WO2007112491A1 - Microreactor - Google Patents

Abstract

The invention is directed to a porous support for conducting one or more chemical reactions comprising a polymeric or silica porous monolith functionalised with an agent. In one embodiment, the agent is a catalytic or scavenging agent. The invention also relates to microreactors comprising such porous supports for use in conducting chemical reactions including organic chemical reactions, metal scavenging, diagnostic reactions or catalysis reactions. The novel porous supports described herein find particular application in small scale high throughput screening.

Description

"Microreactor"

Cross-Reference to Related Applications

The present application claims priority from Australian Provisional Patent Application No 2006901698 filed on 31 March 2006, the content of which is incorporated herein by reference.

Introduction to the Invention

This invention relates to microreactors comprising porous monolithic supports suitable for catalysis and scavenging. In particular, the invention concerns microreactors comprising monolithic polymers prepared by in situ initiated polymerisation of monomers within a capillary or microchannel(s) of a microreactor or scavenging device. The in-situ synthesis of the polymer support provides a simple but unprecedented means to position the reaction support anywhere in the device. Multiple microreactors can be placed in parallel for combinatorial chemistry and combined with microscale bioassays for fully automated, small scale high throughput screening.

This invention also relates to microfluidic devices, agents and supports for use in chemical syntheses and diagnostics where high surface-to-volume ratio supports allow a significant reduction in the consumption of samples and/or reagents used in such reactions.

Background to the Invention

Classical synthetic chemistry has not changed much over the last 100 years, with reactions in general being performed on macroscale in round bottomed flasks. Despite the miniaturisation of chemical analysis having been researched since the early 1990's, only a handful of research groups have investigated the potential of the new, chip-based systems for synthetic chemistry (Feng et al. 2004, Pennemann et al. 2004, Watts & Haswell 2004). The Lab-on-a-Chip concept, however, has the potential to revolutionise synthetic chemistry by changing conventional glassware to microreactors ranging in size from small tubular microreactors to true chip based micro-scale reactors. The small scale of the microreactors results in reduced diffusion times, and therefore improved reaction kinetics. Additionally, the small thermal mass of the microreactors allows rapid transport of heat into or out of a microreactor with little change in local temperature. This allows even highly exothermic reactions to be performed under isothermal conditions. Microreactors can be used as tools for small scale organic synthesis in the conventional laboratory to assess the viability of reactions prior to proceeding to large scale reactions. In the production stage, microreactors could be scaled up to HPLC column size (i.e. 2 - 5 mm internal diameter) for synthesis of larger quantities of compounds (Phan et al. 2004). Microreactors could also be used for producing libraries of compounds for drug discovery. The small amount of compounds synthesised in the microreactor is more than sufficient for preliminary drug screenings, as only a very small amount of highly potent compound is required for initial testing. A microreactor could, for example, be combined with a microfluidic screening device to form a true lab on a chip for drug discovery.

In synthetic organic chemistry, new compounds are made by the formation of new bonds and/or destruction of old bonds. Catalysts might be required to bring together reactive intermediates of simple organic building blocks (hydrogen, alkenes and many others) to form bonds with each other and thereby new products. Catalysts can be heterogeneous, or homogeneous. Homogenous catalysts are in the same phase (e.g. a dissolved catalyst in a liquid reaction mixture), whereas heterogeneous catalysts are present in different phases from the reactants (e.g. a solid catalyst in a liquid reaction mixture). Homogeneous catalysts allow for much greater control of the catalytic reaction, in particular reactant selectivity and stereochemistry of the product, whereas heterogeneous catalysts are easier to be reclaimed after synthesis.

The use of solid supported reagents and scavengers have been increasingly popular in organic chemistry as their use bypasses the purification issues associated with traditional solution-phase reactions (Thomas et al. 2005). hi addition to the easy work-up, attachment of hazardous catalysts or reagents to a solid support reduces the risks of working with the reagent.

Perhaps the most important insoluble support for organic synthesis is cross- linked polystyrene, where derivatised forms are made by copolymerisation of styrene, divinylbenzene and functionalised styrene, or by functionalising a polystyrene starting material (Kwok et al. 2003). Merrifield resin is a polystyrene resin based on a copolymer of styrene and chloromethylstyrene. The benzylic chloride group can be used for attachment of generally any ligand containing an alcohol, phenol, thiol or amine function. Subsequently, the ligand can be complexed with a transition metal to form a organometallic catalyst. Merrifield resins have been used extensively for organic synthesis (Fenger & Le Drian 1998).

In a microreactor for heterogeneous catalysis, the catalysts have to be loaded into and retained within the microchannel. Merrifield resin beads modified with nickel(II) have been packed into glass HPLC columns to form a flow-through microreactor, showing marked improvements in reaction rate over batch methods (Bremeyer et al. 2002, Haswell et al. 2001, Lee et al. 2005, Ley et al. 2002, Phan et al. 2004, Ramarao et al. 2002). When a Kumada reaction was performed using nickel(II) on Merrifield resins in a glass capillary microreactor, an increase in reaction rate of approximately 3400 times was achieved, compared to a traditional batch reaction

(Haswell et al. 2001). Basheer et al. demonstrated Suzuki coupling reactions in capillary-microreactors filled with palladium nanoparticles (Basheer et al. 2004).

Packing beads or particles into narrow bore capillaries or microchannels and retaining them, however, is not straightforward.

In separation science, monoliths have started to replace bead-based systems in areas including chromatography, capillary electrochromatography, and solid phase extraction (Svec & Frechet 2003). Macroporous monolithic polymers are sponge-like three-dimensional support structures that possess a large surface area. These materials are highly porous and are well suited for flow-through applications as a result of their low flow resistance (Svec & Frechet 1992). The main benefit when compared with beads is that monoliths can be synthesised in situ inside the capillary or microchannel, thereby avoiding the packing problem. Other benefits of monoliths include the highly configurable surface chemistry, the configurable porosity, the large surface area and low flow resistance.

To date, monoliths have been used mainly as separation media in liquid chromatography and capillary electrochromatography and solid phase extraction (Hilder et al. 2004, Peterson 2005, Stachowiak et al. 2004). A different application for these modified monolith microreactors is in scavenging of excess reagents, such as reported for reactive amines by Tripp et al.(2000).

The invention described here concerns the use of microreactors containing functionalised porous polymer monoliths for solid-supported catalysis or metal scavenging. This way, advantages of a microreactor can be exploited without the need for packing and retention of beads or nanoparticles inside the microchannel or capillary.

The miniaturisation of chemical reactions for the purpose of synthesis, analysis and diagnosis has many advantages including reduced use of sample and reagent quantities.

The trend towards miniaturisation has now entered an area where chemistry meets microfabrication. Miniaturisation down to the chip level started with the introduction of the miniaturised total analysis system for separation science.

Separations were demonstrated to be faster and more efficient with a significant reduction in the consumption of sample and reagents. The analysis inside the microdevice could be performed in an automated way, and multiple analyses were done in parallel using a multi-channel microdevice. This advance was broadened to the 'Lab- on-a-Chip' concept, a more universal term for chemistry on a microdevice. Downscaling of organic synthesis to the microchip format offers advantages similar to those mentioned for analytical devices. The possibility of doing multiple reactions in parallel in a fast and automated way is especially attractive for combinatorial chemistry, where large libraries need to be synthesised and tested.

Accordingly, miniaturisation of chemical reactors using micro fluidic devices offers fundamental and practical advantages and has the potential to revolutionise drug discovery in the pharmaceutical industry. In combinatorial chemistry, for example, high throughput techniques are required for the synthesis and testing of large libraries of compounds. Transfer of the existing methodology to the microfluidic format would provide the option of running multiple syntheses in parallel in an automated way, with a dramatic decrease in reagent consumption and waste generation.

High throughput synthesis, however, is not the only driving force behind the rapidly increasing interest in synthesis on a chip. Changing the dimensions of a reactor to microchannels drastically changes physical parameters. Typical channels inside a microreactor are between 1 and 10 cm long with a depth and width between 10 and 200 μm. When the microreactor is a chip based device the channel length is preferably between 1 to 100 cm long and 10 μm to 1 cm wide. When the microreactor is a capillary/tube/cartridge based device the channel length is preferably between 1 to 100cm long and 10 μm to 10 cm wide. The control over the reaction is improved by the small thermal mass of microdevices, enabling accurate and fast temperature control. Additionally, reactions on the microscale result in cleaner (green chemistry) and more environmentally-friendly chemistry without compromising on reagents that can be used. A further advantage is that some chemicals that are considered dangerous for laboratory-scale synthesis involve significantly lower risk when used in small amounts in microreactors. Even an explosion of the complete reactor would not cause damage in the laboratory, but only require replacement of the microdevice.

Microfluidic reactors allow improved control in dimensions that are closer to the molecular level, thereby enhancing the study of complex chemical processes. Synthesis on a chip results in relatively pure compounds formed in higher yields and in shorter time than in equivalent bulk reactions. The quantities of the synthesised product are generally sufficient for characterisation using conventional analytical instrumentation (MS), and for screening in drug discovery. Traditionally, organic synthesis in a microreactor or on a chip occurs inside a microchannel network inside a solid substrate. This substrate might consist of any one or a combination of quartz; glass including borofloat, pyrex, borosilicate, fused silica, soda lime, diamond; polymers including PMMA, COC, polycarbonate, PET, TEFLON, TEFZEL, PEEK; metal including nickel, stainless steel, copper, silver, iron, gold, platinum and alloys; ceramics and silicon and silicon-related materials including silicon carbide, silicon nitride, silicon oxide. Typically, the devices consist of two plates bound together, with one of them containing the channels. The channels, typically 10-200 μm wide and deep with a length of 1 to 10 cm, are sealed between the two plates, and connected to the outside world via reservoirs.

To use such microdevices, fluid flows are required. Initially, fluid flows in microfluidic devices were driven by the electro-osmotic flow (EOF). The EOF is a result of the movement of a charged layer in a channel/capillary, and depends strongly on parameters including surface charge, pH, temperature and viscosity. Later, pressure- driven systems were introduced to guarantee a stable and uniform flow independent of surface chemistry. However, in order for such microdevices to function effectively the reaction surface available must be optimised in order to maximise the amount of reaction performed over the limited space available. To this end, it is desirable to use or provide a physical support for the reaction which provides the greatest surface area-to-volume ratio possible in order to maximise the efficient use of the limited physical space available in such devices.

To date, catalysed synthesis inside microreactors has been demonstrated using catalyst-coated beads or particles immobilised between frits. In addition, the use of in- situ synthesised monoliths instead of bead-like supports also solves the problem of how to keep the column securely in place within the channel. Where beads and particles need to be kept in place using frits or tapered channels, monolithic columns are synthesised inside the microchannel and anchor themselves to the channel wall.

Porous polymer monoliths are known for separation technologies. Such monoliths provide a sponge-like amorphous support of high surface area-to-volume ratio. In addition, such monoliths are subject to an array of chemical controls and other remote forms of manipulation.

However, to date such monolith supports have not been used in catalytic, synthetic or diagnostic technologies.

It has been surprisingly found that a support means including such a monolith can be utilized for catalysis, synthesis and diagnostic analysis on a microscale when the components or facets of a chemical reaction are bound to such a monolith using ligands. Statements of the Invention

Accordingly, in a first aspect the invention provides a porous support for catalysing chemical reactions comprising a porous polymer or silica monolith functionalised with a catalyst.

The chemical reaction may be catalysed by a catalytic agent or catalyst which may be directly or indirectly attached to the porous monolithic support. The catalyst may include a ligand as a binding means for attaching said catalyst to the porous polymer monolith. The ligand may in turn utilize a connecting arm for attachment to the monolith. The ligand may be adapted for binding of a metal and the metal may include palladium. The catalysis may include Suzuki type coupling reactions.

In a particularly preferred aspect, the reaction support of the invention includes a catalyst bound to the monolith by a ligand. In another aspect, the ligand is attached to the porous polymer monolith after formation monolith; this may be done using a reaction between a reactive functional group on the monolith and a reactive group on the ligand. In another aspect, the invention provides a reaction support comprising a porous polymer monolith with a ligand incorporated into the base monolith.

In another aspect, the invention provides a porous support for metal scavenging comprising a porous polymer or silica monolith functionalised with a chelating ligand. The scavenging may be conducted by a chelating ligand which may be directly or indirectly attached to the porous monolithic support. The chelating ligand may utilise a connecting arm for attachment to the monolith.

In a particularly preferred aspect, the scavenging support of the invention includes a catalyst bound to the monolith, hi another aspect, the ligand is attached to the porous polymer monolith after formation monolith; this may be done using a reaction between a reactive functional group on the monolith and a reactive group on the ligand. In another aspect, the invention provides a reaction support comprising a porous polymer monolith with a ligand incorporated into the base monolith.

The reaction support as previously described may be created and retained within a microreactor. The microreactor can take the form of a column, narrow-bore capillary, channel, capillary channel, microchip or other like device.

Functionalised monolithic columns may be created inside a microreactor/scavenging device by a polymerisation reaction, and the monolith may anchor itself to the channel wall. Subsequently, ligands may be covalently attached to functional groups in this porous monolithic support. Alternatively, ligands may be incorporated into the monolith during polymerisation, or may be grafted onto a base monolith using ligand-bearing monomers

Similar functionalised porous monolithic columns are already frequently used in chromatography and capillary electrochromatography for separation science because (i) packing of columns is replaced by synthesis of the column inside the capillary; (ii) no frits are required for retaining the column; and (iii) the low flow resistance of the porous support results in low back pressure.

The porosity of the reaction support of the invention provides a large surface area, resulting in an expected increase in the efficiency of catalysed reactions. Accordingly, in another aspect the invention provides a chemical reaction support comprising a porous polymer monolith and a means for binding one or more facets of said reaction to said monolith.

The monolith may include a porous silica monolith.

The reaction is preferably an organic chemical reaction, metal scavenging, diagnostic reaction or a catalysis reaction. The catalysis reaction may include Suzuki type coupling reactions.

The reaction facet may include a catalytic agent or catalyst and will most preferably include a ligand as a binding means for attaching said catalyst to the porous polymer monolith. The ligand may in turn utilize a connecting arm for direct attachment to the monolith.

In a particularly preferred aspect, the reaction support of the invention includes a catalyst bound to the monolith by a ligand. In another aspect, the invention provides a reaction support comprising a porous polymer monolith having a ligand incorporated therewith as an attachment means for one or more facets of the reaction. The reaction may be an organic synthesis, a diagnostic reaction, or a metal scavenging reaction.

The ligand may be adapted for binding of any metal and the metal may include palladium.

The reaction support of the invention most preferably provides a surface area-to- volume ratio of about 2-5 orders of magnitude greater than a conventional batch reaction.

In another aspect the invention provides a reaction support as previously described, retained within a microreactor. The microreactor can take the form of a column, narrow bore capillary, channel, capillary channel, microchip structure in a micro fluidic device or other like device. The microreactor of the invention can take the form of a capillary tube or larger device incorporating the reaction support as previously described, contained within the capillary tube. The microreactor may otherwise comprise a channel where the reaction support as previously described is contained within the channel. Functionalised monolithic columns may be provided inside a micro reactor by a polymerisation reaction, and the monolith will anchor itself to the internal walls of the device. Subsequently, ligands for metal complexation or organocatalysts may be covalents attached to the channel wall. Subsequently, ligands for palladium may be covalently attached to functional groups in this porous monolithic support. Similar functionalised porous monolithic columns are already frequently used in capillary electrochromatography for separation science because (i) packing columns is replaced by synthesis of the column inside the capillary; (ii) no frits are required for retaining the column; and (iii) the low flow resistance of the porous support results in low back pressure. Furthermore, the porosity of the reaction support of the invention provides an even larger surface area than a packed column, since the reagents are pumped through the support rather than around. This results in an expected increase in the efficiency of catalysed reactions.

In another embodiment, the invention provides a cartridge comprising one or a plurality of microreactors as previously described. In another aspect the invention provides a method of organic synthesis comprising the use of a microreactor as previously described as a flow through device for conducting said synthesis.

The method may include the use of multiple micro reactors in parallel for high throughput combinatorial reactions. In another aspect the invention provides a method of diagnosis comprising the use of a microreactor as previously described as a device for conducting said diagnosis. In another aspect the invention provides a method of metal scavenging comprising the use of a microreactor as previously described as a device for conducting said scavenging.

Detailed Description of the Invention

The invention will now be described with reference to particularly preferred embodiments as detailed below, figures 1.1 to 1.6, 2.1 to 2.21, 3.1 to 3.3, table 2.2 and 4.1 ; and appendices. M icrofluidic devices for organic synethsis

As previously detailed miniaturisation of organic synthesis to the microchip where format offers fundamental and practical advantages over conventional synthesis micro fluidic devices and capillaries can be used for synthesis as microreactors. Microreactors have been demonstrated to result in relatively pure compounds formed in higher yields and shorter time than equivalent bulk synthesis. The improved reaction control results partially from the small thermal mass of the microchip, which enables fast and accurate temperature control. The microchip format also affords greater control over the transport and mixing of reactants and products, sometimes even without requiring moving parts (eg. EOF). In separation science, microchips have been used for multiple separations in parallel. Equivalently, multiple reactors could be placed in parallel for high throughput combinatorial chemistry, potentially revolutionising drug discovery in the pharmaceutical industry. Detection of chemical species can be On- chip', for example microscopic imaging of absorbance or fluorescence signals, 'off- chip' by either reservoir sampling or suitable connection to analytical instrumentation, or 'in-chip' by integrating, for example electrochemical detectors.

While the amount of material produced in a chip-based micro reactor is insufficient for bulk purposes, it does allow for characterisation using conventional analytical instrumentation (such as gas chromatography - mass spectrometry (GCMS)) and screening assays in drug discovery. A microreactor may be hyphenated or integrated with other microdevices for purification and analysis to form a true Lab-on- a-chip. With multiple Lab-on-a-chips in parallel, large libraries of compounds could be synthesized and analysed quickly and cheaply. Ultimately, the microreactors should be integrated with a chip-based bioassay for screening purposes in drug discovery. Most synthesis reactions that have been performed on a chip to-date have been conducted to demonstrate proof of principle.

Liquid-phase reactions

For solution-based chemistry on microdevices, the channel networks are connected to a series of reservoirs containing chemical reagents, products and/or waste. Reagents can be brought together in a specific sequence, mixed and allowed to react for a specified time in a controlled region microreactor network.

Quantitative conversion of silyl enol ethers to β-hydroxy ketones was observed in 20 mins, whereas this reaction can take up to 24 hrs to complete for traditional batch synthesis. Catalysis

Catalyst based microreactors are of particular interest in organic synthesis as the surface area-to-volume ratio is 4-5 orders of magnitude greater than a conventional batch reactor. This means that heat can be rapidly transported into or out of a microreactor with little change in local temperature. This allows even highly exothermic reactions to be performed under isothermal conditions. Performing catalytic reactions in the microchip format also improves reaction kinetics, allowing for significantly accelerated catalytic cycles.

Catalysts can be heterogeneous, or homogeneous. Homogenous catalysts are in the same phase (e.g. a dissolved catalyst in a liquid reaction mixture), whereas heterogeneous catalysts are present in different phases from the reactants (e.g. a solid catalyst in a liquid reaction mixture). While several different homogeneous catalytic reactions have been demonstrated in microreactors to-date, these have been in capillaries due to difficulties in retaining homogeneous catalysts in the microchip format. Suzuki coupling reactions have been demonstrated in capillary-microreactors. A 0.4 mm diameter glass capillary filled with palladium nanoparticles and reagents was connected to an alternating current (AC) power source. An AC was applied for continuous agitation of the reaction solution through the nanoparticles using EOF. Reaction yields of 82-95% were obtained compared with 11-23% using conventional methods. The disadvantage of homogeneous catalysts is they have to be reclaimed from the reaction mixture and cannot be reused in some cases. Heterogeneous catalysts allow for much greater control of the catalytic reaction, in particular reactant selectivity and stereochemistry of the product. The basic scheme of the Suzuki coupling reactions of interest is shown in Figure 1.1. Heterogeneous catalysts can also be retained within the reactor by anchoring on a solid substrate, for example Merrifield resin which has been used extensively for organic synthesis. The polystyrene beads of the Merrifield resin contain benzylic chloride functionality to which organometallic catalysts can be attached. Suitable ligands have been developed for attachment to the Merrifield resin and in general, any ligand containing an alcohol, phenol, thiol or amine function can be attached to the benzylic chloride group. An example of a ligand that has been shown to have good catalytic activity on Merrifeld resin is l,10-phenanthrolin-5-ol which is shown in Figure 1.2. After anchoring the catalyst on the resin beads, the beads have to be retained within a microreactor. In capillaries frit-like structures have been used. Fabrication of microfluidic devices

Independent of the application, microfluidic devices need to be made to enable chemistry on a chip. Early microfluidic devices were fabricated using technology developed from the electronics industry, which involved using photolithography to create a pattern on the wafer surface and subsequent wet etching to produce microstructures.

Substrates

For most microfluidic applications, the ideal substrate is chemically inert, robust and transparent in the visible part of the spectrum. Silicon, the traditional microfabrication substrate, is not chemically inert, opaque and not suitable for electrophoresis because it is a semi-conductor. Subsequent further developments used microfluidic devices fabricated in glass and quartz substrates. The main fabrication method for glass involves the use hydrofluoric acid (HF) as etchant, which is a very hazardous substance. Patterning of a glass substrate also requires a serial fabrication process, which increases fabrication time, cost and risk of errors which may lead to a certain degree of irreproducibility between structures.

Plastics and polymers have recently become increasingly popular substrates for the fabrication of microfluidic devices since they are available with different chemical and physical properties, in pure form, at low cost and can be machined and replicated easily using a variety of methods. There are two types of polymers used for the fabrication of microchips, thermoplastics and elastomers. Thermoplastics are linear or branched polymers that are melted upon heating to a certain temperature. Examples of thermoplastics frequently used as substrates for microfluidic devices are poly methyl methacrylate (PMMA) and cyclic olefin copolymer (COC). Elastomers are weakly cross-linked polymers that can be easily stretched but will adopt their original state when relaxed. Elastomers do not melt before reaching their decomposition temperature. Poly(dimethylsiloxane) (PDMS) is an example of a common elastomer that is widely used for the fabrication of microfluidic devices. The fabrication of polymer microchips can be divided into two categories: direct fabrication and replication. During direct fabrication, individual polymer surfaces are structured to form required features. This includes such methods as laser ablation and machine milling. The inconvenience of direct fabrication is that every device is machined separately, requiring longer processing times and introducing a source of possible irreproducibility. Replication methods

Replication methods involve the use of a precise template or master from which many identical polymer structures can be fabricated. These methods include injection moulding, hot embossing and casting. During injection moulding, a polymer is melted and injected under high pressure into a cavity containing the master. This method is better suited for mass production than rapid prototyping as at least the first 100 units are wasted each run. During hot embossing, a planar polymer and master are heated separately in a vacuum chamber to the softening point of the polymer. The master and substrate are then brought together and embossed with a controlled force, usually 0.5-2 kN/cm2. During casting, a pre-polymerisation liquid is poured over a master and polymerised, usually by heating. The process for fabricating a microchip using casting is shown in Figure 1.3.

Master Fabrication As mentioned above, replication methods rely on a master template to form microstructures in the polymer surface. The simplest template is a small diameter metal wire that is bent into the required pattern and pressed into the polymer surface. More accurate and reproducible templates can be made based on silicon micromachining using standard photolithographic methods. A basic outline of this process is shown in Figure 1.4. During photolithography, structures are patterned on a surface using a light sensitive material called a photoresist. The pattern is transferred to the surface by exposing the photoresist to UV light through a mask which consists of transparent and opaque regions to define the required design. Traditionally, the mask was made from a quartz plate with the pattern defined in a thin chromium layer. The price of a mask for a 4" wafer is roughly $1,500. The group of Whitesides introduced an alternative method where a high resolution image printed on a simple transparency is used as a mask. The price of transparency masks varies depending on the resolution of the printer between $20 (2600 dpi) and $150 (4400 dpi). For the fabrication of microfluidic devices with structures greater than 10 μm, transparency masks provide sufficient detail, if structures less than 10 μm are required, the use of the more expensive chromium mask is recommended.

Photoresists are classified as positive and negative, depending on how they respond to radiation. Positive photoresists consist of three components: a photosensitive compound, a base resin and a solvent. Upon exposure to radiation, the photosensitive compound becomes soluble in the developer solution. Thus, the exposed areas are removed upon development. Negative photoresist are polymers. Exposure to radiation initiates cross-linking reactions between the polymer chains, increasing their molecular mass making them insoluble in the developer solution. Thus upon development, the unexposed areas are removed.

Silicon microstructures themselves are good masters for casting, but are too brittle for routine use during more forceful processes as hot embossing or injection moulding. A more rugged template may be formed using LIGA (Lithgraphic Galvano

Formung Abformung, which in English is lithography electroplating molding), where a nickel or nickel alloy master is electroplated on a silicon template. While LIGA methods produce accurate and robust masters, it requires slow and complicated processing resulting in a relatively high price per mould.

Bonding

After trenches in a solid substrate have been fabricated, they need to be sealed by attaching the substrate plate to a cover plate. Although conceptually trivial, one of the key challenges in making microfluidic devices is effective sealing between the structured layer and a cover plate. There are a variety of methods available for polymerbased devices. The most basic level is gluing the substrates together, which affords an adequate seal but has a high risk of channel blockages. Thermal lamination with a polyethylene terephthalate (PET) / polyethylene (PE) film (20 - 40 μm) is another commonly used bonding method. Lamination can be achieved at temps around 100°C using standard industrial lamination equipment. Other methods include pressureinduced sealing at elevated temperatures and laser welding. For elastomer polymers, sealing is a trivial operation. Conformal contact between the structured plate and cover provides for a reversible seal, whilst plasma oxidation affords a permanent (and leakagefree) bond with a variety of smooth materials.

Porous Polymer Monoliths

Macroporous monolithic polymers are sponge-like three-dimensional support structures that possess a large surface area. These materials are highly porous and are well suited for flow-through applications as a result of their low flow resistance. To date, monoliths have been used most often as separation media in liquid chromatography and capillary electrochromatography. The benefits of monoliths are ease of fabrication, reproducibility, highly configurable surface chemistry, and synthesis in-siru. In polymer monoliths fluid passes through the pores of the support rather than around packed beads, resulting in increased contact area with the support surface. The pre-polymerisation mixture is a homogeneous solution of monomers, porogens, and free radical initiator. The monomers determine the surface chemistry of the monolith formed after polymerisation, and the mixture generally contains one or more monomers and a cross-linker. Functionality can be introduced to the surface of the monolith either by using functionalised monomers to copolymerise in the support, or by introducing functionality to the surface after monolith formation via grafting. Functional groups on the surface of the monolith can be used as they are, or used for further modification after the monolith has been formed. Functionalised monolithic columns have been synthesised using either chloromethylstyrene (CMS) or glycidyl methacrylate (GMA) to obtain reactive functional groups inside the monolith. These monomers are shown in Figure 1.5. These reactive functional groups can be used for the covalent attachment of the ligands.

Synthesis of monolithic columns is performed by either UV or thermal-initiated radical polymerisation within the desired medium. The use of a UV initiator allows the region of monolith formation to be easily controlled using masking and allows the use of volatile solvents. The porogenic solvents control the pore size and can be tuned to give a required pore size. The porogenic solvent generally consists of two solvents, one more polar than the other. The trade-off for an increase in pore size, which leads to a decrease in the back-pressure produced when pumping fluids through the monolith, is an exponential reduction of the available surface area.

The process of monolith formation based on radical polymerisation is described as follows. Upon decomposition of the initiator, polymer chains begin to form. The growing polymers precipitate at some stage from the reaction mixture and form nuclei. Polymerisation continues both inside the nuclei and in the remaining solution to form microglobules. These grow further in size by coalescing with polymer species from the reaction medium and by capture of newly precipitated nuclei. The growing units reach a stage when they interconnect with each other by polymer chains threaded through neighbouring microglobules, forming larger clusters of globules. These globular clusters can be seen in Figure 1.6. If the porogenic solvent is a good solvent for the polymer, phase separation will occur later and the solvent competes with the monomers in solvating the nuclei. Hence, the concentration of monomer within the nuclei is relatively low and the attraction between individual nuclei that results in coalescence is limited. This leads to a larger number of smaller microglobules, resulting in a pore size distribution shifted towards smaller pores. Addition of a poor solvent to the polymerisation mixture is preferred and results in earlier phase separation and swelling of the nuclei with monomers remaining in the solution. As a result, formation of larger microglobular units is promoted which translates into formation of larger pores (or voids).

Polymer monoliths are described in some detail in US patent 6,887,384 which is herein incorporated by reference.

Towards Polydimethylsiloxane (PDMS) Devices for Monolith- Supported

Catalysis

50.8 Template Fabrication

Microchip Design Templates for the PDMS chips where fabricated by photolithgraphically patterning structures of SU-8 (Microchem, USA) on silicon wafers. Early template designs tried to make use of the circular constraint by having the channels point radially out from the centre, with input wells in the centre and output wells near the edge of the disk. The idea being that when the disk like chip is spun, the centripetal acceleration provides the force for fluid transport. This type of fluid transport is known as G-force transport. The first mask design (dubbed MkI) based on this type of fluid transport was created and was used to finetune the SU-8 template fabrication process. All subsequent designs discussed below were created in Corel® Draw® and are given in Appendix 1. In the MkI design, six y-shaped channels of equal width pointed radially out from the centre. Each y-shaped channel consisted of two wells for reagent input and a single well for product collection. Three major modifications to this design were incorporated in the Mk2 design: (i) the y-shaped channel was replaced with a straight-channel lay-out; (ii) the channel width varied between the 6 different structures on the single wafer and (iii) flow restrictors were incorporated. The first modification eliminated one of the inlet channels since a straight channel is sufficient for a purely catalytic system. This greatly simplified the design and allowed for easier interfacing with the chip. The channel lengths in the Mk2 design were slightly longer to allow both the MkI and Mk2 chips to be used in the CHl chip holder. The second modification involved varying the channels widths of the six channels from 100 to 350 μm in 50 μm increments to investigate the optimal channel width. Flow restrictors were introduced as the third modification to retain the monolith within the channel without the need for anchoring it to the surface due to blocking and to reduce the flow rate in the wider channels by increasing the flow resistance. After replication, the separate single channel structures were cut out of the PDMS slab. Only two or three of the six single channel devices could be recovered from the radial design replicate since the cut around a single device damaged the adjacent devices, hence the MkI and Mk2 designs were very wasteful. The next design (Mk3) was therefore based on four channels in parallel. This design is shown in Figure 2.1. The four channel manifold was placed centrally on the wafer to allow uniform height of the SU-8 on the wafer. It was found that as a result of the spin coating the thickness of the SU-8 decreased close to the wafer edge. The four single channel units could be either cut out and used individually, or kept together to form a four-channel chip. The four-channel chip was preferred, as the preparation of the four channels could be done simultaneously when using a four channel chip holder, saving a significant amount of time. While flow restrictors were incorporated in the Mk3a mask and not the Mk3b mask, there was no difference between the two devices as the flow restrictors were removed during the drilling of the reservoirs. Therefore all PDMS devices made based on the Mk3a and Mk3b masks will be referred to as Mk3 devices.

The flexibility of the Mk3 PDMS/PDMS devices was problematic as the reversible bonding between the PDMS plates was easily disturbed due to bending of the device. Initially the PDMS/PDMS sandwich was placed on a glass microscope slide (51 x 71 mm) to provide rigidity. Since it was not considered to be essential for the project to have all four channel walls made of PDMS, the PDMS slab containing the channels was later sealed directly to the microscope slide to form a PDMS/glass hybrid. The final chip design (Mk4) was very similar to the Mk3, the only differences being that the four channels were distributed to fit on a 75 x 50 mm glass slide and the flow restrictors were doubled in length. During redistribution of the channels, compatibility with the CH2 chip holder was taken into account. Multiple width channels were reintroduced into the Mk4 design. The Mk4 design was split into two separate designs as with the Mk3 design. The Mk4a (flow restrictors) and Mk4b designs are shown in Figure 2.2. The multiple channel widths resulted in different flow rates when using the chips in the chip holder under the same vacuum because of the differences in flow resistance. After formation of a polymer monolith inside the narrowest channels, it was impossible to flush the channels because of the high flow resistance. In future chip designs, the four-channel design should be maintained but every manifold of four 17 channels should have equal channel width to prevent differences in flow resistance. Different masks should be made with different channel widths to investigate the optimal channel width. Additionally, incorporation of a border around the four- channels would make cutting the chip manifold out as well as alignment of the PDMS and microscope slide significantly easier. Spin-coating

The wafer should be placed in the centre of the spin coater to ensure equal distribution of the photoresist. The spin coating process of the silicon wafers incorporated two spin cycles, a spread cycle and a spin cycle. During the spread cycle (1500 rpm for 15 sees), the pool of SU-8 was spread over 60 mm diameter. Normally, the SU-8 was spun over the wafer surface during the spin cycle (3000 rpm for 25 sees). When the temperature in the clean room increased to over 25 °C, the viscosity of the SU-8 dropped to the point where thickness of the spin coated layer was significantly less than on cooler days. This reduction could not be measured because of the stickiness of the SU-8, but could be observed with the bare eye. At these elevated temperatures, the spin speed during the spin cycle was varied between 1500 rpm to 2500 rpm to obtain a satisfactory result. For optimal processing conditions, a climate- controlled clean room would be required.

Hotplate

When heating the SU-8 templates on the hotplate during the pre-bake and post- bake, any particles falling onto the surface of the wafer became firmly stuck to the sticky SU-8. Although performed in a clean room, the wafer needed to be covered with a Pyrex dish to prevent dust from collecting on it during the pre-bake and post-bake phases.

Exposure

Exposure of the SU-8 templates was carried out using both the DUV lamp and the UV Crosslinker. SU-8 was developed by the manufacturer for optimal sensitivity in the near UV (350-400 run). The spectra of both lamps are given in Appendix 4. The intensity of the DUV lamp was apparently relatively weak in the near UV range and an exposure time of 15 mins was required for a 100 μm thick film of SU-8. The UV Crosslinker was equipped with 365 nm tubes, and the high intensity in the near UV resulted in an optimal exposure time of 60 sec. The light in the UV Crosslinker, however, is diffuse and SU-8 templates made using the UV Crosslinker for exposure contained uneven structures without the vertical walls characteristic of SU-8 lithography. The DUV lamp produces a parallel beam of light, allowing the production of well defined structures in SU-8. The UV Crosslinker would be the preferred exposure source if a large culminating lens was used to parallelise the exposure beam. Efforts to procure a collimating lens with diameter of at least 4" have been unsuccessful to date. Development

Development of the SU-8 required 30-45 mins in a fresh solution of the developer. The developer is reasonably expensive and should be used for more than one development. Unfortunately, the developer became less efficient in the removal of the unexposed SU-8 if it was used repeatedly. Even after a single use, the developer was unable to remove all the unexposed SU-8 from the next template. To overcome this issue, a two bath system was introduced. In the first bath, the majority of the SU-8 was removed, while the remainder of unexposed SU-8 closer to the formed structures was removed in the cleaner second bath. The completeness of development or removal of the SU-8 between the first to the second bath was verified using isopropanol. Any unexposed SU-8 immediately forms a milky white substance when in contact with isopropanol, whereas exposed and developed SU-8 does not give any reaction. Using the two bath system, the same volume of developer could be used for the fabrication of 2-3 times the number of templates that would normally be fabricated using a single developer bath. The structures formed in SU-8 were inspected by SEM. The ridges were higher than expected at approximately 200 μm, when compared with the expected height of 50 μm at 3000rpm coat speed. Also the ridges were spread out at the top, forming a T-shape as shown in Figure 2.3. This is called T-topping and is caused by excessive exposure of the SU-8 below 350 μm. The T-topping does not affect the function of the template as the elastomer replicates can deform around the ridges during removal. This may, however, reduce the lifespan of the SU-8 template, so to stop T- topping from occurring a filter to block out wavelengths below 350 μm should be used.

Mask Design

As discussed previously a mask of the required design is required to make the masters. The masks used here were 2600 dpi transparencies which was the highest resolution available in Australia at the time. This resolution corresponds to a pixel size of 9.6 μm, which is illustrated in Figure 2.4. The template designs were constrained to the area of the circular Silicon wafers which forms the base of the masters.

Replication of PDMS devices

PDMS replicates were formed using the templates previously described. Initially the reservoirs of the microchips were cut out using a 3 mm diameter hole-puncher, which tended to tear the PDMS. The preferred method of creating the well was to mark-off the wells using a whiteboard marker and then freeze the microchip in liquid nitrogen. The wells had to be marked-off using a black whiteboard marker, otherwise they were not visible once the microchip was removed from the liquid nitrogen and water condensated on the surface. Drilling the PDMS had to be carried out very quickly and one well at a time, as the PDMS thaws very quickly once taken out of the liquid nitrogen due to its small heat capacity. The process of drilling the wells made the microchips unavoidably dirty and needed to be thoroughly cleaned before use. Since PDMS is relatively adhesive, keeping the PDMS clean was difficult. Sealing the open channels using a PDMS cover plate made the microchip far too flexible, so glass was used for sealing most of the chips, which made the chips significantly easier to work with. While pressure-driven chips were being investigated, the bonding strength between the microchip and the glass cover plate was increased using the stamp-and- stick method. This procedure called for a glass plate to be spin coated at 8000 rpm with unset PDMS, and then pressed against the replicate and baked. Under these conditions a PDMS/glass microchip with only marginally improved bonding was produced. The procedure was repeated using spin coating speeds ranging from 1000 rpm to 7000 rpm in 1000 rpm increments, and 5000 rpm. was found to give the strongest bond without excessively filling the channels and wells. This method was no longer necessary when pressure-driven flow was abandoned during this project. While investigating the stamp- and-stick method, the PDMS replicates were produced using the hard-bake method described in the experimental section. This produces a slightly harder and more rigid replicate, which is undesirable when using vacuum-driven flow as the deformation of the PDMS inwards helps create a seal around the monolith within the channel. In Figure 2.5 a cross-section of a channel is shown. The T-topping from the SU-8 template is visible and seems to have little effect on the microchip. Moreover, it even seems to aid in anchoring of the monolithic column within the channel.

Interfacing and Fluid Flow

Fluid transport in the microchannels was driven by vacuum because of two reasons. Firstly, the bonding strength of the reversible seal of PDMS was found to be insufficient for pressure-driven flow even when the two slabs of PDMS were clamped together in a chip holder. Secondly, a PDMS channel can expand to 2-3 times its normal size under pressure. In case of the monolith-filled microreactors, the fluid flow would be forced around the monolith rather than through the pores. Using vacuum also has the benefit that it helps seal the channel around the monolith. Interfacing to the PDMS chips was initially done by pipetting fluid into the input well and creating a vacuum at the other channel using an auto-pipette tip connected to a syringe via rubber hosing. This is illustrated in Appendix 2, Figure 1.

The plastic of the auto-pipette tip did not reversibly bind to PDMS so that a reasonable amount of force was required to obtain a sealed connection. For flushing the channel, this pressure was applied by keeping the pipette tip in place by hand. Since this was impractical for flushing longer than a few minutes, the first chip holder (CHl) was designed to eliminate the manual syringe interface. CHl was designed to be used for both the MkI and Mk2 chip designs and featured two inlets and one outlet. CHl was manufactured from PMMA (Perspex) and HPLC fittings were used to interface to the chip wells. A photograph of CHl is given in Figure 2.6. PDMS was found not to reversibly bind to PMMA which resulted in no sealing between the chip holder and the channel wells. A rubber O-ring could not be used to seal the wells to the chip holder because of insufficient space between the two channel inlets, and therefore CHl was not used. For further work with the Mk2 devices, the syringe interface was improved greatly by replacement of the auto-pipette tip with a glass Pasteur pipette and by positioning the pipette with a retort stand. Figure 2.7 shows the optimal setup using the glass pipette interface with four-channels microchips. Since Pasteur pipettes are made of glass, an excellent seal was created around the channel well. The retort stand was positioned to apply a small, constant downward force on the pipette to ensure that the pipette and chip remained sealed. Yellow auto-pipette tips could be squeezed inside reservoirs drilled with the 1/32" drill to increase the size of the input well hereby increasing the solvent volume at the inlet and allowing the channel to be flushed for increased periods of time. The fluid inlet was further improved by connecting the auto- pipette tip via tubing to a 20 mL glass vial so that channel could be flushed unattended over night, or even over the weekend, without exhausting the solvent. Using this interface, a polymer monolith could be prepared inside a PDMS channel in three days, mainly as a result of the long flushing times required.

Interfacing to each of the four channels individually using the pipette interface was significantly more time consuming in the Mk3 and MK4 devices than using single channel chips. Therefore, the four-channel chip holder (CH2) was designed for use with the Mk3 and Mk4 devices. The CH2 chip holder is shown in Figure 2.8. CH2 was, like CHl made of PMMA and HPLC fittings were used to interface the holder with tubing. This time, rubber O-rings were incorporated to seal the channel wells to the holder. The tubing from the four outlets was combined to a single line using splitters and connected to a 60 mL syringe to provide the vacuum. CH2 was used extensively in conjunction with water, methanol, ethanol, and isopropanol as the solvents. Unfortunately, the solvent compatibility of PMMA and DMSO was not explored. After the chip holder was used to flush the DMSO ligand solution through the chips (see Section 2.5), DMSO partly dissolved the PMMA and permanently closed the well interfaces making the chip holder unusable. A photograph of the damaged reservoir is given in Appendix 2, Figure 4. To investigate the compatibility of other organic solvents that were likely to be used in the microreactor with PMMA, several solvents were pipetted into reservoirs drilled in a block of PMMA and left for 20 days with observations everyday. After only a few days, the largest extent of the damage had already taken place. The effects of the solvents are shown in Figure 2.9.

Each of the solvents deformed the PMMA, demonstrating its incompatibility as a chip holder material for an organic microreactor. The fine hairs visible in wells I, II, IV, and V indicated that the solvent has soaked into the PMMA and expanded, forming cracks. The noticeable swelling at the wells (wells I, III, IV, V) indicates that the PMMA is soluble in that solvent. A new chip holder (CH3) was made based on the design of CH2, but the bottom plate was made from aluminum and the top plate with fluid connections was made from Teflon. CH2 is shown in Figure 2.10.

Aluminium was chosen for the bottom plate because of its high thermal conductivity, so that the chip could be heated, for example for cleaning or ligand attachment. The top plate was made from Teflon so that it would be chemically inert. This chip holder proved to be unusable as an effective seal between the HPLC fittings and the top plate could not be made because the treaded holes were slightly too large. This leakage could not be overcome by wrapping Teflon tape around the fittings. A future version of the chip holder (CH4) would be all Aluminium with an interchangeable PMMA top plate for use in conjunction with appropriate solvents.

Microchip reactor:

A fully functioning microchip microreactor was fabricated from borosilicate microchips (Micronit), each with a channel length of ~30 cm and a cross-sectional area of 20,000 μm². The GMA monolith was prepared in the channel and the Palladium/phenanthroline complex attached as per the methodology previously described. The Suzuki-Miyaura reaction was used as a model reaction to demonstrate the potential of these glass microreactors for palladium-mediated carbon-carbon coupling reactions. The reaction of iodobenzene and tolyl boronic acid was performed at 80°C and the formed product was analysed by GC-MS or GC/-FID. The obtained yield of biphenyl of 68% is comparable to the yields obtained in control experiments. The reaction yield was monitored with increasing flow rate with only a small drop in reaction yield (-60%) at flow rates of 60μl/min. The microreactors are recyclable and numerous reactions may be performed in the one device.

Capillary Micoreactors:

A new catalyst system using N-heterocyclic carbenes was fabricated and tested for the Suzuki reaction in both bulk and capillary reactions.

These new systems have the same reactivity of previous catalysts systems but are much easier and less expensive to prepare. 1 -methylimidazole is passed over the monolith formed from chloromethyl styrene/divinyl benzene. Passing a palladium chloride solution or palladium acetate solution over the imidazolium salt gives the palladium carbene complex. The palladium loadings are also much higher with a value Of 0.71 mmol/g of palladium compared to 0.02 mmol/g for the phenanthroline system.

The suzuki reaction with new carbene/palladium catalyst.

Heck reaction:

The Heck reaction has been was trialed using bulk palladium/carbene monolith to determine appropriate conditions for flow through catalysis. The Heck reaction between an aryl halide and an acyrlate gives substituted acyrlates.

Heck reaction:

After testing numerous variations of solvent, base and temp it was found that N₅N- dimethyl acetamide as the solvent and triethylamine as the base at 120C gives the desired acyrlate in 89% yield after 72h.

Bulk monolith Heck reaction.

A mixture of 4-bromoacetophene (178 mg, 1.5mmol) N-butyl acrylate (600 μL) and triethylamine (304 μL, 2.2 mmol) in dimethylacetamide (5mL) with a palladium supported monolith was heated at 12O⁰C for 72 h. GC analysis of the reaction mixture showed a 89% yield of butyl 3-(4-acetylphenyl)-propenoate.

Monolith Synthesis Selection of Monomers

Poly(GMA-co-EDMA) monoliths were chosen as the epoxide ring of the GMA allows the palladium ligand to be attached via a simple condensation reaction under basic conditions. Also GMA and EDMA are transparent to UV radiation allowing poly(GMAco-EDMA) to be synthesised by photoinitiation. Poly(BuMA-co-EDMA) monoliths were also synthesised and photografted with GMA but were abandonded as it was found that the photografted monoliths have lower catalytic activity than the functionalised monolith. The mechanism of the monolith formation is given in Appendix 5.

Bulk Monolith

The polymer monoliths were synthesised in bulk to generate a sufficiently large amount to determine their pore size distribution. The pre-polymerisation mixture was prepared and transferred to a container for exposure in the DUV lamp. Initially quartz tubes with the ends sealed with Parafϊlm and Teflon tape were used to hold the mixture. The removal of the solid polymer monolith from the quartz tubes proved to be difficult as it required breaking the tube and removing as much of the quartz as possible. Using this process, the mass of monolith formed could not be determined since it was impossible to recover the entire monolith and to guarantee that what was recovered was not contaminated with quartz. An alternative container for the polymerisation of the bulk polymer monoliths was fabricated from PDMS. It consisted of two parts, a dish made from using a small watch glass or a weigh boat as the template and a 3 mm thick plate to seal the dish. The solid monolith was easily removed from this container and allowed 27 calculation of the yield. This container was suitable for bulk polymerisation using the methanol/ethanol as porogens, but unsuitable for repeated use with the hexane/methanol porogens. The lifetime of the container was found to be dependant on the porogenic solvents used. For the analysis of a greater number of bulk monolith samples with a variety of porogens, a reusable container made from inert materials like Teflon® glass and UV transparent quartz would be preferred.

Porosimetry

The pore sizes obtained by Mercury intrusion porosimetry for the bulk poly(GMA-co-EDMA) monoliths are given in Table 2.2.

Synthesis of poly(GMA-co-EDMA) monoliths using the porogenic solvents listed in Table 2.2 have not been previously reported. Further characterisation is therefore required. These porogens were chosen to approximate literature porogenic solvents which where unavailable. Although the average pore sizes obtained were appropriate for relatively high flow through the monoliths, in practice the flow rates were much lower than expected. The distribution of the pore sizes was quite broad, meanings that the pore sizes in the formed monoliths were not homogeneous. Thus low flow rates were observed as the overall flowthrough the monolith is only as fast as its slowest component. It is unclear if the pore sizes measured for the bulk monoliths reflect the pore sizes of monoliths grown in microchannels. A possible method for testing this would be to design a PDMS microdevice which is a microchannel spiral that covers the entire 4" silicon wafer, allowing the synthesis of enough monolith material for porosimetry analysis.

Monoliths Synthesis in PDMS Microchannels

A microchannel inside a PDMS chip was filled with a monolith pre- polymerisation mixture and exposed to UV light to produce a polymer monolith inside the channels. When exposing the chip in the DUV lamp, the wells needed to be covered with UV absorbent material to prevent monolith formation in the wells. Even with the wells covered, monolith was formed in the wells due to reflection of the UV light underneath the chip. To prevent the monolith forming in the reservoirs, the first 5 mm of the channels and the wells where masked off with black electrical tape and the chip placed on a base of black electrical tape. Figure 2.11 shows a PDMS microchip with monolith filled channels. After polymerisation of the monolith, the porogens and other waste material were removed by flushing the channel with methanol. Initially this was difficult because there was no chip holder and the channel was flushed by manually pipetting solvent into the wells and creating a vacuum at the opposite end with a syringe. Since the monoliths have to be rinsed for 1 hr, this was a tedious process. To avoid this process, the sealing plate was removed and the PDMS slab containing the monolith-filled channel was placed in the Soxhlet to rinse the monolith overnight. However, since the monolith was not chemically attached to the channel wall, some of the monolith was washed out of the open channel during rinsing in the Soxhlet. The tedious rinsing procedure was found to be an interface problem which was overcome by using the pipette based interface. This interface allowed the channel to be flushed without intervention. The flow rate for the poly(GMA-co-EDMA) monolith while flushing with methanol at room temperature was approximately 1 ml/min, and could be increased by reducing the viscosity of the methanol by heating the chip at 60°C on a hotplate. These conditions more closely resembled the conditions in the Soxhlet. Initially, flushing was only carried out using methanol, but it was found that flushing with water as well as with methanol significantly increased the flow rates, probably because more impurities were removed from the pores in the monolith. Polymer monoliths shrink during formation, which was observed in the microchannel. A photograph of a monolith formed inside a PDMS channel is given in Figure 2.12. In some regions of the channel the monolith occupied less than 50% of the channel volume, leaving a large void with a significantly lower flow resistance than the monolith, where solvents and reagents would flow around the monolith rather than through the monolith. In other formats where monoliths have been used, such as capillaries or COC chips, the problem of monolith shrinkage is overcome by chemically anchoring the monolith to the surface of the substrate.

PDMS Surface Modification

In order to chemically anchor a polymer monolith to the surface of a PDMS channel, methacrylate functionality must be introduced to the surface so that this may link to the polymer chains during the monolith formation. This is shown in Figure 2.13.

Modification of the surface chemistry in PDMS devices has been demonstrated by oxidation of the surface to create silanol groups. Hu et al. demonstrated surface modification of PDMS channels with water-soluble, functional monomers. The water soluble monomers were photografted onto the surface of the PDMS to reduce the hydrophobicity of surface. This method was based on radical formation at the PDMS surface when exposed to deep UV radiation for prolonged periods of time. The radicals formed at the PDMS surface could combine with a monomer to bind to the surface. This procedure was adapted here using GMA as the monomer, because GMA will also be used as a functional monomer to couple to the ligand. The procedure, based on the use of an aqueous solution, was found to be unsuitable for the hydrophobic GMA monomer. During radical formation, the GMA formed a film on top of the water and could not react with the PDMS. The experiment was repeated using methanol as the solvent, but no surface modification was observed. EDMA was used to replace GMA as the monomer using both water and methanol as solvents, but the surface modification was still unsuccessful.

Since GMA contains an epoxy ring, a simple condensation reaction could be used to get the required surface modification when a hydroxy functionality is introduced to the surface. This hydroxy functionality could be obtained by treating the PDMS with bromine and then water, as illustrated in Figure 2.14. This reaction was first trialed using a 10% Bromine in dichloromethane (DCM) solution. The solution was pippetted onto a PDMS slab and left in light for 30 mins for the reaction to take place. During this period, the solution soaked into the PDMS changing it from colourless and transparent to light yellow. During the evaporation of the bromine and DCM, cracks were formed throughout the PDMS slab leaving a very brittle and opaque material. The reaction was repeated using pure bromine with the same result.

A simple solvent-less photografting method was used. This method used a photografting mixture, containing 48.5 %wt GMA, 48.5 %wt EDMA and 3 %wt Benzophenone (a UV photoinitiator), to directly graft methacrylate monomers to the surface of the PDMS. Using this surface treatment on a slab of PDMS yielded no visible indication of grafting. The hydroxy functionality can be introduced to the PDMS surface by treating it with concentrated sodium hydroxide (NaOH). The effect of NaOH of PDMS is illustrated in Figure 2.15. This PDMS treatment has been used to create an EOF inside PDMS channels. A PDMS slab was treated with 5M NaOH for 1 hr, after which it was rinsed with water then methanol. To test the success of the treatment, droplets of water were added to the surface to give an indication of its hydrophobicity. The contact angle of the droplets was reduced indicating a more hydrophilic surface as a result of the treatment. Once the PDMS has a hydroxy functionality, its chemical structure is very similar to glass, which means that surface modification procedures for glass capillaries should be applicable for the treated PDMS. The surface modification process for attachment of monoliths to fused silica capillaries was adapted for surface modification of PDMS microchannels. The procedure for the capillaries requires several steps to clean the surface, which was not considered necessary for the freshly made PDMS. Additionally, acetone was used in the capillary method to remove the excess surface modification reagent. Since acetone is readily absorbed by PDMS, isopropanol was used for removal of the excess modification reagent in PDMS. After flushing the surface with NaOH to create the deprotonated Si-O- groups on the surface, the surface was treated with 3- (trimethoxysilyl)propyl methacrylate in a pH 5 ethanol solution to attach the methacrylate to the PDMS surface. This reaction is shown in Figure 2.16. This surface treatment was first carried out on a PDMS plate and worked very effectively (see Appendix 3, Figure 1 for an image of the treated plate). Confirmation of the monolith anchoring was tried using both transmission and reflection infrared (IR) spectroscopy. The spectrum of the surface modified plate was compared with that of a blank piece of PDMS and the solvent-less photografted PDMS. The reflection spectra are shown in Figure 2.17, the transmission spectra are given Appendix 6.

By transmission IR spectroscopy, no difference between the surface modified and normal PDMS could be observed. In the reflection IR spectra (given in Figure 2.17), none of the expected peaks were observed in the 'finger-print' region of the spectrum. EDMA has C=O groups present after polymerisation, which are not present in PDMS, that should be indicated by a sharp peak at around 1700 cm-1, which is not observed. This could be due to the layer of monolith present of the surface being too thin. This indicated that the surface had indeed been altered, but could not confirm the presence of the methacrylate on the surface. Successful surface modification was later confirmed by SEM images of a channel cross section which had monolith attached to the channel walls. On the PDMS slabs, the surface treatment with 5 M NaOH followed 34 by the methacrylate method was found to be most effective based on reflection IR and visual inspection. This method was used for the surface modification of the PDMS channels. The efficacy of the surface treatment was observed by visual inspection of the channel, using either the digital video microscope or the inverted microscope. A photograph of a monolith successfully anchored to the channel walls is given

Figure 2.18(a). The effect of the surface modification is clear when comparing Figure 2.18(a) and Figure 2.18(b). The binding of the monolith to the PDMS surface was also visualised using the electron microscope. In Figure 2.18(c), the remainders of the monolith on the PDMS surface can be observed. The monolith broke out off the PDMS when breaking the chip after freezing it with liquid nitrogen. Another measure for the effectiveness of the surface modification for anchoring the monolith was the flow resistance of the channel. In channels containing a non-anchored monolith, the flow rate was quite high, approximately 1 mL/hr. For channels containing a monolith anchored to the channel wall, the flow rates were estimated to be less than 40 μL/hr, a reduction of more than 25 times. This implies that in the, fluid flow is through the monolith when the monolith is anchored to the channel wall, whereas most of the flow will be guided around the monolith when the monolith is not attached to the channel wall. With the 4 μm pore size poly(GMA-co-EDMA) monolith anchored in PDMS microchips required flushing times to be increased to several days. This indicated that the pressure gradient provided by the 60 mL syringe was only slightly greater than the back-pressure generated by the monolith.

Ligand Attachment

The ligand is attached to the expoide ring of the GMA via a condensation reaction under basic conditions as shown in Figure 2.19. In this method, the ligand was dissolved in DMSO with 1.1 equivalents of NaOH to form an orange coloured solution. This solution was unstable and to prevent the ligand from precipitating the solution needed to be kept under nitrogen and stored at -4°C. When a monolith inside a PDMS channel was flushed with the DMSO/ligand solution, the solution would initially flow though, but the flow stopped after approximately a minute. Once the flow was interrupted, the ligand precipitated out of the solution in the channel and reservoirs. The initial hypothesis was that the flow stopped due to the ligand precipitating out of solution and clogging up the monolith. To prevent precipitation, the DMSO solution was pumped though a monolith inside a PDMS channel under various conditions including temperatures ranging from 20-80°C, decreased ligand concentrations, and reduced vacuum for pumping. Under all these different conditions the ligand precipitated.

The next step was to verify if it was the precipitating ligand that caused the blocking. Therefore, pure DMSO was pumped through a channel containing a monolith, and the channel got blocked as well. DMSO is incompatible with PDMS35. DMSO was pumped through an empty PDMS channel, and after 1 minute this channel was blocked as well.

During inspection of the blank channel under a microscope, it was observed that the DSMO was deforming the PDMS to the point where it closed off the channel, hereby preventing flow. The damage done by the DMSO was irreversible. It was concluded that DMSO is an unsuitable solvent for PDMS devices. Since the phenanthroline-based ligand is insoluble in most other 'PDMS compatible' organic solvents, but soluble in water under slightly basic conditions, an aqueous solution of KOH was used as the solvent system. In this basic aqueous solution the ligand was stable and did not require any special storage to prevent precipitation. However, the problem with this approach for binding the ligand to the monolith was that there were now two molecules, namely the hydroxide and the phenol-ligand, in solution that can potentially react with the epoxide ring on the poly(GMA-co-EDMA) monolith. The amount of KOH in solution needed to be just enough to dissolve the ligand, but insufficient to effectively compete with the ligand for attachment to the monolith. To dissolve 10 mg of the ligand (5.1 x105 mol) in 1.5 mL of water, 3 mg of KOH (5.6x105 mol) was required. In order to demonstrate that the aqueous system does lead to attachment of the ligand, a bulk poly(GMA-co- EDMA) monolith was prepared and used for testing the ligand attachment. The aqueous ligand solution was added to the bulk monolith to attach the ligand to the monolith. The success of the attachment was verified by passing an aqueous solution of Fe(II)S04 over the bulk monolith which made it turn pink, since the Fe2+ binds to the ligand to form a coloured complex. See Figure 2.20 for picture of Fe2+ complex on bulk monolith. This colour was retained after repeatedly rinsing the bulk monolith with methanol and water, confirming that the Fe2+ phenanthroline complex was attached to the monolith. Since the ligand forms highly coloured complexes with specific metals, colour changes are a good indicator of metal contamination of the functionalised monolith. Using the aqueous ligand solution, ligand attachment was again attempted repeatedly inside a monolith filled microchannel. Initially the solution would flow through the channel, but over the course of several hours, the ligand precipitated out of solution blocking the flow. On passing the ligand solution through a monolith PDMS channel it occasionally would turn bright red, indicating a metal contamination on the monolith. To minimise contamination of the monolith, the channels were only flushed with AR grade solvents and Milli-Q water. When attaching the ligand to the 4 μm pore size poly(GMA-co- EDMA) anchored monolith, the back-pressure was increased to the point where is exceeded the pressure differential generated by the 60 mL syringe. In-order to compensate for this a second syringe was introduced to apply a small amount of pressure to the input well of the channel, increasing the pressure difference across the monolith. This worked reasonably well, but is not recommended as is swells the channel, potentially damaging the binding of the monolith to the channel wall.

Figure 2.21 show ligand attachment to a anchored poly(GMA-co-EDMA) monolith inside a PDMS microchip. The faint red line in Figure 2.21 is where flushing the monolith with the Fe(II)S04 solution was attempted. The flow resistance of the monolith could be reduced by reducing its length, or by increasing its porosity. Decreasing the length of the monolith will result in a decrease in absolute catalyst loading, which might adversely affect the reaction yields. For future work, the better solution would be to use a different porogenic solvent mixture to increase the pore size of the poly(GMA-co-EDMA) monolith to 5 μm or above. This will require a significant amount of experimentation with various porogenic solvents, such as isooctane or toluene.

Nickel Template Fabrication Robust Microchips for Organic Synthesis

Since PDMS is not ideal as a chip substrate for organic synthesis and a chemically and physically more robust substrate is required. A polymer substrate that has the potential to overcome the disadvantages of PDMS is a thermoplastic cyclic olefin copolymer (COC). In particular, a special grade of COC that is transparent in low UV for monolith formation. For the fabrication of devices from thermoplastics at a small scale, hot embossing is the replication technique of choice. Hot embossing requires the template to withstand large pressures and temperature differences, which implies that the silicon wafer templates used for the fabrication of the PDMS devices are not likely to be sufficient.

Fabrication of Nickel Templates

Templates made from metal are ideal as these would be able to stand the pressure of hot embossing, and can be processed down to the micrometer scale. Nickel was the metal of choice because it is a reasonably strong metal, it is available in pure form and it can be etched to form defined microstructures. Traditionally, Nickel templates are made by electroplating Nickel on a silicon negative. In this thesis, etching of Nickel to create templates will be introduced. Since nickel plates have not been used previously for photolithography, the processes for spin-coating, pre-baking, post- baking, exposure, development and etching had to be established. The difficulty of adapting the procedure from the Megaposit® SPR®220 Series Photoresist data-sheet was that problems incurred at each individual step did not become apparent until the development stage. This made correcting the procedure rather difficult.

Spin Coating Only a 1 μm thickness of the SPR®220-3.5 photoresist is needed as the photoresist has a high tolerance to the nickel etchant. To obtain a 1 μm thickness, the nickel plate would need to be spun at 6000 rpm, which was unsafe and impractical due to the weight of the plate. The maximum safe spin rate was estimated at 2500 rpm and resulted in a theoretical photoresist thickness of approximately 4 μm, which is more than sufficient. Initially the pre-bake was carried out according to the photoresist datasheet, 90 sees at 115⁰C on a hotplate. This proved to be inadequate of the 5 mm thick nickel plates. The pre-bake time was extended to up to 10 mins, with little improvement observed in the photoresist adhesion with increased baking times. The intention of the pre-bake is to evaporate off the photoresist solvent. The main solvent of the SPR®220 photoresist is water. Therefore, baking the coated nickel plate was achieved in an oven 80⁰C for 1 hr. This resulted in excellent adhesion of the photoresist to the nickel surface.

Exposure

For testing purposes, the positive mask of the Mk3a chip design was used for the first nickel templates. The exposure was initially carried out under the same conditions as the SU-8 templates, but this surprisingly was insufficient for the SPR®220-3.5 photoresist which absorbs between 250 nm and 450 nm. Since exposure in the DUV lamp was insufficient, exposure was attempted in the UV Crosslinker at 550 mJ/cm2, which resulted in excellent structures in the photoresist. Of course the structures in the SPR®220-3.5 suffered from the same light diffusing problems incurred when exposing SU-8 in the UV Crosslinker, but these were less pronounced as a result of the thinner film thickness.

Development

Development of the photoresist was carried out according to the SPR®220 datasheet without modification.

Etching

Once the photoresist had been successfully patterned, the plate was etched using Nickel Etchant Type I. Theoretically, the etchant is capable of etching Nickel at 800 μm/hr, whereas the observed etch rate was 150 μm/hr. The optimal temperature conditions for the etchant are between 40-60°C. The etchant needed to be heated to within this range and agitated at the same time. To agitate the solution, a gyro-rocker was used and placing a hotplate on top of it would have created a significant hazard. Instead the etchant was heated to 60°C on a hotplate, and then the nickel plate was placed in the etchant, and agitated on the gryorocker. After 30 mins etching time, the structures in the nickel were reasonably sharp. The etch depth at 30 mins was approximately 80 μm which was inferred from SEM images (see Appendix 8). Figure 3.2 shows the under-etching the photoresist results in non-continuous ridges. After 1 hr etching time, the resist was severely under-etched, and in some regions all the Nickel had been removed from beneath the resist. Figure 3.1 shows that longer etch times result in thinner channels. Using a mask with wider channels would allow for a deeper etch without completely under-etching the photoresist. It is important to keep part of the polished Nickel intact otherwise the channel becomes non-transparent due to light scattering as a result of the rough etching surface.

In general there are two classes of etching, wet etching and dry etching. Wet etching is where the material is dissolved when immersed in a chemical solution. Dry etching is where the material is sputtered or dissolved using reactive ions or a vapour phase etchant. Wet etching is the simplest type of etching as all it requires in terms of apparatus is a container filled with a solution that will dissolve the require material. To form microstructures on a surface using wet etching, the patterned material is simply immersed in a liquid etching solution chosen specifically to selectively etch the desired material and not others exposed in solution. The depth of the structures is determined by controlling the reaction conditions. Materials that are crystalline can exhibit two types of wet etching, isotropic and anisotropic. Simply, isotropic etching is the same in all directions. Whereas anisotropic etching has different etch rates in different directions in the material depending on its crystal structure. The result of both processes can be seen Figure 3.3. Isotropic etching suffers from underetching of the resist but gives repeatable structures independent of the crystal structure. Whereas anisotropic etching gives sharper structures, but can be unpredictable in substrates that have not been treated to have a uniform crystal structure.

As the Nickel etchant is isotropic under-etching is unavoidable. Optimising the etching conditions can reduce under-etching, but for high aspect ratio structures better etching processes are available. Anisotropic wet etchants such as HF have been shown to produce well defined structures in Nickel, but due to the hazards associated with HF this type of etching is not appropriate. Direct reactive-ion etching (DRIE) using the Bosch process has been used to produce vertical side walls of depths up to 200 μm. This is where a patterned material is exposed to plasma, which reacts with the exposed surface of the substrate to for volatile compounds. These compounds are then desorbed from the surface and pumped out of the system. Using a time-multiplexed etching scheme with alternating etching and sidewall perseveration steps to prevent lateral under-etching, nearly vertical sidewalls and high aspect-ratio structures are possible.

In accordance with the previously described details the invention provides in a first example a reaction support made up of a porous polymer monolith to which is bound a ligand.

In a second example, a capillary is used to contain the polymer monolith. Reagents can be pumped into the capillary at one side, and product comes out at the other end. Multiple syntheses could be conducted in parallel, or the output could be increased by bundling multiple capillaries. The functionalised monolithic material could also be made in a larger manifold for reactions at a larger scale. The microreactors could be placed inside a cartridge for convenience. A commercial outcome for this part of the project could be achieved in a short amount of time.

A third example is based on the functionalised monolith, but using a microfluidic device or chip instead of a capillary. The above examples have demonstrated a positive yield inside a capillary filled with the reaction support of the invention.

In another embodiment, polymeric microdevices are provided using replication techniques. Initially, a negative of the desired design will be made by patterning SU8, a glass-like photoresist material, on silicon wafers. These masters will be used for casting devices in poly(dimethylsiloxane) (PDMS). PDMS is a flexible, UV transparent silicone rubber that reversibly seals leakage-tight to a large variety of materials. More rigid devices will be made from cyclo olefin copolymer (COC), a hard UV transparent plastic by hot embossing using a nickel mould. A scanning electron microscope (SEM) picture of a microchannel in COC is available. This channel was made by hot embossing, and sealed with the top plate by a solvent-assisted lamination process at the Institute for Microtechnology, Mainz.

Organic synthesis in the microfabricated devices of the invention finds important application in combinatorial chemistry. A set of microreactors in parallel can be used to quickly synthesise a library of compounds. Even though each individual compound will be synthesised in a small amount, this is generally sufficient for analysis. Ultimately, the device of the invention can be linked to an online bio-assay to determine its activity and facilitate rapid screening.

The small dimensions of a microreactor imply that relatively small amounts of product will be formed. Microreactors are therefore not very likely to find an application in the synthesis of bulk chemicals. However, the high degree of automation inside the microreactors of the invention makes them very attractive for the on-site synthesis of unstable and/or potentially dangerous products. This could avoid degradation before use, and eliminate the transport of hazardous, toxic compounds. Microreactors could also be used for synthesis on-demand, where small amounts are synthesised directly onsite, instead of in a remote chemical plant. In one particularly preferred aspect the invention provides a more systematic approach investigating palladium-catalysed synthesis in microfluidic devices.

Palladium-catalysed coupling reactions play an important role in organic synthesis especially in combinatorial chemistry, the main application area for microreactors. In combinatorial chemistry and drug discovery, large libraries need to be synthesised and tested, a process that involves numerous similar reactions in a multiple parallel format. Transfer of the synthesis to the microchip format leads to a significant reduction and consumption of reagents and waste production, thereby minimising the impact on the environment. Furthermore, an enormous reduction in man-hours could be obtained if these libraries were synthesised in parallel on microfluidic devices, thereby reducing the costs of the synthesis.

In another embodiment the microreactors of the invention are formed as a microchip using palladium catalysed reactions as one particularly preferred embodiment.

Various traditional reactions, such as Suzuki and Kumada coupling, esterification and dehydration, have been demonstrated on the microchip format, with relatively pure products formed in high yields. One application of microreactors of the invention is combinatorial chemistry and drug discovery, where palladium-catalysed reactions play an important role.

As an example, there has been great interest in the synthesis of derivatives of the highly active anticancer agent combretastatin A-4, a functionalised cis-stilbene. New derivatives or analogues are required to improve solubility, stability and activity. This can be achieved by a combinatorial approach utilising palladium catalysed coupling reactions.

A library of combretastatin A-4 analogues could be synthesised using the microreactors of the invention. The palladium catalyst is immobilised on a porous monolithic support where the large surface area will result in fast, efficient and high yielding reactions. The synthesised library could then be transferred online to a drug response assay system for direct screening.

Experimental

Figure 1: Combretastin A-4 and stilbem,d£nvatwe&._t, General Experimental Programmable Hotplate

A Torrey Pines Scientific ECHOtherm™ Model HS40 programmable hotplate was used for the heating stages of the SU-8 Template fabrication. The heating processes programmed into the hotplate are shown in Table 4.1

Ultraviolet Light (UV) Exposure

Two UV exposure apparatus with different peak intensities were used. The deep UV (DUV) lamp, fitted with a 500 W HgXe-lamp, was used for photo-initiation polymerisation and SU-8 photoresist exposure. Exposures with the DUV lamp were carried out at a constant intensity of 20.0 mW/cm2 unless otherwise stated. The other exposure apparatus was a Spectronics Corporation UV Crosslinker equipped with 15 W 365 nm UV tubes. The UV Crosslinker was used for exposure of the SPR™ 220 photoresist.

Spin Coating

The spin-coater was equipped with two vacuum chucks, one for 4" wafers and one for smaller substrates. During spin coating, the spin coater should be closed using the lid and the vacuum should be switched on to hold the substrate in place. Most optimal results are obtained when the substrate is placed in the centre of the chuck.

Agitation during Development and Wet Etching

For constant agitation during development and wet etching, a gyro rocker was used.

Porosimetry

Pore size measurements of bulk synthesised polymer monoliths were performed by mercury intrusion porosimetry using a 3 μm3 glass pentrometer for powder analysis.

Measurements were conducted on 0.1 g of polymer monolith samples. The sample was weighed into the penetrometer head, sealed with a small amount of vacuum grease, and the lid securely fastened with a spring clamp. Pressure measurements were taken from

0.5 psi (3.45x103 Pa) to 30,000 psi (2.07x108 Pa), to measure pore sizes ranging from

0.003 to 360 μm.

Microdevice Fluid Transport The CH2 chip holder was used to interface to the PDMS devices. The chip holder was connected to external devices using Nanoport® fittings and plastic capillary tubing. A 60 mL syringe was used to provide vacuum for fluid transfer and 20 mL glass sample vials were used as reagent/solvent reservoirs.

Infrared Spectroscopy

Transmission infrared spectroscopy was carried out using a Perkin Elmer Paragon 1000 FT-IR Spectrometer.

Reflection Infrared spectroscopy was performed using a Bruker Optics IFS 66 series FT-IR spectrometer.

Microdevice and Monolith Imaging MicroChannel and Monolith Inspection

Visible imaging of microchannels and monoliths within microchannels was carried out using a Nikon Eclipse TSlOO Inverted Transmission Microscope and a LabSmith SVM340 Synchronized Video Microscope.

Scanning Electron Microscopy (SEM)

SEM images were taken on a FEI Quanta 600.

Microfabrication

Devices in poly(dimethylsiloxane) (PDMS)

SU-8 Template Making

The method used for making the SU-8 template was adapted from processing instructions provided by Microchem and Groves. The process is described below. A 4" silicon wafer (orientation : <100> , 100 mm diameter, 525 +/- 25 μm thickness, one side polished, test grade, SWI Semiconductor Wafer Inc, Taiwan) was placed centrally on the vacuum chuck of the spin-coater with the polished side up. The vacuum was switched on to hold the wafer in place. A pool with a diameter of about 30 mm of the highly viscous photoresist SU-8 (SU-8 2000) was poured in the centre of the wafer. After placing the lid on the spin coater, the wafer was spun at 500 rpm for 15 seconds and then at 3000 rpm for 25 sees unless indicated otherwise. After spin coating, the wafer was placed on the programmable hotplate and the photoresist was pre-baked using program 1 (see Table 4.1). The wafer was then transferred to the DUV lamp and covered with a negative mask and exposed three times for five minute with breaks of 10 mins in-between. After exposure, the wafer was transferred back to the programmable hotplate for the post-bake using program 2 (see Table 4.1). The un- exposed SU-8 was removed by placing the wafer in a developer solution for 30-60 minutes with constant agitation using the gyro rocker, and then transferred to a second developer bath for one minute. The completeness of development was verified using isopropanol. After complete development, the wafer was thoroughly rinsed with isopropanol and blow dried using nitrogen gas. The wafer was placed on the hotplate for the hard bake using program 3 (see Table 4.1).

PDMS Device Fabrication Bulk Monolith Synthesis Container

The container consisted of two parts, a dish-like base and a lid, both made from poly(dimethylsiloxane) (PDMS). For the fabrication of a single container, 60.0 g of elastomer base and 6.0 g of curing agent (Sylgard 184 elastomer kit, Dow Corning, Michigan, USA) were weighed into a disposable mixing container and mixed thoroughly for approximately five minutes. For larger quantities of PDMS, the 10:1 elastomer base to curing agent ratio was maintained. The mixing container was placed in a large vacuum desiccator under vacuum to degas the PDMS. Care was taken not to overflow the unset PDMS from the container. Once most air bubbles had been removed, about 70% of the unset PDMS was poured over a plastic weighboat (25 x 25 mm) placed upside-down in a plastic Petri dish to form the container. The remaining unset PDMS was poured into another plastic Petri dish of the same size to form the lid. The Petri dishes were placed in the vacuum desiccator to remove trapped air bubbles under vacuum. The dishes were then heated in an oven at 80°C for one hour to cure the PDMS. The container and lid were cut out of the plastic Petri dishes using a scalpel.

Fabrication of PDMS Devices from an SU-8 Template

For replication of the structures on a silicon wafer, 70.0 g of elastomer base and 7.0 g of curing agent were weighed into a disposable mixing container and mixed thoroughly for approximately five minutes. For larger quantities of PDMS, the 10:1 elastomer base to curing agent ratio was maintained. The mixing container was placed in a large vacuum to degas under vacuum, while taking care not to overflow the unset PDMS from the container. Once most air bubbles were removed, the unset PDMS was poured over the silicon wafer template in a plastic Petri dish. The dish was placed in the vacuum desiccator to remove trapped air bubbles under vacuum. This dish was then heated in an oven to cure the PDMS at 8O⁰C for one hour for a basic chip. The chip was cut out of the dish using a scalpel and peeled off the template. For a hard baked chip, the curing process was done at 12O⁰C for 25 minutes. After removal from the template, the replicate was placed back in the oven to bake for a further 25 minutes at 90°C.

When using the four-parallel chip template, the four channels were cut out into a single chip using a 75^χ50mm glass microscope slide as a guide. The wells were then drilled after freezing the chip in liquid nitrogen using a 1/16th bit and a drill press. After drilling, the chips were cleaned by placing them in an ultrasonic bath for 10 mins. The chips were then wiped with lint-free tissues using first methanol and followed by isopropanol, and dried under vacuum over overnight. The four-channel chip was reversibly bound to a 75 x 50 mm glass microscope slide. The bond strength between the PDMS and glass can be increased by using the stamp-and-stick method developed by Satyanarayana et al. This process is described below.

10.0 g of elastomer base and 1.0 g of curing agent were weighed into a disposable mixing container and mixed thoroughly for approximately five minutes. The mixing container was placed in a large vacuum desiccator and pump down using an oil pump, taking care not to overflow the unset PDMS overflowed from the container. A 75x50 mm glass microscope slide was placed centrally on the vacuum chuck of the spin-coater. The vacuum was switched on to hold the slide in place. The glass was covered with 1O g of the unset PDMS and spin coated at 5000 rpm for 120 sees. The coated slide was placed in a plastic Petri dish (150 * 20 mm) and the PDMS replicate placed on the slide with the channel side down. Trapped air bubbles were removed by putting the dish in vacuum desiccator and pumping down with an oil pump. The unset PDMS was cured in an 8O⁰C oven for 30 mins.

Procedure for PDMS Surface Modification The method for surface modification was adapted from capillary surface modification procedures described by Rohr et al. The channel was subsequently rinsed with methanol and Milli-Q water for 1 hour each. Then the channel was flushed with 10 M sodium hydroxide for 1 hour followed by rinsing it with water and ethanol for 1 hour each. The channel was then flushed with the derivatisation reagent (3- (trimethoxysilyl)propyl methacrylate 20 wt% in ethanol (190 proof adjusted to pH 5 with acetic acid) for 1 hr then rinsed with isopropanol. Air was drawn through the channel to dryness. The channel was left for 24 hrs at room temperature to allow the condensation reaction of the silanol groups to complete.

Nickel Master Fabrication A high purity (300 x 1500 x 3 mm) Nickel (Nickel 201) plate was sourced from Austral Wright Metals (Sydney, Australia). The plate was cut into 50 mm x 75 mm pieces and machined as flat as possible using a magnetic chuck in the vertical milling machine. The individual plates were ground and polished by Simon Stephens using available equipment and consumables. The grounding and polishing procedure is given in Appendix 7. A 70 mm x 50 mm polished nickel plate was placed centrally on the small vacuum chuck of the spin coater. SPR®220 positive photoresist was poured over the plate. After covering with the spin coater lid, the plate was spun at 2500 rpm for 40 sees. After spin coating, the plate was placed in a plastic Petri dish, covered with aluminum foil and soft-baked in an oven at 80°C for one hour to evaporate the solvents. Once the plate had been removed from the oven and cooled to room temperature, it was covered with a positive mask and exposed to 550 mj/cm2 of UV light in the UV Crosslinker. After exposure, the plate was placed in a plastic Petri dish with a moist tissue for 30 mins to allow the polymerisation reaction to complete. The plate was post- baked on a hot plate at 115°C for 90 seconds to evaporate the absorbed water. After cooling to room temperature, the plate was developed in CD-26 developer for two mins in the first bath of developer and 10 sees in the second bath of developer. The plate was rinsed with water and blow dried with air. Remaining photoresist that was not part of the template design was removed with acetone and lint-free tissue paper. The plate was etched in Nickel etchant type I at 60°C for 30-60 mins with constant agitation using the gryo rocker. Finally, the photoresist was removed using a photoresist remover.

Porous Polymer Monolith Synthesis

Bulk Porous Polymer Monoliths by UV-initiated Polymerisation The procedure for the synthesis of porous polymer monoliths described below is adapted from the process developed by B. Preinerstorferet et al. The quoted porous polymer monolith pore sizes are based on mercury intrusion porosimetry measurements. 1 μm poly(glycidyl methacrylate-co-ethylene glycol dimethacrylate) monolith: The progenic solvent mixture was prepared by making a 1.50 g (50/50 wt%) mixture of methanol and hexane. The monomer/crosslinker mixture was prepared my making a 1.00 g (60/40 wt%) mixture of glycidyl methacrylate (GMA) and ethylene glycol dimethacrylate (EDMA). The pre-polymerisation mixture was prepared by making a 2:3 wt mixture of the monomer/crosslinker and porogen mixtures, adding 10 mg DMPAP as the UV initiator, and shaking well. 4 μm poly(GMA-co-EDMA) monolith: The porogenic solvent mixture was prepared by making a 1.5O g (50/50 wt%) mixture of methanol and ethanol. The monomer/crosslinker mixture was prepared my making a 1.00 g (60/40 wt%) mixture of GMA and EDMA. The pre-polymerisation mixture was prepare by making a 2:3 wt mixture of the monomer/crosslinker and porogen mixtures, adding 10 mg DMPAP as the UV initiator, and shaking well. Poly(BuMA-co-EDMA) monolith:

The porogenic solvent mixture was prepared by making a 1.50 g (5:1 wt) mixture of decyl alcohol and cyclohexanol. The monomer/crosslinker mixture was prepared my making a 1.00 g (60/40 wt%) mixture of butyl methacrylate (BuMA) and EDMA. The pre-polymerisation mixture was prepare by making a 2:3 wt mixture of the monomer/crosslinker and porogen mixtures, adding 10 mg DMPAP as the UV initiator, and shaking well.

Each of the pre-polymerisation mixtures were degassed by bubbling nitrogen through them before each use and stored wrapped in aluminum foil in a freezer at -4°C. The pre-polymerisation mixture was transferred to a PDMS UV-initiated polymerization reaction container and exposed in the DUV lamp for 10 mins. The polymer monolith formed was transferred to a thimble and rinsed with warm methanol in a Soxhlet for 8-12 hrs to remove unreacted reagents and excess solvents. The polymer monolith was then transferred to a sample vial and dried in a vacuum oven at 50°C overnight. The bulk polymer monoliths were then used for porosimetry measurements.

Photografting of poly(BuMA-co-EDMΛ) Monolith

The photografting procedure below was adapted using the process developed by T. Rohr et al. The UV-initiated grafting of poly(GMA) to the surface of a poly(BuMA- co-EDMA) monolith required the preparation of three mixtures.

Mix 1: combined benzophenone (0.025 g), t-butanol (1.500 g) and water (0.500 g) in a 4 mL sample vial.

Mix 2: combined t-butanol (1.500 g) and water (0.500 g) in a 4 mL sample vial. Mix 1 (0.1167 g), mix 2 (0.450 g) and GMA (0.100 g) were combined in a 4 mL vial, to make the photografting mixture. The poly(BuMA-co-EDMA) monolith was flushed with the photografting mixture until no more air bubbles were observed at the end, and then continued flushed for a further 30 mins. The polymer monolith was grafted by exposure in the DUV lamp for 60 sees and then rinsed with water for 1 hr. It was necessary to keep the grafted monolith wet, thus it was stored immersed in water. In-situ Monolith Synthesis in PDMS Channels

The preparation of the pre-polymerisation mixtures is described above (see x.3.1). The pre-polymerisation mixture was pipetted into a well of PDMS chip and filled the channel as a result of capillary action. The wells and flow-restrictors where masked off using black electrical tape to prevent the polymer monolith from forming in that region of the chip. Also the chip was placed on a mat of black electrical tape to prevent reflection of the UV light under the chip. The chip was exposed to UV light in the DUV lamp for 10 mins. The polymer monolith was rinsed with methanol and water for one hour each, in opposite directions, to remove unreacted reagent and the porogens. The procedure is repeated to fill the voids created by the shrinkage of the polymer monolith .

In-situ Monolith Synthesis in Capillaries Fused silica capillary surface modification The inner surface of a fused silica capillary (100 or 250 μm ID) was surface modified by briefly rinsing the capillary (by hand with a glass syringe) with acetone, then water, then NaOH (0.2 mol/L) until basic pH was detected to the terminal end of the capillary. The NaOH was then pumped through the capillary (0.25 μL/min) for a further 30 min. The capillary was washed with water until neutral pH detected, then rinsed with hydrochloric acid (HCI, 0.2 mol/L) until acidic pH was detected. The HCI was then pumped through (0.25 μL/min) for a further 30 min. The capillary was rinsed with water until neutral pH detected, then rinsed with ethanol (0.5mL, manual pressure). The surface modifying agent (solution of 20 wt% 3-(trimethoxysilyl)propyl methacrylate in ethanol adjusted to pH 5 [with acetic acid]) was pumped through the capillary (0.25 μL/min) for 1 hour. The capillary was rinsed with acetone, and then dried by passing air through the capillary using an empty syringe (0.20 μL/min) for 24 hours.

Poly gfycidyl methacrylate-co-ethylene glycol dimethacrylate monolith in capillary

The GMA/EDMA pre-polymerisation mixture was prepared as stated in 3.5.1. This solution was purged with nitrogen for 10 min then pumped into a surface modified capillary (250 μm ID) (from 3.6.1) using a 250 μL glass syringe, until the capillary was totally filled with the mixture and no air bubbles were apparent (inspected via magnifying glass). The filled capillary was sealed at both ends with rubber HPLC septa and placed in a water bath at 60°C for 20 hours. The monolith-filled capillary was then flushed with MeOH (pumped at 5μL/min, lhour).

Mass of monolith = 1.7mg monolith / 5cm capillary [250 μm ID].

Functionalisation of the Monolithic Support Ligand Attachment

The 5-hydroxy-l,10-phenanthroline ligand was used. The ligand solution was prepared by dissolving 5 -hydroxy- 1,10-phenanthroline (10 mg) in 1.5 mL of water with 1.1 equivalents of potassium hydroxide (2.9 mg) to produce a light orange coloured solution. The solution was purged with nitrogen for 10 mins before use. For ligand attachment, the polymer monolith was first flushed with water for 15 mins and then flushed with the ligand solution overnight at 60°C. Flushing with the ligand solution was repeated, but in the opposite direction with respect to the polymer monolith. Finally, the ligand-bound polymer monolith was rinsed with water until the expelled water appeared colourless, and then was flush for a further 30 minutes.

Iron(II) Complexation

A saturated solution of Fe(II)SO4 (aq) was sourced. The ligand bound polymer monolith was rinsed with water for 30 mins, and then flushed with the Fe2+ solution for 1 hr. Finally the ligand-bound polymer monolith was rinsed with water until the expelled water appeared colourless, and then was flush for a further 30 minutes.

The ligand bound iron (II) could then be used as a catalytic iron (II) source for organic synthesis such as demethylation of tertiary amine N-oxides as has been reported in the literature. Therefore, by passing a solution of a tertiary amine N-oxide over the iron

(II) ligated monolith a demethylation should result.

[K. McCamley, J. A. Ripper, R. D. Singer and P. J. Scammels, "Efficient N-

Demethylation of Opiate Alkaloids using a Modified Nonclassical Polonovski Reaction", Journal of organic Chemistry, 20033, 68, 9847-9850]

Suzuki Miyaura Reaction

Bulk monolith Suzuki-Miyaura reaction

The reaction mixture was produced by stirring iodobenzene (11.12 μL, 0.1 mmol), tolylboronic acid (20.4 mg, 0.15 mmol), tetrabutylammonium methoxide solution (20% in MeOH) (167.2 μL, 0.1 mmol), in toluene and methanol (9:1, 2 mL), along with diphenyl ether (15.9 μL, 0.1 mmol) as an internal standard.

Bulk GMA monolith (100 mg) with 5-hydroxy-l,10-phenathroline and Pd¹¹ attached (from 3.7.2) was placed in a 5 mL Reacti-Vial along with the above reaction mixture (2 mL). This was stirred vigorously at 8O⁰C for 90 min, then cooled in ice.

The reaction was filtered through a Bϋchner funnel to separate the monolith and the reaction products, the products were further filtered through a plug of cotton wool. The reaction yield was analysed by GC-MS (75% yield of 4-phenyltoluene).

Capillary Suzuki-Miyaura reactions

Each capillary was flushed (2 μL/min) with toluene and methanol (9:1) for 1 hour at 8O⁰C prior to reaction to equilibrate the conditions.

The reaction mixture was prepared by adding iodobenzene (11.1 μL), tolylboronic acid (20.4 mg), and tetrabutylammonium methoxide solution (20% in MeOH) (167.2 μL, 0.1 mmol) to a mixture of toluene and methanol (9:1, 2 mL). To this diphenyl ether was added as an internal standard (15.9 μL, 0.1 mmol).

The reaction mix was passed through the capillary over 18 h at 8O⁰C (0.0507 μL/min, equivalent of an approximate contact time of 90 min) and the product collected from the opposing end in small sample vials. The capillary was then flushed with toluene and methanol (9:1) for 1 hour to remove the remaining reagents (2 μL/min).

The reaction was analysed-by GC-MS (68% yield of 4-phenyltoluene).

Note: The same yield was obtained after a continuous flow-through reaction over four days under the same conditions.

Organocatalysts:

Two organocatalysts monoliths were synthesised and tested for reactivity to catalyse organic reactions. These were formed on bulk material initially for ease of analysis and rapid testing. The aim was to take reactions into flow though synthesis in capillaries. The monolith from Chloromethyl styrene/divinylbenzene was used as the scaffold to provide a highly reactive benzyl chloride function for the attachment of the catalysts.

Carbene Organocatalysts-

These were formed by reacting the chloromethyl styrene monolith with 1- methylimidazole. Initially an imidazolium salt forms which on treatment with a base gives the N-heterocyclic carbene which is a useful organocatalyst.

CMS monolith N-heterocyclic carbene catalyst

These carbenes have been used in carbon-carbon bond forming reactions. The most basic example of this is the Benzoin reaction from aromatic aldehydes. It has been shown that p-chlorobenzaldehyde undergoes the benzoin condensation when exposed to the monolith supported carbene and the soluble base lithium hexamethyldisilazide. Initial attempts to take this into a capillary have not been successful.

Lithium hexamethyldisylazide tetrahydrofuran

benzoin product

It was found that N-heterocyclic carbenes supported on monoliths behave differently to polymer bead supported reagents as a change of base results in a change of reactivity. When sodium hydride was used an ester was formed from a competing Tischenko reaction (1 :1 ratio with benzoin product). The corresponding polystyrene support yields the benzoin product.

---rid--e^

benzoin product Tischenko product

Synthesis of N-heterocyclic carbene organocatalyst supported on a monolith.

Capillaries : 1-Methylimidazole was passed over a monolith prepared from chloromethyl styrene/divinyl benzene in a 250 μm capillary at a flow rate of 0.5 μL/min for 8 hours at 5O⁰C. The capillary was then reversed and the solution pumped for a further 18 hours before washing with a solution of chloroform by pumping for 2 hours at a flow rate of 2.0 μL/min. Treatment of the monolith in-situ with a base generates the active catalyst.

Bulk: A solution of 1-methylimidazole (0.3 g) in chloroform was added to the chloromethyl styrene/divinyl benzene monolith (Ig) and the mixture heated at 5O⁰C for 24h. The monolith was collected by filtration and washed with more chloroform to give the catalytic monolith. Treatment of the monolith in-situ with a base generates the active catalyst.

5 Bulk monolith Benzoin reaction-

A solution of lithium hexamethyldisilazide (1.5 Ml of a 1.5M solution in THF) was added to a suspension of the N-heterocyclic carbene monolith and 4- chlorobenzaldehyde and the solution stirred at room temperature for 1 h. Water was added and the products extracted into dichloromethane to give p,p'-dichlorobenzoin in 10 64% yield.

Amine monoliths-

The reaction of N-methylpiperidine gives a monolith with a tertiary amine which 15 should be a substitute for organic bases such as triethylamine which is a common catalyst in organic synthesis.

0 This monolith was used in the aldol reaction between dimethyl malonate and 4- chlorobenzaldehyde. Thus stirring a solution of the aldehyde, dimethyl malonate and the amine monolith in a solution of THF at room temp yields the aldol adduct in 50% yield after 5 days.

* H₃CO₂C CO₂CH₃ tetrahydrofuran

5 aldol product

Synthesis of tertiary amine monolith. 0 Bulk: A solution of N-methylpiperidine (0.5 mL) in chloroform (2mL) was added to the chloromethyl styrene/divinyl benzene monolith (0.2 g) and the mixture heated at 6O⁰C for 3.5h. The monolith was collected by filtration and washed with more chloroform followed by ether to give the catalytic monolith.

Bulk monolith aldol reaction- The tertiary amine monolith formed from chloromethyl styrene/divinyl benzene monolith and n-methylpiperidine was added to a solution of dimethyl malonate and 4- chlorobenzaldehyde in THF and the mixture stirred at room temperature for 5 days. Removal of the monolith by filtration and evaporation of the solvent gave dimethyl (4- chlorophenylhydroxymethyl) malonate in 50% yield.

Organic synthesis on a microchip of the invention has numerous advantages over prior art devices. In addition to the reduced reagent consumption and waste generation, increased reaction rates in combination with higher yields have been demonstrated. Applications in the pharmaceutical industry provide an attractive future for microreactors, justifying investment in fundamental, systematic research on microfluidic devices for catalysed synthesis.

The use of palladium mediated reactions in combinatorial chemistry is one of the driving forces for the recent surge in the development of supported palladium catalysts, where the coordination environment at the metal centre is similar to that for soluble homogeneous catalysts but the metal centre is bound to a support material. Well-defined solid-bound palladium catalysts are sought mainly because of the improved ability to reuse the metal and the ligands, thereby reducing the costs in industrial processes. The invention also greatly facilitates the workup and purification of the products compared to homogeneous catalysts as tedious extractions are reduced to a simple filtration. Consequently, this invention includes the development of a new class of supported catalysts by immobilising palladium onto a macroporous polymer monolith.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive. Appendix 1 PDMS Microchip Designs

Figure 1 Original MkI design by Dr Rosanne Guijt and Dr Thomas Rodemann.

Figure 2 Design Mk2a which incorporates flow restrictor and multiple channel widths.

Figure 3 Mk2b chip design featureing multiple channel widths.

Figure 4 Mk3a design featuring four channels in parallel, flow restrictors and 350 μm channel diameter.

Mk3b (non-flow restrictors)

Figure 5 Mk3b design incorporating a four-channel chip layout and 350 μm channels.

Figure 6 Mk4a design featuring flow restrictors, four- channel layout, and multiple channel widths.

Figure 7 Mk4b design featuring a four-channel layout and multiple channel widths.

Appendix 2 Additional PDMS Microchip Interfaces

Appendix 3 Attempted Surface Modification of PDMS Plates

Appendix 4 Transmission IR Spectra

I X: 4 scans, 4.0cα-l

Figure 1 Transmission IR spectrum of photografted PDMS.

I X: 4 scans, 4.0oα-l

Figure 2 Transmission IR sectrum of PDMS with anchored monolith on surface.

f X: 1 scan, 4.0eι»-l

Figure 3 Transmission IR spectrum of untreated PDMS.

Appendix 5 Steps for formation of PoIy(GMA-CO- EDMA) monolith

Step I: Radical Formation

2,2-dimethoxy- 1 ,2-diphenylethanone

Decomposition

Forms four radicals denoted R-

Step II: Chain Linking

Glycidyl Methacrylate

Step III: Cross-linking

Step IV: Radical Termination

R OU- ^R — ► R — F

Appendix 6 UV Lamp Spectra

Figure 1 Lamp spectrum for DUV Lamp, show regions of absorption for the photoresist used.

Appendix 7 Nickel Plate Polishing

The following description of the nickel plate polishing was complied by Simon Stephens (Lapidary, School of Geology, University of Tasmania)

The trouble is that polishing anything is a black art. There is quite a lot of literature and anecdotal stuff about, but applying it to a specific situation takes some skill and trial and error. The main problems with the nickel sheet provided were:

• Uneven surface with residual tension so that only light pressure could be used for initial coarse grinding and stock removal.

• Mutual attraction of diamond and nickel so that coarse diamond abrasive becomes embedded in the metal to be later released at a finer polishing stage and causing large scratches.

The optimal process using available equipment and consumables consists of 4 stages

Coarse Grinding:

Kemet 15 grinder/polisher, 400grit SiC abrasive, water slurry on Meehanite (nodular cast iron) plate, 60 rpm, pressure 50gm./sq cm. Time 5 to 4 hours depending on initial flatness.

Fine Grinding:

Logitech LP30 lapping and polishing machine, 9 micron alumina abrasive, water slurry, Meehanite plate, 60 rpm, pressure 150gm/sq cm , time 20 min.

Coarse Polishing:

Buehler Ecomet IV polisher with Euromet power head, Hyprez type K 3 micron liquid diamond abrasive, low viscosity oil slurry, Kemet MSF surface (syn. silk, laminated), 200 rpm, Pressure 100 kpa , 20 min. Fine polishing:

Buehler Ecomet as above, Keraet type K Std 1 micron liquid diamond, low viscosity oil slurry, Kemet MSF surface, 75 rpm, pressure 50 kpa, 5 min.

Many other combinations were tried at different stages and the processing in Sydney for coarse grinding turned out to be of no net gain because of the embedded abrasive problem and the plates were damaged on the return journey.

Appendix 8 SEM Images of Etched Nickel Templates

Claims

CLAIMS:

1. 1. A porous support for conducting one or more chemical reactions comprising a porous monolith functionalised with an agent.

2. A porous support for conducting one or more chemical reactions wherein the agent is a catalytic or a scavenging agent.

3. A porous support according to claim 1 or claim 2 wherein the porous monolith is a polymer.

4. A porous support according to claim 1 or claim 2 wherein the porous monolith comprises silica.

4. A porous support according to any one of the preceding claims wherein the catalytic or scavenging agent is directly or indirectly attached to the porous monolith.

5. A porous support according to any one of the preceding claims wherein the one or more chemical reactions include organic chemical reactions, metal scavenging, diagnostic reactions or catalysis reactions.

6. A porous support according to any one of the preceding claims wherein the catalytic or scavenging agent is attached to the porous monolith by means of a ligand.

7. A porous support according to claim 6 wherein the ligand is a chelating ligand.

8. A porous support according to any one of claims 6 or 7 wherein the ligand utilises a connecting arm for attachment to the porous monolith.

9. A porous support according to any one of claims 6 to 8 wherein the ligand is attached to the porous monolith after formation of the porous monolith.

10. A porous support according to any one of claims 6 to 8 wherein the ligand is attached to the porous monolith by being incorporated into the porous monolith prior to, or during, the formation of the porous monolith.

11. A porous support according to any one of claims 6 to 10 wherein the ligand is adapted to bind any metal.

12. A porous support according to any one of claims 6 to 10 wherein the agent is palladium.

13. A porous support according to any one of claims 6 to 10 wherein the agent is N- heterocyclic carbene.

14. A porous support according to any one of claims 6 to 10 wherein the agent is a primary, secondary or tertiary amine.

15. A porous support according to any one of the preceding claims wherein the porous support provides a surface-to-volume ratio of about 2-5 orders of magnitude greater than a conventional batch reaction.

16. A porous support according to any one of the preceding claims wherein the porous support is formed within, or on, a microreactor.

17. A porous support according to any one of the preceeding claims wherein the porous support is prepared in a mould and repackaged into a cartridge or similar housing.

18. A porous support according to any one of the preceding claims wherein the porous support is formed within, or on, a microreactor via a polymerisation reaction.

19. A porous support according to any one of the preceding claims wherein the porous monolith anchors itself to the microreactor during formation.

20. A porous support according to any one of claims 16 to 19 wherein the microreactor is in the form of a column, a narrow-bore capillary, a channel, a capillary channel, a microchip or other like device, or a combination of two or more such devices.

21. A reaction support comprising a porous support according to any one of the preceding claims.

22. A microreactor comprising a porous support according to any one of claims 1 to 15.

23. A microreactor according to claim 22 in the form of a column, a narrow bore capillary, a channel, a capillary channel, a microchip or other like device or a combination of two or more such devices.

24. A microreactor according to claim 22, wherein said microchip reactor comprising a base of chip material selected from any one or a combination of Chip materials include but are not limited to quartz; glass including borofloat, pyrex, borosilicate, fused silica, soda lime, diamond; polymers including PMMA, COC, polycarbonate, PET, TEFLON, TEFZEL, PEEK; metal including nickel, stainless steel, copper, silver, iron, gold, platinum and alloys; ceramics and silicon and silicon-related materials including silicon carbide, silicon nitride, silicon oxide.

25. A microreactor chip based device according to claim 21, having a channel length of between 1 to 100 cm long and 10 μm to 1 cm wide.

26. A microreactor capillary/tube/cartridge based device according to claim 21, having a channel length of 1 to 100cm long and 10 μm to 10cm wide.

27. A cartridge comprising one or a plurality of microreactors according to any one of claims 22 to 24 and 26.

28. A method of organic synthesis comprising the use of a microreactor according to any one of claims 22 to 25 in a flow through device for conducting said synthesis.

29. A method according to claim 28 including the use of multiple micro reactors in parallel for high throughput combinatorial reactions.

30. A method of diagnosis comprising the use of a microreactor according to any one of claims 22 to 25 as a device for conducting said diagnosis.

31. A method of metal scavenging comprising the use of a microreactor according to any one of claims 22 to 25 as a device for conducting said scavenging.

32. A method of demethylation comprising the use of a microreactor according to any one of claims 22 to 25 as a demethylization device.

33. A porous support according to any one of claims 1 to 20, substantially as hereinbefore described in the examples.

34. A microreactor according to any one of claims 22 to 26, substantially as hereinbefore described in the examples.

35. A cartridge according to claim 27, substantially as hereinbefore described in the examples.

36. A method according to any one of claims 28 to 32, substantially as hereinbefore described in the examples.