Applied Nanoscience https://doi.org/10.1007/s13204-018-0692-1

ORIGINAL ARTICLE

An experimental study of the composite CNT/copper coating Valentin Ye. Panarin1 · Nikolai Ye. Svavil′nyi1 · Anastasiya I. Khominich1  Received: 16 January 2018 / Accepted: 16 February 2018 © Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract This paper presents experimental results on the preparation and investigation of the carbon nanotubes–copper composite material. Carbon nanotubes (CNTs) were synthesized on silicon substrates by the chemical vapor deposition (CVD) method and then filled with copper by evaporation from a melting pot in a vacuum. Copper evenly covered both the surface of the entangled tubes and the free substrate surface between the tubes. To improve the adhesion of tubes and matrix material, a carbon substructure was grown on the surface of tubes by adding working gas plasma to the CNT synthesis area. It is proposed to use a copper coating as a diffusion barrier upon subsequent filling of the reinforcing CNT frame by a carbide-forming materials matrix with predetermined physico-mechanical and tribological properties. Keywords  Composite materials · CNT synthesis · Plasma component · Diffusion barrier Abbreviations CNTs Carbon nanotubes CVD Chemical vapor deposition PAD Plasma-arc device

Introduction Intensive development of engineering contributes to the emergence of new materials and technologies for their processing. An intensive search for materials with unique mechanical, tribological, electrical, magnetic, and other properties is conducted in modern materials science. In particular, one promising direction is the development of composite materials, where individual unique properties of composite components are combined to obtain the desired characteristics of the final material. Due to a number of unique properties, carbon nanotubes (CNTs) are widely used as one of the composite components (Moghadam et al. * Anastasiya I. Khominich [email protected] Valentin Ye. Panarin [email protected] Nikolai Ye. Svavil′nyi [email protected] 1



G.V. Kurdyumov Institute for Metal Physics, National Academy of Sciences of Ukraine, Acad. Vernadsky Blvd, 36, Kiev‑142 UA‑03680, Ukraine

2015). A lot of publications have been devoted to production technology development and study of the properties of such composites (Bakshi et al. 2010). Herewith, CNTs are a reinforcing component (the “skeleton” of a composite), and the matrix is selected from a material with adjusted physicomechanical properties. The studies represent a specific interest of composite copper–CNT mixture, where copper acts as a matrix and CNTs are added to the mixture from units of weight percent of the total mass to tens of percent (Guiderdoni et al. 2013; Lim et al. 2006; Guiderdoni et al. 2011). As usually, samples for studying of such composites properties are manufactured by mixing copper fine powders with appropriate CNTs (singlewalled, double-walled, or multi-walled). Furthermore, the mixtures are subjected to pressing or sintering at high temperatures. It is difficult to influence the interphase boundary appearance and state in such mixtures, although it is one of the fundamental conditions for the formation of composite properties. With this approach, questions remain regarding the distribution of matrix atoms over the nanotubes surface, effect of the nanotubes morphology on this distribution, and temperature influence on the diffusion of copper atoms over the CNTs surface: is it possible to regulate cohesion and adhesion? If a carbide-forming metal is selected as the composite matrix material, then it reacts with the formation of carbide during contact with CNTs and its unique properties disappear. On the other hand, the use of a carbide-forming metal

13

Vol.:(0123456789)



Applied Nanoscience

as a matrix significantly expands the possibilities of creating composites with unusual properties. It is possible to solve the arisen problem by creating a diffusion barrier on CNTs’ surface from metal that is inert to the interaction with carbon and, correspondingly, with CNTs, and can prevent the formation of carbides of the metallic matrix of the composite. As a metal for the diffusion barrier, it is logical to choose copper, because it does not chemically interact with carbon (Guiderdoni et al. 2011). If copper is uniformly applied with a continuous layer on the surface of all CNTs, this coating will be prevented from the chemical interaction of carbide-forming metals with CNTs. Then, the metal of the composite matrix can be chosen on the assumption of its physico-mechanical properties, without worrying about the interaction with CNTs. Naturally, it is necessary to take into account a possible interaction of copper coating with the metal of the composite matrix. In the present work, the process of copper barrier layer applying on the surface of cold (< 100 °C) CNTs by condensing vapor, which appears above the molten copper in the melting pot, was experimentally developed. The feature of the proposed technology is that the atomic, i.e., the electrically neutral component of the copper vapor, easily penetrates into the voids between the CNTs, and condenses on the surface regardless of their geometry, distribution in space, shape, and size. The coating creates the necessary conditions for preventing interaction of CNTs with carbideforming metal of the composite. It is important that the developed technology of copper barrier layer applying is consistent with the growth of CNTs. It does not require depressurization of the vacuum chamber and is, in fact, an indissoluble continuation of the process of composite coating formation. The CVD method was chosen as a method of CNTs synthesis, because it is the most optimal from the point of view of the predictability of synthesis results. CNTs with a given diameter, morphology, and density of distribution over the substrate surface can be grown by the method of CVD (Shah and Tali 2016; Svavil′nyi et al. 2013). It is important from the point of view of ensuring the adhesion of copper directly to grown tubes, as well as to other functional materials Fig. 1  Scheme of the modernized ion-plasma sputtering NNV-6,6 type “Bulat” facility, equipped with an additional source of plasma: 1—vacuum chamber, 2—substrates, 3— heating table, 4—cathode, 5, 6—cathode and anode, respectively, of Penning plasma source, 7—multigrid separator

13

deposited later on this copper. The results of the present experiments were briefly presented in (Panarin et al. 2017a).

Experimental facility and CNTs production The experiments were carried out using facility of the scheme that is shown in Fig. 1. The CNTs were synthesized by CVD method on substrates located on a heated table in the center of the vacuum chamber. The plasma-arc device (PAD) for spraying of solid-state substances (Fe, Ni) used as catalysts is to the right of the table. For this purpose, a vacuum-arc discharge between anode and cathode of PAD was ignited. The device in the form of a Penning cell (Gabovich 1972) consigned for the ionization of working gases is to the left of the table. Thus, it was possible to supply the C2H2 working gas to the CNTs synthesis zone not only in the form of neutral molecules, but also in ionized form. The ratio of the ionized and neutral components of the working gas in the CNTs synthesis zone could be controlled by means of the special multigrid electrostatic device (Panarin et al. 2012) located at the outlet of the Penning source. It was possible to supply the potential independently to the electrically insulated heated table for controlling the plasma components of the working gas. The synthesis of CNTs was carried out at pressures of acetylene in a vacuum chamber within the range ­10−1–1 Pa, at table temperatures in the range 500–1000  °C, and the ratio of plasma and neutral acetylene components in the range 0.01–1%. We used oxidized silicon with oxide layer thickness ~ 0.3 μm as substrates, as well as silicon with a deposited TiN layer ~ 150 nm thick. TiN films were synthesized in the same facility by ignition of the vacuum-arc discharge on the PAD with the addition of nitrogen as the working gas. Ti cathode was used in the PAD in modes of TiN synthesis. The detailed description of the modes of TiN film synthesis on substrates in this setup is given in Panarin et al. (2017b) and Svavil′nyi (2016). Catalytic centers for CVD synthesis were formed from thin catalyst films on the substrate. Such films were obtained by depositing plasma streams from the evaporable cathode

Applied Nanoscience

of the PAD. The cathode was made of catalyst material (Ni, Fe). The working gas in this case was Ar. Furthermore, catalyst material thin films were annealed on the work table at a temperature of about 800 °C (and, therefore, at a temperature close to that of the substrates) for 10 s to 5 min in different experiments. After CNTs synthesis on a substrate, they were examined on a SEM using JSM-6490LV and TESCAN MIRA 3 LMU microscopes. Then, the substrates, if necessary, were attached to the working table for the deposition of copper films. The separate thermal evaporator was produced for copper deposition on the CNTs. The scheme of evaporator is shown in Fig. 2. In experiments on copper deposition on grown and investigated CNTs, this evaporator was placed in a vacuum chamber in place of a heating table. The table with attached substrates on it was installed above the evaporator. Devices continuously monitored all electrical parameters of the technological process. The table temperature was measured with a thermocouple, and the plasma parameters of the working gas flows above the work table were measured by the standard methods of plasma probe characteristics processing (Sereda and Tseluiko 2015). Langmuir electrical probe for controlling the plasma parameters was placed above the heating table and could move along and across the table.

Fig. 2  Scheme of the device for thermal evaporation of copper: 1— heating table, 2—substrate, 3—thermocouple, 4—shutter, 5—melting pot, 6—emitter, 7—evaporable material, 8—thermal screen, 9— molybdenum glass

Results and discussion Polycrystalline silicon plates 300 μm thick oxidized to a depth of ~ 0.3 μm and silicon plates on which a TiN film was synthesized to protect them from chemical interaction of the catalyst (in most experiments it was Ni) with Si-substrate were used as substrates with grown nanotubes. A characteristic view of catalytic centers, which were formed after annealing at 800 °C from a thin (~ 8 nm) nickel film, is shown in Fig. 3. The resulting catalytic centers have different sizes due to the roughness of the substrate surface and various conditions for the surface diffusion of nickel atoms. The different areas of the catalytic centers determine the different thicknesses of CNTs synthesized on them (Mohamed et al. 2006). Multi-walled CNTs with an average diameter of about 20 nm were synthesized on the resulting catalytic centers (see Fig. 4). The large or smaller fractures, giving a general morphology in the form of CNTs entanglement, are observed almost on each of the tubes. It should be noted that the morphology of the grown CNTs is determined to a large extent by the specific features of the physical processes

Fig. 3  Catalytic nickel centers formed on a Si/SiO2 substrate: a typical view of catalytic centers; b diagram of the catalytic centers distribution over areas

13



Applied Nanoscience

Fig. 4  Characteristic view of carbon nanotubes obtained by the CVD method at the NNV-6,6 type “Bulat” facility: a top view, b cross section

of carbon atoms diffusion inside of the catalytic particles (it is planned to devote a separate work to this problem). It is important that the observed fractures along the length of the grown CNTs should promote their excellent adhesion to the prospective matrix material. Therefore, the substrate with the CNTs shown in Fig. 4 was fixed on the heating table, which was mounted above the copper evaporator (see Fig. 2) and a flow of evaporated chemically pure copper was directed to the substrate (with the tubes grown on it). The initial temperature of the substrate and the table was ~ 30 °C. The precipitation of copper, carried out under conditions of a technical vacuum (the residual pressure in the vacuum chamber was not better than ~ 10−2 Pa), did not significantly change the initial temperature of the table for ~ 1 min (it increased to ~ 50 °C), since there was a damper between the evaporator and the table, and the table did not heated by the evaporator during

13

Fig. 5  Typical view of CNTs with a deposited layer of copper: a ×3700, b ×50000

the heating of the melting pot to the evaporation temperature of copper. However, the temperature of the substrate surface with tubes on it could notably increase during the deposition of the evaporated material flow (possibly up to ~ 100 °C), due to the low thermal conductivity of the silicon substrate and, therefore, its weak cooling from the table. The type of tubes with a layer of copper deposited on them is shown in Fig. 5. It can be seen that, due to the uniform distribution of copper atoms over the CNT surface, the resulting coating is uniform throughout the entire length of the nanotubes. In places where nanotubes are bending, there is no noticeable thickening of the copper coating. The thickening could be caused by a reduced surface energy due to structural defects, in which condensed copper atoms should predominantly accumulate (Robinson et al. 2006). Perhaps, this heterogeneity of the CNT structure affects only the initial moments of the copper coating formation,

Applied Nanoscience

i.e., at small thicknesses. In this experiment, the thickness of the copper layer on CNTs exceeds their mean diameter by an average of three times. Hence, it can be concluded that the length of copper atoms diffusion over the CNTs surface greatly exceeds the circumference of the cross section of each tube. Therefore, the copper coating is very uniform both on the length and on the diameter of individual tubes. From the analysis of the cross section of the coating on the substrates (Fig. 6), one can deduce that the copper layer thickness on the substrate surface, i.e., between individual CNTs, is very close to the thickness of the copper coating on the tubes. The copper coating contacting the substrate surface (Fig. 6a) is continuous, without breaks between the tubes. It can also be seen that the copper layer on the substrate has the same thickness both directly at the root of each individual CNT, and at a distance from the growth site of CNT. In our opinion, this is a non-trivial result, because one could expect some thickening of the copper coating on the substrate near the root of each CNT. After all, the surface temperature of the substrate (probably close to ~ 100 °C) and each individual tube should be markedly different, since the substrate is pressed to a relatively cold table (not more than 50 °C), and the deposited copper atoms bring energy to the tube (and substrate) equal to the atoms evaporation heat (according to different data ~ 4800 kJ/kg). Under these conditions, the tube must be hotter than the substrate, because it is cooled almost entirely by radiation from its heated surface. In this case, atoms are freely diffusing along the surface of the hot tube, when reaching the cold surface of the substrate, must be effectively intercepted by it. Let us note one more important fact. As it can be seen from the data in Fig. 6, each tube is not only surrounded by a layer of copper, but it is “immersed” in copper by the thickness of the copper film deposited on the substrate. This should help increase the adhesion of CNTs to the substrate and the applied layer of the metal matrix. The strengthening of the cohesive bond matrix-nanotube effect can be achieved by creating carbon atoms substructure (relief) on the surface of nanotubes. Such a substructure was obtained by adding the working gas plasma component during the CNTs’ synthesis. Active components of the plasma are the dissociation, excitation, and ionization products of ­ 2H*, ­C2H2+, C, molecules and atoms of C ­ 2H2 precursor: C * + 2+ * + ­C , ­C , ­C , CH, C ­ H , and C ­ H . Carbon and mixed carbonpolymer substructures were formed directly on the surface of CNTs from these components. One of such substructure is shown in Fig. 7. The filling of a space between branched CNTs by chemically non-interacting metal matrix will undoubtedly make it possible to create composite coatings with high physicomechanical and tribological properties.

Fig. 6  Cross section of the CNT–copper composite: a view of the composite torn from the Si/SiO2 substrate; b composite on TiN film deposited on silicon; c enlarged image of the composite on TiN film

13



Applied Nanoscience

them with various metals to form composite coatings with specified properties. Authors’ contributions  All authors contributed to the ideas behind the project. The idea of the thermal spraying of copper was developed by NYS. All authors have contributed to the writing of the text. All experiments were performed by NYS in assistance with AIK. All authors read and approved the final manuscript. Funding  This study was supported by the National Academy of Sciences of Ukraine (project ID RK No 0113U002667 from 2013).

Compliance with ethical standards  Conflict of interest All authors have approved the manuscript and agree with submission to Applied Nanoscience Journal. We have no conflicts of interest to disclose. The authors declare that they have no competing interests.

Fig. 7  Surface of CNTs with a carbon substructure on them. Si/SiO2 substrate, Fe catalyst, ~ 0.05% of the plasma component was added to the synthesis zone

Conclusions The performed experiments on formation of a copper barrier layer on the CNT surface demonstrated the possibility of creating composite coatings with metal filling between the grown CNTs on the substrate, without chemical interaction between CNT and metal. Due to the created diffusion barrier of copper, the alternative of metals and alloys as a matrix of composite coatings is widened, which makes it possible to vary the properties of such coatings in a wide range. The introduction of a regulated plasma component of the working gas made it possible to create carbon substructures on the surface of CNTs during the synthesis. This helps to increase the mechanical connection the mechanical between the matrix and CNTs in the composite. CNTs grown by the CVD method, in principle, can be uniformly distributed both over the substrate surface and over the volume of material filling the space between the CNTs. Such a composite reinforcing CNTs additive should harmoniously distribute the stresses that can arise between the CNTs and the matrix material during various mechanical loads during the composite exploitation. The modernization of the ion-plasma sputtering facility made it possible to combine in a single technological cycle operations on the creation of catalytic centers on substrates, growing of CNTs on them, as well as diffusion barriers spraying on their surfaces, and then filling the space between

13

Availability of data and materials section  The data sets supporting the conclusions of this article are available in Ukrainian Patents Database http://uapat​ents.com/8-10386​9-sposi​b-otrim​annya​-kompo​zicij​jnogo​ -pokri​ttya-z-nanos​trukt​urnim​-vugle​cevim​-zmicn​yuvac​hem.html.

References Bakshi SR, Lahiri D, Agarwal A (2010) Carbon nanotube reinforced metal matrix composites—a review. Int Mater Rev 55(1):41–64. https​://doi.org/10.1179/09506​6009X​12572​53017​0543 Gabovich MD (1972) Fizika i tekhnika plazmennyh istochnikov ionov. Atomizdat, Moskva (in Russian) Guiderdoni Ch, Estourne`s C, Peigney A, Weibel A, Turq V, Laurent Ch (2011) The preparation of double-walled carbon nanotube/ Cu composites by spark plasma sintering, and their hardness and friction properties. Carbon 49(13):4535–4543. https​://doi. org/10.1016/j.carbo​n.2011.06.063 Guiderdoni C, Pavlenko E, Turq V, Weibel A, Puech P, Estourne`s C et al (2013) The preparation of carbon nanotube (CNT)/copper composites and the effect of the number of CNTwalls on their hardness, friction and wear properties. Carbon 58:185–197. https​ ://doi.org/10.1016/j.carbo​n.2013.02.049 Lim B, Kim C, Kim B, Shim U, Oh S, Sung B et  al (2006) The effects of interfacial bonding on mechanical properties of single-walled carbon nanotube reinforced copper matrix nanocomposites. Nanotechnology 17(23):5759–5764. https​://doi. org/10.1088/0957-4484/17/23/008 Moghadam AD, Omrani E, Menezes PL, Rohatgi PK (2015) Mechanical and tribological properties of self-lubricating metal matrix nanocomposites reinforced by carbon nanotubes (CNTs) and graphene—a review. Compos B Eng 77:402–420. https​://doi. org/10.1016/j.compo​sites​b.2015.03.014 Mohamed NM, Tan YC, Kadir AK (2006) Synthesis of narrow diameter multiwalled carbon nanotubes (MWNTs) by catalytic chemical vapor deposition for mechanical reinforcement applications. Nanotech 1:125–8. http://www.nsti.org/publi​catio​ns/Nanot​ ech/2006/pdf/540.pdf Panarin VY, Svavil′nyi NY, Khominych AI (2012) Device for vacuum synthesis of carbon nanostructures. Patent of Ukraine No98909. Published: 25 Jun 2012. http://uapat​e nts.com/8-98909​- prist​

Applied Nanoscience rijj-dlya-vakuu​mnogo​-sinte​zu-vugle​cevik​h-nanos​trukt​ur.html (in Ukrainian) Panarin VY, Svavil′nyi NY, Khominich AI (2017a) Nanocomposite coatings with CNTs strengthening component. V International research and practice conference “Nanotechnology and nanomaterials” NANO-2017, 23–26 Aug 2017. Yuriy Fedkovych Chernivtsi National University, Chernivtsi, Ukraine. http://www.iop. kiev.ua/~nano2​017/files​/abstr​acts/Khomi​nich.pdf Panarin VY, Svavil′nyi NY, Khominich AI (2017b) The behavior dynamics of a droplet phase in vacuum arc discharge plasma in chemically active gas nitrogen. J Mater Res Technol 6(3):284– 288. https​://doi.org/10.1016/j.jmrt.2017.04.004 Robinson JA, Snow ES, Bǎdescu ŞC, Reinecke TL, Perkins FK (2006) Role of defects in single-walled carbon nanotube chemical sensors. Nano Lett 6(8):1747–1751. https​://doi.org/10.1021/nl061​ 2289 Sereda IN, Tseluiko AF (2015) Probe methods for plasma diagnostics (2015) Kharkov: HNU imeni VN Karazina (in Russian) http:// karazi​ nbook​ .univer​ .kharko​ v.ua/sites/​ defaul​ t/files/​ book_texts/​ sered​ a7.pdf Shah KA, Tali BA (2016) Synthesis of carbon nanotubes by catalytic chemical vapour deposition: a review on carbon sources, catalysts

and substrates. Mater Sci Semicond Process 41:67–82. https:​ //doi. org/10.1016/j.mssp.2015.08.013 Svavil′nyi NY (2016) Dynamics of behaviour of the droplet phase in the plasma flows formed in discharge gaps of vacuum—arc discharges with working argon gas. Metallofiz Noveishie Tekhnol 38(2):247–265. https​://doi.org/10.15407​/mfint​.38.02.0247 (in Russian) Svavil′nyi NY, Panarin VY, Khominych AI (2013) Method for obtaining of the composite coating with a nanostructured carbon reinforcing agent. Patent of Ukraine No103869. Published: 25 Nov 2013. http://uapat​ents.com/8-10386​9-sposi​b-otrim​annya​-kompo​ zicij​jnogo​-pokri​ttya-z-nanos​trukt​urnim​-vugle​cevim​-zmicn​yuvac​ hem.html (in Ukrainian) Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

13