magnitude of concern allows this kind of approximation. This report is organized as follows. The foundations and limits of the WOS theory are recalled in the first paragraph of section 2, where a simplified adaptation of WOS theory is also proposed for the mixtures under study; the second paragraph details the use of the Gibbs model and describes how water excesses are got through association of the WOS theory with experimental data. Section 3 presents the experimental values and the results of calculations. At last, we discuss how the water behavior is affected by the respective nature of anions and organics at the two case-studies of interfaces and the possible ionic arrangements resulting in the surface.

development relies on a previous paper by Wagner,11 is an adaptation of the Debye-Hu¨ckel theory of electrolyte to the interface description. This theory21,22 is itself based on a PoissonBoltzmann formalism. Electrolytic solutions are characterized by long-range, Coulombic interactions, which are repulsive between ions of similar sign and attractive between those of opposite signs. Because of the ionic mobility and Coulombic interactions, each ion is surrounded by ions of opposite signs, which implies the existence of short-range forces. To simplify the description of this complicated situation, Debye and Hu¨ckel described ions as impenetrable charged spheres in a solvent treated as a continuous and structureless dielectric; these features are also found in Wagner’s formalism. By applying several simplifications together with a linearization of the primitive equations, Debye and Hu¨ckel gave the following form to their combination of Poisson and Boltzmann equations:

∇2Φi ) κ2 Φi(R) (|R| g a)

(1)

where Φi(R) is the average electric potential arising from the ion i whose position is taken as the origin of R vector, and from the other ions of the system, R is the position vector of the point where this potential is observed, κ is the reverse of the Debye length, and a is the radius of ions considered as impenetrable spheres. As the ions are spherical in the model, the only distancedependence is through the ionic separation R, which leads to the ordinary form of Debye-Hu¨ckel equation

(1/R)[d2(RΦi)/dR2] ) κ2Φi (R g a)

(2)

where the form chosen for ∇2 is typically of spherical symmetry, because solution bulk is isotropic. An essential feature of a surface layer is to be anisotropic. This anisotropy is, usually, considered as localized in a direction perpendicular to the surface plane. The consequences are numerous: the rupture of the dielectric constant induces a different energetic situation for the ions, and the symmetry of the problem becomes cylindrical instead of spherical. The WOS treatment takes into account both aspects. In the primitive Boltzmann distribution used in bulk by Debye and Hu¨ckel, the term, wij, i.e., the potential of the mean force acting between i and j ions, is purely Coulombic. In the original scheme by Wagner, this term is replaced by the potential of the force exerted on the ion, i, (charge q ) Zie) by the image charge, q′, located at a distance 2z, from q. Coordinate z is the distance between q and the separation plane between both media; the dielectric constants of the media containing i and q′ are respectively 1 and 2. Then

2. Theory

q′ ) [(1 - 2)/(1 + 2)]q

2.1. WOS Description of Ionic Adsorption. The so-called Onsager-Samaras theory of electrolytic adsorption,12,13 whose

Let us assume that 1 ) w for water and 2 ) a for the air; then the force between q and q′ is

(11) Wagner, C. Physik Z. 1924, 25, 474. (12) Onsager, L.; Samaras, N. N. T. J. Chem. Phys. 1934, 2, 528. (13) Buff, F. P.; Stillinger, F. H., Jr. J. Chem. Phys. 1956, 25, 312. (14) Bhuiyan, L. B.; Bratko, D.; Outhwaite, C. W. J. Chem. Phys. 1991, 95, 336. (15) Dang, A. J.; Chang, T.-M. J. Phys. Chem. B 2002, 106, 235. (16) Jungwirth, P.; Tobias, O. J. J. Phys. Chem. B 2002, 106, 6361. (17) Vrbka, L.; Mucha, M.; Minofar, B.; Jungwirth, P.; Brown, E. C.; Tobias, D. Curr. Opin. Colloid Interface Sci. 2004, 9, 67. (18) Petersen, P.; Saykally, R. J. Chem. Phys. Lett. 2005, 397, 51. (19) Jones, G.; Ray W. A. J. Am. Chem. Soc. 1937, 59, 187. (20) Petersen, P. B.; Johnson, J. C.; Knutsen, K. P.; Saykally, R. J. Chem. Phys. Lett. 2004, 397, 46.

F)

(w - a) (Z e)2/16π0wz2 ) -[dw(z)/dz] (w + a) i

(3)

(4)

and the corresponding potential, which clearly depends on only one of the spatial variables, is expressed as (21) Debye, P.; Hu¨ckel, E. Physik Z. 1923, 24, 185. (22) Berry, R. S.; Rice, S. A.; Ross, J. Physical Chemistry; Wiley: New York, 1980.

8792 Langmuir, Vol. 22, No. 21, 2006

(

w(z) ) (Zie)2

)[

w - a w + a

Acharid et al.

]

exp(-2κz) 16π0wz

(5)

This method of “image forces” is classical in electrostatics.23 It constitutes a very short way to take into account the different aspects of the energetic effects issued from the polarization effects at the contact between two dielectrics when one of them contains a charge. However, relation (5) does not take into account the screening of the electrostatic force by the other ions. From eq 5, Wagner11 adapted the Poisson-Boltzmann formalism used by Debye and Hu¨ckel by expressing ∇2 in a symmetry adapted to the considered surface, i.e., the cylindrical one. It ensues that

(∂2Φ/∂z2) + (1/F)(∂/∂F)(F∂Φ/∂F) ) κ2(z)Φ

(6)

Let us, now, consider two punctual charges, Ze located at z and representing the ion of concern and its image located at -z. By using the following approximation

κ2(z) ∼ κ2(∞) ) κ2

(7)

then the solutions of eq 6 are

Φ ) Ze(exp - κr1)/4π0wr1 + [(w - a)/(w + a)] Ze(exp - κr2)/4π0ar2 (8) where Φ is the electrostatic potential in the ion-containing medium at a point located at the distance r1 from the ion and the distance r2 from the image; the second term arises from the fictitious surface charge. One should note that Onsager and Samaras also considered 2 (written as a above) negligible as compared to 1 (written as w above), which is inadequate, here. Φ leads to the Coulombic force, F, corresponding to the energy, w(z), which is simply calculated by Wagner as

w(z) ) (1/2)qΦ(r2) for

r2 ) 2z

(9)

By considering also the size of ions, and introducing the DebyeHu¨ckel factor, a, which is small with respect to z, w(z) takes, finally, the following form:

w(z) ) [(w - a)/(w + a)] {(exp κa)/ (1 + κa)}{Z2e2 [exp(-2κz)]/16π0wz} (10) This method of calculation constitutes a very direct way to get the concentration profiles in the surface phase. By introducing eq 10 into Boltzmann’s equation and using the cylindrical symmetry, the ionic concentrations can be written as

{

ci(z) ) ci(∞) exp -

}

w - a exp κa Z2e2[exp(-2κz)] w + a 1 + κa 16π0wzRT (11)

where ci(∞) is simply the bulk molarity of the ionic species, i. Adsorption excesses for each ionic species, i, can be calculated by using the Gibbs definition together with relation (11)

Γi )

∫∞GP[ci(z) - ci∞] dz ) ci∞ ∫∞GP[{ci(z)/ci∞} - 1] dz

(12)

where GP is the position of the Gibbs plane taken at ci ) 0. The Onsager-Samaras theory being an adaptation of the Debye-Hu¨ckel analysis suffers from the same limitations. Among (23) Durand, E. Electrostatique III. Me´ thodes de calcul. Die´ lectriques; Masson: Paris, 1966.

Figure 2. Cationic profiles and adsorption isotherms in absence and presence of benzene, according to the WOS model. (a) Cationic profiles; (b) adsorption isotherms at 25 °C, calculated point by point by integration of profiles such as those shown in (a); ): 1-1 electrolyte, 4: 1-2 electrolyte, * 1-2 electrolyte in the presence of benzene.

them, as pointed out by Robinson and Stokes,24 the most drastic one is the exponential nature of the Boltzmann distribution; indeed, this form forbids the application of the fundamental law of electrostatics, i.e., the linear superimposition of fields. Practically, the Debye and Hu¨ckel description overestimates the density of counterions around a given central ion. However, these criticisms become far less topical on the condition that the solution be a symmetrical electrolyte at low concentration ( 0, Γw (or the corresponding Γ0W, as defined in Table 1 caption) is always negative because Γs is also always negative. When Γs,w < 0, Γw can become >0 on condition that |Γs| < |Γsw|. Finally, Γw is null when the experimental values are exactly those given by the WOS model. It is worth noting that the last two cases are quite possible, but uneasily observed because of the uncertainties on the very low values of Γs and Γs,w. From a single concentration, Table 2 gives the equivalent set of data for lead nitrate together with Γs,w calculated for a bulk concentration of 1.83 × 10-2 M (Figure 4). It is worth noting the high number of experimental data around this value; so, a tangent to the supposed curve can be easily drawn to determine

Table 1. KCl and Water Adsorptions Calculated from Surface Tension Measurements or Simple Theoriesa C/mol L-1 -9

-1

Γsalt,W/10 mol L Γsalt/10-9 mol L-1 ΓW/ 10-5 mol m-2 Γ0W/10-5 mol m-2

4 × 10-4

10-3

2 × 10-3

10-2

2.47 ( 0.01 -0.49 ( 0.05 -41.0 ( 1.0 -6.8 ( 0.6

0.00 -1.00 ( 0.10 -5.50 ( 0.55 -5.50 ( 0.55

-0.97 ( 0.01 -1.95 ( 0.20 -2.70 ( 0.54 -5.4 ( 0.27

-4.87 ( 0.02 -7.37 ( 0.74 -1.40 ( 0.42 -4.10 ( 0.14

a KCl adsorption: Γsalt,w values were calculated from surface tension measurements by Ray and Jones,19 Γsalt from WOS model. Water adsorption: Γw and Γ0w values were, respectively, calculated by using Γsalt,w experimental determinations19 and the commonly admitted relation Γsalt,w ) 0, both combined with Γsalt through eq 33.

8796 Langmuir, Vol. 22, No. 21, 2006

Acharid et al.

Figure 7. Relative adsorption of salts versus salt and carbofuran concentrations at the air/solution interfaces. In light gray: Pb(NO3)2. In dark gray: PbCl2.

Figure 5. Surface tensions of mixtures of Pb(NO3)2 and carbofuran in water versus salt concentrations, at 25 °C and constant concentrations of carbofuran (), 10-5 M; 4, 5 × 10-5 M; O, 10-4 M; 0, 1.3 × 10-4 M; +, 1.4 × 10-4 M). Dotted vertical lines indicate salt concentration that induce the carbofuran demixing at the studied carbofuran concentration.

Figure 8. Relative adsorption of carbofuran versus salt and carbofuran concentrations at the air/solution interfaces. In light gray: Pb(NO3)2 system. In dark gray: PbCl2 system.

Figure 6. Surface tensions of mixtures of PbCl2 and carbofuran in water versus salt concentrations, at 25 °C and constant concentrations of carbofuran (), 10-5 M; 4, 5 × 10-5 M; O, 10-4 M; 0, 1.3 × 10-4 M; +, 1.4 × 10-4 M). Dotted vertical lines indicate carbofuran solubility limit according to the salt concentration. Table 2. Pb (NO3)2 Adsorption from a Single Salt Concentration and Calculated by Different Ways

c/10-2 mol L-1 1.83

ΓS,W/ 10-8 mol m-2 (experimental)

ΓS ≡ Γ+/ 10-8 mol m-2 (WOS)

ΓW/10-5 mol m-2 (from exptl Γs,w)

Γ0w/10-5 mol m-2 (if Γs,w)0)

-1.14 ( 0.11 -1.82 ( 0.18 -2.1 ( 0.9 -5.50 ( 0.34

Γs,w. Γs was taken identical to the Γ+ provided by the WOS model for the reasons explained above and was found to be close to Γs,w. Unexpectedly, this result was very similar to the one found for KCl, although the WOS theory was considered as inapplicable to 1-2 electrolytes. This finding is in favor of the method used hereafter to calculate Γs: first, the WOS theory seems to better stand for unsymmetrical electrolytes than expected, and second the choice of Γ+ to represent Γs sounds correct. 4.1.2. Mixtures of Salts and Organic Molecules. Concerning the mixtures of salts with organic molecules, Figures 5 and 6 display the surface tensions for mixtures of carbofuran with leadnitrate and -chloride, respectively. The experimental points were

Figure 9. Relative adsorption of Pb(NO3)2 in the presence of carbofuran, at the air/solution interfaces, versus the salt content and the saturation fraction of carbofuran.

drawn with respect to salt concentrations but at constant concentrations of carbofuran. Though the reciprocal sets, i.e., surface tension versus carbofuran concentrations at constant salt contents, are not shown here, they were also used to get the adsorptions presented in Figure 8. Adsorptions were calculated from eqs 28-31. Figure 7 compares the relative adsorption of lead-nitrate and -chloride with respect to carbofuran- and salt-concentrations. It is worth noting that (i) the salt relative adsorption, which was roughly negative with no organic molecules, is positive now and enhanced by increase in carbofuran concentrations and, thus, in its adsorption; (ii) the relative adsorption of lead nitrate is higher than that of lead chloride. Figure 8 shows that the relative carbofuran adsorption varies with the contents of the organics’ and nonorganics’ mixture. Once again, the effect of nitrate is stronger than that of chloride. Both figures evidence an enhancement of all effects near the salt-concentration-dependent solubility limit of carbofuran. This finding is clearly highlighted by Figures 9 and 10 dealing with the comparison of lead nitrate adsorption in the presence of either carbofuran or benzene (data from reference10), because the organics’ content is given relative to the solubility limit. It is

Water BehaVior ReVisited

Langmuir, Vol. 22, No. 21, 2006 8797

Figure 10. Relative adsorption of Pb(NO3)2 in the presence of benzene, at the air/solution interfaces, versus the salt content and the saturation fraction of benzene, according to ref 10.

worth noting the very high salt adsorption (10-5 against 10-7) though the benzene concentrations are higher than carbofuran ones with nearly alike salt concentration scales. Figure 11 compares water excesses for lead nitrate and lead chloride/carbofuran mixtures. As observed for single solutions of salts, excesses are negative; it is worth recalling that, for these values, the Gibbs plane is taken at the abscissa of null ionic excess. The use of water excesses combined with the corresponding carbofuran relative adsorption in eq 27 under the form expressed in (34) leads to positive carbofuran excesses

Γcr ) Γcr,w +

xcr Γ xw w

Figure 11. Water adsorption in the presence of carbofuran and salt, with respect to salt concentration, at different carbofuran contents (), 10-5 M; 4, 5 × 10-5 M; O, 10-4 M; 0, 1.3 × 10-4 M; +, 1.4 × 10-4 M); solid line: Pb(NO)3; dashed line: PbCl2.

particle arrangement as concentration profiles. Figure 12 shows that the ionic profile is “WOS-like” and that its zero defines the Gibbs plane for the calculation of the whole set of adsorption excesses. According to Figure 12, water density is zero until ions take place in the layer, which leads to an overall negative value for the excess. The profile for the organic molecule is the outermost with respect to the solution bulk and was thus drawn as a step profile in agreement with a frequent habit for such a component. Fitting experimental data requires us to differentiate the anionic profiles from the cationic ones instead of only showing the salt adsorption, but, unfortunately, doing it from a thermodynamic analysis and WOS model is quite impossible. 4.2. Comparison with the Silica/Solution Interface. Figure 13 gives the relative adsorption, with respect to water, of carbofuran and lead nitrate from simple solutions onto silica. It is worth noting that the results are almost opposite to those at the air/solution interface despite the use of identical concentrations. Furthermore, carbofuran adsorption is slightly negative, whereas the salt one becomes frankly positive.

(34)

Thus water, with respect to the carbofuran adsorbed layer, is repelled inside the solution and probably farther than most of the ions present in the surface layer. The values of Γw listed in Table 3 are issued either from several systems or from a single system studied under different conditions. The only constant parameter between them is the molarity of salts, 10-3 M, which allowed us to take Γs ) 0 for solutions of single salts, according to Figure 4. Table 3 shows that, whatever the system, Γw is negative. For the same salt, Γw is more negative in the presence of organic molecules, and carbofuran is a more efficient water-repellent than benzene. As already noticed in Figure 11, under similar conditions, water repulsion is greater with nitrates than with chlorides. From these views of the surface layer, we sketched

Table 3. Comparison of Water Adsorptions from Several Case Studies KCl alone [salt]/mol L-1 1 × 10-3

Γs,w/mol m-2

Γsalt/mol m-2 (Onsager 1-1) -1.0 × 10-9

Γw/mol m-2

Pb(NO3)2 or PbCl2 alone [salt]/mol L-1 1 × 10-3

Γs,w/mol m-2

Γw/mol m-2

0

Γsalt/mol m-2 (Onsager 1-2) -2 × 10-9

Pb(NO3)2 + carbofuran [salt]/mol L-1 1 × 10-3

Γs,w/mol m-2 ([carb] ) 10-4 mol L-1) 2.17 × 10-6

Γsalt/mol m-2 (Onsager 1-2) -1.2 × 10-9

Γw/mol m-2

PbCl2 + carbofuran [salt]/mol L-1 1 × 10-3

Γs,w/mol m-2 ([carb] ) 10-4 mol L-1) 3.59 × 10-7

Γsalt/mol m-2 (Onsager 1-2) -1.2 × 10-9

Γw/mol m-2

(Pb(NO3)2 + benzene [salt] /mol L-1 1 × 10-3

Γs,w/mol m-2 ([benzene] ) 2 10-2 mol L-1) 2 × 10-8

Γsalt/mol m-2 (Onsager 1-2) -2 × 10-9

Γw/mol m-2

0

-5.55 × 10-5

-11.1 × 10-5

-12 × 10-2

-19.9 × 10-3

-12.2 × 10-4

Table 4. Adsorption Excesses at the Silica Surface in Contact with Aqueous Mixtures of Carbofuran and Lead Nitratea Ccr 10-3 a

CPb(NO3)2 3.83 ×

10-3

Γcr,w -1.88 ×

Γs,w 10-9

4.95 ×

Concentrations are in mol L-1 and adsorptions in mol m-2.

10-7

Γs (WOS) -3.5 ×

10-9

Γw -7.2 ×

Γcr 10-3

-1.28 × 10-7

8798 Langmuir, Vol. 22, No. 21, 2006

Acharid et al.

Figure 12. Speculative concentration (or density) profiles of the three components at the interface between air and solution: areas represent adsorption excesses. In black: salt. Black and white checked: organic molecule. In white and gray squares: water. The bulk limit concentrations for each i component are shown by horizontal lines, labeled C∞i . The sign of each piece of area is given on the graph.

Figure 14. Adsorption of carbofuran (a) and lead nitrate (b) from mixtures 1-7, at the silica/solution interfaces. The initial concentration of carbofuran was 10-3 M for each mixture. Equilibrium concentrations are given below, as Mcr and MPb for carbofuran and salt respectively (in mmol L-1). 1: Mcr ) 1; MPb)3.83. 2: Mcr ) 1.045, MPb ) 3.42. 3: Mcr ) 1.032, MPb ) 2.71. 4: Mcr ) 1.028, MPb ) 2.01. 5: Mcr ) 1.053, MPb ) 1.20. 6: Mcr ) 1.027, MPb ) 0.595, 7: Mcr ) 1.022, MPb ) 0.099.

Table 4 shows that the water excess is more negative than at the other interface; from a physical point of view, this finding is puzzling, except if the role of the Gibbs plane position is correctly analyzed. In our model, it was taken at the null adsorption plane of the salt. Silica surface being hydrophilic, it sounds natural to assume the existence of a water layer between silica and ions. This water layer is likely responsible for the very weak adsorption of carbofuran, but gives a water density profile probably lying on both sides of GP and leading to the values observed experimentally.

5. Conclusion

Figure 13. Adsorption of lead nitrate (a) and carbofuran (b) on silica at 25 °C, from solutions where each of them is alone. The concentration scales are those of the solutions in equilibrium with silica suspension (supernatant).

Panels a and b in Figure 14 respectively display the relative adsorption of carbofuran and salt onto silica from mixtures. The signs of relative adsorptions are conserved by comparison with simple solutions; both compounds have their adsorption enhanced, particularly salts: the mixing effects are similar to those observed with mixtures at the air/solution interface. We used the same model as previously described to calculate the excesses of water and other components for the first point of the series of data (Table 4).

In this study, adsorption data of several salts dissolved in water as well as those about the coadsorption of these salts with organic molecules were carefully analyzed from the Gibbs adsorption and WOS statistical model. Their close examination through the calculation of water excesses in the different systems together with the assumption of a well-defined position of the Gibbs plane drove us to elaborate some hypotheses about particle organization in the surface layer: water would be repelled toward the solution bulk, whereas ions would likely take place between the bulk and a layer of organic molecules. The differences observed between coadsorption data by changing the nature of the anion under study suggest that simulation data such as those in refs 16 and 17 give a correct description of physical behavior; one should also note that they are consistent with other experimental data.18 Due to its roughness, it would be worth replacing the analysis through the WOS model by more elaborated representations, but the verifications made in this study showed that this model is better than thought by some people. To our

Water BehaVior ReVisited

opinion a model based on quantum mechanics should allow one to check the validity of the interactions suggested here, particularly between solutes and water. It could constitute a basis for further statistical simulations. Concentration profiles as suggested in Figure 12 provided a description of environmental systems helpful to foresee reactivity problems in particulate patterns such as (30) Koelsh, P.; Motschmannn, H. Langmuir 2005, 21, 3436.

Langmuir, Vol. 22, No. 21, 2006 8799

aerosols.5-7 The case of adsorption onto silica is more complicated and constitutes an interesting challenge for simulations. Another challenge is about ionic surfactant layers where the ionic charge is present onto the organic chain itself.30 Acknowledgment. The authors gratefully acknowledge “Re´gion Bretagne” for financial support. LA061160C