This invention relates to semiconductor thin films, particularly but not exclusively for use in the construction of photo-sensitive devices such as photovoltaic cells. More particularly the invention concerns a process for the production of a semiconductor thin film structure, and preferably one incorporating an organic semiconductor.

The basic design of an organic photovoltaic (PV) cell consists of an organic layer sandwiched between two electrodes. In order to allow light into the cell, typically one electrode must be transparent and conductive, for example consisting of a thin film coating of indium tin oxide (ITO) or SnO ₂ on a glass substrate.

Early work was based on single molecular organic layers, typically made of phthalocyanines (Pc) or polyacenes, positioned between two electrodes. In the 1980s a device comprising a bilayer (planar) heterojunction was developed, using copper phthalocyanine (CuPc) and a perylene derivative, resulting in an order of magnitude improvement in power conversion efficiency. The discovery of C ₆ o in 1985 and its use in a CuPc/C ₆ o planar heterojunction device further increased the power conversion efficiency achievable.

The majority of organic PV devices operate by combining organic materials which have donor and acceptor properties and providing a heterojunction between two such organic layers, where one layer is an electron transporter (acceptor) and the other is a hole transporter (donor). In particular, known organic solar cells are based on thin films of organic semiconducting materials, such as phthalocyanines and fullerenes, or conjugated polymers and fullerenes. The donor-acceptor films are typically 100 nm in thickness.

Upon absorption of light into an organic PV device an exciton, i.e. a bound electron - hole pair, is generated. The electron and hole are bound together by electrostatic attraction and are strongly localised. The exciton is able to migrate or diffuse to a lower energy state. This exciton must reach a donor-acceptor interface in order to dissociate efficiently into free charge carriers. This dissociation is essential in solar cells such that when an exciton reaches an interface between the donor material and acceptor material, the electron of the electron hole pair (exciton) may be transferred to the acceptor material. The electron in the acceptor material is transported to the cathode, and the hole, remaining in the donor material, is transported to the anode. The diffusion length of an exciton is of the order of 10 to 50 nm; for example in copper phthalocyanine (CuPc) it has been found experimentally to be about 30 nm. Beyond this length the probability of the electron and hole recombining increases. It may therefore appear desirable to reduce the film thickness to less than 30 nm in order that the exciton reaches a donor-acceptor interface and dissociates. However, in order to absorb light efficiently and hence create excitons, film thicknesses of typically 100 nm are required.

To effect a compromise between the short diffusion length of excitons, and still achieve good light absorption, organic devices have been developed having mixed blend layers. A typical example of such a mixed blend device has a transparent electrode and a conductor electrode situated on opposite sides of a mixed blend layer made up of donor material and acceptor material. The donor and acceptor materials form a random distributed heteroj unction.

By increasing the number of interfaces between the donor and acceptor material, efficiency is increased despite the relatively short exciton diffusion length in these materials. Whereas this arrangement may achieve efficient dissociation, it is inefficient for subsequent free charge transport. This is because the electrons must be transported to the appropriate electrode by the acceptor material, and the holes must diffuse towards the other electrode in the donor material. However, the blended nature of the donor and acceptor materials leads to discontinuity of each of these respective materials, making charge transfer more difficult. Charges can be trapped in isolated domains, which reduces the total charge collection and therefore the overall power conversion efficiency of the device.

A further variation of such a mixed blend organic device has been developed, having multiple mixed blend layers. In such a multilayer device, the composite is arranged in planar layers which are composed of blended donor material and acceptor material. The composition of the layers is graded, the composition of the layer nearest one electrode being made of 100 percent acceptor material, the proportion of acceptor materials then decreasing to zero percent acceptor material and 100 percent donor material in the layer adjacent to the second electrode. This type of device is, however, extremely difficult to manufacture.

To overcome this limitation, a 3D corrugated interface structure has been proposed. Two electrodes are situated on opposite faces of an organic layer. The organic layer is made up of corrugated "fingers" of acceptor material and donor material. The maximum thickness of these fingers should be about two times the exciton diffusion length in that material. All excitons will thus be formed within the diffusion distance of the acceptor-donor material interface and charge transfer to the electrodes is efficient.

In order to improve the efficiency of organic solar cells, it has been proposed to increase yet further the interfacial area between the acceptor and donor phases, so as to minimise the exciton diffusion path, while having a structure thick enough to absorb light efficiently and ensuring that there is continuity between each phase and its respective electrode. By maximising the interfacial area, the exciton diffusion path is minimised. This leads to greater probability that the excitons dissociate at a heterojunction, resulting in a current being generated by the cell, rather than the

1 excitons recombining and the current being lost.

In WO 2008/029161 , the contents of which are incorporated herein by reference, there is disclosed a thin film structure for use in photovoltaic cells, comprising first and second continuous interpenetrating lattices of semiconductor materials acting as respective electron donor and acceptor materials. The creation of continuous interpenetrating lattices for each of the donor and acceptor materials is said to maximise the interfacial area between the materials and hence the exciton dissociation efficiency, minimising the exciton diffusion path.

In the arrangement disclosed in WO 2008/029161 , a substrate such as an ITO coated transparent glass electrode is coated with a layer of a first phase material which is either donor or acceptor material. The donor material can for example be a phthalocyanine, e.g. a metal phthalocyanine, such as copper phthalocyanine, and the acceptor material can for example be a fullerene or a perylene. To form an active layer for the device more than one layer of monodisperse particles such as - A -

colloidal spheres is then deposited on the layer of first phase material, for example by controlled self-assembly deposition from a colloid suspension. Typically, several layers of spheres are deposited, according to the device thickness required. The particles may be of any monodisperse, removable material, which is able to produce an ordered, hexagonally or cubic or other geometry close-packed structure. As this structure is ordered or structured and not random, each particle will be in contact with all of its neighbours, and hence be interconnected. The removable particles may for example be polystyrene spheres.

If used, any carrier solvent is then removed, leaving the layers of spheres in contact with each other but with the interstitial spaces empty. The interstitial spaces are then infiltrated by a further amount of the first phase material. This may be done for example by solution infiltration, by dipping or by deposition from the vapour phase.

The spheres are then removed from the structure by a suitable means so as to leave no residue, resulting in empty space where the spheres were. This may be by combustion, or by a low temperature process, preferably room temperature solution processing; for example the spheres may be removed by solvent extraction, or sonication.

This leaves a skeleton of the donor first phase material, which is a lattice corresponding to and shaped as the interstitial spaces and which is continuous and connected to the substrate electrode. The voids created by removal of the spheres are interconnected because the spheres were originally in contact with each other. The voids form a continuous lattice that interpenetrates the skeleton of the first phase material.

The interconnected lattice comprising the empty space previously occupied by the spheres in the composite structure is then infiltrated by the second phase material which will be an acceptor material if the first phase material was a donor material, or vice versa. A continuous layer of the second phase material is then formed at the upper face of the composite structure and a second electrode, which may be of any appropriate material, for example a metal such as aluminium, gold or copper, is then applied to the continuous layer of the second phase material. McLachlan et al, Journal of Materials Chemistry, 2007,17,3773-3776, the contents of which are incorporated herein by reference, discloses the fabrication of three- dimensionally ordered macroporous thin film structures using organic semiconductors but relatively large template particles were used with diameters of 250 -400 nm.

For applications involving organic semiconductor materials, e.g. for sensing and photovoltaics, a large interface area, combined with a high degree of open-cellular interconnectivity is required, necessitating the use of small template particles having dimensions of, for example, no more than about 100 nm. In WO 2008/029161 it is stated that the spheres may be between 10 and 500 nm in diameter and that a typical diameter would be 50 nm.

WO 2008/029161 discloses that in the process described there are three possibilities, namely: both the donor and acceptor materials are organic semiconductors; or one of the donor and acceptor materials is an organic semiconductor and the other is an inorganic semiconductor; or both of the donor and acceptor materials are inorganic semiconductors.

There is thus disclosed in WO 2008/029161 a process for the production of a semiconductor thin film structure in which nanoparticles of a removable material are deposited on a substrate and by self assembly form a structure defining a network of interstitial spaces; the interstitial spaces are infiltrated by a first semiconductor material which will serve as a donor or acceptor material; and the nanoparticles are removed so as to leave a three dimensional structure of the first semiconductor material including a network of interconnected pores.

Viewed from one aspect, the present invention is characterised over this process in that the nanoparticles and the first semiconductor material are co-deposited so that the first semiconductor material infiltrates the interstitial spaces during nanoparticle self-assembly.

Preferably the mean diameters of the nanoparticles (and the pores they leave) is no more than about 100 nm, or ^' less than about 100 nm. Whilst the nanoparticles have no particular upper limit in terms of the ability to self assemble, and for example could have a diameter of up to at least about 500 nm, in the context of organic photovoltaic device in particular it is desirable to have diameters of no more than about 100 nm given the low exciton diffusion lengths of organic semiconductors. With typical exciton diffusion lengths for organic semiconductors being the range of about 10 nm to about 30 nm, that would be a goal for the nanoparticle / pore size. In some embodiments, the nanoparticle / pore diameters are preferably a few tens of nanometres. However, the smaller the diameter of the nanoparticles, the greater the disorder there is in, the self assembled structure. As more disorder is introduced, the connectivity of the different regions is disrupted and this may limit the performance of a device due to transport problems.

In some embodiments of the invention, the mean diameters of the nanoparticles are in the range of about 50 nm to about 100 nm. In some embodiments of the invention the nanoparticles have a diameter in the range of about 60 nm to about 100 nm. In some embodiments of the invention the nanoparticles have a diameter in the range of about 70 nm to about 100 nm. In some embodiments of the invention the nanoparticles have a diameter in the range of about 80 nm to about 100 nm. In some embodiments of the invention the nanoparticles have a diameter in the range of about 90 nm to about 100 nm. In some embodiments of the invention, the mean diameters of the nanoparticles are in the range of about 10 nm to about 30 nm. In some embodiments of the invention, the mean diameters of the nanoparticles are in the range of about 10 nm to about 50 nm . In some embodiments of the invention, the mean diameters of the nanoparticles are in the range of about 30 nm to about 50 nm.

The nanoparticles may be of any suitable material that can form a structure by self assembly and can be removed afterwards by a method that will not also remove the first semiconductor material. The nanoparticles are preferably able to produce an ordered, or pseudo or partially ordered structure. A typical ordered structure would be a hexagonal or cubic or other geometry close-packed structure. As the structure is at least partially ordered or structured and not random, each sphere or at least. a substantial number of spheres will be in contact with all or a substantial number of its neighbours, so that there is an interconnected array of spheres. When the nanoparticles are removed, there will be defined a structure with interconnected pores. In preferred embodiments of the invention there is provided a solid lattice of the first semiconductor material, defining a lattice of interconnected spaces.

The nanoparticles are preferably substantially monodisperse, i.e. particles whose variation in size is small or extremely small.

In some preferred embodiments, the nanoparticles are removed using solvent vapour. Preferably, the solvent vapour is hot, i.e. substantially above ambient temperature. The solvent may be boiled, preferably under reflux, to create the vapour. For, example, in the case of polystyrene nanoparticles vapour from boiling tetrahydrofurane (THF) may be used. THF has a boiling point of 66 ⁰ C at standard pressure bur preferably nanoparticle extraction is carried out under pressure so this will change.

The nanoparticles may for example be of polystyrene but other nanoparticle materials may be used, for example another polymer such as polymethylmethacrylate. In general, preferred nanoparticles will be of a polymer that can be removed by exposure to an organic solvent, preferably in hot vapour form. For, example, in the case of polystyrene nanoparticles vapour from boiling tetrahydrofurane (THF) may be used. THF has a boiling point of 66 ⁰ C. Suitable organic solvents, in addition to THF, could be toluene or xylene, for example.

The combination of using relatively small nanoparticles with sizes no greater than about 100 nm, and preferably smaller, co-deposition with the first semiconductor material to enhance infiltration of the interstitial spaces, and removal by means of hot solvent vapour, results in a well defined structure. The hot solvent provides deeper penetration into small spaces to remove the nanoparticles. The hot solvent may also provide hot solvent annealing of the material, leading to higher or different material crystallinity.

As noted above, in WO 2008/029161 there is disclosed a process for the production of a semiconductor thin film structure in which nanoparticles of a removable material are deposited on a substrate and by self assembly form a structure defining a network of interstitial spaces; the interstitial spaces are infiltrated by a first semiconductor material which will serve as a donor or acceptor material; and the nanoparticles are removed so as to leave a three dimensional structure of the first semiconductor material including a network of interconnected pores.

In WO 2008/029161 it is stated that the nanoparticles, which may be polystyrene, may be removed by combustion or by a low temperature process, preferably room temperature solution processing.

Accordingly, viewed from another aspect the invention is characterised over WO 2008/029161 in that the nanoparticles are removed by the use of solvent in vapour form. Preferably the vapour is hot, and thus the vapour is preferably obtained by increasing the temperature of the solvent beyond ambient temperature to its boiling point. This is preferably done under reflux. Preferably nanoparticle removal is carried out under pressure.

Preferable and optional features of this second aspect of the invention, such as the nature and size of the nanoparticles, and the solvent used, are as discussed above in respect of the first aspect of the invention.

In embodiments of the above aspects of the invention, to achieve co-deposition of the nanoparticles and the first semiconductor material, the substrate may be exposed to a solution of the first semiconductor material which also contains the nanoparticles in colloidal form. Preferably the first semiconductor material is water soluble. In a typical process, the substrate is immersed in the solution and then there is controlled evaporation so that there is generated a thin film comprising the nanoparticles which have self-assembled, and the first semiconductor material occupying the interstitial spaces between the spheres.

It will be appreciated that to produce a product such as a photovoltaic cell, a second semiconductor material infiltrates the pores left by the nanoparticles. One of the two semiconductor materials is a donor material and the other is an acceptor material. Infiltration of the second semiconductor material results in there being a solid lattice of the first semiconductor material, interpenetrated by a lattice of the second semiconductor material.

In preferred embodiments, the first semiconductor material is an organic semiconductor material. The organic semiconductor material could be, for example, an organic semiconducting polymer, such as sodium poly[2-(3- thienyl)ethoxy-4-butylsulfonate] (PTEBS), or a molecular semiconductor such as copper(ll) phthalocyanine-tetrasulfonic acid tetrasodium salt (TS-CuPc).

When a structure in accordance with such embodiments, using an organic first semiconductor material, is infiltrated by a second semiconductor material in some preferred embodiments the second semiconductor material is also an organic semiconductor material. However, in some embodiments the second semiconductor material may be an inorganic material.

In some embodiments, the substrate is an electrode and after infiltration by a second semiconductor material, a second electrode is placed on the other side of the structure. There may be one or more intermediate layers of material. For photosensitive applications, at least one of the electrodes will be transparent.

In accordance with embodiments of the above aspect of the invention, the nanoparticles are removed and a second semiconductor material is infiltrated into the structure. However, co-deposition is applicable to other arrangements also, such as other arrangements as disclosed in WO 2008/029161. For example, non- sacrificial nanoparticles of a first semiconductor material may be co-deposited with a second semiconductor material which infiltrates the interstitial spaces during self assembly of the nanoparticles. The nanoparticles are left in place and are used as either the donor or acceptor material. This can be used in inorganic-organic hybrid devices, wherein one of either the donor or the acceptor material is an inorganic material. In such a device, the deposited nanoparticles could for example be of titanium oxide or zinc oxide, which are able to act as semiconductors. These are co-deposited with a second semi-conducting material which infiltrates the interstitial spaces. These may act as an acceptor and the co-deposited donor material could be, for example, P3HT (poly-3(hexylthiophene). Alternatively the deposited nanoparticles could be an organic material, examples of which are given above, which would not be removed.

Accordingly, viewed from another aspect, the invention provides a process for the production of a semiconductor thin film structure in which nanoparticles of a first semiconductor material are deposited on a substrate and by self assembly form a structure defining a network of interstitial spaces; and the interstitial spaces are infiltrated by a second semiconductor material so as to provide a three dimensional structure of interpenetrating regions of the first and second semiconductor materials; one of the first and second semiconductor materials serving as a donor and the other serving as an acceptor, material; characterised in that the nanoparticles and the second semiconductor material are co-deposited so that the second semiconductor material infiltrates the interstitial spaces during nanoparticle self-assembly.

Preferable and optional features of this further aspect of the invention, such as the nature a of the nanoparticles, and the selection of the second semiconductor material, such as an organic semiconductor material, are as discussed above in respect of the first aspect of the invention.

It will be appreciated that the expression "nanoparticles" covers a range of particles, some of which may be referred to as nanospheres or nanocapsules. Whilst it may be that in many cases, the particles will be generally spherical, the nanoparticles may have other shapes. Where the expression diameter is used, this does not imply that there is a sphere, and the expression can for example encompass the maximum or mean dimension across a particle. As set out in WO 2008/029161 , the particles may be any geometric shape which is capable of packing to form a continuous interpenetrable lattice, for example non-tessellating shapes or tessellating shapes laid out in a non-tessellating orientation, and any packing geometry may be adopted.

Some embodiments of the invention will now be described and with reference to the accompanying figures, in which:

Figures 1 (a), (b), (c) and (d) are images from field emission scanning electron microscopy (FE-SEM);

Figures 2 (a) and (b) are images from field emission scanning electron microscopy (FE-SEM); Figure 3 is a diagram showing ultra violet / visible absorption spectra;

Figure 4 is a second diagram showing ultra violet / visible absorption spectra;

Figure 5 illustrates diagrammatically steps in a process in accordance with the invention; and

Figure 6 is a diagram illustrating a co-deposition process.

Looking now at a first example in accordance with the invention, there is a process for producing a porous large area organic semiconductor thin film.

Example 1

Synthesis of 3D ordered macroporous (3DOM) structures in this embodiment of the invention, generally inverse opal structures, can be broken down into two main fabrication steps: (1 ) vertical or convective self-assembly, followed by (2) nanoparticle removal.

Polystyrene nanoparticles, i.e. nanospheres, were synthesised by radical initiated soap-free emulsion polymerisation with either cationic or anionic initiator and dialysed against water prior to use. See K. Tauer et al, Colloid and Polymer Science, 2008, 246, 499 - 515. Size and polydispersity were determined by dynamic light scattering and transmission electron microscopy. In this example, colloids of mean diameters between 50-100 nm were used with a polydispersity index range of 0.034-0.010.

Compact composite structures of close-packed, self-assembled monodisperse polystyrene nanoparticles infilled with water-soluble organic semiconductors were fabricated in an one-step process using vertical co-deposition, a controlled evaporation-driven thin film deposition process in which colloidal crystal growth and vacancy infiltration with a second material are combined. To generate the 3DOM structures either an organic semiconducting polymer, sodium poly[2-(3- thienyl)ethoxy-4-butylsulfonate] (PTEBS) or a molecular semiconductor, copper(ll) phthalocyanine-tetrasulfonic acid tetrasodium salt (TS-CuPc), were used.

The structures were grown on either plain glass slides or indium-tin oxide (ITO) coated glass substrates pre-cleaned by sonicating in appropriate solvents followed by ultraviolet/ozone or oxygen plasma treatment. After immersing the substrate in a mixture of colloidal dispersion and organic semiconductor solution, the structures were grown in a temperature-stable incubator at 60 ^° C ± 0.4 ^° C and a relative humidity <15 %.

In a second step the nanoparticles were selectively removed from the composite structure by exposure of the sample to vapour from hot solvent boiling under reflux, in this case tetrahydrofurane. Purified, fresh vapour deeply penetrates the composite structure, dissolving any remaining polystyrene and results in the formation of well defined 3DOM organic thin films.

Penetration of pure THF vapour into the composite thin film, condensation and dissolution of the polystyrene spheres followed by gravity-induced draining of the polymer solution resulted in the formation of well-defined 3DOM organic thin films. This process is equivalent to continous washing in high purity warm solvent, although it requires only a very small amount of solvent. The degree of template removal is simply a function of exposure time.

The resulting 3D0M structures were analysed by field emission scanning electron microscopy (FE-SEM) and ultraviolet/visible (UV/vis).

Example 2

Again, the polystyrene nanospheres were synthesised by radical initiated soap- free emulsion polymerisation using either a cationic or anionic initiator and dialysed against water prior to use. The average particle size and polydispersity was determined by dynamic light scattering and transmission electron microscopy. Polystyrene sphere latexes with two distinct mean particle diameters (100 nm and 60 nm) were prepared with polydispersities of 0.02 and 0.06 respectively. Typically, 0.10-0.15 mL of latex was added to 20 ml_ of water containing 0.02-0.15 mg ml_-1 of the water-soluble polymeric semiconductor, PTEBS, or the molecular semiconductor, TS-CuPc. Colloidal stability upon blending was preserved in both cases. Composite films of close-packed, self- assembled monodisperse polystyrene spheres infilled with the water-soluble organic semiconductors were fabricated in a single-step process using vertical co-deposition as discussed in Example 1. Again, the polystyrene spheres were selectively removed from the composite structure by exposure of the sample to vapour from refluxing tetrahydrofurane (THF). The resulting 3DOM structures were analysed by field emission scanning electron microscopy (FE-SEM) and ultraviolet/visible (UV/vis).

Figure 1 (a) shows an FE-SEM image of a film from co-deposition of PTEBS with 100 nm polystyrene latex nanoparticles, showing a crack free film covering a large substrate area. Figures 1 (b) and 1 (c) are larger scale images. Figure 1 (d) is a larger scale image, but in this case for a film made from co-deposition of TS-CuPc with the 100 nm polystyrene nanoparticles.

The FE-SEM images of the 3DOM films created in accordance with both examples, using 100 nm nanoparticles and both PTEBS and TS-CuPc organic semiconductors showed uniform crack-free domains of up to 10 x 10 μm. An interconnected open cellular structure was realised for both PTEBS and TS- CuPc. Typically, in each cavity there are three pores visible, consistent with a close packed cavity array mirroring that of the template spheres. It can therefore be inferred that below the film surface each cavity is connected to all six adjacent cavities via a pore. Notably, the structure is not a perfect inverse opal, since there are irregularities in the cavity shapes and separating wall thicknesses. These departures from ideality can be attributed to a combination of the size distribution of the template spheres and relaxation of the organic semiconductor matrix upon template removal.

It is important to note that during co-deposition the soluble filling material acts as a surfactant, modifying the inter-sphere interactions and capillary forces which drive nanoparticle self-assembly. In some cases the presence of the filling material may destabilise the latex leading to random clustering and agglomeration of the nanoparticles, although this is not observed to any significant extent when using 100 nm diameter spheres. The use of co-deposition creates a strong composite film which helps to release built up tension in the film during the drying process which otherwise might result in film cracking.

Figure 2 (a) is an FE-SEM image of a film from co-deposition of PTEBS with 60 nm polystyrene latex nanoparticles. Figure 2 (b) is an FE-SEM image of a film from co-deposition of TS-CuPc with 60 nm polystyrene latex nanoparticles.

With nanoparticles of 50 nm or 60 nm diameter the total surface area of the macroporous cellular structure is increased upon sphere removal. For both PTEBS and TS-CuPc the FE-SEM images showed the presence of porous structures but they were less regular and not as well defined as those templated using the 100 nm spheres. Furthermore, in the open cellular structure there are fewer pores between adjacent cavities, although sufficient to ensure continous interconnectivity. Whilst PTEBS films showed a higher degree of regularity in their porous network than those fabricated from TS-CuPc, both are much more defective than those fabricated using the 100 nm spheres. This can in part be ascribed to the broader sphere size distribution (0.06 for 60 nm spheres; 0.02 for 100 nm spheres). However, for sub-100 nm nanoparticles, the self-assembly driving capillary attraction energy decreases to a level comparable to the thermal energy of the, particles, thus counteracting particle ordering. Since the latter operates to disrupt regular array formation, the level of disorder in the films fabricated using 50 nm or 60 nm diameter spheres is to be expected. Useful structures can be obtained with with sphere (nanoparticle) diameters of no more than 60nm or no more than 50 nm. It is also believed that viable structures can be obtained with smaller nanoparticle diameters, i.e. diameters of no more than about 40 nm, or no more than about 30 nm, or no more than about 20 nm or no more than about 10 nm.

Figure 3 shows UV/vis electronic absorption spectra of the composite (nanoparticle/organic semiconductor) films and the resulting 3DOM films formed after sphere removal, for (a) PTEBS and (b) TS-CuPc. For reference spectra are also shown for the polystyrene colloidal particles and a two-dimensional (2D) film of each organic semiconductor formed by spin coating. Whilst the lowest energy electronic transition in polystyrene is in the UV region of the spectrum, the polystyrene sphere assemblies (and cavities formed after sphere removal) appear to absorb light due to scattering, the efficiency of which increases as the wavelength of the light approaches that of the polystyrene sphere (or cavity). The 2D PTEBS film made via spin coating shows the expected behaviour with an absoption peak at -430 nm. This peak is preserved in the composite PTEBS/polystyrene sphere film, before and after removal of the polystyrene template. The composite spectrum correlates relatively well as a linear combination of the 2D PTEBS film and the polystyrene sphere assembly. However, the absorption spectrum of the final 3DOM structure is not the same as that of the solid film of PTEBS due to scattering of light, especially at low wavelengths. The primary absorption in TS-CuPc is located at -610 nm where the effects of light scattering by 100 nm transparent particles/cavities is less significant (Fig 3b).

The embodiments of the invention allow for the fabrication of continuous, large area 3DOM structures of organic semiconductors using vertical co-deposition of templating polystyrene spheres of 100 nm and sub - 100 nm diameter in conjuction with water soluble small molecule (i.e. TS-CuPc) or polymeric (i.e. PTEBS) organic semiconductors. Subsequent post-deposition treatment with hot solvent vapour is an efficient means of sphere removal, generating 3DOM organic semiconductor films of tunable pore size between about 50 to about 100 nm depending on the size of the polystyrene latex spheres used. Despite minor defects and partially distorted packing order - particularly when using the smallest sphere templates - the resulting interconnected cellular networks of organic semiconductor provide a suitable platform for subsequent infiltration of a variety of materials to form interpenetrating network nanocomposites with broad application potential.

Figure 4 is a diagrammatic view of a process in accordance with the invention.

At step A a 1 substrate in the form of an indium-tin oxide (ITO) coated glass electrode is immersed in a mixture 2 of a colloidal dispersion of polystyrene nanoparticles, and organic semiconductor solution. At step B the structures are grown in a temperature-stable incubator to produce self assembled layers 3 of the polystyrene spheres, with the organic semiconductor 4 in the interstitial spaces.

At step C the nanoparticles are selectively removed from the composite structure by exposure of the sample to vapour from a source 5 of hot solvent boiling under reflux, in this case tetrahydrofurane. The vapour penetrates the composite structure, dissolving the polystyrene and resulting in the formation of a thin film of the organic semiconductor 4 with well defined pores 6 as shown at step D. There is effectively a skeleton of the organic semiconductor 4 in the form of a lattice defining interconnected spaces. The spaces are interconnected because the nanoparticles were originally in contact with each other.

In this example, there is further processing to produce a semiconductor device. Thus, at step E a second organic semiconductor material 7 is infiltrated into the interconnected spaces defined by the lattice of organic semiconductor material 4. This provides a lattice of the semiconductor material 7, with the two lattices interpenetrating. A second indium-tin oxide (ITO) coated glass electrode 8 is then placed on the structure at step F. The semiconductor device could be, for example, a photo-sensitive device, such as a photovoltaic cell or a photo detection device.

The first organic semiconductor material such as TS-CuPc or PTEBS may act as a donor, with the second semiconductor material being an acceptor material such as fullerene or fullerene derivative such as PCBM ([6, 6]-phenyl-C61 -butyric acid methyl ester).

Figure 6 shows a preferred co-deposition process. A substrate 9 is directed upwardly, at an inclination. A mixture 10 of a colloidal dispersion of polystyrene nanoparticles 1 1 , and a solution of an organic semiconductor, undergoes convective flow in the direction of arrow 12, upwardly and towards the substrate 9. Evaportation of the liquid takes place from the meniscus 13, as indicated by arrows 14, this leading to self assembly of the nanoparticles into an array 15, with the organic semiconductor infiltrated into the interstitial spaces . In accordance with the above embodiments of the invention, there is provided a process for the production of an organic semiconductor thin film structure. Nanoparticles of a removable material such as polystyrene are co-deposited with an organic semiconductor on an electrode substrate from a mixture of a colloidal suspension of the nanoparticles and a solution of the organic semiconductor. By self assembly, the nanoparticles form an ordered structure defining a network of interstitial spaces which are infiltrated by the organic semiconductor material. The nanoparticles are removed by hot solvent vapour so as to leave a three dimensional ordered macroporous structure of the organic semiconductor material. A second semiconductor material can then be infiltrated into the pores left by removal of the nanoparticles and a second electrode added, to create a photo-sensitive device.

Viewed from another aspect of the invention, there is provided a process which comprises vertical co-deposition of a water soluble organic semiconductor with polystyrene nanospheres as a template, followed by solvent vapour nanosphere removal, so as to generate a macroporous large area thin film of the organic semiconductor having a network of pores of pore size less than 100 nm. In the preferred embodiment, the film has low number crack and defect disorder and is of broad application in different fields.

Viewed broadly from another aspect, the invention provides a process for producing a multiphase thin film structure in which there is co-deposition on a substrate of (i) a first phase material in the form of nanoparticles which self assemble such that nanoparticles are in contact with neighbouring nanoparticles, so as to define a first lattice of first phase material and a second lattice of interconnected interstitial spaces between the nanoparticles and (ii) a second phase material which infiltrates interstitial spaces between the nanoparticles during self assembly of the nanoparticles, so as to occupy at least partially the second lattice. The nanoparticles may be removed subsequently so as to vacate the first lattice, which can then occupied by a third phase material. In such an arrangement the first and third phase materials may be semiconductors, one acting as a donor and the other as an acceptor.