Experimental investigation of interfaces in graphene materials/copper composites from a new perspective

Abstract

Copper matrix composites reinforced with three types of graphene materials, namely graphene nanoplatelets, nickel-plated GNPs and reduced graphene oxide, were separately fabricated using a molecular-level mixing process. Different from nanoparticles and nanofibers, graphene materials exhibit a unique two-dimension structure. The interface between graphene and the matrix could be divided into two types, namely graphene plane–Cu (D_p) and graphene edges–Cu (D_e). The interface microstructure between the constituent phases in graphene/Cu composites were observed for the first time from the two directions by means of transmission electron microscopy (TEM). TEM analysis results showed a big difference in interface microstructure and bonding mechanism in the three kinds of composites. The graphene materials were coherently bonded with the copper matrix in terms of mechanical, metallurgical or chemical bonding.

---

Introduction

Copper matrix composites are widely used in the microelectronics and automobile industry due to their excellent comprehensive properties.^1–3 Many nanomaterials with outstanding mechanical and physical properties have been applied as reinforcements to reinforce the copper matrix.^4–7 Graphene, a single-layer graphite sheet, has been regarded as a promising candidate of the next generation reinforcement for copper matrix composites because of it's big specific surface area (2630 m² g⁻¹),⁸ high thermal conductivity (5 × 10³ W m⁻¹ K⁻¹),^9,10 good electrical conductivity (10⁶ S m⁻¹),¹¹ and exceptional tensile strength (125 GPa).^9,12

Graphene can be synthesized by different approaches, such as mechanical exfoliation,^13,14 chemical vapor deposition,^11,15–17 chemical reduction^18,19 and large-area graphene growth method.²⁰ The structure of graphene strongly depended on the synthesized methods, different synthesized method led to the big difference in structure and properties, which was subsumed under the concept of “graphene derivatives” or “graphene materials”. Nickel plated graphene nanoplatelets (Ni-GNPs) were also mentioned in the literature, therefore, graphene materials should be incorporated into the matrices through different methods to achieve good interface bonding and uniform distribution. Several processing methods, such as powder metallurgy,^21–27 disintegrated melt deposition,²⁸ molecular-level mixing (MLM),²⁹ electrochemical method^30,31 and surface modification,^32,33 have been developed to uniformly incorporate graphene materials into the metal matrices. So far, powder metallurgy or disintegrated melt deposition have been used to incorporate graphene nanoplatelets (GNPs) or ultra-thin graphite sheets (containing multilayer graphene with thickness of less than 100 nm) into copper matrix successfully.^23,24,28 GNPs showed a good strengthening efficiency in the composites due to their excellent mechanical properties. For instance, AZ61-3GNP composite prepared by disintegrated melt deposition method showed 26% and 11.7% increase in yield strength and ultimate tensile strength, compared to monolithic alloy.²⁸ Graphene oxide (GO), containing many hydrophilic functional groups on graphene layer, was the most widely used as graphene precursor to fabricate composites. GO can be reduced to graphene (RGO) by chemical or thermal reduction.^18,34 It has been confirmed that solution-based methods were suitable for synthesizing RGO/Cu composites.^25,29–31 RGO/Cu nanocomposites containing 2.5 vol% RGO synthesized by the MLM method showed a 30% increase in tensile strength and a 30% increase in elastic modulus compared with pure copper.²⁹ Moreover, graphene plated with metal nanoparticles has been incorporated into metal matrices to enhance the weak interface of graphene/metal composites made by powder metallurgy. The copper composites with only 1.0 vol% Ni-graphene showed a 61% increase in elastic modulus and a 94% increase in yield strength.³³

Graphene materials have showed excellent strengthening effect in copper matrix composites. Strengthening mechanisms included grain-size refinement, Orowan strengthening, load transfer effects as well as thermal expansion mismatch. The interface bonding strength between graphene materials and matrix in graphene/Cu composites may vary due to the different fabrication methods as well as different strengthening mechanisms. So far, the main efforts have been focused on developing new processing route to disperse graphene homogeneously into matrices and enhance the interface bonding between the constituent phases in composites. To understand the graphene/matrix interface structure is still necessary in order to provide the fundamental knowledge of controlling the microstructure, so as to optimize the preparation methods and parameters. Some interface bonding types presented between graphene materials and the metal matrices were mentioned in the published papers. For example, mechanical bonding usually existed in GNPs/copper composites fabricated by ball milling due to the non-wetting of the GNPs and matrix.²⁴ Oxygen-mediated bonding was deduced in oxygen-containing graphene materials reinforced copper or nickel matrix composites.^35,36

The present study is undertaken in an effort to understand the interface microstructure of graphene materials/Cu composites by means of transmission electron microscopy (TEM). Copper matrix composites reinforced with GNPs, Ni-GNPs and RGO were separately fabricated using a MLM process. The purposes of this work are twofold: firstly, to observe the interface microstructure between graphene and Cu from two directions since graphene has a typical two-dimension structure, namely graphene plane–Cu and graphene edges–Cu, which were hereafter called the D_p and D_e; secondly, to understand the interface structure and strengthening mechanisms of constituent phases in graphene materials/Cu composites.

Experimental

Sample preparation

As-received raw materials were used in the present work. The GNPs purchased from XFNANO Material Technology Co., Ltd. were 98.9% in purity (1.1 wt% O) and 1–5 μm in lateral dimension. The thickness of raw GNPs was 5–10 nm, corresponding to approximately 15–30 layers of monolayer graphene.²³ Ni-GNPs were synthesized by a typical electroless nickel plating process. The GNPs were first concussed by ultrasonication (100 W) for 2 h in deionized water. Then the GNPs were immersed in 200 ml of SnCl₂ solution (10 g l⁻¹ SnCl₂, 40 ml l⁻¹ HCl) under ultrasonication for 30 min for sensitization. After being washed and filtered with deionized water, the sensitized GNPs were put into 200 ml of PdCl₂ solution (0.5 g l⁻¹ PdCl₂, 25 ml l⁻¹ HCl) under ultrasonication for another 30 min. The activated GNPs were added into the solution of NiSO₄·6H₂O (0.076 mol l⁻¹), NaH₂PO₂·H₂O (0.283 mol l⁻¹), Na₃C₆H₅O₇·2H₂O (0.034 mol l⁻¹), NH₃·H₂O (keep the pH value at ∼8.75) to complete Ni plating at a water-bath temperature of 35 °C. The GO aqueous solution with concentration of ∼2.86 mg ml⁻¹ was purchased from JCNANO Material Technology Co., Ltd. The GO was ≥99% in purity, 1–5 μm in lateral dimension, 0.8–1.2 nm in thickness and ≥99% in single layer ratio. Cu(CH₃COO)₂·2H₂O used as precursor of Cu was in analytically pure grade. Fig. 1 showed the representative morphology of GNPs, Ni-GNPs and GO powders. The X-ray photoelectron spectra (XPS) of raw powders shown in Fig. 1(a)–(c) were given on the right. Fig. 1(a) showed the morphology of GNPs that exhibited a stacking flake structure, suggesting ultrasonic pre-treatment was necessary to overcome the attraction between graphene layers. Fig. 1(b) presented that the Ni-GNPs exhibited flake shape after electroless nickel plating. The coating layer was dense and smooth. Fig. 1(c) showed that GO powder exhibited an ultra-thin flake structure.

Fig. 1 SEM micrographs of GNPs (a) and Ni-GNPs (b); TEM image of GO powders (c). The images on the right are the XPS spectra of powders in (a–c).

The GNPs/Cu composite powders with GNPs content at 0.5 vol% were synthesized by a MLM process. Steps involved in the MLM method were as described in the previous study.³⁷ The sensitization and activation process was not performed during the preparation of Ni-GNPs/Cu and RGO/Cu composite powders. The as-prepared graphene materials/Cu composite powders were consolidated into bulk nanocomposites with full densification by spark plasma sintering (SPS), which could provide high heating rate and rapid consolidation through high-joule heating and spark plasma generated between the powders. The composite powders were sintered at 600 °C for 5 min with a heating rate of 50 °C min under a uniaxial pressure of 45 MPa.

Characterization

The microstructure of the raw materials and composite powders was examined using a scanning electronic microscopy (SEM) and TEM, both of which were equipped with energy-dispersive X-ray spectroscopy (EDS) system. The powders were fully dispersed in ethanol using an ultrasonic oscillator and then collected on a 200-mesh copper grid to make powder TEM samples. The collected powders were thin enough to be observed by TEM at 200 kV. XPS and Raman spectroscopy were applied to investigate the chemical composition and characterize the structure of graphene materials. Interface microstructure of the bulk composites was also characterized by TEM. TEM samples of the bulk composites were firstly mechanically polished and then thinned to 20 μm by mechanical polishing, followed by ion milling. Fig. 2 showed the schematic diagram of the TEM observation directions of D_p and D_e. The mechanical properties of the composites were characterized by tensile tests. Tensile tests were conducted at ambient temperature (25 °C) on a universal testing machine (TH5000) with the crosshead speed of 0.6 mm min⁻¹. The test samples with a gauge length of 10 mm and cross-section of 2 mm × 1.5 mm were machined from the sintered billets. Three samples were tested from each sintered billet for tensile tests to minimize the error. The average tensile strength, yield strength (0.2%) and total elongation were determined from the tensile curves.

Fig. 2 A schematic diagram of the TEM observation directions of D_p and D_e (a); the dimensions (in mm) of tensile samples (b).

Results and discussion

Fig. 3 showed the typical SEM micrographs and Raman spectra of elemental powders after reduction by high purity hydrogen (MLM process). Graphene materials could be identified from the composite powders (Fig. 3) based on their unique appearance. The existence of graphene materials indicated little decomposition or damage of graphene structure was made during the MLM process, which was further proved by Raman analysis below. Fig. 3(a) presented the morphology of GNPs/Cu composite powders. The GNPs were subtranslucent and did not aggregated obviously in copper particles. However, the GNPs showed a lateral size of 20–30 μm, which was 4–6 times as large as raw material. It might be caused by that edges of some multilayer graphene adsorbed on the in-plane of another graphene resulting an increase of lateral size but no obvious increase in thickness was observed. Meanwhile, some copper particles were adsorbed on the surface of GNPs as shown in Fig. 3(a). The adsorbed particles inhibited the aggregation of GNPs and would improve wettability between these constituent phases leading to good interface bonding of sintered composites. Fig. 3(b) and (c) showed the morphology of the Ni-GNPs/Cu and RGO/Cu composite powders. Agglomerate of Ni-GNPs or RGO could hardly be seen. Importantly, the Ni-GNPs were well wetted by the copper particles, proving that the plated Ni successfully decreased the contact angle of copper on the GNPs. Raman spectroscopy of the raw and composite powders characterized the structure change of graphene materials, as shown in Fig. 3(d). Raman spectra of the GNPs/Cu and Ni-GNPs/Cu powders showed typical characteristic peaks of GNPs, while the Raman spectra of the RGO/Cu powders were the signatures of GO structure, insulating that main graphene structure was maintained during the MLM process. It could be obviously found that GO and RGO were more defective compared with GNPs materials based on the intensity ratio of the D band to G band, I_D/I_G. Moreover, G bands of GO showed right shift compared with G bands of GNPs, indicating that RGO exhibited fewer graphene layers than that of GNPs.³⁸

Fig. 3 SEM micrographs of GNPs/Cu (a), Ni-GNPs/Cu (b) and RGO/Cu (c) composite powders; (d) Raman spectra of activated GNPs, GNPs/Cu, Ni-GNPs/Cu, GO and RGO/Cu powders. The insets are the EDS corresponding to the marked points.

The interfacial bonding played a key role in the load transfer between the matrix and graphene, as well as in the inhibition of dislocation movement, which are the most important strengthening mechanisms of graphene/metal composites.³⁹ The following sections focused on the discussion of the interface microstructure between the three types of graphene materials and the copper matrix.

Interfaces in GNPs/Cu composites

The distribution feature of GNPs in the composites was characterized by TEM, as shown in Fig. 4. According to the selection of electron diffraction pattern (inset in Fig. 4(a)), the light-colored phase was GNPs, while the dark phase was the copper matrix. The GNPs located either at grain boundaries or in grain interiors of the copper matrix. As shown in Fig. 4(a), the GNPs with diameters of below 1 μm distributed at the copper grain boundaries, which maintained the flake structure like the raw powder. The GNPs in grain interiors were spherical in shape with size of several decades nanometers, as shown in Fig. 4(b). The special distribution was attributed to the separation of fine graphene pieces from raw GNPs during the ultrasonic oscillation and mechanical stirring process in MLM method.

Fig. 4 TEM micrographs of GNPs/Cu at low magnification.

Fig. 5 showed high resolution transmission electron microscopy (HRTEM) images and schematic diagrams of the interface between GNPs and Cu in directions of D_p and D_e. The GNPs in Fig. 5(a) were about 1 μm in lateral dimension and about 30 nm in thickness. Fig. 5(a) presented a bright-field TEM image of the interface between GNPs basal plane and the copper matrix in D_p. The interface was free of voids, gaps or impurities, which was further illustrated in Fig. 5(c), a HRTEM image taken from the marked area in Fig. 5(a). The graphene layers were clearly stacked and the spaces between lines were about 0.34 nm, close to the theoretical value for graphene stacking. As shown in Fig. 5(c), the interface between GNPs and the copper matrix was clean and well bonded, illustrating a mechanical bonding. There was a 2 nm wide step at the interface. The uneven interface could provide a mechanical lock between GNPs and the copper matrix. The interface microstructure in D_e was shown in Fig. 5(b) and (d). The left image in Fig. 5(d) was a HRTEM image taken from the interface of marked area in Fig. 5(b), while the right one corresponded to the surface of GNPs in the marked area. It was also clear that the GNPs edges were coherently bonded with the copper matrix without voids or gaps. Different from D_p, the interface in D_e was uneven and the two phases were interdigitated at an atomic scale, which was estimated as metallurgical bonding.⁴⁰ The cause of the interfacial difference in the D_p and D_e might be that the carbon atoms at graphene edges differed from those away from the edges. There are many dangling bonds and vacancies at graphene edges, and the uneven edges were observed using annular dark field of a low-voltage scanning transmission electron microscope.⁴¹ During the decomposition and reduction of copper acetate, Cu atoms coherently grew on the uneven graphene edges, resulting atomic-scale diffusion at the interface. Moreover, the localized plasma heating might also promote this quite low diffusion when the green body underwent through SPS.²⁹ Some copper oxides nanoparticles were observed on the surface of GNPs, as shown in Fig. 5(d). The formation of copper oxides was mainly attributed to the reaction between the copper reduced from copper acetate and native oxygen of raw GNPs. These oxides complicated the interfacial microstructure. The existence of native oxygen on GNPs offered numerous sites of oxygen-mediated bonding which could enhance the interfacial bonding.^29,35,36 The formation of copper oxides would reduce the oxygen-mediated bonding, leading to a decrease of interface bonding strength. Based on the observation result, schematic diagrams in D_p and D_e of the GNPs/Cu composite were illustrated in Fig. 5(e) and (f). The interfacial bonding of GNPs/Cu fabricated by MLM was a combination of mechanical bonding (D_p) and metallurgical bonding (D_e). Normally, the metallurgical bonding was not seen in the interaction between GNPs and the copper matrix.²⁴

Fig. 5 Interface TEM and HRTEM micrographs of the GNPs/Cu composite in D_p (a and c) and in D_e (b and d); interface schematic diagrams in D_p (e) and D_e (f).

Interface investigation of Ni-GNPs/Cu composites

Fig. 6 showed low-magnification TEM micrographs of Ni-GNPs/Cu composites. Ni-GNPs distributed both at grain boundaries and grain interiors of the copper matrix. As shown in Fig. 6(a), the Ni-GNPs at grain boundaries were about 1 μm in diameter, which was in accordance with raw GNPs. Ni-GNPs distributed in grain interiors presented irregular flake shape (Fig. 4(d)), which was different from GNPs. It might be because of the reduction of surface tension of GNPs after the electroless nickel plating.

Fig. 6 TEM micrographs of Ni-GNPs/Cu at low magnification.

Fig. 7 showed HRTEM images and typical EDS of the interface of Ni-GNPs/Cu composite. The insets were the diffraction patterns derived from the Fast-Fourier Transform (FFT) corresponding to Fig. 7(c) and (d). The interfacial microstructure in D_p was shown in Fig. 7(a) and (c), while the one in D_e was shown in Fig. 7(b) and (d). The Ni-GNPs in Fig. 7(a) were about 250 nm in lateral dimension and about 40 nm in thickness. As shown in Fig. 7, the interface between Ni-GNPs and the copper matrix was coherently bonded without any cracks or voids. Unlike the interfacial microstructure in the GNPs/Cu composite, a narrow transition zone of about 1 nm was found between Ni-GNPs and the copper matrix (Fig. 7(c) and (d)). The results of FFT diffraction patterns and EDS (Fig. 7(e)–(g)) proved that the interface transition zone was a layer of nickel. As reported, electroless plating could improve the wettability between graphene and the copper matrix.³² The nickel particles adsorbed on GNPs partially dissolved in the copper matrix near the interface, as shown in Fig. 7(d) and (g). Thus, a good chemical bonding formed both in D_p and D_e after decorating GNPs with nickel particles. It should be noticed that the proportion of interface in D_e was much smaller compared with the one in D_p. Consequently, the improvement of interfacial bonding in D_p highly enhanced the strengthening effect of GNPs. That is, chemical bonding between Ni-GNPs and the copper matrix could withstand higher load capacities compared with the mechanical and metallurgical bonding between GNPs and the matrix.

Fig. 7 HRTEM micrographs of the Ni-GNPs/Cu composite in D_p (a and c) and D_e (b and d); EDS (e–g) of points marked in (b). The insets are the diffraction patterns derived from the Fast-Fourier Transform (FFT) corresponding to the selected regions.

Interface investigation of RGO/Cu composites

Fig. 8 showed the distribution of RGO in the copper matrix. Like GNPs and Ni-GNPs, RGO also distributed both at grain boundaries and grain interiors of the copper matrix. As shown in Fig. 8(a), the size of RGO at grain boundaries was rather smaller compared with GNPs and Ni-GNPs. This mainly benefited from small thickness of GO and the inhibition of agglomeration by polar groups on the surface of GO. The ultra-thin structure of RGO made the TEM interface observation of RGO/metal composites difficult. Although interfacial structure of RGO/metal composites has been mentioned, very few HRTEM images of interface between RGO and metal matrices have been presented.^23,42,43

Fig. 8 TEM micrographs of RGO/Cu at low magnification.

The morphology of RGO embedded in the copper matrix and interface microstructure of RGO/Cu were shown in Fig. 9. The RGO in Fig. 9(a) was about 100 nm in lateral dimension. Fig. 9(a) presented a bright-field TEM image of the RGO–Cu interface in D_p. Thin graphene layers in the middle were etched off under the exposure of high-voltage TEM electron beam. The remained RGO was still bonded tightly with the copper matrix. It demonstrated the interface bonding between RGO and Cu was stronger than the bonding of graphene interlayers. Fig. 9(b) showed a bright-field TEM image of the RGO–Cu interface in D_e. The diffraction pattern corresponding to the selected region in Fig. 9(b) displayed the hexagonal lattice structure of graphene layer as well as the structure of the copper matrix. RGO was bonded tightly with the matrix. Fig. 9(c) showed both sides of a sandwich structure of Cu–RGO–Cu selected in Fig. 9(a). The thickness of the multilayer RGO was about 10–15 nm, corresponding to 30–50 graphene layers. The coherently bonded interface was free of voids and gaps. Compared with the microstructure of GNPs/Cu composite, the interface in D_p between RGO and the copper matrix was blurry (right in Fig. 9(c)), which was inferred as a transition zone.⁴⁴ The blurry interface indicated interaction between atoms of carbon and copper. Fig. 9(d) showed that the interface in D_e of RGO/Cu composite (rectangular selection in Fig. 9(b)) was roughly the same as that of GNPs/Cu. But there was about a 0.5 nm wide transition zone between RGO and the copper matrix. EDS (Fig. 9(e)–(g)) corresponding to the points A, B and C marked in Fig. 9(b) was used to analyze the element distribution near the interface in D_e. Fig. 9(e) revealed some oxygen existed on the RGO surface, indicating the reduction of GO was incomplete. Reports on the preparation of RGO and RGO based composites showed the complete reduction of GO was difficult and time-tested.⁴⁵ Fig. 9(f) illustrated the RGO–Cu interface was an oxygen-rich area, which was in accordance with the diffraction contrast shown in Fig. 9(d). The oxygen signal was rather weak in the copper matrix (Fig. 9(g)). The EDS results suggested the oxygen-mediated bonding formed between RGO and the copper matrix. Compared with the GNPs/Cu composite, the interaction between carbon of RGO (with many oxygen of residue polar functional group) and copper atoms should be stronger as oxygen-mediated bonding formed both in D_p and D_e.²⁹ The formation of oxygen-mediated bonding helped to improve the interface bonding in the RGO/Cu composite, such result has been also found in the RGO/Cu and RGO/Ni composites fabricated by other chemical methods.^35,36

Fig. 9 HRTEM micrographs of the RGO/Cu composite in D_p (a and c) and D_e (b and d); EDS (e–g) of points marked in (b). The inset in (b) is the diffraction pattern derived from the FFT corresponding to the selected region.

Mechanical properties analysis

Fig. 10 showed tensile strength of the composites and pure copper. The tensile strength of pure copper increased by 52 MPa after adding 0.5 vol% GNPs. When using 0.5 vol% Ni-GNPs as reinforcement instead of GNPs, the tensile strength of composite further increased by another 25 MPa. As reported, strength improvement of metal matrix composites reinforced by graphene could be attributed to grain-size refinement, Orowan strengthening, load transfer effects as well as thermal expansion mismatch.^23,26 The only difference between GNPs/Cu and Ni-GNPs/Cu composites was whether electroless nickel plating was performed on the surface of GNPs before MLM process. As discussed above (Fig. 5 and 7), the interface bonding between Ni-GNPs and Cu was stronger than that of GNPs-Cu. Good interface bonding lead to excellent load transfer, and accordingly the tensile strength improved.

Fig. 10 Tensile strength of the composites and pure copper.

As shown in Fig. 10, the tensile strength of 0.5 vol% RGO/Cu increased by 22 MPa compared with 0.5 vol% GNPs/Cu composites. Based on the differences in structure, distribution as well as interface bonding, the strength increase was a comprehensive result of many factors, such as Orowan strengthening, load transfer effects and thermal expansion mismatch. Moreover, the improvement of strength of Cu composites reinforced with GNPs was achieved at the expense of ductility, as shown in Table 1. However, 0.5 vol% RGO/Cu composite exhibited ductility as good as pure copper, which was in accordance with previous report.⁴⁶ The behavior might be arisen from the wrinkle structure of graphene and good interface bonding between graphene and the matrix. As shown in Fig. 5(c), 7(c) and 9(c), carbon layers of RGO showed more obvious wavy structure than that of GNPs. The wrinkled RGO was straightened during load transfer from the matrix to RGO. Thus, the ductility of the composite was maintained or even improved. Strengthening efficiency R is widely used to determine the strengthening effect of reinforcement on composites, which can be calculated by equation of R = (σ_c − σ_m)/V_fσ_m, where σ_c and σ_m are the yield strength of composite and matrix respectively, and V_f is volume fraction of the reinforcement. The R values of graphene materials on pure copper in the present work (Table 1) were above 44. The strengthening efficiencies were much larger than traditional reinforcements on metal matrix composites, such as Al₂O₃, SiC, carbon fiber and carbon nanotube.⁴⁷ It demonstrated that graphene materials were promising as reinforcements in copper composites. There are some reports on interface microstructure between the constituent phases in metal composites reinforced with graphene materials. For instance, Li et al.⁴⁰ characterized the interface microstructure between aluminium and graphene using HRTEM. Lin et al.⁴³ also investigated the GO–Fe interfacial structure using HRTEM. The former report focused on the interface in D_p while the latter focused on the interface in D_e. The present work studied the interface microstructure in both directions (D_p and D_e) for the first time, which may provide some information on the microstructure design of graphene composites.

Table 1 Tensile property of Cu composites and pure Cu

Materials	0.2% yield strength (MPa)	Tensile strength (MPa)	Elongation (%)	R
Cu	136 ± 2.3	226 ± 1.6	24.3 ± 1.2
0.5 vol% GNPs/Cu	174 ± 5.5	256 ± 4.3	18.1 ± 3.8	55.9
0.5 vol% Ni-GNPs/Cu	175 ± 8.6	281 ± 5.8	11.9 ± 4.4	57.4
0.5 vol% RGO/Cu	166 ± 10.5	278 ± 4.7	27.9 ± 3.5	44.1

---

Conclusions

Copper matrix composites reinforced with GNPs, Ni-GNPs and RGO were fabricated by a MLM method respectively. The interface microstructure was discussed based on the directions of graphene basal plane–Cu and graphene edges–Cu for the first time by the TEM observation. All of the three types of graphene materials were coherently bonded with the copper matrix without any cracks or voids. Characteristic dimensions of graphene materials can be determined in D_p. Typically, GNPs were ∼1 μm in lateral dimension and ∼30 nm in thickness; Ni-GNPs were 250–1000 nm in lateral dimension and ∼40 nm in thickness; RGO was ∼100 nm in lateral dimension and 10–15 nm in thickness in the composites. A combination of mechanical bonding in D_p and metallurgical bonding in D_e was observed between GNPs and the copper matrix. Benefiting from the decoration of nickel particles on GNPs, a 1 nm-thick transition zone containing nickel formed at the interface between Ni-GNPs and the copper matrix. The good interfacial adhesion between Ni-GNPs and the matrix was attributed to chemical interaction both in D_p and D_e. For the RGO/Cu composite, a transition zone different from the one at the Ni-GNPs/Cu interface was characterized to be an oxygen-enriched region. The interfacial adhesion between RGO and the copper matrix was oxygen-mediated chemical bonding. The improvement of interface bonding strength increased tensile strength of composites containing the same graphene material. The detail observation of interface microstructure is advantageous to the interface design of graphene material/Cu composites.

Acknowledgements

This work was supported by the National Basic Research Program of China (2010CB71600).

References

1. Q. Liu, X. He, S. Ren, C. Zhang, T. Liu and X. Qu, J. Alloys Compd., 2014, 587, 255–259.
2. S. C. Tjong and Z. Y. Ma, Mater. Sci. Eng., R, 2000, 29, 49–113.
3. X. Jiang, W. Liu, J. Li, Z. Shao and D. Zhu, J. Alloys Compd., 2015, 618, 700–706.
4. S. Cha, K. Kim, S. Arshad, C. Mo and S. Hong, Adv. Mater., 2005, 17, 1377–1381.
5. S. M. Uddin, T. Mahmud, C. Wolf, C. Glanz, I. Kolaric, C. Volkmer, H. Höller, U. Wienecke, S. Roth and H. Fecht, Compos. Sci. Technol., 2010, 70, 2253–2257.
6. E. Neubauer, M. Kitzmantel, M. Hulman and P. Angerer, Compos. Sci. Technol., 2010, 70, 2228–2236.
7. D. Zhou, F. Qiu and Q. Jiang, Mater. Sci. Eng., A, 2015, 622, 189–193.
8. M. D. Stoller, S. Park, Y. Zhu, J. An and R. S. Ruoff, Nano Lett., 2008, 8, 3498–3502.
9. A. A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao and C. N. Lau, Nano Lett., 2008, 8, 902–907.
10. J. D. Renteria, D. L. Nika and A. A. Balandin, Appl. Sci., 2014, 4, 525–547.
11. K. S. Kim, Y. Zhao, H. Jiang, S. Y. Lee, J. M. Kim, K. S. Kim, J. H. Ahn, P. Kim, J. Y. Choi and B. H. Hong, Nature, 2009, 457, 706–710.
12. C. Lee, X. Wei, J. W. Kysar and J. Hone, Science, 2008, 321, 385–388.
13. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666–669.
14. K. S. Novoselov, D. Jiang, F. Schedin, T. J. Booth, V. V. Khotkevich, S. V. Morozov and A. K. Geim, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 10451–10453.
15. Z. Z. Sun, Z. Yan, J. Yao, E. Beitler, Y. Zhu and J. M. Tour, Nature, 2010, 468, 549–552.
16. X. S. Li, W. W. Cai, J. H. An, S. Kim, J. Nah, D. X. Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc, S. K. Banerjee, L. Colombo and R. S. Ruoff, Science, 2009, 324, 1312–1314.
17. P. Goli, H. Ning, X. Li, C. Y. Lu, K. S. Novoselov and A. A. Balandin, Nano Lett., 2014, 14, 1497–1503.
18. S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y. Jia, Y. Wu, S. T. Nguyen and R. S. Ruoff, Carbon, 2007, 45, 1558–1565.
19. X. B. Fan, W. C. Peng, Y. Li, X. Y. Li, S. L. Wang, G. L. Zhang and F. B. Zhang, Adv. Mater., 2008, 20, 4490–4493.
20. S. Amini, J. Garay, G. Liu, A. A. Balandin and R. Abbaschian, J. Appl. Phys., 2010, 108, 094321.
21. M. Rashad, F. S. Pan, A. Tang, M. Asif and M. Aamir, J. Alloys Compd., 2014, 603, 111–118.
22. Z. Xu, X. Shi, W. Zhai, J. Yao, S. Song and Q. Zhang, Carbon, 2014, 67, 168–177.
23. W. J. Kim, T. J. Lee and S. H. Han, Carbon, 2014, 69, 55–65.
24. K. Chu and C. Jia, Phys. Status Solidi A, 2014, 211, 184–190.
25. J. Wang, Z. Li, G. Fan, H. Pan, Z. Chen and D. Zhang, Scr. Mater., 2012, 66, 594–597.
26. M. Rashad, F. S. Pan, A. Tang and M. Asif, Progress in Natural Science: Materials International, 2014, 24, 101–108.
27. M. Rashad, F. S. Pan, M. Asif and A. Ullah, Mater. Sci. Technol., 2015, 31, 1452–1461.
28. M. Rashad, F. S. Pan, D. Lin and M. Asif, Mater. Des., 2016, 89, 1242–1250.
29. J. Hwang, T. Yoon, S. H. Jin, J. Lee, T. S. Kim, S. H. Hong and S. Jeon, Adv. Mater., 2013, 25, 6724–6729.
30. C. L. Pavithra, B. V. Sarada, K. V. Rajulapati, T. N. Rao and G. Sundararajan, Sci. Rep., 2014, 4, 4049.
31. F. Zhang, C. Hou, Q. Zhang, H. Wang and Y. Li, Mater. Chem. Phys., 2012, 135, 826–831.
32. M. Li, H. Che, X. Liu, S. Liang and H. Xie, J. Mater. Sci., 2014, 49, 3725–3731.
33. Y. Tang, X. Yang, R. Wang and M. Li, Mater. Sci. Eng., A, 2014, 599, 247–254.
34. D. Li, M. B. Muller, S. Gilje, R. B. Kaner and G. G. Wallace, Nat. Nanotechnol., 2008, 3, 101–105.
35. C. Zhao and J. Wang, Phys. Status Solidi A, 2014, 211, 2878–2885.
36. C. Zhao, Appl. Phys. A, 2015, 118, 409–416.
37. D. Zhang and Z. Zhan, J. Alloys Compd., 2016, 654, 226–233.
38. C. Cong, T. Yu, K. Sato, J. Shang, R. Saito, G. F. Dresselhaus and M. S. Dresselhaus, ACS Nano, 2011, 5, 8760–8768.
39. L. Chen, H. Konishi, A. Fehrenbacher, C. Ma, J. Xu, H. Choi, H. Xu, F. E. Pfefferkorn and X. Li, Scr. Mater., 2012, 67, 29–32.
40. J. L. Li, Y. C. Xiong, X. D. Wang, S. J. Yan, C. Yang, W. W. He, J. Z. Chen, S. Q. Wang, X. Y. Zhang and S. L. Dai, Mater. Sci. Eng., A, 2015, 626, 400–405.
41. K. Suenaga and M. Koshino, Nature, 2010, 468, 1088–1090.
42. Y. Kim, J. Lee, M. S. Yeom, J. W. Shin, H. Kim, Y. Cui, J. W. Kysar, J. Hone, Y. Jung, S. Jeon and S. M. Han, Nat. Commun., 2013, 4, 2114.
43. D. Lin, C. R. Liu and G. J. Cheng, Acta Mater., 2014, 80, 183–193.
44. R. Pérez-Bustamante, D. Bolaños-Morales, J. Bonilla-Martínez, I. Estrada-Guel and R. Martínez-Sánchez, J. Alloys Compd., 2014, 615, S578–S582.
45. D. R. Dreyer, R. S. Ruoff and C. W. Bielawski, Angew. Chem., Int. Ed., 2010, 49, 9336–9345.
46. S. J. Yan, S. L. Dai, X. Y. Zhang, C. Yang, Q. H. Hong, J. Z. Chen and Z. M. Lin, Mater. Sci. Eng., A, 2014, 612, 440–444.
47. S. I. Cha, K. T. Kim, S. N. Arshad, C. B. Mo and S. H. Hong, Adv. Mater., 2005, 17, 1377–1381.

---