GB2556051A - Methods and apparatus for use in producing carbon nanotube/metal composites - Google Patents

Abstract

A metalorganic chemical vapour deposition method of functionalising carbon nanotubes comprises arranging a quantity of nanotubes 36 in contact with an organometallic compound of a transition metal 38 on a substrate 34 in a reaction chamber 12, purging the chamber, filling with a process gas 28 and heating the chamber to a temperature equal to or greater than the melting point of the organometallic compound. The transition metal may be nickel or chromium and the compound may be nickelocene, dibenzenechromium or chromocene. The compound is preferably arranged on top of the nanotubes. The process gas may be a mixture of argon and hydrogen or a mixture of helium and hydrogen. A method of forming a metal matrix composite comprises disposing the functionalised carbon nanotubes within molten metal and solidifying the metal. Also disclosed is an apparatus suitable for performing said method. A method of plating copper particles with a transition metal comprises mixing the copper particles with an aqueous solution comprising a source of a transition metal, a reducing agent to produce ions of the transition metal and a complexing agent.

Description

(54) Title of the Invention: Methods and apparatus for use in producing carbon nanotube/metal composites

Abstract Title: CVD methods and apparatus for producing functionalised carbon nanotubes and metal matrix composites comprising said nanotubes (57) A metalorganic chemical vapour deposition method of functionalising carbon nanotubes comprises arranging a quantity of nanotubes 36 in contact with an organometallic compound of a transition metal 38 on a substrate 34 in a reaction chamber 12, purging the chamber, filling with a process gas 28 and heating the chamber to a temperature equal to or greater than the melting point of the organometallic compound. The transition metal may be nickel or chromium and the compound may be nickelocene, dibenzenechromium or chromocene. The compound is preferably arranged on top of the nanotubes. The process gas may be a mixture of argon and hydrogen or a mixture of helium and hydrogen. A method of forming a metal matrix composite comprises disposing the functionalised carbon nanotubes within molten metal and solidifying the metal. Also disclosed is an apparatus suitable for performing said method. A method of plating copper particles with a transition metal comprises mixing the copper particles with an aqueous solution comprising a source of a transition metal, a reducing agent to produce ions of the transition metal and a complexing agent.

1/16

Figure 1

2/16

Figure 3

3/16

Figure 4

4Γ\6

Figure 5

Figure 6

Figure 7

5/16

Figure 8

6/16

TEMPERATURE

-»τγγ^γγτύ^»ττγτ^»ϊτ !·νιντ^τινττΐϊττΓττιχτ,ττ·π~ΤΓΓϊΎΤΓί

Jr φ^35?ζφ35^Γ^Ι

Φ % ΪΗ9Ι3Μ 31dlAIVS

7/16 (a) (b)

Figure 10

Figure 11

8/16

%'1H9I3M 31dlAIVS

TEMPERATURE

9/16

Figure 13

a) functionalised CNTs carpet

0 00 0 <wa<v

I b b g ο.

0 0 00

II I J B < ) ( 5 f .) f 1 ( 8 1 B

()0()

Step I

FUNCTIONALIZATION OF CNTs VIA GF-MOCVD

b) c} surfactant

Step II

DEPOSITION OF CNTs

DISPERSION

Figure 14 ζ~\

V v, A >0 r>0 r-0

AP

O <> \

Step III

REMOVAL OF SURFACTANT

10/16

Figure 15

11/16

Figure 16

Figure 17

Figure 18

12/16

particles

Figure 19

13/16

Figure 20

14/16

Copper powder particles

Carbon nanotube functionalized using

GF-MOCVD (metal, e.g. Ni coated)

2)

Metal (e.g. Ni) coated carbon nanotubes (functionalized using GF-MOCVD) mixed with copper powder particles

3)

Cu/carbon nanotube interface (e.g. Ni)

Carbon nanotube

Figure 21

15/16

Pristine carbon nanotubes

Metal (e.g. Ni) coated copper powder particle

Cu/carbon nanotube interface (e.g. Ni)

Pristine carbon nanotubes mixed with metal (e.g. Ni) coated copper powder particles

Carbon nanotube

Figure 22

16/16

Figure 23

Figure 24

METHODS AND APPARATUS FOR USE IN PRODUCING CARBON NANOTUBE I METAL COMPOSITES

FIELD OF THE INVENTION

The present invention relates to methods and associated apparatus for use in the production of metal matrix composites in which carbon nanotubes (CNTs) are disposed within a metal matrix. By way of example, the metal matrix may comprise copper, although other metals may alternatively be used.

The term “copper” as used herein should be interpreted broadly, to encompass copper alloys as well as substantially pure copper. Similarly, with other matrix metals mentioned herein, the name of the metal should be interpreted broadly, to encompass alloys of the metal as well as the metal in a substantially pure form.

BACKGROUND TO THE INVENTION

So far, CNT/metal matrix composites have mainly been researched for structural applications, while their electrical potential still remains underestimated. For electrically conductive applications, CNT/metal matrix composites in which the metal comprises copper are of particular interest, not least since CNTs are considered to be more conductive than copper.

However, combining copper and CNTs in order to create an improved electrical (or thermal) conductor still remains a challenge. The main obstacle is contact resistance occurring at the interface between the CNTs and the metal, resulting from poor wettability of the CNTs by the metal. Indeed, it is known that most molten metals do not wet CNTs. There is therefore a desire to modify the CNTs or to alloy the matrix metal in order to improve the wettability of the CNTs by the molten metal, and thereby improve the electrical (or thermal) conductivity of the resulting composite.

Background art - control of interaction between carbon nanotubes and metals The main obstacle when it comes to combining the unique properties of CNTs with electrically and thermally conductive metals, like copper, is the contact resistance occurring at the CNT/metal interface, which may exceed 1kO and reduce the usability of CNT/metal matrix composites [3,4], In order to sustain contact at the interface between the CNTs and the metal matrix, wetting needs to be provided. It is known that most molten metals do not wet carbon nanotubes, and therefore modification is needed in order to enhance wettability and achieve better contact at the interface between the CNTs and the metal matrix [1,5],

Differences in surface energies between the metal and the CNTs are the key obstacle to achieving good contact at the interface, and result in the wetting (or non-wetting) phenomenon at the interface [4], With reference to Figure 1, wetting of CNTs occurs if the contact angle (Θ) between the CNT and a drop of liquid metal is below 90°. This follows from the Young-Laplace equation, where the pressure difference (Δ/’) is positive and drives forces inducing wetting [1,7], According to the Young-Laplace equation, the pressure difference (Δ/’) depends on the contact angle (Θ), the surface tension of the liquid (/), and the radius of curvature (r) - see equation (1) below [7], = (1) r

The surface tension value (/, or y_Lv in Figure 1) is a determining factor for the wettability of CNTs by metals. Wetting can be observed if the surface tension for the metal is compliant with a cut-off point which is between 100 and 200 mN/m [4,6,7], No wetting is observed if the surface tension of the metal is beyond the cut-off value [7], While the surface tension value for commonly used metal matrixes (such as copper, aluminium and iron) is above the range of wetting for CNTs (Cu-1270 mN/m, AI-865 mN/m, Fe-1700 mN/m), good electrical contact and strong adhesion cannot be provided [4], As the contact angle depends on the interfacial tensions and is not easily predictable, it can be presumed that, in order to provide good wetting, polarizability of a solid state substance needs to be higher than for the liquid [6],

It has also been reported in the literature that a difference in work function between CNTs and selected metals has a direct influence on the contact resistance as well. Contact resistance has not been implicated to have a correlation with the work function if good wettability is achieved (Ti, Cr, Fe); however if poor wettability is the case, it increases according to the growth of the work function difference (Au, Pd, Pt). What is more, for metals characterised by low work function and poor wettability (Hf, Ga, Al), surface oxides form very easily and play a more important role in contact resistance [8],

The work function value for as-purified multi-walled nanotubes is 4.3 eV. According to Ago [9] the electronic structure of the nanotube can be affected by oxidative treatments and has a direct influence on the work function value. The introduction of carbonyl and hydroxyl groups by oxidation of CNTs results in an increase of the work function value up to 4.4 eV, while plasma treatment and functionalization of CNTs in acids produce 4.8 eV and 5.1 eV work function values, respectively. Higher work function values after oxidative treatments are a consequence of the π-conjugation disruption and formation of surface dipole moments [9],

Contact resistance at the interface between CNTs and metals may be reduced by CNT doping, functionalization, or an interfacial layer catalysed by different metals. Chai [10] has managed to reduce the interface resistance by deposition of a graphitic-carbon interfacial layer coated subsequently with a nickel layer and heattreated in argon or hydrogen. The phenomenon has been attributed to good wettability of the carbon to both CNTs and metal, graphitization of amorphous carbon in the presence of the transition metal, bridging the vacuum gaps between the metal and CNT, a larger interface area, and more conduction channels [10], So far CNT/metal composites have been researched mainly for their improved mechanical properties; electrical properties of this type of composite have gained attention to a lesser extent [11],

Carbon-carbon covalent bonds are exceptionally strong and, together with the one-molecule structure of every nanotube, contribute to the unique strength of

CNTs [3], When their extremely low weight to specific-surface-area ratio is taken into account, CNTs appear to be promising with respect to the mechanical reinforcement of metals [2], However, the above-described wetting issue still needs to be addressed, as good adhesion between the metal matrix and the CNTs is important in order to sustain mechanical load transfer [2,4], Reactions at the interface between the CNTs and the metal matrix, and chemical stability of the CNTs, as well as the degree of wetting of the CNTs, have been associated directly with final properties of the composite [11],

In order to improve interaction between CNTs and metals, and subsequently reduce contact resistance at the interface, two general approaches may be considered. As most metals do not wet CNTs, the carbon structure may be functionalized in order to enhance wetting and reduce contact resistance. It has been reported that, in case of CNT/metal composites, where no wetting occurs, functionalization of CNTs with other metal moieties may enhance the interfacial bonding [11], The second approach is to enhance the wettability by modification of the metal matrix to enable wetting of CNTs. Matrix-alloying methods in which alloying is carried out using metals having a strong affinity to form carbides, like Cr or Al, has been found to result in improved interfacial contact between CNTs and the metal matrix, as well as improved mechanical strength and hardness [12,13],

While as-synthesized CNTs do not dissolve in organic solvents or aqueous solutions and, due to van der Waals forces, tend to agglomerate in composites, functionalization of CNTs is very important in order to obtain a uniform dispersion and enable wetting by metals [3], If good wettability is sustained, CNTs do not agglomerate in the metal matrix [11],

CNT/metal composites may be manufactured in a number of ways, though temperature of the processing needs to be tuned in order to avoid damage to the CNTs [4], For this reason not all conventional metallurgical processes, especially those involving molten metal processing, may be applied when producing CNT/metal composites [4], Commonly used processes for CNT/metal composite manufacturing include thin-film evaporation, electrodeposition, electrostatic interactions, hot pressing, sputtering, chemical reduction, electroless plating, extrusion of powders, chemical vapour deposition, in situ deposition, ex situ deposition, powder metallurgy, sandwich processing, casting and many others [4,11,13], However, due to difficulties in control over nanometre scale processing and the resulting properties of the composites, manufacture of the composites (which have great potential application in modern electronics, e.g. for interconnects, transistors, optoelectronics, sensing and high performance nanodevices) still remains a challenge [14],

SUMMARY OF THE INVENTION

According to a first aspect of the present invention there is provided an improved metalorganic chemical vapour deposition (MOCVD) method of functionalizing carbon nanotubes with a transition metal in order to enhance their wettability, the method comprising: arranging, on a substrate within a reaction chamber, a quantity of carbon nanotubes in contact with an organometallic compound of the transition metal; purging oxygen from the reaction chamber and filling the reaction chamber with a process gas; and heating the reaction chamber to a process temperature equal to or greater than the melting point of the organometallic compound; whereby a reaction takes place in which the organometallic compound releases the transition metal and the transition metal functionalizes the carbon nanotubes. In the present work this improved MOCVD method is referred to as “Gravity Fed MOCVD” or “GF-MOCVD”.

By virtue of the carbon nanotubes being in contact with the organometallic compound, this enables the carbon nanotubes to be functionalized in an efficient manner, without the need for pre-evaporation of the transition metal, and without the need for the transition metal vapour to be transported from an evaporator into the deposition chamber. Furthermore, the present method eliminates cooling of the transition metal vapour before the CNTs are reached, and also provides high vapour concentration in the direct vicinity of the CNTs. Moreover the present method does not require the use of a vacuum pump, as deposition can be carried out at atmospheric pressure.

In presently-preferred embodiments the transition metal is nickel. For example, the organometallic compound may be nickelocene.

However, in alternative embodiments the transition metal may be chromium. For example, the organometallic compound may be dibenzenechromium or chromocene.

In yet further alternative embodiments the transition metal may be titanium, vanadium, manganese, iron, cobalt, zirconium, niobium or molybdenum, for example.

According to presently-preferred embodiments the carbon nanotubes may be arranged in the form of an aligned array or carpet.

Preferably the organometallic compound is arranged on top of the carbon nanotubes in the reaction chamber, thereby exploiting the effect of gravity acting on the organometallic compound.

The process gas may comprise, for example, a mixture of argon and hydrogen, or a mixture of helium and hydrogen.

Preferably the method further comprises maintaining a flow of the process gas through the reaction chamber during the reaction.

More particularly, in presently-preferred embodiments, the substrate is porous, and the process gas flows through the substrate during the reaction, thereby providing a relatively unhindered flow of gas through the reaction chamber. However, in alternative embodiments the substrate may be non-porous, as long as suitable gas channels or the like are provided, to maintain a relatively unhindered flow of gas through the reaction chamber during the reaction.

In one embodiment the process gas flows upwards through the substrate during the reaction. In another embodiment the process gas flows downwards through the substrate during the reaction.

By way of example, the substrate may be made of a porous ceramic material or a porous glass material (e.g. porous borosilicate glass).

The process gas may be used for purging oxygen from the reaction chamber, as well as for the reaction process. Alternatively, a different gas, or a vacuum pump, may be used for purging the reaction chamber.

The process temperature may be, for example, one of: T_meit + 0°C, T_meit + 10°C, T_meit + 20°C, T_meit + 40°C, T_mei_t + 80°C; where T_mei_t is the melting point of the organometallic compound.

Advantageously, the reaction may take place under atmospheric pressure. Alternatively, it may take place under sub-atmospheric pressure or a vacuum, or under slight super-atmospheric pressure.

Having functionalized the carbon nanotubes with the transition metal to improve their wettability, the functionalized carbon nanotubes may be disposed within molten metal (e.g. molten copper), which may then be solidified to form a metal matrix composite. Alternatively, a powder metallurgy route may be used to form a metal matrix composite, whereby the functionalized carbon nanotubes are mixed with metal powder (e.g. copper powder) to produce a mixture, and then the mixture is compacted and sintered to produce the composite. Either way, the present work enables metal matrix composites (particularly, but not exclusively, Cu-CNT composites) to be produced that have good interfacial bonding between the metal matrix and the CNTs.

According to a second aspect of the present invention functionalized carbon nanotubes are provided, produced by a method in accordance with the first aspect of the invention.

According to a third aspect of the present invention there is provided a metal matrix composite produced by a method in accordance with the first aspect of the invention, or using functionalized carbon nanotubes in accordance with the second aspect of the invention.

According to a fourth aspect of the present invention there is provided metalorganic chemical vapour deposition apparatus (referred to herein as GFMOCVD apparatus) for functionalizing carbon nanotubes with a transition metal, the apparatus comprising: a reaction chamber; a substrate inside the reaction chamber, for holding a quantity of carbon nanotubes in contact with a quantity of an organometallic compound; one or more heaters for heating the reaction chamber to a process temperature equal to or greater than the melting point of the organometallic compound; and a supply of a process gas; wherein the reaction chamber further comprises a gas inlet for receiving the process gas from the gas supply, and an exhaust through which gas can leave the reaction chamber.

As discussed above in relation to the first aspect of the invention, preferably the substrate is porous. For example, the substrate may be made of a porous ceramic material, or a porous glass material (e.g. porous borosilicate glass).

In one embodiment the gas inlet is beneath the substrate and the exhaust is above the substrate, such that the process gas flows upwards through the substrate during the reaction.

In another embodiment the gas inlet is above the substrate and the exhaust is below the substrate, such that the process gas flows downwards through the substrate during the reaction.

As an alternative technique for improving the wettability of CNTs by a molten metal, according to a fifth aspect of the present invention there is provided a method of plating copper particles with a transition metal, the method comprising mixing the copper particles with an aqueous solution comprising: a source of a transition metal, a reducing agent to produce ions of the transition metal, and a complexing agent.

By plating the copper particles with a transition metal in this manner, this improves the wetting of carbon nanotubes.

In accordance with presently-preferred embodiments, the transition metal is nickel. The source of the transition metal may be nickel sulphate. The reducing agent may be sodium hypophosphite. The complexing agent may be sodium citrate.

The method may further comprise heating the solution to a temperature between 30°C and 90°C, for example to a temperature of around 80°C.

The method may further comprise decanting and rinsing the plated copper particles.

The copper particles are preferably dendritic copper particles.

Having prepared the plated copper particles, the method may further comprise combining them with carbon nanotubes to form a metal matrix composite (e.g. via a powder metallurgy route, by mixing the plated copper particles and the carbon nanotubes to form a mixture, and then compacting and sintering the mixture to produce the composite).

According to a sixth aspect of the present invention plated copper particles are provided, produced by a method in accordance with the fifth aspect of the invention.

According to a seventh aspect of the present invention there is provided a metal matrix composite produced by a method in accordance with the fifth aspect of the invention, or using plated copper particles in accordance with the sixth aspect of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention will now be described, by way of example only, and with reference to the drawings in which:

Figure 1 is a schematic diagram showing the contact angle between a drop of molten metal and the surface of a carbon nanotube (L = molten metal, S = surface of CNT, γ = surface tension, Θ - contact angle);

Figure 2 is a scanning electron microscope (SEM) image of a CNT carpet after magnetron sputtering using a nickel target, showing that sputtering has limited ability to coat long, densely packed nanotubes;

Figure 3 is a schematic diagram of an embodiment of the present inventors’ Gravity Fed Metalorganic CVD (“GF-MOCVD”) apparatus for metal coating of CNTs;

Figure 4 is an EDX map of a CNT carpet Ni-coated via two variants of MOCVD:

(a) a conventional approach based on transportation of Ni vapour to the CNT carpet via a carrier gas, characterised by very low deposition efficiency; and (b) the present GF-MOCVD involving short-range vapour transportation for direct metallocene/CNT contact;

Figure 5 is an EDX map of a CNT carpet Ni-coated by the GF-MOCVD process and SEM images of individual nanotubes in different areas of the CNT carpet (obtained using Parameters No. 4, Table 1);

Figure 6 is an EDX map of a CNT carpet Ni-coated by the GF-MOCVD process and SEM images of individual nanotubes in different areas of the CNT carpet (obtained using Parameters No. 5, Table 1);

Figure 7 is an EDX map of a CNT carpet Ni-coated by the GF-MOCVD process and SEM images of individual nanotubes in different areas of the CNT carpet (obtained using Parameters No. 6, Table 1);

Figure 8 is an EDX spectrum and associated quantitative analysis recorded from a central part of a Ni-coated CNT carpet (obtained using Parameters No. 6, Table

1);

Figure 9 is a thermogravimetric analysis (TGA) graph of Ni-coated CNT (obtained using Parameters No. 6, Table 1);

Figure 10 shows SEM images of Ni-functionalised CNTs via GF-MOCVD (a) immediately after functionalization, and (b) after vacuum post heat-treatment;

Figure 11 shows SEM images made after sonication of Ni-functionalised CNTs via GF-MOCVD (a) immediately after functionalization, and (b) after vacuum post heat-treatment;

Figure 12 is a TGA graph obtained for a water-based dispersion comprising ascoated Ni-MWCNTs and surfactant, the graph illustrating data recorded during analysis performed in an argon atmosphere;

Figure 13 shows SEM images of nickel functionalized CNT coatings deposited by spraying on quartz slides (a) before surfactant removal, and (b) after surfactant removal;

Figure 14 illustrates steps for the preparation of samples as used for evaluation of electrical properties of functionalised CNTs;

Figure 15 presents (a) an EDX map of a CNT carpet Cr-coated by the GF-MOCVD process (obtained using Parameters No. 3, Table 6), and (b) an SEM image of the top part of this Cr-coated CNT carpet;

Figure 16 is an EDX map of a CNT carpet Cr-coated by the GF-MOCVD process and SEM images of individual nanotubes in different areas of the CNT carpet (obtained using Parameters No. 1, Table 7);

Figure 17 is an EDX map of a CNT carpet Cr-coated by the GF-MOCVD process and SEM images of individual nanotubes in different areas of the CNT carpet (obtained using Parameters No. 2, Table 7);

Figure 18 is an EDX map of a CNT carpet Cr-coated by the GF-MOCVD process and SEM images of individual nanotubes in different areas of the CNT carpet (obtained using Parameters No. 5, Table 7);

Figure 19 illustrates the formation of a Cu/CNT composite using a powder metallurgy route, where the feedstock consists of pristine carbon nanotubes and dendritic copper powder, and showing that, due to poor Cu/CNT interaction, after compacting and sintering, there are voids at the Cu/CNT interface;

Figure 20 is an exemplary SEM image showing voids at the Cu/CNT interface of a powder metallurgical sample obtained by mixing, compacting and sintering of pristine CNT and dendritic Cu powder;

Figure 21 illustrates improved Cu/CNT interaction resulting from deposition of selected transition metals (carbide forming, such as Ti or Cr, or dissolving relatively large amounts of carbon, such as Ni) on carbon nanotubes via the GFMOCVD process;

Figure 22 illustrates improved Cu/CNT interaction resulting from deposition of selected transition metals (carbide forming, such as Ti or Cr, or dissolving relatively large amounts of carbon, such as Ni) on copper powder particles via immersion/electroless plating;

Figure 23 shows typical SEM images of dendritic copper powder Ni plated using a thiosulphate bath; and

Figure 24 shows typical SEM images of dendritic copper powder Ni plated using a hypophosphite bath.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

The present embodiments represent the best ways known to the Applicant of putting the invention into practice. However, they are not the only ways in which this can be achieved.

The present work provides two alternative routes for improving the wettability of CNTs by a molten metal (in particular, but not exclusively, copper), and thereby improving the electrical (or thermal) conductivity of the resulting CNT/metal composite. An improvement in the mechanical properties of the CNT/metal composite, such as ultimate tensile strength and Young’s modulus, can also be achieved in such a way. The two alternative routes are:

(i) functionalization of the CNTs with a transition metal using the so-called Gravity Fed Metalorganic Chemical Vapour Deposition (GF-MOCVD) technique; and (ii) plating of matrix metal particles, in particular dendritic copper particles, with a transition metal, to improve their ability to bond to CNTs.

Route (i) - Functionalization of CNTs by Gravity Fed Metalorganic Chemical

Vapour Deposition (GF-MOCVD)

A first route for improving the wettability of CNTs by molten metal is a functionalization technique for CNTs called Gravity Fed Metalorganic Chemical Vapour Deposition (GF-MOCVD).

As set out below, experimental work has proven the effectiveness of this method and the usefulness of functionalizing carbon nanotubes with selected transition metals in order to achieve a reduction in contact resistance.

Further background

By way of further background, sputtering was considered as a possible technique for functionalizing CNTs. However, sputtering has a number of limitations, including low efficiency, the small surface area which can be treated, and the lack of possibility to cover 3-dimentional objects like CNT carpets, to name a few. In some cases, specially designed sputtering chambers allow coating of powders (also CNTs). However, the efficiency of this method remains very low and, in the case of CNT agglomerates/chunks/carpets, does not provide coating uniformity (i.e. an inability to coat nanotubes inside of a bigger assembly). This is illustrated in Figure 2, which is an SEM image of a CNT carpet after magnetron sputtering using a nickel target. Region A of the image denotes the inner part of the CNT carpet, that is not affected by sputtering. Region B donates the sputtered surface of the CNT carpet. Thus, it can be seen that sputtering has limited ability to coat long, densely packed nanotubes, as only the top side of the CNT carpet is metal coated after sputtering.

Gravity Fed Metalorganic Chemical Vapour Deposition (GF-MOCVD)

With all of the limitations of sputtering in mind, the present inventors’ Gravity Fed Metalorganic Chemical Vapour Deposition (GF-MOCVD) method was developed. The GF-MOCVD method is a modification of classic Metalorganic Chemical Vapour Deposition process (MOCVD), which has been reported to be suitable for functionalization of CNTs for electrical, thermal and mechanical applications [97],

Metalorganic chemical vapour deposition (MOCVD), also known as metalorganic vapour phase epitaxy (MOVPE) or organometallic vapour phase epitaxy (OMVPE), is a chemical vapour deposition method used for formation of thin metal films, where organometallic precursors are used as a source of metal. Precursors are transferred via gas to the reaction chamber. Thin single crystal (epitaxial) films are grown at the substrate surface as a result of chemical reaction. MOCVD utilizes subatmospheric pressure, which makes the transport of precursor vapours to the reaction chamber possible [15],

There are few precursors which can be used in MOCVD - amongst them are sandwich compounds i.e. metallocenes (bis(cyclopentadienyl)metal complexes), bis(benzene) complexes and other arenes, those containing inorganic ligands as well [16-18], Literature presents just a few examples of successful deposition of metal films using MOCVD; however its simplicity is encouraging for more detailed investigation.

The present Gravity Fed Metalorganic Chemical Vapour Deposition method simplifies classic MOCVD by elimination of precursor pre-evaporation and therefore does not require separation of transporting and reactive gasses. Moreover the GF-MOCVD method does not require the use of a vacuum pump, as deposition is done at atmospheric pressure.

GF-MOCVD apparatus and method

An embodiment of the GF-MOCVD apparatus is illustrated in Figure 3. The GFMOCVD apparatus 10 comprises a reaction chamber 12 with one or more heaters 14, a temperature control unit 16 arranged to supply electrical power to the heaters 14, and a gas feeding system. The gas feeding system includes a gas supply 18 (a cylinder containing an inert or reducing gas or gas mixture - for example, a mixture of argon and hydrogen, or a mixture of helium and hydrogen) and a flowmeter or gas mass flow controller 20. A gas filtration unit 22 (e.g. comprising a gas washing bottle and chemical/HEPA filters) is connected to an exhaust 24 of the reaction chamber 12.

The reaction chamber 12 is gas tight and comprises two parts (a lower part 26 and an upper part 28) which are reversibly separable in order to enable access to the interior of the chamber 12. In use, the upper part 28 is fastened to the lower part 26 in a gas tight manner by means of flange 30, one or more seals (not shown) and bolts 32 (or other fastening means).

A porous substrate 34 is located inside the reaction chamber 12, and may, for example, be in the shape of a disc. The porous substrate 34 may be made, for example, of porous ceramic or porous glass, e.g. borosilicate glass.

The porous substrate 34 serves as a substrate for a quantity of CNTs 36 to be functionalized and a quantity of powder particles of an organometallic compound 38 of the desired transition metal that is to be applied to the CNTs 36 to functionalize them. For example, the organometallic compound may be nickelocene powder, such that nickel is the transition metal that is applied to the CNTs to functionalize them. As schematically illustrated, the CNTs 36 are preferably in the form of an aligned array.

The porous substrate 34 can be removed from reaction chamber 12 if needed, for example in order to load the samples, or for cleaning the chamber 12. The pore size of the substrate 34 should preferably be smaller than the particle size of the organometallic compound powder 38 or arrays/agglomerations of the carbon nanotubes 36. For example, the porosity grading of the substrate 34 may be 0 (pore size 160 to 250 pm), 1 (pore size 100 to 160 pm), 2 (pore size 40 to 100 pm), 3 (pore size 16 to 40 pm) or 4 (pore size 10 to 16 pm).

In use, first a quantity of carbon nanotubes 36 (preferably in form of an aligned array) is placed on the porous substrate 34, within the lower part 26 of the reaction chamber 12, and then a quantity of powder particles of the organometallic compound of the desired transition metal is placed directly on the top of the CNTs

36. The reaction chamber 12 is then assembled in a gas tight manner by fastening the upper part 28 in place.

In the presently-preferred embodiment, the lower part 26 of the reaction chamber 12 is connected to the gas feeding system including a flowmeter or gas mass flow controller 20. The porosity of the substrate 34 and the quantity of the organometallic compound powder 38 should be such as to enable stable gas flow through the chamber 12. The upper part 28 of the reaction chamber 12 (fastened to the lower part 26 in a gas tight manner) is connected to the gas filtration unit 22 via the reaction chamber exhaust 24.

In an alternative embodiment, the upper part 28 of the reaction chamber 12 is connected to the gas feeding system, and the lower part 26 of the reaction chamber 12 is connected to the exhaust 24 and gas filtration unit 22.

In both cases, the porous substrate 34 should be mounted horizontally, in order to prevent pouring of the powder 38 towards a side of the reaction chamber 12.

The assembled reaction chamber 12 is purged with the inert or reducing gas or gas mixture from the cylinder 18 (preferably Ar/H₂ or He/H₂ mixture) in order to remove oxygen. The gas flow rate should be limited to prevent scattering of the powder 38 placed on the porous substrate 34. In an alternative embodiment, oxygen removal may be performed using a vacuum pump.

After the reaction chamber 12 has been purged, the chamber 12 is fed with a process gas (e.g. Ar/H₂ or He/H₂ mixture) and heated to a target process temperature that is at least (and preferably exceeds) the melting point (T_meit) of the organometallic compound of the desired transition metal. Depending on the type of organometallic compound used, the process temperature can be considered as Tmeit + 0 C, T_meit + 10 C, T_meit ⁺ 20 C, Tmeit ⁺ 40 C, T_meit ⁺ 80 C, etc. The heating rate is unrestricted, although it is recommended to reach the process temperature in a short period of time.

At the process temperature, during the reaction process, the molten organometallic compound 38 bears down on the CNTs 36 due to the effect of gravity (hence the process being described as “gravity fed”). The organometallic compound releases the transition metal, which directly functionalizes the CNTs.

In the presently-preferred embodiment the reaction is carried out under atmospheric pressure, although in a modified embodiment the reaction may be carried out under sub-atmospheric pressure, a vacuum, or even slight superatmospheric pressure.

The duration of the reaction depends on type of organometallic compound used, and may vary between a couple of minutes and around an hour. After a specified length of time, the heating of the reaction chamber 12 is switched off. Gas flow is maintained until the reaction chamber 12 reaches room temperature. The upper and lower parts 28, 26 of the reaction chamber 12 can then be disassembled and the arrays/agglomerations of CNTs 36 (which are now functionalized with the transition metal from the organometallic compound) can be extracted from under the melted/reacted organometallic compound powder 38.

An important feature of the GF-MOCVD process (in comparison to conventional MOCVD processes) is the elimination of vapour transportation from an evaporator into the deposition chamber. In the GF-MOCVD process, metal vapour/CNT interaction takes place immediately after the reaction of the metal precursor covering the CNT carpet, in the hot zone of the reaction chamber. Such an approach eliminates vapour cooling before the CNTs are reached, and also provides high vapour concentration in the direct vicinity of the CNTs.

Thus, GF-MOCVD allows functionalization of CNTs, with the organometallic compound (a source of solid metallocenes) covering the top surface of the CNT carpet.

To illustrate the advantages of the present GF-MOCVD technique in comparison to conventional MOCVD techniques for the functionalization of CNTs, Figure 4 is an EDX map of a CNT carpet Ni-coated via (a) a conventional MOCVD approach based on transportation of Ni vapour to the CNT carpet via a carrier gas, characterised by very low deposition efficiency; and (b) the present GF-MOCVD approach involving short-range vapour transportation for direct metallocene/CNT contact.

Characterisation and analysis of functionalized CNTs

An investigation based on sonication of heat-treated copper-, nickel- and chromium-coated CNTs revealed strong Ni/CNT and strong Cr/CNT interaction, and very poor Cu/CNT interaction. Accordingly, testing of functionalized CNTs, as produced by the GF-MOCVD process, was focused on CNTs functionalized with nickel and chromium. The following section covers functionalization of the CNTs, their characterisation, and analysis of the usefulness of the GF-MOCVD process for reduction of contact resistance as a result of functionalization of the CNTs.

Nickel coating of CNTs

Assuming a strong interaction between CNTs and nickel (resulting from nickel’s ability to dissolve carbon), the use of the GF-MOCVD process to functionalize CNTs with nickel was tested.

In the case of deposition of nickel films using the classic MOCVD approach, a commonly used precursor is nickelocene, an organonickel compound with the formula Ni(C5H₅)2. Nickelocene is far less toxic than other nickel-containing metalorganic compounds, sufficiently volatile, relatively inexpensive, and can be used for deposition of relatively pure metallic coatings [19,20], The biggest drawback of nickel deposition via MOCVD is carbon incorporation, which, however, can be reduced by increasing the deposition pressure and the flow rate of hydrogen during the process [20,21], According to the literature, a deposition temperature of around 200°C results in the lowest carbon contamination of metal films as well as the lowest resistivity comparing to films deposited below 190°C and above 225°C [16], Moreover, addition of hydrogen enhances crystallinity and diminishes carbon incorporation [16], Literature reports that deposition of thin films utilising nickelocene would be impossible without the presence of hydrogen at a temperature below 350°C or even 550°C. Nickel deposition at a higher temperature without the aid of hydrogen is possible; however, in any case results in heavily carbon contaminated films [20], Relatively low contamination of nickel films with heteroatoms when using hydrogen makes nickelocene superior to other nickel-containing organometallic precursors.

The present GF-MOCVD process takes advantage of the same nickel precursor, nickelocene. However, parameters such as temperature, pressure and flow rate of the gas used are modified in order to enable metal deposition to take place at atmospheric pressure. The GF-MOCVD testing carried out involved analysis of the influence of various parameters such as temperature, gas flow rate and reaction duration on the efficiency of the CNT functionalization.

GF-MOCVD testing

A list of GF-MOCVD parameters used for functionalization of CNTs with nickel is provided in Table 1 below. During the course of all the experiments the heating rate was fixed at approximately 40°C per minute and 4% of hydrogen in argon was used as the reaction/carrier gas (the process gas).

- Parameters No. 1-3, Table 1

It was found that at a temperature below the melting point of nickelocene (which is 171-173°C [22]) there is no metal deposition regardless of the gas flow rate set.

- Parameters No. 4, Table 1

At a temperature corresponding to the nickelocene melting point, that is 172°C, and a gas flow rate typical for classic MOCVD process, that is 0.1L per minute, nickel deposition was obtained only in the top region of the CNT carpet, as demonstrated by elemental mapping and SEM images in Figure 5.

- Parameters No. 5-8, Table 1

It was found that increasing the gas flow rate to 0.2L per min results in infiltration of approximately half of the thickness of the CNT carpet (Figure 6), while a gas flow rate fixed at 0.7L per minute provides full infiltration of the CNTs carpets (Figure 7), and the highest weight fraction of deposited nickel (Figure 8). Further increase of the gas flow rate to 1L per minute deteriorates infiltration ability and leads to barely coated nanotubes. Using the same gas flow rate, it was also found that at 190°C, which is the minimum temperature recommended to use in classic MOCVD process, nanotubes were found not functionalised (Parameters No. 8 in Table 1).

Parameters No.	Sample name	Target temp, °C	Gas	Gas flow rate, L min'1	Holding at target temp, min
1	9 E,10 E	66		0.2
2	11 E	147		0.2
3	12 E	147		0.7
4	18 E	172		0.1
5	19 E	172	Ar-4%H2	0.2	30
6	13 E, 16 E, 17 E	172		0.7
7	15 E	172		1.0
8	14 E	190		1.0

Table 1 - Parameters of GF-MOCVD procedure which have been used to optimise the process.

Further investigation was focused on determining the time needed to carry out functionalization with results as obtained for Parameters No. 6, Table 1. It was found that even 5 minutes holding at the target temperature provides uniform deposition of nickel across the cross-section of the CNT carpet. A longer duration (that is, 10, 20 and 40 minutes) at the target temperature leads to a slight increase in the amount of deposited metal - see Table 2.

Parameters No.	Sample name	Target temp, °C	Gas	Gas flow rate, L/min	Holding at target temp, min
7	M5-25				0 min
8	M5-24				5 min
9	M5-19, M5- 22	172	Ar-4%H2	0.7	10 min
10	M5-17				20 min
11	M5-16				40 min

Table 2 - Parameters of GF-MOCVD procedure as a function of time.

Processing and further characterisation of functionalised CNTs

In order to complete the characterisation of the GF-MOCVD functionalised CNTs, the following experiments were carried out on the material produced using the optimal parameters, i.e. Parameters No. 6, Table 1:

1) Thermogravimetric analysis (TGA) in air.

2) SEM analysis after vacuum heat-treatment.

3) SEM analysis of as-functionalised and heat-treated nanotubes after sonication.

4) Evaluation of electrical properties of functionalised CNTs.

Based on TGA analysis for as-functionalised carbon nanotubes, produced in four 15 experiments using the same parameters, it can be concluded that the nickel weight fraction varies between 11 and 75wt% (Figure 9). The metal weight fraction variation was common for all four series with similar distribution within all batches.

An aim of further investigation was to define the influence of post heat-treatment on the morphology of the functionalised CNTs. It was found that after heating up to 800°C in vacuum (heating at 20°C per minute, no holding at target temperature) an initially uniform layer of metal was transformed into very fine nanoparticles, that appear to be penetrating the outer walls of multi-walled CNTs (Figure 10).

Next, as-synthesised and post heat-treated nickel functionalised CNTs were sonicated in order to evaluate the strength of the CNT/metal bonding and thereby the ability to disperse both types of nanotubes, as needed for evaluation of the electrical properties of the functionalised material. After 30 minutes sonication in acetone both types of functionalised nanotubes remained unaffected, which proves good metal/carbon bonding (Figure 11).

In order to enable resistance measurements, as functionalised CNT carpets and the same carpets following heat-treatment were used to make dispersions. Later on, coatings deposited on quartz slides were vacuum heat-treated in order to remove surfactant, which, based on TGA analysis, constituted approximately 60wt% of deposits (Figure 12). SEM images of nickel functionalized CNTs coatings sprayed on quartz slides before and after surfactant removal are presented in Figure 13.

A summary of all the above processes is schematically presented in Figure 14.

Resistance of coatings deposited using Ni-functionalized CNTs was measured after spraying (deposition), then surfactant removal and subsequently after heat treatment. Results of measurements done for thin and thick Ni-CNT coatings are presented in Table 3 and Table 4 respectively.

It was found that the resistance of as-sprayed Ni-CNT coatings was 10 times higher than measured for non-functionalised nanotubes; however, after surfactant removal the resistance of the functionalised nanotube coating was much lower than for pristine material. To give an illustration, surfactant removal in the case of non-functionalised nanotube coating resulted in 10-fold resistance decrease while in case of Ni-functionalised nanotubes the same heat-treatment resulted in over 130-fold and almost 300-fold resistance decrease for thin and thick coatings, respectively (Table 3 and Table 4). It can be presumed that, in the case of coatings containing nickel functionalised CNTs, the huge resistance decrease is an effect of reduction of contact resistance due to presence of nickel at the interface of CNTs, rather than surfactant removal exclusively.

Additional heat-treatment of the same coatings does not entail further significant 5 resistance decrease. Resistance decrease after post heat-treatment of thin and thick surfactant free coatings was measured as only 1.1% and 3.6% respectively.

Sample No.	Resistance
As sprayed, kO	Surfactant removed*	Additional HT**
kO	% decrease	kO	% decrease
1	226.3	1.778	99.21%	1.761	0.96%
2	173.0	1.465	99.15%	1.466	-0.07%
3	253.5	1.589	99.37%	1.558	1.95%
4	226.1	1.596	99.29%	1.553	2.69%
5	223.1	1.512	99.32%	1.507	0.33%
6	234.6	1.645	99.30%	1.629	0.97%
7	201.4	1.713	99.15%	1.695	1.05%
Arith. Avr.	219.7	1.614	99.26%	1.596	1.13%

*Heat-treatment in vacuum for 1 h at 800°C **Heat-treatment in vacuum; heating up to 800°C

Table 3 - Resistance of Ni-MWCNTs paint as sprayed, after holding for 1 hour at 800°C in vacuum (surfactant removal), and as a following step after heating up to 800°C without holding at the target temperature (MWCNTs = Multi-Walled CNTs).

Sample No.	Resistance
As sprayed, kO	Surfactant removed*	Additional HT**
kO	% decrease	kO	% decrease
1	219.9	0.716	99.67%	0.680	5.03%
2	212.4	0.627	99.70%	0.599	4.47%
3	221.2	0.676	99.69%	0.646	4.44%
4	132.1	0.654	99.50%	0.623	4.74%
5	179.7	0.615	99.66%	0.585	4.88%
6	170.8	0.627	99.63%	0.591	5.74%
7	140.5	0.618	99.56%	0.588	4.85%
8	186.1	0.699	99.62%	0.693	0.86%
9	212.1	0.729	99.66%	0.729	0.00%
10	254.1	0.762	99.70%	0.754	1.05%
Arith. Avr.	192.89	0.6723	99.64%	0.6	3.61%
;at-treatment in vacuum	for 1 h at 8	O o O o

**Heat-treatment in vacuum; heating up to 800°C

Table 4 - Resistance of thick layer of Ni-MWCNTs paint as sprayed, after holding 5 for 1 hour at 800°C in vacuum (surfactant removal) and as a following step after heating up to 800°C without holding at the target temperature.

Further investigation aimed to justify the usefulness of heat-treatment of GFMOCVD functionalised CNT carpets before their further processing, as needed for electrical measurements. Resistance of sprayed coatings comprising Nifunctionalised CNTs which were heat-treated before formation of dispersion for spraying was measured in several steps: right after spraying, after surfactant removal and subsequently after heating up to 800°C without holding at the target temperature. The high resistance of surfactant-free coatings, made using heat15 treated nanotubes (Table 5), compared to coatings consisting of not heat-treated nanotubes (Table 3 and Table 4), indicates an unfavourable influence of the above treatment.

Sample No.	Resistance
As sprayed, kO	Surfactant removed*	Additional HT**
kO	% decrease	kO	% decrease
1	93.0	4.64	95.01%	4.76	-2.56%
2	60.4	4.15	93.13%	4.24	-2.14%
3	41.5	3.41	91.79%	3.47	-1.79%
4	110.5	7.67	93.06%	7.89	-2.87%
5	169.0	6.22	96.32%	6.34	-1.93%
6	84.9	5.69	93.30%	5.86	-2.99%
7	118.8	5.98	94.97%	6.03	-0.84%
Arith. Avr.	96.9	5.39	93.94%	5.51	-2.16%

*Heat-treatment in vacuum for 1 h at 800°C **Heat-treatment in vacuum; heating up to 800°C

Table 5 - Resistance of sprayed coatings consisting of Ni-functionalised CNTs 5 which were heat-treated before formation of dispersion for spraying.

Measurements done for as sprayed coating, after holding for 1 h in 800°C in vacuum (surfactant removal) and as a following step after heating up to 800°C without holding at the target temperature.

Chromium coating of CNTs

Despite the fact that chromium sputtered samples were not characterised by lowered resistance, unlike samples sputtered with nickel, CNTs functionalized by chromium using the GF-MOCVD were tested in order to investigate the versatility of the GF-MOCVD method. Chromium functionalized CNTs, even if not suitable for electrical applications, may be adopted in metal matrix composites utilising other-than-electrical properties of carbon nanostructures. Here, chromium nanoparticles strongly bonded with outer walls of MWCNTs may improve load and heat transfer in thermal and structural applications [11],

Amongst various organometallic compounds, Cr(CO)6, Cr(C6H₆)2, Cr(C6H₅C3H7)2 and Cr(C5H₅)₂ are commonly used as a Cr precursor in the classic MOCVD process [23], Testing in the present work was done using chromocene Cr(C5H₅)₂ and dibenzenechromium Cr(C6H₆)2- Chromocene requires a higher reaction temperature (above 600°C) than dibenzenechromium (450°C) but provides a lower level of carbon contamination in the deposited metal films. Cr(CO)6 and

Cr(C6H₅C3H₇)2 compounds were not used due to their tendency to form Cr films heavily contaminated with oxygen and carbon, respectively [23],

GF-MOCVD investigation was initially tested using parameters which were acknowledged as optimal for nickel functionalization of carbon nanotubes, that is a temperature of 172°C (corresponding to the melting point of the metal precursor [24]), Ar + 4%H₂ gas at a flow rate of 0.7L per minute, holding at the target temperature for 30 minutes (Parameters No. 1, Table 6). Due to a lack of deposition, further experiments were carried out at higher temperatures, which were 300°C and 350°C. For 300°C no deposition was observed, while at 350°C chromium was deposited only on the surface of the CNT carpet, where CNTs were in direct contact with chromocene powder (Figure 15). It can be presumed that higher temperature (closer to the chromocene processing temperature in standard MOCVD, that is around 600°C) could potentially improve deposition ability; however, due to limitations of the functionalization setup, testing had to be restricted to 350°C. Additional investigation, based on reduction of flow rate to 0.2L per minute, turned out to be unsuccessful as well.

No.	Sample name	Temp, °C	Gas	Flow rate, L min'1	Time, min
1	C1	172
2	C2	300	Ar+4%H2	0.7	30
3	C3,C5	350
4	C4	350		0.2

Table 6 - Parameters of GF-MOCVD process with use of chromocene as a 25 function of temperature and gas flow rate of the procedure.

Further GF-MOCVD investigation was done using dibenzenechromium Cr(C6H₆)2, which in the classic MOCVD process requires much lower processing temperature, that is 300°C and more. According to the literature MOCVD coatings deposited at 300°C are characterised by an amorphous structure. To get a crystalline structure of the deposited metal, a temperature of at least 450°C is required. Nevertheless the processing temperature for dibenzenechromium is much lower than for chromocene, which makes it much more promising for GFMOCVD.

Testing with use of Cr(C6H₆)2 was done for two temperatures, 300°C and 350°C, a gas flow rate of 0.2 and 0.7L per minute, and three different times, as listed in Table 7.

It was found that, for the same flow rate (0.7L per minute) and time (20 minutes at target temperature), a temperature of 350°C provides more efficient CNT functionalization than at 300°C (Figure 16 and Figure 17). Similar to CNT functionalization using nickelocene, higher gas flow rate as well as longer deposition time resulted in higher yield of deposition (Figure 18).

Parameters No.	Sample name	Temp, °C	Gas	Flow rate, L min'1	Time, min
1	CR2	300		0.7	20
2	CR3	350		0.7	20
3	CR5	300	Ar-4%H2	0.7	40
4	CR4	300	0.2	20
5	CR6	350		0.7	60
6	CR7	350		0.2	60

Table 7 - Parameters of GF-MOCVD process as a function of temperature and gas flow rate of the procedure.

GF-MOCVD - Summary and conclusions

This part of the present work concerns the newly-developed method of functionalization of CNTs referred to as Gravity Fed Metalorganic Chemical Vapour Deposition (GF-MOCVD), which is superior to sputtering due to its high scalability and possibility to cover three-dimensional objects such as CNT carpets. Nickel-functionalized CNT carpets, obtained as a result of optimisation of GFMOCVD parameters, were used for preparation of conductive coatings on quartz slides. The sample preparation involved a procedure optimised during initial studies on as-made nanotubes. It was found that, in the case of nonfunctionalised CNTs, heat-treatment resulted in a 10-fold resistance decrease, whereas in the case of Ni-functionalised nanotubes the same heat-treatment resulted in an almost 300-fold resistance decrease.

To summarise, all the studies reported here indicate the effectiveness of functionalisation of CNTs with nickel as a method of providing a reduction in contact resistance (e.g. with respect to a copper matrix, in a metal matrix composite).

With regard to the type of CNTs used, it should be noted that the GF-MOCVD process is particularly useful for multi-walled CNTs (MWCNTs) because the reactive metals used in GF-MOCVD may, in some cases, lead to damage of single- and double-walled CNTs.

Metal precursors other than Ni and Cr may be used. Indeed, GF-MOCVD is fully expected to achieve functionalization of CNTs with other transition metals such as Ti, V, Mn, Fe, Co, Zr, Nb, Mo. It can be presumed that sandwich compounds (arene complexes - a group of organometallic compounds commonly used as precursors in the classic MOCVD method) will also work well in GF-MOCVD.

Just to give examples we can name several arene complexes of selected transition metals to be used in GF-MOCVD, as follows:

• V precursor can be vanadocene, also called bis(cyclopentadienyl) vanadium. Another precursor of this metal can be bis(benzene)vanadium.

• Cr precursor can be chromocene, also called bis(cyclopentadienyl)chromium(ll). Another precursor of this metal can be bis(benzene)chromium also called dibenzenechromium.

• Mn precursor can be manganocene, also called bis(cyclopentadienyl)manganese(ll).

• Fe precursor can be ferrocene, also called bis(-cyclopentadienyl)iron. Another precursor can be bis(benzene)iron.

• Co precursor can be cobaltocene, also known as bis(cyclopentadienyl)cobalt(ll).

• Ni precursor can be nickelocene, also called bis(cyclopentadienyl)nickel. Another source of nickel can be dicationic triscyclopentadienyl dinickel.

• Ti precursor can be titanocene-dichloride, also called bis(cyclopentadienyl)titanium dichloride. Another source of titanium can be cyclopentadienyl titanium trichloride.

Route (ii) - Plating of copper particles with a transition metal

A second route for improving the wettability of CNTs by matrix metal involves plating of metal particles, in particular dendritic copper particles, with a transition metal.

Further background

Problems with the formation of CNT/copper composites result from difficulties in obtaining a strong interphase boundary, which is required for effective load transfer and lossless electric power transmission. As discussed above, a significant obstacle to creating a good interface is lack of wetting of CNTs by molten metals. The phenomenon of macroscopic wetting of graphite by liquid metals is well-known and broadly discussed in the literature. Non-reactive metals such as copper do not wet graphitic materials, unless relatively large amounts of carbon are dissolved (a good example is Ni). In classic applications electroplating or electroless plating of graphite provides good adhesion of Ni particles to carbon, while the mutual solubility of Cu and Ni enhance the effective load transfer ability. Carbide forming metals and their alloys with non-reactive metals wet graphite due to the formation of reaction products (carbide layer with thickness of few to tens of microns) more wettable than the substrate. Uniform dispersion of CNTs in the metal matrix is the most significant challenge in the field of CNT-reinforced composites. The presence of CNT clusters and non-uniform dispersion of CNTs in the metal matrix leads to incomplete utilization of the reinforcement. The quality of dispersion is a crucial factor which determines the homogeneity and final mechanical, thermal and electrical properties of the composite.

One method of synthesis of copper matrix composites reinforced with CNTs is powder metallurgy. Powder metallurgy uses numerous techniques that can be utilized for the dispersion of CNTs in the starting powder mixture; however, without optimisation of carbon nanotube structures (or alternatively the Cu matrix) there is no integration between dissimilar materials, leading to weakening of final composite (as illustrated schematically in Figure 19 and shown in Figure 20).

The ability of metal atoms to form bonds with carbon nanotubes increases with the number of unfilled d orbitals of the metal atoms. Transition metals such as Au and Pd do not have unfilled d orbitals, and consequently their activity is low. Metals having an unfilled d orbital such as Ni, Fe, Co are characterized by higher activity and a certain solubility of carbon. Metals with unfilled 3d or 4d orbitals such as Ti or Cr are able to form strong bonds with carbon, which may lead to formation of stable carbides. The mechanism of wetting is determined not only by the type of carbon-metal system but also by thermodynamic factors. Wetting angle, and thereby work of adhesion, depend largely on the temperature which is reflected in the composite formation process.

Summarizing, the quality of the interphase boundary in a metal-CNT system (in particular, Cu-CNT) can be controlled by surface modification of the carbon nanotubes via GF-MOCVD (as per Route (i) above, and as illustrated schematically in Figure 21) or by modification of the metal matrix via powder plating (the present Route (ii), as illustrated schematically in Figure 22), as follows.

Plating considerations

Deposition of metal coatings on metal powders can be done in a similar way to how it is normally done in the case of bulk metallic components. Classic plating methods are based on immersion or electroless plating techniques utilizing redox reactions during which one element is oxidised and another is reduced, and therefore no external energy (e.g. electric current) is required. The main difference between immersion and electroless plating is the chemical role of the base metal. In immersion plating the base metal is oxidised, i.e. metal atoms are transferred to the solution as metallic ions, while in electroless plating the base metal is a catalyst of the reaction and an additional reductor is added to the solution.

In order to produce metal coated (e.g. Ni-coated) copper powders, the present inventors have investigated the applicability of both approaches (immersion plating and electroless plating) for coating dendiritic Cu powders. It was found that deposition of nickel on copper powders via immersion plating using a thiosulphate bath results in very low purity of coatings (due to the formation of nickel sulphate) whereas electroless plating using a hypophosphite bath produces relatively pure Ni coatings with little phosphorus contamination.

Deposition of nickel on copper powders via immersion plating using a thiosulphate bath

Based on the literature analysis a first bath for immersion plating of Ni on Cu was prepared by mixing the following compounds:

• 60g nickel sulphate (NiSO₄-6-7H₂O) • 60g ammonium nickel sulphate (Ni(NH₄)2SO4-6H₂O) • 120g sodium thiosulphate (NaS₂Os) • 1L water (H₂O)

Nickel sulphate and ammonium nickel sulphate each provide a source of nickel ions, which are later reduced to pure metallic nickel. Additionally, ammonium nickel sulphate is a source of ammonia, which is a complexing agent (forming complex ions with both Ni²⁺ and Cu²⁺) Sodium thiosulphate is a source of another complexing agent (the thiosulphate anion, which also form complexes with both ions). Addition of complexing agents is necessary. According to the electrochemical series of metals, nickel is more reactive than copper, so in normal conditions without addition of complexing agents we are able to deposit copper onto nickel, but not nickel onto copper. Complexing agents change the electrochemical potentials and therefore allow us to deposit nickel onto copper. The reaction is carried out at a temperature in the range of approximately 38°C to 65°C.

Prior to evaluation of copper powder coatability, copper strips were used to investigate the influence of different plating conditions. The procedure was devised as follows:

1. Cleaning of base material.

2. Immersing in the bath at given conditions.

3. Plating.

4. Rinsing with pure RO (reverse osmosis) water and drying.

Different approaches of Ni-plating were investigated:

• CuNiA - plating in the solution for 1 h at room temperature • CuNiB - plating in the solution for 1 h at 45°C • CuNiC - plating in the solution for 1 and 5 mins at room temperature • CuNiD - plating in base Ni plating solution (based on RO water) • CuNiE - plating in the solution where Na₂S₂O3 concentration was reduced by 50% • CuNiF - plating in the solution where Na₂S₂O3 concentration was doubled (200%) • CuNiG - plating in the base Ni plating solution (based on RO water) - Cu strips not cleaned • CuNiH - plating in diluted Ni-plating solution (75% solution 25% water) • CuNil - plating in diluted Ni-plating solution (50% solution 50% water) • CuNiJ - plating in diluted Ni-plating solution (25% solution 75% water) • CuNiK - plating in base Ni-plating solution plus application of surfactant Pluronic P123

Initial investigation showed that samples plated at room temperature had a green coating, which did not stick to the surface. Also the copper surface was brick red, which suggests severe oxidation. The oxidation was even more clear for samples plated at 45°C. Almost all the green coating peeled off during the washing procedure. The strip surface was deep brick with black spots. Coatings obtained after 1 minute and 5 minutes were not peeling off. Similar results were obtained for plating in the solution for 1, 5, 30 minutes and 3 hours at room temperature where pure RO water was used; however, SEM/EDX investigation did not prove deposition of Ni but nickel sulphide.

Due to deposition of nickel sulphide instead of pure metallic nickel several modifications of the bath composition were investigated. Plating in the solution where sodium thiosulphate (Na₂S₂O3) concentration was reduced by 50% did not show any difference in comparison to standard plating solution. However, increase of thiosulphate concentration caused the coating to become green even after only 1 minute of immersion. We checked also if surface cleaning has any influence on the process. Therefore three samples (ref. CuNiG) were immersed in the bath without any cleaning. It was found that lack of cleaning the strips did not have a significant influence on the coatings. In all cases nickel sulphide was formed on the copper surface.

In the next series of experiments diluted Ni-plating solution was investigated, using 75% (ref. CuNiH), 50% (ref. CuNil) and 25% (ref. CuNiJ) of the initial plating solution concentration. A lower concentration of plating solution caused slower coating formation. The coatings after 1 minute of plating became less noticeable as the concentration decreased. Additionally, coatings after 5 minutes, and in the case of the least concentrated solution even after 10 minutes, are still silvery metallic, not green. However, longer immersion times resulted in green coatings. In all cases nickel sulphide was formed on the copper surface.

The influence of surfactant addition (Pluronic P123) on the coatings was also investigated. To the initial Ni-plating solution 0.1% of P123 was added. Surfactant was used to improve copper wetting; however, the only noticed effect was reaction slowdown.

The last series of experiments covered adjustment of plating solution pH as a possible route to reduce formation of sulphides. Thiosulphate decomposes to sulphide and sulphate at lower pH. The original Ni-plating solution was pH 4.5 The following investigations were performed:

• CuNiL - pH=7, room temperature CuNiM - pH=7, 45°C • CuNiN - pH=7.3, room temperature CuNiO - pH=7.3, 45°C

Change of pH did not have significant impact on the process. Like in the previous tests, heating caused quicker formation of green coating. For the experiments carried out at room temperature, metallic black coating was obtained after 1 minute. Longer coating times caused formation of green coating.

Following initial investigations on plating copper strips, the procedure for selected conditions was repeated for dendritic copper powder. The procedure was devised as follows:

1. Preparation of Ni plating solution.

2. Mixing of dendritic copper powder with Ni-plating solution using magnetic stirrer.

3. Plating for pre-defined time.

4. Decantation.

5. Rinsing with pure RO water followed by decantation.

6. Rinsing with acetone followed by decantation and drying at 45°C.

Due to poor wettability of copper the powder plating time was increased to 6h. An exception was plating in a solution where an additive surfactant (Pluronic 123) was used.

The following samples were made:

CuNil - 10g Cu + 60ml Ni-plating solution 45°C 6h CuNi2 - 2.5g Cu + 60ml Ni-plating solution 45°C 6h

- CuNi3 - 2.5g Cu + 60ml Ni-plating solution RT 6h

CuNi4 - 2.5g Cu + 40ml Ni-plating solution RT 6h

CuNi5 - 2.5g Cu + 30ml Ni-plating solution + 10ml 1% P123 solution in water RT 20 mins

CuNi6 - 2.5g Cu + 30ml Ni-plating solution + 10ml 1% P123 solution in 10 water 45°C 20 mins

SEM/EDX analysis showed that in all cases the deposited coating consists of nickel sulphate. Additionally, the deposit is not uniform and forms grains on the surface of the copper powder (Figure 23 and Table 8).

IT	AS	/SeYieg	tlSj·}.. ; Ο/	atslm. -C .:11	Atom-. C Er	igi (:1.....Sigma)
c	6:	S-series	3.68	3.56	12.3;	0.98
0	S;	E-series	:10:,11	10.OS	26.16	::i,;@l
s	is:	A—series,	1,10:	6.96	9.02	0.31
W:	28	Elseries.:
Cu	2:0:	M-'sefeies-·:	72:. 31	S S. 0 S		: 2.35

Table 8 - Results of EDX analysis, typical for all samples of dendritic copper powder Ni plated using thiosulphate bath.

Deposition of nickel on copper powders via electroless plating using a hypophosphite bath

Following the unsuccessful plating using the thiosulphate bath, a second approach was devised, based on electroless platting using a hypophosphite bath. Based on the literature analysis first a bath for electroless plating of Ni on Cu was prepared by mixing the following compounds:

• 2g nickel sulphate (N1SO4 6H₂O) • 2.6g sodium hypophosphite (NaH₂PO2) • 0.03g sodium citrate (NaC₆H₇O7) • 0.1 L water (H₂O) • Reaction conditions: temperature ca. 80°C; pH=5.1

Nickel sulphate is a source of nickel; sodium hypophosphite is a reducing agent and sodium citrate is a complexing agent. In this reaction base metal (copper) acts only as a catalyst to the reaction. Therefore, hypophosphite is used as a reductor of nickel (II) ions. But the reaction of Ni²⁺ reduction by H2PO2' ions may occur not only at the copper surface but also in the solution. In order to slow down this reaction citrate is used as a stabiliser. When an appropriate concentration of the stabiliser is used, the reaction occurs only at the metal surface and not in the bath. Too high concentration of stabiliser may even stop the reaction. The following two reactions are the most important in the method:

NiSO₄ + 3 NaH₂PO₂ + 3 H₂O Ni + 2 H₂ + H₂SO₄ + 3 NaH₂PO₃ 3 NaH₂PO₂ NaH₂PO₃ + 2P + 2 NaOH + H₂O

The above reactions take place parallel in the bath, but the first one is much quicker. The more gas that is evolved during the process, the quicker the first reaction is, and less phosphorus is deposited on the copper surface.

Initial investigation was made for a copper strip, by immersing it in the hypophosphite plating solution for 15 minutes at 80°C. The experiment resulted in the formation of a silver metallic coating. SEM/EDX analysis indicated formation of nickel phosphorus deposits with relatively low phosphorus concentration (Table 9).

	AN Series	iifiii:; Ό G 'ABsmA G>TEritiSji iU	S i gma)
V,	6. K--series	6.24		23.5 4
d;	8 K-series	0.83	0.83		0.23
P	1? K-.se ties	6.99	:S .9 3-		W-Ao
	28 K-ssries		45.82	35.35
OB	29 K-seri-es	ASyfS		28.6 4

Table 9 - EDX analysis of copper strip immersed in hypophosphite plating solution for 15 minutes at 80°C.

For preparation of Cu-Ni powders the following procedure was applied:

1. Adding of Cu powder to Ni-plating solution and mixing using magnetic stirrer (2g of Cu per 100 ml or Ni-plating solution).

2. Heating mixture up to 80°C.

3. Plating - reaction take place as long as gas is evolving.

4. Decantation.

5. Rinsing with pure RO water followed by decantation.

6. Rinsing with acetone followed by decantation and drying at 45°C.

SEM/EDX analysis of Ni-coated Cu powder removed from the electroplating bath after completion of H₂ evolution (~ 18 minutes) revealed deposition of approximately 13.8wt% nickel and ~ 2.4wt% phosphorus (Table 10).

Ei	, AH	SeEigS-	am®, C Atom. C ErAfitL	K1+ sigma)
			iw+oSl	! W L . % ]
<3	A	TS-sAties	2.03	1.90	8.54	0,66
O	3	E ··siries	2. OS	1.96	6.61	0.48
E	AS		2.39	2.23	3.90	0.15
si	sa	K- series	14.79	13.85	12.77	0.55
Cu	29		85.51	80.06	68.17	2.84

Table 10- EDX analysis of dendritic copper powder Ni plated in hypophosphite plating bath.

In order to determine the controllability of the Ni deposition, six batches were left 20 to react for different times: 1, 3, 5, 10, 15 and 20 minutes. In each case reaction was stopped by cooling the plating bath, followed by decantation. SEM/EDX analysis indicated deposition of 3-5wt% Ni after 1 min of reaction, 5-6wt% after 3 min, 10wt% after 5 min, and approximately 13-14% for reaction time over 10 minutes. Broadly similar results were obtained via ICP-MS (inductively coupled plasma mass spectrometry), which is generally more reliable than SEM/EDX analysis (Table 11).

Reaction time	Ni concentration based on SEM/EDX analysis	Ni concentration based on ICP-MS elemental analysis
1 min	3-5%	2.145%
3 min	5-6%	3.143%
5 min	10%	4.845%
10 min	«13%	5.661%
15 min	«14%	6.845%
20 min	«14%	6.846%

Table 11 - Dependence of reaction time on Ni deposition for standard plating conditions (2g nickel sulphate, 2.6g sodium hypophosphite, 0.03g sodium citrate, 100mL water, temperature ca. 80°C; pH=5.1).

Further investigation was focused on evaluation of the influence of the Cu powder concentration in the plating bath on Ni deposition (Table 12). The results indicate that, in the case of high Cu powder loading, there can be insufficient Ni in the plating solution, resulting in low wt% values of deposited Ni. This can readily be compensated by increasing the amount of Ni in the plating solution.

Cu powder loading	Reaction time	Ni concentration based on ICP-MS elemental analysis
4g/100ml_		1.378%
8g/100ml_	1 min	1.123%
16g/100ml_		0.0286%
4g/100ml_		2.581%
8g/100ml_	3 min	1.360%
16g/100ml_		0.0482%
Cu (reference)	-	<0.0001%

Table 12- Dependence of Cu powder loading on Ni deposition for 2 different times and standard plating conditions (2g nickel sulphate, 2.6g sodium hypophosphite, 0.03g sodium citrate, 100mL water, temperature ca. 80°C;

pH=5.1).

Figure 24 shows typical SEM images of dendritic copper powder Ni plated using a hypophosphite bath.

Production of metal matrix composites

With the functionalized carbon nanotubes prepared via Route (i) above, a metal matrix composite may be produced by disposing the functionalized carbon nanotubes within molten metal (e.g. molten copper), which may then be solidified to form the composite.

Alternatively, a powder metallurgy route may be used to form a metal matrix composite, whereby the functionalized carbon nanotubes are mixed with metal powder (e.g. copper powder) to produce a mixture, and then the mixture is compacted and sintered to produce the composite.

Similarly, with the plated copper particles prepared via Route (ii) above, a metal matrix composite may be produced by combining the plated copper particles with carbon nanotubes. For example, a powder metallurgy route may be used, whereby carbon nanotubes are mixed with the plated copper particles to produce a mixture, and then the mixture is compacted and sintered to produce the composite.

Possible modifications and alternatives

Detailed embodiments and some possible alternatives have been described above. As those skilled in the art will appreciate, a number of modifications and further alternatives can be made to the above embodiments whilst still benefiting from the inventions embodied therein. It will therefore be understood that the invention is not limited to the described embodiments and encompasses modifications apparent to those skilled in the art lying within the scope of the appended claims.

It will also be appreciated that, if desired, the above techniques (i.e. Routes (i) and (ii)) may be used in combination. That is to say, the CNTs may be functionalized with transition metals using the GF-MOCVD process, the copper particles may be plated with transition metals, and the functionalized CNTs and the plated copper particles may be combined to form a metal matrix composite.

REFERENCES [1] K. A. Shah et al, Synthesis of carbon nanotubes by catalytic chemical vapour deposition: A review on carbon sources, catalysts and substrates, Materials Science in Semiconductor Processing 41, 67-82, 2016 [2] M. Burghard et al, Chemically Functionalized Carbon Nanotubes, Small, 1, No. 2, 180-192, 2005 [3] T. Chowdhury et al, Carbon Nanotube Composites for Electronic Interconnect Applications, Syntheses and Applications of Carbon Nanotubes and Their Composites, InTech, 16, 367-393, 2013 [4] E.V. Barrera et al, Carbon Nanotubes: Science and Applications, Applications: Composites, 11.4, 266, CRC Press, 2004 [5] Chu et al, Improvement of interface and mechanical properties in carbon nanotube reinforced Cu-Cr matrix composites, Materials and Design, 45, 407-411, 2013 [6] E. Dujardin et al, Capillarity and Wetting of Carbon Nanotubes, Science, Vol. 265, 1850-1851, 1994 [7] E. Dujardin et al, Wetting of Single Shell Carbon Nanotubes, Advanced Materials, 10, 17, 1998 [8] S.C. Lim et al, Contact resistance between metal and carbon nanotube interconnects: Effect of work function and wettability, Applied Physics Letters, 95, 264103, 2009 [9] H. Ago et al, Work Functions and Surface Functional Groups of Multiwall Carbon Nanotubes, The Journal of Physical Chemistry B, 103, 8116-8121, 1999 [10] Y. Chai et al, Low-Resistance Electrical Contact to Carbon Nanotubes With Graphitic Interfacial Layer, IEEE Transactions On Electron Devices, Vol. 59, 1, 2012 [11] S.R. Bakshi et al, Carbon nanotube reinforced metal matrix composites - a review, International Materials Reviews Vol. 55, 1, 2010 [12] H. Kwon et al, Investigation of carbon nanotube reinforced aluminum matrix composite materials, Composites Science and Technology 70, 546-550, 2010 [13] Y. Hu et al, Carbon Nanotubes and Carbon Nanotubes/Metal Oxide Heterostructures: Synthesis, Characterization and Electrochemical Property, Carbon Nanotubes - Growth and Applications, 1, Intech, 2011 [14] Y. Sun, Mechanical Properties Of Carbon Nanotube/Metal Composites, PhD dissertation, Department of Mechanical, Materials, and Aerospace Engineering, University of Central Florida, 2010 [15] P.D. Dapkus, Metalorganic Chemical Vapor Deposition, Annual Review of Materials Science, Vol. 12, 243-69, 1982 [16] J-K. Kang et al, Metalorganic Chemical Vapor Deposition of nickel films from Νί(θ5Η₅)/2Η₂, Journal of Materials Research, Vol. 15, 8, 2000 [17] A.Douard et al, Chromium-based coatings by atmospheric chemical vapor deposition at low temperature from Cr(CO)6, Surface & Coatings Technology, 200 1407-1412, 2005 [18] R. Gaur et al, Ruthenium complexes as precursors for chemical vapordeposition (CVD), RSC Advances, 4, 33785-33805, 2014 [19] L. Brissonneau et al, MOCVD-Processed Ni Films from Nickelocene.

Part I: Growth Rate and Morphology, Chemical Vapor Deposition, 5, 4, 1999 [20] L. Brissonneau et al, MOCVD-Processed Ni Films from Nickelocene.

Part II: Carbon Content of the Deposits, Chemical Vapor Deposition, 5, 4, 1999 [21] L. Brissonneau et al, Investigation of nickelocene decomposition during Chemical Vapor Deposition of Nickel, Journal of The Electrochemical Society, 147 (4) 1443-1448, 2000 [22] Bis(cyclopentadienyl)nickel(ll), Safety Data Sheet, Sigma Aldrich, Version 5.1, Revision Date 09.10.2012 [23] F. Schuster et al, Influence Of Organochromium Precursor Chemistry On The Microstructure Of MOCVD Chromium Carbide Coatings, Surface and Coatings Technology, 43/44, 185-198, 1990 [24] Bis(cyclopentadienyl)chromium(ll), Safety Data Sheet, Sigma Aldrich, Version 5.0, Revision Date 19.09.2012

Claims (53)

1. A metalorganic chemical vapour deposition method of functionalizing carbon nanotubes with a transition metal, the method comprising:

arranging, on a substrate within a reaction chamber, a quantity of carbon nanotubes in contact with an organometallic compound of the transition metal;

purging oxygen from the reaction chamber and filling the reaction chamber with a process gas; and heating the reaction chamber to a process temperature equal to or greater than the melting point of the organometallic compound;

whereby a reaction takes place in which the organometallic compound releases the transition metal and the transition metal functionalizes the carbon nanotubes.

2. A method as claimed in claim 1, wherein the transition metal is nickel.

3. A method as claimed in claim 2, wherein the organometallic compound is nickelocene.

4. A method as claimed in claim 1, wherein the transition metal is chromium.

5. A method as claimed in claim 4, wherein the organometallic compound is dibenzenechromium or chromocene.

6. A method as claimed in claim 1, wherein the transition metal is titanium, vanadium, manganese, iron, cobalt, zirconium, niobium or molybdenum.

7. A method as claimed in any preceding claim, wherein the carbon nanotubes are arranged in the form of an aligned array.

8. A method as claimed in any preceding claim, wherein the organometallic compound is arranged on top of the carbon nanotubes.

9. A method as claimed in any preceding claim, wherein the process gas comprises a mixture of argon and hydrogen, or a mixture of helium and hydrogen.

10. A method as claimed in any preceding claim, further comprising maintaining a flow of the process gas through the reaction chamber during the reaction.

11. A method as claimed in claim 10, wherein the substrate is porous, and the process gas flows through the substrate during the reaction.

12. A method as claimed in claim 11, wherein the process gas flows upwards through the substrate during the reaction.

13. A method as claimed in claim 11, wherein the process gas flows downwards through the substrate during the reaction.

14. A method as claimed in any of claims 11 to 13, wherein the substrate is made of a porous ceramic material.

15. A method as claimed in any of claims 11 to 13, wherein the substrate is made of a porous glass material.

16. A method as claimed in claim 15, wherein the substrate is made of porous borosilicate glass.

17. A method as claimed in any preceding claim, wherein the process gas is used for said purging.

18. A method as claimed in any preceding claim, wherein the process temperature is one of:

Tmeit + 0 C, Tme|t + 10 C, Tmelt + 20 C, Tmelt + 40 C, Tmelt + 80 C, where Tmeit is the melting point of the organometallic compound.

19. A method as claimed in any preceding claim, wherein the reaction takes place under atmospheric pressure.

20. A method as claimed in any of claims 1 to 18, wherein the reaction takes place under sub-atmospheric pressure or a vacuum.

21. A method as claimed in any of claims 1 to 18, wherein the reaction takes place under slight super-atmospheric pressure.

22. A method as claimed in any preceding claim, further comprising: disposing the functionalized carbon nanotubes within molten metal; and solidifying the molten metal to form a metal matrix composite.

23. A method as claimed in claim 22, wherein the molten metal comprises copper, such that the resulting metal matrix composite comprises carbon nanotubes in a copper matrix.

24. A method as claimed in any of claims 1 to 21, further comprising:

mixing the functionalized carbon nanotubes with metal powder to produce a mixture; and compacting and sintering the mixture to form a metal matrix composite.

25. A method as claimed in claim 24, wherein the metal powder comprises copper, such that the resulting metal matrix composite comprises carbon nanotubes in a copper matrix.

26. Functionalized carbon nanotubes produced by a method according to any of claims 1 to 21.

27. A metal matrix composite produced by a method according to any of claims 22 to 25, or using functionalized carbon nanotubes according to claim 26.

28. Metalorganic chemical vapour deposition apparatus for functionalizing carbon nanotubes with a transition metal, the apparatus comprising:

a reaction chamber;

a substrate inside the reaction chamber, for holding a quantity of carbon nanotubes in contact with a quantity of an organometallic compound;

one or more heaters for heating the reaction chamber to a process temperature equal to or greater than the melting point of the organometallic compound; and a supply of a process gas;

wherein the reaction chamber further comprises a gas inlet for receiving the process gas from the gas supply, and an exhaust through which gas can leave the reaction chamber.

29. Apparatus as claimed in claim 28, wherein the substrate is porous.

30. Apparatus as claimed in claim 29, wherein the substrate is made of a porous ceramic material.

31. Apparatus as claimed in claim 29, wherein the substrate is made of a porous glass material.

32. Apparatus as claimed in claim 31, wherein the substrate is made of porous borosilicate glass.

33. Apparatus as claimed in any of claims 29 to 32, wherein the gas inlet is beneath the substrate and the exhaust is above the substrate.

34. Apparatus as claimed in any of claims 29 to 32, wherein the gas inlet is above the substrate and the exhaust is below the substrate.

35. Apparatus as claimed in any of claims 28 to 34, wherein the process gas comprises a mixture of argon and hydrogen, or a mixture of helium and hydrogen.

36. A method of plating copper particles with a transition metal, the method comprising mixing the copper particles with an aqueous solution comprising:

a source of a transition metal, a reducing agent to produce ions of the transition metal, and a complexing agent.

37. A method as claimed in claim 36, wherein the transition metal is nickel.

38. A method as claimed in claim 37, wherein the source of the transition metal is nickel sulphate.

39. A method as claimed in any of claims 36 to 38, wherein the reducing agent is sodium hypophosphite.

40. A method as claimed in any of claims 36 to 39, wherein the complexing agent is sodium citrate.

41. A method as claimed in any of claims 36 to 40, further comprising heating the solution to a temperature between 30°C and 90°C.

42. A method as claimed in claim 41, wherein the solution is heated to a temperature of around 80°C.

43. A method as claimed in any of claims 36 to 42, further comprising decanting and rinsing the plated copper particles.

44. A method as claimed in any of claims 36 to 43, wherein the copper particles are dendritic copper particles.

45. A method as claimed in any of claims 36 to 44, further comprising combining carbon nanotubes with the plated copper particles to form a metal matrix composite.

46. Plated copper particles produced by a method according to any of claims 36 to 44.

47. A metal matrix composite produced by a method according to claim 45 or 5 using plated copper particles according to claim 46.

48. A metalorganic chemical vapour deposition method of functionalizing carbon nanotubes with a transition metal substantially as herein described with reference to and as illustrated in any combination of the accompanying drawings.

49. Functionalized carbon nanotubes substantially as herein described with reference to and as illustrated in any combination of the accompanying drawings.

50. A metal matrix composite substantially as herein described with reference 15 to and as illustrated in any combination of the accompanying drawings.

51. Metalorganic chemical vapour deposition apparatus substantially as herein described with reference to and as illustrated in any combination of the accompanying drawings.

52. A method of plating copper particles with a transition metal substantially as herein described with reference to and as illustrated in any combination of the accompanying drawings.

25

53. Plated copper particles substantially as herein described with reference to and as illustrated in any combination of the accompanying drawings.

Intellectual

Property

Office

Application No: GB1619280.9