WO2012123717A2

WO2012123717A2 - Composite material

Info

Publication number: WO2012123717A2
Application number: PCT/GB2012/050513
Authority: WO
Inventors: Benjamin Lionel Farmer
Original assignee: Eads Uk Limited
Priority date: 2011-03-11
Filing date: 2012-03-08
Publication date: 2012-09-20
Also published as: WO2012123717A3; GB201104131D0

Abstract

A method of manufacturing a composite material. A catalyst material is deposited on a substrate, the catalyst material being patterned to form an array of catalyst regions. Each catalyst region has an outer periphery which has no rotational symmetry and/or no mirror symmetry. A first array of bundles of filaments is grown on the substrate, growth of the filaments being catalysed by the catalyst material. Each bundle and each filament has a base attached to one of the catalyst regions and a free tip. Each bundle comprises a plurality of filaments spaced apart from adjacent bundles in the array by an interstitial gap. The bundles are impregnated with a capillary forming liquid which is then evaporated so that the free tips of the filaments are drawn together within each bundle and the bundles become distorted. These steps are repeated to provide a second array of distorted bundles of filaments which are positioned or grown so that the bundles of one array are positioned at least partially within the interstitial gaps of the other array. The bundles of filaments and the interstitial gaps are impregnated with a matrix material. The bundles typically become distorted due to their lack of symmetry. For instance they may become tilted, twisted, or distorted in some other way. The nature and orientation of the distortion can be controlled easily by tailoring the shape and orientation of the catalyst region, enabling the composite material to possess a desired anisotropic mechanical or electromagnetic property.

Description

COMPOSITE MATERIAL

FIELD OF THE INVENTION

The present invention relates to a method for manufacturing a composite material, and a composite material.

BACKGROUND OF THE INVENTION

A method of manufacturing a composite material is described in WO 2008/029178. Two or more layers of carbon nanotubes (CNTs) are grown in-situ; and each layer is impregnated with a matrix before growing the next layer. WO 2008/029178 also describes a method of aligning the CNTs by applying a plasma at an angle to the layer to cause the growth of a layer of nanofibres, aligned with the direction of the electromagnetic field. A problem with this method is that it can be complex and difficult to control.

Another method of manufacturing a composite material is described in WO 2009/019510. In this case the CNTs are grown ex-situ and dipped into a liquid layer of matrix in a previously impregnated layer of CNTs.

A method of capillary forming CNTs is described in Diverse 3D Microarchitectures Made by Capillary Forming of Carbon Nanotubes; by De Voider, Michael; Tawfick, Sameh H.; Park, Sei Jin; Copic, Davor; Zhao, Zhouzhou; Lu, Wei; Hart, A. John; Advanced Materials; 2010; volume 22; pages 4384-4389 (hereafter referred to as "De Voider et al").

SUMMARY OF THE INVENTION

A first aspect of the invention provides a method of manufacturing a composite material, the method comprising: a. depositing a catalyst material on a substrate, the catalyst material being patterned to form an array of catalyst regions, and at least some of the catalyst regions being asymmetrical regions having an outer periphery which has no rotational symmetry and/or no mirror symmetry; b. growing a first array of bundles of filaments on the substrate, growth of the filaments being catalysed by the catalyst material, each bundle and each filament having a base attached to one of the catalyst regions and a free tip, each bundle comprising a plurality of filaments spaced apart from adjacent bundles in the array by an interstitial gap; c. impregnating the bundles of filaments with a capillary forming liquid; d. evaporating the capillary forming liquid so that the free tips of the filaments are drawn together within each bundle and the bundles grown on the asymmetrical catalyst regions become distorted; e. repeating steps a. to d. to provide a second array of bundles of filaments; f. positioning or growing the arrays so that the bundles of one array are positioned at least partially within the interstitial gaps of the other array; and g. impregnating the bundles of filaments and the interstitial gaps with a matrix material.

The bundles grown on the asymmetrical catalyst regions typically become distorted in the capillary forming stage due to their lack of symmetry. For instance they may become tilted, twisted, or distorted in some other way. The nature and orientation of the distortion can be controlled easily by tailoring the shape and orientation of the catalyst region, enabling the composite material to possess a desired anisotropic mechanical or electromagnetic property. In one example each bundle has a centre line between the centre of its base and the centre of its tip, and the distortion of the bundle during step d. causes the centre line to tilt. In another example the distortion of the bundle during step d. causes the bundle to twist without tilting. A combination of tilt and twist may also be possible.

The shape and/or orientation of the catalyst regions may vary between the first array and the second array. This enables the bundles to be tilted or twisted in a different direction between the arrays, or it enables the magnitude of tilt to vary from array to array. This enables the mechanical and/or electromagnetic properties of the material to be varied through the thickness of the material

The shape and/or orientation of the catalyst regions may also vary across the first array and/or the second array, enabling the mechanical and/or electromagnetic properties of the material to be varied across the width of the material. This enables the bundles to be twisted in an opposite sense between the bundles of one of the arrays, or it enables the magnitude or direction of tilt of the bundles to vary between the bundles of one of the arrays. Thus for example some of the bundles in the first array may be twisted in a clockwise sense, and others in an anti-clockwise sense.

In the embodiments described below each catalyst region in both the first and second layers is an asymmetrical catalyst region with an outer periphery which has no rotational symmetry and/or no mirror symmetry. However it will be appreciated that not necessarily all catalyst regions in all layers must have this asymmetry - in other words some of the layers may consist entirely of symmetrical (e.g. circular) catalyst regions and/or some of the layers may contain a mixture of symmetrical (e.g. circular) catalyst regions and asymmetrical catalyst regions.

Preferably a proportion of the deposition area occupied by the catalyst regions is greater than a proportion of the deposition area which is substantially free of catalyst material, resulting in a higher density of filaments than is possible using the methods described in WO 2008/029178 and WO 2009/019510. Preferably the proportion of the deposition area occupied by the catalyst regions is greater than a proportion of the deposition area occupied by the gaps between the catalyst regions.

Each catalyst region may have a solid or continuous shape with no internal voids which are substantially free of catalyst material. However if such internal voids are present then the proportion of the deposition area occupied by the catalyst regions is typically greater than the proportion of the deposition area occupied by the gaps and the voids summed together. Typically the matrix material forms a continuous structure which substantially completely fills the space in the composite material which is not occupied by the filaments, thereby fixing the filaments in space relative to each other.

Typically the matrix material is capable of transferring load from each array to adjacent arrays, from each bundle to adjacent bundles and from each filament to adjacent filaments.

Typically the matrix material is formed for a material which is less strong and less stiff than the filaments.

The composite material may be impregnated with matrix material in a single step. However more typically the matrix material is applied as a series of layers, each layer being applied at a different time. In this case the first array is impregnated with a first layer of matrix material in a first impregnation step, and the second array is impregnated with a second layer of matrix material in a second impregnation step.

Typically the matrix material is cured as a series of layers, each layer being cured at a different time. The benefit of such a layer-by- layer curing approach is that each cured matrix layer may have a different cross-sectional shape, size, or pattern, enabling a "net shape" part to be grown by additive fabrication. However, the invention also extends to processes in which all of the matrix material in the composite is cured at the same time. That is, each successive layer of matrix material remains uncured until the part is complete, and the part is then heated to cure the matrix throughout in a single curing step.

The matrix material may cured by exposure to electromagnetic radiation, such as a scanning laser beam or other radiation beam such as an electron beam. This enables the matrix to be cured selectively - that is with a desired shape, size or pattern.

The bundles of the second array may be positioned in a separate step f. by dipping the free tips of the second array into a layer of liquid matrix material in the interstitial gaps of the first array, as described in WO 2009/019510. This enables the second array (and optionally also the first array) to be grown ex-situ, that is remotely from the liquid matrix material, and optionally on the same substrate as the first array.. This allows the filaments to be grown at high temperatures, up to ~1400°C, which is significantly higher than the temperatures required to cure certain types of liquid matrix material such as liquid epoxy resin. In this case the method typically further comprises: growing the first array on a substrate; transferring the first array onto a build platform with the tips adjacent the build platform and the bases remote from the build platform; and impregnating the inter- filament gaps and interstitial gaps of the first array with the layer of liquid matrix material after it has been transferred to the build platform. The composite material may have only two layers, or steps a. to d. can be repeated to provide a third array of bundles of filaments; impregnating the inter-filament gaps and interstitial gaps of the second array with a second layer of liquid matrix material; and dipping the free tips of the third array into the second layer of liquid matrix material in the interstitial gaps of the second array. Typically the first layer of liquid matrix material is cured before the second layer of liquid matrix material impregnates the inter- filament gaps and interstitial gaps of the second array, for instance by scanning a radiation beam across it. Typically the tips of the first array and second array each extend in a plane. As the tips of the second array are dipped into the layer of liquid matrix material in the interstitial gaps of the first array, the second array may be moving perpendicular to the plane of those tips, or at an acute angle to the plane if required.

Alternatively the inter- filament gaps and interstitial gaps of the first array may be impregnated with a matrix layer before growth of the second array; and at least part of the second array grown in the interstitial gaps of the first array by depositing a catalyst material on the matrix layer in the interstitial gaps of the first array and growing the second array of bundles of filaments on the catalyst regions in the interstitial gaps of the first array. Thus in this case the substrate for each repeat of steps a. and b. is the matrix layer in the interstitial gaps, and the bundles are grown within the interstitial gaps so there is no separate step f.

In some embodiments each repeat of steps a. and b. is carried out on the same substrate. In other embodiments each repeat of steps a. and b. is carried out on a different substrate, which may be a previously formed layer of matrix material as described above. Preferably the number of catalyst regions in the first array is different to the number of catalyst regions in the second array. This enables a series of reinforcement layers to be grown with different shapes under the control of a three-dimensional computer model of a part so that a "net shape" part is grown by additive fabrication.

A further aspect of the invention provides a composite material comprising: a. a first filament layer comprising an array of bundles of filaments, each bundle being spaced apart from adjacent bundles by an interstitial gap; b. a second filament layer comprising an array of bundles of filaments, each bundle being spaced apart from adjacent bundles by an interstitial gap; and c. a matrix material occupying the interstitial gaps between the bundles; wherein each bundle and each filament has a base, a stalk and a tip, and each filament is spaced apart from adjacent filaments in the bundle by an inter- filament gap which is occupied by the matrix material and is smaller at the tip of the filament than at the base of the filament; wherein the first and second filament layers are partially overlapped with each other so that the stalks of one array are positioned with part of their length within the interstitial gaps of the other layer and pass between either the tips or bases of the other layer; and wherein the base of at least some of the bundles has an outer periphery which has no rotational symmetry and/or no mirror symmetry. Due to the partial overlap between the layers the composite material has a total thickness which is greater than the individual thickness of each filament layer.

A further aspect of the invention provides a composite material comprising: a. a first array of bundles of filaments, each bundle being spaced apart from adjacent bundles by an interstitial gap; b. a second array of bundles of filaments, each bundle being spaced apart from adjacent bundles by an interstitial gap; and c. a matrix material occupying the interstitial gaps between the bundles; wherein each bundle and each filament has a base and a tip, and each filament is spaced apart from adjacent filaments in the bundle by an inter- filament gap which is occupied by the matrix material and is smaller at the tip of the filament than at the base of the filament; wherein the first and second filament layers are at least partially overlapped with each other so that the stalks of one array are positioned with at least part of their length within the interstitial gaps of the other layer; wherein the stalks of the first array are generally oppositely oriented to the stalks of the second array; and wherein the base of at least some of the bundles has an outer periphery which has no rotational symmetry and/or no mirror symmetry.

Typically the bases of the bundles occupy an area which is greater than the area occupied by the inter-base gaps.

Typically the bases of the first array of bundles lie in a plane, and the bases of the first array of bundles overlap with the second array of bundles when viewed at a right angle to said plane.

The filaments may comprise single walled CNTs; multi-walled CNTs, carbon nanofibres; CNTs coated with a layer of amorphous carbon, or any other suitable filament material. Typically the filaments have an aspect ratio greater than 100, preferably greater than 1000, and most preferably greater than 10⁶.

In the embodiments described below each bundle in both the first and second layers is an asymmetrical bundle with a base having an outer periphery which has no rotational symmetry and/or no mirror symmetry. However it will be appreciated that not necessarily all bundles in all layers must have this asymmetry - in other words some of the layers may consist entirely of symmetrical (e.g. circular) bundles and/or some of the layers may contain a mixture of symmetrical (e.g. circular) bundles and asymmetrical bundles.

Possible applications for the composite material include flat lenses, or three-dimensional meta-materials with negative refractive index.

BRIEF DESCRIPTION OF THE DRAWINGS Embodiments of the invention will now be described with reference to the accompanying drawings, in which:

Figure 1 is a schematic view showing an ALM chamber and a CVD-CNT chamber;

Figure 2 shows the deposition of a first catalyst layer;

Figure 3 is a plan view of the catalyst layer;

Figure 4 is a side view of an array of bundles of CNTs;

Figure 5 is a schematic plan view of one of the bundles of CNTs;

Figure 6 shows a first stage of a capillary forming process;

Figure 7 shows the capillary- formed CNTs suspended above a layer of uncured matrix material;

Figure 8 shows the CNTs being dipped into the layer of uncured matrix material;

Figure 9 shows a second layer of CNTs suspended above the first layer;

Figure 10 shows the second layer of CNTs being dipped into a layer of uncured matrix material;

Figure 11 shows the composite with the second matrix layer cured;

Figure 12 shows a three-layer composite piece;

Figure 13 is a sectional view taken along a line A- A in Figure 12;

Figure 14 is a side view of a composite material with interlocking bundles;

Figure 15 is a plan view of the composite material of Figure 14;

Figure 16 shows a two-layer composite piece with 100% overlap;

Figure 17 shows a two-layer composite piece with less than 100% overlap; Figure 18 is a schematic view showing a combined ALM and CVD-CNT chamber;

Figure 19 shows a first stage of an in-situ manufacturing process;

Figure 20 shows a second stage of the in-situ manufacturing process;

Figure 21 shows a third stage of the in-situ manufacturing process; and

Figure 22 shows a three-layer composite piece manufactured by the process of Figures 19- 21;

Figure 23 shows a catalyst deposition pattern in which the orientation of the catalyst regions varies across the deposition region;

Figure 24 shows a catalyst region which lacks mirror symmetry and generates a twisted shape.

DETAILED DESCRIPTION OF EMBODIMENT(S)

Figure 1 is a schematic diagram showing an additive layer manufacturing (ALM) chamber on the left hand side of the figure, and a chemical vapour deposition-CNT (CVD-CNT) growth chamber on the right hand side. These two chambers are separated by a door 10. The ALM chamber comprises a vat 1 containing an un-cured liquid photo curing resin 2. A build platform 3 is mounted in the vat 1 and can be moved up and down as required. The CVD-CNT chamber contains a silicon transfer body 4 which is connected to an electrical heating circuit 6. The chamber has a gas input 7, a gas output 8 and a door 9.

Referring to Figure 2, a catalyst deposition system 5 deposits catalyst material 12 onto the silicon transfer body 4 in a predefined shape, pattern and density. The system 5 may comprise a printing head which sprays an array of colloid drops onto the transfer body 4, and as the colloid evaporates, metal catalyst particles suspended in the colloid drops are deposited. The catalyst particles may be, for example a metal, preferably transition metals Fe, Ni or Co, or alloys thereof; and the colloid liquid may be, for example alcohol, water, oil, or a mixture thereof. Alternatively the system 5 may deposit catalyst by another process such as evaporation of a metal. Figure 3 is a plan view of an array of catalyst regions 12 printed over a deposition area on the substrate 4. The catalyst material is patterned within the deposition area to form an array of square catalyst regions 12 which are spaced apart by gaps 13 substantially free of catalyst material.

The regions 12 have a diameter L and the gaps 13 have a width I. For the regions shown in Figure 3 the area Al occupied by the regions 12 is 25L² and the area A2 occupied by the gaps 13 is 16I²+40LI. For clarity of illustration, the area Al is only marginally greater than the area A2 in Figure 3. However, in practice the width I can be reduced relative to L to make A2 much less than Al if desired.

After the catalyst has been deposited, and subsequent conditioning of the catalyst by a combination of heat and oxidation and reduction using oxygen and hydrogen gases, a layer of carbon nanotubes (CNTs) is grown on the catalyst regions 12 by a chemical vapour deposition process. Carbonaceous gas is introduced into the CVD-CNT chamber via the gas input 7 and the substrate 4 is heated locally by the electrical heating circuit 6. More specifically, the circuit 6 induces an electrical current in the transfer body 4 which heats it resistively. Growth of CNTs is enhanced by generating a plasma in the chamber using an electrode 20 powered by a power supply 21.

Figure 4 is a side view of an array of CNT bundles on the substrate 4. Each bundle has a base 14 attached to a catalyst region 12 on the substrate and a free tip 15. Each bundle comprises a plurality of CNTs, each CNT also having a base attached to the substrate and a free tip. Figure 5 is a schematic plan view of one of the bundles, showing individual CNTs. Note that Figure 5 is highly schematic and the diameter of the CNTs is not to scale relative to the diameter L of the bundle or the spacing between the CNTs. The CNTs 16 at the periphery of the bundle define an area occupied by that bundle (in this case L²). Each CNT is spaced apart from adjacent CNTs in the bundle by an inter- filament gap 17.

After the CNTs have been grown, they are capillary formed by the process shown in Figure 6. In a first step the CNTs are suspended over a bath of liquid solvent 18. The solvent 18 is heated so that it evaporates and condenses on the CNTs, wetting the CNTs and impregnating the inter- filament gaps 17. This draws the free tips of the CNTs together slightly within each bundle so the bundles distort slightly as shown in Figure 6. Next the capillary forming liquid solvent on the CNTs is evaporated so that the free tips of the filaments are drawn together to form the highly densified shapes shown in Figure 7. The inter- filament gaps at the tip of each bundle are now significantly smaller than at the base 14 of each bundle where the filaments remain attached to the catalyst region 12. The diameter L' of the tips of the CNT bundles is now slightly less than the width I of the gaps between the catalyst regions, and significantly less than the width L at the base of the bundle.

Since the catalyst regions 12 have no rotational symmetry, the bundles tilt to one side as shown in Figure 7. The mechanism for this tilting is described in detail in De Voider et al, and results from asymmetric lateral force distribution due to the asymmetric location of the centroid of the region 12. Specifically, each bundle has a centre line 20 between the centre of its base 14 and the centre of its tip 15 which is tilted at an acute angle relative to a line 21 passing through the centre of the base 14 at a right angle to the base. The centre lines 20 are also tilted relative to the parallel planes defined by the substrate 4 and the tips 15.

After the first layer of CNTs has been grown and densified, the doors 9,10 between the chambers are opened to allow the substrate 4 to be decoupled from the resistive heating circuit 6, rotated by 180 degrees, and moved into the ALM chamber.

Referring to Figure 8, the substrate 4 is then lowered such that the tips of the CNTs penetrate into a thin base layer 19 of liquid polymer above the build platform 3. A strong surface interaction derives a capillary action effect, wicking the liquid polymer into the inter- filament gaps 17 between the CNTs and the interstitial (or inter-bundle) gaps 30 between the bundles. The first layer of CNTs is penetrated only partially so that the bases of the fibres protrude from the surface of the polymer layer 19 as shown in Figure 8.

In the next step the substrate 4 is removed. The CNTs remain embedded in the polymer layer 19 due to the surface interactions. The substrate 4 is then returned to the CVD-CNT chamber and the doors 9, 10 are closed. A laser 11 is then activated and scanned over the surface of the layer of CNTs to selectively cure areas of resin, resulting in a cured base layer 31 (Figure 9) comprising a cross-linked and hardened polymer matrix layer surrounding the CNT layer. The build platform 3 is then lowered into the bulk of the liquid resin 2, allowing a flow of liquid resin over the surface of the base layer, into the interstitial gaps 30 between the CNT protruding from the cured matrix layer 31 as shown in Figure 10.

The process is then repeated to provide a second array of CNTs on the substrate 4 with narrow stalks 32 which are dipped into a layer of liquid matrix material 33 in the interstitial gaps of the first array as shown in Figure 10.

The bases 14 of the first array of bundles are spaced apart from each other by inter-base gaps 36 (one of which is labelled in Figure 9) which are just wide enough to admit the stalks 32 of the second array. The bases 14 of the bundles occupy an area which is greater than the area occupied by the inter-base gaps 36 in the same plane (their areas being dictated by the pattern of catalyst regions from which the bundles have been grown).

The second array of CNTs are moved down at an acute angle to the plane of the substrate 4 (as indicated by the arrows in Figure 9) to avoid any clash with the first array of CNTs.

In the next step the substrate 4 is removed and the CNTs 32 remain embedded in the polymer layer 33 due to the surface interactions. The substrate 4 is then returned to the CVD-CNT chamber and the doors 9, 10 are closed.

The laser 11 is then activated and scanned over the surface of the second layer of CNTs 32 to selectively cure areas of resin, resulting in a cured layer 34 (Figure 11) comprising a cross-linked and hardened polymer matrix layer surrounding the CNT layer. The build platform 3 is then lowered again into the bulk of the liquid resin 2, allowing a flow of liquid resin over the surface of the second layer 34, into the interstitial gaps 35 between the CNTs protruding from the matrix layer 34.

The process can then be repeated any number of times. For example Figure 12 shows a three layer composite material in which the process has been repeated one more time only. Figure 12 shows the first and third arrays of CNTs as 5X5 square arrays (with square outer shapes) and the second array of CNTs as a 4X4 square array (with a square outer shape). In general each subsequent array can be grown with the same outer shape or more typically a different outer shape to other arrays. Building up a series of reinforcement layers with different shapes under the control of a three-dimensional computer model of a part enables a "net shape" part to be grown by additive fabrication. Similarly each layer of matrix can be cured selectively with a different shape under the control of a three-dimensional computer model to match the shape of the array which it is impregnating.

Figure 13 is a cross-sectional view taken along a line A- A in Figure 12. The bases 14 of the first array of bundles lie in the plane A-A of the cross-section. The bases 14a of the second array of bundles lie in a second plane which is parallel but not coplanar with the plane A-A. As can be seen in Figure 13, the bases 14 overlap with the bases 14a when viewed at a right angle to the plane A-A. Figure 13 also shows that the second array is offset horizontally in two directions relative to the first array.

Figure 14 is a side view of an alternative composite material in which the second array of CNT bundles are tilted in an opposite sense to the CNT bundles of the first array. Figure 15 is a plan view of the composite material showing the bases 22 of the second array in solid line and the bases 14 of the first array in dotted line. In this case the arrays are only offset horizontally in one direction (the left/right direction in Figure 15). This results in an interlocking or keying interaction between the stalks 23 of the second array and the bases 14 of the first array (since they both lie in the plane of Figure 14).

Figure 16 is a side view of a two-layer composite film comprising a first array of bundles of CNTs 40 which have been grown and densified on the substrate 4 by the process described above and placed on the build platform 3 with their tips pointing up. A second array of bundles of CNTs 41 have been grown and densified on a second substrate 42 and dipped into a liquid layer of matrix material 43 in the interstitial gaps between the bundles of the first array. The substrate 42 is then removed and the matrix material 43 cured. The stalks of the second array of bundles 41 are positioned in the inter-stalk gaps 44 between the stalks of the first array of bundles 40. The stalks of the first array 40 are generally oppositely oriented to the stalks of the second array 41, although the magnitude of tilt may be varied slightly between the arrays to they are not precisely oppositely (180°) oriented.

The first and second arrays in Figure 16 overlap by 100%. Due to the 100% overlap between the layers the composite material of Figure 16 has a total thickness which is equal to the individual thickness of each filament layer. Figure 17 shows an alternative arrangement in which there is less than 100% overlap between the arrays 40,41. Due to the partial overlap between the layers the composite material of Figure 17 has a total thickness which is greater than the individual thickness of each filament layer.

Figures 18-22 show an alternative system in which the CNTs are grown in-situ. The apparatus shown in Figure 18 is housed within a process chamber (not shown). A negative plasma source electrode 52 and a positive plasma source electrode 53 are connected by a power source 54. A laser 55 is positioned above the positive plasma source 53, and is associated with a raster scanning mechanism (not shown). A gas supply 56 can be turned on and off to supply a pre-heated process gas to the chamber, such as CH₄/H₂. A second gas supply 57 can be turned on and off to supply an inert gas such as N₂ to the process chamber. The inert gas is preheated to a temperature at or just below the melting point of the matrix material. The electrode 52 is also heated by a heating element (not shown) to a similar temperature.

A heated hopper 58 and a cooled ink-jet printing head 59 are mounted on a transport mechanism (not shown) which can move the hopper 58 and printing head 59 from left to right in Figure 18 (that is, from one end of the negative plasma source 52 to the other). A transport mechanism (not shown) is provided for driving the negative plasma source 52 up and down.

In a first process step, the hopper 58 is filled with a polymer powder such as polyetheretherketone (PEEK). The hopper 58 is moved across the negative plasma source 52, and a dispensing orifice (not shown) in the hopper 58 is opened to deposit a layer 60 of polymer powder. Thus the source 52 also acts as a bed or platform for the additive layer manufacturing process. The orifice is then closed. The inert gas prevents oxidation of the polymer. The laser 55 is turned on and the raster mechanism scans the beam across the layer 60 to consolidate the layer 60. The heating effect of the laser beam causes the polymer layer 60 to melt. A shutter (not shown) in the path of the laser beam is opened and closed selectively to modulate the beam as it is scanned over the layer 60. Thus the layer 60 is consolidated only in the areas required to form a desired shape. More specifically, the shutter is opened and closed in accordance with a computer-aided design (CAD) model which defines a series of slices through the desired three-dimensional shape.

In a second process step, the printing head 59 is moved across the layer 60 to deposit a patterned array of catalyst particles. The printing head 59 sprays an array of colloid drops onto the layer 60, and as the colloid evaporates in the high temperature inert gas environment, metal catalyst particles suspended in the colloid drops are deposited. The catalyst particles may be, for example a metal, preferably transition metals Fe, Ni or Co, or alloys thereof; and the colloid liquid may be, for example alcohol, water, oil, or a mixture thereof. A fluid-based cooling system (not shown) cools the printing head 59 and a reservoir (not shown) containing the printing fluid to prevent the colloid liquid from boiling before it is printed. The printing orifice of the printing head 59 (which emits the spray of droplets) is positioned sufficiently close to the layer 60 to ensure that the colloid liquid does not evaporate deleteriously in flight, before hitting the layer 60.

Optionally the catalyst material may be conditioned as part of the second process step, through a process of spherulisation and/or oxidation and/or reduction, depending on the catalyst type. This conditioning is performed by the combination of heating and supply of an oxidising and/or reducing gas, depending on the catalyst type

In a third process step, carbonaceous feed stock is introduced from the gas supply 56 and the power source 54 is turned on to generate a plasma between the electrodes 52, 53. This causes the in-situ growth of a layer of nanofibres, aligned with the direction of the electromagnetic field between the electrodes 52,53. The growth mechanism is as described by Baker (Baker, R.T.K., Barber, M.A., Harris, P.S., Feates, F.S. & Waire, R.J. J J Catal 26 (1972). It is generally accepted that the carbonaceous gas is dissociated on the surface of the metal catalyst particle and carbon is adsorbed onto the surface where it is then transported by diffusion to the precipitating face forming a carbon filament with the catalyst particle at the tip. Discussion is ongoing with regards to whether this diffusion is through the bulk of the catalyst or along its surface(s) and to whether the diffusion is driven by a carbon concentration or thermal gradient. Thus when the growth process is complete, a "forest" of nanofibres bundles 62 is produced, each nanofibre carrying a catalyst particle at its tip.

The catalyst particles and plasma enable the nanofibre growth to occur at a relatively low temperature, lower than the melting point of the matrix.

Once nanofibres of a suitable length have been grown, the plasma power source 54 and gas supply 56 are turned off, the inert gas is purged, and in a fourth process step the CNT bundles 62 are densified by the capillary forming method described above, reducing the diameter of the CNT bundles as shown in Figure 19.

Next the platform 52 is lowered and the hopper 58 is moved along the layer of nanofibres to deposit a further layer 63 of polymer powder.

In a fifth process step, the laser 55 is turned on and the raster mechanism scans the beam across the layer 63 to form a consolidated layer 63 ' shown in Figure 19. During the raster scan, the shutter is opened and closed as required to form the consolidated layer 63 ' in a desired shape.

The thickness of the unconsolidated polymer layer 63 is selected so that the layer of CNTs is only partially impregnated with the matrix through a lower part of its thickness, leaving an upper part of the layer of CNTs exposed as shown in Figure 19. By way of example, the thickness of the unconsolidated layer 63 shown in Figure 18 may be in the range of 0.2- 0.5 mm, and the thickness of the consolidated layer 63' shown in Figure 19 may be in the range of 0.1-0.25 mm. Although the ratio between the length of the CNT bundles 62 and the thickness of the consolidated layer 63' is of the order of 2: 1 in Figure 19, this is for illustrative purposes only and in practice a much smaller degree of overlap (for instance a ratio of 1.05 : 1) will be required to give significant interlayer reinforcement. Next a patterned layer of catalyst is deposited on the cured matrix layer 63' in the gaps 65 between the bundles 62, and a second array of CNT bundles 66 is grown on the catalyst regions as shown in Figure 20. Next the CNT bundles 66 are densified by the capillary forming method described above, reducing the diameter of the CNT bundles as shown in Figure 21. The inter- filament gaps and interstitial gaps of both arrays are then impregnated with a second layer of matrix material 67 as shown in Figure 22, and the process can be repeated indefinitely to produce a bulk composite material.

In the catalyst deposition pattern shown in Figure 3, the orientation of the catalyst regions 12 is constant across the deposition region. Figure 23 shows a catalyst deposition pattern in which the orientation of the catalyst regions 70 vary across the deposition region. The spacing between the catalyst regions is tailored to avoid any clash as the second array (not shown) is inserted. Figure 23 shows a pattern in which only the orientation of the regions 70 varies across the deposition region, but optionally the shape of the regions may also vary.

In the catalyst deposition patterns described above, the catalyst regions 12 have mirror symmetry but lack rotational symmetry. This lack of rotational symmetry causes the CNT bundles to tilt during capillary forming. Figure 24 shows a catalyst region 71 which has rotational symmetry but lacks mirror symmetry. This lack of mirror symmetry causes the CNT bundles to twist during capillary forming. The rotational symmetry results in the centroid being symmetrically positioned so the CNT bundles do not tilt during capillary forming. The castalyst region shapes described above can be replaced with the shape shown in Figure 24, resulting in a composite material with two or more layers of CNT bundles, where each layer may twist in the same sense (clockwise or anticlockwise) or the twisting of the bundles can be tailored between layers or across layers to achieve a desired mechanical or electromagnetic property.

Other shapes are possible, including shapes which lack both mirror and rotational symmetry. Although the invention has been described above with reference to one or more preferred embodiments, it will be appreciated that various changes or modifications may be made without departing from the scope of the invention as defined in the appended claims.

Claims

1. A method of manufacturing a composite material, the method comprising: a. depositing a catalyst material on a substrate, the catalyst material being patterned to form an array of catalyst regions, and at least some of the catalyst regions being asymmetrical regions having an outer periphery which has no rotational symmetry and/or no mirror symmetry; b. growing a first array of bundles of filaments on the substrate, growth of the filaments being catalysed by the catalyst material, each bundle and each filament having a base attached to one of the catalyst regions and a free tip, each bundle comprising a plurality of filaments spaced apart from adjacent bundles in the array by an interstitial gap; c. impregnating the bundles of filaments with a capillary forming liquid; d. evaporating the capillary forming liquid so that the free tips of the filaments are drawn together within each bundle and the bundles grown on the asymmetrical catalyst regions become distorted; e. repeating steps a. to d. to provide a second array of bundles of filaments; f. positioning or growing the arrays so that the bundles of one array are positioned at least partially within the interstitial gaps of the other array; and g. impregnating the bundles of filaments and the interstitial gaps with a matrix material.

2. The method of claim 1 wherein step f. comprises dipping the free tips of the second array into a layer of liquid matrix material in the interstitial gaps of the first array.

3. The method of claim 2 further comprising: transferring the first array onto a part bed with the tips adjacent the part bed and the bases remote from the part bed; and impregnating the interstitial gaps of the first array with the layer of liquid matrix material after it has been transferred to the part bed.

4. The method of claim 2 or 3 further comprising repeating steps a. to d. to provide a third array of bundles of filaments; impregnating the interstitial gaps of the second array with a second layer of liquid matrix material; and dipping the free tips of the third array into the second layer of liquid matrix material in the interstitial gaps of the second array.

5. The method of any of claims 2 to 4 further comprising curing the liquid matrix material by scanning a radiation beam across it.

6. The method of any of claims 2 to 5 wherein the tips of the second array extend in a plane and move at an acute angle to that plane as they are dipped into the layer of matrix material.

7. The method of any preceding claim wherein each bundle has a centre line between the centre of its base and the centre of its tip, and wherein the distortion of the bundle during step d. causes the centre line to tilt and/or causes the bundle to twist.

8. The method of any preceding claim wherein the distortion of the bundle during step d. causes the bundle to twist.

9. The method of any preceding claim wherein each asymmetrical catalyst region has an outer periphery which has no rotational symmetry.

10. The method of any preceding claim wherein each asymmetrical catalyst region has an outer periphery which has no mirror symmetry.

11. The method of any preceding claim wherein the shape and/or orientation of the catalyst regions varies between the first array and the second array.

12. The method of any preceding claim wherein the shape and/or orientation of the catalyst regions varies across the first array and/or the second array.

13. The method of any preceding claim wherein the number of catalyst regions in the first array is different to the number of catalyst regions in the second array.

14. A composite material comprising: a. a first filament layer comprising an array of bundles of filaments, each bundle being spaced apart from adjacent bundles by an interstitial gap; b. a second filament layer comprising an array of bundles of filaments, each bundle being spaced apart from adjacent bundles by an interstitial gap; and c. a matrix material occupying the interstitial gaps between the bundles; wherein each bundle and each filament has a base, a stalk and a tip, and each filament is spaced apart from adjacent filaments in the bundle by an inter- filament gap which is occupied by the matrix material and is smaller at the tip of the filament than at the base of the filament; wherein the first and second filament layers are partially overlapped with each other so that the stalks of one array are positioned with at least part of their length within the interstitial gaps of the other layer and pass between either the tips or bases of the other layer; and wherein the base of at least some of the bundles has an outer periphery which has no rotational symmetry and/or no mirror symmetry.

15. A composite material comprising: a. a first array of bundles of filaments, each bundle being spaced apart from adjacent bundles by an interstitial gap; b. a second array of bundles of filaments, each bundle being spaced apart from adjacent bundles by an interstitial gap; and c. a matrix material occupying the interstitial gaps between the bundles; wherein each bundle and each filament has a base and a tip, and each filament is spaced apart from adjacent filaments in the bundle by an inter- filament gap which is occupied by the matrix material and is smaller at the tip of the filament than at the base of the filament; wherein the first and second filament layers are at least partially overlapped with each other so that the stalks of one array are positioned with at least part of their length within the interstitial gaps of the other layer; wherein the stalks of the first array are generally oppositely oriented to the stalks of the second array; and wherein the base of at least some of the bundles has an outer periphery which has no rotational symmetry and/or no mirror symmetry.

16. The material of claim 14 or 15 wherein the stalks of one layer pass between the tips of the other layer.

17. The material of claim 14 or 15 wherein the stalks of one layer pass between the bases of the other layer.

18. The material of any of claims 14 to 17 wherein each bundle has a centre line between the centre of its base and the centre of its tip, and wherein each asymmetrical bundle is twisted and/or its centre line is tilted relative to a line passing through the centre of the base at a right angle to the base.

19. The material of any of claims 14 to 18 wherein each asymmetrical bundle is twisted.

20. The material of any of claims 14 to 19 wherein the asymmetrical bundles are tilted or twisted in an opposite direction between the arrays and/or between the bundles of one of the arrays.

21. The material of claim 20 wherein the magnitude of tilt of the asymmetrical bundles varies between the arrays and/or between the bundles of one of the arrays.

22. The material of any of claims 14 to 21 wherein the number of bundles in the first array is different to the number of bundles in the second array.