Self-assembly of ionic and non-ionic surfactants in type IV cerium nitrate and urea based deep eutectic solvent

Ionic liquids and deep eutectic solvents (DESs) have gained interest as alternatives to traditional solvents in many applications from material synthesis^1–6 and electrochemistry^7–13 to pharmaceuticals^14–16 due to their low vapor pressure, low flammability, wide liquid range, and tunable nature. In recent times, there has been an upsurge in the field of studying self-assembly of ionic^17–23 and non-ionic^23–27 surfactants in ionic liquids and DESs, ever since the first study reporting self-assembly of surfactants in ethylammonium nitrate in 1982.²⁸ Studies of self-assembly in these novel solvents give an insight into both fundamental interactions in these alternative media and offer unique potential for uses in the fields of emulsification and templating applications, e.g., synthesis of polymers and polymer membranes^29–32 or porous materials^33–35 and use in transdermal delivery systems.³⁶ 

In addition to their shared characteristics, DESs offer some unique advantages in comparison to ionic liquids. They often comprise cost-effective, widely available, and environmentally benign components making them biocompatible and have therefore attracted great interest as ionic liquid analogs. DESs are generally made by complexation of a hydrogen bond donor with a salt capable of sustaining a hydrogen bond network, which results in a significant dip in their melting point, which is lowest at a particular molar ratio called the eutectic point.³⁷ Their hydrophobicity, physicochemical properties, and nanostructure can be altered by changing the constituents or their ratios, making them highly tunable for specific applications. Studies have shown that both DES nanostructure and composition play a crucial role in self-assembly of various amphiphiles.^38,39

Surfactant micelle templated synthesis of mesoporous materials, such as silica^40,41 and metal oxides,⁴² has widely been exploited in traditional solvents. Recently, Hu et al. reported surfactant template-assisted synthesis of hierarchical ZIF-8 particles using sodium dodecyl sulfate (SDS) in choline chloride/urea (ChCl:U) DESs.³⁵ They reported that the porosity in the ZIF-8 particles can be tuned by varying the amount of SDS and water in the reaction mixture. Hammond et al. also reported solvothermal synthesis of iron oxide⁵ and cerium oxide⁶ nanoparticles from ChCl:U DESs where the morphology and porosity are controlled by the water content of the mixture. Furthermore, it is understood that when synthesizing these oxides, the urea component of the DES is crucial to the solvothermal reaction mechanism. Therefore, self-assembly of surfactants in urea containing DESs offers an interesting potential for templating metal oxides in solvothermal synthesis.

Until now, the most commonly studied DESs are ones that comprise a quaternary ammonium halide salt with a hydrogen bond donor, e.g., choline chloride (ChCl) with urea (U), glycerol (Gly), ethylene glycol (EG), or some carboxylic acid. Most of these fall into the category of type III DESs.⁴³ However, there are other types of DESs comprising metal salts or their hydrates with quaternary ammonium salts, type I (ZnCl₂:ChCl) and type II (ZnCl₂·3H₂O:ChCl), respectively, or metal salts with a hydrogen bond donor, type IV (ZnCl₂:urea, AlCl₃:urea, and CrCl₃·6H₂O:urea). Recently, a halide-free type IV DES was reported by Hammond et al. comprising hydrated salts of lanthanide metals (cerium, praseodymium, and neodymium) with urea.⁴⁴ These are low melting point and low viscosity DESs with very high density and surface tension. Their Gordon parameter is similar to that of water, suggesting potential for surfactant aggregation/micellization due to solvophobic interactions.⁴⁵ Lanthanide oxides can be prepared from these solvents by controlled combustion, showing potential use of these solvents in efficient synthesis applications. Surfactant self-assembly in DESs besides type III DESs is a relatively unexplored field and offers the unique opportunity to understand intermolecular interactions in a solvent comprising metal salts along with organic molecular components as well as the potential for templated synthesis applications.

In this work, we have explored the self-assembly of a range of surfactants: cationic (C₁₂TAB and C₁₂TANO₃), anionic (SDS), and non-ionic (C₁₂EO₆ and C₁₆EO₈); in a type IV DES comprising cerium (III) nitrate hexahydrate and urea in the molar ratio 1:3.5, dodecyl chain (C₁₂) surfactants were the main focus in this study as they have a low Krafft temperature facilitating measurements at room temperature and a wide range of data are available for the self-assembly of these surfactants in various solvents for comparison. The bromide and nitrate variants of the C₁₂ alkyl chain cationic surfactant allow us to discern any differences arising in the DES interaction with the solvent due to complexation of different anions with the cerium salt. In addition to the dodecyl chain lengths, we have also studied C₁₆EO₈ to understand the effect of change in both alkyl chain length and polyethoxy head group length on the surfactant self-assembled structures. Surfactant self-assembly is studied using interfacial tension to establish the surface activity of these surfactants in the DES along with small angle neutron scattering (SANS) measurements to demonstrate the existence of micelles and determine their structures.

Cerium nitrate hexahydrate [Ce(NO₃)₃·6H₂O, 99.5%] was purchased from Acros Organics, and urea (h-U, ≥99.5%) was purchased from Sigma-Aldrich. Deuterated urea-d4 (d-U, 99% atom, 98% D) was purchased from Cambridge Isotope Laboratories.

Dodecyl trimethylammonium bromide (C₁₂TAB, ≥99%), sodium dodecyl sulfate (SDS, ≥99%), hexaethylene glycol monododecyl ether (C₁₂EO₆, ≥98%), and octaethylene glycol monohexadecyl ether (C₁₆EO₈, ≥98%) were purchased from Sigma-Aldrich. Dodecyl trimethylammonium nitrate (C₁₂TANO₃) was prepared by using an ion exchange procedure⁴⁶ using Amberlite® IRN78 hydroxide (Acros Organics) from C₁₂TAB. The ion exchange resin was stirred with 1M NaOH for 2 h and then rinsed with water until a pH of 7 was obtained. Thereafter, 10 g of C₁₂TAB was added to a flask containing 100 g of the ion exchange resin and 150 ml of deionized water. The solution was stirred for a few hours until the entire C₁₂TAB dissolved. The solution was filtered into a new batch of 100 g of ion exchange resin and 150 ml of water and stirred for a further 2 h and repeated once more. The strongly basic solution is separated from the resin by filtration and rinsing with copious amount of water. This results in C₁₂TAOH, which is then titrated with dilute nitric acid until a pH ≈6 was obtained to get C₁₂TANO₃. The solution was tested for any remaining C₁₂TAB using silver nitrate as any bromide will result in silver bromide, which will precipitate out of the solution. Once checked, the solutions were freeze dried to obtain dry C₁₂TANO₃ powder.

Isotopically labeled C₁₂TAB (d34, d-C₁₂TAB) was supplied by the ISIS Deuteration Facility. Deuterated C₁₂TANO₃ (d34, d-C₁₂TANO₃) was prepared from d-C₁₂TAB using the ion exchange method described above and using D₂O instead of deionized H₂O.

The Ce(NO₃)₃·6H₂O:urea (Ce:U) DES was prepared by combining the components in molar ratios of 1:3.5. The mixtures were stirred at room temperature until a clear, homogeneous liquid was obtained, which was subsequently sealed and equilibrated overnight. Once formed, the mixtures were stable and were used within 2–3 weeks from preparation. For the DES containing deuterated urea, Ce:d-U, Ce(NO₃)₃·6H₂O, and d4-urea were mixed in the molar ratio of 1:3.5.

The surfactant in DES solutions containing C₁₂TAB, C₁₂TANO₃, SDS, C₁₂EO₆, and C₁₆EO₈ was mixed and equilibrated at room temperature until homogeneous mixtures were obtained. Concentration series of various surfactants were obtained by diluting high concentration (10 wt. %) solutions.

The density of the DES was determined from a triplicate average of measurements on an Anton Paar DMA 4500M at 25 °C. Differential scanning calorimetry (DSC) measurements to determine the melting/transition temperature of the neat DES were carried out on a TA Instruments DSC-Q20 differential scanning calorimeter. The sample was first equilibrated at 25 °C and held for 5 min and then cooled to −75 °C at a ramp rate of 5 °C min⁻¹ and held for 10 min before heating to 30 °C at the same ramp rate. The viscosity of the DES was measured using a TA Instruments HR-3 Discovery Hybrid Rheometer operating in a flat plate geometry with Peltier temperature control. Room temperature viscosity data for the DES were obtained for the applied shear rate ranging from 0.1 to 100 s⁻¹ at 25 °C and a Newtonian behavior was observed. Therefore, the temperature sweep data were measured at a shear rate of 1 s⁻¹ for the temperature ranging from −10 to 40 °C.

Surface tension measurements were made using the drop-shape-analysis method⁴⁷ with a FTA1000 drop shape analyzer. A concentration series of the surfactants in DESs ranging from 0.002 to 1 wt. % (0.1–60 mM, depending on the surfactant) were prepared and measured at room temperature. Each concentration was measured five times using a fresh drop, with each measurement comprising 100 data points taken within 1 min. The data reported are the average value of the five measurements with the standard error.

Small angle neutron scattering (SANS) measurements were carried out on the SANS2D⁴⁸ instrument at the ISIS Pulsed Neutron and Muon Source, UK (RB1920676⁴⁹). Using a sample-to-detector distance of 4 m and neutrons with wavelengths of 2–16.5 Å, a q-range of 0.0042–0.82 Å⁻¹ was obtained. The samples were loaded into 1 mm path length rectangular quartz cuvettes (Hellma GmbH) and placed on the automatic sample changer on the beamline. The measurements were performed at 25 °C. Data reduction was performed according to the standard procedures using the routines within Mantid,⁵⁰ resulting in output converted to absolute units of the scattering intensity [I(q), cm⁻¹] vs the momentum transfer (q, Å⁻¹). Subtraction of the scattering from the pure solvents was performed afterward using the NIST NCNR macros in Igor Pro⁵¹ to account for the incoherent contribution to each sample.

Surfactant mixtures in DESs were prepared at concentrations higher than the CMC, as determined by the surface tension, at 1 wt. % in different isotopic mixtures of the DES to aid in resolving the micelle structure. Protonated surfactants (C₁₂TANO₃, C₁₂TAB₃, SDS, C₁₂EO₆, and C₁₆EO₈) were measured in Ce:d-U and Ce:U DES to provide information on the size of the micelle core as the scattering is dominated by the micelle core–solvent scattering length density contrast plus some information about the shell structure. Furthermore, to get the details of the micelle structure and solvent surfactant interactions, d-C₁₂TANO₃ and d-C₁₂TAB were also measured in Ce:U and Ce:d-U DES. In addition to the 1 wt. % concentrations, for C₁₂TANO₃ and C₁₆EO₈, a series of measurements were performed with concentration ranging from 0.5 to 10 wt. % to study the concentration dependence of the surfactant micelle structure. The neutron scattering length densities (SLDs) for the constituent components of the surfactants and DES and the SLDs of the DES at the two contrasts are given in the supplementary material (Tables S1 and S2).

The SANS data were analyzed by co-fitting the various isotopic contrasts using standard spherical and cylindrical form factors and their core–shell variants in SasView.⁵² In the case of high concentration data, where inter-particle interactions were observed, the Percus–Yevick hard sphere structure factor⁵³ was used to account for the structure factor contribution to the scattering as employed in previous work on similar systems.^21,38 The details for the fitting models and procedure can be found in the supplementary material (Sec. S3).

Cerium nitrate hexahydrate and urea form a homogeneous room-temperature liquid at a molar ratio of 1:3.5 as previously reported by Hammond et al.⁴⁴ The glass transition temperature as determined using extrapolation of the viscosity/temperature relationship with the Vogel–Fulcher–Tammann (VFT) equation⁵⁴ and differential scanning calorimetry (DSC), shown in the supplementary material, was found to be ∼−60 °C, consistent with the studies by Hammond et al. Table I summarizes the main physical properties of the Ce:U DES along with the corresponding values for Ce:U DESs reported by Hammond et al., a typical type III DES ChCl:U, water and formamide, and a polar organic solvent, for comparison. Molar volumes for the DES were calculated using their measured densities and average molar masses, and the Gordon parameter was calculated using the surface tension and molar volume.⁴⁵ This parameter is a description of the “solvophobicity” of a solvent and provides a measure of the cohesiveness of solvent molecules. It may therefore be used to predict the capability of a solvent to promote the self-assembly of amphiphilic molecules.^19,55,56

The Ce:U DES shows physical properties, which are significantly different from typical type III DESs, i.e., ChCl:U. While the differences in density, average molar mass, and volume can be attributed to the presence of relatively heavy cerium ions, the more important and interesting differences are the ones observed in viscosity and surface tension. Based on structural investigations, Hammond et al.⁴⁴ suggested that the low viscosity and high surface tension for the Ce:U DES result from the presence of strongly bound transient oligomers comprising [–Ce–NO₃–], which form a molten salt-like network, in flux with an interpenetrating H-bonded network of urea, water, and some nitrate ions. The Ce:U DESs have significantly high surface tensions (∼85 mN m⁻¹) when compared to typical type III DESs (40–70 mN m⁻¹),⁶² ionic liquids (50–60 mN m⁻¹),⁶³ and even higher than water (72 mN m⁻¹). The high surface tension and the small molar volume result in a Gordon parameter, which is higher than in most DES and ionic liquids^19,20 but comparable to that of water, suggesting that their cohesive energy density is comparable to or even slightly higher than water. In addition to the comparatively high Gordon parameter, the polarizability of the solvent is also expected to be high due to the ionic nature of the constituents, which should also increase the tendency of amphiphiles to form micelles.⁶⁴ A combined effect would suggest that the Ce:U DES could promote self-assembly in a range of surfactants.

The surface tension vs molar concentration curves for C₁₂TAB, C₁₂TANO₃, SDS, C₁₂EO₆, and C₁₆EO₈ in Ce:U DESs are shown in Fig. 1. Since our conversion to the molar concentration is dependent on the measured density of the DES assuming that the volume of the DES is not influenced by the presence of surfactants, we have also plotted the data against the percentage weight concentration and mole fraction, and they are shown in the supplementary material (Fig. S6). The surface tension shows a clear variation with concentration, indicating that all the surfactants measured here remain surface-active in the DES solutions. For each surfactant, the shapes of the surface isotherm exhibit the classic behavior of a surfactant in aqueous solution⁶⁵ and other DESs.^20–22 Initially, the surface tension of the solution gradually decreases with the addition of surfactant until it reaches the critical micellar concentration (CMC) beyond which the surface tension essentially remained the same and a plateau is observed. We estimated the CMC from the intersection of the two straight lines: one obtained by a linear approximation to the first three data points in the region of the graph just below the break point concentration and the second being the surface tension in the plateau region above the break point concentration. The black lines (solid and dashed) in Fig. 1 represent the trends that help determine the CMC. Additionally, the surface excess (Γ) was calculated from the surface tension (γ) vs ln concentration ([conc]) graph using the following relation:

where $dγd⁡ln[conc]$ is the tangent to the surface tension vs ln concentration graph just before the CMC, R is the gas constant, T is the absolute temperature, and n takes the value of 1 for non-ionic surfactants and 2 for ionic surfactants. The surface excess can then be used to calculate the area per molecule (Apm) in Ų using the following relation:

where N_A is Avogadro’s number. Table II shows the CMC values in mM for the various surfactants (CMC values in wt. % and mole fraction are available in the supplementary material, Table S7), the surface tension at CMC (γ_CMC), surface excess (Γ) and the area per molecule (Apm) in the Ce:U DES along with the literature CMC, γ_CMC, surface excess and area per molecule values for the surfactants in water, and the CMC values in an organic polar solvent, formamide, for comparison.

The CMC of cationic surfactants, C₁₂TAB and C₁₂TANO₃, in Ce:U DESs is less than that in water and even lower than that observed for ionic liquids^19,23,28 and also ChCl-based DESs.^21,22,77 This is possibly due to the salting out effect in the DES owing to its ionic nature in addition to a high Gordon parameter. This is similar to the observations in aqueous solutions where the CMC of CTAB surfactants decreases as the salt concentration of the solution increases.^78–80 In Ce:U DESs, the entire solvent comprises ions and a combination of this along with the high Gordon parameter drives micellization at a lower concentration. There is no measurable effect of counterions on the CMC of dodecyltrimethylammonium (C₁₂TA⁺) surfactants in Ce:U, an effect also observed in ionic liquids, e.g., ethanolammonium nitrate;¹⁸ however, this is contrary to measurements of the CMC for cationic surfactants in aqueous solutions^81–83 and some other ionic liquids.¹⁸ This could potentially be explained by counterion exchange with the nitrate ions in the Ce:U DES resulting in little effect of counterions having a similar size at low surfactant concentrations. This is also observed for hexadecyltrimethylammonium (C₁₆TA⁺) in ChCl:Gly where the CMCs of C₁₆TAB (15.2 mM) and C₁₆TACl (15.4 mM) are very similar; however, that of C₁₆TATS, which has a significantly larger counterion, is higher (20.8 mM).⁸⁴ In the case of the anionic surfactant SDS in Ce:U, the CMC is lower than that in water and comparable to the CMC observed in ChCl:U,²⁰ which is attributed to the high polarizability parameter for these eutectic solvents and the presence of counterions, i.e., [Choline]⁺ in ChCl:U or [–Ce–NO₃–]⁺ polycations in the Ce:U DES.

On the other hand, in the case of non-ionic surfactants in Ce:U, a significantly higher CMC is observed as compared to water. This is similar, although smaller in magnitude, to the effect observed in ionic liquids,^19,25,26 where CMCs that are 3–5 orders of magnitude higher than water are observed. In ionic liquids, this is explained due to the weaker solvophobic interaction in the solvent compared to the hydrophobic effect in water owing to the presence of alkyl chains. A similar effect on CMC is also observed for non-ionic surfactants in formamide⁷⁶ and ethylene glycol.^19,25 These solvents, though polar in nature, exhibit some organic character, which increases the solubility of the surfactant alkyl chains and reduces the solvophobic interactions with the non-ionic surfactants leading to a higher CMC, an effect also observed upon addition of urea to aqueous solutions of non-ionic surfactants.^85,86 For C₁₂EO₆ and C₁₆EO₈ in Ce:U, the presence of urea has the potential to increase the solubility of the surfactant alkyl chain, while Ce(NO₃)₃ could result in a salting in/out effect on both the hydrophilic and hydrophobic moieties in the surfactant.⁸⁷ The increased solubility of the alkyl chains and the salting in effect on the polyethyoxy head group result in an increase in CMC, whereas the salting out effect on the hydrophobic moieties decreases the CMC.⁸⁸ A combination of the two reduced the solvophobic interactions of the surfactant with Ce:U when compared to water, though not to the same extent as that in polar organic solvents or ionic liquids. Therefore, for non-ionic surfactants in Ce:U, a CMC in between that in water and polar organic/ionic liquids is observed. The Gibbs free energy of micellization (ΔG = RT ln CMC) for C₁₂EO₆ was calculated to be −16.7 kJ/mol in Ce:U DES, compared to −23.3 kJ/mol in water and −8.7 kJ/mol in formamide. In addition, we see a decrease in CMC with an increase in the alkyl chain length/polyethoxy group length. This can be expected due to an increase in the solvophobic interaction as the alkyl or polyethoxy chain length increases and is consistent with observations in water;^70,72 polar organic liquids, such as formamide and ethylene glycol;^25,76 ionic liquids;^19,26 and aqueous solutions in the presence of urea^85,86 or salts.⁸⁷ As expected, with a decrease in CMC, the Gibbs free energy of micellization for C₁₆EO₈ also decreases and was calculated to be −18.2 kJ/mol in Ce:U DES, compared to −33.3 kJ/mol in water and −11.2 kJ/mol in formamide. The magnitude of the decrease in CMC and ΔG in Ce:U are less than that in water and comparable to the effect observed in formamide and ionic liquids;^19,25,26 $(ΔGC16EO8−ΔGC12EO6)$ was calculated as −1.5 kJ/mol for Ce:U DES, −10 kJ/mol for water, and −2.5 kJ/mol for formamide. The trends in the change in CMC and ΔG are another indication that the interactions of the non-ionic surfactants with the DES constituents are more akin to polar organic solvents or ionic liquids and both the organic (urea) and ionic [Ce(NO₃)₃] constituents of the DES impact the CMC.

The surface excess for all surfactants in the Ce:U DES is positive, indicating that the concentration of the surfactants is higher at the air/DES interface compared to the bulk, representative of a surface-active monomer with a hydrophobic chain that prefers to be at the interface compared to in the bulk solution. As expected, the surface excess and the area per molecule of the two cationic surfactants and the two non-ionic surfactants are in good agreement with one another. The area per molecule of the surfactants in Ce:U DES is greater than that observed in water and similar to some polar organic solvent and ionic liquids.⁷¹ This suggests that the solvophobic interactions, potentially governed by the hydrophobic chain, in the Ce:U DES are weaker than those in water and similar to organic polar solvents or ionic liquids, possibly explained by the presence of urea in the DES.

SANS was measured from 1 wt. % solutions of C₁₂TAB and C₁₂TANO₃ in Ce:U for h-C₁₂TA⁺ in hh-DES and hd-DES and d-C₁₂TA⁺ hh-DES and hd-DES contrasts to resolve the structure of the C₁₂TA⁺ micelles in the DES. The four SANS contrasts were co-fitted to a spherical form factor and its core–shell variant (see the supplementary material, Fig. S3 and Table S3) with the dodecyl chains forming the core and the trimethylammonium cationic headgroup forming the shell. It is not possible to fit all the contrasts to a spherical model with a single radius for either of the two cationic surfactants, as can be seen in Fig. S3. However, using the core–shell model, if we allow the shell SLD to vary to account for the changes in the SLDs arising due to the solvation of the micelle head group by the ions in the DES, all four contrasts can be fit to a single core-radius and shell thickness with reasonable confidence. This is also reflected in the reduced χ² parameter, which drops from 10.2 to 2.5 for C₁₂TAB and 6.05 to 1.67 for C₁₂TANO₃ when the data are fitted to the core–shell spherical form factor instead of a homogenous spherical model (see the supplementary material, Table S3). The SANS data are shown along with the best fit core–shell spherical model for the various contrasts in Fig. 2, and the parameters are summarized in Table III. In Table III, the core-radius and shell thickness obtained from the combined fit using the h-C₁₂TA⁺ in Ce:U, h-C₁₂TA⁺ in Ce:d-U, d-C₁₂TA⁺ in Ce:U, and d-C₁₂TA⁺ in Ce:d-U contrasts are reported along with the SLD of the shell at the four contrasts and the calculated solvent volume fraction of the DES in the shell region (ϕ).

The radii of the micelles are 15.0 ± 1.1 Å for C₁₂TAB and 16.2 ± 0.9 Å for C₁₂TANO₃, which are consistent with extended C₁₂ chain lengths,⁸⁹ and the shell thicknesses are 8.3 ± 0.7 Å for C₁₂TAB and 5.8 ± 0.6 Å for C₁₂TANO₃. Both the radius and shell thickness are in good agreement with values for alkyltrimethylammonium surfactants in other deep eutectic solvents,^21,38 ionic liquid ethanolammonium nitrate,¹⁸ and in water.^83,90 A more recent study by Hargreaves et al. of C₁₀TAB micelles in water using wide q-range neutron diffraction and atomistic modeling gives a hydrated shell thickness of 7.5 Å.⁹¹ Our values of solvated shell thickness, obtained using co-refinement of multiple contrasts, in SANS are in good agreement with this value and provide an average solvent volume fraction of 0.62 ± 0.07 for C₁₂TAB and 0.73 ± 0.15 for C₁₂TANO₃ in the shell region, possibly as a result of interaction of the trimethylammonium head group with the nitrate counterions in the DES. The cross section of the micelles is similar to that reported in type III DES-like ChCl:Gly²¹ and ChCl:U:Gly;³⁸ however, unlike type III DESs where elliptical micelles are reported for C₁₂TAB, spherical micelles are observed for C₁₂TA⁺ in the Ce:U DES. It must be noted here that for the elliptical micelles of C₁₂TAB in type III DESs, the axial ratio <2,^21,38 and arguably, they can be fitted to spherical form factors as well and therefore are generally treated as spheroidal. The aggregation number for the micelles was calculated to be 37 ± 8 for C₁₂TAB and 45 ± 7 for C₁₂TANO₃, giving an area per molecule of 79 ± 15 and 74 ± 12 Ų, respectively. These are in good agreement with the values calculated from the surface tension data. Both the structure and size of the micelles show no particular dependence on the counterion. This indicates that the ions in this DES behave similarly to a swamping amount of electrolytes in a solution, resulting in all counterions being exchanged with the solvent, and therefore, no effect of counterion is observed on the shape and size of the micelles, an effect also observed for these cationic micelles in ionic liquid ethanolammonium nitrate.¹⁸ 

SANS data were collected for 1 wt. % SDS in Ce:U in both hh-DES and hd-DES contrast and are shown in Fig. 3. The data show the scattering characteristic of elongated structures where I(q) ∝ q⁻¹. The two SANS contrasts were co-fitted to a cylindrical form factor and its core–shell variant (see the supplementary material, Fig. S4 and Table S4) with the dodecyl chains forming the core and the sulfate headgroup forming the shell. The SLD of the shell region was allowed to fit in order to take into account solvation of the sulfate head group by the DES. The core–shell cylindrical form factor gives a marginally better fit in this case; reduced χ² = 3.19 as compared to 3.44 for cylindrical model (see the supplementary material, Table S4). It also allows for a better comparison with the data obtained for other surfactants and is therefore presented in Fig. 3, and the fit parameters are given in Table IV. In Table IV, the core-radius, shell thickness, and length obtained for SDS micelles from the combined fit using the hh-DES and hd-DES contrasts are reported along with the SLD of the shell at the two contrasts and the calculated solvent volume fraction of the DES in the shell region (ϕ).

SDS forms extremely elongated micelles in Ce:U DESs. The radius of the micelles is 15.1 ± 0.25 Å, shell thickness is 4.5 ± 0.22 Å, and length is ≫1500 Å. The aggregation number for the SDS micelles was calculated to be 7315 ± 350, giving an area per molecule of 53 ± 3 Ų, consistent with the surface tension measurements. The solvent volume fraction in the shell region, calculated from the SLD of the shell, is 60%–75%, indicating highly solvated shells possibly due to interactions between the cations in the DES with the sulfate head group. The cross section of the micelles is similar to that reported in type III DES-like ChCl:U,²⁰ ChCl:Gly,⁹² and ChCl:U:Gly³⁸ and is consistent with extended C₁₂ chain lengths;⁸⁹ however, the elongation of the micelles is much greater in this case. For ionic liquids and DES, the surfactant–solvent counterion interaction is used to explain the morphology of micellar structures, such as the elongation of SDS micelles in aqueous solutions and other solvents. For SDS, spherical micelles are observed in water;⁹⁰ however, in salt + water, extremely long micelles are observed,⁹³ and in ChCl:U, cylindrical micelles with length ∼100 Å are observed. The elongation of the micelles is attributed to Coulombic interactions between the salt ions or DES [cholinium]⁺ with the surfactant SO⁴⁻ headgroups, which provides a charge screening effect, allowing for the headgroups to pack more closely and causing micelle elongation.⁹⁴ In our case for Ce:U DES, the [Ce–NO₃–]⁺ polycations that form the structural network in the DES interact with the SO⁴⁻ headgroup of the SDS and provide the charge screening effect. The ionic nature of the DES components offers greater charge screening than in the case of type III DES and leads to micelles that are longer in size. The salting out effects and strong interactions of the [Ce–NO₃–]⁺ polycations with the SO⁴⁻ head group also cause SDS to form a gel when dissolved in the DES at concentrations >1 wt. %. Formation of a gel phase was also reported by Matthews et al. in ChCl:Gly DES at room temperature.⁹⁵ Microscopy, SANS, and rheology studies by the group indicate self-assembled SDS fractal dendrites with dimensions of up to ∼mm in ChCl:Gly, which lead to the formation of a gel phase at SDS concentration above 1.9 wt. %. The formation of these aggregate morphologies is attributed to the condensation of choline ions around the SDS aggregates. In our case, it could be the condensation of the [Ce–NO₃–]⁺ polycations around the SDS aggregates that results in the formation of a gel.

SANS was measured from 1 wt. % solutions of C₁₂EO₆ and C₁₆EO₈ in Ce:U for hh-DES and hd-DES contrasts and is shown in Fig. 4. The data show the scattering characteristic of spherical micelles; therefore, the two SANS contrasts for each surfactant were fitted to a spherical form factor and its core–shell variant (see the supplementary material, Fig. S5 and Table S5) with the alkyl chains forming the core-radius and the polyethoxy headgroup forming the shell of the structure. The shell SLD was once again allowed to vary to account for solvation of the polyethoxy head group by the DES components. The core–shell spherical model gives a significantly improved fit in this case; reduced χ² = 5.36 for C₁₂EO₆ and χ² = 3.6 for C₁₆EO₈ as compared to 9.21 and 7.1, respectively, for the homogenous spherical model (see the supplementary material, Table S5). Therefore, this is considered a better representation of the structure of these non-ionic surfactants in Ce:U DES and is presented in Fig. 4, and the fit parameters are given in Table V. In Table V, the core-radius and shell thickness for the C₁₂EO₆ and C₁₆EO₈ micelles obtained from the combined fits using both contrasts are reported along with the SLD of the shell at the two contrasts and the calculated solvent volume fraction of the DES in the shell region (ϕ).

The radius of the micelles was found to be 18.7 ± 1.8 Å for C₁₂EO₆ and 23.5 ± 1.5 Å for C₁₆EO₈, and the shell thickness was found to be 7.2 ± 1.0 Å for C₁₂EO₆ and 7.8 ± 0.9 Å for C₁₆EO₈. The aggregation number was calculated to be 69 ± 20 for C₁₂EO₆ and 107 ± 21 for C₁₆EO₈, giving an area per molecule of 64 ± 18 and 65 ± 12, respectively, in good agreement with the values obtained from the surface tension measurements. The fitted core-radius for both C₁₂EO₆ and C₁₆EO₈ micelles while slightly on the higher side is consistent with fully extended C₁₂ and C₁₆ chain lengths,⁸⁹ respectively, within the fit error. Micelles of core-radius of up to 18.1 Å have been reported previously for C₁₂EO_n micelles in ionic liquid ethylammonium nitrate.²⁴ The larger size could imply that some of the “headgroup” is in the core due to poor solvation in these liquids, also manifested in the fact that the headgroup region (shell thickness) is smaller than the size of the polyethoxy head group for both C₁₂EO₆ and C₁₆EO₈. The solvent volume fraction in the shell region, calculated from the SLD of the shell, is 40%–55%, lower than the solvation of ionic micelles. The smaller shell thickness may be due to loss of scattering contrast beyond a certain distance resulting from the solvation of the shell region by the DES or coiling of the polyethoxy head group in the DES⁹⁶ compared to water. In this case, the coiling of the polyethoxy head group could be due to the presence of a large number of ions in the DES⁹⁷ or due to the fact that the interactions of the polyethoxy head group with the DES are less solvophilic as compared to that in water because of the presence of the organic species, an effect also observed for C₁₂EO_x micelles in ethylammonium nitrate.²⁴ The overall size of the C₁₂EO₆ micelle (∼2.6 nm) is smaller than that observed for water (∼3.2 nm)⁹⁸ but larger than that observed in ionic liquids (1.5–2.0 nm).^24–27 A similar observation can be made for C₁₆EO₈ where the micelle size (∼3.2 nm) is smaller than water (∼3.7 nm)⁹⁸ but comparable to some ionic liquids (∼3.2 nm).²⁶ The solvophobic interactions influence not only the CMC, as previously discussed, but also the size of the micelles. The urea in the DES along with the ionic components results in solvophobic interactions for non-ionic surfactants in Ce:U that are intermediate compared to the ones observed for water and polar organic solvents/ionic liquids, resulting in both sizes of micelles and CMCs, which lie in between the two.

In addition to scattering data from low surfactant concentration solutions, SANS was also measured from h-C₁₂TANO₃ and h-C₁₆EO₈ for concentrations ranging from 0.5 to 10 wt. % in hh-DES and hd-DES to investigate the effect of concentration on the surfactant micellar structure. The data are shown in Fig. 5. It is not possible to do a similar concentration study for SDS due to the formation of a gel at concentrations above 1 wt. %.

As can be seen from the appearance of a shoulder-like feature at q ∼ 0.1 Å⁻¹, in the SANS data for solutions above 3 wt. %, intermicellar interactions are observed in the scattering pattern. Therefore, these data were fitted using a core–shell form factor multiplied with a hard sphere structure factor with a radius independent of the micelle size. This is a methodology applied to micelles in DES so as to take into account interactions that may be mediated via structured solvent.^21,38 The core-radius, shell thickness, and SLD of the shell for the micelle were fixed at the values obtained from the fit at 1 wt. %. The effective radius was co-fitted for the two contrasts and various concentrations (>3 wt. %) with the volume fraction varied to account for the change in concentration. The fits obtained are shown as the dashed lines in Fig. 5 with the calculated and fitted structure factor volume fractions summarized in the supplementary material, Table S6. For both surfactant concentration series, the above model provides reasonable fits with the fit volume fraction showing good agreement with the calculated ones. This suggests that the surfactant micellar structure (shape and size) is not affected by concentration and a structure factor with a radius independent of the concentration can explain the intermicellar interaction. In order to verify this, the concentrations were each fitted to individual radii. De-constraining the fits resulted in radii, which were within 2% of the value obtained for the combined fit and did not result in any significant improvement to the fit. Therefore, a single radius and varying volume fractions between different concentrations are considered the best suited model to describe the intermicellar interaction contribution to these scattering patterns.

The value of the interaction radius obtained from the structure factor fits at high concentration (>3 wt. %) for C₁₂TANO₃ was found to be 22.4 ± 0.1 Å. Even though no fitting bounds were applied to the intermicellar interaction radius, this value is consistent with the size of the C₁₂TANO₃ micelles, which was found to be 22.0 ± 1.1 Å from the fits for 1 wt. % data at multiple contrasts. This is analogous to the approach used for C_nTAB micelles in ChCl:Gly²¹ and ChCl:U:Gly³⁸ DES where the effective interaction radius with the same second virial coefficient as the scatterer is used in the case of ellipsoidal micelles. Here, we have additionally demonstrated the equivalence of the two radii by obtaining a consistent value for the two despite, allowing the interaction radius to vary.

A similar approach as the one above applied to C₁₂TANO₃ gives an interaction radius of 36.6 ± 0.1 Å for the high concentration (>3 wt. %) C₁₆EO₈ micelles. This is slightly larger than the micelle size, 31.4 ± 1.8 Å, obtained from 1 wt. % C₁₆EO₈ in Ce:U. The difference can be explained by the fact that the size of the micelle is potentially an underestimate due to the solvation of the shell region by the DES resulting in loss of scattering contrast beyond a certain distance, manifested also in the shell thickness for the C₁₆EO₈ micelles being smaller than the size of the polyethoxy (EO₈) head group expected in pure water. Assuming that the shell thickness obtained from the fitting of multiple contrast at 1 wt. % is an underestimate, the actual shell thickness would be 13.1 ± 1.2 Å as opposed to 7.9 ± 0.9 Å. Another reason for the higher interaction radius could be that the interaction is mediated by structured solvent between the micelles, a more likely explanation given we suspect polyethoxy chains to be collapsed in the core due to less favorable interactions with the solvent.

The type IV DES studied in this work, Ce(NO₃)₃·6H₂O:3.5*Urea (Ce:U), has unique characteristics and physicochemical properties, such as high density, small molar volume, low viscosity, and high surface tension when compared to typical type III DESs. This results in the solvent having a Gordon parameter comparable to, even slightly higher than, that of water. The high Gordon parameter along with the highly ionic nature of the solvent results in self-assembly of a range of surfactants in this DES.

The cationic (C₁₂TAB and C₁₂TANO₃), anionic (SDS), and non-ionic (C₁₂EO₆ and C₁₆EO₈) surfactants studied in this work remain surface-active in the Ce:U DES as shown by the surface tension data. They form micelles at concentrations above the CMCs, as determined by changes in the DES/air surface tension upon increasing surfactant concentration, as summarized in Table II. The CMCs of cationic and anionic surfactants, C₁₂TAB, C₁₂TANO₃, and SDS, in Ce:U DESs are lower than that observed in water^18,68,99 and ionic liquids.^19,23,28 This is attributed to the presence of highly charged ions in the DES along with the high Gordon parameter. For C₁₂TA⁺ surfactants, this results in a CMC, which is even lower than that observed in type III DESs.^21,22,77 In the case of non-ionic surfactants, the CMC lies in between that of water^70,72 and ionic liquids.^19,25,26 We attribute this to a combined effect of the organic and highly ionic character of the solvent, which offers some reduction in the solvophobic interactions between the surfactant and the DES but not to the same extent as that in ionic liquids or polar organic solvents, such as formamide.

The structure of the surfactant micelles at concentrations above the CMC was investigated using Small Angle Neutron Scattering (SANS). The scattering data show the presence of micelles with a core–shell structure that comprise the alkyl tail core and the trimethylammonium/sulfate/polyethoxy head group shell. The size and shape of the micelles are summarized in Tables III–V. The scattering curves for dodecyl surfactants and a visualization of the micellar structures are shown in Fig. 6. The core-radius in each case is consistent with the corresponding C₁₂/C₁₆ chain lengths and is in good agreement with values obtained for these surfactants in other solvents. The shell region, on the other hand, comprises highly solvated headgroups, and co-fitting the various contrasts gives an indication of the dominant interaction mechanism for the particular surfactant. For cationic and anionic surfactants, the interaction is primarily electrostatic modulated by counterion binding, as is the case for ionic surfactants in water. However, in the case of non-ionic surfactants, in addition to the ions, urea in the DES offers polar organic interactions resulting in a different pathway to surfactant morphology modulation. The micelles are spherical in the case of cationic (C₁₂TAB and C₁₂TANO₃) and non-ionic (C₁₂EO₆ and C₁₆EO₈) surfactants and cylindrical in the case of the anionic surfactant SDS.

For surfactant/ionic liquid solutions, it has been shown that a surfactant–solvent counterion exchange can occur in the system, where surfactant counterions are incorporated into the bulk solvent, while ions in the solvent interact with the micelle surface.¹⁸ As has been observed for aqueous systems,¹⁰⁰ the morphology of ionic surfactant micelles in type III DES is strongly dependent on the interaction of the surfactant headgroup and counterion with the solvent.^{20–22,38,94} These types of interactions can also explain the morphology of ionic surfactant micelles in the type IV Ce:U DES. For SDS, Coulombic interactions between the DES [–Ce–NO₃–]⁺ polycations [Ce₃$(NO3)72+$]⁴⁴ with the surfactant SO⁴⁻ headgroups provide a charge screening effect, which allows for the headgroups to pack more closely, causing elongation of the micelles, whereas in the case of C₁₂TA⁺, spherical/spheroidal rather than elongated micelles are observed due to the presence of nitrate/bromide ions within the DES, which interact with the positively charged headgroups but do not cause any significant disruption to the micelle shape for C₁₂TA⁺ micelles. This effect is also observed for C₁₂TA⁺ surfactants in water where spherical micelles are observed until extremely high concentrations are reached.¹⁰¹ Little difference in micelles size and shape is observed between C₁₂TAB and C₁₂TANO₃ in the Ce:U DES; Br⁻ and $NO3−$ are next to each other in the Hofmeister series and give approximately equivalent micelle properties when binding to micelles in water.¹⁰² Similar observations of micelle morphology have also been made for SDS and C_nTAB micelles in ChCl-based DES, where the interacting ions are [cholinium]⁺ and chloride/bromide.^20,21,38,94

SANS studies show no effect of surfactant concentration on C₁₂TANO₃ and C₁₆EO₈ micelle structure (shape or size) in Ce:U, apart from an intermicellar interaction contribution to the scattering for concentrations >3 wt. %. The intermicellar interaction radius was found to be in good agreement with the micelle size for both C₁₂TANO₃ and C₁₆EO₈, despite being allowed to fit independently. This is in contrast to studies on cationic surfactants in ChCl:Gly²¹ and ChCl:U:Gly³⁸ where small differences in micelle structure with surfactant concentration are observed. The authors attribute this to small fluctuations in the fit arising from the fitting procedure. In this work, each individual concentration was not fitted to a different core-radius and shell thickness, as was the case for studies in ChCl:Gly or ChCl:U:Gly, but instead, the values were fixed at those obtained from 1 wt. % surfactant co-fitted using multiple contrast and only the interaction radius was co-fitted for the different concentrations.

Self-assembly of cationic, C₁₂TAB and C₁₂TANO₃, anionic, SDS, and non-ionic, C₁₂EO₆ and C₁₆EO₈, surfactants in the type IV DES comprising metal salt, cerium(III) nitrate hexahydrate, and hydrogen bond donor, urea, in the molar ratio of 1:3.5 is investigated using interfacial tension and SANS measurements. The Gordon parameter for the DES is comparable to that of water, which drives micellization for a range of surfactants. Surface tension measurements have shown that the CMC of ionic surfactants in this DES is less than that of water, whereas the CMC of non-ionic surfactants is higher than that of water. These studies show that both the ionic character of the DES, which drives interactions in the case of ionic surfactants, and the presence of organic moieties in the DES, which along with the ions affect interactions with non-ionic surfactants, play a role in micellization. The structure and morphology of surfactant micelles were investigated using contrast variation SANS. The results show the formation of core–shell micelles with a spherical (C₁₂TAB, C₁₂TANO₃, C₁₂EO₆, and C₁₆EO₈) or cylindrical geometry (SDS) depending on the nature and strength of interaction of the surfactant headgroups with the solvent components. Counterion binding plays an important role in the morphology modulation in the case of ionic surfactants; [–Ce–NO₃–]⁺ polycations bind to the SO⁴⁻ headgroups in SDS leading to charge screening and close packing of the headgroup resulting in cylindrical micelles, whereas the nitrate/bromide ions that bind to the cationic surfactant headgroup do not cause any significant disruption in headgroup packing or micelle shape, resulting in spherical micelles. On the other hand, for non-ionic surfactants, both the organic (urea) and ionic [Ce(NO₃)₃] components play a role in micellization leading to a CMC and structure that lies between that in aqueous solutions and organic polar solvent/ionic liquids. To the best of our knowledge, this is the first reported work on micellization of non-ionic surfactants in a deep eutectic solvent. The results of this study have provided an insight into the mechanism of micelle formation in these novel highly ionic but hydrogen-bonded solvents, leading to an important understanding about the interaction of DES components with amphiphiles and how this influences their self-assembly. This raises the opportunity to use these soft structures as liquid crystalline templates for generating porosity in synthesis applications.

See the supplementary material for characterization data for the Ce:U DES (DSC data, viscosity measurements with temperature, and glass transition temperatures calculated from viscosity); scattering length densities for the Ce:U DES, its components, and surfactant components used to fit SANS data; model fitting trials and details of parameters from the fitting of the SANS data; surface tension measurements and the calculated critical micellar concentrations in different surfactant concentration units (mM, wt% and mole fraction).

I.M. acknowledges funding from EPSRC (Grant No. EP/S020772/1). M.R.A. acknowledges EPSRC for funding (Grant No. EP/S021019/1). R.S.A. acknowledges the EPSRC Centre for Doctoral Training in Sustainable Chemical Technologies for funding (Grant No. EP/L016354/1), and J.H. acknowledges the University of Bath for Ph.D. funding. The authors acknowledge the ISIS Neutron and Muon Source for neutron beamtime (Experiment No. RB1920676). This work benefited from the use of the SasView application, originally developed under NSF Award No. DMR-0520547. SasView contains the code developed with funding from the European Union’s Horizon 2020 research and innovation program under the SINE2020 project, Grant No. 654000.

The authors declare no conflicts of interest.

The data that support the findings of this study are openly available in the Bath research data archive system at https://doi.org/10.15125/BATH-01049.