Investigation of the Stability of Graphite Particle Dispersion and the

AFM images were obtained using an SPA 400 (Seiko Instruments Co., Japan) instrument. Crystalline graphite (HOPG) was glued to the microscope slides us...
7 downloads 0 Views 223KB Size
16746

J. Phys. Chem. B 2004, 108, 16746-16752

Investigation of the Stability of Graphite Particle Dispersion and the Hemimicelle Formation Process at Graphite/Solution Interfaces Using Atomic Force Microscopy Hideya Kawasaki,* Kazuya Ban, and Hiroshi Maeda Department of Chemistry, Faculty of Science, Kyushu UniVersity 33, Hakozaki, Higashi-ku, Fukuoka, 812-8581, Japan ReceiVed: March 12, 2004; In Final Form: August 20, 2004

The correlation between the dispersion stability of graphite particles suspended in aqueous solutions of dodecyldimethylhydroxylammonium chloride (C12DMAOH‚Cl) and the structure of the adsorption layer on the graphite surface was investigated by a combination of data from AFM images, measurements of interaction forces, an adsorption isotherm, and turbidity measurements of graphite suspensions. A model for the twostep adsorption mechanism was successfully applied to describe the adsorption isotherm. (The first step was the formation of the horizontal adsorbed monolayer, and the second step was the formation of the hemicylindrical aggregate.) It was suggested that there is a transition regime where the horizontal adsorbed monolayer coexists with the hemicylindrical aggregate. A dispersion of graphite particles is unstable in the presence of the horizontal adsorbed monolayer. Surface force measurements suggest that the surface charges of the horizontal adsorbed monolayer are mostly neutralized by the bound counterions. The critical hemimicelle concentration (hmc) was about 0.8 mM, which was about 1/10 of the bulk critical micelle concentration (cmc ≈ 6 mM). The hmc value was close to the critical concentration at which the suspension stability started to increase dramatically with the surfactant concentration. We found that a stable dispersion of graphite particles can only be achieved once the hemicylindrical aggregates have formed on the surface at full surface coverage, leading to the development of the electric double layer forces from the positively charged aggregate.

Introduction It is well known that surfactant adsorption to solid surfaces and surfactant self-assembly at interfaces enhance the stability of particle dispersions. The self-assembly at interfaces has been employed in a wide range of industrial processes such as flotation, lubrication, wetting, and paint technology. Control of particle dispersions depends on the properties of the selfassembly at interfaces. Compared to the studies on self-assembly in bulk solutions,1 self-assembly at interfaces is not well understood. Gaudin and Fuerstenau first proposed the concept of hemimicelles in 1955.2 The concentration at which a rapid increase in the adsorption occurs is known as the critical hemimicelle concentration (hmc). Traditionally, adsorption isotherms combined with techniques such as contact angles and ζ potentials have been used to investigate the self-assembly behavior of surfactants at solid/solution interfaces, leading to many important insights into hemimicellization processes, critical aggregation numbers, and adsorption kinetics.3-5 A major disadvantage of these techniques, however, is the lack of information on the structures of adsorbed layers. Atomic Force Microscopy (AFM) has provided a new insight into the structures of micellar aggregates at solid/solution interfaces6-8 besides those provided by reflection techniques and spectroscopy.5 Reviews about the micellar aggregates at solid/solution interfaces have been published.5,9,10 Manne et al.6 demonstrated the first direct observation of hemimicelles at graphite (crystalline hydrophobic surfaces)/solution interfaces by AFM. The AFM studies indicated that the dominant aggregate structures were hemicylinders. Cationic, anionic, zwitterionic, and nonionic surfactants with an alkyl chain of * Corresponding author. Tel: 81-92-642-4367. Fax: 81-92-642-2607. E-mail: [email protected].

the dodecyl-group length or larger all form hemicylindrical aggregates on graphite,8-15 in contrast to the observations for various types of micellar aggregates at mica/solution interfaces.5,9-11 It has been proposed that the graphite surface templates the formation of hemicylindrical aggregate because of a good fit between the surfactant alkyl chain and the graphite lattice.6,16,17 The horizontal adsorbed monolayer on the graphite surface has a head-to-head and tail-to-tail arrangement, which templates further adsorption and finally leads to the formation of the hemicylindrical aggregate on graphite. This process is supported by a thermodynamic analysis of calorimetric enthalpies.18,19 As is the case for smooth amorphous hydrophobic surfaces, several surfactants formed globular structures20 or laterally unstructured layers21,22 at the hydrophobic solid/solution interfaces. Thus, the hydrophobicity of the graphite surface is also important in the formation of the hemicylindrical aggregate. Many investigations of the structures formed by adsorbed surfactants at the graphite/solution interface have been reported,5,8-14 but they have yielded little information on the relationship between the dispersion stability of graphite colloidal particles and the aggregate morphology of the adsorbed surfactant. To investigate the dispersion stability of graphitized carbon particles in the surfactant solutions, we used AFM techniques coupled with an adsorption isotherm in this study. We examined the adsorption isotherm of dodecyldimethylhydroxylammonium chloride (C12DMAOH‚Cl) on graphitized carbon blacks, yielding information about the formation process of the hemicylindrical aggregate at the graphite/solution interface. Using the AFM images and the surface force measurements, we investigated the morphology of the adsorbed layer and the forces between the adsorbed layers. The AFM results were analyzed in combination with the adsorption isotherm.

10.1021/jp0488970 CCC: $27.50 © 2004 American Chemical Society Published on Web 10/01/2004

Stability of Graphite Particle Dispersion

J. Phys. Chem. B, Vol. 108, No. 43, 2004 16747

With this knowledge, the dispersion behavior of graphite particles as a function of the surfactant concentration is explained. Experimental Section Materials. Water was prepared by distillation and then passed through an ultrapure water system consisting of an ion exchanger, an activated carbon cartridge, and a 0.2-µm filter (Barnstead Co.). The resulting water had a resisitivity of 18 MΩ cm at 25 °C. Dodecyldimethylamine oxide (C12DMAO) (GERBU Co.) was recrystallized three times from hot acetone. After recrystallization, there was a single peak in the high-performance liquid chromatography (HPLC) chromatograms that were produced using an ODS-120T (Tosoh Co., Japan) column (MeOH/ H2O ) 7/3). Dodecyldimethylhydroxylammonium chloride (C12DMAOH‚Cl) was obtained by mixing equimolar solutions of nonionic C12DMAO and HCl and then freeze-drying. Methods. (a) AFM Imaging. AFM images were obtained using an SPA 400 (Seiko Instruments Co., Japan) instrument. Crystalline graphite (HOPG) was glued to the microscope slides using epoxy glue with a low melting temperature and then freshly cleaved with adhesive tape in a laminar flow bench. Prior to use, the cantilevers were irradiated with ultraviolet light (8 mW/cm2 at 253.7 nm) in a laminar flow bench for about 20 min and then washed with distilled ethanol. The C12DMAOH‚ Cl surfactants were adsorbed onto the surface of the freshly cleaved graphite by immersing the substrate in surfactant solutions of desired concentrations for at least 12 h. AFM images were obtained for the substrates in the surfactant solutions at 25 ( 2 °C in both tapping and soft contact modes. In the soft contact mode, deflection images (showing the error in the feedback signal) were collected with scan rates of 5-10 Hz using silicon nitride tips (Olympus Co., Japan) with a nominal spring constant of 0.09 N m-1 and a fluid cell. No filtering of images was performed other than that inherent in the feedback loop. Zero distance in the force curves was determined from a constant high gradient of force where the tip is assumed to be touching the substrates. Most AFM imaging was performed using a repulsive force from the adsorbed films with the tip separated from the substrates by about 1-3 nm (so-called soft contact imaging). In the tapping mode, we used cantilevers with nitride tips (Olympus Co., Japan) with a small spring constant of 0.09 N m-1 to observe surface aggregates at the graphite/ solution interface. Tapping-mode imaging conditions were as follows: scan rate, 0.5-1.5 Hz; 512 points collected per line; integral gain at 0.2 and proportional gain at 0.1; and resonance frequency 8-9 kHz in the aqueous solutions. The amplitude set point was 0.5 V. (b) Surface Force Measurement. Forces between a HOPG plate and a carbon sphere were measured on an atomic force microscope SPA 400 (Seiko Instruments) following the procedures by Ducker et al.23,24 A carbon sphere of 5-10 µm in diameter (supplied by Ishicawa Carbon Co., Japan) was used as the colloidal probe and was attached to the top of a rectangular cantilever that was 100 µm in length (RC800 PSA, Olympus) using an epoxy resin (Epikote 1004, Shell), and force-distance profiles between the HOPG plate and the carbon sphere in aqueous solutions were recorded. The HOPG plate was mounted on the scanner table. Before gluing to the cantilever, the carbon sphere was cleaned in distilled ethanol. The spring constant (k) of each cantilever was determined by the resonance method25 and was found to be around 1 N m-1. The measured force (F) was normalized using the radius of the sphere (R) to give F/R which is proportional to the interaction

Figure 1. Adsorption isotherm for C12DMAOH‚Cl (pH 2) at 25 °C on graphitized particles showing the four regimes of surfactant adsorption behavior.

Figure 2. Adsorption isotherm for C12DMAOH‚Cl (pH 2) at 25 °C on graphitized particles according to eq 1. Experimental data Γ. Solid curve from eq 1 Γmax ) 3.95 µmol/m2, k1 ) 3 × 105, k2 ) 1 × 106, and n ) 3.3.

energy between the corresponding flat surfaces (Gf) by F/R ) 2πGf.26 The radius of the carbon sphere was determined by the reverse imaged procedures proposed by Craig et al.27 The AFM images of the carbon sphere (500 × 500-nm2 region) indicated that the rms roughness of the typical surface roughness is 1.3 nm. (c) Suspension Stability. The suspension stability was determined by measuring the turbidity of 2.5 wt % suspensions of graphitized carbon particles purchased from Tokai Carbon Co. in Japan. The solutions containing the suspensions of graphitized particles were stirred gently for 30 min, and then the turbidity of the supernatant of the solution was measured after 20 h of settling, using a Jasco Ubest-50 UV-vis spectrophotometer at 400 nm. Stable suspensions showed small decreases in the turbidity after 20 h of settling, whereas the turbidity of unstable suspensions decreased significantly because of rapid sedimentation or coagulation. (d) Adsorption. The adsorption of C12DMAOH‚Cl surfactants onto graphitized carbon particles (Tokai Carbon Co.) was measured by solution depletion methods. The mean particle diameter of the particles was about 70 nm, as determined by transmission electron microscopy, and the surface area, as determined by nitrogen adsorption, was 30 m2/g. A known concentration of surfactant was mixed with the particle suspensions, equilibrated for a week, and then centrifuged or filtered to separate the supernatant from the surfactant-coated particles. Residual surfactant concentrations of the solutions were then measured by the cationic surfactant-selective electrode according to the method of Takisawa et al.28 When the sample concentration was above the cmc, the sample was diluted to a concentration less than the cmc. Results and Discussion Adsorption Isotherm. Figure 1 shows the adsorption isotherm of C12DMAOH‚Cl on the surface of graphitized particles in the aqueous solutions at 25 °C. In Figure 3, the isotherm data is plotted against the logarithmic scale of the surfactant

16748 J. Phys. Chem. B, Vol. 108, No. 43, 2004

Kawasaki et al. surface to form horizontal monolayers. It is assumed that interactions among adsorbed surfactant molecules are negligible. The equilibrium constant of the following reaction

site + monomer S (monomer-site) complex

(1)

is denoted by

a1

k1 )

(2)

(asa)

where a is the activity of the surfactant monomers in the solution (for dilute solutions a ) C, the surfactant concentration) and a1 and as are activities of the complex and the free surface sites, respectively. In the second step, the adsorption increases dramatically through the hydrophobic interaction to form hemicylindrical aggregates, and each of the surfactant molecules of horizontal monolayers provides a possible active site for the hemicylindrical aggregate. The equilibrium constant of the following reaction Figure 3. Schematic diagram of the horizontal monolayer on the graphite surface.

concentration. The pH values of the solutions were fixed to be 2 because, at this pH value, the C12DMAOH‚Cl molecules behave as cationic surfactants. The shape of the adsorption isotherm indicates a two-step type adsorption. Four regions in the adsorption isotherm may be distinguished. At low surfactant concentrations, less than 2 × 10-5 M (regime I), the amount of C12DMAOH‚Cl adsorbed increases gradually with the surfactant concentration Cd, reaching Γ1 ≈ 1.2 µmol/m2 at around Cd ) 2 × 10-5 M. Gradual adsorption continues in the Cd range of (1-6) × 10-4 M (regime II), and the isotherm shows a significant increase above Cd ) 8 × 10-4 M (regime III), reaching a saturation level of Γmax ≈ 3.95 µmol/m2 around Cd ) 3 × 10-3 M (regime IV). As is the case for the horizontal adsorption of a surfactant molecule to a graphite surface as shown in Figure 2, Groszek proposed a geometric model where each methylene group occupies one hexagon (0.0524 nm2), and a methyl group occupies two hexagons on the graphite basal plane.18,29,30 Assuming that the occupation area of a dimethylamine oxide headgroup is 5 hexagons (i.e., 1 hexagon for the nitrogen atom and 4 hexagons for the 2 methyl groups, with the hydroxyl group being exposed to the aqueous phase) and that the occupation area of the dodecyl tail group is 13 hexagons, the area occupied by a C12DMAOH molecule is 0.94 nm2. The geometrical model gives a value of Γ1 ) 1.77 µmol/m2 for the horizontal adsorbed monolayer of C12DMAOH‚Cl at full surface coverage on graphite. This calculated Γ1 value is close to the experimental Γ1 value (∼1.2 µmol/m2), although the geometric model overestimates it somewhat. The experimental Γ1 value is very close to the reported values of Γ1 ) 1.2-1.3 µmol/m2 for the surface coverage on the formation of the horizontal adsorbed monolayer of dodecyltrimethylammonium bromide (C12TAB) on graphite.18 It is likely that the horizontal adsorbed monolayer of C12DMAOH‚Cl is formed in the first step of adsorption (regime I). In the present study, we apply the two-step adsorption isotherm equation by Zhu et al.31,32 to our experimental adsorption isotherm. The basic theoretical assumption is that the adsorption of surfactants at solid/solution interfaces occurs in two steps. Here, we consider that in the first step surfactant molecules adsorb onto the graphite surface through the interactions between individual surfactant molecules and the solid

(n - 1) monomers + (monomer-site) complex S hemicylindrical aggregate (3) is denoted by

k2 )

ahm

(4)

(a1an - 1)

where ahm is the activity of hemicylindrical aggregates. The aggregation number of surface aggregates per site is denoted as n. As an approximation, the amount of adsorbed monomer, Γ1, the amount of hemicylindrical aggregate, Γhm, and the number of sites, ΓS, can be used instead of a1, ahm, and as, respectively. Equations 2 and 4 may be given by

Γ1

k1 ) k2 )

(5)

(ΓsC) Γhm

(6)

(Γ1Cn - 1)

The following equation can be derived by combining eqs 5 and 6

Γ ) Γ1 + nΓhm

(7)

Γmax ) n(Γ1 + Γhm + Γs)

(8)

where Γ is the amount of surfactant adsorbed at concentration C, and Γmax is the limiting adsorption at high concentration. On the basis of the two-step adsorption mechanism and the mass action treatment, the following adsorption isotherm is obtained31,32

Γ)

(n1 + k C )

Γmaxk1C

n-1

2

1 + k1C(1 + k2Cn - 1)

(9)

Both Γmax and Γ1 can be obtained from the adsorption data (Γmax ≈ 3.95 µmol/m2 and Γ1 ≈ 1.2 µmol/m2). In the case of hemicylindrical aggregate, the aggregation number N is equal to 2n. The n value may be obtained (n ) 3.3) by using the following equation18

Stability of Graphite Particle Dispersion

n)

Γmax Γ1

J. Phys. Chem. B, Vol. 108, No. 43, 2004 16749

(10)

As shown in Figure 3, the theoretical curve given by eq 9 represents the experimental data well. The aggregation number per cross section in the hemicylindrical aggregate is estimated to be 7 because the N value () 2n) is 6.6. This value indicates that the surfactant packing parameter (P) is 0.48 (Appendix), which fulfills the condition 1/3 < P < 1/2 for the formation of hemicylindrical aggregates on solid surfaces.33 The hmc is defined as the concentration at which the adsorption increases dramatically due to the formation of the hemimicelle at solid/solution interfaces.32 Accordingly, the hmc value may be determined from the inflection point of the isotherm in the second step. However, the hmc value that was estimated from the adsorption isotherm is empirical and somewhat arbitrary. Therefore, we estimated the hmc value from the theoretical adsorption isotherm of C12DMAOH‚Cl, as shown in Figure 3. In this way, the hmc value is estimated to be 0.8 mM. AFM Imaging. To examine the formation of the hemicylindrical aggregate above the hmc, we observed the morphology of the adsorption layer at the graphite/solution interface as a function of the surfactant concentration, using soft contact AFM imaging. The AFM images indicated that the hemicylindrical aggregates form on the surface at full surface coverage at 4 mM, which is above the hmc (Figure 4b), whereas the AFM images are featureless at 0.5 mM, which is below the hmc (Figure 4a). The period of the structure of the hemicylindrical aggregates was 5.8 ( 0.2 nm. It is likely that the hemicylindrical aggregates of C12DMAOH‚Cl form on the graphite in the process of the cooperative adsorption of regime III. It has been reported that there is a transition regime between the formation of the hemicylindrical aggregate and the formation of the horizontal adsorbed monolayer.18,19 It was proposed that, in the transition regime, the formation of the hemicylindrical aggregate can occur before the completion of the horizontal adsorbed monolayer.19 In this study, it is suggested that regime II in the adsorption isotherm of C12DMAOH‚Cl corresponds to the transition regime. The isotherm data plotted on the logarithmic scale (Figure 3) shows the existence of regime II. To examine the morphology of the adsorption layer in the transition regime (regime II), we performed AFM imaging on the adsorbed layers in soft contact mode, but it was difficult to obtain AFM images of the layer structures due to the absence of a large repulsive force in regime II; therefore, we used tapping-mode AFM to examine the morphology of the adsorption layer in regime II, using the cantilever with a small spring constant of 0.09 N m-1. The tapping-mode AFM images indicated the coexistence of featureless parts and a fragment of the parallel stripes characteristic of hemicylindrical aggregates at 2.5 × 10-4 M, as shown in Figure 5a. The corresponding force curves between the AFM tip and the graphite surface show that the interaction is usually attractive but that its values depend on the position of the area (Figure 5b), suggesting inhomogeneity of the surface aggregates on the graphite. These AFM images and the force curves indicate that the horizontal adsorbed monolayer coexists with the hemicylindrical aggregate in regime II (i.e., the transition regime). From the viewpoint of the gradual adsorption in the isotherm, the structural change from the horizontal adsorbed monolayer to the hemicylindrical aggregate may occur gradually in regime II. The calculated Γ1 value (∼1.77 µmol/m2) from the geometric model for the close-packed horizontal adsorbed monolayer is

greater than the Γ1 value (∼1.2 µmol/m2) from the adsorption isotherm, suggesting the absence of close-packing in the horizontal monolayer. A portion (ca. 30%) of the graphite surface may remain covered by water molecules or by the counter chloride ions.18,19 In fact, the AFM images indicated periodicities of hemicylindrical aggregates (5.8 ( 0.2 nm) larger than twice the length of the fully extended ionic surfactant molecules (4.1 nm).34 Taking into account the intermicellar distance(1.7 nm) of the hemicylindrical aggregates, the calculated Γ1 value for the horizontal adsorbed monolayer is corrected to 1.25 µmol/m2 (4.1/5.8 × 1.77 ) 1.25). This value is close to the Γ1 value (∼1.2 µmol/m2) from the adsorption isotherm. From the AFM images and the adsorption isotherm, it is concluded that the first step in the adsorption isotherm (regime I) corresponds to the formation process of a horizontal monolayer on the graphite surface. The horizontal adsorbed monolayer templates further adsorption, and cooperative formation of hemicylindrical aggregates occurs in the second step (regime III). There is a transition regime (regime II) between the hemicylindrical aggregate (regime III) and the horizontal adsorbed monolayer (regime I). The hmc value is estimated to be 0.8 mM from the AFM images and the adsorption isotherm, which is about 1/10 of the bulk cmc (∼6 mM) determined from the surface tension measurement.35 It should be noted that the hmc value proposed by the present study (hmc/cmc ≈ 0.1) is much higher than those of other surfactant organizations at hydrophobic surface/water interfaces. For the adsorption of ionic surfactants with a dodecyl chain, for example, a recent model predicts that the hmc is 2 orders of magnitude less (hmc/cmc ) 0.002) than the cmc.36 Additionally, the hmc/cmc value was reported to be less than 0.017 in the organization of nonionic decyldimethylamine oxide at the graphite/water interface at its natural pH.37 These different hmc/cmc values may originate from the uncertainty in the position of the hmc on the adsorption isotherm. In the transition regime (i.e., regime II), the hemicylindrical aggregate can occur before the completion of the horizontal adsorbed monolayers. If the hmc value is defined as the concentration at the onset of the transition regime, then the hmc value of the C12DMAOH‚ Cl hemicylindrical aggregate is about 10-4 M (hmc/cmc ≈ 0.016). Suspension Stability and Surface Force Measurement. We are interested in the influence of the surfactant adsorption processes of the hemicylindrical aggregates on the stabilization of graphite particle dispersion. Figure 6 shows the suspension turbidity, a measure of stability, of graphite particles as a function of C12DMAOH‚Cl concentration at pH 2, together with the corresponding adsorption isotherm. It is observed that the suspension stability of graphite particles starts to increase dramatically above a certain concentration of C12DMAOH‚Cl (∼1 mM), which is close to the hmc (0.8 mM). A stable dispersion of graphite particles can be attained at surfactant concentrations above 4 mM, where the hemicylindrical aggregate forms on graphite at full surface coverage, as seen in the AFM image (Figure 4b). The results indicate that the critical concentration of the suspension stability corresponds to the hmc. The improvement in the dispersion stability above the hmc is clearly a result of the formation of the hemicylindrical aggregates of C12DMAOH‚Cl on the surface of graphite particles and the consequent enhancement of the repulsion of the graphite particles. We emphasize the fact that the surfactant adsorption on the surface of graphite particles below the hmc cannot produce a stable suspension, despite the formation of the horizontal adsorbed monolayer.

16750 J. Phys. Chem. B, Vol. 108, No. 43, 2004

Kawasaki et al.

Figure 4. AFM images showing the aggregate structure of cationic C12DMAOH‚Cl at the graphite/solution interface (200 × 200 nm2). The surfactant concentrations are (a) 0.5 mM and (b) 4 mM.

Figure 5. (a) AFM images showing the aggregate structure of cationic C12DMAOH‚Cl at the graphite/solution interface at a surfactant concentration of 2.5 × 10-4 M. The area of A indicates the horizontal monolayer, whereas the area of B shows hemicylindrical aggregates on graphite. The bar corresponds to 50 nm. (b) Force profiles between the graphite plate (HOPG) and the silicon nitride tip at a surfactant concentration of 2.5 × 10-4 M. Data from two independent measurements are shown.

To interpret this result, we measured surface forces between a graphite (HOPG) plate and a hydrophobic carbon sphere at different concentrations of the C12DMAOH‚Cl surfactant at pH 2. In degassed water, the force-distance curves between the graphite plate and the carbon sphere exhibit purely attractive interactions with a jump around 10 nm to contact as the force gradient exceeds the spring constant (k ) 17 N m-1) as shown in Figure 7a. A large pull-off force was observed (50-200 mN m-1). The attraction is considerably larger than that expected from the van der Waals forces alone (solid line), implying that a long-range hydrophobic force is present. The presence of the long-range hydrophobic attraction also suggests that any surface charge is low. These results are in accord with previous reports on hydrophobic bulk polymer surfaces in degassed NaCl solutions.38 Figure 7b shows the force-distance curves between the graphite plate and the carbon sphere in 0.5 mM (< HMC) and 4 mM (> HMC) aqueous solutions of C12DMAOH‚Cl at pH 2. Repeated measurement cycles at the same point on the surface, which had a number of surface contacts, did not alter the force profile, demonstrating that the C12DMAOH‚Cl

Figure 6. Turbidity (transmittance at 400 nm) of graphitized particles after 20 h of settling in the C12DMAOH‚Cl solutions at pH 2 as a function of concentration. The corresponding adsorption isotherm is also shown. The critical concentration of the suspension stability is close to the hmc.

Stability of Graphite Particle Dispersion

J. Phys. Chem. B, Vol. 108, No. 43, 2004 16751

Figure 7. (a) Force profiles between the graphite plate (HOPG) and the hydrophobic carbon sphere in degassed water. Data from two independent measurements are shown. The solid curve is the theoretical curve of the van der Waals forces with a Hamaker constant of H ) 3.3 × 10-20 J, assuming the H value for graphite/water/graphite. (b) Force profiles between the graphite plate (HOPG) and the hydrophobic carbon sphere in the C12DMAOH‚Cl solutions of 0.5 mM and 4 mM at pH 2. The solid curve is the fit DLVO theoretical curve at a constant surface potential with the following parameters: surface potential φ0 ) 35 mV, Debye length κ-1 ) 3.1 nm, radius of the carbon sphere R ) 6 µm, and Hamaker constant H ) 3.3 × 10-20 J, assuming the H value for graphite/water/graphite.

monolayers were stable enough for the force measurements. In the 4 mM solution, where the hemicylindrical aggregates are formed at full surface coverage on the graphite surface, we observed a long-range repulsion described by an exponential function. The observed forces correspond to the electric double layer forces. The obtained decay length (3.1 nm) agrees well with the Debye length (κ-1) for the corresponding salt concentration, 2.6 nm. It is clear that the electric double layer forces are attributed to positively charged hemicylindrical aggregates. The surface potential φ of the hemicylindrical aggregates was calculated from the double layer forces, assuming a constant surface potential of 35 mV. The effective surface charge density of the hemicylindrical aggregates that was estimated from the measured force profile is 8.5 × 10-3 C/m2. This effective surface charge density is noticeably smaller than the surface charge density (∼3.7 × 10-1 C/m2) estimated from the value (ΓmaxF) from the adsorption isotherm. Here, F is the Faraday constant. This indicates that the surface charges of the hemicylindrical aggregates are mostly neutralized by bound counterions. In the 0.5 mM solution, where the close-packed horizontal adsorbed monolayers are formed at full surface coverage on the graphite surface, however, no electric double layer repulsion described by an exponential function as shown in Figure 7b is observed, although a high surface charge density (∼1.2 × 10-1 C/m2) is estimated from the value (Γ1F) from the adsorption isotherm. This suggests that the effective surface charge density in the horizontal adsorbed monolayer is extremely low. In the case of the horizontal adsorbed monolayers, it is likely that most of the surface charges are neutralized by bound counterions. As a result, the formation of the horizontal adsorbed monolayer is not sufficient to produce the stable suspension of graphite particles, as shown in Figure 6. In scanning tunneling microscopy (STM) studies, the STM images showed the direct observation of the horizontal adsorbed monolayers arranged in a head-to-head configuration on graphite for hexadecyltrimethylammonium bromide (C16TAB) and dodecyltrimethylammonium bromide (C12TAB).39 Regarding the close proximity of ionic headgroups in the horizontal adsorbed monolayers, the following two mechanisms have been proposed by Manne et al.:6 (1) the surface charge is neutralized by bound counterions and (2) attractive forces are present between the cationic headgroup and semiconductive graphite. Our data for the surface force between the horizontal adsorbed monolayers supports the proposition that the surface charges are neutralized by bound counterions. We note that the pull-off force is 0.7 ( 0.3 mN m-1 at 0.5 mM, which is considerably smaller than that observed

in water (50-200 mN m-1). Because of the presence of the hydrophilic headgroups in the horizontal adsorbed monolayers on graphite, the hydrophobicity of the substrate may decrease. In this context, the studies on suspensions of surfactantstabilized carbon nanotubes (CNT) are very interesting.40 Recently, it has also been reported that sodium dodecyl sulfonate (SDS) forms hemicylindrical aggregates on the surface of carbon nanotubes (CNT).41 The suspension stability of CNT in aqueous solutions was investigated with various types of surfactants and polymers such as sodium dodecylbenzene sulfonate (NaDDBS), SDS, Triton X-100, C12TAB, dextrin, and poly(styrene)-poly(ethylene oxide) (PS-PEO) diblock copolymer.40 It was found that the NaDDBS with a benzene ring moiety enhances the stability of CNT in water by a factor of the order of 10 to 100 compared to that of other surfactants and polymers examined. The NaDDBS-stabilized CNT remained dispersed for at least 3 months; neither sedimentation nor aggregation of nanotube bundles was observed. In the present study, we found that the formation of horizontal adsorbed monolayers on a graphite surface remarkably decreases the attractive interaction between the hydrophobic surfaces as well as the pull-off force (adhesive force) in the aqueous solutions. Thus, it is considered that the horizontal adsorbed monolayers stabilized by the π-π interaction of the benzene rings with the graphite surface weaken the attractive interactions between the CNT tubes. In this context, the stabilization of the horizontal adsorbed monolayers on the graphite surface may play an important role in preventing the formation of CNT bundles. Summary In the present study, AFM techniques (AFM imaging and surface force measurements) coupled with an adsorption isotherm were used to investigate the dispersion stability of graphitized carbon particles in surfactant solutions as a function of the concentration. It was found that the adsorption isotherm could be described by a two-step mechanism, taking into account the formation of hemicylindrical aggregates. The aggregation number per cross section of the hemicylindrical aggregates was estimated to be 7, which fulfills the packing parameter condition for solid surfaces. The horizontal monolayer is formed on the graphite surface in the first step of adsorption (regime I). The formation of horizontal adsorbed monolayers remarkably decreases the attractive interaction between the hydrophobic surfaces and the pull-off force (adhesive force) in the aqueous solutions; however, it is not sufficient to produce the stable suspension of graphite

16752 J. Phys. Chem. B, Vol. 108, No. 43, 2004 particles, despite a relatively large amount of adsorbed surfactant (∼1.2 µmol/m2). It is suggested that the surface charges of the horizontal adsorbed monolayer are mostly neutralized by bound counterions. It is proposed that there is a transition regime (regime II) between the formation of the horizontal adsorbed monolayer and that of the hemicylindrical aggregate, where horizontal adsorbed monolayers coexist with hemicylindrical aggregates. From the viewpoint of the gradual adsorption in the isotherm, the structural change from horizontal adsorbed monolayers to hemicylindrical aggregates may occur gradually in regime II. The cooperative formation of hemicylindrical aggregates occurs in the second step of adsorption (regime III). The hmc was estimated to be 0.8 mM, which is about 1/10 of the bulk cmc (∼6 mM). The hmc value corresponds to the critical concentration at which the suspension stability starts to increase dramatically with the surfactant concentration due to the development of the electric double layer forces from the positively charged hemicylindrical aggregates on graphite. Stable dispersion of graphite particles can be obtained only at surfactant concentrations higher than 4 mM, where the formation of the hemicylindrical aggregate occurs at full surface coverage. Appendix Surfactant Packing Parameter for Hemicylindrical Micelles at Graphite/Solution Interface. The packing parameter (P) is a leading concept in the treatment of self-assembly in bulk solutions.26 It has been reported that the packing parameter should fulfill the condition 1/3 < P < 0.5 for the formation of hemicylindrical aggregates on solid surfaces.33 Here, we consider the packing parameter for the hemicylindrical aggregate of C12DMAOH‚Cl at the graphite/solution interface. The surfactant packing parameter is defined as

P)

V (A0R)

where R is defined as the radius of the core of the hemicylindrical aggregate, which is roughly equal to the fully extended length of the hydrocarbon chain, lc, V is the volume of the hydrocarbon tail of the C12DMAOH‚Cl molecule, and A0 is the area per surfactant of the aggregate at the hydrocarbon/water interface. To simplify matters, we assumed that the cross section of the headgroup is circular with a diameter D

A0 ) π

(D2 )

2

D)

πlc (N ) 7) N

where N is the aggregation number per cross section of the hemicylindrical aggregate. In the case of C12DMAO H‚Cl1, V ) 0.35 nm3 and R ≈ lc ) 1.67 nm, thus

D ) 0.75 nm, A0 ) 0.44 nm2, and P ) 0.48 Acknowledgment. We thank Dr. H. Katsuura for giving us methods for the cationic surfactant-selective electrode. This work was partially supported by a Grant-in-Aid for Scientific Research

Kawasaki et al. (no.15750121) from the Monbukagaku-shou Japan, the CREST of Japan Science and Technology Corporation (JST). References and Notes (1) (a) Hunter, R. J. Foundations of Colloid Science; Clarendon Press: Oxford, U.K., 1986. (b) Wennerstrom, H.; Evances, D. F The Colloidal Domain: Where Physics, Chemistry, Biology, and Technology Meet , 2nd ed.; Wiley-VCH: New York, 1999. (2) Gaudin, A. M.; Fuerstenau, D. W. Trans. AIME 1955, 202, 958. (3) Adamson, A. W.; Gast, A. P. Physical Chemistry of Surfaces, 6th ed.; John Wiley & Sons: New York, 1997. (4) Hough, D. B.; Rendall, H. M. Adsorption of Ionic Surfactants. In Adsorption from Solutions at the Solid-Liquid Interfaces; Parfitt, G. D., Rochester, C. H., Eds.; Academic Press: London, 1983. (5) Atkin, R.; Craig, V. S. J.; Wanless, E. J.; Biggs, S. AdV. Colloid Interface Sci. 2003, 103, 219. (6) Manne, S.;Cleveland, J. P.; Gaub, H. E.; Stucky, G. D.;Hansma, P. K. Langmuir 1994, 10, 4409. (7) Manne, S.; Gaub, H. E. Science 1995, 270, 1480. (8) Wanless, E. J.; Ducker, W. A. J. Phys. Chem. B 1996, 100, 3207. (9) Warr, G. G. Curr. Opin. Colloid Interface Sci. 2000, 5, 95. (10) Tiberg, F.; Brinck, J.; Grant, L. Curr. Opin. Colloid Interface Sci. 2000, 4, 411. (11) Liu, Jun-Fu.; Ducker, W. A. J. Phys. Chem. B 1999, 103, 8558. (12) Wanless, E. J.; Ducker, W. A. Langmuir 1997, 13, 1463. (13) FitzGerald, P. A.;Warr, G. G. AdV. Mater. (Weinheim, Ger.) 2001, 13, 967. (14) Lamont, R.; Ducker, W. A. J. Colloid Interface Sci. 1997, 191, 303. (15) Kawasaki, H.; Syuto, M.; Maeda, H. Langmuir 2001, 17, 8210. (16) Grant, L. M.; Tiberg, F.; Ducker, W. A. J. Phys. Chem. B 1998, 102, 4288. (17) Rabe, P. J.; Buchhloz, S. Science 1999, 253, 424. (18) Kiraly, Z.; Findenegg, G. H. J.Phys. Chem. B 1998, 102, 1203. (19) Kiraly, Z.; Findenegg, G. H.; Klumpp, E.; Schlimper,H.; Dekany, I. Langmuir 2001, 17, 2420. (20) Wolgemuth, J. L; Workman, R. K; Manne, S. Langmuir 2000, 16, 3077. (21) Grant L. M.; Tiberg, F.; Ducker, W. A. J. Phys. Chem. B 1998, 102, 4288. (22) Grant, L. M.; Edeth, T.; Tiberg, F. Langmuir 2000, 16, 2285. (23) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Nature 1991, 353, 239. (24) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Langmuir 1992, 8, 1831. (25) Cleveland, J. P.; Manne, S.; Bocek, O.; Hansma, R. K. ReV. Sci. Instrum. 1993, 62, 403. (26) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: New York, 1992. (27) Neto, C.; Craig, V. S. J. Langmuir 2001, 17, 2097. (28) Takizawa, N.; Hall, D. G.; Wyn-Jones, E.; Brown, P. J. Chem. Soc., Faraday Trans. 1 1988, 84, 3059. (29) Liphard, M.; Glanz, P.; Pilarski, G.; Findenegg, G. H. Prog. Colloid. Polym. Sci. 1980, 67, 131. (30) Groszk, A. Proc. R. Soc. London, Ser. A 1970, 134, 473. (31) Bu-Yao Zhu.; Tiren, Gu. J. Chem. Soc., Faraday Trans. 1, 1989, 85, 3813. (32) Bu-Yao Zhu.; Tiren Gu. J. Chem. Soc., Faraday Trans. 1, 1989, 85, 3819. (33) Retter, U. Langmuir 2000, 16, 7752. (34) Garamus, V.; Kameyama, K.; Kakehashi, R.; Maeda, H. Colloid. Polym. Sci. 1999, 277, 868. (35) Imaishi, Y.; Kakehashi, R.; Nezu, T.; Maeda H. J. Colloid Interface Sci. 1998, 197, 309. (36) Johmson, R. A.; Nagarajan, R. Colloids Surf., A 2000, 167, 31. (37) Kiraly, Z.; Findenegg, G. H. Langmuir 2000, 16, 8842. (38) Meagher, L.; Craig, V. S. J. Langmuir 1994, 10, 2736. (39) Xu, S. L.; Wang, C.; Zeng, Q. D.; Wu, P.; Wang, Z. G.; Yan, H. K.; Bai, C. L. Langmuir 2002, 18, 657. (40) Islam, M. F.; Roajas, E.; Bergey, D. M.; Johnson, A. T.; Yodh, A. G. Nano Lett. 2003, 3, 269. (41) Richard, C.; Balavoine, F.; Schultz, P.; Ebbesen, T. W.; Miskowski, C. Science 2003, 300, 775.