Adsorption of Cationic Gemini Surfactants at Solid Surfaces Studied by

Publication Date (Web): May 24, 2011. Copyright © 2011 American Chemical ..... The silica surfaces were cleaned prior to use by simple UV ozone treat...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/Langmuir

Adsorption of Cationic Gemini Surfactants at Solid Surfaces Studied by QCM-D and SPR: Effect of the Rigidity of the Spacer Leila Mivehi, Romain Bordes,* and Krister Holmberg Chalmers University of Technology, Department of Chemical and Biological Engineering, SE-412 96 G€oteborg, Sweden ABSTRACT: Two small series of cationic gemini surfactants with dodecyl tails have been synthesized and evaluated with respect to selfassembly in bulk water and at different solid surfaces. The first series contained a flexible alkane spacer and is denoted 12-n-12, with n = 2, 4, and 6. The second series had a phenylene group connected to the quaternary nitrogens in either the meta or para position and the surfactants are referred to as 12-m-Φ-12 and 12-p-Φ-12, respectively. The phenylene group is a rigid linker unit. The critical micelle concentration (cmc) was determined both by tensiometry and by conductometry, and the packing density of the surfactants at the airwater interface was calculated from the Gibbs equation. The cmc values for the geminis with a rigid spacer, 12-m-Φ-12 and 12-p-Φ-12, were of the same order of magnitude as for 12-4-12, which is the flexible surfactant that most closely matches the phenylene-based surfactants with respect to hydrophobicity, measured as log P, and distance between the positively charged nitrogen atoms. The adsorption of flexible and rigid surfactants was investigated on gold, silicon dioxide (silica), gold made hydrophobic by the self-assembly of hexadecanethiol, and gold made hydrophilic by the self-assembly of 16hydroxyhexadecanethiol. On all of the surfaces, there was a reverse relationship between the adsorbed amount at the cmc and the length of the spacer (i.e., 12-2-12 gave the highest and 12-6-12 gave the lowest amount of adsorbed material). The adsorption pattern was similar for all of the surfactants when recorded at 25 °C. Thus, one can conclude that a rigid spacer does not render the self-assembly of a gemini surfactant difficult, neither in bulk water nor at solid surfaces. However, on one of the surfaces—untreated gold—the adsorbed amount of the geminis with a rigid spacer at 40 °C was approximately twice the values obtained at 25 °C. This is interpreted as the formation of an interdigitated bilayer at 25 °C and a regular bilayer without interpenetration of the alkyl chains at 40 °C.

1. INTRODUCTION Dimeric surfactants consisting of two amphiphiles connected by a spacer unit at, or close to, the polar headgroups are subject to considerable current interest. Such structures have been described in the literature for at least 40 years.1 In 1991, Menger coined the term “gemini surfactant”,2 and Menger,3 Zana,4 Rosen,5 and others later synthesized a wide range of such amphiphiles and characterized their physical chemical behavior. Particular attention has been directed toward cationic geminis with the two moieties linked by a short alkane spacer. The most fundamental difference in behavior between these geminis and their monomeric counterparts is that the dimeric surfactant selfassembles in the bulk at a concentration that is more than 10 times lower than the monomeric species with the same length of the hydrophobic tail. Thus, in Rosen’s terminology the geminis are much more “effective” than conventional surfactants.6 The majority of reported investigations of gemini surfactants concerns self-assembly and other solution properties.18 Less attention has been devoted to adsorption at solid surfaces, even if early studies were carried out on silica914 and on mica1517 mainly using the depletion method, AFM, or optical reflectometry. More recently, quartz crystal microbalance with dissipation r 2011 American Chemical Society

monitoring (QCM-D) has been employed to study the adsorption of gemini surfactants on different surfaces.1820 In the present work, we have used QCM-D in combination with surface plasmon resonance (SPR) to investigate the adsorption of a series of gemini surfactants on different surfaces. QCMD and SPR are often used in combination because they provide complementary information about adsorption. SPR measures the change in refractive index that occurs in the vicinity of the surface during adsorption. It provides information about the amount of dry mass that builds up on the surface. QCM-D is a gravimetric technique that measures the amount of material in the adsorbed layer (i.e., the values obtained include the hydration water or the entrapped water accompanying the adsorbate). An important limitation of SPR is that the substrate needs to have a plasmon band, which in reality limits the surface to gold, silver, and a few other metal surfaces. The vast majority of SPR studies have been performed on gold. QCM-D is much more flexible with respect to the surface and metallic surfaces, oxides, and Received: February 11, 2011 Revised: May 9, 2011 Published: May 24, 2011 7549

dx.doi.org/10.1021/la200539a | Langmuir 2011, 27, 7549–7557

Langmuir

ARTICLE

C NMR (100 MHz, CDCl3): δ 13.89, 22.45, 22.86, 26.1, 29.13, 29.33, 29.36, 29.43, 29.45, 31.68, 50.78, 56.57, 65.76. 2.1.3. N,N0 -Didodecyl-N,N,N0 ,N0 -tetramethyl-N,N0 -butanediyl-diammonium Dibromide [12-4-12]. N,N,N-Dodecyldimethylamine (9.39 g, 44 mmol, 10% excess) was dissolved in acetone (30 mL) in a roundbottomed flask. 1,4-Dibromobutane dissolved in 20 mL of acetone (4.32 g, 20 mmol) was added dropwise under stirring. The mixture was refluxed for 24 h and cooled to room temperature, and then a white solid was collected by filtration and washed with acetone. The product was recrystallized from a 1:9 mixture of ethanol and diethyl ether. Yield: 81% 1 H NMR (400 MHz, CDCl3): δ 0.86 (t, 6H), 1.191.30 (m, 32H), 1.34 (m, 4H), 1.75 (m, 4H), 2.14 (m, 4H), 3.27 (s, 12H), 3.41 (t, 4H), 3.94 (t, 4H). 13 C NMR (100 MHz, CDCl3): δ 13.90, 19.72, 22.47, 22.72, 26.2, 29.06, 29.12, 29.28, 29.31, 29.41, 31.69, 50.77, 63.33, 64.99. 2.1.4. N,N0 -Didodecyl-N,N,N0 ,N0 -tetramethyl-N,N0 -hexanediyl-diammonium Dibromide [12-6-12]. N,N,N-Dodecyldimethylamine (9.39 g, 44 mmol, 10% excess) was dissolved in acetone (30 mL) in a roundbottomed flask. 1,6-Dibromohexane dissolved in 20 mL of acetone (4.88 g, 20 mmol) was added dropwise under stirring. The mixture was refluxed for 24 h and cooled to room temperature, and then a white solid was collected by filtration and washed with acetone. The product was recrystallized from a 1:9 mixture of ethanol and diethyl ether. Yield: 78% 1 H NMR (400 MHz, CDCl3): δ 0.85 (t, 6H), 1.191.30 (m, 32H), 1.33 (m, 4H), 1.55(m, 4H), 1.69 (m, 4H), 1.98 (m, 4H), 3.35 (s, 12H), 3.42 (t, 4H), 3.68 (t, 4H). 13 C NMR (100 MHz, CDCl3): δ 13.94, 21.61, 22.50, 22.75, 24.48, 26.19, 29.12, 29.15, 29.26, 29.32, 29.42, 31.72, 50.88, 64.02, 64.58. 2.1.5. Bis-N,N,N-dodecyldimethyl-m-phenylenediammonium Dibromide [12-m-Φ-12]. R,R0 -Dibromo-m-xylene (1.00 g, 3.79 mmol) was dissolved in ethanol (50 mL) in a round-bottomed flask. N,N,NDodecyldimethylamine (1.98 g, 9.28 mmol, 20% excess) was added dropwise under stirring. The mixture was refluxed for 48 h. The solvent was removed under vacuum, and the product was recrystallized from a 1:9 mixture of ethanol and diethyl ether. Yield: 76% 1 H NMR (400 MHz, CDCl3): δ 0.87 (t, 6H), 1.211.3 (m, 32H), 1.31.4 (m, 4H), 1.83 (m, 4H), 3.23 (s, 12H), 3.55 (t, 4H), 5.02 (s, 4H), 7.51 (t, 1H), 7.85 (d, 2H), 8.60 (s, 1H). 13 C NMR (100 MHz, CDCl3,): δ 13.8, 22.45, 22.9, 26.11, 29.2, 29.38, 29.41, 29.50, 31.7, 52.8, 66.6, 68.8, 126.3, 129.8, 135.0, 137.5. 2.1.6. Bis-N,N,N-dodecyldimethyl-p-phenylenediammonium Dibromide [12-p-Φ-12]. R,R0 -Dibromo-p-xylene (1.00 g, 3.79 mmol) was dissolved in ethanol (30 mL) in a round-bottomed flask. N,N,NDodecyldimethylamine (1.98 g, 9.28 mmol, 20% excess) was added dropwise under stirring. The mixture was refluxed for 48 h. Then the solvent was removed under vacuum, and the product was recrystallized from a 1:9 mixture of ethanol and diethyl ether. Yield: 73% 1 H NMR (400 MHz, CDCl3): δ 0.85 (t, 6H), 1.181.28 (m, 36H), 1.31.4 (m, 4H), 1.84 (m, 4H), 3.21 (s, 12H), 3.52 (t, 4H), 5.18 (s, 4H), 7.80 (d, 4H). 13 C NMR (100 MHz, CDCl3): δ 14.0, 22.4, 22.9, 26.24, 29.2, 29.36, 29.4, 29.5, 31.8, 51.2, 64.6, 66.2, 103.3, 133.0. 2.2. NMR. 1H NMR and 13C NMR analyses were carried out on a Jeol 400 MHz spectrometer at 25 °C. Deuterated solvents were purchased from Armar Chemicals. 2.3. Tensiometry. Surface tension measurements were carried out on a Sigma 70 tensiometer (KSV) using the du No€uy ring method. The temperature was kept at 25 °C ((0.01 °C) by a cryostat Neslab RTE200. The glassware was cleaned with chromosulfuric acid, and the ring was flame treated prior to use. The measurements were conducted from low to high concentrations. 2.4. Conductometry. The cmc and the degree of ionization were determined by conductivity measurements according to the method described by Zana.22 The conductometer was a CDM 210 (Radiometer 13

Figure 1. From left to right, dodecyltrimethylammonium bromide (DTAB), the general structure of the gemini surfactants, the spacers used in the study, and the names given to the surfactants.

minerals, and spin-coated polymers have been used as a substrate. In this work, SPR was used on gold and QCM-D was employed on four widely different surfaces: gold, silica, gold made hydrophobic by the self-assembly of an alkane thiol, and gold made hydrophilic by the self-assembly of a hydroxylterminated alkane thiol. The surfactants studied are cationic geminis with bromide as the counterion prepared by reacting a bis-dimethylamino spacer with 2 equiv of dodecyl bromide or a dibromoalkane spacer with 2 equiv of docecyldimethylamine. Two types of spacer units have been used: a linear alkane with 2, 4, or 6 methylene groups (named 12-n-12 with n = 2, 4, and 6) and a meta- or paraphenylene group (denoted 12-m-Φ-12 and 12-p-Φ-12, respectively). The structures are given in Figure 1. The alkane spacer is flexible, and the phenylene unit is rigid. The main aim of this work is to evaluate to what extent the rigidity of the spacer affects the adsorption characteristics. The results are also compared with the adsorption of the corresponding monomeric surfactant, dodecyltrimethylammonium bromide (DTAB).

2. EXPERIMENTAL SECTION 2.1. Material and Synthesis. 2.1.1. Materials. Acetone (Aldrich, g99%), diethyl ether (Aldrich, g99%), ethanol (Kemetyl AB, 99.5%), dodecylbromide (Aldrich, 97%), N,N,N-dodecyldimethylamine (Acros, 95%), N,N,N0 ,N0 -tetramethyl-1,2-ethylenediamine (Aldrich, g99%), 1,4dibromobutane (Aldrich, 99%), 1,6-dibromohexane (Aldrich, g97%), R, R0 -dibromo-m-xylene (m-phenylene dibromide, Aldrich, g97%), R,R0 dibromo-p-xylene (p-phenylene dibromide, Aldrich, g98%), and hexadecanethiol (Aldrich, 99%) were used as purchased. N,N,N,-Dodecyldimethylammonium bromide (DTAB, Aldrich, g98%) was recrystallized twice from ethanol and dried under vacuum prior to use. 16-Hydroxyhexadecanethiol were from prepared from 15-carboxyhexadecanethiol (Aldrich, 90%) according to a protocol described elsewhere.21 Milli-Q water (resistance >18 MΩ 3 cm) was used for the preparation of aqueous solutions. 2.1.2. N,N0 -Didodecyl-N,N,N0 ,N0 -tetramethyl-N,N0 -ethanediyl-diammonium Dibromide [12-2-12]. Dodecyl bromide (11.96 g, 48 mmol, 20% excess) was dissolved in acetone (30 mL) in a round-bottomed flask. N,N,N0 ,N0 -Tetramethyl-1,2-ethylenediamine dissolved in 20 mL of acetone (2.32 g, 20 mmol) was added dropwise under stirring. The mixture was refluxed for 48 h and cooled to room temperature, and then a white solid was collected by filtration and washed with acetone. The product was recrystallized from a 1:9 mixture of ethanol and diethyl ether. Yield: 58% 1 H NMR (400 MHz, CDCl3): δ 0.86 (t, 6H), 1.211.3 (m, 32H), 1.36 (m, 4H), 1.8 (m, 4H), 3.49 (s, 12H), 3.68 (t, 4H), 4.70 (s, 4H).

7550

dx.doi.org/10.1021/la200539a |Langmuir 2011, 27, 7549–7557

Langmuir Copenhagen), and the solutions were thermostatted with a Haake Fison DC1 cryostat ((0.1 °C). The surfactant solution at a concentration above the cmc was diluted by the addition of 200 μL of water every 20 s with a 765 Dosimat (Metrohm) automatic buret. The conductometer and the automatic buret were computer-controlled via an RS232 interface and homemade software written in Python (version 2.5). 2.5. Krafft Temperature Determination. The Krafft temperature was determined by recording the change in the conductivity of the solution during an increase in temperature. A surfactant solution at 5cmc was placed in a double-jacketed beaker, and the temperature of the circulating water was thermostatted with a Neslab RTE-200 cryostat ((0.01 °C). The heating rate was set to 0.1 °C/min. The conductivity and the temperature were measured with a CDM 210 (Radiometer Copenhagen). The conductometer and the cryostat were computercontrolled via an RS232 interface and homemade software written in Python (version 2.5). 2.6. Surface Plasmon Resonance (SPR). Measurements of the variations in the surface plasmon resonance were made with an SPR Biacore X from Biacore SIA (Uppsala, Sweden). Details of the technique and the apparatus can be found elsewhere.21 After recording a baseline in Milli-Q water, we injected the surfactant solution (40 μL) in order to reach a steady state in the analysis chamber under a continuous flow of 25 μL/min. The gold chips from Biacore SIA (Uppsala, Sweden) were cleaned prior to use by the following procedure: UV ozone treatment for 10 min, immersion in a 5:1:1 mixture of H2O/ammonia (25%)/H2O2 (30%) at 75 °C for 5 min, rinsing with Milli-Q water, drying with N2, and finally 10 min of UV ozone treatment.

2.7. Quartz Crystal Microblance with Dissipation (QCM-D) Monitoring. A QCM-D instrument (model D300) from Q-Sense (G€oteborg, Sweden) was used. To avoid crystal perturbations during the shear oscillation of the crystals, the measurements were carried out under nonflowing conditions. The AT-cut crystals coated with a 100 nm thin gold layer were also from Q-Sense. The cleaning procedure, prior to use, was carried out as follows: UV-ozone treatment for 10 min, immersion in a 5:1:1 mixture of H2O/ammonia (25%)/H2O2 (30%) at 75 °C for 5 min, rinsing with Milli-Q water, drying with N2, and finally 10 min of UV ozone treatment. The silica surfaces were cleaned prior to use by simple UV ozone treatment. The hydrophobic and hydrophilic self-assembled monolayers (SAMCH3 and SAMOH, respectively) were prepared by immersing the cleaned gold surfaces in solutions of hexadecanethiol and 16hydroxyhexadecanethiol (2 mM in ethanol) for 16 h. The crystals were then rinsed with ethanol and sonicated in ethanol for 5 min to remove loosely adsorbed alkanethiols. Finally, the surfaces were rinsed with water and dried with nitrogen prior to use.21 The contact angle of water drops was checked on the modified surfaces and found to be 105° for the SAMCH3 surfaces and 30° for the SAMOH surfaces, which is in agreement with previous measurements.21 The measurements were made at 20 °C with a baseline corresponding to the damping of the crystal by Milli-Q water. The surfactant solution was injected into the crystal chamber using a homemade device, going from the lowest to the highest concentration without rinsing the surface between additions. The crystal was finally rinsed with Milli-Q water to remove poorly adsorbed surfactants. The automatic injection device is based on electric valves that open different flow channels on a collector connected to the inlet of the QCMD cell. The valves are controlled via an electronic interface using a remote control by the com port of a computer. The software was developed in Python (version 2.5). Prior to use, all of the flow channels were rinsed with ethanol and water. 2.8. Molecular Modeling. Structures were optimized in vacuum with Chem3d Pro 12 (CambridgeSoft) under the MM2 forcefield. The dodecyl tails were reduced to methyl groups. Distances expressed are the distance between the carbonyl carbons of the two carboxylate groups.

ARTICLE

Optimization was carried out starting from different conformations in order to assess as accurately as possible the lowest-energy conformation. 2.9. Assessment of Hydrophobicity. The log P values of the surfactant and of the spacer were determined with ALOGPS 2.1 software, which is based on a neural network principle.23

3. RESULTS AND DISCUSSION 3.1. Bulk Behavior. The synthesis and purification of the surfactants were straightforward and followed procedures described in previous work.4,24 Except for 12-p-Φ-12, all of the surfactants had a Krafft temperature of below 20 °C and were readily water-soluble. The Krafft temperature for this surfactant but with chloride as the counterion has previously been reported to be 23 °C.24,25 With bromide as the counterion, the Krafft temperature was found to be 38 °C using conductivity measurements. Conductometry is a suitable method for assessing the Krafft point of ionic surfactants because the conductivity of a solution in the dilute regime is proportional to the salt concentration. This means that the sudden increase in solubility that occurs when the Krafft temperature is reached can be accurately recorded. The difference in Krafft temperatures seen here for the chloride and the bromide salts is similar to what has been reported for other cationic surfactants (e.g., for alkylpyridinium salts26). Gemini surfactants are generally known to have low Krafft temperatures; in fact, that has been put forward as one of the advantages of this class of surfactants. The Krafft point of 38 °C found for 12-p-Φ-12 (bromide salt) is therefore surprisingly high. It is interesting that whereas 12-p-Φ-12 exhibited this high Krafft temperature, its isomer, 12-m-Φ-12, had an usually low Krafft temperature. This obviously reflects a more favorable molecular packing in the solid state of the para isomer. Because of its high Krafft temperature, 12-p-Φ-12 was studied only at 40 °C. The critical micelle concentration (cmc) of the gemini surfactants and of the corresponding monomeric amphiphile, dodecyltrimethylammonium bromide (DTAB), was determined by both tensiometry and conductometry, and the values are given in Table 1. The values of the area per molecule at the air/water interface were calculated from the surface excess concentration, Γ, determined from the Gibbs equation and using Avogadro’s number to calculate the molecular area:27 ! 1 Dγ Γ¼  2:303nRT D log C



1023 NA Γ

In the above expressions, γ is the surface tension in mN 3 m1, T is the absolute temperature in K, R = 8.31 (J 3 mol1 3 K1), NA is Avogadro’s number, Γ is the surface excess in mol 3 km2, and A the area per molecule in Å2. For the gemini surfactants, a value of n = 3 was used whereas n = 2 was used for DTAB.28,29 The cmc values determined by conductometry were taken at the break in the conductivity curves.22,30 Although the absolute values differ between the two methods of measuring the cmc— tensiometry and conductometry—the trends observed are the same. The ratio between the slopes above and below the cmc in the measurement of conductivity is indicative of the degree of ionization of the micelle, often referred to as R.22,30 (The counterion binding, often expressed as β, is related to R by β = 1  R.) The R values are also collected in Table 1. 7551

dx.doi.org/10.1021/la200539a |Langmuir 2011, 27, 7549–7557

Langmuir

ARTICLE

Table 1. cmc Values for the Gemini Surfactants Determined at 25 or 40°C by Tensiometry and by Conductometrya cmc by

area per

number of

surface tension above

cmc by conductometry

degree of

tensiometry (mmolar)

molecule (Å2)

molecules per nm2

the cmc (mN 3 m1)

(mmolar)

ionization, R

12-2-12 (25 °C)

0.95

86

1.16

29.3

0.88

0.21

12-4-12 (25 °C)

0.99

83

1.20

36.3

1.23

0.23

1.29

0.24

surfactant

12-4-12 (40 °C) 12-6-12 (25 °C)

0.85

108

0.93

37.7

12-m-Φ-12 (25 °C)

0.98

0.46

0.92

0.18

12-m-Φ-12 (40 °C)

0.78

100

1.00

35.5

0.94

0.27

12-p-Φ-12 (40 °C) DTAB (25 °C)

0.70 15

79 62

1.27 1.61

36.8 37.5

1.05 15.1

0.52 0.29

a

The area per surfactant molecule at the airwater interface was calculated from the slope of the surface tension versus log surfactant concentration curve using the Gibbs equation and represents the maximum in packing density. The surface tension above the cmc was taken to be the plateau of the tensiometry curve. The degree of ionization was determined from the conductivity curves.

Table 2. Distance between Charged Groups and log P Values of the Spacers Determined in Silico

The different spacers used will influence the molecule in a variety of ways. The distance between the charged nitrogen atoms will affect the charge density of the micelle and thus the degree of ionization. One can expect that a longer distance between charged headgroups will lead to a higher degree of ionization, and this is also seen in Table 1 for the series of 12-n12, although the difference between 12-2-12 and 12-4-12 is small. A rigid linker unit, such as a meta- or para-phenylene group, is likely to produce severe packing constraints as compared to a flexible alkane chain and a longer alkane spacer, such as in 12-612, and should provide more flexibility than a shorter spacer, such as in 12-2-12. Finally, the spacer unit will also influence the lipophilicity of the molecule: the more carbons, the stronger the lipophilic contribution. We have used a molecular mechanics force field in vacuo to calculate the distance between the two charged nitrogen atoms in the spacer units. In doing this, we have simplified the molecules in two ways: the dodecyl chain has been replaced by a methyl group and the counterions have been omitted. Because all of the geminis carry two dodecyl chains and all have the same counterion (bromide), we believe that the relative values of distance between the charges will still be accurate after these simplifications. Table 2 gives nitrogennitrogen distances for the different spacers.

The lipophilicity of the spacers was quantified by a log P value, which was calculated using a neural network approach.23 log P values of the spacers are also given in Table 2. As can be seen, 12-2-12 has the least hydrophobic spacer and 12-6-12 has the most hydrophobic spacer. The phenylene unit makes approximately the same contribution to the hydrophobicity as does the n-butylene group of 12-4-12. As can be seen in Table 1, the cmc values for the 12-n-12 series exhibit a maximum for 12-4-12, which is in agreement with data found in the literature.4 The slightly higher cmc value for 12-4-12 at 40 °C than at 25 °C is according to expectations. The degree of ionization determined by conductometry increases with the spacer size, which is also in accordance with previous work.4 Also, the values of the area per polar headgroup for the 12-n-12 surfactants are in line with literature data.31 The approximately 15-fold-lower cmc values of the gemini surfactants compared to those of the corresponding monomeric species, DTAB, is a confirmation of the high effectiveness of the geminis, as discussed in the Introduction. The two geminis with a rigid phenylene spacer may be compared with 12-4-12, which is the closest equivalent out of the 12-n-12 series with respect to both the size of the spacer unit and log P. Because 12-p-Φ-12 has a Krafft point of 38 °C and its values should be compared with those of 12-m-Φ-12 and 12-4-12, all three surfactants were studied at 40 °C. Table 1 shows that the cmc values of the phenylene surfactants are slightly lower than the value for 12-4-12 in spite of the fact that the log P values indicate that 12-4-12 is somewhat more hydrophobic. Thus, the rigid spacer seems not to render selfassociation in the bulk difficult. The degree of ionization is approximately the same for 12m-Φ-12 as for 12-4-12. The value for 12-p-Φ-12 is much higher, however. In fact, an R value of 0.52 is unusually high and probably reflects a large distance between the cationic groups at the micelle surface. This is reasonable in view of the structure of 12-p-Φ-12, with the two headgroups situated on opposite sides of the para-phenylene linker. An inspection of the molecular models shows that the situation is different for the meta isomer, 12-m-Φ-12. This molecule is capable of orienting its quaternary ammonium groups relatively close to each other. Table 1 also reveals an interesting difference between the two phenylene surfactants when it comes to the area per molecule at the airwater interface. Whereas the value is 100 Å2 for 12-m-Φ7552

dx.doi.org/10.1021/la200539a |Langmuir 2011, 27, 7549–7557

Langmuir

ARTICLE

12, which is almost as high as for 12-6-12, 12-p-Φ-12 has the lowest value of all of the geminis, 79 Å2. Most likely, this reflects the fact that the para isomer can pack tightly, which the meta isomer cannot do. The ability to form tightly aligned assemblies is probably also the reason behind the high Krafft temperature of 12-p-Φ-12, as discussed above. 3.2. Adsorption at Solid Surfaces. 3.2.1. Adsorption on Gold Monitored by SPR. In the SPR technique, monochromatic p-polarized light is passed through an optical unit made of a prism linked to a glass support coated with a thin layer of gold, which is in contact with the solution in the flow cell. Light directed above a critical angle of the incident light will cause an evanescent field to penetrate into the gold film. This evanescent field can couple to an electromagnetic surface wave, which is called a surface plasmon. Therefore, the reflected light will not be totally reflected, and the surface plasmon, which is excited at the goldliquid interface and thus absorbs light, is measured by a photodiode array as a minimum in reflected light. The change in the refractive index in the vicinity of the surface, resulting from the adsorption of surfactant, is monitored as a change in the position of the minimum intensity of the reflected light. This angle variation is in the BIAcore terminology expressed in resonance units (RUs), where 1 RU is equal to a 0.0001° change in the intensity minimum. This variation, ΔRU, is directly proportional to the mass adsorbed, Δm, according to Δm ¼

ΔRU  CSPR β

where CSPR is a factor containing an instrument constant, dn/dc is the variation of the refractive index with concentration for the adsorbent, and β is a factor compensating for the decrease in the SPR signal with distance from the gold substrate. CSPR was calculated to 0.094 ( 0.008 ng/cm2 using an average dn/dc for 18 different surfactants, and β was set to 1, which is the case for a plain gold surface.32 The factor β will differ from 1 when the surface layer is thick. In this work, one can anticipate the formation of a very thin layer and β will be close to 1. Furthermore, in the present case, because of the very low cmc values of the gemini surfactants, the change in refractive index measured by the SPR technique can be regarded as being caused only by the adsorption of the amphiphile on the gold surface. The gold surface is often seen as a controversial substrate, and its true nature is still debated.3335 Nevertheless, halide ions are well known to adsorb readily on gold, providing a negative charge to the surface.36,37 Thus, one can anticipate that the electrostatic interaction between the polar headgroups of the surfactants and bromide ions chemisorbed on the gold surface will play a role in the adsorption process. The adsorption isotherms for the 12-n-12 series as well as for 12-m-Φ-12 are shown in Figure 2. The solution concentrations are normalized to the cmc. As can be seen, the adsorption pattern at low concentrations is relatively similar for all of the surfactants. At higher concentrations, the curves deviate. The plateau values for the 12-n-12 series are inversely proportional to the spacer length (i.e., 12-2-12 > 12-4-12 > 12-6-12). 12-m-Φ-12, which from a spacer length and hydrophobicity point of view is similar to 12-4-12, gives somewhat higher adsorption above the cmc than anticipated, in between the values of 12-2-12 and 12-4-12. As mentioned above, the gold surface can be regarded as being strongly negatively charged because of the chemisorption of bromide ions. The values obtained around the cmc can therefore

Figure 2. Mass adsorbed on gold determined by SPR at 25 °C as a function of surfactant concentration normalized to the cmc.

be assumed to relate to amphiphiles self-assembled in a double layer or as closely packed micelles. The difference in the adsorbed amount between the three surfactants in the 12-n-12 series is large. Around the cmc, 12-212 gives approximately double the amount of adsorbed material as compared to 12-6-12. This is reasonable and it obviously reflects a closer packing of the dodecyl chains for the short gemini spacer. However, the same difference was not found at the airwater interface. As shown in Table 1, 12-2-12 occupies an area of 86 Å2 and 12-6-12 occupies 108 Å2. This difference in molecular area is smaller than that found at the solid surface. It seems that the two tails of 12-6-12 come closer together at the airwater interface than at the solid surface. The reason that we experience this difference in the packing pattern between the airwater interface and the negatively charged goldwater interface is probably that at the airwater interface the relatively long alkane spacer (of 12-6-12) does not align along the surface but folds into the air, thus enabling a closer packing of the long hydrocarbon tails. This behavior has been described before.38 Such folding is unlikely at the polar gold surface; hence, the density of the long alkyl chains becomes low, as does the amount of adsorbed material. As can be seen in Figure 2, the curve for the monomeric surfactant, DTAB, is somewhere in the middle of the adsorption curves. As will be seen below in the discussion of the QCM experiments, the adsorption curve for DTAB at the various surfaces is at least as high as for the average gemini surfactant. However, this does not mean that the true adsorption isotherm of DTAB is high compared to the that of geminis. All of the plots shown in this article on mass adsorbed versus surfactant concentration have been normalized to the cmc, and as we saw earlier, the cmc of the geminis is around 15 times lower than the cmc of DTAB. Hence, a plot of mass adsorbed versus absolute molar concentration (i.e., not normalized to the cmc) would provide the message that a gemini surfactant gives total coverage of the surface at a 15-fold lower concentration than does DTAB. Again, this demonstrates the effectiveness of the geminis. 3.2.2. Adsorption Monitored by QCM-D. QCM-D is a gravimetric method that is well suited to studies of surfactant adsorption at solid surfaces.20,39,40 The measuring principle is 7553

dx.doi.org/10.1021/la200539a |Langmuir 2011, 27, 7549–7557

Langmuir

ARTICLE

based on the fact that the resonance frequency of a quartz crystal depends on the mass of the crystal. The adsorption of a solute will thus change the frequency. Employed in the early 1950s to measure the adsorption of gases,41 QCM was adapted to liquid environments in the 1980s thanks to the work of Kanazawa.42 Around 10 years ago, a theoretical development enabled insight into the structure of the adsorbed film by correlating the dissipation of the oscillation to the viscoelastic nature of the adsorbed layer.43 The dissipation, which corresponds to the decay of the intensity of the oscillation versus time resulting from a loss of energy, can be defined as D¼

Edissipated 2πEstored

where Edissipated and Estored are the dissipated and stored energy of the adsorbed layer, respectively. The variation of frequency for the gemini surfactants was monitored at the third, fifth, and seventh overtones, and the dissipation was also recorded. The adsorption values normalized to the overtone number overlapped whereas the dissipation remained below 1  106, which is an indication that the change in bulk viscosity and density remain negligible upon an increase in surfactant concentration.44 The mass adsorbed can then be determined using the Sauerbrey relation45 Δm ¼ 

Δf  C n

where Δm is the adsorbed mass, Δf is the variation in frequency observed at overtone n, and C = 17.7 ng 3 cm2, which is a constant characteristic of the equipment. After first recording a baseline in pure water, surfactant solutions of increasing concentrations were injected into the measurement cell, leading to a step-by-step change in frequency caused by adsorption. For each step, the frequency value was converted to mass using the above relationship. 3.2.2.1. Gold. Figure 3 shows the adsorption curves on a gold surface at 25 °C for DTAB, 12-2-12, 12-4-12, 12-6-12, and 12-mΦ-12 and at 40 °C for 12-m-Φ-12 and 12-p-Φ-12. At 25 °C, the four gemini surfactants lie close together with a plateau adsorption in the range of 140180 ng 3 cm2. At the cmc, the adsorption decreases as 12-m-Φ-12 > 12-2-12 > 12-4-12 > 126-12 but the differences are small. The level of adsorption is higher than that obtained by SPR, also on gold, but the order of the adsorbed amount for the different surfactants is approximately the same. Higher values of the adsorbed amount with QCM than with SPR are normally experienced when the two methods are used and compared.39,40 This is at least partially due to QCM also taking into account the water that accompanies the adsorbed layer. The adsorbed amount for the surfactants with a rigid spacer, 12-m-Φ-12 or 12-p-Φ-12, run at 40 °C is much higher than the amount recorded for the 25 °C experiments. We interpret this as being due to differences in the structure of the adsorbed layer, as will be discussed below. The DTAB curve in the experiments shown in Figure 3, as well as in all of the other QCM experiments (Figures 46), deviates from the other curves by progressing upward beyond the cmc. This does not reflect continued adsorption above the cmc, however. As we have shown in previous papers, the continued

Figure 3. Mass adsorbed on gold determined by QCM-D at the third overtone at 25 and 40 °C as a function of surfactant concentration normalized to the cmc.

rise of the adsorption curve is due to a high concentration of surfactant in the solution. For surfactants with a high cmc, as is the case for DTAB, the unimer concentration is high at the plateau value of adsorption. This gives rise to a considerable bulk effect, which in the interpretation of QCM experiments must be deducted from the observed value in order to obtain the true value of the adsorbed amount.40,44 No attempt has been made in this work to subtract the bulk contribution from the values of DTAB adsorption. 3.2.2.2. Silicon Dioxide (Silica). Figure 4 shows the adsorption curves on a silicon dioxide surface. 12-p-Φ-12, monitored at 40 °C, is also included. The order of the adsorbed amount at the cmc for the geminis is 12-m-Φ-12 > 12-2-12 > 12-4-12 > 12-6-12 (i.e., roughly the same as on the gold surface). The lower adsorption of 12-p-Φ-12 can be attributed to the experiments being conducted at higher temperature. Surfactant adsorption generally decreases with increasing temperature.46 The adsorption pattern on silicon dioxide is relatively similar to that on gold, as can be seen by comparing Figure 4 with Figure 3. The curve for DTAB is the highest, and the order of the adsorbed amount for the geminis is approximately the same. Such similarities are not unexpected, however, because both surfaces are in reality negatively charged. The gold surface acquires a negative charge by the chemisorption of bromide ions (which are counterions of the cationic surfactants), and the silicon dioxide surface carries negative charges because of the deprotonation of silanol groups. It is conceivable that the higher level of the adsorbed amount of surfactants on the silicon dioxide surface compared to that on the gold surface is due to a higher density of negative charges on the SiO2 surface. This issue will be further discussed below. 3.2.2.3. Adsorption on Hydrophobized Gold. The adsorption behavior of the gemini surfactants was determined on gold made hydrophobic by treatment with hexadecanethiol. The selfassembled monolayer that is formed will expose methyl groups to the aqueous phase; thus, the surface will have a pronounced hydrophobic character. Figure 5 shows the adsorption isotherms on the hydrophobized gold surface. At the cmc, the adsorption pattern is quite similar to that for the previously discussed surfaces, with DTAB 7554

dx.doi.org/10.1021/la200539a |Langmuir 2011, 27, 7549–7557

Langmuir

Figure 4. Mass adsorbed on silicon dioxide determined by QCM at the third overtone as a function of surfactant concentration normalized to the cmc. The adsorption of 12-p-Φ-12 was recorded at 40 °C, and the other surfactants were studied at 25 °C.

giving the highest value and the adsorbed amount of the geminis decreasing in the order of 12-2-12 > 12-4-12 > 12-m-Φ-12 > 12-6-12. 3.2.2.4. Adsorption on Hydrophilized Gold. The gold surface was treated with 16-hydroxyhexadecanethiol, which is known to give a hydroxyl-functional surface. The adsorption pattern, shown in Figure 6, is again very similar to that seen for the other surfaces, with 12-2-12 adsorbing the most and 12-6-12 adsorbing the least at concentrations around the cmc. 3.2.2.5. Comparison of the QCM Results for the Different Surfaces. For the four surfaces studied—gold, silicon dioxide, hydrophobized gold, and hydrophilized gold—the order of the adsorbed amounts for the four gemini surfactants that could be studied at 25 °C ranged from 140 to 260 ng 3 cm2. Within the 12-n-12 series, 12-2-12 adsorbed the most and 12-6-12 adsorbed the least on all of the substrates. 12-m-Φ-12 usually appeared somewhere in the middle. One may therefore conclude that the ability to self-assemble at solid surfaces is not hampered by a rigid spacer in the gemini surfactant. Table 3 summarizes the adsorption levels for the four gemini surfactants at the different surfaces. The highest adsorbed amounts are obtained on silicon dioxide, followed by hydrophilized gold (although very widespread) > hydrophobized gold > pure gold. It is not obvious how to translate these results into the organization of the geminis at the surfaces. Gold treated with hexadecanethiol is very hydrophobic, giving a water contact angle of 105°.40 All surfactants would be expected to form a tightly packed monolayer or closely aligned hemimicelles on such a surface. Considering their special geometry, gemini surfactants are likely to assemble in a monolayer rather than as hemimicelles. Thus, it is likely that the range of 160190 ng 3 cm2 corresponds to a closely aligned monolayer of surfactants. QCM-D takes bound water into account, and a comparison of the values of the adsorbed amount on gold obtained with SPR, which records only the organic adsorbate, shows that the adsorbed water represents a substantial amount of the total amount recorded by QCM-D. This has also been observed before. We have assumed in the following discussion that the

ARTICLE

Figure 5. Mass adsorbed on hydrophobized gold determined by QCM at the third overtone as a function of surfactant concentration normalized to the cmc. The adsorption of 12-p-Φ-12 was recorded at 40 °C, and the other surfactants were studied at 25 °C.

Figure 6. Mass adsorbed on hydrophilized gold determined by QCM at the third overtone as a function of surfactant concentration normalized to the cmc. The adsorption of 12-p-Φ-12 was recorded at 40 °C, and the other surfactants were studied at 25 °C.

ratio of surfactant to bound water in the surface layer is the same on all surfaces studied. Silicon dioxide is a strongly negatively charged substrate and is definitely likely to give a bilayer adsorption of cationic amphiphiles. It is therefore reasonable to assume that the level of the adsorbed amount on this surface, 230260 ng 3 cm2, corresponds to a bilayer. This coverage is obviously less than twice the monolayer coverage on the hydrophobized gold surface (160190 ng 3 cm2), but this is not unusual. Also, the gold surface is known to become negatively charged as a result of the chemisorption of counterions to cationic surfactants,36,37 but the charge density may not be very high. Without chemisorbed ions, the gold surface, when freshly cleaned, is hydrophobic. We have previously recorded a water contact angle of 88° on such a substrate.40 Because the adsorbed amount recorded at 25 °C on the naked gold surface is 7555

dx.doi.org/10.1021/la200539a |Langmuir 2011, 27, 7549–7557

Langmuir

ARTICLE

Table 3. Range of Adsorbed Amount at 25 °C for Geminis 122-12, 12-4-12, 12-6-12, and 12-m-Φ-12 adsorbed amount at

adsorbed amount at

half cmc (ng 3 cm2)

cmc (ng 3 cm2)

silicon dioxide gold

140170 100140

230260 140180

hydrophobized gold

90140

160190

hydrophilized gold

7090

150260

substrate

Figure 7. Possible modes of organization of gemini bilayers. Left, noninterdigitated bilayer; right, interdigitated bilayer.

approximately the same as the amount measured on hydrophobized gold, the first interpretation that comes to mind is that there is monolayer adsorption on both surfaces. However, an alternative structure for the adsorbate on the gold surface is that of an interdigitated double layer. A structure with fully interdigitated surfactants, alternating with head up and head down, would give approximately the same surfactant loading on the surface as a monolayer with all of the surfactants head up. The reason that a fully developed bilayer would form on the silica surface whereas on a gold surface adsorption leads to an interdigitated bilayer is that the density of negative surfaces charges is likely to be lower on gold than on silica. Figure 7 illustrates the modes of adsorption. It is well known that doubletailed surfactants such as phospholipids may form interdigitated double layers.4749 Recently, an interdigitated structure of the solid-supported bilayer of cationic gemini surfactants was demonstrated by the calculation of the electron density profile along the normal to the bilayer.50 It is also known that subtle differences in the experimental conditions may cause a change in the adsorption pattern from regular to interdigitated bilayers.51 The unexpectedly high adsorbed amount at 40 °C of 12-m-Φ12 and 12-p-Φ-12 on the gold surface can then be explained by the structure of the bilayer formed. As can be seen from Figure 3, these two surfactants give approximately twice the adsorbed amount at 40 °C compared to the other gemini surfactants at 25 °C. It is particularly noteworthy that for 12-m-Φ-12 the adsorption at 25 °C is approximately half that at 40 °C. (Because of the high Krafft point, 12-p-Φ-12 could not be studied at 25 °C.) These results can be explained by a structural difference in the bilayers formed at the gold surface. At 40 °C, 12-m-Φ-12 and 12-p-Φ-12 form a regular bilayer without interpenetration of the hydrocarbon tails whereas at 25 °C all of the surfactants, including 12-m-Φ-12, give an interdigitated bilayer. It is known that a higher temperature renders the formation of an interdigitated adsorbed layer less favorable and the interpenetration of the chains is often lost when the transition from a gel to a fluid bilayer occurs.52 The hydrophilized gold surface is in effect a hydroxyl-functional substrate. The driving force for the adsorption of a cationic amphiphile on such a surface is likely to be much smaller than

on a negatively charged or a hydrophobic surface. The high level of the adsorbed amount obtained on this surface is therefore surprising. A lower driving force for adsorption on this surface compared to that on the three other surfaces can be detected from the QCM-D data, however. The adsorption at low surfactant concentration on the hydrophilized gold is lower than for the other surfaces. This is quantified in Figure 3 by the values of the adsorbed amount at half the cmc. Whereas at half the cmc the amount of adsorbed surfactant on hydrophilized gold is only around 40% of the amount at the cmc, for the other surfaces the level at half the cmc is between 66 and 94% of the amount at the cmc.

4. CONCLUSIONS In this work, cationic gemini surfactants with flexible or rigid spacers were compared with respect to self-assembly in water and on a variety of solid surfaces. It was found that for surfactants with spacers of the same hydrophobicity and giving the same distance between headgroups, the cmc values were approximately the same. meta-Phenylene and para-phenylene were employed as the rigid spacer unit. The para isomer was found to pack very tightly at the airwater interface. It also exhibited a high Krafft temperature of 38 °C. Such high Krafft points are unusual among geminis and most likely reflects a favorable packing of the amphiphile also in the solid state. The adsorption of the different gemini surfactants was evaluated on hydrophobized, hydrophilized, and untreated gold as well as on silica, which is strongly negatively charged. For flexible geminis of the 12-n-12 type, the adsorbed amount increased with decreasing spacer length regardless of the nature of the substrate. At 25 °C, rigid gemini surfactant 12-m-Φ-12 showed an adsorption pattern similar to that of the 12-n-12 series. Thus, a rigid linker unit in a gemini surfactant does not render self-assembly difficult in bulk water or at solid surfaces. The amount adsorbed at 25 °C on gold is approximately half that obtained on silica in spite of the fact that gold can also be regarded as a negatively charged surface under the conditions used. Bromide, which is the counterion of the gemini surfactants, is known to chemisorb on gold and impart negative charges. We believe that the difference in the adsorbed amount is due to differences in the structure of the adsorbed layer. Whereas on gold an interdigitated bilayer is formed, on silica the bilayer is noninterdigitated. When the adsorption experiments are performed at 40 °C (which is needed for 12-p-Φ-12 because of its high Krafft point), the adsorbed amount is approximately twice that obtained at 25 °C. Also, this can be explained by the formation of a bilayer without interpenetration of the alkyl chains at 40 °C. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ REFERENCES (1) Bunton, C. A.; Robinson, L. B.; Schaak, J.; Stam, M. F. J. Org. Chem. 1971, 36, 2346–2350. (2) Menger, F. M.; Littau, C. A. J. Am. Chem. Soc. 1991, 113, 1451–1452. (3) Menger, F. M.; Littau, C. A. J. Am. Chem. Soc. 1993, 115, 10083–10090. (4) Zana, R.; Benrraou, M.; Rueff, R. Langmuir 1991, 7, 1072–1075. 7556

dx.doi.org/10.1021/la200539a |Langmuir 2011, 27, 7549–7557

Langmuir (5) Zhu, Y.-p.; Masuyama, A.; Kirito, Y.-i.; Okahara, M.; Rosen, M. J. Am. Oil Chem. Soc. 1992, 69, 626–632. (6) Rosen, M. J. Surfactants and Interfacial Phenomena, 3rd ed.; Wiley: New York, 2004; pp 415427. (7) Menger, F. M.; Keiper, J. S. Angew. Chem., Int. Ed. 2000, 39, 1906–1920. (8) Zana, R. Adv. Colloid Interface Sci. 2002, 97, 203–251. (9) Esumi, K.; Goino, M.; Koide, Y. J. Colloid Interface Sci. 1996, 183, 539–545. (10) Chorro, C.; Chorro, M.; Dolladille, O.; Partyka, S.; Zana, R. J. Colloid Interface Sci. 1998, 199, 169–176. (11) Chorro, M.; Chorro, C.; Dolladille, O.; Partyka, S.; Zana, R. J. Colloid Interface Sci. 1999, 210, 134–143. (12) In, M.; Bec, V.; Aguerre-Chariol, O.; Zana, R. Langmuir 1999, 16, 141–148. (13) Atkin, R.; Craig, V. S. J.; Wanless, E. J.; Biggs, S. J. Phys. Chem. B 2003, 107, 2978–2985. (14) Atkin, R.; Craig, V. S. J.; Wanless, E. J.; Biggs, S. Adv. Colloid Interface Sci. 2003, 103, 219–304. (15) Manne, S.; Schaeffer, T. E.; Huo, Q.; Hansma, P. K.; Morse, D. E.; Stucky, G. D.; Aksay, I. A. Langmuir 1997, 13, 6382–6387. (16) Blomberg, E.; Verrall, R.; Claesson, P. M. Langmuir 2007, 24, 1133–1140. (17) Fielden, M. L.; Claesson, P. M.; Verrall, R. E. Langmuir 1999, 15, 3924–3934.  (18) Boschkova, K.; Feiler, A.; Kronberg, B.; Stalgren, J. J. R. Langmuir 2002, 18, 7930–7935. (19) Sakai, K.; Matsuhashi, K.; Honya, A.; Oguchi, T.; Sakai, H.; Abe, M. Langmuir 2010, 26, 17119–17125. (20) Tehrani-Bagha, A. R.; Holmberg, K. Langmuir 2008, 24, 6140–6145. (21) Oskarsson, H.; Holmberg, K. J. Colloid Interface Sci. 2006, 301, 360–369. (22) Zana, R. J. Colloid Interface Sci. 1980, 78, 330–337. (23) Tetko, I.; Gasteiger, J.; Todeschini, R.; Mauri, A.; Livingstone, D.; Ertl, P.; Palyulin, V.; Radchenko, E.; Zefirov, N.; Makarenko, A.; Tanchuk, V.; Prokopenko, V. J. Comput.-Aided Mol. Des. 2005, 19, 453–463. (24) Laschewsky, A.; Lunkenheimer, K.; Rakotoaly, R. H.; Wattebled, L. Colloid Polym. Sci. 2005, 283, 469–479. (25) Laschewsky, A.; Wattebled, L.; Arotc-arena, M.; Habib-Jiwan, J.-L.; Rakotoaly, R. H. Langmuir 2005, 21, 7170–7179. (26) Heckmann, K.; Schwarz, R.; Strnad, J. J. Colloid Interface Sci. 1987, 120, 114–117. (27) Rosen, M. J. Surfactants and Interfacial Phenomena, 3rd ed.; Wiley: New York, 2004; pp 6064. (28) Tsubone, K.; Arakawa, Y.; Rosen, M. J. J. Colloid Interface Sci. 2003, 262, 516–524. (29) Tsubone, K.; Rosen, M. J. J. Colloid Interface Sci. 2001, 244, 394–398. (30) Rosen, M. J. Surfactants and Interfacial Phenomena, 3rd ed.; Wiley: New York, 2004; pp 139144. (31) Zana, R. R.; Alami, E. O. Gemini Surfactants. In Novel Surfactants: Preparation, Applications, and Biodegradability; Holmberg, K., Ed.; Dekker: New York, 2006; pp 385424. (32) Larsson, C.; Rodahl, M.; H€o€ok, F. Anal. Chem. 2003, 75, 5080–5087. (33) King, D. E. In Oxidation of Gold by Ultraviolet Light and Ozone at 25 °C; Proceedings of the 41st National Symposium of the American Vacuum Society; Denver, CO, 1995; AVS: Denver, CO, 1995; pp 12471253. (34) Ron, H.; Matlis, S.; Rubinstein, I. Langmuir 1998, 14, 1116–1121. (35) Smith, T. J. Colloid Interface Sci. 1980, 75, 51–55. (36) Tao, N. J.; Lindsay, S. M. J. Phys. Chem. 2002, 96, 5213–5217. (37) Wandlowski, T.; Wang, J. X.; Magnussen, O. M.; Ocko, B. M. J. Phys. Chem. 1996, 100, 10277–10287. (38) Alami, E.; Beinert, G.; Marie, P.; Zana, R. Langmuir 1993, 9, 1465–1467.

ARTICLE

(39) Bordes, R.; Tropsch, J.; Holmberg, K. Langmuir 2010, 26, 3077–3083. (40) Bordes, R.; Tropsch, J.; Holmberg, K. Langmuir 2010, 26, 10935–10942. (41) Urbakh, M.; Tsionsky, V.; Gileadi, E.; Daikhin, L. In Piezoelectric Sensors; Steinem, C., Janshoff, A., Eds.; Springer-Verlag: Berlin, 2007; pp 111149. (42) Kanazawa, K.; Gordon, J. G. Anal. Chim. Acta 1985, 175, 99–105. (43) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391. (44) Bordes, R.; H€o€ok, F. Anal. Chem. 2010, 82, 9116–9121. (45) Sauerbrey, G. Z. Phys. 1955, 155, 206. (46) Rosen, M. J. Surfactants and Interfacial Phenomena, 3rd ed.; Wiley: New York, 2004; p 53. (47) McIntosh, T. J.; McDaniel, R. V.; Simon, S. A. BBA-Biomembranes 1983, 731, 109–114. (48) Mou, J.; Yang, J.; Huang, C.; Shao, Z. Biochemistry 1994, 33, 9981–9985. (49) Hui, S. W.; Huang, C. H. Biochemistry 1986, 25, 1330–1335. (50) Caracciolo, G.; Pozzi, D.; Mancini, G.; Caminiti, R. Langmuir 2007, 23, 10040–10043. (51) Sanii, B.; Szmodis, A. W.; Bricarello, D. A.; Oliver, A. E.; Parikh, A. N. J. Phys. Chem. B 2009, 114, 215–219. (52) Wolf, C.; Chachaty, C.; Takahashi, H.; Hatta, I.; Quinn, P. J. Chem. Phys. Lipids 1997, 87, 111–124.

7557

dx.doi.org/10.1021/la200539a |Langmuir 2011, 27, 7549–7557