Nuclear magnetic resonance and small-angle X-ray scattering studies of mixed sodium dodecyl sulfate and N,N-dimethyldodecylamine N-oxide aqueous systems performed at low temperatures

Surfactant crystallisation is important in many applications in the food, consumer product and medical sectors. However, these processes are not well understood. In particular, surfactant crystallisation can be detrimental to the stability of detergent formulations, such as dish liquid products, resulting in a turbid solution that fails appearance criteria. With the rising global demand for detergent products, understanding the factors that influence formulation stability is of increasing importance. To enable industry to build more robust formulations, it is important to understand the underlying chemistry of the crystallisation process. Here, a model system containing anionic (sodium dodecyl sulfate, SDS) and amphoteric (N,N-dimethyldodecylamine N-oxide, DDAO) surfactants, at concentrations typical of dish liquid products, is studied. Variable temperature 1 H nuclear magnetic resonance (NMR) spectroscopy and small-angle X-ray scattering (SAXS) is used to probe the compositional and structural properties of this system, as a function of pH. On cooling, at pH 9, a mixture of hydrated crystals, predominately composed of SDS, and micelles containing both surfactants, have been observed prior to complete freezing. At pH 2, both surfactants appear to undergo a simultaneous phase transition, resulting in the removal of micelles and the formation of hydrated crystals of mixed composition.


Introduction
Surfactants are important components in many applications, including the manufacture of food products, jet fuels, medical treatments and consumer products. 1 At low temperatures, such surfactants can crystallise, which may be either essential or detrimental to their performance in a given application. 2 Despite this importance, there remain relatively few studies on surfactant crystallisation, in particular the kinetics, crystal composition and structure. Of significant interest is surfactant crystallisation at low temperatures, which causes stability issues for liquid detergents. While this process is reversible, it is considered a failure in those products. Such failures can occur at any point during the product lifecycle, especially during transport and shelf-life stages. Hence, it is important to enhance the understanding of surfactant crystallisation in order to improve the stability of these formulations. Such knowledge will also be of significant interest in other applications of surfactants in food, biomedical and consumer product applications.
Common liquid detergents typically contain a mixture of anionic and amphoteric surfactants in an aqueous formulation 3 to optimise soil removal and foaming characteristics. Anionic surfactants are the major surfactant component and are cheap, efficient at removing soils and are largely responsible for the high foaming characteristics of the detergent. 4 The most 3 commonly used anionic surfactant in liquid detergents is sodium dodecyl sulfate (SDS). 5 While amphoteric surfactants comprise the minor surfactant component, they also play an important role by increasing the tolerance of the detergent to increased water hardness. 6 Common amphoteric surfactants found in liquid detergents are those based around the amine oxide functional group. One example of this is N,N-dimethyldodecylamine N-oxide (DDAO). At pH < 5, DDAO is protonated and behaves as a cationic surfactant, whereas at pH > 5 it is non-ionic. [7][8] In hand-dishwashing detergent products, commonly referred to as dish liquid, the typical pH is sufficiently high that amine oxide-based surfactants exhibit nonionic surfactant-like behaviour.
In mixed micelles containing SDS and DDAO, the surfactant headgroups strongly interact, leading to a reduction in the critical micelle concentration, CMC, versus that of either surfactant alone. 9 Ion pairs form between the two surfactants which stabilise the mixed micelles through a shielding of the electrostatic repulsion between the SDS headgroups by those of DDAO. 9 Moreover, it is believed that the presence of SDS causes DDAO protonation to occur at a higher pH than that of pure DDAO. [9][10] Combining these surfactants is also expected to lower the Krafft temperature, T K , in comparison to a pure SDS solution.
This drop in T K can be attributed to the lower CMC 11 of the system and the formation of nonideal mixed micelles. 12 With DDAO present, there is expected to be a decrease in the concentration of SDS monomers and unbound counter-ions, and consequently a decreased tendency for precipitation.
Studies into the precipitation of mixed surfactant systems have been carried out to some extent, but there remains limited literature reporting the nature of the precipitate. 6,13 For example, HPLC was used to determine the composition of the precipitate formed when the non-ionic surfactant nonylphenol ethoxylate (NPE) was added to SDS and found that NPE was not present in the precipitate. 6 Furthermore, the composition of the precipitant from a bi-4 anionic surfactant system containing the surfactants SDS and sodium octylbenzene sulfonate (SOBS) has also been investigated via X-ray diffraction (XRD). 13 In the mixed sample, two sets of peaks were observed with 2θ values corresponding to pure SDS and SOBS crystals, suggesting that mixed crystals were not formed. Aside from these studies, there are few On the other hand, several studies have used SAXS or small-angle neutron scattering (SANS) to provide structural insight into various phase transitions, [17][18][19] and to probe the structure and composition of micelles in dilute systems of both pure and mixed SDS and DDAO. [20][21] In this paper, NMR has been used to observe the crystallisation of a mixed micelle system, comprising SDS and DDAO surfactants, in water. NMR parameters for the two surfactants were monitored as the solution was cooled from 25 °C to -3 °C, at pH 9. By comparing the change in relative peak intensity for the two surfactants, it was possible to identify the crystal composition formed in this mixed micelle system. SAXS was used to probe the structure of the systems at room temperature and at 0 °C. Furthermore, a difference in the behaviour was observed upon lowing the pH of the system. By combining results from these complementary techniques, it has been possible to build a clearer picture of both the structure and composition of the phases that form in these systems under both pH environments. Before processing, uncertainty estimates based on Poisson counting statistics are added to all measurement data, which are subsequently propagated through the image correction steps.

Materials
Each raw background measurement was corrected for the following in order: masking pixels, time, incident beam flux, and transmission. Each sample file was corrected for the following in order: masking pixels, time, incident beam flux, transmission, background, thickness, and scaled to absolute units. The scaling factor for scaling to absolute units was determined using a calibrated glassy carbon sample. 25 After this correction, the data was azimuthally averaged, with the resulting uncertainty assuming the largest of: 1) the propagated uncertainties, 2) the standard error of the mean for the data points comprising a bin, or 3) 1% of the mean intensity in the bin.
The data at 24 °C was analysed using the SASfit software package. 26 There is some discussion in the literature concerning whether SDS micelles are oblate or prolate ellipsoids. [20][21][27][28][29] The two options are difficult to distinguish using scattering techniques, but the prolate model is considered to be more appropriate in denser systems. 28 Several smallangle scattering studies, including one that focused on mixed micelles comprising SDS and DDAO, 20 have favoured the prolate shape. Consequently, a model comprising a delta distribution of charged core-shell prolate ellipsoids, as outlined in the supporting information, was used to analyse the data. A breakdown of the S(Q) and P(Q) contributions to the overall fit is shown in Figure S3 and the analysis parameters are provided in Table S3.
The data at 0 °C for both the pure SDS solution and the mixed SDS + DDAO system was also analysed using SASfit, 26 in both cases using a simple model comprising contributions 8 from one or more Bragg peak(s), the power law and the background scattering (see details in SI). For the 20 wt. % SDS + 3 wt. % DDAO sample, an additional contribution of delta distribution of charged core-shell prolate ellipsoids was added, to account for the increased I(Q) at high Q.   In contrast, at pH 2, both surfactants display a sharp drop intensity between 5 °C and 3 °C ( Figure 3). There is a change in the chemical shifts for the H a (DDAO) peak, compared to the H 1 (SDS) peak at pH 2, because of the change in environment experienced by these protons when the headgroup is protonated. [30][31][32] In addition, at pH 2, there are a greater number of peaks in the region of interest, attributed to interaction between the two charged surfactant species resulting in further proton environments. In Figure 4, the SAXS profiles, at 24 ˚C and 0 ˚C, for the mixed system are compared to the corresponding profiles for a pure SDS system. In Figure 4(a), a cropped Q-range is shown due to the high error in the low Q (< 0.8 nm −1 ) part of the data arising from the weak sample scattering versus background compounded by a high sample volume fraction. The data, collected at 24 °C, is characterised by two merged maxima resulting from a form factor, P(Q)

Results and discussion
corresponding to the electron-dense micelle shell and a structure factor, S(Q) arising from the micelle charge. The addition of DDAO causes a shift in the low Q maxima to a lower Q.
Cooling both samples to 0 ˚C results in a notable change in the SAXS data, shown in Figure   4(b). For the 20 wt. % SDS sample, an increase in I(Q) at low Q is observed followed by a region over which I(Q) ∝ Q −4 and a Bragg peak at Q = 1.9 nm −1 The 20 wt. % SDS + 3 wt. % DDAO data at 0 °C is similar to that for the pure SDS sample, except for an increased I(Q) in the region 0.7 < Q < 2.5 nm −1 , which has a similar shape to that presented in Figure 4(a).    Table S5 (supplementary).
At pH 9, the 1 H NMR data suggests that SDS crystallises first, before the system fully solidifies. The drop in NMR signal for the SDS and DDAO protons is expected to be associated with an increase in viscosity and a reduction in the mobility of the monomers and micelles, leading to a reduction in their T 2 * NMR relaxation time and an increase in the width 15 of their peaks. 33 There is a sudden drop in SDS signal between 10 °C and 13 °C, which indicates SDS is undergoing a phase transition. However, not all of the SDS signal disappears at the lower temperature, indicating a proportion of the SDS remains in solution.  20 and is expected to reduce repulsions between the SDS head-groups in the interphase. Consequently, the shell curvature is able to decrease, allowing a more elongated structure to form.
The formation of a Bragg peak in the pure SDS system at 0 °C indicates the growth of SDS crystallites. 21 The calculated d-spacing, d = 3.3 nm, is closer to the values expected for the mono-or hemi-hydrated crystal forms (2.9 or 3.1 nm respectively), but lower than that of the 1/8 hydrate (3.9 nm). 35 For the mixed SDS + DDAO system, the Bragg peak position is the same as the pure SDS system, indicating SDS is the dominant component of the crystals in both samples. This is in agreement with the drop in intensity of NMR signal observed for the SDS protons (H 1 ) (Figure 2). Aside from the Bragg peak, there is an additional contribution in 16 the SAXS profile for the mixed system, which can be fitted to the remaining micelle phase.
For the remaining micelles, N agg = 240 and a/b = 2.4, which is more in line with data reported for mixed SDS + DDAO micelles than that for DDAO micelles alone. 20  and 17.9 nm 1 ) may arise from the head-head and alkyl peaks. 37 In the pure SDS system, more peaks are apparent in this region. However, as the peak intensities for this sample are approximately 2-3 times more intense, over the entire Q-range, this is more likely to be an observational artefact arising from a low peak intensity versus background. In contrast, at pH 2, a different structure forms, with first, second and third order lamellar d-spacings shifted to higher Q (at 2.13, 4.33 and 6.52 nm 1 respectively) giving d = 2.95 nm. This suggests the formation of crystals containing both surfactants.

Conclusions
By combining variable temperature 1 H NMR and SAXS measurements it has been possible, for the first time, to determine the compositional and structural changes in a mixed SDS and DDAO surfactant system, which is of particular interest in the fast moving consumer goods (FMCG) industry. 38 Crystals formed as the mixed SDS + DDAO system is cooled, a pH 9, are observed to be predominately SDS hydrated crystals, with little or no DDAO. At pH 2, the protonated DDAO, and its subsequent ion pairing behaviour, results in both surfactants crystallising simultaneously. While dish liquid is a highly complex system, this research provides important insight into the crystallisation process of liquid detergent systems upon exposure to cold climates. In turn, this understanding will assist in building more robust formulations and improvements to their accompanying test methods.

Associated content
The following files are available free of charge.
NMR spectral assignments and the SAXS model fitting (PDF)