Langmuir 2006, 22, 3207-3213
3207
Adsorption of Dodecyltrimethylammonium Bromide and Sodium Bromide on Gold Studied by Liquid Chromatography and Flow Adsorption Microcalorimetry Zolta´n Kira´ly,*,†,‡ Gerhard H. Findenegg,‡ and A Ä gnes Mastalir§ Department of Colloid Chemistry, UniVersity of Szeged, Aradi Ve´ rtanu´ k tere 1, H-6720 Szeged, Hungary, Stranski-Laboratory for Physical and Theoretical Chemistry, Technical UniVersity of Berlin, Strasse des 17. Juni 112, D-10623 Berlin, Germany, and Department of Organic Chemistry, UniVersity of Szeged, Do´ m te´ r 8, H-6720 Szeged, Hungary ReceiVed NoVember 24, 2005. In Final Form: February 5, 2006 Here, we report on a new aspect of the adsorption of Br- on the surface of gold. The adsorption of dodecyltrimethylammonium bromide (C12TABr) from aqueous solutions onto macroporous gold particles was studied by continuous flow frontal analysis solid/liquid chromatography and flow adsorption microcalorimetry. The material balance and enthalpy balance of adsorption and the change in the solution pH were measured simultaneously. Initially, Br- is irreversibly bound to high-affinity surface sites counterbalanced by the adsorption of H+ from the aqueous phase. The surface speciation is accompanied by the formation of C12TAOH, which in turn results in a significant pH increase in the bulk solution. The net process was found to be strongly exothermic (-280 kJ‚mol-1), which is indicative of the occurrence of chemisorption. The specific adsorption of Br- is followed by the reversible adsorption of C12TABr to produce a firmly bound monolayer in a head-to-surface arrangement (-53 kJ‚mol-1). In a relatively narrow range of the surface coverage, various composite structures may develop on the top layer and eventually transform to full-cylindrical surface aggregates. The surface aggregation was found to be reversible, with an enthalpy change of -11 kJ‚mol-1. The importance of the specific binding of Br- to the surface of gold was confirmed by measurement of the initial adsorption of NaBr on the microparticles. The initial adsorption was found to be irreversible, with an enthalpy change of approximately -240 kJ‚mol-1. This process involved the formation of an AuBr-/H+ electric double layer at the gold/water interface, accompanied by a dramatic increase in the solution pH due to the release of a copious amount of OH- in the bulk liquid phase.
Introduction Graphite, silica, and mica are regarded as model adsorbents possessing hydrophobic, slightly charged hydrophilic, and highly charged hydrophilic surfaces, respectively. The adsorption properties of these materials have been extensively studied. Alkyltrimethylammonium bromides (CnTABr) are the prototype of cationic surfactants and have been widely used in studies of adsorption from aqueous solutions on a variety of solid substrates. These surfactants form ordered half-cylindrical aggregates on graphite,1-8 spherical surface aggregates on silica,2,3,5,9-11 and full-cylindrical aggregates on mica,2,3,5,9,12-14 as revealed by * To whom correspondence should be addressed. E-mail: zkiraly@ chem.u-szeged.hu. † Department of Colloid Chemistry, University of Szeged. ‡ Stranski-Laboratory for Physical and Theoretical Chemistry, TU Berlin. § Department of Organic Chemistry, University of Szeged. (1) Manne, S.; Cleveland, J. P.; Gaub, H. E.; Stucky, G. D.; Hansma, P. K. Langmuir 1994, 10, 4409-4413. (2) Manne, S.; Gaub, H. E. Science 1995, 270, 1480-1482. (3) Manne, S. Prog. Colloid Polym. Sci. 1997, 103, 2226-2233. (4) Liu, J.-F.; Ducker, W. J. Phys. Chem. B 1999, 103, 8558-8567. (5) Atkin, R.; Craig, V. S. J.; Wanless, E. J.; Biggs, S. AdV. Colloid Interface Sci. 2003, 103, 219-304. (6) Kira´ly, Z.; Findenegg, G. H. J. Phys. Chem. B 1998, 102, 1203-1211. (7) Kira´ly, Z.; Findenegg, G. H.; Klumpp, E.; Schlimper, H.; De´ka´ny, I. Langmuir 2001, 17, 2420-2425. (8) Kira´ly, Z.; Findenegg, G. H.; Mastalir, A Ä . J. Phys. Chem. B 2003, 107, 12492-12496. (9) Wangnerud, P.; Berling, D.; Olofsson, G. J. Colloid Interface Sci. 1995, 169, 365-375. (10) Partyka, S.; Lindheimer, M.; Faucompre, B. Colloids Surf., A 1993, 76, 267-281. (11) Lajtar, L.; Narkiewitz-Michalek, J.; Rudzinski, W.; Partyka, S. Langmuir 1994, 10, 3754-3764. (12) Lamont, R. E.; Ducker, W. A. J. Am. Chem. Soc. 1998, 120, 7602-7607. (13) Ducker, W. A.; Wanless, E. J. Langmuir 1999, 15, 160-168.
atomic force microscopy (AFM) studies at a molecular level1-5,12-14 and by adsorption microcalorimetry studies on the thermodynamic level.6-11 The theoretical model of Johnson and Nagarajan for the self-assembly of surfactants at solid/liquid interfaces complies with these experimental observations on both hydrophobic15 and hydrophilic substrates.16 Although research efforts to study surfactant adsorption on metal surfaces have recently intensified, the major difficulty in such studies stems from the spontaneous formation of a native, thin oxide layer on the surface; in such cases, adsorption occurs on the metal oxide rather than on the parent metal. A further problem arises when carbonaceous contamination adheres to the metal surface upon storage, which renders the surface hydrophobic. Although gold is the most noble of all metals, because of its very low reactivity, there has been considerable controversy in the literature as to whether the gold surface is hydrophilic or hydrophobic; it cannot easily be classified in either category.17 Nevertheless, since Au2O3 has no thermodynamic stability,18 gold appears to be a much better candidate for the study of adsorption phenomena compared with other members of the noble metal series. The adsorption of surfactants on gold19-22 has received little attention so far compared with the adsorption of (14) Schulz, J. C.; Warr, G. G. Langmuir 2000, 16, 2995-2996. (15) Johnson, A. R.; Nagarajan, R. Colloids Surf., A 2000, 167, 21-36. (16) Johnson, A. R.; Nagarajan, R. Colloids Surf., A 2000, 167, 37-46. (17) Smith, T.; J. Colloid Interface Sci. 1980, 75, 51-55. (18) Tsai, H.; Hu, E.; Perng, K.; Chen, M.; Wu, J.-C.; Chang, Y.-S. Surf. Sci. 2003, 537, L447-L450. (19) Jaschke, M.; Butt, H.-J.; Gaub, H. E.; Manne, S. Langmuir 1997, 13, 1381-1384. (20) Stalgren, J. J. R.; Boschkova, K. Langmuir 2002, 18, 6802-6806. (21) Benton, D. P.; Sparks, B. D. Trans. Faraday Soc. 1966, 62, 3244-3252. (22) Benton, D. P.; Sparks, B. D. Trans. Faraday Soc. 1967, 63, 2270-2274.
10.1021/la053184+ CCC: $33.50 © 2006 American Chemical Society Published on Web 02/25/2006
3208 Langmuir, Vol. 22, No. 7, 2006
thiols20,23-27 and inorganic anions, especially halide ions,28-33 on gold. This may be, in part, a consequence of the greater number of experimental difficulties to be overcome with the former system. Manne et al. presented direct images of ionic surfactant aggregates at the macroscopic, planar surface of gold, obtained by using AFM.19 In the postmicellar region of C14TABr, the surface pattern was composed of parallel, flexible stripes, compatible with ordered domains of full-cylindrical aggregates. A model was proposed in which surfactant adsorption in direct contact with the substrate occurs with the ionic trimethylammonium headgroups pointing toward the surface of the metal, and with the Br- being sandwiched between the surface and the heads. The vertical orientation of the alkyl chains favors further associative adsorption of C14TABr molecules to produce fullcylindrical surface aggregates. The high- affinity adsorption of Br- on gold, which plays a crucial role in the above model, has been well established by electrochemical, spectroscopic, probe microscopic, and some other experimental techniques.29-33 Boschkova and Stalgren performed adsorption measurements at a concentration of 1.2 times the critical micelle concentration (cmc) of dodecyltrimethylammonium bromide (C12TABr), using the quartz crystal microbalance technique.20 It was found that micellar adsorption occurs on the surface of gold and that 20% of the surface remains covered by C12TABr molecules, even after elution of the substrate with pure water. Some 40 years ago, Benton and Sparks reported on the adsorption isotherms of a series of CnTABr amphiphiles on gold.21,22 The occurrence of surface aggregation was not foreseen in that study; the postulated structure of the interfacial layer was markedly different from that proposed in the above, more recent AFM study. Nevertheless, a reinterpretation of the adsorption and electrokinetic data of Benton and Sparks led us to conclude that the raw experimental data reported in that pioneering work are in line with the AFM results of Manne et al. The adsorption isotherms of CnTABr in a suspension of gold correlated well with the electrokinetic potential of the gold fines, both measured up to the cmc: the shapes of the isotherms were of the doubleplateau type, and the electrophoretic mobility data exhibited a surface charge reversal from negative to positive as the concentration of the surfactant in the bulk solution was increased.21,22 These observations are consistent with the progressive building-up of a bilayer in which both the first and second layers are in a vertical orientation with facing or interdigitated tailgroups, and consistent with the gradual formation of cylindrical surface micelles in which the headgroups are directed toward the aqueous phase. It has been found that the extents of adsorption of Br- 33 and CnTABr34,35 differ significantly on the edges and faces and on the various crystal planes of gold particles. Further, the presence (23) Zhong, C. J.; Brush, R. C.; Anderegg, J.; Porter, M. D.Langmuir 1999, 15, 518-525. (24) Jung, L. S.; Campbell, C. T. Phys. ReV. Lett. 2000, 84, 5164-5167. (25) Prasad, B. L. V.; Stoeva, S. I.; Sorensen, C. M.; Klabunde, K. J. Langmuir 2002, 18, 7515-7520. (26) Brust, M.; Kiely, C. J. Colloids Surf., A 2002, 202, 175-186. (27) Park, J.-S.; Vo, A. N.; Barriet, D.; Shon, Y.-S.; Lee, T. R. Langmuir 2005, 21, 2902-2911. (28) Hora´nyi, G. Rizmayer, E. M. Joo´, P. J. Electroanal. Chem. 1983, 152, 211-222. (29) Deakin, M. R.; Li, T. T.; Melroy, O. R. J. Electroanal. Chem. 1988, 243, 343-351. (30) Tao, N. J.; Lindsay, S. M. J. Phys. Chem. 1992, 96, 5213-5217. (31) Shi, Z.; Lipkowski, J.; Mirwald, S.; Pettinger, B. J. Chem. Soc., Faraday Trans. 1996, 92, 3737-3746. (32) Wang, J.; Bard, A. J. J. Phys. Chem. B 2001, 105, 5217-5222. (33) Magnussen, O. M. Chem. ReV. 2002, 102, 679-725. (34) Nikoobakht, B.; El-Sayed, M. A. Langmuir 2001, 17, 6368-6374. (35) Gao, J.; Bender, C. M.; Murphy, C. J. Langmuir 2003, 19, 9065-9070.
Kira´ ly et al.
of defect sites makes the surface energetically heterogeneous, and these sites are generally more reactive than the more fully bonded terrace sites.19 Fourier transform infrared (FTIR) spectroscopic studies have revealed the formation of CnTABr bilayer assemblies on gold nanorods and nanospheres.34,35 Experimental evidence of the bilayer structure of CnTABr has also been provided on silver36 and palladium37 nanoparticles. The apparent controversy between the two aggregate morphologies is readily resolved by considering that the bilayer structure on highly curved surfaces (nanoparticles) is the most plausible, complementary structure of full-cylinders on macroscopic, flat surfaces (cleaved planes, wafers, and annealed films). In the present work, we report on a liquid-flow microcalorimetric study of the adsorption of C12TABr from aqueous solution onto gold microparticles (-20 mesh) at 298.15 K. The differential heat of adsorption was determined as a function of the surface coverage, and the results were interpreted in terms of a threestage adsorption model: the specific adsorption of Br-, accompanied by a marked increase in the solution pH; the subsequent monomolecular adsorption of C12TABr; and the formation of full-cylindrical surface aggregates, induced by the surfactant monolayer through a series of composite structures. The adsorption measurements were extended to a study of the binding of Br- to gold from a dilute aqueous solution of NaBr. Experimental Section Materials. C12TABr (>99%) was purchased from Sigma-Aldrich and used as received. The water used had a pH of 5.5 after ion exchange, distillation, and a further treatment with a Milli-Q filtration system. Stock solution was prepared by weight and then diluted volumetrically to the desired concentrations. Gold sponge (-20 mesh, >99.95%) was purchased from Alfa Aesar-Johnson Matthey (Lot No. D23G44), with a Brunauer-Emmett-Teller (BET) surface area to N2 of 0.54 ( 0.01 m2‚g-1 at 77 K (the very low surface area is apparent because of the very high density of gold, 19.3 cm-3‚g-1; the adsorbent has a surface area per unit volume of 10.42 m2‚cm-3). The gold samples were heated to 473 K in a vacuum oven before the adsorption calorimetric measurements. Methods. The experimental setup of the flow adsorption apparatus was similar to that described in detail in our previous studies.6,38,39 The fully automated measuring system consisted of a TAM 2277 isothermal microcalorimeter (Thermometrics, Lund, Sweden), an auxiliary high-performance liquid chromatograph supplied with a differential refractive index detector (RID) (Knauer, Berlin, Germany), and a flat-surface combination pH electrode attached to a low dead-volume (0.05 cm3) flow-through cell (Sensorex, Garden Grove, CA). The calorimeter vessel, a small chromatographic column, was loaded with the gold particles (0.4-0.8 g) and placed inside the measuring block of the calorimeter, and the column packing was conditioned in a stream of pure ethanol. After this pretreatment, pure water was percolated through the column at a constant flow rate of 6 cm3‚h-1. The adsorption experiments were performed by using a set of solutions containing sequentially increasing surfactant concentrations. Sufficient time was allowed for thermal equilibrium to be attained between neighboring concentration steps. For a step ∆c ) c′′ - c′, the change of the amount adsorbed, δΓ, can be related to the break-through (or retention) time, tR, as δΓ )
(Q × tR - VD) ∆c m × aS
(1)
(36) Cheng, W.; Doung, S.; Wang, E. Electrochem. Commun. 2002, 4, 412416. (37) Veisz, B.; Kira´ly, Z. Langmuir 2003, 19, 4817-4824. (38) Kira´ly, Z.; Bo¨rner, R. H. K.; Findenegg, G. H. Langmuir 1997, 13, 33083315. (39) Kira´ly, Z. In Thermal BehaVior of Dispersed Systems; Garti, N., Ed.; Marcel Dekker: New York and Basel, Switzerland, 2001; pp 335-356.
Adsorption of C12TABr and NaBr on Gold
Langmuir, Vol. 22, No. 7, 2006 3209
where Q is the liquid-flow rate, m is the mass of the solid, aS is the specific surface area of the solid, and VD is the dead volume of the system. VD was readily determined in a separate experiment by using an aqueous solution of D2O (2% in water). tR was obtained from the analysis of the concentration profile c(t) continuously monitored by the RID: tR )
1 × c′′ - c′
∫
t∞
t0
(c′′ - c(t))dt
(2)
where to is the time at which the concentration of the feed solution is changed from c′ to c′′, and t∞ is the time at which the new adsorption equilibrium is reached, that is, c(t∞) ) c′′. For each step, the associated enthalpy change appeared as a calorimetric peak. Integration of the calorimeter power signal allowed the enthalpy of adsorption δ(∆aH) to be calculated. Thus, the differential enthalpy of adsorption was measured directly as ∆a h )
δ(∆aH) δΓ
(3)
The standard deviation of a single step in the frontal-flow adsorption experiment was, on average, 5.8%, while, for calorimetry, it was about 4.8%. This accuracy provides more reliable differential molar enthalpy data, calculated directly as ∆ah ) δ(∆aH)/δΓ for each step, than a combination of separately measured adsorption and enthalpy isotherms. Variation of the solution pH upon adsorption was recorded by the flow-through pH monitor, and the data were then converted to OH- concentration as cOH-(t) ) antilog10(pH(t) - 14)
Figure 1. RID signal (a) and calorimeter signal (b) of the stepwise adsorption of C12TABr from aqueous solutions onto gold sponge at 298.15 K. Column packing: 0.7712 g of gold; flow rate: 6.65 cm3‚h-1; concentration sequence: 0, 0.2, 0.5, 1, 2, and 4 mM C12TABr.
(4)
The reliability of the conversion of pH to cOH- was confirmed in a series of blank experiments using 0.2 mM NaOH.
Results and Discussion The Initial Adsorption of C12TABr and NaBr on Gold. Gold sponge is a macroporous material40 that has a relatively low BET surface area. Measurement of the material balance and the enthalpy balance of adsorption at the solid/solution interface for adsorbents with low surface area requires instruments of high sensitivity.41 The present apparatus complies with this requirement.6,38,39 The concentration profile (RID signal) and the calorimeter heat-flow signal in the sequence 0, 0.2, 0.5, 1, 2, and 4 mM C12TABr adsorbed from aqueous solution onto gold sponge are presented in Figure 1. A close inspection of the RID signal and the calorimeter signal reveals anomalous adsorption behavior for the first concentration step: 0/0.2 mM C12TABr. The magnifications of the instrumental response curves for this step are given in Figure 2a,b. A shoulder develops on the sigmoidal break-through curve, indicative of the presence of two different solute species coming off the adsorption column, with an overlap region of the concentration profiles of the two components. Further, a shoulder develops on the descending branch of the heat-flow signal, composed of two closely spaced calorimeter peaks. This characteristic behavior diminishes after a full adsorption/desorption cycle: upon repetition of the 0/0.2 mM concentration step, the shoulder on the break-through curve and the double peak of the calorimeter signal are completely absent (Figure 2c,d). We observed that the amount adsorbed and the heat evolved were higher on the parent surface than on the same surface after use, and that the adsorption process was reversible for concentrations of c > 0.2 mM. It seemed reasonable to assume, therefore, that the nonreversible nature of the (40) Gao, S.; Zhang, H.; Wang, X.; Yang, J.; Zhou, L.; Peng, C.; Sun, D.; Li, M. Nanotechnology 2005, 16, 2530-2535. (41) Van Os, N. M.; Haandrikman, G. Langmuir 1987, 3, 1051-1056.
Figure 2. RID signal (a and c) and calorimeter signal (b and d) of the adsorption of C12TABr from aqueous solutions onto gold sponge at 298.15 K. Experimental conditions are the same as those described in Figure 1. The first adsorption step (0/0.2 mM) on the parent solid material was found to be nonreversible, and both the RID signal (a) and the calorimeter signal (b) displayed anomalous behavior for this step (see arrows). Subsequent steps were found to be reversible (not shown), as was the readsorption of 0/0.2 mM C12TABr after a full adsorption/desorption cycle (c and d).
adsorption observed for the 0/0.2 mM concentration step can be attributed to the formation of a chemical bond between the Brcounterions of C12TA+ and the most active surface sites of the gold substrate. This hypothesis was supported in a repeat experiment, in which a flow-through pH detector was attached to the outlet port of the freshly packed column. Figure 3a,b reveals that the initial adsorption of C12TABr is accompanied by a pronounced increase in the solution pH due to the release of a copious amount of OH- in the bulk solution. Such an effect was negligible for the subsequent concentration steps (c > 0.2
3210 Langmuir, Vol. 22, No. 7, 2006
Kira´ ly et al.
C12TAOH, and eqs 1 and 2 could therefore not be applied for calculation of the amount of C12TABr adsorbed. However, a good estimation of the amount of Br- irreversibly adsorbed is provided by the assumption that this quantity is equal to the amount of OH- produced in the bulk solution during the chemisorption of Br-:
AuS + C12TABr + H2O ) [Au‚Br-‚H+]S + C12TAOH (5)
Figure 3. Variation in pH with time (a and c) and variation in OHconcentration with time (b and d) for the adsorption of C12TABr from aqueous solutions onto gold sponge at 298.15 K. Column packing: 0.3785 g of gold; flow rate: 6.10 cm3‚h-1. The first adsorption step (0/0.2 mM) on the parent solid material was found to be nonreversible (a and b), indicating the formation of a chemical bond between Au and Br- on the surface accompanied by the generation of a large amount of free OH- in the bulk solution. The readsorption of 0/0.2 mM C12TABr after a full adsorption/desorption cycle was found to be reversible (c and d).
mM) and for the repeated 0/0.2 mM adsorption step after one adsorption/desorption cycle (Figure 3c,d). As a possible explanation, we propose that, in the early stage of the adsorption process, Br- is irreversibly adsorbed, mostly on low-coordination defect sites of the gold particles, including different kinds of edge atoms, corner atoms, domain boundaries, terrace vacancies, and adatoms, differing from terrace metal atoms in the degree of coordinative unsaturation. The surface complexes of AuBr- are then neutralized by H+ originating from dissociated water molecules. A charge neutralization of this kind implies either the stoichiometric incorporation of H+ into the Br- adlayer at the gold/water interface, or the formation of a AuBr-/H+ electric double layer in the interfacial region. In either case, the accumulation of H+ at the interface is accompanied by an increase in the OHconcentration in the bulk solution so as to rebalance any shift in the ionic product of water back to Kd ) 10-14, a shift which would be favored by the net adsorption process. It follows that C12TA+ is not directly involved in the chemisorption of Br-, but the surfactant cations leave the adsorption column in the form of C12TAOH, and this compound increases the pH of the flowing bulk solution. The pH (and cOH-) passes through a maximum as a function of time (Figure 3a,b). The restabilization of the pH detector baseline is a very sensitive indication of the attainment of adsorption equilibrium. The surface speciation of the highaffinity surface sites of the polycrystalline gold particles by Bris followed/accompanied by the reversible adsorption of C12TABr on the low-index surfaces. The superposition of physisorption on chemisorption is reflected by the anomaly observed for both the RID signal (Figure 2a) and the calorimeter signal (Figure 2b), and by the nonreversible nature of the adsorption (Figure 2a-d). The determination of the surface concentration of C12TABr on the parent substrate was not straightforward at high dilutions. For the first adsorption step of 0/0.2 mM, the break-through curve involved the refractive indices of the species C12TABr and
where the subscript s refers to the gold surface layer. A variety of hydrated gold-bromo complexes may be formed at the interface, depending on the extent of surface roughness/ heterogeneity and on the solution pH, and these surface complexes have their own stochastic character. The stoichiometry of eq 5 may therefore be an oversimplification of the interfacial binding of Br- to gold. Because of the complex interface structure, however, we shall not attempt to identify the various surface species that might be formed upon adsorption. In any case, the change in the pH is caused by the increase in the OHconcentration, which must be quantitatively related to the increase in the amount of Br- adsorbed. If eq 5 applies, monitoring the pH of the bulk solution allows calculation of the number of moles of OH- and, hence, the surface concentration of irreversibly bound Br-. This calculation was performed by the use of eq 4. Integration of the concentration profile for the 0/0.2 mM C12TABr concentration step (Figure 3b) yielded a value of 0.2636 µmol‚m-2 of Br-. After elution of the column, otherwise for the same concentration step, the amount of reversibly adsorbed surfactant was calculated by the evaluation of Figure 2c in terms of eqs 1 and 2 to give a value of 0.8435 µmol‚m-2 of C12TABr. The enthalpy of adsorption for the total (irreversible plus reversible) adsorption process was measured to be -118.3 mJ‚m-2 (Figure 2b), from which -44.7 mJ‚m-2 is related to the reversible adsorption of C12TABr (Figure 2d), while the remaining enthalpy change of -73.6 mJ‚m-2 is associated with the irreversible adsorption of Br- (Figure 3a,b). The molar enthalpy of adsorption can be calculated from a combination of the material balance and the enthalpy balance of the adsorption according to eq 3. Thus, the enthalpy of formation of the surface chemical bond between Br- and gold is on the order of -280 kJ‚mol-1, and the integral molar enthalpy of adsorption (physisorption) of C12TABr on gold is -53 kJ‚mol-1. The proposed mechanism for the high-affinity adsorption of Br- on gold (eq 5) was further confirmed by measurement of the initial adsorption of NaBr (0.2 mM in water) on a freshly packed column. Panels a and b of Figure 4 show the concentration profile and the heat-flow signal, respectively, for the 0/0.2 mM adsorption step on the parent metal. The variation in the pH and the related change in the OH- concentration in the bulk solution are given in panels a and b of Figure 5, respectively. Figures 4c,d and 5c,d show the instrumental response curves for the readsorption step of 0/0.2 mM of NaBr, that is, after the used column was eluted with pure water. There is a close analogy between the initial adsorption of C12TABr on gold and the initial adsorption of NaBr on gold in the sense that Br- is essentially irreversibly bound to high-affinity (low-coordination) surface sites of the substrate, and the negatively charged AuBr- sites are then compensated by the coadsorption of H+ at the expense of water molecules. The irreversible adsorption of Br- on these sites (Figure 4a,b) is followed/accompanied by the reversible adsorption of NaBr (Figure 4c,d) on low-affinity (high-coordination) surface sites. The irreversible adsorption of Br- is accompanied by a dramatic increase in the solution pH due to the liberation of a large amount of OH- in the bulk liquid phase (Figure 5a,b). As for C12TABr, we estimated the enthalpy of
Adsorption of C12TABr and NaBr on Gold
Figure 4. RID signal (a and c) and calorimeter signal (b and d) of the adsorption of NaBr from aqueous solutions onto gold sponge at 298.15 K. Column packing: 0.4097 g of gold; flow rate: 6.11 cm3‚h-1. The first adsorption step of 0/0.2 mM on the parent solid material (a and b) was found to be nonreversible and strongly exothermic. The readsorption of 0/0.2 mM NaBr after elution with pure water (c and d) was found to be reversible and less exothermic.
Figure 5. Variation in pH with time (a and c) and variation in OHconcentration with time (b and d) for the adsorption of NaBr from aqueous solution onto gold sponge at 298.15 K. Experimental conditions and concentration steps were the same as those described in Figure 4. The adsorption is seen to be nonreversible, indicating the formation of a chemical bond on the surface between Au and Br-, accompanied by the generation of a copious amount of free OH- in the bulk solution.
chemisorption of Br- on gold to be -240 kJ‚mol-1 (irreversible adsorption up to 0.4026 µmol‚m-2 of Br-), and the enthalpy of physisorption of NaBr to be -67 kJ‚mol-1 (reversible adsorption of 0.3405 µmol‚m-2 of NaBr); these enthalpy changes are of the same order of magnitude as that for the adsorption of C12TABr on gold. It may be noted that regeneration of the surfaces of the gold particles after use was difficult to achieve. Flushing of the column
Langmuir, Vol. 22, No. 7, 2006 3211
Figure 6. Adsorption of C12TABr from aqueous solution onto gold sponge (solid symbols) at 298.15 K: adsorption isotherm (a) and enthalpy isotherm of adsorption (b). The experimental conditions are the same as those described for Figure 1. Dashed line: first adsorption path; solid symbols with solid line: second adsorption (and desorption) path. The very first adsorption step of 0/0.2 mM C12TABr was found to be nonreversible and strongly exothermic, which is characteristic of chemisorption. Subsequent adsorption/ desorption steps were found to be reversible and less exothermic, which is characteristic of physisorption. Open symbols: adsorption of C12TABr from aqueous solution onto graphite.6
with a 200 mM solution of NaOH (pH > 13) resulted in a gold surface free of Br-, but this treatment increased the specific surface area of the particles to an appreciable extent, presumably in consequence of surface reconstruction, surface restructuring, and/or surface metal-dissolution processes. The surface complexation reaction is sensitive to the pH of the solution from which the Br- is adsorbed. Further Adsorption of C12TABr on Gold up to the cmc. Figure 6 depicts the cumulative adsorption isotherm and the integral enthalpy isotherm of the adsorption of C12TABr on gold in the concentration region from c ) 0 to c ) cmc. While the solid line with the filled symbols refers to the reversible adsorption path (i.e., when the gold surface was preliminarily treated with a surfactant solution and then eluted with pure water), the dashed line is a good estimation of the nonreversible isotherm, characteristic of the adsorption on the parent substrate (the adsorbent was in contact with pure water only). For the sake of comparison, the isotherms of the adsorption of C12TABr at the graphite/water interface are also indicated in the figure. Although the isotherms on gold and graphite are qualitatively similar to each other, the formation and the structure of the adsorption layer are different in the two cases. On the strongly hydrophobic surface of graphite, the headgroups and the counterions do not adsorb significantly on the hydrophobic basal planes, but the physisorption of the tailgroups predominates, and the surfactant molecules form half-cylindrical surface micelles templated by a flat, ordered monolayer. This configuration was proposed in a series of AFM studies1-5 and was supported by the results of adsorption microcalorimetry studies.6-8 For gold, however, the Br- counterions can preferentially adsorb on the surface, thereby creating a dense, negatively charged surface layer. In the monolayer region at low surfactant concentrations, adsorption is then primarily driven by headgroup-surface electrostatic forces, leading to the formation of full surface cylinders at higher
3212 Langmuir, Vol. 22, No. 7, 2006
Figure 7. Enthalpies of adsorption plotted against surface concentration at 298.15 K. Solid symbols: reversible adsorption for the system C12TABr-water/gold sponge. Irreversible adsorption is not indicated. Open symbols: enthalpies of the adsorption of C12TABr at the graphite/water interface.6 The arrow indicates the critical surface coverage ascribed to the formation of surfactant aggregates induced by the monolayer: vertical monolayer/full cylinders on gold, and horizontal monolayer/half cylinders on graphite. Various microstructures may develop between the two characteristic adsorbate structures.15,16 It should be noted that ∆ahh measured for the lowest surface coverage is an integral molar quantity (dashed line). ∆ah values measured for higher surface coverage are differential molar quantities.
concentrations. This mechanism was proposed in the AFM study of Manne et al.,19 and the present results are consistent with it. The steeply rising section in the adsorption isotherm in Figure 6a covers the monolayer region, after which the adsorption continues to increase in the surface aggregation region until it turns to a plateau, close to the cmc. The shape of the isotherm in Figure 6a is slightly different from that reported by Benton and Sparks.21,22 Nevertheless, the plateau level is the same for the two cases: Γ ) 3.7 µmol‚m-2. The calorimetric enthalpy isotherms of adsorption in Figure 6b run quite parallel to the corresponding adsorption isotherms in Figure 6a. The (differential) molar enthalpies of adsorption in the reversible adsorption region are plotted against surface concentration in Figure 7. Since the adsorption of C12TABr for the first, small concentration step of 0/0.2 mM led to a relatively high surface coverage, the associated enthalpy change per mole of adsorbed surfactant is an integral molar quantity rather than a differential molar quantity. To emphasize this difference, the integral molar enthalpy of adsorption is designated ∆ah, as compared with the notation ∆ah that is applied for the differential molar enthalpy of adsorption. C12TABr monolayer formation is less exothermic on the surface of gold (-53 kJ‚mol-1) than it is on the surface of graphite (-62 kJ‚mol-1), as depicted in Figure 7. On both surfaces, aggregation is induced by the adsorbed monolayer. There is a relatively narrow transition region between the two stages of the adsorption process, in which the enthalpy of adsorption drops from -53 to -11 kJ‚mol-1 on gold, and from -62 to -17 kJ‚mol-1 on graphite. A comparison can be made between the micelle formation at the interfaces and the micelle formation in the bulk liquid phase: at 298.15 K, the enthalpy of the formation of full-cylindrical aggregates on the surface of gold is -11 kJ‚mol-1 (present work), as compared with -17 kJ‚mol-1 for the formation of half-cylindrical aggregates on the surface of graphite,6 -7 kJ‚mol-1 for the formation of
Kira´ ly et al.
globular surface aggregates on the surface of hydrophilic silica,10 and -1.2 kJ‚mol-1 for the formation of spherical micelles in aqueous bulk solution.42 Surface aggregation on gold is induced at a lower surface coverage than that on graphite (Figure 7): this observation may be related to the different orientations of the surfactant molecules in the adsorbed monolayer on the two surfaces. The theoretical considerations of Johnson and Nagarajan suggested that, at low solution concentrations, the transition from a vertical monolayer (with headgroups in contact with the surface) to full surface cylinders may proceed through a series of composite structures (hemicylinder; hemisphere; bilayer; finite disk) that cannot be resolved and identified by AFM or other currently available structure-sensitive experimental techniques.16 In fact, AFM experiments are usually conducted near the bulk cmc, which is well above the concentration regime in which the various microstructures predicted by theory may be formed. For the composite structures, immediate surface saturation on the top layer is a requisite condition, since any aggregate core-water contact is energetically unfavorable. It has further been argued16 that, with increasing solution concentration, the adsorption of even a small amount of surfactant can induce morphological transformations between the composite structures on the surface. The results of the present study are consistent with a gradual transition between the two major morphologies: the (loosely packed) vertical monolayer far below the cmc, and the full surface cylinders close to the cmc. Morphological transformations may well occur in the intermediate solution concentration region of 0.1-6 mM (Figure 6), corresponding to the surface concentration region of 0.8-1.6 µmol‚m-2 in which -∆ah drops from 50 to 11 kJ‚mol-1 (Figure 7). The formation of full-cylindrical aggregates on gold are induced by a head-to-surface vertical monolayer. The half-cylindrical aggregates on the surface of graphite are templated by a headto-head, tail-to-tail horizontal monolayer. There is a delicate difference between the driving forces of the formation of the two kinds of aggregate morphologies, and adsorption microcalorimetry per se cannot distinguish between them. In fact, the present microcalorimetric study relies on theoretical considerations16 and the results of earlier AFM studies19 on closely similar systems. A combination of the two complementary experimental techniques, adsorption calorimetry and AFM, seems to be a powerful way to study the adsorption phenomena at solid/solution interfaces.
Conclusions The adsorption of C12TABr from aqueous solutions on gold is a three-stage process. Initially, Br- is irreversibly (chemically) bound to low-coordination surface sites of the gold microparticles, and the AuBr- surface species are neutralized by H+. The surface speciation process dramatically increases the pH of the bulk solution, due to the generation of C12TAOH in the aqueous phase. In the second step, C12TABr molecules are reversibly (physically) adsorbed on high-coordination surface sites in a head-to-surface orientation. In the third adsorption step, the surfactant monolayer induces the formation of full-cylindrical surface aggregates through a series of composite structures. As for C12TABr, the initial adsorption of Br- on gold from a dilute aqueous solution of NaBr increases the pH in the bulk liquid phase. An electric double layer of the type AuBr-/H+ is formed and the surface speciation is further accompanied by the production of a large amount of OH- in the free solution. It is (42) Bashford, M. T.; Woolley, E. M. J. Phys. Chem. 1985, 89, 3173-3179.
Adsorption of C12TABr and NaBr on Gold
anticipated that more information about the present systems can be obtained by the use of a series of low dead-volume flowthrough ion-selective detectors, including H+, Na+, Br-, and surfactant (C12TA+)-sensitive electrodes. Such an extended study is currently being undertaken in our laboratory.
Langmuir, Vol. 22, No. 7, 2006 3213
Acknowledgment. This work was supported by the Bolyai Ja´nos Foundation, Hungary; the Alexander von Humboldt Foundation, Germany; and the Hungarian Scientific Research Fund (OTKA T042521). LA053184+