All the reactants used in this report were of analytical grade. CTAB was purchased from Sigma-Aldrich Corporation (99%; St. Louis, MO, USA), ferrocyanide salt (K₄[Fe(CN)₆]3H₂O), from J.T. Baker (99%; Deventer, The Netherlands), n-hexane (C₆H₁₄) and NiCl₂6H₂O, from Caledon Laboratories Ltd. (98% and 99%, respectively; Halton Hills, Canada), n-butanol (C₄H₉OH), from Productos Químicos Monterrey (99%; Monterrey, Nuevo León, Mexico), and double distilled water, from Selectropura S.A. de C.V. (σ = 1.5 to 3 μS/cm; Guadalajara, Jalisco, Mexico). For neutron scattering experiments, D₂O (99.9% D; Euriso-Top, Gif-sur-Yvette, France) was used instead of H₂O to increase the contrast and lower the background.

The surfactant CTAFeII was prepared by a direct metathesis reaction in an aqueous phase. The detailed procedure is described in the study of Gutierrez-Bercerra et al. 33. Functional groups of CTAB and CTAFeII were identified by a Fourier transform infrared [FTIR] spectrometer (Spectrum One, PerkinElmer, Waltham, MA, USA). Infrared spectra were recorded in the 400- to 4,000-cm^-1 region, with a resolution of 4.00 cm^-1.

The pseudo-ternary phase diagram for the CTAB + CTAFeII + n-butanol/n-hexane/water system was constructed, considering as surfactant a mixture of CTAB, CTAFeII, and n-butanol, using W _CTAB/W _CTAFeII ratios of 0.95:0.05 and the (W _CTAB + W _CTAFeII)/W _but ratio of 1, where W _CTAB, W _CTAFeII, and W _but are the weights of CTAB, CTAFeII, and n-butanol, respectively. A simple titration technique was used to construct the diagram. Microemulsions were prepared by mixing weighed appropriate amounts of the individual components. The amount of n-hexane (W _hex) in the surfactant mixture determines the H value (H = [W _CTAB + W _CTAFeII + W _but]/[W _hex + W _CTAB + W _CTAFeII + W _but]), while W _w = W _water /(W _water + W _hex + W _CTAB + W _CTAFeII + W _but) represents the weight fraction of water used as the titration component. Water was added in small volumes under permanent stirring in a tightly closed vial to avoid evaporation. Then, the vials were placed in a thermostatic bath (25°C) until a homogeneous media is reached.

Dynamic light scattering [DLS] measurements were performed using an ALV/CGS-3 goniometer with an ALV/LSE-5004 multiple tau digital correlator (ALV-Laser Vertriebsgesellschaft m-b.H., Langen, Germany). The light source was an He-Ne laser operating at a wavelength of 633 nm. The homodyne intensity autocorrelation function g ⁽²⁾(τ) was measured at 90°. Data analysis was performed with the normalized intensity autocorrelation function using a third-order cumulant fit 34 that yielded as key parameter the effective collective diffusion coefficient.

Small-angle neutron scattering [SANS] measurements were done on the instrument PAXY at Laboratoire Léon Brillouin, Gif-sur-Yvette, France. A wavelength of 0.5 nm (FWHM 10%) was selected, and two configurations were used with sample-to-detector distances of 1.25 and 5.05 m.

The synthesis of nickel hexacyanoferrate [Nihcf] nanoparticles was carried out at H = 0.4 and W _w = 0.09. Appropriated amounts of CTAFeII, CTAB, n-butanol, and hexane were mixed until an H value of 0.4 was reached, and then it was maintained under stirring. After that, as an aqueous phase, a solution of 5 mM NiCl₂ was added to the mixture to reach W _w = 0.09. The microemulsion formed was stable for several days and at the same time maintaining a transparent state. Nihcf nanoparticles were separated from the microemulsion media by centrifugation at 9,000 rpm for 10 min. The precipitate was then washed several times with acetone. Despite the washing process, a small quantity of CTAB remained mixed with the nanoparticles. To obtain transmission electron microscopy [TEM] micrographs (JEM-1010, JEOL de Mexico S.A. de C.V., Mexico City, Mexico), a drop of the nanoparticles dispersed in acetone was placed directly on a carbon-coated copper grid. X-ray diffraction [XRD] patterns were recorded with a STOE Theta/theta X-ray diffractometer (STOE & Cie GmbH, Darmstadt, Germany) using a CuKα (λ = 0.15406 nm) at room temperature. FTIR spectra of the Nihcf were carried out in a PerkinElmer Spectrum One spectrometer.

For comparison, the IR spectra of the surfactants CTAB and CTAFeII are shown in Figure 1. The symmetric (υ _s(CH₂), d ⁺) and asymmetric (υ _as(CH₂), d ^-) stretching vibrations of pure CTAB indicate equivalent gauche defects which lie at 2,849 and 2,918 cm^-1 35, as well as those of CTAFeII. The peaks at 3,017 and 1,487 cm^-1, and at 1,473 and 1,462 cm^-1 were attributed to the asymmetric and symmetric C-H scissoring vibrations of CH₃-N⁺ moieties and to the CH₂ scissoring mode, respectively 36. The above mentioned results indicate that both surfactants possess a long aliphatic chain with a positively charged polar head as expected for the hydrocarbon CTAB structure. On the contrary, two peaks only appear in the CTAFeII: at 595 cm^-1 due to the Fe-C vibration and at around 2,000 to 2,100 cm^-1 due to the C≡N stretching 37. Hence, it confirms that indeed the ferrocyanide ion is present in the CTAFeII. The low-spin Fe(II) is diamagnetic and will thus not have electronic transitions. The absorptions near 1,500 and between 1,550 and 1,700 cm^-1 can be attributed to overtones and combination tones of OH^- and H₂O fundamental vibrations. The much lower reflectivity of the CTAFeII is a consequence of the high water content, which produces intense absorption with a broad band near 1,550 and 1,700 cm^-1 because of the water present. In order to quantify this amount of water in CTAFeII samples, Karl Fisher titrations were carried out, obtaining a 2.8% in weight.

The pseudo-ternary phase diagram obtained for the CTAB + CTAFeII + n-butanol/n-hexane/water at 25°C is shown in Figure 2A. The boundary between the microemulsion and the non-microemulsion region was established by a systematic titration based on the clear-turbid observation and conductivity measurements (not shown here). This microemulsion region is transparent because of the small dispersion size of water droplets in the system. Outside this area, the mixture is turbid, indicating that the system reaches the solubilization boundary for water and forms big emulsion droplets. In order to study the influence of alcohol, in addition, a phase diagram with n-pentanol as cosurfactant was obtained (see Figure 2B). The reverse microemulsion region of both phase diagrams is reached at a similar composition but is more extended for the case of pentanol. Obviously, the effectiveness of the cosurfactant increases with increasing chain length, an effect that is typically observed for the formation of microemulsions induced by the addition of a cosurfactant 3839. This indicates that the water solubility has no strong dependence on the type of alcohol and that similar conditions prevail at the amphiphilic interface of the microemulsion aggregates for butanol and pentanol. On the other side, by comparing the phase diagrams in Figure 2C, it is seen that the microemulsion region for the mixture of surfactants examined is much larger than the microemulsion region shown for CTAFeII. This shows that the mixture of surfactants favors larger solubilization of water compared to the pure CTAFeII surfactant.

Figure 3 shows a representative plot of the correlation function g ⁽²⁾(τ)-1 obtained for the microemulsions studied. The solid line is a fit to the data using the cumulant method 34. For a reverse micellar solution, the third-order cumulant expansion of g ⁽²⁾(τ)-1 varies linearly with 2q²τ. From the slope, the effective collective diffusion coefficients [D _eff] were determined. As a first approximation to determine the droplet size, we considered that the microemulsion is formed by non-interacting droplets. In this condition, the hydrodynamic radius [R _h] can be calculated by the Stokes-Einstein equation R_h = KT/6πηD_eff, where k is the Boltzmann constant, T, the temperature, and η, the solvent viscosity (the continuous phase in the case of microemulsions). The obtained radii (2.5 to 4.5 nm) are in the same range as those measured by SANS, proving that the non-interacting supposition can be applied in this system without significant error. D _eff and R _h depend on the H values (see inset in Figure 3, and Table 1), with larger droplets being present for smaller H. An explanation could be that by increasing the hexane content, less butanol is present at the amphiphilic interface (as it becomes dissolved in oil, whereas CTAB and CTAFeII should not be soluble in hexane to any significant extent). Thereby, the total amphiphilic interface available becomes smaller, which then explains the increase in size for a given amount of water. The R _h is also proportional to the relative amount of water contained as demonstrated in earlier works 4041.

SANS was applied for obtaining a more comprehensive structural picture in the relevant size range that, for our given q range, is about 0.5 to 15 nm, where q (q = 4 πsin (θ/2)/λ₀, here, λ ₀ and θ are the wavelength and the scattering angle, respectively) can be interpreted in terms of distances using d = 2 π /q. The obtained scattering curves as a function of q are given in Figure 4. At a low amount of water, the spectra have a low intensity. With increasing W _w, the scattering intensity increases and a pronounced angular dependence develop. Apparently, the water core of the aggregates is now large enough to produce a noticeable scattering and is becoming bigger with increasing water content. The shape of the scattering curves at higher q already indicates that these reverse microemulsions have a globular structure.

In addition, a correlation peak is visible that becomes much more prominent with increasing water content in the reverse microemulsion, and at the same time, its maximum moves from 1.4 to 0.85 nm^-1 (for a fixed H of 0.5). This, together with the intensity increase, shows that the aggregates grow substantially in size with increasing water content, where, however, it should be noted that in SANS basically, only the D₂O core is visible as an aggregate due to the strong contrast between the two isotopes H and D.

The pronounced correlation peak has to be due to steric interactions between the reverse aggregates as electrostatic interactions in the oil-continuous medium should be negligible, but of course, at the concentrations employed, the volume fractions of the amphiphilic material (CTAB, CTAFeII, n-butanol) plus D₂O are in the range of 34% to 59% v/v and therefore high enough to explain effective repulsion already on the basis of purely steric interactions. In addition, it is well known that in reverse microemulsions, the solvent oil molecules are to a certain extent bound to the reverse microemulsion aggregates 4243, thereby enhancing the effective volume fraction further.

The first analysis of the SANS data can be performed using the peak position q _Peak to determine the number density ¹ N of aggregates assuming simple cubic packing: 2π /q_peak = ¹N^-1/3. Then, it can be recalled that ¹ N can be expressed with the volume fraction Φ of the aggregated material and the volume of one aggregate V: N = Φ/V. Assuming spheres, a radius can be deduced. Depending on the choice of material to consider, either only D₂O or the whole aggregated material (D₂O, n-butanol, surfactant), two values are found that can be considered as values for the core (neglecting dissolved butanol) and for the entire droplet (neglecting the contribution from oil swelling the aliphatic chains of the surfactant); the core radius varies from 0.8 to 3.1 nm and is proportional to the water content of the microemulsion. The difference between the core and the overall radius is in all cases around 1.1 nm (1.03 to 1.25 nm), a reasonable value for the surfactant acting as a shell; taking into account the solvation of this shell by hexane, a higher value would be reached; using Tanford's length, the stretched C15 chain is 2.05 nm; the typically retained value of 75% to 80% of this elongation corresponds to lengths of 1.54 to 1.64 nm; the radius of the tetramethylammonium head group is 0.285 nm; the overall thickness expected for the swollen shell would then be 2.11 to 2.21 nm. However, notice that the R _h, obtained by DLS and SANS, increases roughly linear with the water content as typically observed for reverse microemulsion droplets 444546.

Two-dimensional data were reduced using BerSANS accounting for dead time, transmission, and background scattering assimilated to the empty cuvette (which means that the incoherent scattering in the spectra still contains contributions from all compounds in the samples including the solvent), and the scattering from H₂O in a 1-mm cuvette was used to account for the detector pixel efficiency and solid angle variations. Absolute scale was deduced from the evaluation of the direct beam flux. As all corrected scattering patterns are isotropic, they were finally radial-averaged, and data from two configurations were merged.

The whole scattering curves can be described by a model of globular aggregates interacting via an effective hard sphere potential for which the scattering intensity is given by:

I ( q ) = 1 N ⋅ SLD p - SLD s 2 ⋅ P ( q ) ⋅ S ( q ) ,

where ¹ N is the number density of particles, SLD_p and SLD_s, the scattering length densities of the particle and the solvent, respectively, P(q), the particle form factor assuming core-shell spheres, and S(q), the structure factor accounting for the interparticle interactions, keeping the same density number and using an adjustable hard sphere radius. The core is composed of D₂O and butanol; the shell of polar charged moieties is composed of D₂O, butanol, and counterions (bromide is known to adsorb strongly on alkylammonium interfaces, and the amount of ferrocyanide ions is negligible); and the matrix contains hexane and the aliphatic chains of the surfactants (Table 2). The shell thickness was fixed to the dimension of -CH₂-N(CH₃)₃ ⁺ which is 0.57 nm 47. The incompressibility of all the species was assumed as we do not have access to apparent molecular volumes in situ.

To evaluate the feasibility of this model, where butanol is absent from the oil phase and partitions between the core and the shell, a comparison of the experimental invariants INV exp = ∫ 0 ∞ ( I ( q ) - I inc ) q 2 d q with the theoretical invariants

IN V th = 2 π 2 ϕ oil ϕ shell ( SL D oil - SL D shell ) 2 + ϕ oil ϕ core ( SL D oil - SL D core ) 2 + ϕ core ϕ shell ( SL D core - SL D shell ) 2

was performed with the partition coefficient of butanol between the core and the shell as the only adjustable parameter; the volume fraction of water in the shell was fixed to be identical to the volume of the tetraalkylammonium group (ca. four water molecules per group). Identity was found with the partition of butanol toward the water phase increasing with the amount of D₂O, except in two cases where the amount of water was too little to actually allow for a core. The maximum volume fraction of butanol in the core does not exceed 16%, only slightly higher than the solubility limit of the alcohol in bulk water (9.5%). Accordingly, we think that our model is reasonable and self-consistent. In all cases, we observe a very good agreement between the values for the droplet core obtained by the analysis of the peak position and by the full fits (Table 3).

The formation of Nihcf nanoparticles in this system was carried out in four principal stages as depicted in Figure 5. First, the dissociation of the surfactant counterions (ferrocyanide and bromide) is reached when the solution of NiCl₂ is added (dropwise under vigorous stirring) to the reverse microemulsion. Then, the nickel and ferrocyanide ions react to form the first nuclei of Nihcf. Once the nuclei are formed, further growth of the particles is taking place via collisions with other microemulsion droplets containing additional salt precursors. Size and shape of nanoparticles are controlled by the steric stabilization provided by adsorbed surfactant molecules on the surface of the nanoparticles 48. This prospective mechanism still has to be confirmed in more detail by further studies that are currently going on.

The reaction between [Fe(CN)₆]^4- ions and the CTAFeII and nickel(II) ions from the aqueous solution produces a colored (yellowish brown) microemulsion without precipitation, suggesting that this reaction is sufficiently facile to allow for the formation of Nihcf, while suppressing at the same time the further growth due to surfactant stabilization of the nanoparticles. Accordingly, the formed particles remain in the nanometer range and are colloidally dispersed (see Figure 6). The inset in this figure shows the indexing of the electron diffraction pattern of the sample with a 40 direction that coincides with the space group F43m characteristic of the Nihcf 49. Figure 6 confirms the existence of small particles (approximately 6 nm in average size) which have a homogeneous size distribution and correspond to the droplet size of the initial microemulsion droplets (as measured by DLS and SANS).

XRD and FTIR were performed in order to obtain a better characterization of the Nihcf nanoparticles. Figure 7A shows a comparison between the FTIR spectra of the stretching vibration of the cyano group in the Nihcf nanoparticles (solid line) and the surfactant CTAFeII (dashed line). The absorption band at 2,109 cm^-1 can be assigned to the stretching vibration of the C≡N group into the CTAFeII. While for the Nihcf nanoparticles, this absorption band shifts to 2,096 cm^-1 and represents the stretching vibration of the cyano group into the cyanoferrate lattice of Nihcf nanoparticles [Ni²⁺-C≡N-Fe²⁺] 50. On the other hand, Figure 7B shows the XRD pattern of Nihcf nanoparticles mixed with CTAB remained from the washing process. In order to isolate the Nihcf nanoparticle contribution in the diffractogram, the peaks assigned to the diffraction of CTAB 51 were subtracted. The lattice parameter value for the nanoparticles calculated by indexing the peak position using an F43m lattice symmetry is 1.016 nm, close to the 1.000 nm reported elsewhere 49.

An advantage identified in this work for the synthesis of coordination compound nanoparticles is that with this system, different transition Mhcf can be obtained only by varying the transition metal (copper(II), cobalt(II), iron(III), etc.) in the aqueous phase. This constitutes an alternative method using cationic, modified surfactants in reverse microemulsion for the synthesis of this type of nanoparticles.