Unexpected Adsorption Behavior of Nonionic Surfactants from Glycol

A. Pan , B. Naskar , G. K. S. Prameela , B. V. N. Phani Kumar , V. K. Aswal , S. C. Bhattacharya , A. B. Mandal , S. P. Moulik. Soft Matter 2014 10 (3...
0 downloads 0 Views 83KB Size
Langmuir 2006, 22, 11187-11192

11187

Unexpected Adsorption Behavior of Nonionic Surfactants from Glycol Solvents Caroline Seguin,† Julian Eastoe,*,† Sarah Rogers,† Martin Hollamby,† and Robert M. Dalgliesh‡ School of Chemistry, UniVersity of Bristol, Bristol, BS8 1TS, United Kingdom, and ISIS-CCLRC, Rutherford Appleton Laboratory, Chilton, Oxon, OX11 0QX, United Kingdom ReceiVed June 16, 2006. In Final Form: September 8, 2006 Adsorption and interfacial properties of model methyl-capped nonionic surfactants C8E4OMe [C8H17O(C2H4O)4CH3] and C10E4OMe [C10H21O(C2H4O)4CH3] were studied in water and water/ethylene glycol mixtures as well as pure ethylene glycol. Critical micellar concentrations (cmc’s), surface tensions, and surface excess were determined using surface tension (ST) and neutron reflection (NR) as a function of solvent type and surfactant tail length. The ST results show a strong dependence on solvent type in terms of cmc. The NR data were analyzed using a single-layer model for the adsorbed surfactant films. Surprisingly, the adsorption parameters obtained in both water and pure ethylene glycol were very similar, and variations in film thickness or area per molecule are negligible in respect of the uncertainties. Similarly, for C10E4OMe, estimates for the free energies of adsorption and micellization show only a weak solvent dependence. These results suggest that for such model nonionic surfactants dilute solution properties are dictated by solvophobicity, which is quite similar for this class of water, glycol, and water-glycol mixtures. More specifically, the nature of the adsorption layer appears to be hardly affected by the type of solvent subphase. The findings highlight the significance of solvophobicity and show that model nonionic surfactants can behave very similarly in hydrogenbonding glycol solvents and water.

Introduction Adsorption and interfacial behavior of nonionic surfactants in aqueous solutions have been well studied and an extensive literature can be found.1-13 On the other hand, studies in water/ glycol mixtures and pure glycols are less comprehensive and are almost exclusively devoted to ionic surfactants.14-28 However, * Author to whom correspondence should be addressed. E-mail: [email protected]. † University of Bristol. ‡ Rutherford Appleton Laboratory. (1) Penfold, J.; Staples, E.; Tucker, I.; Thompson, L.; Thomas, R. K. J. Colloid Interface Sci. 2002, 247, 404-411. (2) Binks, B. P.; Fletcher, P. D. I.; Paunov, V. N.; Segal, D. Langmuir 2000, 16, 8926-8931. (3) Valkovska, D.; Wilkinson, K. M.; Campbell, R. A.; Bain, C. D.; Wat, R.; Eastoe, J. Langmuir 2003, 19, 5960-5962. (4) Goates, S. R.; Schofield, D. A.; Bain, C. D. Langmuir 1999, 15, 14001409. (5) Lu, J. R.; Li, Z. X.; Thomas, R. K. J. Phys. Chem. 1994, 98, 6559-6567. (6) Lu, J. R.; Su, T. J.; Li, Z. X.; Thomas, R. K.; Staples, E. J.; Tucker, I. J. Phys. Chem. B 1997, 101, 10332-10339. (7) Penfold, J.; Staples, E.; Tucker, I.; Thomas, R. K.; Woodling, R.; Dong, C. C. J. Colloid Interface Sci. 2003, 262, 235-242. (8) Penfold, J.; Staples, E.; Tucker, I.; Thomas, R. K. Colloids Surf., A 1999, 155, 15-26. (9) Penfold, J.; Staples, E.; Tucker, I.; Thomas, R. K. J. Colloid Interface Sci. 1998, 201, 223-232. (10) Penfold, J.; Thomas, R. K.; Lu, J. R.; Staples, E.; Tucker, I.; Thompson, L. Physica B 1994, 198, 110-115. (11) Stoyanov, S. D.; Rehage, H.; Paunov, V. N. Phys. ReV. Lett. 2003, 91, 086102-086101/086102-086104. (12) Lu, J. R.; Li, Z. X.; Thomas, R. K.; Staples, E.; Tucker, I.; Penfold, J. J. Phys. Chem. 1993, 97, 8012-8020. (13) Staples, E.; Thompson, L.; Tucker, I.; Penfold, J.; Thomas, R. K.; Lu, J. R. Langmuir 1993, 9, 1651-1656. (14) Ionescu, L. G.; Fung, D. S. J. Chem. Soc., Faraday Trans. 1 1981, 77, 2907-2912. (15) Backlund, S.; Bergenstahl, B.; Molander, O.; Warnheim, T. J. Colloid Interface Sci. 1989, 131, 393-401. (16) Blokhus, A. M.; Hoiland, H.; Gjerde, M. L.; Backlund, S.; Ruths, M.; Douheret, G. 1990, 82, 243-248. (17) Sjoberg, M.; Henriksson, U.; Warnheim, T. Langmuir 1990, 6, 12051211. (18) Sjoberg, M.; Jansson, M.; Henriksson, U. Langmuir 1992, 8, 409-413. (19) Gharibi, H.; Palepu, R.; Bloor, D. M.; Hall, D. G.; Wyn-Jones, E. Langmuir 1992, 8, 782-787.

compared to the well-documented textbook behavior known in water, less is understood about how CiEj nonionic surfactants respond in glycol solvents. Possibly, this is due to close chemical nature of glycols and the ethylene oxide headgroups, which introduces difficulties for determining critical micelle concentrations (cmc’s) by using standard methods such as electrical conductivity and dye probe solubilization. The aim of this work was to extend knowledge of the adsorption behavior of nonionic surfactants and to study the influence of varying solvent type on interfacial properties as well as to provide new insight into “solvophobicity”. The systems were designed to explore variations of solvent (water, ethylene glycol (EG), and 50/50 wt % water/EG mixtures) and surfactant chain length at constant headgroup (C8E4OMe [C8H17O(C2H4O)4CH3] and C10E4OMe [C10H21O(C2H4O)4CH3]). Ethylene glycol was selected as an alternative H-bonding solvent, owing to good miscibility with water, but a very different surface tension. Two important techniques for investigating surfactant interfaces are surface tension (ST) and neutron reflectivity (NR); when combined, these represent powerful tools to probe adsorption layer densities and structures,1,3,5-10,12,13 especially in EG and water/EG mixtures. In this work, Du Nou¨y surface tension and NR measurements have been employed to reveal (where appropriate) cmc’s, surface tensions, surface excesses (Γmax), and areas per molecule (Acmc) in air-solvent films. Tensiometry provides measures of surface activity via limiting surface tension at the critical micelle concentration γmin and surface excess Γmax, (20) Palepu, R.; Gharibi, H.; Bloor, D. M.; Wyn-Jones, E. Langmuir 1993, 9, 110-112. (21) Callaghan, A.; Doyle, R.; Alexander, E.; Palepu, R. Langmuir 1993, 9, 3422-3426. (22) Gracie, K.; Turner, D.; Palepu, R. Can. J. Chem. 1996, 74, 1616-1625. (23) Carnero Ruiz, C. J. Colloid Interface Sci. 2000, 221, 262-267. (24) Carnero Ruiz, C. Colloid Polym. Sci. 1999, 277, 701-707. (25) Carnero Ruiz, C.; Molina-Bolivar, J. A.; Aguiar, J. Langmuir 2001, 17, 6831-6840. (26) Nagarajan, R.; Wang, C.-C. Langmuir 2000, 16, 5242-5251. (27) Ray, A. J. Am. Chem. Soc. 1969, 91, 6511-6512. (28) Ray, A.; Nemethy, G. J. Phys. Chem. 1971, 75, 809-815.

10.1021/la0617356 CCC: $33.50 © 2006 American Chemical Society Published on Web 11/02/2006

11188 Langmuir, Vol. 22, No. 26, 2006

through the Gibbs adsorption isotherm.29 Neutron reflection is also sensitive to Γ, providing a consistency check on γ measurements, but importantly yielding interfacial film thickness through model fitting analyses.30,31 As mentioned above, two model nonionics were used: the normal h-chain versions for surface tension measurements and chain-deuterated analogues for NR experiments. The NR contrast variation technique was employed so that experiments were conducted in null reflecting solvents (air contrast match water (ACMW) and air contrast match solvent (ACMS)), which are appropriate mixtures of H2O, D2O, and EG. By using these ACM solvents, the neutron reflection signal arises principally from the deuterated adsorbed layer. At first sight, it might be thought that perfect surfactants for such a study would have been C8E4OH and C10E4OH (both hand d-versions). However, the synthesis of common d-CiEj’s (needed for NR) naturally leads to a distribution of headgroup lengths (EO numbers) and therefore a spread of solvophilicities. Indeed, it is very difficult to generate a strictly monodisperse EO surfactant. As the composition of the EO headgroup is known to have an effect on interfacial properties,32,33 it is clearly desirable to obtain samples with well-defined EO lengths for meaningful physicochemical studies. The use of “unusual” CiE4OMe surfactants here presents the important advantage of having a well-defined and monodisperse EO group length. On the other hand, the CH3-ended compounds do introduce some minor complications. Conroy et al. have studied the effect of changing the terminal group of a nonionic surfactant from OH to CH3 on phase behavior.34 It appeared that the major effect is on the cloud point (Tc) and the location of mesophases. In general, Tc was found to decrease by changing OH with CH3 (Tc ) 3.6 °C and 0.2 °C for C12E4OH and C12E4OMe, respectively). These authors also pointed out the existence of two different clouding mechanisms, one due to critical dewetting of EO chains, and the second thought to be due to a micelle shape change from rod to disk. Hence, on balance, it can be claimed that the benefits of employing -OCH3 surfactants for this current study outweigh the limitations. Unsurprisingly, the surface tension and cmc behavior was found to be strongly solvent-dependent. On the other hand, no significant effects were found on the adsorption parameters limiting surface excess Γcmc and area per molecule Acmc. Switching solvent from water to a glycol is known to reduce the surfactant solvophobicity; in general, with ionic surfactants in glycols, micellization is strongly suppressed, and in some cases even extinguished.14-28 Therefore, since both aggregation and adsorption are believed to be driven by the common solvophobic (hydrophobic) effect, it might be expected that the limiting surface excess at the air-solvent interface would also be somewhat reduced in glycols as compared to pure water. For these model nonionic surfactants, a surprising finding is revealed here: solvent effects on the adsorption parameters Γmax and Acmc are apparently only very weak. This points to the significance of a general solvophobicity rather than hydrophobicity alone when considering surfactant properties. Estimation of adsorption and aggregation (29) Gibbs, J. W. The Collected Works of J. W. Gibbs; Longmans, Green: New York, 1931; Vol. I, p 219. (30) Lu, J. R.; Thomas, R. K.; Penfold, J. AdV. Colloid Interface Sci. 2000, 84, 143-304. (31) http://www.isis.rl.ac.uk. (32) Van Os, N. M.; Haak, J. R.; Rupert, L. A. M. Physicochemical Properties Of Selected Anionic, Cationic and Nonionic Surfactants; Elsevier: New York, 1993. (33) Nazrul Islam, M.; Kato, T. Langmuir 2005, 1-6. (34) Conroy, J. P.; Hall, C.; Leng, C. A.; Rendall, K.; Tiddy, G. J. T.; Walsh, J.; Lindblom, G. Prog. Colloid Polym. Sci. 1990, 82, 253-262.

Seguin et al.

free energies (∆G°mic, ∆G°ad) for the model surfactants in these different solvents reinforces this assertion. Supporting information accompanies the work,35 and reference to this section is frequently made in the text. Experimental Section Materials. C8H17O(C2H4O)CH3 (C8E4OMe), C10H21O(C2H4O)CH3 (C10E4OMe), and their chain-deuterated analogues C8D17O(C2H4O)CH3 (dC8E4OMe) and C10D21O(C2H4O)CH3 (dC10E4OMe) were synthesized by literature methods described in the Supporting Information.35 H2O of resistivity 18.2 MΩ cm was taken from an Elga water system and deuterium oxide was from Goss Scientific (distilled prior to use). Ethylene glycol was obtained from Aldrich (99+% GC). Water/glycol solvent mixtures were made up at 50:50 wt % by mass. The purities of h-solvents were checked by surface tension.35 Glassware was cleaned; was soaked in Micro 90 cleaning solution; was rinsed repeatedly with tap water, deionized water, ethanol, and acetone; and then was kept in an oven prior to use. After trial investigations, studies of the C8E4OMe in pure EG were discontinued, owing to a phase separation at 25 °C over the concentration range of interest. The origin of this transition is unknown, but it was not observed in the longer chain length C10E4OMe analogue which gave homogeneous solutions in EG at 25 °C. Surface Tension Measurements. The air/solution surface tensions were measured with a Du Nou¨y ring tensiometer (K12, Kru¨ss) at 25 °C ( 0.2 °C. The instrument was calibrated with aqueous absolute ethanol solutions as a function of mole fraction XEtOH: agreement with literature values was typically (0.2 mN m-1). Experimental details and correction procedures and details of the quantitative analyses of surface tension measurements are given in the Supporting Information.35 Neutron Reflection (NR). Technical details of NR and data analysis can be found in the Supporting Information;35 these experiments were performed on the SURF and CRISP reflectometers at the neutron spallation source ISIS (Rutherford Appleton Laboratory, Didcot, United Kingdom). The specular neutron reflection, R(Q), was measured normal to the air/solvent interface at 25° ( 0.2 °C. Reproducibility was checked for certain repeat samples and also between reflectometers. Solutions of deuterated surfactants dC8E4OMe and dC10E4OMe were prepared in air contrast matched water (ACMW, 92% H2O and 8% D2O volume fraction), air contrast matched solvent (ACMS, 50/50 by weight water/glycol 50.4% H2O, 2.6% D2O, and 47% h-EG volume fraction), and pure h-EG.

Results and Discussion Tensiometry. Plots of the surface tension (γ), showing clear cmc break points, for C10E4OMe in various solvents (water, a binary water-EG mixture, and pure EG) are shown in Figure 1a. Surface excesses versus the reduced concentration ratio c/cmc derived from these tension data are shown in Figure 1b. The cmc’s are listed in Table 1, as well as values for limiting surface tensions (γmin), surface excesses (Γmax), and the effective areas per molecule (Acmc), at the respective cmc’s. These latter three quantities were obtained from analyses of quadratic functions fitted to pre-cmc tension curves, as explained in ref 35. For the C8E4OMe surfactant, γ-ln c data can be found in the Supporting Information,35 and values derived from Gibbs analyses are listed in Table 1; agreement with previous work3 was good. a. General Trends. Unsurprisingly, varying the type of the solvent appears to have a significant effect on the surface tension behavior and cmc, which increases following the order water < water/EG < EG.36,37 Micelle formation in the presence of ethylene glycol would be expected to be driven by a weaker solvophobic (35) See Supporting Information. (36) Rodriguez, A.; Munoz, M.; Graciani, M. d. M.; Chacon, S. F.; Moya, M. L. Langmuir 2004, 20, 9945-9952. (37) Nagarajan, R.; Wang, C.-C. J. Colloid Interface Sci. 1996, 178, 471482.

Adsorption BehaVior of Nonionic Surfactants

Langmuir, Vol. 22, No. 26, 2006 11189

length.14,39-42 The cmc’s obtained for these model nonionics are in good agreement with previous work on related -OH bearing nonionic surfactants of the kind CiEj. In water, for C8E4OH and C10E4OH, respectively, cmc ) 7.1 mM and 0.68 mM32 (see Table 1, for C8E4OMe and C10E4OMe, respectively, cmc ) 6.3 mM and 0.6 mM). The standard free energies of micellization ∆G°mic were calculated as proposed by Rosen et al.35,42 using eq 1

∆G°mic ) RT ln cmc

Figure 1. (a) Surface tensions of C10E4OMe as a function of surfactant concentration, at 25 °C, in H2O, H2O/EG, and EG. (b) Surface excess curves of C10E4OMe at 25 °C, in H2O, H2O/EG, and EG derived from the tension data in Figure 1a. Table 1. Critical Micellar Concentration, Surface Tension, Surface Excess, Area per Molecule, and Standard Free Energy of Micellization of C8E4OMe and C10E4OMe, in Different Solvents, at 25 °Ca solvent surfactant

H2O

H2O/EG

EG

C8E4OMe C10E4OMe C8E4OMe C10E4OMe C10E4OMe

cmc/(10-3 mol dm-3) 6.34 γpure solvent/(mN m-1) 71 γmin/(mN m-1) 26.3 Γmax/(10-6 mol m2) 4.0 Acmc/Å2 42 ∆G°mic/(kJ mol-1) -12.5 ∆G°ad/(kJ mol-1) -23.8

0.60 71 28.8 3.9 43 -18.4 -29.3

11.39 55 29.3 3.8 44 -11.1 -17.9

1.54 55 32.1 3.0 55 -16.0 -23.6

11.60 46 22.9 4.0 41 -11.0 -16.7

a Obtained through surface tension measurements. Uncertainties: ∆γ ( 0.20 mN m-1; γcmc ( 0.5 mN m-1; Γcmc ( 0.5 × 10-6 mol m-2; Acmc ( 6 Å2.

effect as compared to water, accounting for the increase in cmc.21,27,38 The limiting surface tension γmin is weakly solvent dependent, being ∼6 mN m-1 lower in pure EG than in water. However, because of the inherent differences in γsolvent (see Table 1), it is difficult to draw any firm conclusions from the γmin values in different solvents. Critical Micelle Concentrations. As seen in Table 1, for C10E4OMe, the cmc is found to increase when the solvent type changes from pure water to pure EG.28,37 As expected, in a given solvent, the cmc is found to decrease with increasing surfactant chain (38) Aramaki, K.; Olsson, U.; Yamaguchi, Y.; Kunieda, H. Langmuir 1999, 15, 6226-6232.

(1)

where cmc is in mol dm-3, R is in J mol-1 K-1, and ∆G° is in kJ mol-1. The values obtained show that micellization is energetically less favorable in EG than in water (∆G°mic ) -11 kJ mol-1 and -18.4 kJ mol-1, ratio ∆G°mic water/∆G°mic EG 1.7). As expected, ∆G°mic in the H2O/EG mixtures is found to be between the values for both pure solvents (see Table 1). By comparing ∆G°mic for the two different surfactants, it is clear that the free-energy contribution per -CH2 group (solvophobicity) is only weakly dependent on solvent type (2.9 kJ mol-1 and 2.5 kJ mol-1 in H2O and H2O/EG, respectively). These values and trends in ∆G°mic are consistent with previous work40,42 on CiEj surfactants (energy contribution per -CH2 group ∼ 2.85 kJ mol-1). Since the surfactants under study here end with -OCH3, the absolute free energies might to be different to those quoted for normal CiEj-OH analogues. The closest comparison to C8E4OMe and C10E4OMe in water (∆G°mic ) -12.5 kJ mol-1 and -18.4 kJ mol-1, Table 1), generated using this method35 is with C12E4, for which Rosen et al. determined ∆G°mic ) -23.9 kJ mol-1 at 25 °C.42 On the basis of the individual -CH2 group contributions, predicted values for the analogues of C8E4OMe and C10E4OMe (“C8E4” and “C10E4”) would be ∆G°mic -12.5 kJ mol-1 and -18.2 kJ mol-1, which compare favorably with the actual values listed in Table 1. These rough estimates point to only minimal influences on the tendency to micellize on swapping -OH for -OCH3 in the hydrophilic chains. On the basis of comparisons of these thermodynamic parameters, it would appear the limitations of employing -OCH3 surfactants for this study are only second-order effects, and they fit within the general pattern of behavior established for CiEj-OH’s.32 Using these parameters for C10E4OMe, the micellization freeenergy penalty of switching water for EG (∆G°mic EG - ∆G°mic H2O) can be confidently estimated as ∼ +7 kJ mol-1, representing about 40% (7/18) of the thermodynamic solvophobicity that could be attributed to this surfactant in water. The significance of this is that for nonionic ethylene oxide surfactants alternative hydrogen bonding solvents can be selected, in which classic aqueous phase behavior (adsorption, aggregation, and perhaps mesophase formation) would be expected, albeit at higher concentrations compared to that observed in water. Interestingly, despite attempts with different -OH and -OMe surfactants, it has not been possible to observe significant surface pressures or aggregation (by small-angle neutron scattering) in the longer chain length propylene glycol solvent (PG). This is consistent with a higher monomer solubility for the surfactants in PG compared to EG, leading to an absence of micellization (∆G°mic close to zero or positive). Hence, for nonionic surfactants to express classic (39) Hsiao, L.; Dunning, H. N.; Lorenz, P. B. J. Phys. Chem. 1956, 60, 657660. (40) Meguro, K.; Takasawa, Y.; Kawahashi, N.; Tabata, Y.; Ueno, M. J. Colloid Interface Sci. 1981, 83, 50-56. (41) Telgmann, T.; Kaatze, U. J. Phys. Chem. A 2000, 104, 4846-4856. (42) Rosen, M. J.; Cohen, A. W.; Dahanayake, M.; Hua, X.-y. J. Phys. Chem. 1982, 86, 541-545.

11190 Langmuir, Vol. 22, No. 26, 2006

Seguin et al.

Table 2. Fitting Parameters Obtained from the Neutron Reflection Data of 2 × cmc Surfactant Solutions of dC8E4OMe and dC10E4OMe, in Different Solvents, at 25 °Ca solvent surfactant c/(10-2 mol dm-3) τ/Å F(z)/(10-6 Å2) As/Å2 Γ/(10-6 mol m-2)

ACMW

ACMS

EG

C8E4OMe C10E4OMe C8E4OMe C10E4OMe C10E4OMe 1.3 20 2.1 43 3.9

0.1 20 3.0 37 4.5

2.3 22 1.8 48 3.5

0.3 21 2.7 41 4.1

2.3 24 2.4 38 4.4

a ACMW is air contrast matched water; ACMS is air contrast matched water/EG mixed solvent. Uncertainties: τ ( 2 Å on water, (4 Å on solvents; As ( 3 Å2; Γ ( 0.2 × 10-6 mol m-2.

Table 3. Comparison of Area per Molecule and Surface Excess at the cmc of C10E4OMe Derived from Tensiometry and Neutron Reflectiona Γcmc/(10-6 mol m-2)

Acmc/Å2 solvent

ST

NR

ST

NR

H2O H2O/EG EG

43 55 41

37 41 38

3.9 3.0 4.0

4.5 4.1 4.4

a Uncertainties: Γ -6 mol m-2 by ST and 0.2 × 10-6 cmc ( 0.5 × 10 mol m-2 by NR; Acmc ( 6 Å2 (ST) and ( 3 Å2 (NR).

behavior in a glycol solvent, it would appear that a reasonable upper threshold would be ∆G°mic ∼ -10 kJ mol-1. b. Adsorption BehaVior. On the basis of cmc values (Table 1), and the surface excess curves derived from γ-ln(c) profiles (Figure 1b and Figures 6, 8, and 10 in Supporting Information35), maximum adsorption is achieved at lower concentrations in water compared to the other solvents. This is in line with expectations of increased surfactant solvophilicity in H2O/EG and EG compared to pure water. Surprisingly, comparing the complete set of five different solvent and surfactant conditions, limiting surface excesses at the respective cmc’s appear to be essentially independent of solvent type (with one exception discussed below). As suggested by molecular areas listed in Table 1, limiting interfacial packing is hardly affected by solvent type; this is out of kilter with the strong effects of solvent on cmc. The only outlier is the C10 surfactant in H2O/EG (Acmc 55 Å2, Γmax 3.0 × 10-6 mol m-2). Measurements of γ-ln(c) and generation of Γ(c) profiles were repeated with freshly made samples, showing good reproducibility. This discrepancy in C10 H2O/EG could be attributed to a weak chain length effect, since the differences (44 Å2 for C8 versus 55 Å2 for C10) are indeed just outside the errors (see example error bars in Figure 1b). Unfortunately, for C10 in pure EG, Acmc is very similar to the value for pure water, suggesting that the H2O/EG result cannot be ascribed to a solvent effect only. Hence, on the basis of the available results and reproducibility checks, it is not possible to make any significance of the difference in the surface tension-derived Γmax or Acmc between the C10 surfactant in H2O/EG as compared to the other four systems. (Below, it is shown that neutron reflection measurements are more accurate; NR data for the C10 surfactant in H2O/EG fall into line with other samples, Table 3). Nonetheless, despite the large effects on cmc (bulk solubility), there does not appear to be a notable change in surface packing or adsorption densities as a function of solvent type. The standard free energies of adsorption ∆G°ad in kJ mol-1 have been estimated using the expression given by Rosen et al.35,42

∆G°ad ) RT ln cmc - 6.023 × 10-1 πcmcAcmc

(2)

πcmc ) γsolvent - γcmc

(3)

where cmc is in mol dm-3, πcmc is in mN m-1, and Acmc is in nm2. The values obtained suggest that adsorption is less favorable in EG than in water (for C10E4OMe ∆G°ad ) -16.7 kJ mol-1 and -29.3 kJ mol-1 in H2O and EG, ratio ∆G°ad water/∆G°ad EG 1.7). As expected, the values in H2O/EG are found to be between both pure solvents (see Table 1). In terms of adsorption, the thermodynamic solvophobicity penalty on swapping water for EG (∆G°ad EG - ∆G°ad H2O) appears to be ∼ +12 kJ mol-1; comparing with ∼ +7 kJ mol-1 for micellization noted above. In terms of direct comparison with related literature, ∆G°ad for C12E4 at 25 °C was quoted by Rosen et al.35,42 as -35.9 kJ mol-1. Therefore, as before for micellization, these ∆G°ad values listed in Table 1 fit well into the general pattern of behavior and orders of magnitude, documented for standard CiEj surfactants in the literature.42 Having established correlations between thermodynamic properties of dilute solutions, interfacial structures in terms of film thicknesses and surface layer densities (surface excess, area per molecule) have been studied by neutron reflection NR, as described below. Neutron Reflection Measurements. Typical R(Q) data are shown in Figures 2 and 3 along with fits to the single-layer model for C8E4OMe at 2 × cmc in water and in water/glycol mixture. The fitted parameters, thickness τ and area per molecule As, are summarized in Table 2. Figure 3 shows similar results for C10E4OMe in water, in water/glycol mixture, and in pure EG solvent. It is clear from Figures 2 and 3 that the R(Q) reflectivity profiles are hardly affected by solvent, suggesting broadly similar layer structures on the three different types of subphase. This is also borne out by the fitting analyses (results in Table 2). Lu et al. have carried out extensive NR investigations of regular CiEj surfactants;30 the results for surface layer thicknesses and moleclar areas of these -OCH3 surfactants compare favorably (e.g., regular C12E4 at the cmc As ) 44 Å2).30 Therefore, the values of these structural parameters determined by NR for CiE4OMe surfactants compare well with literature on closely related systems. First the layer thickness, which is essentially linked to the rate of the R(Q) decay, is approximately independent of solvent and surfactant type. The implication is that the interfaces are monomolecular layers. Despite the two carbon increase in chain length, the C8 and C10 films appear to be the same thickness; this may be put down to a combination of resolution limitations and the thermal roughening of liquid surfaces, which also contribute to the uncertainties. Another possibility is that the two films may be penetrated to a lesser or greater extent by solvent; this cannot be ruled out in the present case owing to the use of ACM solvents, and only further contrast variation experiments could resolve the issue. As shown elsewhere,35 fitted scattering length density (F) and thickness (τ) values can be combined to yield surface excesses Γ and corresponding molecular area As. The striking thing about these derived values given in Table 2 is the similarity, no matter what solvent or surfactant. This suggests that, for such model surfactants, interfacial film structure and 2-D packing are hardly affected by solvent type. Comparison of Results. Table 3 shows results obtained for C10E4OMe with surface tension (ST) and neutron reflection (NR) in all three solvents. In both pure solvents, water and EG, the values are in good agreement with respect of the uncertainties. However, in the water/glycol mixture, the adsorption parameters differ, depending on the method. These discrepancies may reflect genuine, albeit minor, differences in adsorption. However, it can

Adsorption BehaVior of Nonionic Surfactants

Figure 2. NR data for C8E4OMe at 2 × cmc in different solvents at 25 °C. ACMW is air contrast matched water; ACMS is air contrast matched water/EG mixed solvent. The solid lines are fits to the one-layer model.

Figure 3. NR data for C10E4OMe at 2 × cmc, in different solvents, at 25 °C. ACMW is air contrast matched water; ACMS is air contrast matched water/EG mixed solvent; EG is ethylene glycol. The solid lines are fits to the one-layer model.

be claimed that NR experiments are more reliable, since the method is sensitive to the adsorption density, which is essentially linked to the absolute reflectivity (intensity). On the other hand, interpretation of γ-ln c curves is necessarily reliant on an adsorption isotherm and in the case of the Gibbs equation, the local tangent to the γ-ln c curve. Taking these issues into consideration tends to favor NR over ST for providing more robust adsorption parameters. The surprising conclusion is that to a first approximation cmc surface excesses for these model nonionics are essentially independent of solvent type for water, EG, and their mixture. To emphasize the point, Figure 4 depicts these effective molecular volumes for the systems under study at the cmc. The limiting molecular volumes of the two surfactants can be estimated as 517 and 575 Å3 for the C8E4OMe and C10E4OMe, respectively (Table 2 in Supporting Information35). Therefore, these effective adsorption volumes in Figure 4 lie between 1.3 and 2.0 times the finite molecular sizes (C10E4OMe in water 1.3, C8E4OMe in H2O/EG 2.0). For C10E4OMe, this ratio of adsorption volume to actual volume is 1.3, 1.5, and 1.6 for the water, water/EG, and pure EG, highlighting that the effective surface packing is only weakly solvent dependent. In other words, the surface packing at the EG-air interface is only about 20% less efficient than in the water-air film.

Summary and Conclusions Adsorption behavior and interfacial properties of two nonionic surfactants in H-bonding solvents have been studied using two different methods, surface tension and neutron reflection. The

Langmuir, Vol. 22, No. 26, 2006 11191

Figure 4. Representation of effective molecular volumes Vm for both surfactants C8E4OMe and C10E4OMe at the cmc in different solvents.

purpose was to investigate the effects on surfactant properties of varying solvent from pure water to pure nonaqueous ethylene glycol. These surfactants were nonionics of the type CiEjOMe, bearing a methyl at the end of the hydrophilic headgroup instead of the more common OH. The most surprising observations are the similarity in behavior as a function of solvent type: (a) The free energies of micellization ∆G°mic and adsorption ∆G°ad only weakly depend on the type of solvent studied here. The free-energy penalties on swapping water for EG as the solvent were estimated as ∼ +7 and ∼ +12 kJ mol-1 for micellization and adsorption, respectively. (b) Adsorption parameters, limiting surface excesses, and molecular areas are also broadly the same, showing no notable solvent effect. (c) Surfactant film thicknesses, surface densities, and molecular packing do not depend notably on solvent, being consistent with close-packed macromolecular layers. The results show that ethylene glycol (EG) can be considered a nonaqueous alternative solvent, in which classic surfactant behavior can be observed with nonionics, and broadly these properties are retained on dilution with water. Few studies have been published with nonionic surfactants in EG,25,26,28,36 however, work with ionic surfactants in this solvent is far more extensive.14-28 Most of them suggest that for ionic surfactants, above a certain level of added EG (about 60 wt % EG in water),24,36 aggregation is suppressed, and cmc determinations are very difficult.19,43 Nonetheless, in general, an increase in cmc of common ionic surfactants is observed in these EG/ water blended solvents as well as lower aggregation numbers.24,26,37 More generally, however, the findings shed new light on the “hydrophobic effect”, pointing to deficiencies if surfactant properties are only described with respect to hydrophobicity. In terms of all-important H-bonding solvents, solvophobicity is the more relevant term, and the results presented here could be used as the basis of a “solvophobicity scale”. This may have ramifications for modeling surfactant behavior, since it does not seem necessary to consider water as the only solvent; simple models for H-bonding fluids could be useful to mimic continuous solvent media without loss of generality. These results act as a starting point for a broader-based study to generate a solvophobicity index for assisting rational design of new surfactants for applications in nonaqueous solvents. (43) Binana-Limbele, W.; Zana, R. Colloid Polym. Sci. 1989, 267, 440-447.

11192 Langmuir, Vol. 22, No. 26, 2006

Acknowledgment. C.S. thanks Huntsman and University of Bristol, School of Chemistry for a scholarship. We also thank CLRC and (Rutherford Appleton Laboratory) for allocation of beam time at ISIS and ILL and contributions toward consumables and travel.

Seguin et al.

Supporting Information Available: Details of the synthesis of the nonionic surfactants, theoretical background, and extensive NR data and analysis.35 This material is available free of charge via the Internet at http://pubs.acs.org. LA0617356