WO2017191415A1 - Composite material based on vertically aligned carbon nanotubes and on a metal matrix - Google Patents

Abstract

Process for manufacturing a composite material comprising vertically aligned carbon nanotubes (VACNT) and a copper metal matrix coating said nanotubes, said process comprising at least the following steps: a) a provision of vertically aligned nanotubes (VACNT) deposited on a substrate, b) the electrochemical deposition of said metal matrix on said vertically aligned nanotubes from a solution that is in contact with said VACNTs, said solution comprising at least one precursor of said metal matrix and at least one organic solvent, said process being characterized in that: - said at least one precursor is selected from the group formed by: copper acetate (Cu(CH₃COO)₂), copper formate, copper propionate; and in that - said electrochemical deposition is a galvanostatic electrodeposition.

Description

 Composite material based on vertically aligned carbon nanotubes and a metal matrix

Technical field of the invention

The invention relates to the field of composite metal matrix materials comprising carbon nanotubes. More particularly, it relates to a composite material comprising vertically aligned carbon nanotubes (VACNT) and a metal matrix, its manufacturing method, as well as its use as a thermal interface material.

State of the art

Carbon nanotubes (abbreviated often as "CNT", Carbon NanoTubes) are tubes of nanometric diameter whose walls consist of graphitic sheets (graphene sheets). Whether they have a single-walled wall or multi-walled wall, they have particular mechanical, thermal, electronic and structural properties; these properties reflect their strong structural anisotropy as indicated in the publication by Lukes et al. ("Thermal conductivity of individual single-wall carbon nanotubes", Journal of Heat Transfer, 129 (2007) 705-716). Many applications have been imagined taking advantage of these particular properties.

Thus, polymeric materials loaded with nanotubes were prepared, which were used for the manufacture of tennis rackets, taking advantage of mechanical properties combining strength and flexibility. It has also been envisaged to take advantage of their high electronic conductivity in the direction of the length of the tubes (see Huard et al., "Vertically aligned carbon nanotube-based composite: Elaboration and monitoring of the nanotubes alignment", Journal of Applied Polymer Science, vol.131 (2014) pp. 1-6). Their thermal conductivity is also very anisotropic, high in the direction of the length of the tube. Kim et al. ("Thermal Transport Measurements of Individual Multiwalled Nanotubes", Phys Rev. Lett, 87 (2001) 215502), and Hu et al. ("3-omega measurements of vertically oriented carbon nanotubes on silicon", Journal of Heat Transfer, vol 128 (2006) p.1 109-1 1 13) have announced intrinsic conductivities of the order of 3000 W / mK for an isolated CNT.

For ten years it has been known to deposit CNT vertically aligned on a substrate; this product is known as VACNT (Vertically Aligned Carbon NanoTubes). Prasher ("Thermal interface materials: historical perspective, status, and future directions ", Proc. IEEE vol. 94 (2006) p.1571-1586) and Cola et al. ("Carbon nanotubes as high performance thermal interface materials", Electron Cool, vol.16 (2010) p10-15) describe the possibility of using a VACNT mat as thermal interface materials (TIM). Thermal interface materials are used to evacuate the heat produced by electronic components with which they are in thermal contact. The authors observe on a VACNT mat deposited on a silicon crystal that the thermal conductivity in the thickness direction (ie parallel to the length of the aligned tubes) is much higher than that of commercially available thermal interface materials. The usual thermal interface materials are phase change materials (PCM) or thermal greases whose thermal conductivity does not exceed 5 W / mK (see Otiaba et al., Microelectronics Reliability 51 (201 1) p. 2031-2043). Thermal interface materials based on carbon nanofibers and copper have been described by Ngo et al. (Nano Letters 4 (2004) p.2403-2407). The copper was electrodeposited on a substrate previously coated with carbon nanofibers, to a maximum thickness of the order of 30 μηι.

Thermal interface materials are widely used to dissipate heat from electronic components (such as transistors, integrated circuits, LEDs) to a dissipator. The miniaturization of electronic devices, the compactness of electronic circuits and the increase in the number of components per unit of surface cause an increase in the energy density to be dissipated in the form of heat (see Lasance, "Advances In High-Performance Cooling"). For Electronics, "Electron, Cool, Mag, November 2005, see McNamara et al., International Journal of Thermal Sciences 62 (2012) p2-1 1). Although there are many thermal interface materials, electronic device manufacturers need more efficient thermal interface materials.

For example, B. Vergne ("Shaping carbon / Alumina NanoTubes composites and modeling their thermal conductivity", Thesis N ° 142007, University of Limoges (2007)) underlines the importance of having materials thermal interface having both good thermal conductivity and good thermomechanical cohesion with the materials in contact, in order to optimize the reliability of the chip modules. In the field of power electronics, the heat generated by the electronic chip in operation is of the order of 0.4 W mm ^"2 (Dunn et al., J. Appl. Phys. 73 (4) (1993 To dissipate this heat, a heat sink is added under the ceramic substrate (usually composed of alumina or aluminum nitride) .These drains are made of copper because the thermal conductivity of this metal is high (400 Wm ^{~ 1} .K ^{~ 1} ). The reliability of these Silicon chips / ceramic substrate / copper drain module modules essentially rely on good thermomechanical cohesion between the substrate and the heat sink. In these silicon chips heating structures can reach 150 ° C, and the thermal expansion coefficients of alumina (8.10 ^-6 K ^{~ 1} ) or aluminum nitride (4.10 ^-6 K ^{~ 1} ) being very far from that of copper (17.10 ^{~ 6} K ^{~ 1} ), a thermal interface material (solder joint) must be added between the substrate and the drain. Despite this seal, thermomechanical fatigue related to this difference in expansion coefficients is a major problem for the reliability of these modules. More reliable industrial solutions are known for replacing the copper drain, for example the Al / SiC composites or the Cu / Invar1 / Cu multilayers, but these alternatives have a high manufacturing cost, entail considerable difficulties in the production of copper. machining, and their thermal conductivity does not even reach 200 Wm ^{~ 1} .K ^{~ 1} , and is much lower than that of copper. Improving thermal interface materials on electronic boards can lead to increased component life and make their operation more stable. They can also be operated at higher power: thus the use of more efficient TIM ultimately improves the performance of electronic devices. Depending on the application of the electronic components, the main factors influencing the choice of a TIM are related to the quality of the connection with the electronic component in thermal contact (pressure exerted, applied limit stress, mechanical stability), and intrinsic thermal properties TIM such as the thermal conductivity K _th , the thermal resistance R _th or the coefficient of expansion or thermal expansion CTE (see Otiaba et al., "Emerging Nanotechnology-based Thermal Interface Materials for Automotive Electronic Control Unit Application", 18th IEEE European Microelectronics and Packaging Conference (EMPC), September 201 1, p1-8, Schelling et al., "Managing Beat for Electronics," Materials Today vol.8 (2005) p.30-35, Gwinn et al., "Performance and Testing of Thermal Interface Materials", Microelectron J., 34 (2003) p.215-222). This problem however also arises with VACNT carpets, as recognized in the aforementioned publications by Prasher et al. (2006) and Cola et al. (2010).

An object of the present invention is to overcome at least in part the disadvantages of the prior art mentioned above.

Another objective of the present invention is to propose TIMs with optimized and industrialized performances. Objects of the invention

According to the invention, the above objectives are achieved by using metal matrix composite materials incorporating a VACNT mat, in which the metal matrix coats the VACNTs. CNTs have excellent thermal conductivity in the direction of their length. Vertically aligned on a substrate, they form a "carpet" and give the latter the ability to conduct heat in the direction of its thickness, this thickness being defined by the average length of said VACNT.

According to the invention, the VACNTs deposited on a substrate are coated with a metal matrix to form a metal matrix composite incorporating a VACNT mat. In the metal matrix, the heat flux diffuses in all directions. Copper is preferred because of its high intrinsic thermal conductivity (~ 400 Wm ^-1 .K ^-1 ). Thus, the inclusion of VACNT in a metal matrix which surrounds them not only makes it possible to increase the overall thermal conductivity of the TIM but also to improve the distribution of the thermal flux in a preferred direction, namely perpendicular to its thickness. The inclusion of VACNT in a metal matrix which enrobes them also makes it possible to reduce the coefficient of thermal expansion ("coefficient of thermal expansion" in English or CTE) of the metal matrix, ie to reduce the expansion of the volume of the metal matrix when operating the devices including them.

The present invention relates to a method of manufacturing a composite material comprising vertically aligned nanotubes (VACNT) and a copper metal matrix, coating said nanotubes, said method comprising at least the following steps, namely: a) a supply of vertically aligned nanotubes (VACNT) deposited on a substrate,

 b) the electrochemical deposition of said metal matrix on said nanotubes vertically aligned from a solution which is in contact with said VACNTs, said solution comprising at least one precursor of said metal matrix and at least one an organic solvent,

said method being characterized in that: said at least one precursor is selected from the group consisting of: copper acetate (Cu (CH ₃ COO) ₂ ), copper formate, copper propionate; and in that

 said electrochemical deposit is a galvanostatic electroplating.

In an advantageous embodiment, the metal matrix of the composite material according to the invention completely encompasses vertically aligned nanotubes (VACNT).

In an advantageous embodiment, the metal matrix is based on copper and the electrochemical deposition of copper is made from a bath, by a galvanostatic process, said bath comprising at least one precursor of metallic copper, such as a salt. copper, and at least one organic solvent. Advantageously, the copper salt is selected from copper acetate (Cu (CH ₃ COO) ₂ ), copper formate, copper propionate, a mixture thereof. Advantageously, the organic solvent is acetonitrile (C ₂ H ₃ N). In one embodiment, before step b), the vertically aligned nanotubes (VACNT) on said substrate are, prior to the electrochemical deposition, impregnated with a conductive ink, preferably comprising metallic particles, preferably of nanometric size. Advantageously, these metal particles are chosen from particles of copper, gold, silver and / or metal alloys. The impregnation with a conductive ink VACNT previously deposited on a substrate improves the subsequent electrochemical deposition of the metal matrix.

Preferably, the conductive ink is in aqueous solution. The conductive ink may be dispersed in organic solution, such as in a ketone-based solution such as acetone, in a solution based on alcohols such as ethanol or isopropanol or in a solution comprising an ester, a fatty acid or an oil.

In a preferred embodiment, before step b), the vertically aligned nanotubes (VACNT) on said substrate are impregnated with a conductive ink of metal particles having a nanoparticle size less than or equal to 100 nm in diameter. Advantageously, the VACNTs are multi-layered carbon nanotubes.

Advantageously, the substrate is chosen from electrically conductive substrates.

Advantageously, the substrate is selected from silicon, stainless steel, aluminum and metal alloys. In a preferred embodiment, the electrochemical deposition of said matrix on said VACNT is performed by stationary galvanostatic plating or galvanostatic electrodeposition by pulsed currents.

In a preferred embodiment, after step b) annealing of the obtained composite is carried out.

Advantageously, the annealing of the composite obtained is carried out at 400 ° C. under an inert atmosphere, preferably in an atmosphere in which the cumulative content of O ₂ and H ₂ O is less than 5 mass ppm.

In another preferred embodiment, said annealing is carried out under the action of a pulsed laser pulsed light source, preferably at an incident power P, of between 10 ⁹ and 10 ¹³ W / cm ² for a period of time. t between 10 ^"12 and 10 ^{" 9} seconds.

In another preferred embodiment, said annealing is carried out under the action of a continuous wave laser, preferably at an incident power P, of between 10 ⁵ and 10 ¹⁰ W / cm ² for a time t between 10 ^"4 and 10 ^{" 2} seconds.

In another preferred embodiment, said annealing is carried out in the presence of a reducing gas, preferably in a mixture of hydrogen and nitrogen, preferably in a mixture of hydrogen and nitrogen containing 5 mol% of hydrogen, and preferably at 1000 ° C for 10 min with a heating rate of 50 ° C / min. Advantageously, in this case, said annealing is performed in a rapid thermal annealing furnace, preferably at an incident power P, comprised between 1 and 10 ² W / cm ² for a time t comprised between 1 and 10 ³ seconds.

In another preferred embodiment, the annealing of the composite is carried out by microwave, preferably at an incident power P, of between 1 and 10 ² W / cm ² for a time t between 10 ² and ⁴ seconds.

In another preferred embodiment, the annealing of the composite is photonic annealing.

Advantageously, the photonic annealing is carried out at an incident power P of between 10 ³ and 10 ⁵ W / cm ² for a time t between 10 ^"5 and 10 ^{" 2} seconds. Advantageously, the surface density of carbon nanotubes, relative to the surface of the substrate on which they have been deposited, is between 10 ⁹ / cm ² and 10 ¹¹ / cm ² and preferably between 3 × 10 ⁹ / cm ² and 5 x ^{10 10} / cm ² . In a preferred embodiment, when the electrochemical deposition of said matrix on said VACNT is carried out by pulsed-current electrodeposition: a constant current density J ₀ of between -5 mA cm ^-2 and -50 mA cm ^-2 , preferably between mA.cm -10 ^"and -40 mA.cm ^{2" 2,} and even more preferably between -15 mA.cm ^"2 and -35 ^{mA.cm" 2} is applied during periods (t ₀₎ of duration between 1 second and 10 seconds, preferably between 1, 5 seconds and 5 seconds, and even more preferably of a value of 3 seconds;

a pulse current density J ₀ + J _peak is applied during periods of pulse t _peak between 0.5 seconds and 5 seconds, preferably between 0.5 seconds and 3 seconds, and during which J _peak is between -20 mA.cm ^"2 and -100 mA.cm ^{" 2} , preferably between -25 mA.cm ^"2 and -75 mA.cm ^{" 2} , and even more advantageously between -35 mA.cm ^"2 and -60 mA .cm ^"2 . Preferably, J ₀ is between -20 mA.cm ^"2 and -30 mA.cm ^{" 2} , t ₀ is between 2.5 seconds and 3.5 seconds, t _peak is between 0.7 seconds and 1 , 3 seconds, and J _eak is between -45 mA cm ^"2 and -55 mA cm ^{" 2} .

Advantageously, J ₀ is between -23 mA.cm ^"2 and -27 mA.cm ^{" 2} , t ₀ is between 2.7 seconds and 3.3 seconds, t _peak is between 0.8 seconds and 1, 2 seconds, and J _P e _ak is between -47 mA.cm ^-2 and -53 mA.cm ^-2 .

Another subject of the invention is a composite comprising VACNTs embedded in a metallic copper matrix, capable of being prepared by a process according to the invention.

Preferably, the surface density of carbon nanotubes of the composite, relative to the surface of the substrate on which they have been deposited, is between 10 ⁹ / cm ² and 10 ¹¹ / cm ² and preferably between 3 × 10 ⁹ / cm ² and 5 x ^{10 10} / cm ² .

Advantageously, the length of said VACNT of the composite is greater than 200 μηη, preferably between 200 μηη and 400 μηη.

Advantageously, the volume fraction of the VACNT of the composite is between 5 and 8%. Preferably, the average distance between two adjacent nanotubes of the composite is between 40 nm and 100 nm.

Advantageously, the thermal conductivity of the composite, in the direction of the tube, is greater than 300 W / mK Another object of the invention is the use of a composite according to the invention as a thermal interface material, in particular in electronic devices.

Another object of the invention is the use of a composite according to the invention as an electrically conductive material, preferably in electrical cables.

Description of figures

Figure 1 illustrates various aspects of embodiments of the invention, without limiting its scope.

The nature of the solution serving as a galvanostatic bath influences the morphology of composite VACNT / metal matrix materials. The apparent copper deposit, observed with the naked eye, depends on the impregnation of the germs. More specifically, when a galvanic bath containing copper sulphate is used, the copper is mainly deposited on the surface, giving a copper appearance to the composite material. Conversely, in the case where the copper is deposited in depth, and not on the surface, the composite material remains dark.

The use of a copper acetate solution in place of a copper sulphate solution in the electroplating bath makes it possible to obtain a composite material comprising a higher concentration of crystals, distributed more evenly and penetrating deeper into the VACNT carpet. FIG. 1 presents scanning electron microscopy images of a VACNT carpet alone (FIG. 1 a), VACNT / Cu composite materials obtained by stationary galvanostatic electroplating (FIG. 1b, sample ECU006 '), composite materials VACNT / Cu obtained by pulsed galvanostatic electrodeposition (Figure 1c, sample ECU013). The morphology of VACNT / Cu composite materials observed at SEM confirms the presence of deep copper in the VACNT carpet as well as the alignment of VACNT within the composite material.

Figure 2 shows scanning electron micrographs of VACNT / Cu composite material obtained by potentiostatic electrodeposition (Figure 2). The sample 250 μηη thick, shown in Figure 2A was developed by introducing a substrate previously covered with a carpet of VACNT in a bath containing copper (solution of CuS0 ₄ 0.6 M and d H ₂ S0 ₄ to 1.85 M) and to apply a voltage of -0.9 V for 1 minute (see sample ECU054). The sample shown in Figure 2B was developed as the sample ECU054 in a bath additionally comprising 100 mg of PEG. The SEM images show that the copper is deposited on the surface of the VACNT carpet in a compact manner encapsulating the upper part of the carbon nanotubes without deep impregnation of the VACNT carpet.

FIG. 3 presents optical microscopy images as well as scanning electron microscopy images of VACNT / Cu composite materials obtained by stationary galvanostatic electroplating from a galvanic bath comprising a solution of copper sulphate (FIG. 3a and FIG. 3a). ) or a solution of copper acetate (Figure 3b and Figure 3b ').

The nature of the solution serving as a galvanostatic bath influences the morphology of composite VACNT / metal matrix materials. In FIGS. 3a and 3b, the apparent copper deposit, observed with the naked eye, depends on the impregnation of the seeds. More specifically, when a galvanic bath containing copper sulphate is used (cf Figure 3a), the copper is mainly deposited on the surface, giving this coppery appearance to the composite material. Conversely, in the case where the copper is deposited in depth, and not on the surface, the composite material remains dark (see Figure 3b where copper acetate is used in the electroplating bath).

With regard to SEM images, the use of a copper acetate solution (FIG. 3b ') instead of a solution of copper sulphate (FIG. 3a') in the electroplating bath makes it possible to obtain a composite material comprising a higher concentration of crystals, distributed more homogeneously and penetrating more deeply within the VACNT carpet.

Figure 4 shows a scanning electron micrograph of a VACNT / Cu composite material obtained by stationary galvanostatic electrodeposition from a bath comprising copper acetate (Sample ECU055) for 10 min at a low current density of - 200 mA.cm ^"2 .

FIG. 5 shows a scanning electron microscopy image of a VACNT / Cu composite material obtained by stationary galvanostatic plating from a bath comprising copper acetate (Sample ECU056) for 30 min at a low current density of -.. 32 mA.cm ^"2 Figure 6 shows electron micrographs scanning composites VACNT / Cu obtained by electrodeposition stationary galvanostatic The ECU055 sample is shown in Figure 6A sample ECU056 is shown in Figure 6B. . detailed description

The term "thermal interface material" or TIM any material used to transport and dissipate the heat emitted by an electronic component in contact with the TIM.

The "volume fraction" of the VACNT in the composite VACNT / metal matrix Î _C NT is defined by the following formula: vnl ~ r û arm / rp cm where PCNT represents the density of the multi-layer carbon nanotubes in g / cm ³ measured by pycnometry (it is about 2.2 g / cm ³ ) and p _array represents the density of VACNT within the VACNT composite material / metal matrix in g / cm ³ . p _array is determined by the following ratio ^MCOM P ° ^STTE _OR m _composite , S _composite , e _composite

composite ^x composite

correspond respectively to the mass of the composite, to the composite surface and to the thickness of the composite measured at vernier caliper P _C NT-

The term "electrochemical deposition" means any deposition carried out by an electrochemical method, preferably a potentiostatic electrochemical method or a galvanostatic electrochemical method, in particular of stationary or pulsed type. The galvanostatic electrochemical method is preferred.

The present invention relates to novel metal matrix composite materials incorporating a "nanotube mat", that is, vertically aligned nanotubes (VACNTs) deposited on a substrate. In the context of the present invention, the acronym VACNT includes carbon nanotubes, carbon nanofibers but carbon nanofibers are less preferred as will be explained below. This substrate is substantially flat. Said nanotubes are carbon nanotubes (CNTs), with a single or multiple wall, preferably multiple because these multi-layer nanotubes are easily industrializable. The metal matrix is preferably copper because of its high thermal conductivity. The metal matrix may also be silver, aluminum, tungsten or gold. According to the invention, said VACNT / metal matrix composite materials can be obtained by a method in which a mat of vertically aligned nanotubes (VACNT) deposited on a substrate is supplied, and said metal matrix is deposited on said nanotubes vertically aligned by a technique. electrochemical solution from a solution comprising the precursor (s) of said metal matrix which is in contact with said VACNT. This electrochemical technique is preferably galvanostatic electroplating, and preferably a pulsed galvanostatic technique. Synthesis of vertically aligned nanotubes (VACNT) on a substrate

Carbon nanotubes can be synthesized according to various methods known per se, in particular by electric arc ablation, laser ablation, or chemical vapor deposition (CVD). According to the invention, the vertically aligned carbon nanotubes (VACNT) are deposited by chemical deposition from a vapor phase (CVD). CVD is a fast, inexpensive and industrially available method for producing high-performance, high-purity solid materials.

Two variants of catalytic CVD (CCVD) may be suitable for VACNT deposition. A first multi-step process is known as such (see Kukovitsky et al., "Correlation between metal catalyst particle size and carbon nanotube growth", Chemical Physics Letters vol.355, (2002) p.497-503 ) and most often requires the use of a reaction promoter (see Yasuda et al., "Improved and large area single-walled carbon nanotube forest growth by controlling the flow direction", ACS Nano vol.3, (2009) This first process lends itself less to mass production than the one-step aerosol-assisted CCVD process, in which the catalyst and carbon precursors are simultaneously introduced into the reactor. (see Mayne et al., Chem Phys Lett vol.338 (2001) p.101, see also Andrews et al., "Continuous production of aligned carbon nanotubes: a step doser to commercial realization," Chem Phys Lett Vol.303 (1999). p.467-474), the catalyst nanoparticles are form in the gaseous phase and are then deposited on the substrate (see Castro et al., "The role of hydrogen in the aerosol-assisted chemical vapor deposition process in producing thin and densely packed vertically aligned carbon nanotubes", Carbon vol.61 ( 2013) p.585-594) where they constitute seeds for the nucleation and the continuous growth of carbon nanotubes at atmospheric pressure and without the addition of a promoter. The development of VACNTs deposited on a substrate by the aerosol-assisted CCVD process can also be carried out in the presence of a promoter. This process is robust and quite simple to implement. This aerosol assisted CCVD process is described in documents FR 2 841 233, FR 2 927 619 and FR 3 013 061. The experimental setup is based on the use of aerosol generators which ensure the continuous supply of catalyst precursors and carbon precursors (for example, ferrocene dissolved in toluene) without any special preparation of the substrate. This method makes it possible to typically obtain multi-layer NTC mats aligned directly on the walls of the reactor or on substrates of a different nature (silicon, quartz, carbon, metals) placed in the reactor. The resulting nanotubes are typically free of highly crystalline carbon byproducts, with average diameters adjustable between 20 nm and 50 nm and a density of up to 10 ^10. CNT / cm ^2. The reactor can be used at atmospheric pressure, and the synthesis temperature is typically in the range of 600 ° C to 1090 ° C.

In an advantageous embodiment, the development of a multi-layer NTC mat vertically aligned on a substrate comprises a CVD synthesis step followed by a heat treatment step.

In the context of the process according to the invention, the substrate may be quartz, silicon, stainless steel, aluminum, and preferably an electrically conductive substrate.

The growth temperature of the VACNTs is chosen to be greater than or equal to the catalytic decomposition temperature of the carbon precursor and the thermal decomposition temperature of the catalyst precursor. In the case of the use of ferrocene as precursor of the catalyst and toluene as carbon precursor, a minimum temperature of 650 ° C. is necessary and a temperature of 850 ° C. makes it possible to increase the kinetics of reaction.

The temperature of VACNT growth must also be adapted to the substrate. Indeed, the growth temperature of the VACNT must not exceed the melting temperature of the substrate so as not to alter the latter.

In an advantageous embodiment, a substrate is used which supports a temperature of between 450 ° C. and 1090 ° C. (for example: silicon, quartz, alumina, glass, copper, steel, aluminum). This substrate is placed in a CVD reactor (or furnace) for heating and injecting gases and aerosols. The air present in the reactor is expelled by injecting a neutral gas (for example Argon, Nitrogen or Helium) and / or by pumping. The temperature of the oven is raised to the synthesis temperature, typically between 600 ° C. and 900 ° C. under a stream of neutral gas. A reaction mixture of gases and vapors is then injected. A typical volume composition of this reaction mixture consists of about 37% argon, about 28% hydrogen, about 28% acetylene and about 8% toluene vapors preheated to about 250 ° C in which about 10% is dissolved. % by weight of ferrocene. For example, for a reactor whose working section is 10 cm ² , at a temperature of 615 ° C., the mixture can be obtained by injecting at a pressure close to atmospheric pressure 0.2 slm of argon, 0.15 slm. of hydrogen, 0.15 slm acetylene and 4.8 g / h of a toluene-ferrocene mixture with a ferrocene mass concentration of 10%. Another typical volume composition of this reaction mixture consists of about 1/3 of hydrogen, about 1/3 of argon and about 1/3 of toluene vapors preheated to about 200 ° C. in which about 2.5% is dissolved. mass of ferrocene. For example, for a reactor with a working section of 10 cm ² , at a temperature of 850 ° C, can the mixture be be obtained by injecting at a pressure close to atmospheric pressure 1 slm of argon, 1 slm of hydrogen and 100 g / h of a toluene-ferrocene mixture with a ferrocene concentration equal to 2.5%. Under these conditions, an exposure time of between 25 min and 75 min makes it possible to obtain VACNT mats with a thickness of between 600 μηη and 800 μηη. The synthesis of VACNTs can be carried out on an oxide sub-layer such as Al ₂ O ₃ or SiO ₂ alumina _. The use of an under-layer of oxide can make it possible to limit the diffusion of the catalyst into the substrate. , to promote its reactivity, to prolong its duration of effectiveness and to favor the synthesis of VACNT of small diameter. Thus, under these conditions it is possible to grow vertically aligned carbon nanotubes, the height of which can vary from micrometer to millimeter, as a function of temperature and duration.

After synthesis, the thermal treatment of VACNT is advantageously carried out. According to an advantageous embodiment, the heat treatment of VACNT is carried out at a temperature of between 1900 ° C and 2100 ° C, and advantageously at about 2000 ° C for a time t, under an inert atmosphere of argon. This heat treatment leads to the sublimation of iron (catalyst) and more generally to the purification of nanotubes, and it promotes the structural reorganization of nanotubes; thus, an increase in the atomic order (crystallinity) is observed, which results in the increase of the thermal conductivity. The overall thermal conductivity of VACNT carpets is approximately 3.9 W / m.K. After heat treatment, the thermal conductivity of the VACNT mats is about 1 1 W / m.K.

Advantageously, the heat treatment of the VACNT is carried out for a time t of between 1 hour and 3 hours, preferably for approximately 2 hours with a temperature rise rate of the order of 10 ° C./min and a rate of descent of temperature of the order of 30 ° C / min. The rise and fall rates in temperature have no impact on the structure of the VACNT carpet, they are chosen to optimize the operation of the heat treatment device.

The height of the VACNT mats produced is advantageously between 200 μηη and 1.5 mm. Below 200 μηη, the VACNT mat is not mechanically stable and the mat is friable. The length of the nanotubes obtained is very homogeneous. For example, for a length of about 1000 μηη, the standard deviation is about ± 50 μηη, preferably about ± 25 μηη, and can even be as much as ± 20 μηη. Standard deviation is a measure used to characterize the dispersion of a distribution or sample. A small standard deviation corresponds to a small dispersion. For each substrate, the VACNT synthesis process can be optimized by modulating the proportions of precursors, catalysts such as ferrocene or the flow of inert gas so as to give the VACNT carpet a particular geometry where the carbon nanotubes are distributed. laterally with a spacing between the carbon nanotubes of between 40 and 100 nm. The VACNT obtained are multi-layer nanotubes and their diameter is between 10 nm and 80 nm. Advantageously, the density of carbon nanotubes in the composite VACNT / metal matrix is less than 5 × 10 ¹¹ / cm ² , preferably is between 10 ⁹ / cm ² and 10 ¹¹ / cm ² and even more preferably about 10 ¹⁰ / cm ^2. cm ² depending on the type of substrate used. In this density range, the VACNT carpet is dense enough to give the composite a good mechanical strength while including sufficiently large interstices to allow the precursor of the metal matrix to penetrate deep into them.

The volume fraction of the CNT is between 5 and 8%. One of the problems that the inventors have encountered is the homogeneous penetration of the metal between the nanotubes. The depth of penetration depends both on the average spacing between nanotubes, which should not be too low, and the length of the nanotubes. Moreover, to increase the thermal conductivity in the direction of the carpet thickness it is preferable to have a large number of small diameter nanotubes rather than a small number of large diameter nanotubes. The VACNT geometry has an inter-spacing between tubes of the order of 40 nm and 100 nm and a tube length / tube spacing ratio of between 1000: 1 and 2500: 1. According to the state of the art these geometric characteristics generally make difficult the deposition of copper in the interstices of the carpet. The electrochemical deposition of a metallic copper matrix on the vertically aligned nanotubes can be performed by galvanostatic electrodeposition, in stationary or pulsed mode.

Electrochemical deposition of a metallic cuiyre matrix on nanotubes vertically aligned by stationary galvanostatic electroplating

According to an essential technical characteristic of the invention, the metal matrix is deposited electrochemically (electrodeposition) in galvanostatic mode. This allows good penetration of the metal between the nanotubes. The electrodeposition method has the advantage of easily filling the interstices between the different nanotubes of carbon. This controls the matrix filling rate in the VACNT carpet and thus adjusts the volume fraction of the matrix that is deposited.

The substrate previously covered with a VACNT mat is immersed in a bath containing the elements to be deposited and a direct current J is applied between an auxiliary electrode and a working electrode consisting of said substrate previously covered with VACNT. This method makes it possible to impose a constant particle flux to the VACNT mat, while shifting the potential threshold, and ultimately to deposit the metal in the VACNT mat further in depth with respect to other deposition techniques including the method potentiostat. The current J can be constant throughout the duration of the electroplating, or it can be pulsed. The current density J applied during galvanostatic electroplating is advantageously between -100 mA.cm ^-2 and -10 mA.cm ^-2 , preferably between -75 mA.cm ^-2 and -15 mA cm ^-2 , and even more preferably between -50 mA.cm ^-2 and -15 mA.cm ^-2 . By way of example, the galvanostatic copper plating can be carried out with a stationary current density of approximately -25 mA cm ^-2 , which gives good results.For the copper plating, the galvanostatic bath advantageously comprises at least a metal precursor of said metal matrix such as a copper salt and at least one organic solvent The copper salt may be copper sulphate and / or copper acetate (Cu (CH ₃ COO) ₂ ) and acetonitrile (C ₂ H ₃ N) The inventors have found with measurements by goniometer of wettability that acetonitrile promotes the impregnation of copper in the VACNT carpet.

The concentration of the solution containing Cu ²⁺ ions is advantageously between 0.3 mol / L and 0.6 mol / L, and advantageously about 0.3 mol / L. For Cu ²⁺ concentrations above 0.6 mol / L, the solution containing Cu ²⁺ ions precipitates. With regard to copper sulphate, the use of a copper acetate solution with the same molar concentration makes it possible to obtain a higher concentration of copper crystals, distributed more evenly over the VACNT-coated substrate.

The nature of the solution serving as a galvanostatic bath influences the morphology of composite VACNT / metal matrix materials. Indeed, the use of a copper acetate solution (FIG. 3b ') instead of a solution of copper sulphate (FIG. 3a') in the electroplating bath makes it possible to obtain a composite material comprising a higher concentration of crystals, distributed more homogeneously and penetrating deeper into the VACNT carpet as shown by the SEM images (see Figure 3).

This can be explained by the wetting quality of the solution comprising the precursor (s) on the VACNT carpet. The wettability of VACNT carpets by Copper acetate and copper sulfate solutions were characterized by goniometric contact angle measurements. The smaller the contact angle, the more the liquid wets the VACNT carpet. The contact angle for copper sulphate is 130 °; that of the copper acetate is 100 ° indicating better wettability of the VACNT carpet by the copper acetate solution. The total time of electroplating is a function of the thickness of the desired deposit and the current density applied during electroplating. Thus, for a thickness of the order of 250 μηη at a current density of approximately -50 mA.cm ^-2 , the total duration of electrodeposition is approximately 1 hour.

Electrochemical deposition of a metal matrix on said nano-tubes vertically aligned by a pulsed galvanostatic technique (electrodeposition by pulsed currents)

In a preferred manner, electroplating is carried out in pulsed galvanostatic mode. The pulsed current electrodeposition technique consists of immersing the substrate previously covered with VACNT in a bath containing ions of a metal to be deposited, and applying an electric current pulsed between a counter-electrode and a working electrode constituted by said substrate previously covered with VACNT; which leads to a variation of the potential as a function of time. The pulsed mode is characterized by the alternation between a constant current of a first current density J ₀ , applied for a time t ₀ , and a constant current of a second density J _peak , applied for a duration t _pea k-

More precisely, in this method, the current alternates rapidly between two different values, the initial current density, J ₀ and the current density of the pulse corresponding to J _peak + Jo- At each pulse, the current density J _peak is applied for a time t _peak - Between each pulse, a current density J ₀ is applied for a duration t ₀ . For a given current density J, the optimal duration of the pulse depends on the nature and importance of the phenomena induced by the passage of the current, such as the depletion of the electroactive species diffusion layer or structural modifications. of the deposit.

The electroplating parameters were chosen in order to germinate small crystals in large quantities within the VACNT carpet, then to grow these seeds so as to encapsulate each carbon nanotube by the metal matrix.

Thus, the starting current density, J ₀ , is advantageously between -5 mA.cm ² and -50 mA.cm ^-2 , preferably between -10 mA.cm ^-2 and -40 mA.cm ^-2 , and still more advantageously between -15 mA.cm ^-2 and -35 mA.cm ^-2 , a value of about -25 mA.cm ^-2 is particularly suitable with copper. The current density of the pulse, J _peak , is between -20 mA.cm ^"2 and -100 mA.cm ^{" 2} , preferably between -25 mA.cm ^"2 and -75 m A.cm ^{" 2} , and even more preferably between -35 mA.cm ^-2 and -60 mA.cm ^-2 .

The period between two pulses t ₀ is between 1 sec and 10 sec, preferably between 1, 5 sec and 5 sec, and even more preferably between 2 sec and 4 sec; a value of 3 seconds is particularly suitable.

The period of the pulse, t _peak , is advantageously between 0.5 sec and 5 sec, preferably between 0.3 sec and 3 sec, and preferably about 1 second.

By way of example, the electroplating of the copper can be carried out in pulsed galvanostatic mode with J ₀ = -25 mA cm ^-2 and t ₀ = 3 s, J _pea k = -50 mA cm- ² and t _peak = 1 s.

Thus the layer deposited by this method, whether its structure, its appearance, its thickness can be modulated by parameters such as the starting current density, J ₀ , the current density of the pulse, J _peak , the pulse period, t _peak , the period between two pulses t ₀ and the total duration of electroplating t _to t. The modulation of these parameters makes it possible to influence in particular the germination, the size, the quantity of the deposited crystals and in fine on the composition and the thickness of the deposited layer.

According to the observations of the inventors, to obtain a homogeneously deposited layer composed of numerous small crystals by this method, it is preferable to have pulses of short duration, separated from each other by a relaxation phase of the system. a sufficiently long duration t _off , so that the diffusion layer can be recharged in excess electroactive species.

By this method of electroplating, the nucleation of crystals in large quantities and small sizes as well as the growth of crystals allowing the encapsulation of each carbon nanotube could be observed. By way of example, it can be observed that the copper deposition corresponds to a morphology of beads with a diameter ranging from about 1 μm to 5 μm or even 10 μm in pulsed galvanostatic mode. Without wishing to be bound by this theory, the inventors believe that in this embodiment the metal ions enter the bottom of the carpet, in the interstice between two nanotubes, to create a seed, which will grow laterally until completely filled. space available, which ends the ball's growth. The stationary galvanostatic mode gives a less fine morphology, with balls of about 10 to 20 μηη in diameter, and the metal matrix thus deposited tends to be less dense in the center of the thickness than at the bottom of the carpet and close to its thickness. area. For these reasons, the pulsed-current embodiment of the galvanostatic plating is preferred.

Impregnation of the entire thickness of the VACNT carpet depends on the composition of the plating bath, the nature of the precursor of the metal matrix and the density of the CNTs in the carpet. When galvanostatic electrodeposition, the galvanic bath used must contain in solution at least one copper precursor, ie copper ions Cu ^2+. Specifically, solutions of copper acetate (Cu (CH ₃ COO) ₂ ), copper formate and / or copper propionate are used. The Cu ²⁺ concentration of the solution containing Cu ²⁺ ions is between 0.3 mol / L and 0.6 mol / L, and advantageously about 0.3 mol / L. The inventors have found that, with respect to copper sulphate, the use of a copper acetate solution with the same molar concentration makes it possible to obtain a higher concentration of crystals, distributed more evenly over the substrate coated with VACNT.

According to an advantageous embodiment, the galvanic bath comprises a solution of copper acetate, preferably at 0.3 mol / l.

Advantageously, the galvanic bath may contain acetonitrile. Acetonitrile promotes the impregnation of copper seeds in the VACNT carpet. According to an advantageous embodiment, the concentration of the galvanic acetonitrile bath is between 0.1 mmol / L and 1 mol / L, advantageously 0.5 mmol / L. The concentration of electroactive species in the diffusion layer is also a parameter influencing the quality of the deposit. In fact, the renewal of the galvanic bath makes it possible to recharge the bath with Cu ²⁺ ions and ^finally to obtain a high concentration of copper throughout the VACNT carpet with the presence of spheres, clusters of copper crystals, increasing in a homogeneous and isotropic way. Advantageously, the galvanic bath is renewed at all times t, t being between 30 minutes and 5 hours, preferably between 1 hour and 4 hours, and even more preferably every 3 hours.

The total duration of the electroplating is a function of the thickness of the desired deposit. Thus, for a thickness of about 250 μηι, the total duration of electroplating t _to t is between 2.5 hours and 3.5 hours, preferably about 3 hours. For a sample of 1 mm thick, the total time of electroplating t _to t is between 10 hours and 14 hours, preferably about 12 hours with a renewal of the solution every 3 or 4 hours. The optimization of these different parameters makes it possible to develop a composite material consisting of VACNT in a metal matrix in which the metal deeply impregnates the VACNT carpet in contrast to the potentiostatic methods where the metal is deposited on the surface of the VACNT carpet in a compact manner, encapsulating only the upper part of the carbon nanotubes without deep impregnation.

According to this method a VACNT / Cu composite was made and characterized by EDX. The VACNT / Cu composite has a high concentration of metallic copper and carbon.

The morphology of the deposit observed at SEM (see Figure 1) confirms the presence of copper deep in the VACNT carpet and the alignment of VACNT within the composite material. VACNT alignment within the composite allows the composite material to maintain the unidirectional character of the nanotubes and increase the thermal conductivity of the composite.

The volume fraction of the CNTs in the composite VACNT / metal matrix is between 5 and 8% when the deposition of the metal matrix is carried out by stationary or pulsed galvanostatic electroplating.

The thermal diffusivity of the composite VACNT / metal matrix was measured from a flash analysis method, the differential scanning calorimetry specific heat in modulated mode and the bulk density was determined by measuring the mass of the composite and by measuring the thickness and surface of the composites to determine the volume of the composites. The thermal conductivity of the elaborated VACNT / Cu composites is estimated between 300 and 385 W / mK.

The process according to the invention can also be implemented with vertically aligned carbon nanofibers. This embodiment is less preferred. With the methods of the state of the art, a surface density of vertically aligned carbon nanofibers is not as high as vertically aligned carbon nanotubes. This tends to reduce the thermal conductivity of the composite obtained.

After synthesis of the VACNT / Cu composites, an annealing (or sintering) of the composite can be carried out in order to improve the electrical conductivity, the thermal conductivity and the recrystallization of the metal. This annealing facilitates the coalescence of the grains making it possible to increase the mechanical strength of the composite and to reduce the percentage of air present in the composite. The annealing may be carried out under an inert atmosphere, for example with a percentage of 0 ₂ and H ₂ O less than 5 ppm) at 400 ° C., using a heating plate.

The annealing can be carried out in a rapid thermal annealing furnace at 1000 ° C for 10 min at a rate of 50 ° C / min in the presence of a reducing gas, preferably in a mixture of hydrogen and nitrogen, plus preferably in a mixture of hydrogen and nitrogen containing 5 mol% of hydrogen.

Photonic annealing can be achieved by exposing the VACNT / Cu composites to a high-intensity bright flash applied on the opposite side of the substrate, over an extremely short period (<10ms) and with a fairly wide spectrum that encompasses the visible spectrum. A xenon arc lamp is suitable. This technique allows annealing in an uncontrolled atmosphere without elevation of temperature. This technique is particularly suitable for VACNT / Cu composites whose substrate is flexible and has a thermal resistance below 300 ° C.

Examples

The invention is illustrated below by examples which, however, do not limit the invention. These examples relate to a method for manufacturing metal matrix composite materials made from VACNT deposited on a substantially plane substrate, the characterization of these materials and their uses as thermal interface material.

Development of vertically aligned nanotubes (VACNT)

NTCs vertically aligned on a 1 cm ² surface silicon wafer were deposited by chemical vapor deposition (CVD). The substrate was heated at a temperature of 850 ° C in an oven and then exposed to this temperature for 75 minutes to a gaseous mixture of toluene and ferrocene, which was introduced into the reaction chamber in the form of an aerosol in an argon flow. The pressure was atmospheric. So we got a product called here "VACNT carpet".

After synthesis, a VACNT mat of 1 cm ² to 10 mg was obtained in which the carbon nanotubes have an average external diameter of 45 nm. The density of the carpet (excluding substrate) was measured by pycnometry and was 2.2 g. cm ^"3, and the thickness of the mat (ie the length of the nanotubes) was 450 μηη about. Other mats VACNT were prepared by the same method and used in the context of the manufacture of a composite material VACNT / copper matrix (see Tables 1 and 3 below). Manufacture of VACNT composite material / metal matrix by electroplating in stationary galvanostatic mode

Galvanostatic plating consists of immersing the substrate previously coated with VACNT in a bath containing the elements to be deposited (10 ml) and applying a constant current between an auxiliary electrode and a working electrode consisting of said substrate previously covered with VACNT; which leads to a variation of the potential as a function of time. This method makes it possible to impose a constant particle flux to the VACNT mat, while shifting the potential threshold, and ultimately to deposit the metal in the VACNT mat further in depth with respect to other deposition techniques including the method potentiostat.

The VACNT carpet previously obtained was thus introduced into a galvanizing bath having a concentration of copper acetate or copper sulfate in the presence of acetonitrile as mentioned in Table 1 below.

After introduction of the VACNT carpet into the electroplating bath, copper plating was performed at a constant current density J of -50 mA.cm ² .

The total time of electroplating was 1 hour. It was chosen according to the final thickness of the desired deposit of the order of 250 μηι.

Table 1: Experimental conditions for the elaboration of VACNT / Cu composites by electroplating in galvanostatic mode at a constant current density of -50 mA.cm ^-2

 Renewal of

Additive

 galvanostatic bath

Thickness name [CuS0 ₄ ] [Cu (CH ₃ COO) ₂ ] [C ₂ H ₃ N] (PEG

 (number of the sample (μηι) (mol / L) (mol / L) (mmol / L) (100

 mg renewals)

 made)

ECU011 250 0.6 1

 ECU018 250 0.3 0.5 5

 ECU020 250 0.6 Solvent

 ECU031 200 0.6 Solvent

 ECU008 200 0.6 1

 ECU006 300 0.6 1

 ECU006 '300 0.6 1

 ECU009 200 0.6 2

 ECU010 250 0.3 0.5 2

ECU017 250 0.3 0.5 X After synthesis, the electrodeposition was stopped. The samples were then extracted from the plating bath, rinsed with water and then dried in the open air. The samples were then analyzed by scanning electron microscope and shown in Figure 1b.

Other tests were carried out on VACNT carpets previously obtained and detached from their substrate after shear synthesis according to a method known to those skilled in the art in order, on the one hand, to visualize more easily the deep impregnation of copper. and secondly to evaluate the influence of the electrical contact on the deposit by making a contact on the front face (F. Av) and / or back of the VACNT mats.

The term "backface contact" (hereinafter abbreviated F. Ar) means an electrical contact 10 made on the surface of the VACNT mat detached from their substrate after synthesis, and this on the surface previously in contact with the substrate.

"Front-face contact" (hereinafter abbreviated as F. Av) means an electrical contact made on the surface of the VACNT mat detached from their substrate after synthesis, both on the surface previously in contact with the substrate, on that diametrically opposed to this surface and on a surface contiguous to these two surfaces.

After introduction of the VACNT carpet into the electroplating bath, a stationary galvanostatic copper plating was performed.

Table 2: Experimental conditions allowing the elaboration of VACNT / Cu composites without substrate by electrodeposition in stationary galvanostatic mode

20 After synthesis, the electrodeposition was stopped. The samples were then extracted from the plating bath, rinsed with water (or with acetonitrile for sample ECU041) and then dried in the open air. The samples were then analyzed by scanning electron microscope and shown in Figures 4 and 5, allowing us to evaluate the effect of the constituents of the plating bath.

The inventors have found that a solution of copper sulphate does not give good results for developing VACNT / Cu composite materials by electroplating in a stationary galvanostatic mode because the copper does not impregnate the entire depth of the VACNT carpet. On the other hand, the use of a solution comprising copper acetate and acetonitrile as a galvanic bath makes it possible to obtain VACNT / Cu composites in which the copper deeply impregnates the VACNT carpet (see FIG. sample ECU055 and FIG. 5, sample ECU056) where the presence of spherical crystals in the carpet is observed, distributed homogeneously. Figure 4 (sample ECU055) shows the presence of spheres of large diameter spaced by several micrometers between them. This is due to a large current density (J = - 200 mA.cm ^"2). And a low deposition time (t = 10min) Conversely, Figure 5 (sample ECU056) shows diameter spheres However, this can be explained by a longer deposition time (t = 10 min) and a lower current density (J = -32 mA.cm ^"2 ). As shown by the SEM images of Figures 4 and 5, the size of the crystals is proportional to the duration of the deposition, just as the number of seeds is proportional to the current density applied.

Sample ECU055 was synthesized with front panel contact, and sample ECU056 by rear panel contact. SEM images of these samples are presented respectively in Figure 6A and 6B. A larger deposit on the surface is observed for the sample ECU055 which has been synthesized with a contact on the front face than on the back. The contact on the back side promotes better impregnation of copper throughout the carpet. Manufacture of VACNT composite material / metal matrix by galvanostatic electrodeposition by pulsed currents

The previously obtained VACNT carpet was introduced into a galvanizing bath (10 mL) having a concentration of copper acetate or copper sulfate in the absence or in the presence of acetonitrile as mentioned in Table 1, as well as the galvanostatic electrodeposition parameters by selected pulsed currents.

After introduction of the VACNT carpet into the plating bath, copper plating was carried out according to the parameters J ₀ , J _pe ak, t _on and t _off disclosed in Table 3. The total duration of electrodeposition, presented in FIG. Table 3 was chosen according to the final thickness of the desired deposit of the order of 250 μηι.

Experimental conditions for the elaboration of VACNT / Cu composites by galvanostatic electrodeposition by pulsed currents

In the case of sample ECU019, a renewal of the galvanic bath (10 ml) was carried out every hour in order to reload the solution. The renewal of the galvanic bath has resulted in a high concentration of copper throughout the carpet, with the presence of spheres, clusters of copper crystals, growing homogeneously and isotropically. After synthesis, the electrodeposition was stopped. The samples were then extracted from the plating bath, rinsed with water and then dried in the open air. The samples were then analyzed by scanning electron microscope and presented in Figure 1c. In the case of sample ECU015, 100 mg of polyethylene glycol (PEG) was added to the plating bath. The addition of PEG to the galvanic bath has no effect on the final structure of the VACNT / Cu composite obtained.

Another test was performed on a carpet of VACNT previously obtained and detached from its substrate. After introduction of the VACNT carpet into the galvanic bath, a galvanostatic copper plating by pulsed currents was carried out.

Table 4: Experimental conditions allowing the elaboration of VACNT / Cu composites without substrate by galvanostatic electrodeposition by pulsed currents

After synthesis, the electrodeposition was stopped. Sample ECU007 was then extracted from the plating bath, rinsed with acetonitrile and then dried in the open air.

The copper deposit observed was consistent and consistent, but when rinsing with acetonitrile, all the deposited copper oxidized, and detached from the carpet. The result is a densified carpet without copper. In order to overcome these undesirable effects, it is preferable to use a solvent different from that used in the electroplating bath. The use of water or oily compounds to rinse the VACNT / Cu composites after synthesis makes it possible to avoid the formation of cracks in the composite during drying.

Analysis of VACNT / Cu composites The thermal diffusivity of VACNT / Cu composites was determined by the Flash method. The VACNT / Cu composite was illuminated by a high energy radiation sink of a xenon lamp (XFA). The energy produced by the XFA was absorbed by the front face of the sample. The temperature of the back side of the sample was measured as a function of time. Diffusivity was determined according to the law issued by Parker et al. (Cf. Parker et al., "Flash Method of Determining Thermal Diffusion, Heat Capacity, and

 0 138 xL ^

Thermal Conductivity ", J. Appl Phys, 1961) such that a = _^'2 -.. Where L is the thickness of the analyzed sample, and t1 / 2 is the time required for the temperature on the rear face of sample reaches half of the maximum possible diffusivity for the material.

The thermal diffusivity of the carpet VACNT alone is approximately 9.5 × 10 ⁶ mm ² / s. The thermal diffusivity of the composite VACNT / Cu is estimated between 1 16 and 1 17 mm ² / s, in particular according to the following equation

Composite ^ ⁼ ^ composite- CPcomposite - ^ ^ ^ Composite -Composite SSt thermal COndUCtlVlté composite VACNT / Cu _composite Cp specific heat and p _Composite density of the composite is determined as follows.

The thermal conductivity of the VACNT / Cu composite was determined according to the law of the mixtures as a function of the thermal conductivity of each material (λ) and of their volume fraction in the composite (V _f% ). Given the configuration of VACNTs in a matrix and their anisotropic properties in terms of thermal conduction, the "parallel model" of the law of mixtures was used to evaluate the thermal conductivity of the composite (λ comosite) according to the following formula: λ composite ⁼ <¾ · matrix -Vf% matrix ⁺ -Vf% [1]

In this model, the contact resistance between these two materials is considered to be zero and it is considered that there is no transverse heat exchange between the materials. The volume fraction of vertically aligned nanotubes (CNTs) in the VACNT / Cu composite is determined experimentally by comparing the theoretical density of CNTs with that measured with the helium pycnometer. For the ECU013 sample, the volume fraction of the vertically aligned nanotubes, V _{f% C} was determined experimentally and is 5.2%.

The volume fraction of the CNTs in the composite is between 5 and 8%.

The volume fraction of the copper in the composite is from about 92 to about 95%.

The specific heat Cp and the density of the VACNT / Cu composite were measured. For the ECU013 sample, the specific heat of the VACNT / Cu (Cp) composite is 425 ± 20 J.kg ^-1 .K ^-1 at 25 ° C; and the density of the VACNT / Cu (p) composite is 7540 ± 178 kg.m- ³ . The equation [1] for thermal conductivity can also be written by explaining the specific heat and the density according to equation [2]:

P CPcomposite ⁼ Vf% Matrix-

Pu trice + Vf% CNT- PcNT ^■ + Vf% Air- PAÎT [2] considering for equation [2] the parameters measured for raw VACNT samples with Cp _C NTex. = 697 Kg J. ^{"1 .K" 1} pcNTex = 80 Kg. M ^"3, and the parameters given by Cp _Cu theo literature. = 385 ^{J.Kg" 1} .K ^"\ p _Cu Theo. = 8 960 Kg. M- ³ , Cp _A i _r theo. = 1000 J.Kg ¹ .IC ¹ , pAir thôo. * 1, 2 Kg.m ^"3 .

For the ECU013 sample, the volume fraction V _{f% air} and that of the copper V _f % copper were estimated from the following equation [3]: C composite ⁼ Vf% CNT-C cNT + Vf% _Copper matrix -

[3] considering parameters measured for raw VACNT samples with Cp _C NTex. "697 J.Kg ^{" 1} .K ^"1 , the parameters given by the literature Cp _Cu theo. = 385 J. Kg ^"1 .K ^{" 1} , Cp _A ir theo. = 1000 J.Kg ¹ .K ^"1 , that the sum of the volume fractions of air, copper and CNT is equal to 1 and that the previously determined volume fraction of vertically aligned nanotubes V _f % _C NT is 5 , 2%.

The "volume fraction" could thus be estimated and presented hereafter according to the data presented in the following estimation table, showing the theoretical determination of the specific heat and the density (see Table 5) according to previous parameters and variable parameters that are the volume fractions of air and copper.

Table 5: Theoretical determination of the specific heat and density of the VACNT / Cu composite as a function of volume fractions of air and copper

 Vf% Copper Vf% Air Cp Composite Composite P

%% J / Kg.K Kg / m ³

 94.8% 0% 401 8158

 89.8% 5% 432 7178

 84.8% 10% 463 6328

 79.8% 15% 493 5585

 74.8% 20% 524 4928

 69.8% 25% 555 4344

 64.8% 30% 586 3822

 59.8% 35% 616 3352

54.8% 40% 647 2926 49.8% 45% 678 2539

 44.8% 50% 709 2186

 39.8% 55% 739 1861

 34.8% 60% 770 1563

 29.8% 65% 801 1288

 24.8% 70% 832 1033

 19.8% 75% 862 796

 14.8% 80% 893,576

 9.8% 85% 924 370

 4.8% 90% 955 178

Therefore, according to the Cp and density values measured for the sample ECU013 and those theoretically presented in Table 5, for the VACNT / Cu composite, a volume fraction of air V _{f% Air theo according to Cp or p is} deduced. <5% and a volume fraction of copper V _{% Copper theo according to Cp or p} > 89.8%.

Considering now for equation [1], the parameters measured for raw VACNT samples with an experimental volume fraction of CNT V _{0 / oCNTexp} . ~ 5.2%, thermal conductivity _{CNTexp CNTs} . ^¾ 70 ^ .and those given in the literature, ie the theoretical thermal conductivity of air h _{Air Theo} «0.025 ^ - and that of copper _{Cu Theo} ^¾ h ^ 400 - we can estimate the total thermal conductivity of composite VACNT / Cu (see Table 6) according to the following equation [4]:

^ Composite ⁼ Vf% CNT- ^ CNT + Vf% Matrix _Q ■ ^■ ^ -Brass Copper ^ f% Air- ^ _Air [4]

Table 6: Estimation of the thermal conductivity of composite VACNT / Cu thermal conductivity of the composite

 Vf% Copper Vf% Air calculated according to equation [4]

 composite

 %% W / m.K

 94.8% 0% 383

 89.8% 5% 363

 84.8% 10% 343

 79.8% 15% 323

 74.8% 20% 303

 69.8% 25% 283

 64.8% 30% 263

 59.8% 35% 243

 54.8% 40% 223

 49.8% 45% 203

 44.8% 50% 183

39.8% 55% 163 34.8% 60% 143

 29.8% 65% 123

 24.8% 70% 103

 19.8% 75% 83

 14.8% 80% 63

 9.8% 85% 43

 4.8% 90% 23

Knowing the voluminal fractions of air and copper previously deduced

^(V f% Air Theo. According Pc ^{<5% and V} f% Copper theo. According Cp or p>^{89> 8%)> the} theoretical thermal conductivity of the composite VACNT / Cu can be estimated. It is between 363 W.rrfVK- ¹ <À _th , _nrimiP <383 \ ΛΛ ΓΤΓ ¹ . Κ ^"1

The VACNT / Cu composites obtained by galvanostatic electrodeposition by pulsed currents exhibit a greater copper impregnation than those obtained by stationary galvanostatic electroplating. Tests for manufacturing a VACNT composite material / metallic matrix by potentiostatic electrodeposition

The VACNT carpet previously obtained was introduced into an electrodeposition bath having a copper sulfate concentration in the absence or in the presence of acetonitrile and additives such as polyethylene glycol (PEG), chloride ions (Cl ), bis (3-sulfopropyl) disulfide (SPS) or Janus Green B (JGB) as mentioned in Table 7 below, as well as the electrodeposition parameters chosen.

PEG (Polyethylene Glycol) is known to be a wetting agent, preventing the deposition of copper at the top of the tubes. Moreover, by combining the PEG with chloride ions, the appearance of a passivation film is observed, making it possible to improve the deposition of the copper. SPS (bis (3-sulfopropyl) disulfide) and JGB (Janus Green B) are used as leveling additives, allowing preferential deposition in cavities rather than surface.

After introduction of the VACNT carpet into the bath, a potentiostatic copper plating was carried out at the fixed potentials disclosed (E) in Table 7. The total duration of electroplating (t), shown in Table 7, was chosen depending on the final thickness (e) of the desired deposit between 165 μηη and 600 μηι, preferably between 189 μηη and 260 μηη. Table 7: Experimental conditions for the preparation of composites VACNT /

Cu by potentiostatic electrodeposition

In Table 7, all the tests were carried out with a bath comprising 0.6 mol / l copper sulphate and 1.85 mol / l sulfuric acid. After the electroplating of copper, rinsing with water and drying in the oven at 60 ° C. under primary vacuum (-1 mbar) of the VACNT / Cu composites were carried out to overcome the gaseous release of sulfuric acid which hinders the thermal properties of TIM.

Figure 2 shows the observation by SEM of the sample ECU054 free of additive (see Figure 2A) and that prepared in the presence of PEG (ECU022, see Figure 2B). In both cases, the copper deposit is surface. No impregnation of copper by the carpet was observed.

Claims

claims

1. A process for manufacturing a composite material comprising vertically aligned nanotubes (VACNT) deposited on a substrate and a copper metal matrix coating said nanotubes, wherein electrochemical deposition of said metal matrix is carried out on said vertically aligned nanotubes from a solution which is in contact with said VACNTs, said solution comprising at least one precursor of said metal matrix and at least one organic solvent,

 and said method being characterized in that:

said at least one precursor is selected from the group consisting of: copper acetate (Cu (CH ₃ COO) ₂ ), copper formate, copper propionate; and in that

 said electrochemical deposit is a galvanostatic electroplating.

2. Method according to claim 1, characterized in that the vertically aligned nanotubes (VACNT) on said substrate are, prior to the electrochemical deposition, impregnated with a conductive ink, preferably comprising metal particles, preferably of nanometric size.

3. Method according to any one of claims 1 to 2, characterized in that the conductive ink is in aqueous solution or organic solution, preferably in a solution based on ketone, in an acetone solution, in a alcohol-based solution, in a solution of ethanol or isopropanol or in a solution comprising an ester, a fatty acid or an oil.

4. Method according to any one of claims 1 to 3, characterized in that the VACNT are multi-layer carbon nanotubes.

5. Method according to any one of claims 1 to 4, characterized in that the substrate is selected from electrically conductive substrates.

6. Method according to any one of claims 1 to 5, characterized in that the substrate is selected from silicon, stainless steel, aluminum and metal alloys.

7. Method according to any one of claims 1 to 6, characterized in that the electrochemical deposition of said matrix on said VACNT is carried out by stationary galvanostatic plating or galvanostatic electroplating by pulsed currents.

8. Process according to any one of claims 1 to 7, characterized in that after the electrochemical deposition annealing of the composite obtained is carried out.

9. Method according to claim 8, characterized in that said annealing of the composite obtained is carried out at 400 ° C under an inert atmosphere, preferably in an atmosphere in which the cumulative content of O ₂ and H ₂ O is less than 5 ppm mass.

10. The method of claim 8, characterized in that said annealing is carried out in the presence of a reducing gas, preferably in a mixture of hydrogen and nitrogen, and preferably at 1000 ° C for 10 min with a heating rate of 50 ° C / min.

1 1. Process according to claim 8, characterized in that said annealing of the composite is carried out by microwave, preferably at an incident power P, of between 1 and 10 ² W / cm ² for a time t between 10 ² and 10 ⁴ seconds.

12. The method of claim 8, characterized in that said annealing of the composite is a photonic annealing.

13. The method of claim 12, characterized in that said annealing is performed at an incident power Pi between 10 ³ and 10 ⁵ W / cm ² for a time t between 10 ^"5 and 10 ^{" 2} seconds.

14. Method according to any one of claims 1 to 13, characterized in that the surface density of carbon nanotubes, relative to the surface of the substrate on which they were deposited, is less than 5x10 ¹¹ / cm ² , and preferably between 10 ⁹ / cm ² and 10 ¹¹ / cm ² .

15. Method according to any one of claims 7 to 14, characterized in that, when the electrochemical deposition of said matrix on said VACNT is carried out by galvanostatic electrodeposition by pulsed currents:

a constant current density J ₀ between -5 mA.cm ^"2 and -50 mA.cm ^{" 2} , preferably between -10 mA.cm ^"2 and -40 mA.cm ^{" 2} , and even more advantageously between -15 mA.cm ^"2 and -35 mA.cm ^{" 2} is applied during periods (t ₀ ) of duration of between 1 second and 10 seconds, preferably between 1, 5 seconds and 5 seconds, and even more preferably of a value of 3 seconds;

a pulse current density J ₀ + J _peak is applied during periods of pulse t _peak between 0.5 seconds and 5 seconds, preferably between 0.5 seconds and 3 seconds, and during which J _peak is between -20 mA.cm ^"2 and -100 mA.cm ^{" 2} , preferably between -25 mA.cm ^"2 and -75 mA.cm ² , and even more preferably between -35 mA.cm ^{" 2} and -60 mA. cm ² .

16. A method according to claim 15, characterized in that J ₀ is between -20 mA.cm ^"2 and -30 mA.cm ^{" 2} , t ₀ is between 2.5 seconds and 3.5 seconds, tpea _k is between 0.7 seconds and 1.3 seconds, and J _peak is between -45 mA.cm ^-2 and -55 mA.cm ^-2 .

17. A method according to claim 16, characterized in that J ₀ is between -23 mA.cm ^"2 and -27 mA.cm ^{" 2} , t ₀ is between 2.7 seconds and 3.3 seconds, t ea _k is between 0.8 seconds and 1, 2 seconds, and J _peak is between -47 mA.cm ^"2 and -53 mA.cm ^{" 2} .

18. Composite comprising VACNTs embedded in a metal matrix, capable of being prepared by a process according to any one of Claims 1 to 17.

19. Composite according to claim 18, characterized in that the surface density of carbon nanotubes, relative to the surface of the substrate on which they have been deposited, is between 10 ⁹ / cm ² and 10 ¹¹ / cm ² and preferably between 3 x 10 ⁹ / cm ² and 5 x ^{10 10} / cm ² .

20. Composite according to any one of claims 18 and 19, characterized in that the length of said VACNT is greater than 200 μηη.

21. Composite according to any one of Claims 18 to 20, characterized in that the volume fraction of the VACNT is between 5 and 8%.

22. Composite according to any one of claims 18 to 21, characterized in that the average distance between two neighboring nanotubes is between 40 nm and 100 nm.

23. Composite according to any one of claims 18 to 22, characterized in that its thermal conductivity in the direction of the tube is greater than 300 W / m.K.

24. Use of a composite according to any one of claims 18 to 23 as a thermal interface material, in particular in electronic devices.

25. Use of a composite according to any one of claims 18 to 24 as an electrically conductive material, preferably in electrical cables.