WO2010022441A1

WO2010022441A1 - Extraction processes

Info

Publication number: WO2010022441A1
Application number: PCT/AU2009/001086
Authority: WO
Inventors: John Ralston; Craig Ian Priest; Rossen Velizarov Sedev; Kazuma Mawatari; Takehiko Kitamori
Original assignee: University Of South Australia
Priority date: 2008-08-25
Filing date: 2009-08-25
Publication date: 2010-03-04
Also published as: ZA201205862B

Abstract

A process for extracting a solute from an analyte-containing fluid phase. The process includes passing the analyte-containing fluid phase along a first fluid microchannel of a microfluidic extraction device and passing an extractant fluid phase that is at least partially immiscible with the analyte-containing fluid phase along a second fluid microchannel of the microfluidic extraction device. The analyte-containing fluid phase and the extractant fluid phase contact one another at a contact zone formed between the first and second fluid microchannels so that the solute is able to diffuse from the analyte-containing fluid phase into the extractant fluid phase. The analyte-containing fluid phase and/or the extractant fluid phase is a particle laden phase.

Description

EXTRACTION PROCESSES

Related Applications

This international patent application claims priority from Australian provisional patent application 2008904363, the specification of which is hereby incorporated by reference.

Field

The present invention relates to solvent extraction (SX) processes for extracting analytes from solution. More particularly, the present invention relates to processes for extracting analytes, such as metal ions, from solution using microfluidic devices or apparatus.

Background

Liquid-liquid extraction (referred to herein as "solvent extraction" or "SX") is a process that is commonly used for the recovery or removal of analytes or solutes from solution.

Typical solvent extraction processes involve contacting an analyte-containing aqueous phase with an organic liquid phase having an affinity for a solute and mixing the two phases to distribute small droplets of one phase in the other phase, and subsequently separating the two phases by gravity. When the phases are mixed, there is diffusive transfer of solute from the small droplets into the other phase or vice versa.

On an industrial scale, solvent extraction is usually carried out using a large volume, two stage vessel known as a mixer-settler. At the mixer stage, the two immiscible liquid phases are mutually dispersed under turbulent flow conditions so that the analyte can transfer by diffusion into the organic phase. The mutually dispersed phases then flow into the settler where they are allowed to coalesce and settle by gravity whereupon at least a portion of the analyte is dispersed in the organic phase.

Solvent extraction is used in a number of areas, including the extraction of metals from leach solutions and environmental samples, as well as in synthetic chemistry. Mineral processing, in particular, is a widely used application of solvent extraction. Many mineral processing plants utilise hydrometallurgical processes as part of an extractive metallurgical operation and solvent extraction is an important step in the recovery of economically significant metals (e.g. Cu, Ni, Au, Pt, Pd, Be and U) from ores. A typical solvent extraction process in this context entails preferentially removing a target metal or metal complex from an aqueous phase, and transferring it to an organic phase so that the metal can ultimately be recovered.

In some other solvent extraction processes one or more metal species present in an aqueous solution may be removed so that the aqueous solution itself can be re-used. Solvent extraction processes of this type are used in the remediation of contaminated soils, tannery effluent, and galvanic sludge, which contain harmful levels of heavy metals, such as chromium.

Despite the widespread use of solvent extraction processes, particularly in the mineral processing industry, there remain some difficulties. Firstly, whilst dispersion of the two phases is normally fairly rapid, the efficiency of the separation is partly determined by the rate of coalescence and, thus, phase separation may proceed slowly. This is especially true whenever surfactant or particulates are present at the liquid-liquid interfaces and are able to (partially) stabilize the droplets against coalescence. When this happens it is common for a stable emulsion ("third phase" or "crud") to persist between the two phases, thereby delaying or preventing complete phase separation.

In many cases, the stable emulsions that are formed in particle laden systems are very difficult to break. For example, in some cases, the emulsions cannot be broken using electrocoalescence which is a method that is highly effective in surfactant systems and commonly used in the petroleum industry.

Persistent dispersion of the organic phase traps the extracted species for long periods or, worse still, indefinitely, and the liquids cannot be recovered for recycling. Where the liquids or their contents are costly, e.g. in biomedical applications, or the volumes are large, e.g. industrial applications, total phase separation is crucial.

There is a need for improved solvent extraction processes particularly in relation to particle laden systems such as those found in mineral processing, metal recovery/removal process industries, chemical synthesis, and biological applications. Reference to any prior art in this specification is not, and should not be taken as, an acknowledgment or any form of suggestion that this prior art forms part of the common general knowledge in any country.

Summary

The present invention arises from studies on the microfluidic solvent extraction of complexed metal ions from aqueous solutions. Specifically, we have found that it is possible to extract metal ions from aqueous solutions on an industrial scale in an efficient manner using a microfluidic extraction device. Furthermore, we have found that microfluidic solvent extractions proceed smoothly in particle laden fluid phases for which bulk extraction processes are either inefficient or do not work at all.

The present invention provides a process for extracting a solute from an analyte- containing fluid phase, the process including: passing the analyte-containing fluid phase along a first fluid microchannel of a microfluidic extraction device; passing an extractant fluid phase that is at least partially immiscible with the analyte-containing fluid phase along a second fluid microchannel of the microfluidic extraction device; and contacting the analyte-containing fluid phase and the extractant fluid phase at a contact zone formed between the first and second fluid microchannels so that a solute is able to diffuse from the analyte-containing fluid phase into the extractant fluid phase, wherein the analyte-containing fluid phase and/or the extractant fluid phase is a particle laden phase

In some embodiments, the analyte-containing fluid phase is an aqueous phase and the extractant fluid phase is a water-immiscible non-aqueous fluid. In this context, the analyte-containing fluid phase may be a leach solution formed from leaching a crude ore in a hydrometallurgical process. Alternatively, the analyte-containing fluid phase may be a suspension of nanoparticles that may be a reagent, a catalyst, or a product and which requires the removal of one or analytes from the solution. Furthermore, the analyte-containing fluid phase may be an environmental sample, such as solution a contaminated water sample from a soil sample, a watercourse or groundwater, or an industrial effluent stream from which one or more analytes need to be extracted. In each of these cases, the analyte containing fluid phase is a particle laden phase. - A -

However, it will be appreciated that the extractant fluid phase may be a particle laden phase and the analyte-containing fluid phase may not be particle laden, or both phases may be particle laden.

In alternative embodiments, the extractant fluid phase is an aqueous phase and the analyte-containing fluid phase is a water-immiscible non-aqueous fluid. These embodiments may be suitable for applications that require a metal ion (or complex thereof) to be transferred from a solvent into an aqueous fluid, such as in stripping a metal from an organic solvent during mineral processing.

For the sake of clarity, where applicable the following description will refer to the extractant fluid phase as a water-immiscible non-aqueous fluid phase and the analyte- containing fluid phase as an aqueous phase. However, it will be appreciated that any such reference is for the purpose of clarity only and that the invention also encompasses the alternative of the extractant fluid phase being an aqueous phase and the analyte-containing fluid phase being a water-immiscible non-aqueous fluid.

In most cases, the process will also include separating the non-aqueous fluid phase and the aqueous phase after the contact zone.

The process may also include a further step of recovering the solute from the nonaqueous fluid phase.

In some embodiments, the non-aqueous fluid phase is an organic solvent. The organic solvent may be selected from the group consisting of: alkanes, alkenes, alkynes, alcohols, aldehydes, ketones, acids, esters, and aromatics and their halogen, sulfur, phosphorous, and nitrogen-containing derivatives; silicone oils and their halogen, sulfur, phosphorous, and nitrogen-containing derivatives; petroleum (all commercial grades) and petroleum-based products; and mixtures thereof.

In some embodiments, the solute is a metal ion. The metal ion may be in the form of a metal complex in which the metal ion is bound to a ligand. The metal ion may be selected from one or more ions of the group of metals consisting of: Be, Mg, Al, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Rb, Sr, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Ag, Cd, In, Sn, Sb, Te, Cs, Ba, Hf, Ta, W, Re, Os, Ir, Pt, Au, Hg, Tl, Pb, Bi, Po, At, Rn, Fr, Ra, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Er, Tm, Yb, Lu, Ac, Th, Pr, U, Np, Pu, Am, Cm, Bk, and Cf. The choice of ligand used to bind the metal ion will be dictated, at least in part, by the particular metal ion of interest. For example, extraction of Cr³⁺ ions can be carried out using acetylacetone as the ligand, whilst extraction of Cu²⁺ ions can be carried out using an oxime as the ligand. Of course other ligands could also be used, including (but not limited to): alkyl sulfides, alkyl phosphates, alkyl amines, alkyl phosphoric acids, ketoximes, aldoximes, and derivatives of any of the aforementioned.

In some embodiments, the flow rate of the non-aqueous fluid phase through the second fluid microchannel is about two times the flow rate of the aqueous solution through the first fluid microchannel. However, the flow rate of the non-aqueous fluid phase through the second fluid microchannel may be about 1000 times larger or smaller than the flow rate of the aqueous solution through the first fluid microchannel. Typical flow rates for either fluid may be between 0.001 ml_/h to 1000 ml_/h per microchannel. A ratio of the flow rate of the aqueous solution to the flow rate of the non-aqueous fluid phase may be between about 0.001 and about 250, inclusive. In a typical solvent extraction application the number of microchannels may be several million. In some embodiments, the aqueous and non-aqueous fluid phase flow rates are 0.2 to 10.0 ml_/h and 0.4 to 20.0 ml_/h, respectively.

In some embodiments, the aqueous phase and the non-aqueous fluid phase flow in the same direction though their respective fluid microchannels. In some other embodiments, the aqueous phase and the non-aqueous fluid phase flow in opposing directions relative to one another though their respective fluid microchannels.

The present invention also provides a microfluidic extraction device when used in the process of the present invention. The microfluidic extraction device may include: a substrate having first and second fluid microchannels for carrying the analyte- containing fluid phase and the extractant fluid phase respectively, the first fluid microchannel and second fluid microchannel converging at a contact zone where the analyte-containing fluid phase and the extractant fluid phase are able to contact one another so that the solute is able to diffuse from the analyte-containing fluid phase to the extractant fluid phase.

In some embodiments, the first and second fluid microchannels merge at the contact zone where the fluids contact one another for a given extraction time so that the solute is able to diffuse from the analyte-containing fluid phase to the extractant fluid phase. The first and second fluid microchannels then diverge at a Y-junction downstream from the contact zone to recover the two phases. These processing steps are carried out under continuous flow conditions. Furthermore, a range of processing parameters can be precisely controlled by adjusting flow rate alone, e.g. volumetric throughput, extraction efficiency, and extraction time.

Brief Description of the Figures

Figure 1 (a) shows a series of photographs of raw chromite samples.

Figure 1 (b) shows a schematic diagram of the ore processing for bulk and microfluidic extractions. Ground chromite samples were leached in concentrated sulfuric acid (84.6 wt%) while stirring for 8 h at 175°C, complexed with acetylacetone at pH 4 and 90⁰C for 2.5 h. Extraction was carried out (bulk or microfluidic) and samples were analysed using

UV-vis or a thermal lens microscope.

Figure 2 (a) is a schematic diagram of a microfluidic extraction device showing first (1 ) and second (2) y-junctions, separated by a serpentine channel.

Figure 2 (b) is a schematic diagram showing details of the aqueous and organic phases merging at the first y-junction (1 ), remaining in contact as coflowing streams along the length of the serpentine channel, and then diverging (phase separating) at the second y-junction (2).

Figure 3 is a plot showing a comparison of the extraction efficiency in bulk and microfluidic methods. UV-vis data was normalised by the absorbance for 500 mM metal acetylacetonate complex, determined from the linear fit for the bulk extraction data (for each metal complex). Thus, the solid line represents the bulk extraction, while the points represent the microfluidic extraction data. No significant difference between the two methods could be observed for all of the metal complexes studied. Data is also shown for selective extraction of chromium(lll) acetylacetonate from an aqueous matrix containing 500 μM iron(lll) acetylacetonate, 500 μM magnesium acetylacetonate, excess acetylacetone, and 1% ethanol. Figure 4 shows plots of the extraction efficiency, normalized to the maximum transfer measured, for chromium(lll) acetylacetonate as a function of residence time, t_R . The residence time was adjusted by changing the

Q flow rate at fixed flow rate ratio, -^- = 2 .

*~-aq Figure 5 (a) shows UV-vis absorbance spectra showing the selectivity for chromium(lll) acetylacetonate when a 0.1 M NaOH wash step is carried out. The matrix containing 500 μM iron(lll) acetylacetonate, 500 μM , excess acetylacetone, and 2% ethanol. After extraction and washing with 0.1 M NaOH, the matrix only sample showed no peaks at all, while for the chromium(lll) acetylacetonate plus matrix sample, only the chromium(lll) acetylacetonate peak remained (336 nm).

Figure 5 (b) is a plot of the extraction results for bulk and microfluidic extractions of chromite leach solutions.

Figure 6 (a) is a photograph showing the bulk extraction interface before shaking.

Figure 6 (b) is a photograph showing the formation of crud in the bulk extraction immediately after hand shaking the two liquids, droplets of the organic phase could be observed falling through the liquid and collecting at the bottom of the vessel.

Figure 6 (c) is a photograph showing that the droplets appeared to be indefinitely stable (> 42 h), inset: magnification of the particle-stabilized chloroform droplets.

Figure 6 (d) is an optical micrograph of the emulsion.

Figure 6 (e) is a photograph showing the organic phase after microfluidic extraction showing no generation of crud, no suspended particulates, and a clear interface between the organic phase and the upper water phase (used to prevent evaporation of the chloroform during collection).

Figure 7 (a) is a plot of phase separation with time after cessation of mixing, tPS, vs the relative height of the aqueous phase, h, for a bulk solvent extraction with and without silica particles present. The results for 5 g/L moderately hydrophobic particles are plotted in the inset.

Figure 7 (b) shows photographs of particle-stabilized emulsions formed in the presence of moderately hydrophobic particles.

Figure 7 (c) is a plot of the extraction efficiency of Cu²⁺, E, from aqueous phases containing hydrophilic particles with mixing (contact) time tc.

Figure 8 (a) is a schematic of a microfluidic solvent extraction device. Two liquids meet, flow parallel through the extraction channel, and separate at a Y- junction.

Figure 8 (b) is a cross-section of the extraction channel, showing the channel dimensions and the liquid-liquid interface. Flow is partly stabilized by a guide structure.

Figure 8 (c) is a photograph of a microchip mounted in an aluminium chip holder with fluid connections.

Figure 9 (a) is a plot of the bulk solvent extraction (SX) and microfluidic solvent extraction (μSX) efficiency, E, against contact time tc in the absence of particles (bulk SX and μSX) and with particles (μSX only). The transfer kinetics are unaffected by the presence of particles, irrespective of their hydrophobicity.

Figure 9 (b-f) shows photographs of bulk and microfluidic phase separation: Bulk SXs with (b) no particles and (c) 5 g/L moderately hydrophobic particles. Microfluidic SX with (d) no particles, (e) 61 g/L hydrophilic particles, and (f) 5 g/L moderately hydrophobic particles present, showing no PS emulsion. The silica particles remain in the bulk aqueous phase, indicated by the milky appearance of the aqueous phase in (e).

Figure 10 (a) shows a photograph of an SX microchip with 61 g/L hydrophilic silica nanoparticles in the aqueous phase. The organic phase is difficult to see, since it does not contain the light scattering silica particles. The Y- junctions are shown larger, right, with flow directions indicated. Figure 10 (b-c) shows photographs of the microflow for (b) 61 g/L hydrophilic and (c) 5 g/L moderately hydrophobic particles after > 7 h continuous operation.

Detailed Description

Before proceeding to describe the present invention, and embodiments thereof, in more detail it is important to note that various terms that will be used throughout the specification have meanings that will be well understood by a skilled addressee.

The term "particle laden", and variants thereof, as used throughout the specification means that the phase or fluid contains one or more objects that are small enough to be manipulated in association with the fluid, but large enough to be distinguishable from the fluid. Particles are typically nanoscopic (less than 1 μm) to microscopic (between 1 μm and 1 mm). Non-limiting examples of particle laden fluid phases include leach solutions, environmental samples, solutions containing nanoparticles, etc.

The term "microfluidic", and variants thereof, as used throughout the specification means that the device, apparatus, substrate or related apparatus in which solvent extraction takes place accommodates picolitre, nanolitre, microlitre, or millilitre fluid volumes. Accordingly, the term "microfluidic extraction", and variants thereof, means an extraction in which the volume of fluids involved in the liquid-liquid contact stage of extraction are in the picolitre, nanolitre or microlitre range. However, it will be appreciated that microfluidic extractions can be used to process large volumes (litres or more) of liquid using continuous throughput processing and/or multiple devices in parallel.

The term "immiscible", and variants thereof, as used throughout the specification means that two phases, if mixed together, will separate and not form a homogeneous mixture. It will be appreciated that two phases may be immiscible even if there is some (albeit relatively low) solubility of one phase in the other phase.

The present invention relates to microfluidic solvent extraction that is particularly suitable for use in particle laden systems. We have found that metal ligand complexes can be extracted from particle laden aqueous solutions efficiently and with selectivity for a target metal. The extractions proceed cleanly in the presence of particles which cause conventional bulk extractions to completely fail. This extraction method can be used for extraction in the presence of complex leach solutions, particulate biomaterials, and environmental samples, and also assist synthetic chemistry via particulate catalysts in microfluidic platforms.

As previously described, the present invention provides a process for extracting a solute from an analyte-containing fluid phase, the process including: passing the analyte-containing fluid phase along a first fluid microchannel of a microfluidic extraction device; - passing an extractant fluid phase that is at least partially immiscible with the analyte-containing fluid phase along a second fluid microchannel of the microfluidic extraction device; and contacting the analyte-containing fluid phase and the extractant fluid phase at a contact zone formed between the first and second fluid microchannels so that a solute is able to diffuse from the analyte-containing fluid phase into the extractant fluid phase, wherein the analyte-containing fluid phase and/or the extractant fluid phase is a particle laden phase.

Metal acetylacetonate complexes (for example, chromium acetylacetonate complexes) and copper oxime complexes were used to demonstrate the utility of the present invention. However, other metal ions and ligands could also be used.

The metal ion may be selected from one or more ions of the group of metals consisting of:, Be, Mg, Al, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Rb, Sr, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Ag, Cd, In, Sn, Sb, Te, Cs, Ba, Hf, Ta, W, Re, Os, Ir, Pt, Au, Hg, Tl, Pb, Bi, Po, At, Rn, Fr, Ra, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Er, Tm, Yb, Lu, Ac, Th, Pr, U, Np, Pu, Am, Cm, Bk, and Cf.

The ligand may be selected from the group consisting of (but not limited to): alkyl sulfides, alkyl phosphates, alkyl amines, alkyl phosphoric acids, ketoximes, aldoximes, and derivatives of any of the aforementioned.

Partitioning in the microfluidic solvent extractions described herein was equivalent to that achieved in the bulk extractions for all the complexed metal ions studied for contact times (aqueous phase) of 4 s. This represents a significant reduction in extraction time when compared to typical industrial processing times (e.g. 5 mins). This indicates that:

• microfluidic extraction of chromium(lll) acetylacetonate from chromite leach solutions gave similar partitioning to that achieved in bulk extractions; • microfluidic extractions proceeded without any interference from the silica particles, whereas phase separation in bulk extractions was no longer possible when silica particles were added to the aqueous phase to induce crud formation ;

• the absence of liquid-liquid dispersion in microfluidic extractions has the potential to streamline mineral processing (and other industrial processes) by eliminating the generation of crud.

The process of the present invention is carried out in a microfluidic extraction device. Suitable microfluidic extraction devices are shown in Figures 2(a), 8(a) and 8(c). The microfluidic extraction device includes a substrate having first and second fluid microchannels for carrying the aqueous and non-aqueous fluid phases, respectively. The first fluid microchannel and the second fluid microchannel converge at a contact zone where the aqueous and non-aqueous fluid phases are able to contact one another so that the solute is able to diffuse from the aqueous phase to the non- aqueous fluid phase, or vice versa.

As best seen in Figure 2(a), the substrate incorporates two fluid microchannels. However, it is envisaged that three or more fluid microchannels could be used.

The fluid microchannels merge at the contact zone where the fluids contact one another for a given extraction time so that the solute is able to diffuse from the first fluid to the second fluid. The channels then diverge at a downstream junction to recover the two phases. These processing steps are carried out under continuous flow conditions. Furthermore, a range of processing parameters can be precisely controlled by adjusting flow rate alone, e.g. volumetric throughput, extraction efficiency, and extraction time.

Any suitable substrates can be used. Pyrex glass microchips (Institute of

Microchemical Technology, Japan) may be suitable. Methods for forming fluid microchannel networks, typically embedded in inexpensive materials, are known in the art. For example, the microchips can be fabricated using standard photolithographic and etching procedures [for example, see Shi J., et ai., Chen Y., Applied Physics Letters 91 , 153114 (2007); or Chen Q., et ai., Journal of Microelectromechanical Systems, 16, 1 193 (2007)].

In the embodiment shown in Figure 2(a), two inlet microchannels (42 μm high, 100 μm wide) meet at a y-junction (1 ), where they merge to form the contact zone in the form of a single microchannel divided along the channel by a guide structure. At the contact zone the microchannel is wider (160 μm wide) and continues for 80 mm before diverging into two channels (as before) at a second junction (2). Thus, the flow in the microchannel is initially in separated streams (aqueous and organic) until they meet at (1 ). The two streams then flow concurrently along the wider channel, where the guide structure partially separates the two co-flowing liquids. The guide structure pins the three phase contact line at the edge created by the guide structure, which provides for stabilisation of the liquid-liquid interface and clean phase separation at the second y- junction (2).

Variations of the size, shape and/or configuration of the inlet microchannels from those described above are also envisaged. For example, the inlet microchannels may be from 1 to 1000 μm high or wide. The size of the inlet microchannels may also differ from one another in both dimensions. The merged channel (at the contact zone) can be of similar dimensions and, importantly, can continue over a distance of, for example, up to 500 mm.

An inner surface of one or more of the fluid microchannels may be modified to minimise or prevent adsorption of particles to the surface. For example, the inner surface may be modified with a chemical agent. Suitable chemical agents are known in the art and include, for example, poly(ethylene glycol), chlorosilanes, methoxysilanes, hydroxysilanes, and their amine, hydroxy, fluorine, carboxylic, derivatives, amine compounds, polyelectrolytes such as poly(methacrylic acid), poly(allylamine), poly(N- vinylpyrrolidone) etc.

Alternatively, or in addition, an inner surface of one or more of the fluid microchannels may be modified with nanostructures, such as nanoprotrusion or nanoholes. Methods for modifying microchannels are known in the art.

The flow rate of the non-aqueous fluid phase through the second fluid microchannel is about two times the flow rate of the aqueous solution through the first fluid microchannel. However, the flow rate of the non-aqueous fluid phase through the second fluid microchannel may be about 1000 times larger or smaller than the flow rate of the aqueous solution through the first fluid microchannel. Typical flow rates for either fluid may be between 0.001 ml_/h to 1000 ml_/h per microchannel. A ratio of the flow rate of the aqueous solution to the flow rate of the non-aqueous fluid phase may be between about 0.001 and about 250, inclusive. In a typical solvent extraction application the number of microchannels may be several million. In some embodiments, the aqueous and non-aqueous fluid phase flow rates are 0.2 to 10.0 ml_/h and 0.4 to 20.0 ml_/h, respectively.

Following extraction, the extraction efficiency can be determined using any suitable technique. UV-visible spectroscopy (off-line or on-line analysis) and/or thermal lens microscopy (on-line analysis) may be suitable.

Typically, the analyte-containing fluid phase is an aqueous phase containing the solute. The aqueous phase may be an ore sample (or more specifically a leach solution derived therefrom), mineral tailings, a refinery waste stream, a tannery waste stream, an aqueous soil sample, etc containing metal ions of interest. The extraction process may be used to recover the metal ions (such as in mineral processing) or to remove the metal ions from the aqueous stream so that it can be further processed (such as in remediation of soil or tannery waste streams).

The metal ion may be selected from one or more ions of the group of metals consisting of: Be, Mg, Al, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Rb, Sr, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Ag, Cd, In, Sn, Sb, Te, Cs, Ba, Hf, Ta, W, Re, Os, Ir, Pt, Au, Hg, Tl, Pb, Bi, Po, At, Rn, Fr, Ra, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Er, Tm, Yb, Lu, Ac, Th, Pr, U, Np, Pu, Am, Cm, Bk, and Cf.

Typically the metal ion will be partitioned when it is in the form of a metal complex. The metal complex may be formed by treating the aqueous solution with a ligand for the metal of interest. For example, a sample containing Cr³⁺ may be treated with a solution of acetylacetone to form an aqueous solution containing Cr(acetylacetonate)3 which is then subjected to the extraction process described herein. In another example, a sample containing Cu²⁺ may be treated with a solution of 2-hydroxy-5- nonylacetophenone oxime (LIX) to form an aqueous solution containing Cu(LIX)₂ which is then subjected to the extraction process described herein. Alternatively, the nonaqueous fluid phase may contain the ligand so that when the aqueous solution (containing the metal ion of interest) and the non-aqueous fluid phase come in to contact at the contact zone the ligand is able to diffuse into the aqueous solution or to the liquid-liquid interface where it can form a complex with the metal ion of interest. The metal complex thus formed may then diffuse in to the non-aqueous fluid phase.

Typically, the extractant fluid phase is a non-aqueous fluid phase that is at least partially immiscible with the aqueous solution. In some embodiments, the nonaqueous fluid phase is chloroform. In some other embodiments, the non-aqueous fluid phase is a hydrocarbon liquid. However, other solvents could also be used including: alkanes, alkenes, alkynes, alcohols, aldehydes, ketones, acids, esters, aromatics and their halogen, sulfur, phosphorous, and nitrogen-containing derivatives, silicone oils and their halogen, sulfur, phosphorous, and nitrogen-containing derivatives, petroleum (all commercial grades) and petroleum-based products, and mixtures thereof. The solubility of the metal or metal complex in a particular solvent may guide the choice of solvent.

The solubility of the metal ligand complex in the aqueous phase is strongly dependent on the metal ion. This enables selective separation of low solubility metal complexes, e.g. Cr and Be. Selectivity can also be achieved by selective metal ligand complex formation. The ligand might bind to only one (or a few) of the metal ions.

The process may also include a further step of recovering the solute from the nonaqueous fluid phase. For example, selectivity for Cr³⁺ from other metal ions (Fe³⁺, Mg²⁺) can be achieved by washing the extract (chloroform) with 0.1 M sodium hydroxide.

We have found that partitioning from the aqueous phase to the non-aqueous phase in the microfluidic extractions was equivalent to that achieved in bulk extractions for Cr³⁺, Fe³⁺, and Be²⁺ at typical contact residence times (aqueous phase) of 4 seconds.

In specific embodiments, a chromite ore sample is leached and the Cr³⁺ ions complexed with acetylacetone. Microfluidic extraction of Cr³⁺ from the chromite leach solutions gave equivalent partitioning to that achieved in the bulk extraction, but at reduced residence times and without liquid-liquid dispersion. Avoiding liquid-liquid dispersion using microchips may streamline industrial solvent extraction by eliminating the generation of a "third phase" (crud). In other embodiments, a particle laden aqueous phase containing Cu²⁺ ions is extracted using oximes, such as 2-hydroxy-5-nonylacetophenone oxime, in a petroleum-based organic phase. The oxime extractant is able to rapidly extract copper into the organic phase, i.e. 90% of max transfer within 60 s, and achieve spontaneous phase separation within 60 s in the absence of particles. In the equivalent bulk extraction by conventional solvent extraction, phase separation with and without silica nanoparticles is inhibited by the presence of the silica nanoparticles. In contrast, the presence of nanoparticles had a negligible influence on the rate of transfer using the method of the present invention.

Based on the above, it is clear that the methods of the present invention can be used for handling particle laden solutions. Complexation and extraction of analyte, such as Cu²⁺, proceeds without hindrance in the presence of high concentrations of hydrophilic (61 g/L) and moderately hydrophobic (5 g/L) silica nanoparticles, the latter of which causes conventional bulk extractions to completely fail. Particle-stabilized emulsions, which have a catastrophic effect on bulk solvent extraction do not form using the methods of the present invention due to the absence of liquid-liquid dispersion. This unique behaviour is suited to solvent extraction of complex leach solutions, particulate biomaterials, and environmental samples, and assist synthetic chemistry via particulate catalysts in microfluidic platforms. Furthermore, use of the methods of the present invention may lead to reduced footprints for solvent extraction unit operations and, because the microchannels are closed systems, greater potential for recycling of volatile liquids and reduced human exposure to potentially hazardous chemicals.

The processes and devices described herein can also be used for the purpose of stripping a metal ion (in the form of a complex) from an organic solvent into an aqueous phase.

Various engineering approaches, including "microchip stacking", where many microchips operate in parallel, have been used in microfluidics and can be used with the process of the present invention.

The present invention provides a continuous-flow process for solvent extraction on microfluidic chips. This work can also be extended to liquid-vapour, three-phase and counter-current systems, for which solvent extraction proceeds via diffusion at high transfer rates. The high surface-to-volume ratios required for rapid solvent extraction are achieved in the microchips using stream dimensions at the microscopic scale, rather than via droplet formation.

The microfluidic extraction device may be used in an apparatus for extracting an analyte from a sample. The apparatus may include a microfluidic extraction device as described herein and at least one flow controller for directing the aqueous phase along the first fluid microchannel and for directing the water immiscible non-aqueous fluid phase along the second fluid microchannel of the microfluidic extraction device. The flow controller may include one or more valves, flow diverters, or fluid diodes. The apparatus may also include a waste chamber at the end of the first fluid microchannel for collecting the remaining aqueous phase and a second chamber at the end of the second fluid microchannel for receiving the eluted analyte. The second chamber could be a reaction chamber formed in a separate reaction vessel coupled to the microfluidic extraction device to receive the eluted analyte for further processing or recycling through the microfluidic device.

Description of Specific Embodiments

Specific embodiments of the present invention will now be described in more detail. However, it must be appreciated that the following description is not to limit the generality of the above description.

Example 1 - Extraction of chromite leach solutions

Chemicals

Chromium(lll) acetylacetonate was synthesized by dissolving 0.22 g of OCI3 (B. D. H., England), 1.5 g urea (Ajax Chemicals, Australia) and 0.47 g of acetylacetone in 8 ml. of water. The solution was then heated in a steam bath for 1.5 h, followed by vacuum filtration and recrystallization of the crude product in cyclohexane. Water was purified using a Barnstead NANOpure Diamond purification device (Crown Scientific, Australia). Magnesium acetylacetonate dihydrate and iron(lll) acetylacetonate were purchased from Sigma-Aldrich (Germany). Acetylacetone (99.8 %) was purchased from Ajax Chemicals (Australia). Chloroform (99%), sodium hydroxide (97%) and ethanol (100%) were purchased from Chemsupply (Australia) and sulfuric acid (98%) from Scharlau Chemie SA (Spain) and used without further purification. Mineral Ores

Three chromium ore samples were examined: Raw ore specimens originating in Siberia, Russia (South Australian Museum, Australia), and Guleman, Turkey (BK Minerals, Australia) and a chromite concentrate from the Bushveld complex, South Africa (Anglo Platinum Research, South Africa). The raw ore specimens were cm- scale rock fragments, while the chromite concentrate was in the form of coarse particles, Fig. 1 (a). The composition of the major components of the three chromite samples according to XRF analysis is given in Table 1. The Cr concentration is similar in the three ore samples, while the concentration of the major components varied significantly between ore deposits. Minor components (K, Ca, S, P, Ti, and Mn) were also present in these ore samples, but at concentrations less than 1 %.

Table 1. Characterisation of chromite ore samples determined by XRF (shown as %).

Chromite Ore

Chromite ore samples were ground to a fine powder using a polished ceramic mortar and pestle. Chromite powder (200 mg) was placed in a 10 ml. test tube containing 5 ml. of concentrated sulfuric acid (84.6 wt%) and leached while stirring for 8 h at 175°C. Reported leach efficiencies for the liberation of Cr from chromite (using H₂SO₄) varies greatly, ranging from 58% to 94%. The leach efficiencies obtained in this work (Siberia, 35%; Turkey, 19%; South Africa; 53%), were relatively low; however, optimization of the leach efficiency was not the focus of this work. The green chromium sulphate leach solution was hot filtered using a glass frit, cooled, adjusted to pH ~ 4 with concentrated NaOH, and finally diluted to 100 ml_. The complexation was carried out at pH 4 to avoid precipitation of Cr(OH)₃. Aliquots (1 ml.) of the diluted chromite leach solutions were added to 1 mmol of acetylacetone and 100 μl_ of ethanol in 5 ml. of water. Depending on the experiment, the pH was readjusted using 0.1 M NaOH. These solutions were diluted to 10 mL, heated for 2.5 h at 90⁰C, and diluted to 20 ml. prior to bulk and microfluidic solvent extraction.

Comparative Example - Bulk Extraction

Bulk extractions were performed by vigorously hand-shaking a 1 :2 volume ratio of aqueous to organic (chloroform) phase for 5 min in glass vials sealed with PTFE-lined caps. The two phases were allowed to separate before removing samples from the aqueous and organic phases with a Pasteur pipette for UV-vis analysis.

Example - Microfluidic Extraction

Microfluidic extractions were carried out in Pyrex glass microchips (Institute of Microchemical Technology, Japan). A schematic of the microchip design is shown in

Fig. 2(a). Two inlet channels (42 μm high, 100 μm wide) meet at a y-junction (1 ), where they merge to form a single microchannel divided along the channel by a guide structure. Here the microchannel is wider (160 μm wide) and continues for 80 mm before diverging into two channels (as before) at a second y-junction (2). Thus, the flow in the microchannel is initially in separated streams (aqueous and organic) until they meet at (1 ). The two streams then flow concurrently along the wider channel, where the guide structure partially separates the two co-flowing liquids. The guide structure pins the three phase contact line at the edge created by the guide structure, which is crucial for stabilisation of the liquid-liquid interface and clean phase separation at the second y-junction (2).

Fluid connections to the microchip were made using an aluminium chip holder, with screw-in connectors and PEEK tubing. PEEK tubing was connected to 1 or 2 mL Luer tip gastight syringes (Hamilton Company, USA) using Teflon connectors (IMT, Japan). The syringes were driven by programmable syringe pumps (KDS210P, KD Scientific, USA). The aqueous and organic phase flow rates were 0.2 mL h^"1 and 0.4 mL h^"1, respectively, unless otherwise stated. This flow rate ratio balances the effect of the different fluid viscosities, which determine the position of the liquid-liquid interface. For experiments where different flow rates were used (residence time experiments), the flow rate was adjusted stepwise while maintaining the flow rate ratio. To obtain clean phase separation at the second y-junction the flow resistance in each of the outlet channels/tubing must be equal. The flow resistance was adjusted in the outlet tubing by means of a finely-threaded hose clamp. After stabilizing concurrent flow and phase separation, the organic phase was collected in 1.5 ml. glass vials with PTFE sealed septa (Cole-Parmer, USA). When the organic phase was washed with NaOH, the sample was directly collected in larger vials (4 ml.) containing 3 ml. of 0.1 M NaOH. Microfluidic extractions were monitored using an optical microscope (Model BH-2, Olympus Australia, Australia) and a Moticam 2000 digital camera (Motic, China). All extractions were carried out at ambient temperature (~ 20⁰C).

After solvent extraction, the two phases were separated and the organic phase was analysed by UV-Vis spectroscopy (Cary 1 E Photospectrometer, Varian Australia,

Australia). A quartz microcuvette (10 mm pathlength, 580 μl_ volume, Starna, Australia) was used to minimize the sample volumes and therefore collection times. The cuvette was closed using a Teflon stopper to avoid evaporation of the chloroform solvent. All spectra were baseline corrected for the solvent and collected at 1 nm wavelength resolution from 200-800 nm.

The composition of the chromite ore samples was determined using x-ray fluorescence spectrometry (XRF). Total Cr concentrations in the aqueous leach solutions (before and after extraction) were determined using inductively coupled plasma atomic emission spectroscopy (ICP-AES).

The characteristic UV-vis spectra for the metal acetylacetonate, IVT⁺A_n, complexes were determined. The following absorbance peaks were identified: acetylacetone, HA, 274 nm; BeA₂, 295 nm; CrA₃, 336 nm; FeA₃, 437 nm; MgA₂, 296 nm; NiA₂, 285 nm; AIA₃, 288 nm. Beer's law was obeyed for all of the metal acetylacetonate complexes over the relevant concentration range (5 - 500 μM). This calibration data was used for quantitative UV-vis analysis of the organic and aqueous phases post extraction. Initial bulk (bench-scale) extraction experiments revealed no transfer of the water soluble MgA₂ and NiA₂ complexes, some transfer of FeA₃ and AIA₃, and complete transfer of the water insoluble BeA₂ and CrA₃ to the organic phase (chloroform), within our detection limits.

For the metal acetylacetonate complexes which gave the highest transfer efficiencies, i.e. BeA₂, FeA₃, and CrA₃, we compared the partitioning achieved using the bulk and microfluidic extractions. The extractions were carried out as described earlier using a series of aqueous standard solutions (5 - 500 μM) of the complexes. Ethanol (1 %) was added to avoid crystallization of the complexes in the aqueous phase. The contact (extraction) times for the bulk and microfluidic extractions were 5 min and ~ 3.4 s, respectively. The latter is based on the residence time of the aqueous solution between positions (1 ) and (2) along the microchannel (see Figure 2) for an aqueous flow rate of 0.2 ml. h^"1. The residence time of the organic phase is reduced by a factor of two, ~ 1.7 s, due to the 2:1 flow rate ratio. After each experiment, the organic phase was collected and, in the case of BeA₂, the organic phase was washed with 0.1 M NaOH to remove excess acetylacetone (due to interference from the HA absorbance peak). The organic phase was then analysed by UV-vis spectrophotometry or Thermal Lens Microscopy. Figure 3 compares the UV-vis results from the two extraction methods. The data are normalised according to the absorbance measured in the organic phase after bulk extraction of the 500 μM aqueous solution for each complex. This value was determined from the linear fit of the bulk extraction data. Thus, the solid line represents the normalised data for the bulk extractions, while the data points are the results obtained for the microfluidic extractions. We observed excellent agreement between the partitioning observed in the bulk and microfluidic extractions, despite the microfluidic extraction residence time being 100 times less than that used for the bulk extraction.

Based upon Beer's law, the concentration of complex in the organic phase was calculated and thus the distribution coefficient:

(1 )

where \x\_rg and \x\_q are the equilibrium concentrations of the metal complex, X. The water insoluble complexes, i.e. CrA₃ and BeA₂, gave distribution coefficients of ∞, within detection limits.

The extraction efficiency depends also on the duration of liquid-liquid contact compared to the rate of diffusion for the metal acetylacetonate species. For microfluidic applications, e.g. mineral processing or effluent treatment, relatively high flow velocities are required to achieve sufficient volumetric throughput (even where microchip stacking is employed). In the microfluidic chip, the contact time is defined by the flow velocity and the channel length between positions (1 ) and (2) (see Figure 2). Thus, microfluidic throughput will necessarily be limited by the minimum residence time at which sufficient partitioning is achieved. To study the effect of residence time on the partitioning, the flow rate of the aqueous phase was increased from 0.2 to 10 ml. h^"1 at constant flow

Q rate ratio, -^- =2, corresponding to aqueous phase residence times from ~ 0.07 s to

3.40 s. Figure 4(a) shows the dependence of the extraction efficiency, normalized to the maximum transfer, on the residence time for the CrA₃ complexes. As one intuitively expects, infinitely short residence times (t_R → O ) would result in no transfer at all, while sufficiently long residence times (t_R ≥ 4 s) result in complete partitioning.

For industrial applications, complete partitioning in a single extraction step is usually not required. Instead, 2 or 3 extractions may be carried out in series to ensure complete recovery of the valuable species. Therefore, the efficiency of each extraction step in a microfluidic extraction unit operation may be reduced in a multistage separation to maximise throughput. The extraction efficiency for an n stage extraction process for a given distribution coefficient can be calculated using the following equation:

\_xγ where ^{L Jα«} is the concentration in the aqueous phase after n extraction stages [Coulson J. M., et al., Particle Technology and Separation Processes, 4th ed. Chemical Engineering, Vol. 2. 1996, Oxford: Butterworth-Heinemann]. Typically, 3 stage extraction processes are used, thus residence times of the order of 0.3 s are sufficient to recover at efficiencies comparable to typical transfer efficiencies in industry. The stated residence time corresponds to flow rates of 3.2 ml. h^"1 through a single microchannel.

Clearly, throughput is an important consideration for industrial processing. Typical solvent extraction throughputs may be up to 10,000 m³ h^"1. To address this consideration, it is possible to stack the microchips in a block, with each block of microchips containing many millions of microchannels to achieve the required throughput. Selectivity in the separation of valuable solutes is directly related to the grade, and thus the commercial value, of the end product. A typical feed solution for an industrial solvent extraction is a complex mixture of components liberated from raw materials and upstream processing. The selective separation of Cr(III) is particularly relevant to remediation of process waste (in both solid and liquid form). Chromium is recovered on a commercial scale from chromite, FeCr₂O₄, which may also contain various amounts of magnesium, aluminium, and silica, depending on the particular ore deposit, see Table 1. We have used both model and real chromite leach solutions as feed materials for bulk and microfluidic extractions. For the real ore samples, we have liberated the Cr by acid leaching and complexed the Cr³⁺ with acetylacetone, HA. The resulting solutions were then extracted using both bulk and microfluidic methods and the results compared.

The model leach solutions contained 5-500 μM CrA₃, 500 μM of both FeA₃ and MgA₂, 5 mM HA (excess extractant) and 1% ethanol (to avoid precipitation of CrA₃). The extractions were carried out as previously described; however, now with a post extraction washing step (using 0.1 M NaOH) to selectively remove any of the matrix species present the organic phase (see Figure 5(a)). This washing step does not remove water insoluble complexes, i.e. CrA₃, and therefore chromium can be selectively recovered from the model leach solution. The extraction data for the microfluidic solvent extraction of chromium from the model matrix is shown in Figure 3, where they are compared with the earlier results (both bulk and microfluidic extractions).

Using real ore samples (sourced from Siberia, Turkey, and South Africa), we carried out a complete leach-complex-extraction procedure to determine whether other components in chromite ore, not included in our model leach solutions, would interfere with the efficiency of the microfluidic solvent extraction. The leach and complexation procedures are as described above. The efficiencies of the leach and complexation procedures were determined using ICP-AES and XRF. Despite the chemical complexity of the chromite ore (see Table 1 ), and thus of the leach solutions, the microfluidic extractions were carried out without difficulty. Notably, the microfluidic extractions of the chromite leach solutions consistently gave slightly higher transfer efficiencies than those achieved in the bulk extractions (Figure 5(b)). The reason for this is unclear; however, we can confirm that loss of solvent to evaporation (which would increase the sample concentration) was negligible, because the organic phase was collected in a sealed vial containing the NaOH_(aq) wash solution. The upper aqueous phase and sealed cap prevented evaporation of the lower chloroform phase.

In order to determine the performance of the microfluidic extractions under conditions that lead to crud formation, we introduced partially hydrophobized silica particles (12 nm, 0.5 g/L) to the aqueous phase prior to extraction. Figures 7(a)-(d) show the result of a bulk extraction carried out under these conditions. The organic phase is dispersed as droplets on which the silica particles readily adsorb. These adsorbed particles stabilize the droplets against coalescence. The droplets remain loosely packed at the bottom of the vessel (as the organic phase is denser than the aqueous phase) and are indefinitely stable (image shown for 42 h). This stable layer of crud (or "third phase") did not permit us to carry out UV-vis analysis, as phase separation is not possible. In contrast, the microfluidic extraction proceeded as per our earlier experiments, with the exception that the three-phase contact line appears to pin more readily. This results in a slight distortion of the liquid-liquid interface; however, there was no adverse effect on the phase separation at the second y-junction. Figure 7(e) shows the optically clear organic phase post microfluidic extraction.

Example 2 - Extraction of Cu2+ from particle laden solutions

Aqueous copper solutions (5.1 g/L) were prepared using AR grade CuSO₄-5H₂O (Chem-Supply) and pure water (18.2 MΩ-cm, Barnstead), adjusted to pH 1.5 using H₂SO₄. Cu²⁺ was extracted using 15%v/v LIX84-IC (2-hydroxy-5-nonylacetophenone oxime, Cognis) in Shellsol 2046 (Shell Chemicals). Two types of nanoparticles were added to the aqueous phase: (i) Hydrophilic silica nanoparticles, SNOWTEX-ZL (Nissan Chemical Industries), as a 40 wt % dispersion in water at pH 9 (θ water-air< 10°, primary size 80 - 100 nm) and (ii) moderately hydrophobic silica nanoparticles (Aerosil® R816, Degussa) with primary size of 12 nm and θ water-air = 23°. Moderately hydrophobic silica was dispersed in the aqueous phase by sonicating for 30 min. Moderately hydrophobic and hydrophilic silica particles formed 200-300 nm and 140-180 nm aggregates within 30 min of dispersion, according to light scattering measurements (Malvern Nano Zetasizer).

Bulk extractions were carried out in a 200 mL round bottom flask stirred at 1100 rpm. Within two seconds, 67 ml of 5.1 g/l CuSO₄ solution was added to 43.3 ml of the stirred organic phase (15% v/v LIX84-IC in Shellsol 2046). The volumetric ratio R = 0.65 was fixed near to the viscosity ratio, ~ 0.62, for the most stable microfluidic flow conditions. One second before the sampling time tc, 5 ml of emulsion was removed using a large bore sampling tube and transferred to a glass vial. After phase separation, the aqueous phase was analysed by UV-vis spectroscopy. Samples were collected from 5 to 300 s. The rate of phase separation was determined only for tc = 300 s. Sampling for phase separation experiments was carried out as above, except aliquots were transferred to a 10 ml measuring cylinder for analysis. The height of the aqueous phase was monitored until phase separation was complete or the maximum height of the aqueous phase was reached.

Microfluidic extractions were carried out in Pyrex™ microchips (Institute of Microchemical Technology, Japan). Two microchannels (100 μm x 40 μm) merge at a Y-junction to form a single microchannel (160 μm x 40 μm) that is divided into two by a guide structure where the extraction takes place. The length of the channel varied from 80 mm to 480 mm. Phase separation occurred at a second Y-junction downstream. Liquid flow was driven by a precision syringe pump (KDS210P, KD Scientific) fitted with Hamilton gas tight syringes. The flow rate of the aqueous phase ranged from 0.1 ml/h to 8 ml/h at a fixed organic/aqueous flow rate ratio of 0.65. For experiments with particles present, the aqueous phase was spun at 22 000 rpm in a centrifuge (Hermle Labortechnik, Z36HK) for 30 min after extraction.

UV-vis absorption (Ocean Optics QE65000) was used to determine the concentration of Cu²⁺ or Cu-complex. A Z-Flow Cell (2.5 or 10 mm path length, quartz windows) was directly connected to the outlet of the microchip using PEEK tubing. The tubing contained a T-junction (Upchurch Scientific) to flush the flow cell with solvent to check the baseline of the spectra. Absorbance measurements at 750 nm (i.e. at the shoulder of the 796 nm peak) and 656 nm for the Cu²⁺ and Cu(LIX)₂ complex, respectively, were used according to Beer's law. Flow stability in the microchip was monitored using optical microscopy (Olympus, Model BH2-UMA).

Extraction of copper is achieved in industry using oximes, e.g. 2-hydroxy-5- nonylacetophenone oxime (LIX84-IC), in a petroleum-based organic phase, e.g. Shellsol 2046. The chosen extractant (LIX84-IC) is able to rapidly extract copper into the organic phase, i.e. 90% of max transfer within 60 s, and achieve spontaneous phase separation within 60 s in the absence of particles. Figure 7(a) shows the progress of phase separation for conventional bulk extractions with and without silica nanoparticles, hydrophilic (Qwater-air < 10°, 80 nm) and moderately hydrophobic (θ water-air = 23°, 12 nm). The relative height of the aqueous phase, h, is plotted against time after mixing cessation, tPS. Phase separation is clearly inhibited by the presence of silica nanoparticles. Hydrophilic particles delayed phase separation from only 30 s (particle free solution) to ~ 300 s (61 g/L hydrophilic particles). When moderately hydrophobic particles were added (5 g/L), phase separation was totally arrested at 60% separation (stable for > 6 months), and was two orders of magnitude slower. Optical microscopy revealed a PS emulsion of decreasing droplet size with increasing silica particle concentration, as expected from classical studies, cf. Figure 7(b).

Bulk extraction kinetics were determined using a standard procedure (MCT Redbook: Solvent Extraction Reagents and Applications. 2007: Cognis Group). In brief, the two liquids were dispersed and aliquots of emulsion were removed at various extraction times. The emulsion was allowed to phase separate before UV-vis analysis of the aqueous phase. Longer phase separation times in the presence of hydrophilic particles contributed to a slight increase in the measured transfer efficiency E, cf. Figure 1 (c). Thus, the presence of the hydrophilic particles did not slow down the transfer kinetics. In contrast, the transfer kinetics for moderately hydrophobic particles could not be followed at all due to the inability to rapidly and completely separate the liquids after dispersion.

In this study, oil-in-water (O/W) emulsions formed irrespective of the type of particles present. This is consistent with PS emulsion theory, which predicts that adsorbed particles will prefer to be wet by the continuous phase (all particles were relatively hydrophilic, θ water-air < 30° ). For SX of solute from aqueous to organic phase, O/W emulsions are particularly detrimental due to the entrapment of both the extracted species and the organic phase (which is usually recycled) in the droplets. Furthermore, the effect of silica particles is catastrophic in industrial phase separations, e.g. mineral processing.

The microchip used in this study is shown in Figure 8. The Pyrex™ chip was fabricated using photolithography, followed by wet-etching and thermal bonding. The aqueous and organic phases converge at a Y-junction and flow parallel for a given contact time, tc, which is determined by the length of the main channel, L, and the flow rate of the liquids. The optimal flow rate ratio (organic/aqueous), R = 0.65, was fixed by the viscosity ratio (the cross-section of the aqueous and organic streams similar), and flow was laminar (Re < 50). The cross-section of the extraction channel includes a guide structure which helps to pin the three-phase contact line and, therefore, maintains the position of the liquid-liquid interface, Figure 8(b). The extraction channel terminates at a second Y-junction at which the two phases are separated and flow out of the device for online UV-vis analysis.

Solvent extraction requires high surface-to-volume ratios {S/V) to enhance the rate of transfer. In bulk solvent extraction, this is achieved by dispersing the liquids as small

(sub-millimetre) droplets. In micro solvent extraction, the liquids are contacted in the form of microscopic streams. For the work described herein, we have estimated S/V for bulk solvent extraction (based on 200 μm droplet radius) and micro solvent extraction.

The ratios were similar (S/V ~ 15000 m-1 ), which permits a direct comparison of the transfer kinetics.

Figure 9(a) shows the transfer kinetics for the bulk and microfluidic methods. Very little difference was detected between the rate of extraction, as one might expect for the similar S/V and R ratios. Fitting the second order rate equation to these data gives good agreement, yielding a rate constant, k, of 0.1 1 s-1 (units reflect the normalized concentration). The presence of nanoparticles had a negligible influence on the rate of transfer, which was consistent with our bulk experiments with hydrophilic particles, cf. Figure 7(c). This was expected for hydrophilic particles, as they are not readily adsorbed at the liquid-liquid interface (they are preferentially wet by water). In contrast, we know that moderately hydrophobic particles adsorb at the liquid-liquid interface during bulk extractions (we could not determine E(tc) previously due to slow and incomplete phase separation, cf. Figure 7). In contrast, the extraction kinetics in the presence of moderately hydrophobic particles was accessible from our microfluidic extractions, for which the nanoparticles did not interfere with the phase separation. The rate of extraction in the presence of moderately hydrophobic particles was comparable to that for the particle-free experiments, suggesting that transfer through the interface was uninhibited by the particles present at the interface.

As the nanoparticles flow through the microchip, particles may be transported from the bulk aqueous phase to the liquid-liquid interface, solid-liquid interface (microchannel walls), or the bulk organic phase. Therefore, with the addition of nanoparticles, one could intuitively expect major problems with fouling of the channels or disruption of the flow stability. However, after initializing flow, no particle-induced disturbances were observed, regardless of the hydrophobicity or concentration of particles used. Silica nanoparticles can be observed (courtesy of scattered light) flowing in the aqueous stream throughout the microchip in Figure 10(a), while the light scattered from the organic phase is minimal. Furthermore, the silica particles are recovered in the aqueous phase upon exiting the microchip, rather than in the organic phase. Thus, the behaviour of the particles in the microchip differs greatly from that in the bulk, where the moderately hydrophobic particles were adsorbed at the liquid-liquid interface and are taken up into the emulsion (crud) leaving a transparent aqueous phase below.

Flow stability was tested for endurance under the most challenging conditions studied; 61 g/L hydrophilic and 5 g/L moderately hydrophobic silica, Figure 10(b-c). The endurance test revealed negligible particle adsorption on the microchannel walls and no detrimental effect on phase separation for > 7 h continuous extraction. Similarly, phase separation was unhindered by the presence of particles in the aqueous phase. Compared with the formation of indefinitely stable PS emulsions in our bulk SXs, cf. Figure 7(b), the microfluidic extraction method significantly outperforms bulk methods for particle laden systems.

The examples provided herein indicate that microfluidic solvent extraction can be used for particle laden solutions. Complexation and extraction of analyte proceeds without hindrance in the presence of high concentrations of hydrophilic (61 g/L), and moderately hydrophobic (5 g/L) silica nanoparticles, the latter of which causes conventional bulk extractions to completely fail. Particle-stabilized emulsions, which have a catastrophic effect on industrial bulk solvent extraction, do not form in microfluidic solvent extraction due to the absence of liquid-liquid dispersion.

Finally, it will be appreciated that various modifications and variations of the methods and compositions of the invention described herein will be apparent to those skilled in the art without departing from the scope and spirit of the invention. Although the invention has been described in connection with specific preferred embodiments, it should be understood that the invention as claimed should not be unduly limited to such specific embodiments. Indeed, various modifications of the described modes for carrying out the invention that are apparent to those skilled in the art are intended to be within the scope of the present invention.

Claims

1. A process for extracting a solute from an analyte-containing fluid phase, the process including: - passing the analyte-containing fluid phase along a first fluid microchannel of a microfluidic extraction device; passing an extractant fluid phase that is at least partially immiscible with the analyte-containing fluid phase along a second fluid microchannel of the microfluidic extraction device; and - contacting the analyte-containing fluid phase and the extractant fluid phase at a contact zone formed between the first and second fluid microchannels so that the solute is able to diffuse from the analyte-containing fluid phase into the extractant fluid phase; wherein the analyte-containing fluid phase and/or the extractant fluid phase is a particle laden phase.

2. A process according to claim 1 , further including separating the analyte- containing fluid phase and the extractant fluid phase after the contact zone.

3. A process according to either claim 1 or claim 2, further including recovering the solute from the extractant fluid phase.

4. A process according to any one of claims 1 to 3, wherein the analyte-containing fluid phase is an aqueous phase and the extractant fluid phase is a water-immiscible non-aqueous fluid.

5. A process according to any one of claims 1 to 3, wherein the extractant fluid phase is an aqueous phase and the analyte-containing fluid phase is a water- immiscible non-aqueous fluid.

6. A process according to any one of claims 1 to 5, wherein the non-aqueous fluid phase is an organic solvent.

7. A process according to claim 6, wherein the organic solvent is selected from the group consisting of: alkanes, alkenes, alkynes, alcohols, aldehydes, ketones, acids, esters, and aromatics, and their halogen, sulfur, phosphorous, and nitrogen-containing derivatives; silicone oils and their halogen, sulfur, phosphorous, and nitrogen- containing derivatives; petroleum and petroleum-based products; and mixtures thereof.

8. A process according to any one of claims 1 to 7, wherein the solute is a metal ion.

9. A process according to claim 8, wherein the metal ion is selected from one or more ion of the group of metals consisting of: Be, Mg, Al, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Rb, Sr, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Ag, Cd, In, Sn, Sb, Te, Cs, Ba, Hf, Ta, W, Re, Os, Ir, Pt, Au, Hg, Tl, Pb, Bi, Po, At, Rn, Fr, Ra, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Er, Tm, Yb, Lu, Ac, Th, Pr, U, Np, Pu, Am, Cm, Bk, and Cf.

10. A process according to either claim 8 or claim 9, wherein the metal ion is in the form of a metal complex in which the metal ion is bound to a ligand.

1 1. A process according to claim 10, wherein the ligand is selected from the group consisting of (but not limited to): alkyl sulfides, alkyl phosphates, alkyl amines, alkyl phosphoric acids, ketoximes, aldoximes, and derivatives of any of the aforementioned.

12. A process according to claim 1 1 , wherein the metal is Cr and the ligand is acetylacetone.

13. A process according to claim 1 1 , wherein the metal is Cu and the ligand is an oxime.

14. A process according to any one of claims 1 to 13, wherein the analyte- containing fluid phase is selected from the group consisting of: a leach solution, a contaminated water sample, an industrial effluent stream, and a suspension of nanoparticles.

15. A process according to any one of claims 1 to 14, wherein the flow rate of the aqueous and/or the non-aqueous fluid phase is between about 0.001 mL/h and about 1000 mL/h, inclusive, per microchannel

16. A process according to any one of claims 1 to 15, wherein a ratio of the flow rate of the analyte-containing fluid phase to the flow rate of the extractant fluid phase is between about 0.001 and about 250, inclusive.

17. A microfluidic extraction device when used in the process of any one of claims 1 to 16, the device including: a substrate having first and second fluid microchannels for carrying an analyte-containing fluid phase and an extractant fluid phase respectively, the first fluid microchannel and second fluid microchannel converging at a contact zone where the analyte-containing fluid phase and the extractant fluid phase are able to contact one another so that a solute is able to diffuse from the analyte-containing fluid phase to the extractant fluid phase.

18. A microfluidic extraction device according to claim 17, wherein an inner surface of one or more of the fluid microchannels is modified to prevent adsorption of the particles.

19. A microfluidic extraction device according to claim 18, wherein the inner surface is modified with a chemical agent.

20. A microfluidic extraction device according to any one of claims 17 to 19, wherein an inner surface of one or more of the fluid microchannels is modified with nanostructures.

21. A microfluidic extraction device according to claim 20, wherein the nanostructures are nanoprotrusions or nanoholes.

22. A process according to claim 1 and substantially as hereinbefore described with respect to any one or more of the accompanying examples.