Article pubs.acs.org/Langmuir
The Cloud Point of Alkyl Ethoxylates and Its Prediction with the Hydrophilic−Lipophilic Difference (HLD) Framework Silvia Zarate-Muñoz, Americo Boza Troncoso, and Edgar Acosta* Department of Chemical Engineering and Applied Chemistry, University of Toronto, Wallberg Building, 200 College Street, Toronto Ontario M5S 3E5, Canada S Supporting Information *
ABSTRACT: The hydrophobicity of surfactants has been described through different concepts used to guide the formulation of surfactant− water (SW) and surfactant−oil−water (SOW) systems. An integrated framework of hydrophobicity indicators could provide a complete tool for surfactant characterization, and insights on how their relationship may influence the overall phase behavior of the system. The hydrophilic− lipophilic difference (HLD) and the characteristic curvature (Cc) parameter, included in the HLD, have been shown to correlate with different hydrophobicity indicators including the hydrophilic−lipophilic balance (HLB), packing factor (Pf), phase inversion temperature (PIT), spontaneous curvature (Ho), surfactant partition (Ko‑w), and the critical micelle concentration (CMC). This work aims to investigate whether the HLD can further describe a concomitant hydrophobicity parameter, the cloud point (CP) of alkyl ethoxylates. After applying group contribution models to calculate the Cc of monodisperse (pure) nonionic alkyl ethoxylates, a linear correlation between the calculated Cc and the CP was observed for pure surfactants with 8 ethylene oxide (EO) units or less. Furthermore, using an apparent equivalent alkane carbon number (EACN) to represent the hydrophobicity of the micelle core, the HLD equation was capable of predicting cloud point temperatures of pure alkyl ethoxylates, typically within 5 °C. Polydisperse surfactants did not follow the linear CP-Cc correlation found for pure surfactants. After treating polydisperse samples using a liquid−liquid extraction procedure used to remove the most hydrophobic components in the mixture, the resulting treated surfactants fell in the correlation line of pure alkyl ethoxylates. A closer look at the partition behavior of these treated surfactants showed that their partition, Cc and cloud point are dominated by the most abundant ethoxymers in the treated surfactant. The HLD also predicted the cloud point depression of treated surfactants with increasing sodium chloride concentration. This work shows how the HLD framework could be extended to predict the behavior of SW systems.
1. INTRODUCTION The cloud point is a characteristic feature of nonionic surfactants that denotes the onset temperature at which the surfactant separates from an aqueous solution due to the weakening of hydrogen bonds between the surfactant and water molecules.1,2 At this point, the stronger attractions between surfactant tails increase the micelle aggregation number, producing surfactant assemblies that form a coacervate phase within a diluted surfactant aqueous solution, appearing as a milky dispersion.3 The cloud point temperature is a critical parameter in wetting, cleaning and foaming applications,4 as well as in surfactant-based extraction processes of organic compounds and proteins.5−7 Additionally, nonionic surfactants are found in numerous industrial applications including cosmetics, pharmaceuticals, food, and oilfield formulations, where a higher cloud point temperature is desirable to prevent phase separation. The hydrophobicity of alkyl ethoxylate nonionic surfactants is observed in their cloud point, i.e., surfactants with longer alkyl tails, and a small number of ethoxylate (EO) groups are more hydrophobic and have lower cloud points. The hydrophobicity of surfactants also influences their behavior in © 2015 American Chemical Society
surfactant−oil−water (SOW) systems. For the case of alkyl ethoxylates in the presence of oil, the phase inversion temperature (PIT) would be the equivalent of the cloud point (CP), as qualitatively correlated by Shinoda and collaborators.8 However, instead of producing a coacervate phase, increasing the temperature beyond the PIT produces water-in-oil (w/o) microemulsion systems.9,10 Different approaches have been used to predict the CP of pure alkyl ethoxylate nonionic surfactants, starting from the correlation of Gu and Sjöblom,11 and the modified version from Huibers et al.,12 which relate the contributions of the surfactant molecular structure on the CP: CP = 87.1 × ln NES − 5.78 × NCS − 40.7
(1)
where NES is the number of ethylene oxide (EO) units, and NCS is the number of carbons in the alkyl chain. Ren et al.13,14 also introduced nonlinear quantitative structure property relationships (QSPR) to predict cloud point temperatures of alkyl Received: August 16, 2015 Revised: October 8, 2015 Published: October 14, 2015 12000
DOI: 10.1021/acs.langmuir.5b03064 Langmuir 2015, 31, 12000−12008
Article
Langmuir ethoxylates by including additional molecular structure descriptors. However, the complexity associated with this approach does not allow an easy adaptation of the method. In an effort to predict the cloud point of polydisperse surfactants, Kim et al.15 developed a linear correlation with four fitting parameters that considers the nominal structure of the polydisperse surfactant mixture and its polydispersity index. Additionally, Bendjaballah et al.16 carried out CP predictions of some polydisperse surfactants adapting the Flory−Huggins theory and pseudophase models developed for microemulsions. Nevertheless, lack of data limits the applicability of this model for a variety of nonionic surfactants. Moreover, none of the approaches described above are yet able to predict the effect of salt on the CP. The hydrophilic−lipophilic difference (HLD) has been used to predict the phase inversion temperature (PIT) of nonionic surfactants present in a SOW system.17−19 The HLD is an empirical equation expressing the relationship between formulation variables that produce an “optimal formulation” where the largest cosolubilization of oil and water, and the lowest interfacial tension is attained. Moreover, it has been found to correlate with the change in the surfactant chemical potential when transferred from the oil phase into the aqueous phase.17 For nonionic surfactants, the HLD equation is μws
− RT
μos
Hn′ =
Hn HLD =− 2 2L
(3)
Where L is proportional to the surfactant tail length. Another concept used to describe the hydrophobicity of nonionic surfactants is the distribution or partition coefficient (Kw‑o) between two immiscible oil and aqueous phases at equilibrium (below and above the CMC), and between oil and water excess phases of middle phase microemulsions.17,18,26 A relationship between the partition coefficient of the surfactant in the excess phases of middle phase microemulsions and the HLD was established by Salager et al.17 However, according to Fraaije et al.,27 the correlation between Kw‑o and the HLD is surfactant-specific, due to the different configuration that surfactant molecules acquire at the interface. Recently, Boza et al.28 have also established a connection, via an activity coefficient model, between the partition coefficient of alkyl ethoxylates and the critical micelle concentration (CMC), which is another hydrophobicity indicator used by the surfactant community. According to the predictions from the Universal Functional Activity Coefficient (UNIFAC) model used by Boza et al.,28 the partition coefficient of pure alkyl ethoxylate surfactants correlates with their CMC as log CMC = 0.085 × log K w − o + b
(4)
where b is function of the surfactant tail length and the EACN of the oil. Thus, a relationship between the HLD and the CMC could also be established, given the connection between HLD and Kw‑o.17 An integrated framework of hydrophobicity indicators, centered around the HLD, can provide not only a complete tool for surfactant characterization, but also insights on how the relationship between these indicators may influence the overall phase behavior of the system, a subject that continues to be the focus of extensive research.14,29−35 This work aims to investigate whether the HLD can also describe a concomitant surfactant hydrophobicity concept, the cloud point (CP) of nonionic surfactants. To this end, the first objective of the work is to observe whether there is a relationship between the Cc and CP of monodisperse (pure) alkyl ethoxylate nonionic surfactants, and to determine whether the HLD can reproduce experimental cloud point temperatures. To achieve this goal, the Cc values of pure alkyl ethoxylate nonionic surfactants were calculated using a group contribution model introduced by Acosta:19
∝ HLD = b·S − K (EACN) + Cc + c T(T − Tref ) − Ø(A)
(2)
where S is the electrolyte concentration (g/100 mL), EACN is the equivalent alkane carbon number of the oil (i.e, EACNn‑hexane = 6, and EACNtoluene = 1), T is the temperature of the system, Tref = 25 °C, Ø(A) is a function of the concentration of alcohol or cosurfactants (zero if none is added), and b, K, and cT are proportionality constants that depend on the type of surfactant. The characteristic curvature (Cc) is a normalized curvature defined as the product of the surfactant tail length and the surfactant curvature when exposed to a set of characteristic (reference) conditions, viz., T = 25 °C, EACN = 0, in the absence of salt and cosurfactant.20 Hence, it quantifies the hydrophobicity of a surfactant present in a SOW system, and indicates the type of surfactant assembly formed at these reference conditions. For instance, a hydrophilic surfactant that tends to produce oil-swollen micelles in the presence of an oil with EACN = 0 (e.g., benzene), in the absence of salt or cosurfactant in the system has a negative Cc; whereas a lipophilic surfactant that tends to form water-swollen reverse micelles at the same conditions has a positive Cc. Through the Cc parameter, the HLD has also been shown to correlate with other common hydrophobicity indicators such as the hydrophilic−lipophilic balance (HLB) and, at certain range, with an apparent packing factor (Pf), concepts that provide estimations of the hydrophobicity of nonionic surfactants based on their molecular structure and their geometrical configuration at the oil/water interface.20−22 In the same line, Kunz et al.23 derived correlations between the HLD, a local spontaneous packing factor (Po) and the average spontaneous curvature (Ho), linked by the average and Gaussian curvature framework introduced by Helfrich. Noteworthy is that the net curvature (Hn) introduced by Acosta et al.,24 and the modified net curvature (H′n) by Kiran et al.,25 scale to the spontaneous curvature (Ho) of the Helfrich framework. These curvatures scale to the HLD as
Cc = 0.28 × NCS + 2.4 − NES
(5)
where NCS is the number of carbons in the surfactant alkyl chain, and NES is the number of ethylene oxide (EO) groups in the hydrophilic head. Subsequently, the Cc values of these pure alkyl ethoxylates were compared with their corresponding cloud point temperatures obtained from literature data. In order to determine if the HLD could predict cloud points, an apparent equivalent alkane carbon number (EACNapp), used to describe the lipophilic nature of the alkyl ethoxylate micelle cores was obtained. The calculated cloud point temperatures were compared with experimental values for these pure surfactants, as well as with predictions obtained from the correlation of Gu and Sjöblom modified by Huibers (eq1). A linear relationship was confirmed for pure surfactants with 8 EO units or less. The second objective of this work is to investigate if there is a relation between the Cc and the CP of polydisperse alkyl ethoxylate and glycerol ethoxylate nonionic surfactants, and to 12001
DOI: 10.1021/acs.langmuir.5b03064 Langmuir 2015, 31, 12000−12008
Article
Langmuir
Treated Polydisperse Surfactants. 2 mL of the aqueous phase from the liquid−liquid extraction procedure containing originally 2 wt % of polydisperse surfactant were placed in 2 dram flat bottom vials. Different amounts of sodium chloride were added to each vial to maintain the electrolyte concentration from 0 to 20 gNaCl/100 mL in 5 gNaCl/100 mL increments. All vials were gently mixed and left to equilibrate for 2 h. After this time, the cloud point was assessed following the procedure described above for untreated surfactants. 2.3. Characteristic Curvature Measurements. Untreated Polydisperse Surfactants. The characteristic curvature of the untreated polydisperse nonionic surfactants was measured adapting a procedure described in ref 20 for nonionic surfactants. Briefly, salinity scans were carried out at room temperature by mixing 2 mL of oil phase with 2 mL of aqueous phase in 2 dram flat bottom vials to produce microemulsion systems. The oil phase consisted of cyclohexane (EACN = 3.3), decahydronaphthalene mixture of cis+trans (EACN = 6.3), or hexanes (EACN = 6). The aqueous phase contained a polydisperse nonionic surfactant, or a combination of a polydisperse nonionic surfactant with a reference surfactant (Novel TDA-6 or Dehydol O5) at a total surfactant concentration of 10 wt %. Additionally, the salinity of the aqueous phase was varied from 0 to 27 g/100 mL by adding specific amounts of 30 wt/v% NaCl solution and deionized water. All systems were simultaneously mixed and left to equilibrate for 1 day. The optimum salinities (S*) were determined as the middle phase bicontinuous microemulsions that displayed the fastest coalescence rate after re-emulsification. The supporting material contains detailed examples of this protocol, as well as the measurement of the EACN of the decahydronaphthalene mixture of cis+trans. Treated Polydisperse Surfactants. The characteristic curvature of treated C9E4.5 and C13E6 polydisperse nonionic surfactants was determined following the procedure described for untreated surfactants with some modifications. In short, 2 mL of aqueous phase from the liquid−liquid extraction procedure containing originally 10 wt % surfactant were mixed with 2 mL of decahydronaphthalene mixture of cis +trans as oil phase in 2 dram flat bottom vials. The salinity of the aqueous phase was varied from 0 to 20 wt % by adding different amounts of sodium chloride. The C9E4.5 and C13E6 polydisperse surfactants can form microemulsions without the need of a reference surfactant. For these surfactants, the optimal salinity (S*, where middle phase microemulsions contain the same volume of oil and water solubilized, and HLD = 0) was used in eq 2 to obtain their Cc. Measurement of the Cc of treated C13E8 and C9GE6 was not conducted because these surfactants are too hydrophilic to undergo phase inversion (HLD = 0) on their own, and require the use of a reference surfactant to measure their Cc. Conducting mixtures of reference and test surfactant require knowing the total surfactant concentration of each surfactant, something that could not be fully accomplished for C13E8 and C9GE6 given the poor signal of ESI-MS for CiEO0−2 ethoxymers. 2.4. Partitioning Studies. 20 mL of surfactant solution (2 and 10 w/v %) of each polydisperse surfactant in deionized water were prepared in 50 mL flat bottom vials. An equivalent volume of decahydronaphthalene mixture of cis+trans was subsequently added to each surfactant solution, followed by mixing and equilibration at room temperature for 1 week. The
investigate whether the HLD can predict CP temperatures and the effect of inorganic salts, such as sodium chloride, on their CP. To achieve this goal, the characteristic curvatures of a set of polydisperse alkyl ethoxylate nonionic surfactants were measured along with their cloud points. These values were compared with the values found for pure surfactants. The results showed that polydisperse surfactants did not follow the linear relationship between the Cc and the CP found for pure surfactants. A liquid−liquid extraction method was used to remove the more hydrophobic ethoxymers in the polydisperse samples, yielding treated samples that followed the trend displayed by the pure surfactants. Moreover, the effect of salt on the cloud point of the treated polydisperse samples was assessed. Surfactant partitioning studies were also carried out to understand the relationship among the partition coefficient, the characteristic curvature, and the cloud point behavior displayed by polydisperse alkyl ethoxylates.
2. MATERIALS AND METHODS 2.1. Materials. Commercial-scale samples of polydisperse alkyl ethoxylate nonionic surfactants with nominal structures C9E4.5 (100% active, EO units ranging from 2 to12), C13E6 (100% active, EO units ranging from 3 to 12), C13E8 (100% active, EO units ranging from 5 to 15), and a glycerol ethoxylate nonionic surfactant C9GE6 (100% active, EO units ranging from 3 to 12), were provided by leading manufacturers of alkyl ethoxylated surfactants (brand names withheld to prevent commercial inferences from the work), and used to determine the cloud point and characteristic curvature (Cc) of these polydisperse samples. It must be clarified that the range of EO units indicated above was based on electrospray ionization mass spectrometry (ESI-MS) distributions from unmodified samples, a technique that cannot capture the presence of free alcohol in the samples. Cyclohexane (≥99.7%), and decahydronaphthalene mixture of cis+trans (decalin, 98%), were purchased from Sigma-Aldrich (St. Louis, MO). Hexanes (nhexane ∼62−95%, isomers of methylpentane
