Are Hydrotropes Distinct from Surfactants? - Langmuir (ACS

---

ARTICLE pubs.acs.org/Langmuir

Are Hydrotropes Distinct from Surfactants? Marios Hopkins Hatzopoulos,† Julian Eastoe,*,† Peter J. Dowding,‡ Sarah E. Rogers,§ Richard Heenan,§ and Robert Dyer|| †

School of Chemistry, University of Bristol, Cantock’s Close, Bristol, BS8 1TS, United Kingdom Infineum UK Ltd., Milton Hill Business & Technology Centre, Abingdon, Oxfordshire OX13 6BB, United Kingdom § ISIS-STFC, Rutherford Appleton Laboratory, Chilton, Oxon OX11 0QX, United Kingdom Kr€uss U.K., School of Chemistry, University of Bristol, Cantock’s Close, Bristol, BS8 1TS, United Kingdom

)

‡

bS Supporting Information ABSTRACT: The physicochemical properties of a homologous series of sodium p-nalkylbenzoates have been investigated. The objective was to determine whether there is a clear transition point from hydrotropic to surfactant-like behavior with increasing alkyl chain length n, so as to shed clear light on the aggregation mechanism of so-called “hydrotropes”. Electrical conductivity measurements were used for a first estimation of the critical aggregation concentrations (cac). As for classical surfactants, log(cac) depends on alkyl chain length n, but two branches of behavior were observed: one having a gradient typical of long chain fatty acid salts and the other with a more shallow dependence. Surface tension (γ) measurements of high purity aqueous solutions were used to generate limiting headgroup areas Acac, which were in the range (40
50 Å2) being consistent with monolayer formation. Small-angle neutron scattering conclusively shows that the lower chain length homologues (classed as hydrotropes) exhibit sharp transitions in aggregation as a function of bulk concentration, typical of regular surfactants. As such, there is little to suggest from this study that hydrotropes differ in association behavior from regular surfactants.

’ INTRODUCTION Hydrotropes have been identified as amphiphilic organic compounds bearing a close structural resemblance to classical surfactants. In terms of molecular structure, these so-called hydrotropes comprise small organic groups, for example, an aromatic ring, whereas surfactants normally bear longer C8
C20 alkyl chains. A good example of a class of hydrotropes is the n-alkylbenzoate series (CnBenz), shown in Figure 1: the short chain C0
C4Benz homologues may be considered as hydrotropes, whereas the longer chain C7
C8 analogues resemble more closely true surfactants. For this series, it might be expected that the physicochemical properties of aqueous solutions evolve as a function of chain length. A recent review1 summarizes the present state of understanding about the nature and function of hydrotropes. As outlined below, there remain controversies in the literature concerning the properties and action of hydrotropes in comparison to the well understood behavior of classical surfactants.2
5 Neuberg6 noticed that certain organic salts increased the solubility of organic compounds in solution; this phenomenon was termed hydrotropy. As such, hydrotropy is generally accepted as occurring above a species-specific concentration termed the minimum hydrotrope concentration (MHC), also referred to as a critical aggregation concentration (cac) here, being typically ∼10
1 M. Hydrotropes can be ionic or nonionic and exist in a range of structures. The hydrophobic moieties can be a short alkyl tail,7 saturated, unsaturated, and aromatic rings, and like surfactants they reduce surface r 2011 American Chemical Society

Figure 1. Sodium p-n-alkylbenzoates.

tension. Due to the small hydrophobic units and the high concentrations needed for solubilization, some have considered them to be a separate class of amphiphiles, distinct from classical surfactants. There are three main approaches accounting for hydrotropic action: (i) Formation of a complex between solutes and hydrotropes.8 (ii) Hydrotropes may change local solvent structure and can be therefore considered as structure makers or breakers.9 (iii) The most supported hypothesis is that of self-association, generating aggregates that may act similarly to micelles above the cac.10 Balasubramanian et al.10 support the view of a gradual stepwise association of hydrotropes, as do studies by Loh et al.11 when hydrotropes are used with water-soluble polymers. Both groups Received: July 8, 2011 Revised: August 26, 2011 Published: August 29, 2011 12346

dx.doi.org/10.1021/la2025846 | Langmuir 2011, 27, 12346–12353

Langmuir proposed that changes in properties smeared out over a range of concentrations and were consistent with continual stepwise aggregation, rather than “on
off” association seen with classical surfactants when crossing the cmc. From X-ray diffraction studies of solid hydrotropes, Balasubramanian et al.12 concluded that lamellar-type association structures should be formed in dilute hydrotrope solutions. Triolo et al.13 investigated a range of hydrotropes, including compounds studied by Balasubramanian et al.,12 using small-angle neutron scattering (SANS) to probe aggregation in aqueous solutions. Local clustering and aggregation was detected in D2O solutions; the data could be interpreted in terms a of low aggregation number spheroidal or ellipsoidal form factor (P(Q)) multiplied by a charge repulsion structure factor (S(Q)). These SANS results13 pointed to links between the properties of hydrotropes and surfactants. The problem with previous work is a lack of systematic structural variation of the amphiphiles, which is the main feature of this new paper. The inspiration for this work was taken from a publication of Balasubramanian et al.,12 who posed the question “when does the switch from hydrotropy to micellar behavior occur?” This issue was tackled in a further publication by the same author14 examining the single chain linear sodium alkylbenzenesulfonate homologous series (p-toluenesulfonate to p-octylbenzenesulfonate) by pyrene solubilization, fluorescence, and surface tension measurements. The surface tension versus concentration (γ
ln(c)) curves for the lower homologues do not show sharp break points, whereas for p-pentylbenzenesulfonate and higher homologues classical γ
ln(c) curves characteristic of surfactants were observed. Hydrophobic domain micropolarity, estimated from pyrene fluorescence spectra, showed a marked lowering in polarity for p-pentylbenzenesulfonate with increasing chain length through the series. It was concluded that there may not be a point of departure, where the physical properties of hydrotrope solutions are different from those of surfactants. As stated above, this work seeks to clarify the relationship between surfactants and hydrotropes by using a well-defined pure homologous series (Figure 1) and by performing a systematic study. Electrical conductivity was used to provide a first estimate for any cac, local changes in molecular environments were sensed using proton shift NMR, and aggregation was studied using SANS. The point of these systematic studies is to understand whether hydrotropes of this kind may be considered a distinct class of amphiphiles separate from surfactants. Compared to probe dye measurements, which give secondhand reports of environmental changes and can be unreliable, here direct methods have been employed which are sensitive to nuclear properties of the hydrotrope/surfactant molecules themselves. As such, NMR and SANS give unambiguous reports of environmental changes, solution state structure, and aggregation, and provide irrefutable evidence which can be clearly interpreted.

ARTICLE

Methods. Sodium salts were prepared by refluxing equimolar amounts of acid and NaOH for 8 h, in approximately 1.5 times the minimum amount of ethanol required to dissolve the acid. The hot solutions were passed over carbon black, and the collected filtrate was allowed to boil while distilling excess ethanol. The solutions were then allowed to cool; the products were collected by filtration and washed with cold ethanol. The products were then dissolved in deionized water, any excess acid was extracted by shaking with distilled diethyl ether, and then the products were collected by evaporation of the aqueous phase. The powders were placed in a vacuum oven at 40 °C for 24 h over refreshed phosphorus pentoxide. Electrical conductivities were determined using a Jenway model 4510 conductivity/TDS meter with temperature controlled at 25 °C ( 0.1 (thermostatic water bath). NMR proton shifts were measured using a Varian 400 MHz instrument at 25 °C, relative to HDO at 4.75 ppm. Surface tension measurements were measured using the Wilhelmy plate method with a Kr€uss Easy Dyne instrument at 25 °C. Protocols for purification of the amphiphilic salts are given in the Supporting Information. Solutions were prepared with the necessary minimum amount of the chelating agent EDTA to sequester the presence of divalent ions (Supporting Information). All glassware was prewashed with 50% nitric acid and thoroughly rinsed with deionized water. The platinum plate was held over a blue flame until red hot after every measurement and then cooled to room temperature before reuse. SANS was carried out on the SANS2d instrument at ISIS-STFC, Rutherford Laboratory U.K. The experiments used a slightly modified instrument setup, so as to achieve a greater scattering vector (Q/Å
1) range. The incident wavelength range was 2
14 Å, giving rise to a Q range of 0.007
1.1 Å
1. Q is defined as
 θ 4π sin 2 ð1Þ Q ¼ λ where θ is the scattered angle and λ is the incident neutron wavelength. The detector offset was vertically 150 mm and sideways 300 mm, with the detector distances at L1 = 4 m and L2 = 3 m. D2O was used to prepare the solutions, providing the necessary contrast. Samples were contained in 2 mm path length Hellma fused silica cells. The SANS data were normalized for transmission and subtraction of the empty cell and solvent background. The error in absolute intensities was within 5%, by measuring the intensity of a partially deuterated standard. The FISH multimodel program was employed for data analysis. The scattering profile curve is broadly described by IðQ Þ µ PðQ Þ SðQ Þ

where P(Q) is a form factor, which contains information about the size and shape of the aggregates, and S(Q) is a structure factor reporting on interactions. Trials with different plausible P(Q) functions were carried out. In the cases described here, the ellipsoid P(Q) model was used in final fits, in combination with a Hayter
Penfold charge repulsion S(Q).15 The ellipsoids are characterized by two principal radii R1 and R3 with the aspect ratio X: X ¼

’ EXPERIMENTAL SECTION Materials. All chemicals were purchased from Sigma Aldrich at the highest purity available. The sodium alkylbenzoates were synthesized from parent acids, with the exception of sodium benzoate, which was purchased in the highest available purity (>99.5%) and used as received. Sodium hydroxide was purchased from Fischer Scientific (97%). D2O for NMR and SANS experiments was purchased from Aldrich (99.9%). EDTA sodium salt, used in surface tension experiments, was also purchased from Aldrich (99.5%). Activated carbon was purchased from Lancaster and used as received.

ð2Þ

R3 R1

ð3Þ

Further parameters in the model include the volume fraction ϕ, the Debye length k
1, and the effective radius of the charged micelle RS(Q). An approximate value of k
1 value can be calculated by !1=2 2F 2 FI ð4Þ k¼ ε0 εr RT where F is the Faraday constant, F is the solvent density, I is the ionic strength, ε0 is the permittivity of free space, εr is the solvent dielectric 12347

dx.doi.org/10.1021/la2025846 |Langmuir 2011, 27, 12346–12353

Langmuir

ARTICLE

Figure 2. Electrical conductivity versus concentration for C8Benz and C0Benz (inset).

Table 1. Values of cac for n-Alkylbenzoates Determined by Electrical Conductivity (25 °C) CnBenz

cac/M

0

0.480

1

0.320

2

0.300

3 4

0.280 0.200

5

0.090

6

0.048

7

0.024

8

0.011

Figure 3. (a) Comparison of the behavior of log(cac/M) with hydrophobic chain length n for n-alkylbenzoates (b) and n-alkylcarboxylates (9). (b) Effective chain length (nring) of the benzene ring plotted against log(cac/M). The numbers indicate linear carbon chain lengths.

constant, R is the gas constant, and T is the temperature. More detail on SANS modeling is given in the Supporting Information. First approximation fits used hydrophobic tail lengths calculated using the Tanford equation16 (eq 5) lmax ¼ 1:5 þ 1:265n

ð5Þ

where lmax is the fully extended chain and n is the carbon number. It was assumed that a benzene ring is ∼3.5  CH2 in length. For the final analyses, the micellar radii R1,3 were also minimized.

’ RESULTS AND DISCUSSION Electrical Conductivity Measurements. Figure 2 shows the electrical conductivity profiles of sodium benzoate (C0Benz, “hydrotrope”) and sodium octylbenzoate (C8Benz, “surfactant”) solutions. From first inspection, it is apparent that C8Benz exhibits a break in conductivity behavior typical of classical surfactants.17 C0Benz also shows two distinct gradients; however, the transition is more of a curve than a clear break in conductivity. This curvature in conductivity behavior is noticeable for C2Benz and lower homologues. From C3Benz, and for higher homologues, the typical break in behavior is progressively more pronounced (see the Supporting Information). Evans17 examined a series of n-alkylsulfates by conductivity and noted that for the lower molecular weight compounds there was also a pronounced curvature, behavior attributed to the presence of “half developed” micelles. Mosquera et al.18 studied sodium hexylsulfate by conductivity and surface tension among other techniques and also observed ill-defined transition regions. They concluded

that such behavior can be accounted for by a micellar structure with large solvent exposure, that is, water penetration. Table 1 shows cac's as determined at 25 °C by electrical conductivity. It is interesting to note that an approximate halving of the cac for every methylene group added to the chains, which is behavior typical of classical surfactants, begins from C4Benz onward. Similar behavior was found for the alkylbenzenesulfonate series studied by Balasubramanian and Srinivas.14 Figure 3 shows the dependence of log(cac/M) on n, the carbon number of the hydrophobic tail. Klevens19 observed that this relationship is linear for surfactants and the gradient and intercept are series specific. These plots for n-alkylbenzoates exhibit two gradients, one for C0-C4Benz and another for the C4Benz
C8Benz analogues. The latter class has a similar gradient as found for linear sodium alkylcarboxylates, for which critical micelle concentration (cmc) values were also determined by conductivity.20 The linear fits are linear alkylcarboxylates

logðcmcÞ ¼
0:2832n þ 1:5145 ð6Þ

Cn Benz

logðcacÞ ¼
0:2866n þ 0:4179

ð7Þ

The difference in intercept can be ascribed to the fact that the benzene ring is not included in n, though it has a hydrophobic contribution. Assuming that hydrophobicity is universally additive, then neff could be defined as neff ¼ n þ nring

ð8Þ

where nring is the distinct hydrophobic contribution of the ring and n is the number of linear tail alkyl carbons. By applying the Klevens 12348

dx.doi.org/10.1021/la2025846 |Langmuir 2011, 27, 12346–12353

Langmuir

ARTICLE

Table 2. Parameters Derived from Surface Tension Measurements Using the Gibbs Equation in Terms of Activitya cac

γcac ( 0.1/

Γcac ( 0.1/

CnBenz

activity (α)

(mN m-1)

(10-6 mol m-2)

Acac ( 10%/Å2

3 4

0.270 0.160

47.0 45.5

3.2 3.9

51.4 43.1

5

0.078

43.1

4.0

41.1

6

0.044

40.7

4.4

44.1

7

0.020

41.2

4.1

40.3

8

0.009

37.3

4.0

41.8

a

For comparable analyses using concentrations, in place of activities see Table SI 2 in the Supporting Information.

Figure 4. Surface tension vs ln(α) of CnBenz homologues. Quadratics fitted to “pre-cac” data were used to determine adsorption parameters using the Gibbs equation.

equation (e.g., eq 6) to the cac values and substituting neff in place of n, the hydrophobic contribution of the benzene ring alone can be determined: then all compounds follow a Klevens-type line. Interestingly, this treatment reveals that the effective hydrophobicity of the benzene ring nring is not constant. Previous reports and studies with alkylbenzenesulfonates2,5,14,21
30 established that nring ∼ 3.5
CH2
groups. However, here, for this CnBenz series, the effective hydrophobicity of the ring nring is apparently dependent on n itself. As chain length is shortened from C8Benz to C4Benz, nring ∼3.8
CH2
groups, but there is clear point of departure at C4 where greater effects are seen for the shorter chains. In other words, with shorter substituent alkyl chains (Cn < 4), the aromatic ring packs a bigger and bigger hydrophobic punch, as compared to systems with longer chains (Cn > 4). This kind of behavior tends to suggest that the short chain hydrotropes have a notably different thermodynamic behavior as compared to longer chain “surfactants”. Hence, based on the analysis alone, it could be concluded that hydrotropes are a distinct class from surfactants. Surface Tension Measurements. For C3
C8Benz homologues, the γ
ln(c) plots exhibit typical curves normally identified with classical surfactant behavior: a smooth “pre-cac” tension decrease, up to a sharp break at a cac (cmc), followed by a “post-cac” plateau. Similar data were acquired for the C0
C2Benz compounds: the data show broadly the same features as in Figure 4 but suffered from irreproducibility and variability in derived adsorption parameters. These problems may be ascribed to the high background concentrations (>0.1 M), resulting in high ionic strengths compounded by a lower surface activity. The limiting tension γcac shows a chain length dependence, decreasing with increasing n, consistent with increased alkyl chain density lowering surface energy.31 The cac's from tension measurements (Table 2) are in good agreement with electrical conductivity determined values (Table 1). It is interesting to note the slight decrease in “pre-cac” gradient with decreasing chain length. The pre-cac tensions were analyzed in terms of the Gibbs adsorption equation Γ¼


1 dγ mRT ln α

ð9Þ

where Γ is the surface excess (mol m
2), m is a constant depending on ion dissociation (see below), and R and T represent the gas constant and temperature, respectively. Strictly, activity α should be used in place of concentration C. Here the

Table 3. Parameters Derived from Surface Tension Measurements for Standard Literature Surfactants (25 °C) Γcac/(10-6 mol m-2)

surfactant

Acac/Å2

ref

sodium alkylsufonates C10

3.2

52

37

C12

2.9

57

38

SDS

2.6

63

38

55 49

39 40

sodium alkylbenzenesulfonates 3.0 3.4

C8 C8a

a

C10a

3.9

43

40

C12a

3.7

45

40

Measurements carried out at 70 °C.

activity coefficients were estimated using standard procedures,32 and alternative analyses as a function of concentration C are presented and discussed in Supporting Information. The upshot is that derived adsorption parameters are essentially the same for both methods of analysis. The long-standing controversies over the correct value of m have now been clearly resolved for ionic surfactants in dilute solutions.33 By combination of surface tension and neutron reflectivity, it has been shown for ionic surfactants (10 mM, this value m = 2 has been taken to hold, so that it must be borne in mind when interpreting the curves. To extract the adsorption parameter Γ, the pre-cac curves were fitted to quadratics, and then using the obtained coefficients and associated derivatives surface excesses were derived. The limiting values for Γcac were used to estimate effective headgroup areas Acac given by Acac ¼

1 ΓN a

ð10Þ

with Na the Avogadro number. These values can be compared to with adsorption parameters for literature standard compounds shown in Table 3. Interestingly, it can be seen that the limiting adsorption parameters for the CnBenz series do not change greatly as a function of substituent chain length n. This suggests the molecular surface density is broadly constant; the chain length dependence of cac is identified with increasing hydrophobicity. Furthermore, the reduction in γcac with n follows increased surface chain density,4 12349

dx.doi.org/10.1021/la2025846 |Langmuir 2011, 27, 12346–12353

Langmuir

ARTICLE

Figure 6. Variation in Δδ/ppm of meta protons with concentration.

Figure 5. High chemical shift 1H NMR showing change in ring proton signals for C8Benz below and above the cac.

and longer alkyl chains pack more closely and have lower surface energies. These values may be compared with adsorption parameters for literature standards, as listed in Table 3. It can be readily appreciated that Γcac and Acac for the CnBenz compounds both compare favorably with similar anionic surfactants. Therefore, in contrast to conductivity behavior (Figure 3), there does not appear to be any clear distinction between the surface tension and air
water adsorption behavior of the longer (>C4) and shorter ( C4Benz may be consistent with a break in behavior seen in the conductivity study. However, if it is assumed that the smoother changes in NMR environment are an effect of increased smearing of the break point with decreasing hydrophobicity, as the solvent environment becomes less waterlike with increased organic solute, then there is nothing yet to suggest a point of departure from hydrotropes to surfactants. Small-Angle Neutron Scattering (SANS). SANS is a very good technique for probing directly the nature of amphiphile aggregates and other soft matter. As discussed previously, there has been a longstanding debate around the topic of hydrotrope action. Despite conventional “traditional” techniques suggesting the formation of aggregates, it can be often argued that the interpretations of these results remain speculative. SANS is an advanced technique that can definitively answer as to whether aggregation occurs. Figure 7 shows example SANS data for C4Benz D2O solutions as a function of concentration, expressed in multiples of the cac determined from conductivity measurements (Figure 3). At 4 cac, the scattering profile is characteristic of charged aggregates, with similar features as seen for anionic surfactants such as SDS above the cmc.43 Decreasing the concentration to 2 cac leads to a linked decrease in I(Q). At 1, cac the intensity is very weak and the clear charged S(Q) peak has been lost as seen in Figure 7 inset. (Note the log scale and recall that the concentrations of the study are based on k-derived cac values.) Then, below the threshold concentration at 0.5 cac, no scattering can be detected. A similar concentration dependence of I(Q) was noted for all other CnBenz systems studied (see the Supporting Information). For long chain homologues C8Benz and C6Benz, SANS signals as a function of concentration were consistent with a surfactant-like “on
off” aggregation crossing the cac (cmc). Furthermore, that picture was repeated with short chain homologues C0-C4Benz. 12350

dx.doi.org/10.1021/la2025846 |Langmuir 2011, 27, 12346–12353

Langmuir

ARTICLE

Table 4. Parameters Obtained by Fitting SANS Data to the Charged Ellipsoid Aggregate Model for the Homologous Benzoate Series at 4 caca P(Q)

S(Q)

CnBenz

R1/Å

X

RS(Q)

k/Å
1

8

15.3

1.6

22.3

0.073

6

11.9

1.8

16.7

0.11

4

9.8

1.9

10.2

0.15

2

5.5

2.0

9.2

0.34

0

4.8

1.2

5.5

0.46

a

Figure 7. SANS profiles of C4Benz in D2O at multiples of the cac. Inset: log
log plot of the SANS profile at 1 cac as determined by conductivity (Figure 3).

Figure 8. (a) SANS profiles for n-alkylbenzoates in D2O at 4 cac (C8Benz offset by +1, C6Benz offset by
1, C4Benz offset by
2, and C2Benz offset by
3 log units for clarity). (b) SANS profiles for C0Benz and C2Benz (magnified) in D2O at 4 cac.

Figure 8a shows scattering profiles of C0
C8Benz D2O solutions at 4 cac, showing broadly common behavior across the series. Increasing chain length leads to increasing I(Q), and S(Q) also becomes more pronounced, suggesting stronger aggregate interactions. The data were fitted using the protocol and model described in the Experimental Section, being consistent with charged prolate ellipsoidal aggregates at this elevated concentration of 4 cac (Table 4). The data and analyses taken together show continuous aggregation behavior across the homologous series: from the short chain “true hydrotropes” to the longer chain alkyl-hydrotropes, which would be identified as “true surfactants”.

Parameters for 2 cac can be found in Table SI 2 in the Supporting Information.

For SANS, no signals will be detected if aggregates are absent, and instead there is a chaotic nonstructured molecular solution. Simply put, SANS can detect the onset of aggregation with systems of this kind, and it is capable of distinguishing between the two extremes of a gradual stepwise association and a sharp switch “on
off” aggregation threshold. Recall that the conductivity-derived cac values were used as guides to make up the solutions, and it was a surprise to see how accurately these concentrations mapped onto the cac values detected by SANS. Hence, the utility of SANS as a direct structural probe for systems of this kind has been clearly demonstrated. Significantly, it can be seen that data from a bulk technique such as electrical conductivity and the environmental probe method of NMR as well as other approaches such as probe dye fluorescence14 could lead to misconceptions about the underlying physicochemical behavior in hydrotrope systems. There may be good reasons for the failure of classical techniques to detect clear cac's for short chain homologues as the concentration is raised; approaching the cac, the ionic strength may come into play, obscuring the recorded values or signals suggesting smeared transitions. However, SANS is not sensitive to ionic strength but merely the presence of local H-domains in the D2O solvent. SANS data support the view that compounds of this kind, often termed “hydrotropes”, do exhibit classic features of normal surfactants above threshold concentrations in water. Comparing these results with previous literature, Balasubramanian et al.12 studied X-ray diffraction from solid hydrotropes. They found no evidence of π-stacking and issued warning to this speculation on the aggregation behavior of aromatic hydrotropes. They also concluded that hydrotropes do not form spherical or near-spherical aggregates but instead layered structures. These conclusions were drawn without taking into consideration the findings of Triolo et al.13 who studied a number of small hydrotropes by SANS. This somewhat neglected report clearly shows that aggregation does take place for a number of anionic hydrotropes. The scattering for low concentration samples fitted a spherical P(Q) model, changing to a prolate ellipsoid form at higher concentrations. Further support of aggregation structures formed by hydrotropes was presented in a more recent SANS study of sodium butylbenzenesulfonate by Aswal et al.44 Scattering was detected at 7
8 times the MHC (0.15 M) as determined by Balasubramanian et al.14

’ CONCLUSIONS Physicochemical studies of a well-defined homologous series of amphiphilic sodium p-n-alkylbenzoates reveal close links between the properties of short chain “hydrotropes” (C0-C4Benz) and longer chain analogues (C5
C8Benz), which would be more 12351

dx.doi.org/10.1021/la2025846 |Langmuir 2011, 27, 12346–12353

Langmuir readily identified as “surfactants”. The view that emerges concerning the onset of aggregation, in terms of a stepwise association versus a mass action “on
off” mechanism, is technique dependent. Electrical conductivity k was used to provide a first estimate of any critical aggregation concentrations (cac). For long chain analogues, a welldefined switch over in behavior was noted as a function of concentration (C). The breaks in k
C curves with two distinctly different gradients were clearly evident, as commonly seen with classical ionic surfactants.17 For shorter chain homologues (C0
C4Benz), the break points in k
C curves could be detected, but they certainly were not as clear. Hence, it might be reasonable to invoke a model of continual stepwise growth based on conductivity data alone. Treatment of the cac values assuming a Klevens type relationship reveals two regimes of thermodynamic behavior with a lower gradient d log(cac)/dn for chain lengths < C4 compared to higher chain homologues. For the longer chain lengths, d log(cac)/dn is essentially the same as those of classical surfactants,19,45
48 showing clear parallels in behavior. The reason for the switch over can be appreciated when comparing compounds dominated by
CH2
groups with those bearing mainly aromatic
CH
, which also accounts for the varying hydrophobic contribution of the benzyl ring in the homologous series: not all carbons are thermodynamically equal. Surface tension measurements show typical surfactant behavior with increasing concentration even for the lower chain analogues. The cac values were in good agreement with those derived from the break points of k
C curves for the higher homologues (>C4) and reasonable agreement for the shorter chain systems. The limiting adsorption areas are consistent with monolayer formation when compared to classical surfactants.37
40 The NMR chemical shifts of the aromatic ring protons can be separated into “sharp” transition profiles as a function of concentration for the higher homologues (C5
C8Benz) as with 1H NMR studies of classical surfactants,41 and profiles showing smooth changes in chemical shift for the short chain homologues (C0
C4Benz). It seems reasonable to suggest that short chain “hydrotropes” form ill-defined aggregates, shielding the aromatic protons from the full strength of the external water environment. Interestingly, a different picture emerges from SANS experiments in D2O. The SANS data show that short chain and long chain amphiphiles display common behavior, with the onset of aggregation (micelles) above a well-defined concentration. These conclusions are in contrast to those of previous authors6
10 who have proposed that hydrotropes are a distinct class, being related to, but not precisely the same as, classical surfactants in terms of the mode of aggregation. With reference to this n-alkylcarboxylate series, these data resolve old controversies6
12 and furthermore provide an answer to the question posed by Balasubramanian, stated in the Introduction. Based on SANS data, there is no clear transition between hydrotropes and surfactants as a function of substituent alkyl chain length, but rather the homologues exhibit similarities, with clearly defined switches in aggregation from free molecules to aggregated assemblies above a critical aggregation concentration. Future studies should focus on systems of this kind in the presence of hydrophobic additives (such as pharmaceuticals or agrochemicals) to resolve the remaining questions surrounding the mechanism of “hydrotropy”.

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional experimental results. This material is available free of charge via the Internet at http:// pubs.acs.org.

ARTICLE

’ ACKNOWLEDGMENT The authors would like to thank Dr. Craig Butts for useful discussions on NMR. M.H.H. thanks Infineum UK and the EPSRC through the University of Bristol School of Chemistry Doctoral Training Account for the provision of a Ph.D. studentship. We also acknowledge STFC for the allocation of beam time, travel, and consumables grants to work at ISIS. ’ REFERENCES (1) Hopkins Hatzopoulos, M. T.; Dowding, P. J.; Eastoe, J. Soft Matter 2011, 7, 5917–5925. (2) Cosgrove, T. Colloid Science, Principles, Methods and Applications, 2nd ed.; Wiley: Chichester, 2010. (3) Hunter, R. J. Foundations of Colloid Science, 2nd ed.; Oxford University Press: New York, 2001. (4) Rosen, M. J. Surfactants and Interfacial Phenomena, 2nd ed.; Wiley: New York, 1989. (5) Shinoda, K.; Nakagawa, T.; Tamamushi, B.; Isemura, T. Colloidal Surfactants; Academic Press: New York and London, 1963. (6) Neuberg, C. Biochem. Z. 1916, 76, 107–176. (7) Danielson, I.; Stenius, P. J. Colloid Interface Sci. 1971, 37, 264. (8) Dempsey, G.; Molyneux, P. J. Chem. Soc., Faraday Trans. 1992, 88, 971–977. (9) Ferreira, G.; Perigo, D. M.; Politi, M. J.; Schreier, S. Photochem. Photobiol. 1996, 63, 755–761. (10) Balasubramanian, D.; Srinivas, V.; Gaikar, V. G.; Sharma, M. M. J. Phys. Chem. 1989, 93, 3865–3870. (11) da Silva, R. C.; Spitzer, M.; da Silva, L. H. M.; Loh, W. Thermochim. Acta 1999, 328, 161–167. (12) Balasubramanian, D.; Srinivas, V.; Rodley, G. A.; Ravikumar, K.; Robinson, W. T.; Turnbull, M. M. Langmuir 1997, 13, 3235–3239. (13) Triolo, R.; Ho, P. C.; Johnson, J. S. J. Phys. Chem. 1995, 99, 9581–9588. (14) Balasubramanian, D.; Srinivas, V. Langmuir 1998, 14, 6658–6661. (15) Hayter, J. B.; Penfold, J. Mol. Phys. 1981, 42, 109–118. (16) Tanford, C. The Hydrophobic Effect, 2nd ed.; Wiley: New York, 1980. (17) Evans, H. C. J. Chem. Soc. 1956, 579–586. (18) Suarez, M. J.; Lopez-Fontan, J. L.; Sarmiento, F.; Mosquera, V. Langmuir 1999, 15, 5265–5270. (19) Klevens, H. B. J. Phys. Colloid Chem. 1948, 52, 130–148. (20) Campbell, A. N. Can. J. Chem. 1965, 43, 1729–1737. (21) Greshman, J. W. J. Phys. Chem. 1957, 61, 581–584. (22) Shinoda, K.; Kazuyoshi, K. J. Phys. Chem. 1960, 64, 54–57. (23) Dick, S. G.; Fuerstenau, D. W.; Healy, T. W. J. Colloid Interface Sci. 1971, 37, 595–602. (24) Paquette, R. G.; Lingafelter, E. C.; Tartar, H. V. J. Am. Chem. Soc. 1943, 65, 686–692. (25) Lin, I. J.; Friend, J. P.; Zimmels, Y. J. Colloid Interface Sci. 1973, 45, 378–385. (26) Lindman, B.; Wennerstrom, H. Top. Curr. Chem. 1980, 87, 1–87. (27) Watry, M. R.; Richmond, G. L. J. Am. Chem. Soc. 2000, 122, 875–883. (28) Rosen, M. J.; Nakamura, Y. J. Phys. Chem. 1977, 91, 873–879. (29) Rosen, M. J. J. Am. Oil Chem. Soc. 1974, 51, 461–465. (30) Scamehorn, J. F.; Schechter, R. S.; Wade, W. H. J. Colloid Interface Sci. 1981, 85, 463–478. (31) Pitt, A. R.; Morsley, S. D.; Burbridge, N. J.; Quickenden, E. L. Colloids Surf., A. 1996, 114, 321–335. (32) Atkins, P.; de Paula, J. Atkins Physical Chemistry, 7th ed.; Oxford University Press: New York, 2002. (33) Eastoe, J.; Downer, A.; Pitt, A. R.; Simister, A.; Penfold, J. Langmuir 1999, 15, 7591–7599. (34) Eastoe, J.; Mohamed, A.; Trickett, K.; Chin, S. W.; Cummings, S.; Sagisaka, M.; Hudson, L.; Nave, S.; Dyer, R.; Rogers, S. E.; Heenan, R. K. Langmuir 2010, 26, 13861–13866. 12352

dx.doi.org/10.1021/la2025846 |Langmuir 2011, 27, 12346–12353

Langmuir

ARTICLE

(35) Nave, S.; Eastoe, J. Langmuir 2000, 16, 8733–8740. (36) Li, Z. X.; Lu, J. R.; Thomas, R. K. Langmuir 1997, 13, 3681– 3685. (37) Kling, W.; Lange, H. Proc. Int. Congr. Surf. Act., 2nd 1957, 1, 295. (38) Dahanayake, M.; Cohen, L. A. W.; Rosen, M. J. J. Phys. Chem. 1986, 90, 2413–2418. (39) Gray, F. W.; Gerecht, J. F.; Krems, I. J. J. Org. Chem. 1955, 20, 511–524. (40) Lange, H. Proc. Int. Congr. Surf. Act., 4th 1964, II, 497. (41) Das, S.; Bhirud, N.; Nayyar, K. S.; Narayan, K. S.; Kumar, V. V. J. Phys. Chem. 1992, 96, 7454–7457. (42) Gordon, J. E.; Robertson, J. C.; Thorne, R. L. J. Phys. Chem. 1970, 74, 957–961. (43) Bezzobotnov, V. Y.; Borbely, S.; Cser, L.; Farago, B.; Gladkih, I. A.; Ostanevich, Y. M.; Vass, Sz. J. Phys. Chem. 1988, 92, 5738–5743. (44) Pal, O. P.; Gaikar, V. G.; Joshi, J. V.; Goyal, P. S.; Aswal, V. K. Pramana 2004, 63, 357–362. (45) Shinoda, K. Bull. Chem. Soc. Jpn. 1953, 26, 101–105. (46) Debey, P. J. Phys. Colloid Chem. 1949, 53, 18. (47) Barry, B. W.; Morrison, J. C.; Russel, G. F. J. J. Colloid Interface Sci. 1970, 33, 554–561. (48) Lin, I.; Somasundaran, P. J. Colloid Interface Sci. 1971, 37, 731–743.

12353

dx.doi.org/10.1021/la2025846 |Langmuir 2011, 27, 12346–12353