Facile synthesis of novel octopus-like carbon ...

Diamond & Related Materials 74 (2017) 145–153

Contents lists available at ScienceDirect

Diamond & Related Materials journal homepage: www.elsevier.com/locate/diamond

Facile synthesis of novel octopus-like carbon nanostructures by chemical vapor deposition Muhammad Asif a,b,⁎, Muhammad Rashad c, Fride Vullum-Bruer d, Jiayan Li a, Xiaogang You a, Abdul Sammed Khan b, Chenghao Deng b, Lujun Pan b,⁎⁎, Yi Tan a,⁎⁎ a

School of Materials Science and Engineering, Dalian University of Technology, Dalian 116024, PR China School of Physics and Optoelectronic Technology, Dalian University of Technology, Dalian 116024, PR China Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116024, PR China d Department of Materials Science and Engineering, NTNU, Norwegian University of Science and Technology, 7491 Trondheim, Norway b c

a r t i c l e

i n f o

Article history: Received 1 February 2017 Received in revised form 25 February 2017 Accepted 7 March 2017 Available online 08 March 2017 Keywords: Chemical vapor deposition Carbon nanofibers Carbon nanostructures Carbon materials Thin film Nano devices

a b s t r a c t Growth of novel carbon nanomaterials has been of great interest for researchers due to their potential applications in nano-devices. In the current work, we reported radial growth of novel octopus-like carbon nanostructures (OCNS) by chemical vapor deposition (CVD), using methane as carbon precursor gas, and Cu film sputtered on Si/SiO2 substrate as the catalyst. Annealing at high temperature transforms Cu thin film to catalytic copper nanoparticles (CuNPs), which on exposure to methane results in the radial growth of carbon nanofibers (CNFs) departing from the central CuNPs. The size of OCNS and morphology vary as a function of Cu film thickness, precursor gas concentration, and growth time. High methane concentration boosts up growth kinetics, resulting in long carbon fibers in addition to OCNS, with fiber length varying from a few hundred nanometers to several hundred microns. Effect of substrate on the morphology of carbon nanostructures is also studied using Cu film sputtered on silicon, quartz, and Si/SiO2 substrates. The branch like morphology of OCNS exhibits large surface/contact area for their applications in electronic and electrochemical devices. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Integrated electronic circuits miniaturization requires their corresponding constituent components to catch up on similar development, e.g. large surface area carrying lightweight contacts. To cope with the increasing demands, lightweight conducting carbon nanostructures with large surface area has attracted tremendous research interest [1,2]. The physical and chemical properties of carbon nanomaterials can be controlled by tailoring their morphologies, and play a critical role in determining their respective applications. Thus growth of carbon nanostructures with different morphologies has always been of great interest for the researchers, e.g. flower-like carbon materials [3], hollow carbon spheres [4], Y-junction carbon nanomaterials [5], and helical carbon nanotubes [6]. These nanomaterials have been used for various applications such as, microwave adsorption, logic devices, supercapacitors, and lithium ion batteries [4,7–9]. Among the carbon nanostructures, the octopus-like carbon nanostructures (OCNS) which consist of a number of

⁎ Correspondence to: M. Asif, School of Materials Science and Engineering, Dalian University of Technology, Dalian 116024, PR China. ⁎⁎ Corresponding authors. E-mail addresses: [email protected] (M. Asif), [email protected] (L. Pan), [email protected] (Y. Tan).

http://dx.doi.org/10.1016/j.diamond.2017.03.004 0925-9635/© 2017 Elsevier B.V. All rights reserved.

carbon nanofiber legs radiating from the central catalyst particle, have been considered as ideal candidates due to their low-density and large surface area [3,10–13]. Growth of carbon nanostructures with different morphologies has been carried out using template assisted solvothermal and chemical vapor deposition (CVD) methods. Wang et al. [9] reported growth of carbon nanostructures through a carbonization process on pitch coated ZnO precipitates as growth templates, followed by hydrochloric acid assisted removal of the ZnO template. Xiao et al. [3] synthesized flower like carbon nanomaterials through a hydrothermal method, using Mg(CH3COOH)2·4H2O and polyethylene glycol as reactants in the ethanol as solvent. The solution was kept at 600 °C in the stainless steel autoclave for 24 h. However, CVD techniques have proved more effective for the growth of carbon nanomaterials, such as carbon nanotubes and carbon nanofibers [5–8]. Nevertheless, most of the flower-like nanomaterials reported so far, exhibit spherical morphology. Several inorganic nanomaterials (such as ZnO [14], CuO [15], and In2O3 [16]) have been synthesized through solvothermal methods. Noorduin et al. have recently reported growth of hierarchical flower-like BaCO3-SiO2 microarchitectures with diverse morphologies by adjusting different growth parameters of the reaction solution, i.e. CO2, pH, and temperature [17]. Moreover, there has been a modest number of reports where flower-like carbon nanomaterials with spherical shape have

146

M. Asif et al. / Diamond & Related Materials 74 (2017) 145–153

been synthesized using solvothermal methods as the common choice [3,18–20]. In general, synthesis of flower-like carbon nanomaterials have been restricted to solution based procedures. Therefore, growth of novel carbon nanostructures and alternative dry synthesis techniques should be explored. Varshney et al. used a dry process for the first time to fabricate vertically aligned carbon nanotubes flower-like clusters coated with ultra-nanocrystalline diamond by hot filament CVD [21]. Generally, synthesis of octopus-like carbon nanostructures occurs at temperature higher than 950 °C, when using a thermal CVD method and methane as precursor gas [22–24]. The carbon nanomaterials have been synthesized using nanoparticles which catalyze the nucleation of carbon nanostructures. The morphologies of the products can be controlled by varying the gaseous mixtures carrying carbon feedstock elements, type of catalysts, and their proportions. The gases and catalyst used to grow these carbon nanomaterials mostly comprise of numerous combinations of different materials. The octopus-like morphology of the product is observed to be significantly dependent on the amount of copper in the nickel catalyst. A number of studies have been carried out to analyze the role of hybrid catalysts using various copper contents (ranging from 3% to 80% Cu) in the synthesis of octopus-like nanomaterials. However, the exact growth mechanism and the role of the catalyst still need to be explored [24–27]. We have synthesized octopus-like carbon nanostructures (OCNS) using 100% copper catalyst, which is advantageous over alloy catalysts containing metals which are magnetic in nature, such as nickel and iron [28,29]. In the current work we have synthesized octopus-like carbon nanostructures (OCNS) using a pure copper catalyst via a conventional thermal CVD process. To better understand the effect of the substrate, the growth process was carried out on three different substrates i.e. silicon wafer with a SiO2 layer, quartz plate, and silicon wafer. Effect of copper film thickness, and methane gas concentration was also studied in detail. The growth products were analyzed by FESEM, TEM, Raman and XRD characterization tools. 2. Experiments and methods 2.1. Synthesis of carbon nanostructures Atmospheric pressure chemical vapor deposition was used to synthesize different types of carbon nanostructures on Cu film. The copper film was sputtered on a silicon wafer with a 300 nm thick SiO2 layer via vacuum magnetron sputtering (JCP-200, BTSC563). The Si/SiO2/Cu substrate was loaded in the CVD quartz chamber and was heated to 1035 °C at 33 °C per min heating rate under flowing Ar/H2 (1000/100 sccm). The sample was first annealed at 1035 °C for 1 h, by which the Cu film transformed into spherical shaped Cu nanoparticles (CuNPs). Then, the quartz tube was cooled down to 1010 °C and the sample was exposed to the carbon precursor containing gaseous mixture (Ar/H2/CH4) for a specific time, which resulted in growth of carbon nanostructures with varying morphology and size. The tube was cooled down to room temperature under flow of Ar/H2 gases (1000/10 sccm). Effects of methane concentration were studied by changing methane gas flow rate, while keeping all other parameters constant. Furthermore, the Cu film thickness was varied to study the effect on morphology and size of the carbon nanostructures. Moreover, the effect of the substrates on growth of carbon nanostructures was studied by comparing growth of carbon nanostructures on different substrates, i.e. Si/SiO2, Si wafer, and quartz plate. The detailed growth conditions for the different experiments are illustrated in Table 1. 2.2. Characterizations The morphological characterizations of the synthesized carbon nanostructures were performed by field emission scanning electron microscopy (FE-SEM, NOVA NanoSEM450) and transmission electron

Table 1 Experimental conditions for growth of different carbon nanostructures. Sample name

Substrate

Cu film thickness (nm)

Gases flow rates Ar:H2:CH4 (sccm)

Growth time (min)

SO55M SO55L SO55H SO95M SO120M Q55M S15M S120M

Si/SiO2 Si/SiO2 Si/SiO2 Si/SiO2 Si/SiO2 Quartz Si Si

55 55 55 95 120 55 15 120

1000:100:10 1000:10:02 1000:10:32 1000:100:10 1000:100:10 1000:100:10 1000:100:10 1000:100:10

25 15 15 15 15 15 15 15

SO: silicon dioxide (Si/SiO2), S: silicon wafer, Q: Quartz, H/M/L: high/moderate/low flow rate (CH4).

microscopy (TEM; FEI, Tecnai G2 F30S-Twin). The elemental analysis was performed by energy dispersive spectroscopy (EDS) connected with SEM. The graphitization of carbon nanostructures was investigated by Raman spectroscopy (Renishawin Via plus, operated at 632.8 nm, He–Ne laser, 4.0 mW laser power, and 50 × objective lens). Crystallographic behavior was analyzed by X-Ray Diffraction (XRD, Cu Kα radiation, PANalytical B.V. Empyrean) with 2θ ranging from 10–90°. The thickness of Cu film was measured with the ZYGO 5022 profilometer. 3. Results and discussion 3.1. Growth of carbon nanostructures and influence of methane gas concentration The uncommon carbon nanostructures were synthesized through a CVD process, using copper film deposited on a silicon wafer with an oxide layer (SiO2) on its surface. This process is as an excellent way to synthesize non-magnetic carbon nanostructures, and thus eliminates the need for a purification step. The Cu film was transformed into CuNPs through annealing at high temperature for 1 h, followed by growth step. The CuNPs are very reactive towards methane gas at elevated temperature, thus when exposed to the precursor gas, methane dissociates into CHx radicals in the presence of the H2 gas. These CHx radicals are adsorbed on the CuNPs active surface sites and form nucleation points, which ultimately result in a multidirectional or branched growth with several carbon fiber/filaments from the central catalyst particle. Fig. 1(a, b) depicts FESEM images of very interesting octopuslike carbon nanostructures (OCNS) with multiple carbon fiber legs grown from single catalyst particles, when using SO55M growth conditions (as shown in Table 1). The catalyst particle is lifted above the surface of the Si/SiO2 substrate by the carbon fiber legs. It can be observed that the OCNS are grown in large quantity and the size of these OCNS is around 5 μm and in some cases even larger, with leg fibers curling around each other around the tips. The average number of leg fibers per catalyst nanoparticle are more than 10 and their number vary from particle to particle. It is also noticeable that some OCNS produce very thin but long fiber legs, as illustrated in Fig. 1b. Careful observation revealed that under these OCNS there exist some irregular shapes and immature carbon nanostructures. Furthermore, very large quantities of long carbon fibers are grown near the edge area of the substrate. However, OCNS can be observed while moving away from the edge of the substrate. Effect of methane gas concentration was studied by carrying out two additional growth experiments at low and high methane gas flow rates (SO55L and SO55H growth conditions). Fig. 1(c, d) depicts FESEM micrographs at different magnifications for flower like OCNS synthesized using low methane gas flow rate. The synthesized flower like OCNS show different morphology compared to OCNS grown using SO55M growth conditions. The flower like OCNS are much smaller in size (less than 1 μm) and fibers originating from central CuNPs exhibit thick and flat strip like shape. The top surface of the CuNPs is fully


147

Fig. 1. Growth of OCNS and effect of methane concentration: low and high magnification FESEM micrographs for OCNS synthesized on 55 nm thick Cu film sputtered on a Si/SiO2 substrate at moderate i.e. SO55M (a, b), and low methane concentrations i.e. SO55L (c, d).

covered by growth of flat carbon nanofibers, which is contrary to the naked top surface of the CuNPs using SO55M growth conditions. This could be attributed to the low methane flow rate leading to slow growth kinetics, consequently resulting in growth of thick strip like fibers covering the top surface of the CuNPs. Furthermore, FESEM analysis revealed that the density of carbon products was quite low compared to the SO55M growth conditions.

Effect of higher methane concentration on the growth of carbon products was investigated using SO55H growth condition (as shown in Table 1). High methane flow rate resulted in fast growth kinetics leading to high density of linear and branched carbon nanofibers, as depicted in Fig. 2. Fig. 2(a–h) demonstrate morphology of the growth products at different positions, i.e. from edge of the substrate towards the center of the substrate. It is obvious from the FESEM micrographs

Fig. 2. Growth of carbon nanostructures and effect of methane concentration: low and high magnification FESEM micrographs for carbon nanostructures synthesized on 55 nm thick Cu film sputtered on a Si/SiO2 substrate at high methane concentration i.e. SO55H growth condition, starting from outer edge to center of substrate (a-h).

148


that the growth density and morphology vary depending on the position on the substrate. Fig. 2a shows high density of carbon nanofibers growing in different directions. When moving a little away from the edge towards the center of the sample, carbon nanofiber growth tends to be oriented in a particular direction along the direction of the gas flow (as shown in Fig. 2b). However, it is quite interesting to note that the fibers are curling around each other to form small bundles. The fibers' density reduces when moving even further towards the center of the substrate (as shown in Fig. 2c, d). Fig. 2d shows a magnified FESEM micrograph, illustrating growth of OCNS under long carbon nanofibers. Here, the growth direction of the carbon nanofibers varies, which indicate that flow of the gaseous mixture is not as smooth as near the sample edges. The main reason for high density and irregular shape of the carbon nanostructures around the border areas of the substrate is due to temperature gradient which subsequently results in carbon diffusion gradient in CuNPs. The temperature gradient and fluctuation in global gases flow rate around the edge areas lead to fast growth kinetics which results in long irregular shape carbon fibers. Fig. 2e illustrates growth of OCNS around a hole, spreading in a radial direction. Several similar holes are observed on this sample, which are surrounded by OCNS. These OCNS consists of several carbon fiber legs, lifting CuNPs from the surface of the Si/SiO2 substrate. However, some OCNS also possess a carbon leg floating in vertical direction, as illustrated in the inset image. Moreover, careful observation revealed the existence of immature carbon products at the surface of Si/SiO2 substrate under OCNS. These patterns have more to do with how the structures nucleate and grow, which is determined by gas flow rate and temperature of the substrate. Under this growth condition, much higher concentration of precursor gas seems to highly affect the morphology and formation of nanostructures all across the substrate. There also exists a large variation in morphology of carbon nanostructures, depending on where we are on the substrate. If we look close to the hole in Fig. 2e, we can also observe that the morphology of the structures is quite different than further away from the hole. This might be attributed to the variation in substrate temperature and precursor gas concentration. While moving further towards the center of the sample, growth of quite different carbon structures can be observed, as illustrated from the FESEM micrograph in Fig. 2f. One type of structure consists of carbon fibers originating from the Cu particles spread on the surface of the underlying substrate. These carbon fibers divide into branches, and twist around each other near the fiber tips. Moreover, it is quite interesting to note that in this case CuNPs stick to the substrate and the fiber legs grow in a vertical direction. This is contrary to the growth behavior of OCNS, where the fiber legs lift the CuNPs from the substrate, while the fiber legs stick to the surface of the substrate. There also exists immature octopus-like carbon products with thicker leg fibers, grown on the Cu catalyst particles. When moving further towards the central part of the sample, two different kinds of OCNS were observed, as depicted from Fig. 2g. The first type of OCNS possesses very thick strip like fiber legs, making their appearance more like a flower, while the second type of OCNS are very beautiful flower like shape with fiber legs lifting the central catalyst CuNPs. At the central region of the sample, growth of immature carbon fibers originating from the CuNPs can be observed, as illustrated in the Fig. 2h. These CuNPs are lifted from the surface of the substrate by single or branched carbon fiber legs. The conversion of a single carbon fiber leg into multiple fiber legs can be attributed to the secondary catalyst CuNPs inside the carbon fiber legs segregated from the primary catalyst CuNPs, subsequently initiating growth of secondary fiber legs, as discussed in the later part. The insets in Fig. 2h shows the morphology of CuNPs after the CVD process, which appear as diamond like nanoparticle. The elemental composition of the synthesized products was analyzed by EDS mapping, as illustrated in Fig. S1, in the Supporting information. To further explore the morphology and growth mechanism of the different as synthesized carbon structures containing OCNS, transmission electron microscopy (TEM) characterization was performed, as illustrated in Fig. 3, and Fig. S2, in the Supporting

information. Fig. 3a shows a TEM micrograph of the branched carbon fiber and some other carbon products. It is quite obvious from the TEM image that the single carbon fiber is divided into more than eight fibers. This can be attributed to growth from a secondary catalyst CuNPs, which might have been separated from the primary CuNPs and existed in the primary carbon fiber. When that CuNPs are exposed to methane gas, secondary growth can take place, which lead to a branched carbon fiber network. The existence of the secondary catalyst CuNPs inside the carbon fibers can be observed in the background Yshaped carbon fiber in Fig. 3b. However, on the top of the Y-shaped carbon fiber, there exist an OCNS with a CuNP at the center and carbon fiber legs growing away from the catalyst particle. Some very small diameter carbon fiber/nanotubes can also be observed in the TEM images. It is quite interesting to observe the human humerus bone like carbon fiber/tube filled with spindle like Cu metal (as shown in the Fig. 3c). Furthermore, the main CuNP is also covered by carbon layers with two pointed edges, one of which is in the direction of carbon fiber. Moreover, the inset in Fig. 3c is a SAED pattern taken from the CuNP responsible for the growth of OCNS. Fig. 3d depicts growth of OCNS at an early stage, with fiber legs in a downward direction. It is obvious from the TEM image that small secondary CuNPs are segregated into the carbon fibers, as indicated by arrows. Thus it can be concluded that small secondary CuNPs might have dissociated from the primary CuNPs and moved down with the growth of the carbon fibers. Fig. 3(e, f) depict high magnification TEM images of a graphitic carbon layer on the surface of a CuNP (taken from selected area in Fig. 3d). The thickness of the graphitic layer is around 4 to 5 nm, and thus can be used for the protection of metal nanoparticles, as corrosion resistant coating. Moreover, the secondary CuNPs inside the carbon fibers (indicated by arrows in Fig. 3d) were analyzed by HRTEM. Fig. 3g depicts a HRTEM image of a secondary CuNP, which is around 20 nm in diameter. The inset shows Cu lattice fringes with an interlayer spacing of 0.21 nm, which corresponds to the (111) plane of Cu. 3.2. Effect of Cu film thickness on growth of the carbon nanostructures Effect of Cu film thickness on growth of OCNS was also examined by carrying growth under moderate methane flow rate using thicker Cu films i.e. 95 nm and 120 nm sputtered on the Si/SiO2 substrate, (SO95M and SO120M growth conditions, respectively), as illustrated in Fig. 4. Moreover, H2 gas flow rate was kept at 100 sccm during heating/annealing and growth stage, and all other conditions were kept the same as that of SO55M. For the SO95M growth condition, growth of small size OCNS with relatively short carbon fiber legs originating from the central catalyst CuNPs can be observed, as illustrated in Fig. 4(a, b). The density of the carbon structures is significantly lower compared to that observed for the SO55M growth condition. Fig. 4b depicts high magnification FESEM micrographs of two different size OCNS, with carbon fiber legs lifting the catalyst particle from the surface of the substrate. The tips of the carbon fiber legs appear to stick to the surface. For the SO95M growth condition, average sizes of the CuNPs range from 250 to 600 nm and is larger than that observed for SO55M. Additionally, Fig. 4(c, d) depicts the existence of flower like OCNS with thick carbon fiber legs, different from those observed in other areas of the sample. The morphology of these OCNS predicts that the CuNP catalysts might have been poisoned due to interaction of poisoned gases produced during the CVD reaction process or due to partial contamination of catalyst NPs from amorphous carbon film, happen in case of carbon nanotubes/nanofibers growth [30,31]. When these poisoned catalyst CuNPs get exposed to methane gas, the thick fiber OCNS are synthesized. The carbon fiber legs are strip like in shape and these fiber legs are stacked with each other. Moreover, it is obvious from the FESEM micrographs that the growth of some carbon products was initiated from the Cu particles/residues on the substrate. The elemental composition of the OCNS shown in Fig. 4a was analyzed by EDS mapping, as illustrated in Fig. S3, in the Supporting information.


149

Fig. 3. TEM analysis for carbon nanostructures with different morphologies: branched carbon nanostructures (a), OCNS (b), humerus bone like carbon nanofiber encapsulating the Cu metal (c), immature grown OCNS (d), graphitic carbon coating on CuNP from the selected area (e, f), and secondary CuNP (g). The lattice fringes with a spacing of 0.21 nm in the inset of (g) correspond to (111) plane of Cu.

The FESEM micrographs and EDS mapping for OCNS grown under SO120M growth conditions are illustrated in Fig. 4(e–j). Thickness of the Cu film was increased to 120 nm, and H2 gas flow rate was kept at

100 sccm during heating, annealing and reaction time. The FESEM micrographs demonstrate growth of OCNS similar in shape to that observed in Fig. 4(a, b), but the fiber legs are significantly shorter (less

Fig. 4. Effect of copper film thickness on the growth of OCNS: low and high magnification FESEM micrographs and EDS mapping for the OCNS synthesized on 95 nm (a–d) and 120 nm (e–j) thick Cu film sputtered at Si/SiO2 substrates for 15 min growth under moderate methane flow rate, i.e. SO95M and SO120M growth conditions, respectively. The scale bars in the insets of (f) are 1 μm.

150


than 1 μm) in length. Moreover, the size of CuNPs is almost double that of the SO95M sample, and the sizes vary from around 600 nm to 1.2 μm. It is quite interesting that some of the CuNPs appears to possess a small cavity on their top surface. This indicates that a part of the CuNPs might have burst during the growth process. Moreover, a graphitic carbon layer does exist on the surface of CuNPs, in addition to growth of carbon fiber legs. The elemental composition of the OCNS is analyzed by performing EDS mapping, which demonstrates that fiber legs are carbonaceous in nature originating from the CuNPs which itself is coated with a graphitic carbon film, as illustrated in Fig. 4(g–j). 3.3. Effect of substrates on growth of carbon nanostructures To better understand the growth mechanism and investigate the effect of substrate on the growth of carbon nanostructures, two additional substrates (quartz plate and silicon wafer) were used. A copper film with different thickness was sputtered on these substrates to form copper catalyst nanoparticles. Fig. 5 depicts FESEM micrographs of the formed carbon nanostructures with different morphologies, synthesized on the quartz substrate under Q55M growth conditions. For FESEM characterization, samples were sputtered with 4–5 nm thick gold film to make the surface of the quartz substrate electrically conductive. It is quite obvious from the different magnification FESEM micrographs shown in Fig. 5(a, b) that the OCNS synthesized on the quartz substrate appear to be similar to those grown on the Si/SiO2 substrate. It is evident from the FESEM micrographs that the size of the CuNPs varies from 250 nm to 500 nm, and the length of the fiber legs is around 500 nm. Moreover, the number of fiber legs increases with size of CuNPs. Fig. 5(c, d) shows growth of flower like carbon nanostructures which are slightly different from the previously observed OCNS. The small size flower like OCNS exhibit growth of thick and short fiber legs, whereas the relatively large size flower like carbon nanostructures exhibit large number of bud like carbon legs grown on their surface. The growth of such kind of short fiber legs or bud like structures might be attributed to a large number of active sites (surface defects) on the CuNPs which leads to a larger number of nucleation sites on each particle. In addition to these two types of flower like carbon structures, there also exists long carbon fibers/tubes scattered around. These thin and long carbon fibers/tubes might have been synthesized from very small CuNPs.

Fig. 5. Effect of substrate on the growth of carbon nanostructures: (a–d) low and high magnification FESEM micrographs of the flower like OCNS synthesized on 55 nm thick Cu film sputtered on quartz substrate for 15 min growth under moderate methane flow rate, i.e. Q55M growth condition.

Growth of carbon nanostructures on the Cu film sputtered on a Si wafer substrate under S15M, and S120M growth conditions (Table 1), is quite different from that of the Si/SiO2 and quarts substrates, as illustrated in Fig. 6. The absence of barrier layer e.g. SiO2 may result in increased reactivity of Si and Cu at elevated temperature and thus may lead to formation of copper-silicides (i.e., Cu3Si) [32,33]. The dewetting process [34] may also be affected in the absence of SiO2 layer, leading to the formation of bud like Cu3Si catalyst NPs. When exposed to the CH4 gas, these irregular shape catalyst NPs result in growth of branch like fibers through a tip growth mechanism. The bud shappedCu3Si catalyst NPs itself might undergo catalyst splitting phenomenon therefore could be responsible for growth of branch like carbon fibers network [35]. Fig. 6(a, b) illustrates different magnification SEM micrographs of the branched carbon fiber network. The careful observation reveals that the Si surface is etched during the formation of Cu3Si NPs catalyst, while annealing at 1035 °C. The size of a single catalyst NP is around 200–300 nm in diameter. Fig. 6(c, d), depicts that the higher Cu film thickness produces large size Cu3Si catalyst NPs. However, it is quite interesting to observe that diameter of the carbon fibers synthesized on Cu films of different thicknesses does not change much. The high magnification SEM images of large size catalyst NP revealed that it is actually aggregation of more bud like NPs which are responsible for growth of higher branch order fibers network. Therefore, we may conclude that diameter of carbon fibers is independent of Cu film thickness. The Raman spectroscopic analysis was performed to further confirm the carbonaceous nature of the carbon structures synthesized under different growth conditions, as illustrated in Fig. 7. The Raman spectra for the SO55M sample demonstrate distinct D and G peaks and a low intensity 2D peak (as show in Fig. 7a). The top and middle spectra contain 2D peaks, signifying that these signals might have reflected from the graphitic carbon, either from carbon tubes or a graphitic carbon layer on the surface of the CuNPs. However, the bottom spectra does not show a 2D peak, indicating that the Raman signal might be generated from the carbon fibers. The intensity ratio ID/IG for the top spectra is greater than 1, whereas for the other two spectra its value is less than 1. Fig. 7b illustrates Raman spectra for carbon structures synthesized using SO55H growth conditions, where all three spectra contain distinct D, G, and 2D bands. The I2D/IG ratio for the top two spectra indicate graphitic like carbon which could be from carbon nanotubes, whereas the

Fig. 6. Effect of substrate on the growth of carbon nanostructures: low and high magnification FESEM micrographs for carbon nanostructures synthesized on 15 nm (a, b) and 120 nm (c, d) thick Cu film sputtered on silicon wafer substrate for 15 min growth under moderate methane flow rate, i.e. S15M and S120M growth conditions, respectively.


151

Fig. 7. Raman spectra for the carbon nanostructures synthesized under different growth conditions: SO55M (a), SO55H (b), and S120M growth conditions (c), and corresponding ID/IG ratio (d).

bottom spectra shows double layer graphene like structure which could be either grown on the surface of the Si/SiO2/Cu substrate or on the CuNPs. Fig. 7c demonstrate full range Raman spectra for carbon structures synthesized under S120M growth conditions. The G band for the red colored spectra is broad and an additional small peak can be observed near 1500 cm−1, which might be attributed to the formation of silicon carbide in the synthesized carbon products, which can also be confirmed by X-ray diffraction peaks, as illustrated in Fig. 8. This is further confirmed by a peak near 782 cm−1 corresponding to SiC bond vibrations. However, in the blue spectra this peak is small in intensity and in the green spectra it disappears. Furthermore, all spectra contains two distinct peaks around 299 cm−1 and 623 cm−1 indicating formation of CuO/Cu2O. Fig. 7d illustrates the statistical data for D to G band intensity ratio (ID/IG) for the carbon nanostructures synthesized under three different conditions. The D to G band Intensity ratio (ID/IG) for SO55M sample vary from 0.75 to 1.25, and the average ID/IG ratio is around 0.79. Moreover, with increasing methane gas flow rate and reaction time, the ID/IG ratio increases to 1.12, which could be due to the deposition of amorphous carbon on the surface of the carbon nanostructures. In the case of carbon nanostructures synthesized under S120M growth conditions, the intensity ratio again decreases to around 0.80, which could be attributed to the higher graphitic order or reduction in amorphous carbon deposited on the surface. The XRD analysis was carried out to study the crystallinity of the carbon nanostructures. Fig. 8 shows XRD spectra for the carbon nanostructures synthesized on different substrates. As the carbon nanostructures are not uniformly distributed on the substrates, it is hard to get XRD spectra on particular OCNS. The XRD spectra show different peaks, confirming the existence of CuO, Cu2O, Cu, SiC, and Si, which is consistence with the Raman spectroscopy results. However, no obvious peak for carbon can be observed, except in the top spectra for Q55M sample, which confirms the existence of crystalline/graphitic carbon nanostructures. This spectra also contains strong peak for Cu, which indicate that this spectra is originated from a carbon tube/fiber containing metallic

Cu inside (as shown in Fig. 3c) or from a CuNP of the OCNS encapsulated by the graphitic carbon (as shown in Fig. 3d–f). Furthermore, the intensity ratio of D to G band (ID/IG) in the Raman spectra also indicate the existence of graphitic as well as amorphous nature of the carbon nanostructures, as shown in Fig. 7. The TEM images also revealed that some carbon nanostructures have poor crystallinity and others are amorphous in nature. 3.4. Growth mechanism The Cu film deposited on silicon substrate exhibits high migration resistance, low resistivity (1.63 μΩ·cm), and high melting temperature,

Fig. 8. X-ray diffraction patterns of carbon nanostructures grown under different growth conditions, i.e. SO55L, SO120M, SO55H, Q55M, and S120M.

152


thus the Cu-Si interface has attracted tremendous research interest [36]. However, annealing Cu film at elevated temperatures, Cu and Si show high reactivity and diffusion rate of Cu into Si substrate increases [33, 36]. Therefore, a dielectric layer with high thermal stability such as SiO2 is required to control diffusion of Cu into Si substrate [37–41]. Moreover, as deposited metal thin films are mostly metastable or unstable, and thus undergo dewetting process to form metal islands, while heating at elevated temperatures. This phenomenon occur through surface diffusion, well below the melting temperature of the thin film and thus lower their surface energy [34]. The copper thin film exhibits a relatively low melting point, thus may start melting at 1035 °C. Annealing a Cu thin film at elevated temperature could result in dewetting and accumulation of Cu in the form of liquid islands. As a consequence of the Kirkendall effect [22] during the catalyst formation process, these Cu islands aggregate further in the form of liquid drops, and move on the surface of the silicon dioxide substrate. At the growth temperature i.e. 1010 °C, these drops transform into faceted spherical CuNPs. To lower the energy of the system, Ostwald Ripening will occur where some of the smaller catalyst particles diffuse into the larger ones, subsequently resulting in an increase of their size [42]. The greater the size of the CuNPs, the larger the available surface area would be for growth of wide diameter leg filaments, consequently resulting in the wide range of differently sized fiber legs. When these CuNPs were exposed to the gaseous precursor mixture, catalytic decomposition of methane occurs at the catalyst surface in the presence of H2 gas. Subsequently, this might result in adsorption of carbon radicals (CHx) on the preferential step-edges leading to nucleation and crystallization into graphitic layers [43–45]. In case of the octopus growth process, the surface step edges and defects formed due to Cu segregation could be responsible for the nucleation process. This is evident from the previous results where octopus-like structures grew only in a Cu-Ni alloy system [28]. The driving force responsible for the octopus-like structures is mainly attributed to the fluctuation of growth conditions [35,46–48]. Growth of different number of carbon fiber leg filaments occur depending upon size of the catalyst NPs, with varying fiber diameter and lengths. The carbon fiber legs grown on the bottom side lift the catalyst CuNPs from the surface, whereas fiber tips stick with the substrate. The growth mechanism for octopus-like nanostructures is illustrated in Fig. 9. In case of the growth process on the bare silicon wafer, the growth products are entirely different, and appeared as branched fiber network. The reactivity of Cu film and Si substrate increases while annealing at elevated temperature, which result in the formation of Cu3Si catalyst NPs [32,33]. Moreover, in case of Cu film deposited on bare Si wafer, dewetting of Cu film may be different to that containing SiO2 as diffusion barrier, as discussed above. The dewetting principle used to form the catalyst nanoparticles proceeds differently on the different substrates, i.e., the Si/SiO2 and quarts substrates produce similar products,

as both have SiO2 on the outer surface, while Si have more different properties. During the growth process these Cu3Si NPs grow carbon fibers through a tip growth mechanism. The Cu3Si catalyst NPs may exhibit catalyst splitting during the growth process leading to branched like fiber network, as illustrated in Fig. 6. As example, Li et al. [46] reported Y-junction straight carbon nanotubes by thermal fluctuation induced catalyst splitting. Goswami et al. [35], reported catalyst splitting phenomenon leading to growth of branched like carbon nanotubes. Moreover, small size catalyst NPs constituting larger catalyst NPs posses' weak adhesion forces and thus carbon extrusion results in splitting of smaller sized catalyst NPs, which subsequently lead to growth of small diameter carbon nanofibers [47]. The growth of OCNS own an advantage over the conventional carbon nanotubes/nanofibers due to the fact that their multiple carbon legs are free of bundles and therefore present high surface area to volume ratios and higher diffusion speeds for the adsorption of larger molecules, thus are useful for numerous applications such as surface contacts in electrodes for supercapacitor and Li ion batteries. The porous nature of hollow carbon spheres reduces the lithium ions diffusion length and thus can be used as anode material for lithium ion batteries [4]. Similarly, spherical flower like carbon materials can be used as electrode material for supercapacitors due to their high surface area [9]. We can produce even greater surface area by multiple layers stacking which would be perfect as contacts for organic light emitting diodes (OLED) and organic photovoltaics (OPV) applications. In addition, these nanostructures could also be used for designing various electronic, mechanical and chemical sensor devices. For instance, Chiral morphology of the helical carbon nanotubes makes them exceptional contenders and thus has been used for microwave adsorption applications [7]. Y-junction carbon nanotubes have been employed in logic devices due to their switching and logic functional capabilities [8]. 4. Conclusions We have successfully synthesized different types of carbon nanostructures through chemical vapor deposition (CVD). The Cu film was sputtered on different substrates (Si/SiO2, quartz, and Si wafer) and was subjected through annealing at 1035 °C temperature, followed by growth of carbon nanostructures at 1010 °C temperature using methane as precursor gas. The growth on Si/SiO2 substrate produced octopus-like carbon nanostructures (OCNS). However, higher methane flow rate produced long straight carbon nanofibers (CNFs) in addition to OCNS grown at the central area of the substrate. The growth products on a quartz substrate exhibited similar morphology i.e., OCNS. However, growth products on silicon wafers exhibited a branched carbon nanofiber network, thus confirming that the SiO2 substrate plays a crucial role for the growth of OCNS.

Fig. 9. Schematic diagram illustrating growth mechanism for the OCNS.


Acknowledgement This work was supported by the National Natural Science Foundation of China (Nos. 11274055 and 61137005), and the Program for Liaoning Excellent Talents in University. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.diamond.2017.03.004. References [1] G.G. Wallace, J. Chen, D. Li, S.E. Moulton, J.M. Razal, Nanostructured carbon electrodes, J. Mater. Chem. 20 (2010) 3553–3562. [2] F.S. Boi, G. Mountjoy, M. Baxendale, Boundary layer chemical vapor synthesis of selforganized radial filled-carbon-nanotube structures, Carbon 64 (2013) 516–526. [3] Y. Xiao, Y. Liu, L. Cheng, D. Yuan, J. Zhang, Y. Gu, G. Sun, Flower-like carbon materials prepared via a simple solvothermal route, Carbon 44 (2006) 1589–1591. [4] S. Yang, X. Feng, L. Zhi, Q. Cao, J. Maier, K. Müllen, Nanographene-constructed hollow carbon spheres and their favorable electroactivity with respect to lithium storage, Adv. Mater. 22 (2010) 838–842. [5] E.-X. Ding, J. Wang, H.-Z. Geng, W.-Y. Wang, Y. Wang, Z.-C. Zhang, Z.-J. Luo, H.-J. Yang, C.-X. Zou, J. Kang, L. Pan, Y-junction carbon nanocoils: synthesis by chemical vapor deposition and formation mechanism, Sci. Rep. 5 (2015) 11281. [6] W. Wang, K. Yang, J. Gaillard, P.R. Bandaru, A.M. Rao, Rational synthesis of helically coiled carbon nanowires and nanotubes through the use of tin and indium catalysts, Adv. Mater. 20 (2008) 179–182. [7] G. Wang, Z. Gao, S. Tang, C. Chen, F. Duan, S. Zhao, S. Lin, Y. Feng, L. Zhou, Y. Qin, Microwave absorption properties of carbon nanocoils coated with highly controlled magnetic materials by atomic layer deposition, ACS Nano 6 (2012) 11009–11017. [8] P.R. Bandaru, C. Daraio, S. Jin, A.M. Rao, Novel electrical switching behaviour and logic in carbon nanotube Y-junctions, Nat. Mater. 4 (2005) 663–666. [9] Q. Wang, J. Yan, Y. Wang, T. Wei, M. Zhang, X. Jing, Z. Fan, Three-dimensional flowerlike and hierarchical porous carbon materials as high-rate performance electrodes for supercapacitors, Carbon 67 (2014) 119–127. [10] Y. Saito, Nanoparticles and filled nanocapsules, Carbon 33 (1995) 979–988. [11] L.B. Avdeeva, O.V. Goncharova, D.I. Kochubey, V.I. Zaikovskii, L.M. Plyasova, B.N. Novgorodov, S.K. Shaikhutdinov, Coprecipitated Ni-alumina and Ni Cu-alumina catalysts of methane decomposition and carbon deposition. II. Evolution of the catalysts in reaction, Appl. Catal. A Gen. 141 (1996) 117–129. [12] C. Pham-Huu, R. Vieira, B. Louis, A. Carvalho, J. Amadou, T. Dintzer, M.J. Ledoux, About the octopus-like growth mechanism of carbon nanofibers over graphite supported nickel catalyst, J. Catal. 240 (2006) 194–202. [13] Y. Li, J. Chen, Y. Ma, J. Zhao, Y. Qin, L. Chang, Formation of bamboo-like nanocarbon and evidence for the quasi-liquid state of nanosized metal particles at moderate temperatures, Chem. Commun. (1999) 1141–1142. [14] B. Li, Y. Wang, Facile synthesis and enhanced photocatalytic performance of flowerlike ZnO hierarchical microstructures, J. Phys. Chem. C 114 (2010) 890–896. [15] Y. Duan, X. Liu, L. Han, S. Asahina, D. Xu, Y. Cao, Y. Yao, S. Che, Optically active chiral CuO “nanoflowers”, J. Am. Chem. Soc. 136 (2014) 7193–7196. [16] A. Narayanaswamy, H. Xu, N. Pradhan, M. Kim, X. Peng, Formation of nearly monodisperse In2O3 nanodots and oriented-attached nanoflowers: hydrolysis and alcoholysis vs pyrolysis, J. Am. Chem. Soc. 128 (2006) 10310–10319. [17] W.L. Noorduin, A. Grinthal, L. Mahadevan, J. Aizenberg, Rationally designed complex, hierarchical microarchitectures, Science 340 (2013) 832–837. [18] L. Xu, J. Liu, J. Du, Y. Peng, Y. Qian, A self-assembly template approach to form hollow hexapod-like, flower-like and tube-like carbon materials, Carbon 43 (2005) 1560–1562. [19] J.-M. Shen, Y.-T. Feng, Formation of flower-like carbon nanosheet aggregations and their electrochemical application, J. Phys. Chem. C 112 (2008) 13114–13120. [20] J. Liang, S. Chen, M. Xie, Y. Wang, X. Guo, X. Guo, W. Ding, Expeditious fabrication of flower-like hierarchical mesoporous carbon superstructures as supercapacitor electrode materials, J. Mater. Chem. A 2 (2014) 16884–16891. [21] D. Varshney, A.V. Sumant, O. Resto, F. Mendoza, K.P. Quintero, M. Ahmadi, B.R. Weiner, G. Morell, Single-step route to hierarchical flower-like carbon nanotube clusters decorated with ultrananocrystalline diamond, Carbon 63 (2013) 253–262.

153

[22] L. A., R. Abbaschian, R.E. Reed-Hill, Physical Metallurgy Principles, Cengage Learning, Stamford CT, 2009 355–357. [23] M.A. Ermakova, D.Y. Ermakov, A.L. Chuvilin, G.G. Kuvshinov, Decomposition of methane over iron catalysts at the range of moderate temperatures: the influence of structure of the catalytic systems and the reaction conditions on the yield of carbon and morphology of carbon filaments, J. Catal. 201 (2001) 183–197. [24] I. Alstrup, A new model explaining carbon filament growth on nickel, iron, and NiCu alloy catalysts, J. Catal. 109 (1988) 241–251. [25] C.A. Bernardo, I. Alstrup, J.R. Rostrup-Nielsen, Carbon deposition and methane steam reforming on silica-supported Ni-Cu catalysts, J. Catal. 96 (1985) 517–534. [26] M.T. Tavares, C.A. Bernardo, I. Alstrup, J.R. Rostrup-Nielsen, Reactivity of carbon deposited on nickel-copper alloy catalysts from the decomposition of methane, J. Catal. 100 (1986) 545–548. [27] Y. Ma, X. Sun, N. Yang, J. Xia, L. Zhang, X. Jiang, Shape-controlled growth of carbon nanostructures: yield and mechanism, Chem. Eur. J. 21 (2015) 12370–12375. [28] M.S. Saavedra, G.D. Sims, L.N. McCartney, V. Stolojan, J.V. Anguita, Y.Y. Tan, S.L. Ogin, P.A. Smith, S.R.P. Silva, Catalysing the production of multiple arm carbon octopi nanostructures, Carbon 50 (2012) 2141–2146. [29] D. Er-Xiong, G. Hong-Zhang, W. Jing, L. Zhi-Jia, L. Guangfen, W. Wen-Yi, L. Lin-Ge, Y. Hai-Jie, D. Shi-Xun, W. Jie, J. Hua, I.K. Esko, Hierarchical chrysanthemum-flower-like carbon nanomaterials grown by chemical vapor deposition, Nanotechnology 27 (2016) 085602. [30] M. Argyle, C. Bartholomew, Heterogeneous catalyst deactivation and regeneration: a review, Catalysts 5 (2015) 145. [31] L. Yu, L. Sui, Y. Qin, F. Du, Z. Cui, Catalytic synthesis of carbon nanofibers and nanotubes by the pyrolysis of acetylene with iron nanoparticles prepared using a hydrogen-arc plasma method, Mater. Lett. 63 (2009) 1677–1679. [32] L. Stolt, F.M. D'Heurle, J.M.E. Harper, On the formation of copper-rich copper silicides, Thin Solid Films 200 (1991) 147–156. [33] B. Cao, G.-P. Li, X.-M. Chen, S.-J. Cho, H. Kim, Atomic diffusion in Cu/Si (111) and Cu/ SiO2/Si (111) systems by neutral cluster beam deposition, Chin. Phys. Lett. 25 (2008) 1400. [34] C.V. Thompson, Solid-state dewetting of thin films, Annu. Rev. Mater. Res. 42 (2012) 399–434. [35] G.K. Goswami, R. Nandan, K.K. Nanda, Growth of branched carbon nanotubes with doped/un-doped intratubular junctions by one-step co-pyrolysis, Carbon 56 (2013) 97–102. [36] B. Cao, Y.-H. Jia, G.-P. Li, X.-M. Chen, Atomic diffusion in annealed CU/SiO2/Si (100) system prepared by magnetron sputtering, Chin. Phys. B 19 (2010) 026601. [37] S.-Y. Jang, S.-M. Lee, H.-K. Baik, Tantalum and niobium as a diffusion barrier between copper and silicon, J. Mater. Sci. Mater. Electron. 7 (1996) 271–278. [38] J. Hong, S. Lee, S. Lee, H. Han, C. Mahata, H.-W. Yeon, B. Koo, S.-I. Kim, T. Nam, K. Byun, B.-W. Min, Y.-W. Kim, H. Kim, Y.-C. Joo, T. Lee, Graphene as an atomically thin barrier to Cu diffusion into Si, Nano 6 (2014) 7503–7511. [39] S.-Q. Wang, Barriers against copper diffusion into silicon and drift through silicon dioxide, MRS Bull. 19 (1994) 30–40. [40] B.T. Mayumi, S. Masaru, A. Eiji, N. Atsushi, Characterization of TiHfN ternary alloy films as a barrier between Cu plug and Si, Jpn. J. Appl. Phys. 53 (2014) (02BC04). [41] C.A. Chang, Formation of copper silicides from Cu(100)/Si(100) and Cu(111)/ Si(111) structures, J. Appl. Phys. 67 (1990) 566–569. [42] S. Sakurai, H. Nishino, D.N. Futaba, S. Yasuda, T. Yamada, A. Maigne, Y. Matsuo, E. Nakamura, M. Yumura, K. Hata, Role of subsurface diffusion and Ostwald ripening in catalyst formation for single-walled carbon nanotube forest growth, J. Am. Chem. Soc. 134 (2012) 2148–2153. [43] G.Y. Chen, B. Jensen, V. Stolojan, S.R.P. Silva, Growth of carbon nanotubes at temperatures compatible with integrated circuit technologies, Carbon 49 (2011) 280–285. [44] S. Helveg, C. Lopez-Cartes, J. Sehested, P.L. Hansen, B.S. Clausen, J.R. Rostrup-Nielsen, F. Abild-Pedersen, J.K. Norskov, Atomic-scale imaging of carbon nanofibre growth, Nature 427 (2004) 426–429. [45] V. Stolojan, Y. Tison, G.Y. Chen, R. Silva, Controlled growth-reversal of catalytic carbon nanotubes under electron-beam irradiation, Nano Lett. 6 (2006) 1837–1841. [46] W.Z. Li, J.G. Wen, Z.F. Ren, Straight carbon nanotube Y junctions, Appl. Phys. Lett. 79 (2001) 1879–1881. [47] J. Cheng, X. Zou, H. Zhang, F. Li, P. Ren, G. Zhu, Y. Su, M. Wang, Growth of Y-shaped carbon nanofibers from ethanol flames, Nanoscale Res. Lett. 3 (2008) 295. [48] P. Ghosh, M. Subramanian, R.A. Afre, M. Zamri, T. Soga, T. Jimbo, V. Filip, M. Tanemura, Growth of Y-junction bamboo-shaped CNx nanotubes on GaAs substrate using single feedstock, Appl. Surf. Sci. 255 (2009) 4611–4615.