Chapter 1: Inorganic Chemistry Within Nanoreactors

The confinement provided by colloidal systems, in particular by droplets in emulsions, has been extensively exploited in polymer science for the synthesis of organic polymers through so-called emulsion polymerisation techniques.^1–5 In an analogous way, emulsions can also be employed for confining inorganic reactions, typically of precipitation, oxidation/reduction or sol–gel processes.⁶ Several publications in recent years, including some from us, have periodically reviewed the use of droplets and nanodroplets for the synthesis of inorganic materials.^6–9 Our aim in this chapter, rather than to provide an exhaustive research review, is to cover in an explanatory fashion the possibilities offered by microemulsions and, especially, by miniemulsions for obtaining nanostructured inorganic materials, and providing thereby examples of different synthetic strategies. Based on a didactic perspective, we will also review some related general concepts of colloidal chemistry, but specifically focused on the field of inorganic reactions within nanodroplets.

Before starting the description of the synthetic strategies, it is convenient to make a few initial remarks about the confusing terminology in the field of emulsions. Traditionally, emulsions have been classified in macroemulsions (or simply emulsions), miniemulsions and microemulsions.¹ This nomenclature has been followed by polymer scientists and part of the colloid community.^5,10 However, another partially overlapping term is very often found in the literature, especially—although not only—in the fields of food science and pharmacy: nanoemulsions.^11,12 We will revisit here the most relevant aspects of the differences between these emulsion terms, which are discussed in more detail elsewhere.^5,9

Macroemulsions are thermodynamically unstable emulsions with droplet sizes above 500 nm (usually >1 µm). Microemulsions, in contrast, are characterised by being thermodynamically stable and have droplet sizes below 50 nm. Miniemulsions are in between these two types and are kinetically stabilised emulsions but thermodynamically unstable (like macroemulsions), with droplet sizes in the range between 50 and 500 nm.^2,3 The term nanoemulsion (systematically and intentionally hyphenated by some authors^13–15 as nano-emulsion) should in principle be used only for miniemulsions with sizes below 100 nm (or 200 nm, at most). Although the term ‘nanoemulsion’ is often used interchangeably with ‘miniemulsion’, nanoemulsions are a subdivision of miniemulsions, corresponding to nanosized (below 100 nm) and thermodynamically unstable emulsions. This classification of emulsions is schematically represented in Figure 1.1. The inclusion of microemulsions within the label ‘nanoemulsions’ is discouraged, because microemulsions are supposed to be thermodynamically stable, by definition. Although some contradictory examples are found from time to time in the literature, we encourage the reader to follow this recommendation, as it is the consensus preference of most of the research groups working in the field.

where γ is the surface tension and r is the radius of the droplet. As an example, according to this expression, for water droplets with a radius of 1 mm, the pressure difference is 145.6 Pa = 1.09 mm Hg (considering γ = 72.8 mN m⁻¹ at 20 °C for water). By reducing the radius by a factor of 10 (i.e., 0.1 mm = 100 µm), the pressure is multiplied by the same factor, so that ΔP = 10.9 mm Hg. If we go to droplets with a size of 100 nm (r = 50 nm), the pressure difference would be as high as 2.2 × 10⁵ mm Hg (ca. 29.1 bar). This simple calculation makes very obvious the great effect that the formation of a curved interface has on the pressure of the confined space within a nanodroplet.

In addition to the advantages related to the formation of smaller particles in smaller spaces (higher specific surface area)¹⁶ and the possible finite-size effects in crystallisation and selection of polymorphs,¹⁷ the high pressure within nanodroplets can also allow the formation of crystal phases of inorganic materials that otherwise would require higher temperatures.¹⁸ A shift in the phase diagram through the pressure resulting from the space confinement could even allow the stabilisation of certain polymorphs.⁹ Also, other properties of the solvent (typically water in inorganic synthesis), which are relevant for solubilisation and crystallisation phenomena (e.g., viscosity, relative permittivity), dramatically change when confined in constrained environments.^19–21 Emulsions containing nanodroplets (i.e., miniemulsions and microemulsions) have, therefore, significant potential in inorganic synthesis. It is very important to remark that, in most cases, in miniemulsion and microemulsions, crystallisation of the materials takes place under very mild conditions, often at room temperature or close to it.²² 

As mentioned in Section 1.1, the use of emulsions in synthesis has been reviewed in different works. Some representative publications are contained in Table 1.1, with a brief indication of the focus in each case.

The title of this chapter contains the term ‘nanoreactor’, which has been very commonly linked to nanodroplets in miniemulsions.^3,23,24 The term can also be used for other types of colloidal confinement, such as micelles, colloidal assemblies (see the review by Swisher et al.²⁵) or nano-/mesopores, but we will focus our attention here on nanodroplets. If we move away for a moment from the inorganic synthesis to emulsion polymers, and we compare conventional emulsion polymerisation with miniemulsion polymerisation, the former is characterised by diffusion from monomer droplets, acting as reservoirs, to the growing particles being formed. By contrast, in miniemulsion polymerisation, there is no diffusion between droplets, and the polymerisation reaction takes place (at least ideally) within the droplets, which act as independent ‘nanoreactors’. To maintain the stability of the system, two types of agent are normally used: effective surfactants, which stabilise the droplets (sterically, electrostatically or electrosterically) and minimise the coalescence between them; and osmotic-pressure agents, which help to avoid Ostwald ripening.^2–7 For inorganic synthesis, the idea is essentially the same: in an ideal case, the formation of the desired compound should take place within a droplet containing the precursor, while avoiding diffusion of the contained matter towards other ‘nanoreactors’ (see cartoon in Figure 1.2). In a real system, however, this situation is very difficult to reach, and diffusion and interaction between droplets and growing particles occurs in most of cases, and is more likely than in polymerisation reactions for organic polymers.

Since both γ and ΔA are positive (the formation of an emulsion involves an increase of the overall surface area), ΔG_interf is always positive and it becomes increasingly positive with decreasing droplet size (the interfacial area increases with decreasing size). By contrast, the configuration free energy, −TΔS_config, becomes increasingly negative with decreasing droplet size (the number of different ways in which the droplets can be organised increases). For spontaneous formation of an emulsion, ΔG_form has to be <0, so that ΔG_interf will oppose the formation of the emulsion, whereas the configuration term will favour it. Since the interfacial term dominates the configuration term, the formation of an emulsion should be thermodynamically less favourable when the droplet size decreases. However, this is only true if we assume that the interfacial tension remains constant with the curvature radius, which is not the case when the size decreases. Indeed, the interfacial tension does not have a trivial dependence on the droplet size. Therefore, the dependence of the overall formation free energy with the size is more complex than initially assumed. For detailed explanations about the contributions of the interfacial and the configuration entropic terms, the work by McClements²⁶ is recommended.

For macroemulsions, the interfacial term clearly dominates (the increase in the interfacial area is much more relevant than the increase in the possible configurations), so that γΔA ≫ TΔS_config, and ΔG_form > 0. For miniemulsions, the interfacial area increases, but so does ΔS_config; eventually, the former continues to dominate the latter, and γΔA > TΔS_config, so that ΔG_form > 0 also. Accordingly, macroemulsions and miniemulsions are both thermodynamically unstable and work has to be done to form them. From a kinetic point of view, both need to be stabilised, but miniemulsions achieve much longer stability than macroemulsions.

The situation for microemulsions is more complex. When the droplets become sufficiently small, the interfacial tension of a droplet can no longer be assumed to be the same as for planar interfaces. As mentioned previously, the interfacial tension depends on the curvature of the surfactant monolayer and so at a certain small droplet size (typically <10 nm), the value of the interfacial tension as a function of the droplet radius reaches a minimum. Around that minimum the interfacial free energy term, which usually dominates, is substantially reduced. Under these conditions, γΔA < TΔS_config and ΔG_form becomes negative, so that the formation of microemulsions is spontaneous. In other words, the resulting system is thermodynamically stable.

According to this eqn (1.6), general ways to reduce sedimentation/creaming are either by decreasing the droplet size or increasing the viscosity of the continuous phase.

In addition to gravity, under certain conditions, temperature (affecting solubility) and concentration of the disperse phase with respect to the continuous one, may be the origin of a phase inversion. Furthermore, attractive forces between droplets, typically from van der Waals interactions or introduced by charges in the system resulting from components of the formulation (charged precursors or additives), can lead to aggregation (flocculation, coagulation). Finally, two other effects are also crucial in the destabilisation of emulsions: coalescence and Ostwald ripening. Since they are especially relevant, we will discuss in more detail the driving forces behind them and what strategies can be used to minimise them.

which is smaller than double the area given by eqn (1.9), $2(4πr12)$⁠. The new droplet has about 20.9% less area than the sum of the two initial droplets.

With the exception of surfactant-free systems (which are rather unusual) and cases of spontaneous emulsification or the ‘ouzo effect’^27,28 (also rare for inorganic compounds), any emulsion strategy implies the use of surfactants, which stabilise the emulsions against coalescence, but also against aggregation and flocculation. Non-ionic surfactants provide a steric stabilisation, while ionic surfactants provide electrostatic or electrosteric (electrostatic and steric simultaneously) stabilisation. The mechanism of emulsification is schematically depicted in Figure 1.4. In a first step, a pre-emulsion is obtained by mechanical or magnetic stirring (droplets with sizes of several micrometres and polydispersed are initially formed). Afterwards, disruption (i.e., comminution) of the droplets is achieved by the application of mechanical energy. The formation of miniemulsions requires high shear forces (ultrasonication or high-pressure homogenisation are the most common methods) to reach stable nanosized droplets. Achieving a stable emulsion will require efficient droplet breakup and minimisation of coalescence.

Representative examples of surfactants used in micro- and miniemulsions are shown in Figure 1.5. Poly(ethylene-co-butylene)-block-poly(ethylene oxide) (P(E/B-b-PEO)), often referred to as ‘KLE’ (an acronym comprised of ‘KL’, for Kraton Liquid, the company that produced the ethylene/butylene block, and ‘E’ for ethylene oxide), has been widely used for stabilising, very effectively, water-in-oil systems. Unfortunately, in spite of its great stabilisation ability, this copolymer is not commercially available, requires a complex and tedious synthesis,²⁹ and so is not a very convenient option for most research groups. In this respect, surfactants such as polyglycerol polyricinoleate (PGPR) or sorbitan oleate (commercially known as Span 80) are much more readily available for stabilising inverse systems, even if they are sometimes less effective than P(E/B-b-PEO). The stabilisation of oil-in-water (direct) systems is generally easier. Although ionic (anionic or cationic) surfactants are normally very efficient, they may interact easily with other charges in the systems, leading to coagulation. Therefore, especially when metal complexes are involved as precursors, non-ionic surfactants may be preferred. In this context, polyoxyethylene alcohols linked to apolar chains are a common choice.

where M_hydro is the molar mass of the hydrophilic segment of the surfactant and M_total is the total molar mass, including the hydrophilic and the lipophilic segments. An index between 0 and 8 indicates that the surfactant is oil soluble and suitable for stabilising water-in-oil systems, and surfactants with indices above 11 are water soluble and normally good as oil-in-water emulsifiers. Indices between 8 and 11 indicate poor solubility in both oil and water, and may result in the formation of lamellar structures or vesicles.

where m and n are the number of hydrophilic and lipophilic groups, respectively, and H_hydro and H_lipo are arbitrary values assigned to the different chemical groups, of hydrophilic and lipophilic (i.e., hydrophobic) nature, respectively. Table 1.2 lists the H values proposed by Davies for some hydrophilic and lipophilic groups, making some special considerations for ether groups (ethylene oxide is hydrophilic, while propylene oxide is considered lipophilic). It should be noted that eqn (1.14) may sometimes be found in the literature with a negative sign before the lipophilic term; in this case, the values in Table 1.2 for lipophilic groups should be positive instead of negative, so that the final result for the HLB index is the same.

Although HLB values provide indications related to application in direct and inverse systems, in practice the identification of suitable surfactants is slightly more complex, especially for microemulsions. Surfactants such as cetyltrimethylammonium bromide (CTAB), which is ionic and with a HLB index above 20, can be used in water-in-oil microemulsions.^32–34 In miniemulsions, in contrast to microemulsions, high or low HLB values are quite a good estimation of the suitability for oil-in-water or water-in-oil systems, respectively.

where S(r) is the solubility of a particle or droplet of radius r, S(∞) is the solubility of the bulk (r → ∞), γ is the surface tension of the liquid, $VmL$ is the molar volume of the liquid, R is the gas constant and T is the absolute temperature. For efficiency, the added osmotic-pressure agent should have a lower solubility in the continuous phase than the predominant compound of the disperse phase, so that the overall solubility will be slightly lower than without it. When the predominant compound diffuses from smaller droplets to bigger ones, the solubility of the small ones will be further reduced. Consequently, the increased solubility from the Kelvin effect will be balanced by the decrease in solubility when adding the additional compound.³⁶ In this way, the addition of a third component to emulsion systems is able to minimise Ostwald ripening. In the case of inorganic reactions within droplets, inorganic precursors themselves (e.g., inorganic salts in the case of water-in-oil emulsions) may directly have the effect of osmotic-pressure agents.

As an attempt at systematisation of the possible routes for synthesising inorganic nanoparticles from emulsions, we classify the strategies into three groups, which are schematically represented in Figure 1.6: (i) the ‘two-emulsion method’; (ii) external addition of a precipitating agent; and (iii) combination of precursors in the disperse phase, or use of single-source precursors. For completeness, some examples of strategies falling outside this classification will also be discussed at the end of this section.

This method involves mixing two separate miniemulsions, one containing a first reactant (typically a metal precursor) and another containing a second reactant (normally a precipitating agent). Upon application of high shear forces (produced, for example, by ultrasound), the droplets of both components fuse, coalesce and their contents react, leading to the formation of the desired compound, or of an intermediate that eventually converts to it (e.g., hydroxide species that transform to an oxide). This strategy has been used, for example, to prepare ZnO nanoparticles by mixing two emulsions containing a zinc salt and a base, such as NaOH or ammonia.^37,38

In this case, the second reactant (precipitating agent) is added externally to an emulsion in which the first reactant (metal precursor) is contained. The added reactant dissolves in the continuous phase and diffuses to the droplets containing the first component, driving the reaction. Since the interface between the droplet and the continuous phase is the first place where the two components meet, hollow structures can also be formed, as discussed in Section 1.5, depending on the reaction kinetics and how fast the diffusion takes place.^22,39–41 The concentration of the precursor within the droplets can be crucial for obtaining solid or hollow particles, as reported for Ce_1−xCu_xO₂ systems, in which hollow structures are formed at lower precursor concentrations and solid particles at higher concentrations.⁴⁰ The chemistry of the precipitating agent (more or less soluble in the disperse phase, depending on polarity), as well as the rate of addition, can also determine the formation, or not, of solid particles.⁴¹ 

In the previous reactions, the peroxo species on the left-hand side decompose to produce molybdenum trioxide hydrates. Such processes have been investigated in miniemulsion systems to produce molybdic acid and, analogously, tungstic acid.⁴⁴ 

Another example of the use of a complex single-source precursor in a miniemulsion was reported for the preparation of Au/TiO₂ materials starting with a gold-containing titanium peroxo complex with the structure AuCl₄(NH₄)₇[Ti₄(O₂)₄(cit)(Hcit)₂]₂·12H₂O, which was photochemically decomposed in the confined space of the droplets.⁴⁵ 

In addition to the three previously described groups of strategies, a different approach for the formation of inorganic nanoparticles is the use of polymer–inorganic hybrid nanoparticles as precursors that can transform to the targeted material by oxidation or reduction of the metal (e.g., through thermolysis or plasma steps). For example, different organotin compounds were encapsulated in polystyrene nanoparticles by a miniemulsion polymerisation procedure; after thermolysis of the films formed with the resulting suspension of organotin/polystyrene nanoparticles, tetragonal SnO₂ (cassiterite) nanoparticles were obtained.⁴⁶ In a slightly more sophisticated way, plasma treatment can be applied to hybrid nanoparticles containing metal complexes to transform them into highly ordered arrangements of metal nanoparticles.⁴⁷ In this process, hexagonal close-packed nanoparticles deposited on a surface are initially treated with oxygen plasma to remove the organic parts of the precursor particles, and subsequently thermally annealed to obtain the metal nanoparticles. This approach, depicted graphically in Figure 1.7, has been used for the preparation of platinum^47–50 and Fe–Pt alloy^49,51 nanoparticles. An analogous process was also reported for the preparation of ZnO nanoparticles from hybrid precursor particles containing zinc complexes.⁵² 

Most of the work in the field of inorganic synthesis in emulsions has been conducted in microemulsions. In this context, it is important to mention the pioneering work of Pileni in the application of nanodroplets and colloidal soft templates (micelles and colloidal associations) to prepare inorganic nanoparticles.^53–55 Many other groups have also made significant advancements in this area, with especially remarkable contributions from Feldmann^34,56–60 and Sánchez-Domínguez^61–68 and their respective teams. Both groups have been very active in recent years and have prepared a wide range of materials by using microemulsions.

Since many inorganic precursors are water soluble, it is common to use inverse emulsions, that is, water-in-oil microemulsions in which the disperse phase is aqueous (or based on a polar solvent or mixture of polar solvents) and the continuous phase is apolar. In this case, the inorganic precursor is solubilised in the aqueous phase, with inorganic salts being very common reagents. However, oil-soluble precursors, such as metal acetylacetonates, can also be used in direct (i.e., oil-in-water) emulsions for the synthesis of inorganic compounds.^{61–63,69,70}

As reported in Table 1.3, a large number of inorganic systems have been prepared in microemulsions, including metal oxides, mixed oxides, metal hydroxides, chalcogenides and more complex systems.^59,65,71 For the case of miniemulsions, the diversity of inorganic substances prepared is slightly more modest, but it has been increasing in the last few years, in parallel with the development of the technique for inorganic synthesis. Most of the substances prepared have been metal oxides (see Muñoz-Espí and Landfester⁹), but some chalcogenides (e.g., CuS⁷² and ZnS⁷³) have also been reported. The preparation of ternary systems is also possible, including, among others, Ca(OH)₂ : Ln and Mg(OH)₂ : Ln (Ln = Eu^III, Sm^III, Tb^III),⁷⁴ lanthanide-doped CaF₂,⁷⁵ MFe₂O₄ (M: Co^II, Cu^II, Ni^II, Zn^II),¹⁸ YCrO₃ ⁷⁶ or Ce_1−xCu_xO₂.⁴⁰ 

The majority of work conducted in miniemulsions has applied water-soluble inorganic precursors (typically metal salts). Consequently, inverse (i.e., water-in-oil) miniemulsions have almost always been used. Common apolar solvents in the continuous phase are cyclohexane (b.p. 80.7 °C, which allows for easy evaporation), toluene (b.p. 100.6 °C) or oils, such as n-decane or Isopar M (b.p. >150 °C, which allows for higher synthesis temperatures, but makes it very difficult to remove the solvent).

Table 1.3 contains a collection of the inorganic compounds prepared so far at the nanoscale by using emulsion systems (both microemulsion and miniemulsion), together with a selection of representative references. We include here nanoparticles formed by precipitation within droplets. Particular cases involving interfacial precipitation to form hollow structures are considered in Section 1.5 (see Table 1.4).

Most studies on the use of micro- and miniemulsions for confining inorganic synthesis have been conducted under ambient pressure. In the last few years, some studies have started to explore the combination of miniemulsions with hydrothermal conditions (i.e., temperatures above room temperature and pressures above the ambient one). In other words, miniemulsions are placed under hydrothermal conditions, so that there is a double pressure effect: the intrinsic pressure resulting from the formation of the curved interface of the droplets (Laplace pressure) and the external one applied during the hydrothermal reaction. Different spinel ferrites, including Fe₃MnO₈, CoFe₂O₄, CuFe₂O₄, NiFe₂O₄ and ZnFe₂O₄ have been prepared under hydrothermal conditions by adding NaOH to an inverse emulsion of the metal precursors (metal salts) and ultrasonicating again to allow coalescence and precipitation.¹⁸ This study demonstrated that for some of the zinc ferrite systems, the materials obtained by combining miniemulsion and hydrothermal conditions resulted in much more crystalline ferrites than the analogous ones obtained by miniemulsions at ambient pressure or under bulk conditions, either at ambient pressure or under hydrothermal conditions.

Nanocapsules of inorganic materials can be produced when the formation reaction of the target substance takes place at the liquid–liquid droplet interface. A variety of metal oxides, metal hydroxides and metal chalcogenides have been produced in microemulsion in the form of hollow particles.^{33,56–58,60,77–79}

In miniemulsions, one of the most common inorganic capsular systems is silica, prepared by a sol‒gel process starting at the droplet interface. In a typical oil-in-water formulation, a silicon alkoxide, such as tetraethyl orthosilicate (TEOS), is dissolved in an oil (e.g., octane) and dispersed in a continuous aqueous phase.⁸⁰ The alkoxide groups of the silane precursor are hydrolysed to silanol groups upon contact with water under acidic or basic catalysis. A silica network is produced when the silanol groups condense with each other, leading to the capsule formation. The surface chemistry of the capsule can be tuned by adding functionalised silanes (such as 3-aminopropyltriethoxysilane or 3-mercaptopropyltrimethoxysilane) during the process.⁸¹ The strategy is similar to the copolymerisation procedures used to prepare functionalised polymer particles with hydrophilic monomers in miniemulsion polymerisation.⁸² The silanes with polar functional groups tend to be placed at the surface of the resulting particles.

An analogous strategy for the preparation of silica can also be carried out in inverse systems.⁸³ In this case, the silicon alkoxide is added to the apolar continuous phase, and hydrolyses upon contact with the water droplets of the disperse phase. This situation, together with the formation reactions of silica by hydrolysis and condensation of silanes, is presented in Figure 1.8. By adding silanes with different polarities during the synthesis, the surface hydrophobicity of the resulting capsules can be adjusted.⁸⁴ 

For typical sol–gel systems, such as silica, zirconia, hafnia or titania, the materials produced are amorphous and only crystallise upon calcination. By contrast, crystallisation pathways can occur under mild miniemulsion conditions for various transition metal oxide and lanthanide oxide systems, including γ-Fe₂O₃, CeO₂ and CuO.^22,39

Some of the most representative systems reported in the literature for the production of capsules in micro- and miniemulsions are listed in Table 1.4. This compilation complements and completes the inorganic systems given in Table 1.3.

In this chapter we have seen that nanodroplets in miniemulsions and microemulsions can create enclosed areas where chemical reactions can occur. For this reason, nanodroplets can be considered as ‘nanoreactors’, especially for miniemulsion systems, in which the diffusion between droplets is supposed to be ideally suppressed. The formation of the curved interface at the droplet surface provides an intrinsic increase of the inner pressure, which may be advantageous for obtaining certain crystal phases under milder temperature conditions. Unusual crystallisation pathways may be pursued under such conditions, even allowing the potential stabilisation of certain metastable phases. The solubility of the precursor and some properties (e.g., vapour pressure, viscosity, relative permittivity) of the solvent in the disperse phase, most commonly—but not only—water in the case of inorganic systems, may also be significantly affected by the very small size of the droplets and the related confinement. The confinement can be utilised to tune the size of the particles obtained. In the simplest case, the crystalline size and the number of crystalline domains that can aggregate to form polycrystalline particles will be influenced by the restricted space available for crystallisation.

We have classified the methods for producing inorganic materials in emulsions in three groups: (1) the ‘two-emulsion method’, based on the coalescence of droplets containing the two reactants leading to the desired compound; (2) addition of the second reactant to an emulsion containing the first reactant in the disperse phase; and (3) emulsions in which both reactants are simultaneously present in the disperse phase or, alternatively, emulsions containing a single-source precursor that yields, upon thermally or photochemically induced decomposition, the target materials. Furthermore, besides these strategies, we have also seen that highly structured arrangements of metal oxide and metal nanoparticles can be produced by oxidation or reduction of films of polymer/inorganic precursor particles, after thermolysis or plasma treatment to remove the organic matter, to achieve the desired compound.

In addition to the confinement within droplets to form nanoparticles, the space constriction can also take place at the droplet interface, leading to the formation of hollow structures or capsules. The variety of inorganic compounds reported in the literature is very broad, including silica, metal oxides, hydroxides, chalcogenides, phosphates or even ternary or more complex systems.