The synthetic principes of microemuisions have been widely described in the literature. For recent work see for example: F. J. Arriagada and K. Osseo-Asare, "Synthesis of Nanosized Silica in Aerosol OT Reverse Microemulsions,"Journal of Colloid and Interface Science 170 (1995) pp. 8-17 ; V. Chhabra, V. Pillai, B. K.

Mishra, A. Morrone and D. 0. Shah,"Synthesis, Characterization, and Properties of Microemulsion-Mediated Nanophase TiO2 Particies. "Langmuir 11 (1995) pp. 3307- 3311; H. Sakai, H. Kawahara, M. Shimazaki and M. Abe, "Preparation of Ultrafine Titanium Dioxide Particles Lising Hydroiysis and Condensation Reactions in the Inner Aqueous Phase of Reversed Micelles : Effect of Alcohol Addition," Langmuir 14 (1998) pp. 2208-2212; J. Tanori and M. P. Piìen !"Control of the Shape of Copper Metallic Particles by Using a Colloïdal System as Template,"Langmuir 13 (1997) pp.

639-646. A microemulsion is a sufficiently thermodynamically stable solution of two normally immiscible liquids (for example, oil and water) consisting of nanosized droplets (or cores) of one phase in another"continuous"phase, stabilised by an interfacial film of a surfactant with or withou, a co-surfactant. Examples of surfactants include ionic ones such as Aerosol OT and cetyidimethylethylammonium bromide, and non-ionic ones such as the ooiyoxyethylene ether and ester surfactants. Examples of co-surfactants include medium to long alkyl-chain alcohols such as 1-nexano'. i xamDies of oiis inciuae nvzr.. zarbons such as cyclohexane and isooctane. The surfactant and co-surfactant molecules reduce the interfacial tension so that stable dispersions can be formed.

Forming nanoparticles by the microemulsion route typically involves preparing a reaction mixture as a water-in-oil reverse micellar system using a ternary phase mixture containing high oil and surfactant contents, but low water content. This allows discrete but thermodynamically-stable nanometer-sized"water pools"or "water cores"to develop in the reaction mixture. In a typical water-in-oil microemulsion, the water cores are around 1 to 10nm in diameter. One reactant for the nanoparticle formation can be initially housed in these water cores. The second reactant can subsequently diffuse into and react inside these"nano-reactors"in the normal course of microemulsion dynamics. In this way, microemulsions provide a versatile route to the controlled synthesis of a wide array of oxide and non-oxide types of nanoparticle. In the water pools a metal salt can be reduced to the free metal, or metathesis reactions can be included, to obtain a controlled nucleation and growth of the desired nanoparticle material. The surfactant also acts as a coating to prevent unwanted flocculation (agglomeration) of the growing particles. Many of the fundamental principes governing such micellar chemistry, such as reaction rate and final growth size, are still largely unknown. Most experiments are done by trial-and- error and the data interpreted empirically.

Much of the work on nanoparticles has concentrated either: (i) on demonstrating that nanometer-sized particles have indeed been created (for instance by using transmission electron microscopy (TEM) or ultraviolet-visible (UV-Vis) spectroscopy) ; or (ii) on subsequently sintering the nanoparticles to prepare a sintered body. This work has involved relatively crude techniques for handling the nano-sized material.

For aspect (i) there has generally been no need to isolate or further manipulate the nano-size material. For aspect (ii) the formed material has typically been recovered from the emulsion by bulk precipitation upon addition of a destabilising solvent, or by vacuum removal of the reaction solvent. The material is then sintered at high temperatures to obtain the desired nano-grained article after"burning off'of the surfactant coating. Since the particles are to be sintered into a solid mass there is no need to counteract their tendency to aggregate.

Some work has been done on other uses for nanoparticles. S Carter, J C Scott and P J Brock, Appl. Phys. Lett, 71,1997, 1145-1147 describe the use of polydispersed Ti02, Si02 and Al203 nanoparticles in the form of a blend in polymer LED devices with the aim of enhancing the forward emission of light generated in the LED and/or improving carrier injection and recombination. The route by which the nanoparticles are obtained is not described, but the particles are described as having relatively large sizes: 30 to 80nm, especially in comparison to the device thickness of 110nm.

It appears from the presence of light scattering that the nanoparticle material suffers from agglomeration. Thus the nanoparticular nature of the material cannot be fully exploited.

In some other works, for instance V. L. Colvin, M. C. Schiamp & A. P. Alivisatos Nature 370,6488 (1994) "Light-emitting-diodes made from cadmium diselenide nanocrystals and a semiconducting polymer, the use of CdSe nanocrystals as a form of transport iayer (deposited neat either by spin-coating or electrostatic self- assembly) in a multilayered device with organic light-emitting polymers has also been described. Another reference relating to nanoparticle polymer composites is J.

Schmitt, G. Decher, W. J. Dressick, S. L. Brandow, R. E. Geer, R. Shashidhar, and J.

M. Caivert,"Metal nanoparticle/polymer superlattice films : fabrication and control of layer structure,"Adv. Mater., vol. 9, pp. 61,1997.

Organic materials are used for a wide range of applications, including the formation of light emissive devices (see PCTIW090/13148 and US 4,539, 507, the contents of both of which are incorporated herein by reference). There is often a need to tune the properties of such an organic material. For example, in the manufacture of optoelectronic devices there is a need for control over various properties of the materials to be used, including conductivity (and/or mobility), refractive index, bandgap and morphology. Some examples of known techniques for tuning various properties are as follows : 1. Conductivitv. This has been tuned by adding a chemical compound that acts as a donor or acceptor (namely an electronic dopant), see C. K. Chiang, C. R. Fincher, Y.

W. Park, A. J. Heeger, H. Shirakawa, E. J. Louis, S. C. Gau, and A. G. MacDiarmid "Electrical Conductivity in Doped Polyacetylene,"Phys. Rev. Lett., vol. 39, pp. 1098- 1101,1977.

2. Charge generation and photo-voltaic response. This has been tuned by blending two materials with appropriate electronic levels so that electrons prefer to reside on one and holes on the other. The blends have been either of two organic materials such as polymers (J. J. M. Halls, C. A. Walsh, N. C. Greenham, E. A. Marseglia, R.

H. Friend, S. C. Moratti, and A. B. Holmes, Nature 376,498 (1995),"Efficient photodiodes from interpenetrating polymer networks ") or of organic material with a nano-particle (N. C. Greenham, X. G. Peng, and A. P. Alivisatos, "Charge separation and transport in conjugated-polymer/semiconductor-nanocrystal composites studied by photoluminescence quenching and photoconductivity,"Phys. Rev. B - Cond.

Matt., vol. 54, pp. 17628-17637,1996) to achieve exciton dissociation at the interface.

3. Band-qap and emission colour. This has been tuned by mixing organic compounds in the form of blends or co-polymers (see co-pending UK patent application number 9805476. 0).

4. Scatterinq. Highly aggregated or very large size particles have been blended into polymers to take advantage of multiple internal light-scattering so as to increase the effective length the light travels within the polymer and hence, enhance amplified stimulated emission processes (F. Hide, B. J. Schwartz, M. A. Diazgarcia, and A. J. Heeger, "Laser-emission from solutions and films containing semiconducting polymer and titanium-dioxide nanocrystals,"Chem. Phys. Lett., vol. 256, pp. 424-430, 1996).

Scattering might have also been used to enhance external efficiency of LEDs (S Carter, J C Scott and P J Brock, Appl. Phys. Lett, 71,1997, 1145-1147).

According to a first aspect of the present invention there is provided a method for preparing nanoparticles for use, from a mixture of nanoparticles with another material, the method comprising washing the mixture with a solvent in which the nanoparticles are soluble to remove the said other material and form a solution of nanoparticles in the solvent. The solvent is preferably one in which the said other material is not soluble. Alternatively, the solvent may be one in which the other material is soluble but the nanoparticles are not.

The method preferably comprises separating at least a first fraction of the nanoparticles from a mixture of the solvent and the said other material. These separated nanoparticles may then be used, for example in the applications described below. These nanoparticles are preferably only weakly bound (e. g. unaggregated or only weakly aggregated), and not strongly bound, so that they suitably exist in a disaggregated state. This can assist in subsequent processing steps, such as forming a substantially uniform dispersion of the nanoparticles in another material.

The solvent is preferably one that is capable of holding the dissolved nanoparticles in a disaggregated state.

The method preferably includes a step of maintaining the pH of the solvent at a predetermined level. This suitably maintains a charge on the nanoparticles. An acid or a base and/or a suitable buffer may be added to the solvent to maintain the desired pH.

The method preferably includes dialysis through a suitable membrane to remove soluble low molecular weight material (for example surfactant molecules) from the nanoparticle solution. Continuous or intermittent sonification could be performed during dialysis.

The separation may be performed by filtration and/or dialysis and/or centrifugation.

Preferably the separation step or another step of the method also allows for the separation of the said first fraction of the nanoparticles from another fraction of the nanoparticies. For example, the nanoparticles of the said other fraction may be a set of particles that are relatively small in comparison to the nanoparticies of the first fraction. Thus, the separation step may also serve to narrow the size distribution of the retained nanoparticles.

The said other material may be a by-product of the formation of the nanoparticles, (examples include reaction products from the formation of the nanoparticles) and materials used to maintain process conditions during the formation of the nanoparticles, such as surfactants which could have been used in a microemulsion process for forming the nanoparticles.

The solvent may be an organic or an inorganic solvent. The solvent may be a polar solvent such as water or methanol (which may improve the solubility of the nanoparticles with polar surfaces) or a non-polar solvent. The solvent may be a polar non-hydrogen bonding solvent. It is preferred that the said other material is soluble in the solvent, and most preferably that the solvent is one in which the said other material is preferentially soluble to the nanoparticles. This can assist in the separation of the components.

The size range of the nanoparticles is preferably, but not necessarily, within the range from 1nm to 100nm. Preferably all or substantially all of the nanoparticles are smaller than 50nm, 30nm or 10nm in diameter. Preferably all or substantially all of the nanoparticles are larger than 1 nm, 5nm or 1 Onm in diameter.

The method preferably comprises surface modifying the nanoparticies. This may suitably be achieved by adsorbing a material to the surface of the particles. The material may be added as surface modifying agent to the solution of nanoparticles.

The material may, for example, be a silylating agent or a dye or a chemical functional material. The material may promote specific interactions with other materials such as polymers. Alternatively, or additionally, the nanoparticles may already have a surface coating. This may be a coating of a surfactant.

The nanoparticles may be of metallic, semiconducting or insulating material.

Examples of suitable materials include inorganic oxides such as Si02, Ti02, AI203 or Zr02, or ternary or other binary inorganic materials such as BaSC>4, YbFs, ZnS or other organic materials, especially polymer materials, such as PTFE, poly- methyimethacrylate (PMMA) or polystyrene (PS « . The nanoparticles are preferably light transmissive and most preferably optically transparent. Therefore, the material of which the nanoparticles are formed is preferably a wide optical bandgap material.

The nanoparticles may have been formed by any suitable route. Examples include the microemulsion route and the sol-gel route.

A further step in the processing of the nanoparticies is preferably to incorporate them into a body of material. To achieve this the material of the body, or a precursor of it, is preferably added to the solution of nanoparticles. A uniform (or substantially uniform) non-aggregated (or substantially non-aggregated) dispersion of the nanoparticles in the final body is achieved by ensuring that they are held in a substantially disaggregated state until fixed in place in the body, e. g. by removal of the solvent by a step such as evaporation. it is therefore preferable that the material of the body (or its precursor) is soluble in the solvent in which the nanoparticles are dissolved, and does not have undesirable interactions with the nanoparticles that may lead to severe aggregation or phase separation.

The material of the body could be (but need not be) an organic material. Examples are polymers, oligomers and materials of small organic molecules. If the material is a polymer material it may be a conjugated polymer such as poly (p-phenylenevinylene) (PPV). Alternatives include poly (2-methoxy-5- (2'-ethyl) hexyloxyphenyiene-vinylene) ("MEH-PPV"), a PPV-derivative (e. g. a di-alkoxy or di-alkyl derivative), a polyfluorene and/or a co-polymer incorporating polyfluorene segments, PPVs and/or related co- polymers, poly (2,7- (9, 9-di-n-octylfluorene) - (1,4-phenylene- ( (4-secbutylphenyl) imino) -<BR> 1, 4-phenylene)) ("TFB"), poly (2,7- (9, 9-di-n-octylfiuorene) - (1, 4-phenylene- ( (4-<BR> methylphenyl) imino) -1, 4-phenylene- ( (4 - methylphenyl) imino) - 1, 4-phenylene)) ("PFM"), poly (2,7 - (9,9 - di-n-octylfluorene) - (1, 4-phenylene- ( (4- methoxyphenyl) imino) -1,4-phenylene- ( (4-methoxyphenyl) imino) -1, 4-phenyiene)) ("PFMO"), F8 or F8BT. Alternative materials include organic molecular materials such as Alq3. The material is suitably light-transmissive and/or light-emissive.

The presence of the nanoparticles (including optionally any material attached to the surface of the nanoparticles) in the material of the body preferably influences at least one material property of the material of the body. This could be an optical property such as refractive index or an electrical property such as conductivity. Thus the nanoparticles could be dispersed in the material of the body to tailor its refractive index - either increasing or decreasing it depending on the relative refractive indices of the material of the body and the nanoparticles. The presence of the nanoparticles, and the interaction between the nanoparticles and the polymer (including optionally any material attached to the surface of the nanoparticles) in the material of the body could also influence the morphology of the material of the body, for instance by inhibiting crystallisation.

The volume fraction of the nanoparticles in the body is preferably greater than 1,5 or 10 volume %. The volume fraction of the nanoparticies in the body is preferably less than 50,30 or 30 volume %. The density of the nanoparticle distribution in the body is preferably greater than 10-17 and/or less than 10-'91cl2.

It is preferred that the nanoparticles exist in the body in a disaggregated state. This suitably promotes uniformity of the properties in the body. Furthermore, in some circumstances aggregates of the particles could scatter incident light, and it is preferred that the particles are of a sufficiently small size and disaggregated nature that they substantially do not scatter incident light.

The body could be a layer of a device such as an electronic and/or optical device.

Preferred non-limiting examples of such devices are as follows : 1. A device comprising a stack of layers defining a light-reflective structure, with at least one of the layers comprising a dispersion of nanoparticles as described above.

Preferably alternating layers of the device comprise a dispersion of nanoparticles as described above. This device is suitably a distributed Bragg reflector.

2. A light emissive device in which a light emissive layer or a layer adjacent to the light emissive layer comprises a dispersion of nanoparticles as described above. In this case the nanoparticles may carry a fluorescent dye (suitably as a surface layer).

This dye can suitably be stimulated to fluoresce by energy transfer or light emission from the emissive layer; thus the dye can act to modify the colour of light emission from the emissive layer. The device may also comprise a waveguide structure defined by a relatively high refractive index layer located between two relatively low refractive index layers. One of those three layers preferably comprises a dispersion of nanoparticles as described above which suitably modifies its refractive index to help define the waveguide structure. The waveguide is preferably located outside and/or separately from and/or independent of the light-emissive region of the device (the energy level profile of the device may be arranged to encourage light emission other than in the waveguide). This can permit independent tuning of the material properties of the emissive layer and the waveguide layer. The device may have a pair of mirrors located on either side of it - either mirrors of the type described for device 1 above or mirrors of another type such as cleavage surfaces. These mirrors may define a microcavity which can spectrally redistribute light generated by the device. The device may be a laser, for instance a microcavity or waveguide laser. A mirror could be provided by a DBR grating superimposed on the waveguide structure (on a substrate or any subsequent layer).

Other aspects of the present invention include any or all of the articles described above. For example, according to a second aspect of the present invention there is provided an optical and/or electronic device including any of the features described above. According to a third aspect of the invention there is provided a method for forming an optical and/or electronic device including any of the features described above. According to a fourth aspect of the invention there is provided a solution of nanoparticles in other than a strongly bound state, suitably including any of the features described above. According to a fifth aspect of the invention there is provided a solution of a polymer material (or a polymer precursor material) and nanoparticles in other than a strongly bound state. According to a sixth aspect of the invention there is provided an organic material containing a substantially uniform dispersion of nanoparticles ; preferably the organic material is a semiconductive and/or a polymer material. According to a seventh aspect of the invention there is provided a method for tailoring at least one property of an organic material, the method comprising forming a substantially uniform dispersion of nanoparticles in the material.

The present invention will now be described by way of example with reference to the accompanying drawings, in which: figure 1 shows the results of energy dispersive X-ray spectrometry of Ti02 nanoparticles ; figure 2 shows the results of Raman spectroscopy of TiO2 nanoparticles ; figure 3 shows the results of attenuated total reflection infrared spectroscopy of'fi02 nanopaticles ; figure 4 shows the UV-Vis spectrum of TiOz nanoparticles ; figure 5 shows the results of transmission electron microscopy of TiO2 nanoparticles deposited as a film ; figure 6 shows an electron diffraction pattern for TiO2 nanoparticles; figure 7 shows the results of energy dispersive X-ray spectrometry of SiO2 nanoparticles ; figure 8 shows the results of Raman spectroscopy of Si02 nanoparticles ; figure 9 shows the results of attenuated total reflection infrared spectroscopy of SiO2 nanoparticles ; figure 10 shows the UV-Vis spectrum of SiO2 nanoparticies ; figure 11 shows the results of transmission electron microscopy of SiOz nanoparticles deposited as a film ; figure 12 shows an electron diffraction pattern for Si02 nanoparticles ; figure 13 plots refractive index measurements for PPV: Ti02 thin films; figure 14 illustrates the structure of a distributed Bragg reflector; figures 15 and 16 each show reflection spectra for two distributed Bragg reflectors showing that the reflection peak can be tuned; figure 17a shows the structure of an edge emitting organic light emitting device; figure 17b shows the refractive indices (n) and energy levels (E) through the device of figure 17a; figure 18 shows difference UV-Vis spectra for PPV containing dye-loaded nanoparticles ; figure 19a shows the result of atomic force microscopy of a PPV film ; figure 19b shows the result of atomic force microscopy of a PPV: Tir2 film with<BR> 30% vol. Ti02 ; and figure 20 shows UV-Vis spectra for PPV : TiO2 films with differing amounts of Ti02 nanoparticles.

Processes for the formation of nanoparticles, together with some exemplary applications for those nanoparticles (or for other suitable nanoparticles), will now be described. The processes allow for the formation of nanoparticles in a sufficiently non-aggregated state that they can be dispersed relatively evenly into a matrix such a polymer body. In that state the nanoparticles can be employed to tailor the properties of the matrix in desired ways.

Processes for the synthesis and isolation of Ti02 and Si02 nanoparticles will be described. These processes can readily be extended to the synthesis and isolation of nanoparticles of other materials, including metallic and semiconductive materials.

Svnthesis of TiO) NanoDarticles As a first step, a microemulsion of H20+NH3/AOTlcyclohexane was prepared. To prepare the microemulsion 17. 2g of finely divided sodium dioctylsulfosuccinate (AOT) (39mmol, Aldrich) was dried at 120°C in a vacuum oven for 2h to remove adsorbed moisture (3% weight loss), and then added to 200ml of HPLC grade cyclohexane (Aldrich) in a three-necked round bottom flask quipped with a magnetic stirrer to give a 0.2M AOT solution. Then 3. 8ml of concentrated aqueous NH3 (35w/v%, Aldrich) was added to this solution to give a H20: NH3: AOT molecular ratio of 1. 8. 3. 2 : 1.0 and the solution was cooled to 5°C under N2 flow in an ice-bath to give an optically clear single-phase microemulsion. The total molar ratio of H20+NH3 : AOT (5.0 : 1.0) fixed the radius of water cores in the emulsion at around 25A (F. J. Arriagada and K. Osseo-Asare, "Synthesis of nanosize silica in Aerosol OT reverse microemulsions", Journal of Colloid and interface Science 170 (1995), pp. 8-17).

To prepare a Ti02 precursor solution, 2 0ml TiC14 (18mmol, Aldrich) was added by syringe into 10ml HPLC grade cyclohexane (previously dried for 18h over 3A molecular sieves, Aldrich) in a vial equipped with a rubber septum. This Ti02 precursor solution was added in two portions to the AOT solution with vigorous stirring. Rapid evolution of HCI gas was observed together with the formation of a cloudy suspension of NH4CI in the reaction medium, according to the following equations: TiCI4 + H20 -> Ti02 + 4HCI HCI + NH3 < NH4CI HCI + NaAOT o NaCI + HAOT After half an hour, the bath temperature was raised to and then maintained at 80°C for 3h to complete the reaction, and finally cooled to room temperature overnight.

The amount of H20 in the reaction medium (120mmol) was well in excess of the amount needed to hydrolyse the TiCU. However, the amount of NH3 was not sufficient to neutralise all the HCI produced, so the reaction medium became acidic and a positive surface charge developed on the Ti02 nanoparticles. (The iso-electric point of bulk Ti02 is about pH 4 to 5).

Isoiation of TiO NanoDarticles The cyclohexane solvent was slowly evaporated away to give a cloudy, viscous gel containing the Ti02 and NH4CI by-product. 20ml HPLC-grade CHCI3 (Aldrich) was added to this gel to precipitate out the NH4CI and NaCI which could then be readily separated from the clear supernatant by centrifugation. Electron beam elemental analysis by energy dispersive x-ray spectrometry (EDX) of the CHC13 supernatant gave no detectable Cl (the height of the Cl Ka peak was less than 1 % of the height of the Ti Ka peak), indicating it was essentially free from contamination by the chlorides.

The CHC13 supernatant was then concentrated in vacuo to give a viscous gel which was then taken up in 25ml MeOH to give the crude TiO2-containing liquor.

To purify the TiOz further from excess AOT, 1 Oml of the crude liquor was placed in a 6-8k MWCO cellulose dialysis tubing (Spectra/Por3) and dialyse against twice-daily refreshed 200ml quantities of HPLC grade MeOH each containing 0. 3ml concentrated aqueous HCI. The Ti, S and Cl content of both the dialysate and retentate were monitored daily by EDX, and the dialysis was stopped when the Ti and S signals in the dialysate fell below that of Cl (which was treated as indicating a background level). This dialysis, in addition to removing the excess AOT contamination, also improved the size distribution of the particles by removing the smallest particles, which could readily pass through the pores of the dialysis tubing.

In this way, an optically clear dispersion of about 0.9w/v% Ti02 in a MeOH solution of pH=2 was obtained. The Ti02 nanoparticles at this point were para-crystalline and coated by a monolayer of surfactant derivative.

Characterisation of TiO NanoDarticiles Electron probe energy dispersive x-ray spectrometry (see figure 1) shows the final product to be pure Ti02 with a surface coating of AOT and Cl. It is believed that the AOT (which is the origin of the S signal) is strongly absorbed onto the TiO2 surface while Cl is believed to have existed in the form of Cl- acting as a weakly-bound counterion for the positively-charged TiO2 surface. This coating appears to be stable against further dialysis. Raman spectroscopy (see figure 2) confirms the structure as being that of Ti02, with broad vibration peaks at 170cm~', 440cm~'and 610cm~'.

Attenuated total reflection Fourier-transformed infrared spectroscopy (see figure 3) provides further confirmation of structure. Ultraviolet-visible spectrometry (see figure 4) shows the product to be non-scattering in the visible, and provides evidence for the highly dispersed (non-aggregated) nature of the Ti02 nanoparticies achieved via the above described synthesis and isolation. Transmission electron microscopy (see figure 5) shows that a particle size of 2-8nm has been achieved, while electron beam diffraction (see figure 6) gives the expected d-spacings for Ti02.

Synthesis of SiO2 nanoDarticles An H2o+NH3lAoTlcyclohexane phase was prepared as described above.

To prepare a Si02 precursor sokution, 2 1ml SiC14 (18mmol, Aldrich) was added by syringe into 10ml HPLC grade cyclohexane (previously dried for 18h over 3A molecular sieves, Aldrich) in a vial equipped with a rubber septum. This Si02 precursor solution was added in two portions to the AOT solution with vigorous stirring. Rapid evolution of HCI gas was observed together with the formation of a cloudy suspension of NH4CI in the reaction medium, according to the following equations: SiCI4 + H20 -> SiO2 + 4HCI HCI + NH3 -4 NH4CI HCI + NaAOT < NaCI + HAOT After half an hour, the bath temperature was raised to and then maintained at 80°C for 3h to complete the reaction, and then cooled to room temperature overnight.

The amount of H20 in the reaction medium (120mmol) was well in excess of the amount needed to hydrolyse the Silt.

Isolation of SiO ? nanoparticles The cyclohexane solvent was slowly evaporated away to give a cloudy, viscous gel containing the SiO2 and NH4CI (together with NaCI) by-product. Then 20ml HPLC- grade CHCI3 (Aldrich) was added to this gel to precipitate out the Si02 together with NH4CI and NaCI by-product as a slurry which could then be readily separated from the clear supernatant which contained mainly excess AOT by centrifugation. EDX of the CHC13 supernatant gave no detectable Si (the height of the Si Ka peak was less than 1 % of the height of the S Ka peak), indicating that it was essentially free from Si02. The CHCI. , supernatant was then removed. and the slurry washed (with sonication) in several 25ml lots of HPLC-grade MeOH to remove as much NH4CI and NaCi by-product as possible. After each washing the amount ouf Si in solution was checked. After three washings the centrifugate containing the SiO2 was then taken up in 30ml MeOH, to give the crude Si02 as a translucent gel.

To purify the crude Si02 further from the inorganic chlorides, 10ml of the gel was placed in a 6-8k MWCO cellulose dialysis tubing (Spectra/Por (E) ) and dialyse against two change of 200mL HPLC grade MeOH containing 0. 3ml concentrated aqueous HCI. The retentate was then sonicated for 30min to break up the weak Si02 aggregates and then dialysis continued with twice-daily refreshed MeOH+HCI. The appearance of the retentate became clear after a while. The Si, S and Cl content of the dialysate was monitored daily, and the dialysis was stopped when the Cl signal fell to and stabilise at about 1.7 times that of S (from the AOT) which acted as a background. Sonication could be performed during dialysis. This dialysis, in addition to removing the chloride and AOT contamination, also improved the size distribution of the particles by cutting off the smallest particles, which could readily pass through the pores of the dialysis tubing. The dialysis took three to five days. A further dialysis step, with a larger molecular weight cut-off tube, could be used to separate off the largest particles and thus give an even tighter size distribution. <BR> <BR> <BR> <BR> <BR> <P>In this way, an almost optically clear dispersion of SiO2 of about 1. 1w/v% in a MeOH solution of pH=2 was obtained.

Characterisation of SiO7 nanoDarticles Electron probe energy dispersive x-ray spectrometry (see figure 7) shows the final product to be pure Si02 with essentially no surface coating. Raman spectroscopy (see figure 8) confirms the structure as being that of Si02 with broad vibration peaks at 490cm', and 830cm-'. Attenuated total reflection Fourier-transformed infrared spectroscopy (see figure 9) provides further confirmation of structure. Ultraviolet- visible spectrometry (see figure 10) shows the product to be essentially non- scattering in the visible, providing evidence for the highly-dispersed (non-aggregated) nature of the SiO nanoparticies achieved in the described synthesis and isolation. Transmission electron microscopy (see figure 11) shows that a particle size range of 2-8nm was been achieved, while electron beam diffraction (see figure 12) gives the expected d-spacings for Si02.

General Nanoparticle Considerations The nanoparticle synthesis process can be suited to formation of nanoparticles of a desired type by selection of the materials involved in the process, for instance the precursor materials and surfactant, and optimisation of chemical parameters such as oil composition, temperature, water-surfactant ratio, surfactant-oil ratio, co-surfactant composition and reactant ratio. Nucleation and growth of the particles could be monitored by TEM and/or UV-Vis spectroscopy to ensure suitable reaction rates and size distribution are achieved.

It is preferable that the steps be performed in a controlled atmosphere to minimise undesirable reaction in the nanoparticle solution. The temperature of the nanoparticle formation step could be fixed to favour the desired balance between nucleation and growth rates to give a preferred particle size distribution. It will be noted that the emulsion conditions described above have been chosen to yield especially small water cores.

The synthesis processes described above make use of the microemulsion route, which has been found generally to offer good control of particle nucleation, growth, aggregation, size distribution, and surface capping that could be judiciously exploited to tailor specific material and processing properties. As an alternative the nanoparticles could be synthesised by other routes, such as the sol-gel route based on direct hydrolysis in water or alcohols, (see for example T. Moritz, J. Reiss, K. Diesner, D. Su, and A. Chemseddine,"Journal of Physical Chemistry B 101 (1997) pp. 8052-8053) and then isolated in a non-strongly aggregated state generally as described above.

Applications for the Nanooarticles Some non-limiting examples of applications of the nanoparticies as prepared above will now be described. Nanoparticies prepared by other routes could be used in these applications, provided they were of the appropriate size and composition and had a level of disaggregation sufficient to allow them to be dispersed relatively uniformly in the polymer matrices.

Two general modes have been identified by which the nanoparticles can provide an effect in a host matrix material.

1. Non-interferina mode. The particles may be selected and/or treated so that they do not significantly alter the intrinsic properties of the host material. For most applications it would be preferred that such particles are of a size below around 20nm (to allow them to be easily dispersed uniformly in the host) and above 5nm (suitably to avoid them acting like molecules and possibly interfering with the host structure - for example by chain separation in a polymer matrix).

2. Interfering mode. The particles may be selected and/or treated so that they do interact significantly with the host material to affect its intrinsic properties. This may occur if: a. The particles themselves or especially their attached surface functions are of a material that interacts with the host material, through chemical or physical interactions. The particles may carry an additive agent that interacts with the host material, for example to change the chemical properties of the host (e. g. electrical/optical doping) and/or its physical properties (e. g. polarisation or alignment or energy transfer). The additive agent may be bound to the surface of the particles or at least partially included in the particles. b. The particles are sufficiently small and numerous that they significantly perturb the arrangement of the molecules in the body - for instance to affect the morphological structure of a polymer host (e. g. by perturbing the chain trajectories). Such particles could also change the environment of molecules of the host material (e. g. polymer chains) thus affecting many of its properties (band-gap, electron affinity, exciton binding, radiative lifetime etc.).

Examples 1,2 and 3 below use nanoparticles predominantly in the non-interfering mode. Examples 4 and 5 use nanoparticles in the interfering mode.

Example 1 : Tunina of Polymer Film Refractive Index In this example the refractive index of films of PPV is tuned by the addition to the film of a dispersion of Ti02 nanoparticles. Since the nanoparticles are light transmissive and sufficiently small to avoid scattering of incident light the resulting film has good optical properties.

Four samples of PPV containing varying amounts of Ti02 (referred to herein as <BR> <BR> <BR> <BR> "PPV: Ti02") were fabricated by blending the appropriate volume of the 1.4w/v% precursor PPV-MeOH solution (Cambridge Display Technology) with a 1.8w/v% Ti02-MeOH solution to give the same final concentration of precursor PPV but different concentrations of Ti02. The size range of the Ti02 particles was approximately 2-8nm. A nominally 1000A film was obtained by spin-coating on to a glass substrate and then thermally-converted at 180°C under dynamic vacuum (less than 10-5 mbar) for 8h. The actual Ti02 content in the film was measured by a combination of electron probe x-ray spectrometry (for Ti02 content) and visible spectrometry (for PPV content) and the film thickness was measured by surface profilometer. Film PPV: MeOH: TiO2 Measured TiO2 Measured Refractive Vol. Ratio Vol% in Final Film Index at 633nm A 3 : 4: 0 0 2. 60 B 3 : 1: 3~~\5'2"37 ~C3: 2: : 2 | 30 í 1. 94 D 1 3 0 4 | 50 1 1. 77 ! The refractive indices of the films over the whole visible window were then measured by spectroscopic eliipsometry (see figure 13). Transmission spectrometry shows the resultant films to be highly non-scattering (scatter at 633nm was estimated at less than 2%). These results demonstrate that high optical-quality organic films can be obtained using highly-dispersed nanoparticles as a refractive index modifier. <BR> <BR> <P>Similar results were obtained for samples of PPV blended with Si02 (PPV : Si02). In the films detailed in the following table, the precursor PPV and Si02 concentrations were both 1. 1w/v%. Nominally 1000A films were obtained by repeated spin-coating and conversion on glass substrates and then thermally-annealing at 160°C under dynamic vacuum (less than 10-5 mbar) for 8h. The other experimental procedures were similar to those given earlier. Film PPV:MeOH : SiO2 Measured SiO2 Measured Refractive Vol. Ratio Vol% in Final Film Index at 633nm E 3 : 16.5 : 0"02. 39 F 3 : 15.5 : 1 6 2. 18 G 3 : 14.5 : 2 10 2. 10 H 5 : 12: 4 19 2. 05 To further check the effect on refractive index, films of similar thickness were placed between two silver mirrors to form a Fabry-Perot cavity. Since the resonance frequency of the cavity is a function of the optical length it serves as a measure for the effective refractive index.

The precursor PPV polycation and the negatively-charged SiO2 nanoparticles interact to give a precipitate when mixed together. However, concentration conditions could be selected to overcome this problem (with low PPV and SiO2 concentrations relative to the MeOH). and any slight precipitation could be removed by centrifugation. The optically-clear centrifugate could then be used to fabricate films with almost no scattering.

Photoluminescence measurements indicate that high photoluminescence efficiency is retained in the PPV: Si02 films, in contrast to PPV: Ti02 films since, unlike TiO2 particles, Si02 particles do not quench the PPV excited states. PPV: Si02 films could therefore be preferred for use as emissive layers of electroluminescent devices.

(Alternatively, quenching nanoparticles such as Ti02 could be treated to carry a surface isolation or spacing layer of a material such as alkyl chains or Si02). The basic structure of such devices is well known and generally comprises a light- emissive organic layer, for instance a film of PPV, sandwiched between two electrodes. One of the electrodes (the cathode) injects negative charge carriers (electrons) and the other electrode (the anode) injects positive charge carriers (holes). The electrons and holes recombine in the organic layer generating photons.

In PCT/W090/13148 the organic light emissive material is a polymer. In US 4,539, 507 the organic light emissive material is of the class known as small molecule materials, such as tris- (8-hydroxyquinolino) aluminium ("Alq3"). In a practical device, one of the electrodes is typically transparent, to allow photons to escape the device.

This technique for varying refractive index could be used for matrices of other materials, especially other organic materials. The organic materials could be polymers, oligomers or small organic materials or blends of two or more such materials. As illustrated above the refractive index of the resulting material could be fixed by appropriate choice of the material of the nanoparticies and their volume fraction in the matrix. The effective refractive index could be estimated by the Bruggeman effective medium approximation. The refractive index is thus adjustable depending on the refractive index of the organic matrix and the nanoparticle, the loading volume fraction, and the dispersion morphology.

A prior approach to adjusting the refractive index of organic materials has been to blend two or more organic materials together. However, the effectiveness of this approach has been limited because organic materials tend to have very similar refractive indices, and are often incompatible, leading to large-scale phase separation.

Example 2: Polvmer distributed Braaq reflectors (Polymer mirrors) A distributed Bragg reflector (DBR) consists of a stack of regularly alternating higher- and lower-refractive index dielectrics (light transmissive materials) fabricated to fulfil the Bragg condition for reflection at the design wavelength. This occurs when the optical path of the periodicity in the dielectric stack corresponds to half a wavelength, and the reflectivity is further optimised when the DBR stack obeys the following equation: 1/2X=n, d1+n2d2, where n1, n2 are the respective refractive indices; du, dus are the corresponding component film thicknesses in the DBR; and A, is the design wavelength.

Figure 14 shows a polymer distributed Bragg reflector (polymer mirror) in which the altemate layers are formed of PPV and PPV modified by the dispersion of Ti02 nanoparticles to tailor its refractive index. To form the polymer DBR a 0.4w/v% precursor PPV-MeOH solution (Cambridge Display Technology) containing 70ppm <BR> <BR> <BR> AOT as surface tension modifier and a 0. 7w/v% (total solids) precursor PPV: Ti02- MeOH solution (in which the ratio of precursor PPV to TiO2 is 1: 1.8 by weight) were alternately spun on glass and thermally converted at 180°C under Ar for 1h to gave 600-700A thick films of each material. The PPV material 10 forms the higher refractive index layers while the PPV : TiOz 11 forms the lower refractive index layers.

The exact spin conditions which determine the thicknesses d, and d2 were set after measurement of the reflection spectra upon the completion of each pair of high-and low-refractive index layers. Feedback is thus provided to allow a DBR to be engineered with the desired value of n, d, +n2d2.

The spin coating could be performed on to a substrate of, for example, glass or plastic, which could in addition have a transparent conducting layer for charge injection.

Reflection spectra for three- and six-pair polymer layers with a first order reflection maximum at about 600nm (see figure 15) and 550nm (see figure 16) show the high level of control that can be achieved.

Since the DBR is formed of conjugated material it could be electrically-pumped to generate photons in addition to reflecting.

Example 3: Separate Confinement Heterostructures: Edqe-emittina Oraanic Liaht Emittinq Diodes The ability to vary the refractive index of a material by dispersing nanoparticles in it provides an important enabling technology for the fabrication of photonic structures.

An example is a photonic structure that makes use of refractive index contrasts (variations) to confine optical photon modes, for example to form waveguide structures and/or separate confinement heterostructures (see for example: S. M. Sze,"Semiconductor Devices: Physics and Technology,"John Wiley & Sons, New York, 1985). The inability to conveniently vary refractive index has hitherto been a major obstacle to the full exploitation of organic materials in optoelectronic device technologies.

Figure 17a shows an edge-emitting organic light-emitting diode (EEOLED) that emits in the plane of the LED rather than vertically through it as in conventional surface emitting LEDs.

The EEOLED is fabricated by spin coating to give a multilayer structure as shown in figure 17a. There is a 3500A thick iow-refractive index layer of PPV: Ti02 12, followed by a 2500A thick high-refractive index wave-guiding layer of PPV 13, and then a 1000A thick emitter layer of MCP 14 which is fabricated with a number of thin charge-transport"hurdle"layers of PPV: Si02 14a buried within the MCP layer. These PPV : SiO2 thin layers provide a refractive index almost matching that of the adjacent MCP layers, but impose a weak electron and hole transport"hurdle"which can be exploited to enhance and direct recombination to occur largely in the MCP layer.

Hence, during operation of the device, excitons are produced in this layer. Light- emitted from this layer is then wave-guided along the low-refractive index layer 12, which confines the photons, and emitted along the edge of the polymer film (see figure 17b). The excitons and photons are thus confined to separate regions of the device. In such a device the material properties of the region to which the excitons are confined can be tailored separately from those of the region to which the photons are confined: charge response and optical response properties of the conjugated material can be separated and individually modified or engineered to produce the desired electronic and photonic structure.

One advantage of this structure is that since the light is emitted from the edge of the device rather than the side there is no need for one of the electrodes of the device to be transparent, as in prior devices. Another advantage is that this structure forms the architecture for a wave-guiding electrically-pumped injection laser.

Example 4: Nanoparticles as a fluorescent dve carrier By attaching functional materials to nanoparticles and dispersing those nanoparticles in a matrix the properties of the matrix can be adjusted. As an example of this technique, a device in which fluorescent die is attached to nanoparticles and dispersed in a PPV matrix will be described.

A (sub) monolayer of the well-known red laser dye Rhodamine 101 is first adsorbed on to SiO2 nanoparticles prepared as described above. To achieve this equal volumes (1ml) of a dilute 1fiv% solution of Rhodamine 101 (a Xanthene) in EtOH was mixed with 1. 1w/v% SiO2 in MeOH, and the homogeneous mixture dialysed against 200ml HPLC grade MeOH through a 6-8k MWCO cellulose dialysis tubing (Spectra/Por) twice. Rhodamine 101 molecules not bound to the surface of the Si02 particles were dialyse out with the dialysate. whereas bound dye molecules were retained on the surface of the dispersed nanoparticles in the retentate. The retentate was then analyse by ultraviolet-visible spectrometry (see figure 18) to illustrate the principle that the surfaces of nanoparticles could indeed be used as a dye carrier.

The nanoparticles prepared in this way could then be used as for the nanoparticles of example 1 to disperse the fluorescent dye into a layer of PPV. This layer could form the active layer of an organic light emitting device, in which the fluorescent dye could be stimulated by emission from the PPV to emit light at another frequency. This could be particularly significant in the manufacture of tri-colour displays.

Further examples of the surface modification of nanoparticle surfaces with fluorescent dye molecules are given in B. Oregan and M. Gratzel, "A low-cost, high efficiency solar-cell based on dye-sensitised colioidal TiO2 films, "Nature 353 (1991) pp. 737-740. The surface modification could be preceded by surface modification by a suitable chemical, such as a silylation reagent, to cap the nanoparticles'surfaces with a layer carrying appropriate chemical functional groups (such as -COOH, -NH2, - SO3H, -PO3H or -NR3+) on which covalent or ionic or hydrogen bonds could be formed with an appropriate functional group on the dye molecule. This could help the dye to anchor to the surface of the nanoparticles. This may require modification of the dye molecules themselves to produce the desired functional group.

A wide range of fluorescent dye molecules are available, to provide sensitised emission over a wide range of frequencies, especially when used as a"guest"in a "guest-host"active layer structure where the host is a light emissive material which can stimulate the guest. In these structures, the matrix "host"layer, which could be of an organic material, performs the role of charge transport and acts as first-stage recombination centres, but subsequently the energy is to be transferred to the"guest" which then emits the desired colour. It is preferred that the energy transfer from the initial excitation created in the matrix to the final excitation to be created in the dye should be efficient, which in general demands that the emission spectrum of the initial ("host") excited state should overlap significantly with the absorption spectrum to the final ("guest") excited state. It is also preferred that the guest should exhibit a high photoluminescent efficiency : and that the density of the"guest"sites in the "host"matrix should be sufficiently high (preferably 10-17 to 10-'9/cl3) if the effective radius of energy transfer (the Forster radius) is small, e. g. about 3-10nm.

Example of suitable dyes may include the class of molecules widely known as the laser dyes (such as coumarins, xanthenes and oxazines).

In the past a major stumbling block to the incorporation of dye molecules in matrices such as organic materials has been rapid diffusion and segregation (recrystallisation) of the dye molecules in the host layer or at the interfaces during storage, and especially during operation, so that emission efficiency falls rapidly. This problem can be addressed by the process described above of immobilising the dye by binding it to the surface of a nanoparticle which is relatively immobile in the matrix. Even with a diameter of 1-10nm, the diffusion coefficient of a nanoparticle through a typical organic host could be expected to be several orders of magnitude lower than that of the dye molecules themselves.

Example 5: Nanoparticles as a PolymerMorphology Modifier This example demonstrates that nanoparticles that are intimately dispersed in a polymer matrix can modify the morphology (e. g. the chain conformation) of the polymer. <BR> <BR> <BR> <BR> <BR> <BR> <BR> <P>Polymer films of PPV: Ti02 (having the same composition as the film C described above) and PPV (as control) were spun and subjected to atomic force microscopy examination. Both films were thermally converted at 180°C. Figures 19a and 19b show the results of the examination. The surface of the control film exhibits a number of nanoscale protruding domains (about 10-20nm high and 20-40nm wide) that are believed to be related to micro-crystallisation of the PPV chains in the film thickness direction (M. A. Masse, D. C. Martin, E. L. Thomas and F. E. Karasc, "Crystal morphology in pristine and doped films of poly(p-phenylene vinylene)," Journal of Materials Science 25 (1990), pp. 311-320). In contrast, films fabricated with the nanoparticies showed a relatively smooth featureless surface over the same length scale. This is believed to be due to the suppression of micro-crystallisation by the nanoparticles, probably by steric constraints and the introduction of kinks in the chain conformations in the interfacial layer next to the nanoparticle surface.

Further evidence for this belief is given by ultraviolet-visible spectrometry of a set of films with a range of PPV: Ti02 volume ratios, which showed a progressive blue shift of the Ti-Ti* transition with increasing nanoparticle content above 15% (see figure 20).

Thus, it appears that the presence of nanoparticles can induce perturbation of local polymer morphology, changing the effective conjugation length and/or crystallinity of the polymer matrix and thus its optical refractive index and other related properties.

This feature of organic/nanoparticle composites could be used to inhibit the recrystallisation of the organic during storage or in operation. Recrystallisation is generally undesirable as, taking optical devices as an example, it can cause changes in electrical behaviour, colour stability and luminous efficiency, with a deleterious impact on device performance. Altering morphology could also alter many"intrinsic" properties such as transport, binding energies (as for excitons) and efficiency.

General Issues Nanoparticles for the above applications and other applications could be formed of materials other than SiOs or TiO2. Depending on the application some properties of the particles that may have to be borne in mind are: 1. Transparency. To achieve this the particles would have to have a sufficiently large band-gap. For instance, where the nanoparticles are dispersed in a matrix of an emissive polymer and are required to be transparent to emissions from that polymer the particles should have a larger band gap than the polymer material. One possibility is to use transparent, non-metallic, inorganic particles. It is preferred that the optical bandgap of the nanoparticle material is equal to or larger than that of the matrix material, so that the absorption edge is not squeezed into the optical transparency window.

2. fnsolubilitv (e. g. in water). This could be assessed in the first approximation by behaviour towards water as detailed in the CRC Handbook.

3. Resistance to acid corrosion. This is especially relevant where the nanoparticles are to be used in an acid environment, e. g. in PPV blends. This property could be assessed in the first approximation by bulk behaviour towards inorganic acids as detailed in the CRC Handbook.

4. Non-mobile ion formina. A cation-exchange process could be used to remove unwanted ions such as Na from the preparation if they would otherwise interfere with performance.

5. Size. If uniform charge transport and high optical quality of the resulting body is a requirement then a preferred size range is about 5-1 Onm. Particles greatly smaller than this may be undesirable because of lowered stability due to their higher surface-to-volume ratio. For transparency of bodies containing the nanoparticies it is preferred that the nanoparticles'diameters are generally less then half of the relevant light wavelength. If the body is small, such as a typical thin film of an optical device, it is preferred that the nanoparticies'diameters are less than half the thickness of the film, most preferably less than one fifth or one tenth of the thickness of the film. For a uniform (statistical) distribution the size of the nanoparticies should be 1/5 (1/10) of the relevant size (wavelength or film thickness).

6. Dispersion. To improve dispersion of the nanoparticles, and reduce agglomeration, a surfactant coating could be provided on the surface of the nanoparticles. One example of a potentially suitable surfactant is sodium decussate (AOT). This is an ionic surfactant with short hydrophobic tails in the form of 2- ethylhexyloxy ester side-chains that may be expected to be compatible with and hence dispersible in MEH-PPV and a variety of polar casting solvents like MeOH and CHCI3.

7. Suitable refractive index.

8, Processable into homogenous blends with the matrix material. for example a desired emissive polymer.

Nanoparticles may be "bfended" with sublimed molecule materials by forming a dispersed layer of particles and than subfimina molecules on top of that layer. In this way the molecules (oligomers etc.) can fill in the gaps set by the particles. The formation of the first layer could be through spin coating or self assembly. This technique may also allow nanoparticies to be blended with materials that can be deposited by evaporation at high temperature as inorganic materials are deposited in a CVD (chemical vapour deposition) or MBE process or any variant or combination.

The present invention may include any feature or combination of features disclosed herein either implicitly or explicitly or any generalisation thereof, irrespective of whether it relates to the presently claimed invention. In view of the foregoing description it will be evident to a person skilled in the art that various modifications may be made within the scope of the invention.