SEARCH

SEARCH BY CITATION

Keywords:

  • graphene nanoplatelets;
  • metal–matrix composites;
  • powder metallurgy;
  • yield strength

Abstract

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

Graphene nanoplatelets (GNPs) exhibit ultra-high strength and elastic modulus. Therefore, they are potential ideal reinforcements in metal–matrix composites (MMCs). In this work, we report the use of GNPs to strengthen the bulk Cu-matrix composites. GNP reinforced Cu-matrix (GNP/Cu) composites were prepared by a combination of the ball milling and hot-pressing processing, and their mechanical properties were investigated. Microstructure studies indicated that the GNPs with 0–8 vol.% contents were well dispersed in the Cu matrix by ball milling. Compared to unreinforced Cu, the GNP/Cu composites showed a remarkable increase in yield strength and Young's modulus up to 114 and 37% at 8 vol.% GNP content, respectively. The extraordinary reinforcement is attributed to the homogeneous dispersion of GNPs and grain refinement. However, the mechanical improvement of GNP/Cu composites was still below the theoretical value. The possible reasons for this deviation were discussed and the methods for further mechanical improvement of GNP/Cu composites were proposed.

1 Introduction

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

Metal–matrix composites (MMCs) today are extensively used in automobile and aerospace applications [1]. Among the promising reinforcements for MMCs, there is a choice of different carbon allotropes and nanostructures: fullerene, carbon nanotube (CNT), nanofiber, and graphene. In the past decade, owing to the excellent mechanical and physical properties, CNT is receiving the significant attention as a novel reinforcement material for the production of advanced engineering MMCs [2]. Extensive experiments indicated that CNT can provide a high efficiency in properties enhancement of MMCs [2]. Until recently, CNT is the dominant nano-sized carbon reinforcement for MMCs. However, the development of CNT-reinforced MMCs is hindered by the complexity of their dispersion in metal matrix and high cost [3].

Graphene, a recently discovered two-dimensional (2D) platelet consisting of carbon atoms, has attracted tremendous attention from the scientific communities [4]. Its unique mechanical and physical properties make graphene a promising nanofiller to improve mechanical, electrical, and thermal properties of the composites [5, 6]. Recent experiments demonstrated an industrially viable procedure for the large-scale production of few-layered graphene nanoplatelets (GNPs) [7, 8], indicating that the production cost for GNP in large quantities is much lower than that for CNT. Therefore, GNP might be more suitable relative to CNT as an effective and economical reinforcement material for the development of new-generation MMCs. From the other side, since the discovery of graphene, there have been a considerable number of researches carried out on the reinforcement of a variety of polymers by GNP [9, 10]. However, only limited researches have been done on GNP-reinforced MMCs [11-13]. This can be attributed primarily to a series of troublesome problems, including the difficulty in homogeneous GNP distribution and full densification with metal powders as well as the interfacial problems between the GNP and most metallic matrices compared with polymer-based composites. Even so, early available results demonstrated that the introduction of GNPs into Al or Mg matrix can dramatically improve the matrix mechanical properties [12, 13]. To date, surprisingly no experimental studies, to our knowledge, have involved the preparation and characterization of GNP-reinforced bulk Cu-matrix composites.

In this study, we report the work on examining the feasibility of fabricating bulk Cu-based composites by incorporation of the GNPs. Mechanical ball-milling, which has been confirmed to be very effective in the dispersion of CNTs in MMCs [14, 15], is also applied in attempt to uniformly disperse GNPs in Cu powders. A particular goal is to evaluate the effect of the GNP introduction on the mechanical properties of Cu-matrix composites.

2 Experimental

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

Cu powder (99.9% pure, 15–20 µm) and GNPs were used to produce Cu-based composites. Reagents used were of analytical grade or the highest commercially available purity (∼99.9%). The exfoliated GNPs used in this work were obtained from natural graphite flakes according to the modified Brodie's method [16]. In this method, graphite flakes were introduced into a solution of H2SO4 (98%) under vigorous stirring. Then KCl was added to the mixture slowly over 30 min on the temperature condition of 0 °C. The mixture solution was naturally reached to room temperature with stirring for 96 h. After completion of reaction, the reaction mixture was repeatedly washed and filtered in distilled water to achieve an aqueous mixture solution until the pH of the supernatant reached 7. When the acidic and impurities were removed, the graphite oxide was achieved through sedimentation and finally dried at 60 °C. The GNPs were finally obtained by the thermal reduction of graphite oxide at 1050 °C for 15 min in argon gas atmosphere. A dispersion of GNPs in isopropyl alcohol ((CH3)2CHOH) was conducted by sonication for 2 h to achieve a further exfoliation.

In order to obtain a high dispersion of GNPs in matrix powder, the GNPs with volume fractions of 0, 3, 5, 8, and 12% were mixed with Cu powders by ball-milling [17]. Mixed powders were each placed in a steel vial and milled by a SPEX mixer in a rotary speed of 1200 rpm for 3 h under an argon atmosphere. A ball-to-powder weight ratio of 10:1 was used, and petroleum ether was added as a process control agent. As-milled powders were first compacted to a green density of 75% theoretical density and then consolidated using a hot pressing technique. The compact powders were sintered at 800 °C for 15 min. The heating rate was 50 °C min−1 and a pressure of 40 MPa was applied. After the sintering process, the disk-shaped samples were obtained after getting rid of the graphite felt left on the surface of the composites.

The microstructure of the materials was characterized by optical microscope (OM), field-emission scanning electron microscope (FESEM) equipped with energy-dispersive X-ray (EDX). The phase structure was identified by X-ray diffraction spectrum taken with Cu Kα radiation at room temperature. The size of the GNPs was identified by an atomic force microscope (AFM). The AFM measurements were conducted by deposition of the GNPs on a mica substrate from an aqueous dispersion. Raman spectroscopy was conducted to evaluate the structure change of the GNPs using a Bruker dispersive Raman spectrometer with 532 nm Nd: YAG laser. A maximum laser power of 3 mW was applied on three different points of each sample for 30 s of the accumulation time.

The density (ρ) of the composites was measured by Archimedes' principle. The theoretical densities of Cu (8.96 g cm−3) and GNP (2.2 g cm−3) were used to calculate the relative density of the samples. Tensile test (ASTM D638) was performed using a REGER-3010 apparatus under a crosshead speed of 0.5 mm min−1 at room temperature. The test sample was a dog-bone shape with a gage length of 20 mm and width of 5.5 mm. Tensile stress–strain curves for the reference and composite materials were obtained, and yield strength and Young's modulus values were measured from the curves.

3 Results and discussion

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

Figure 1 shows the microstructure of as-received GNPs. It is clear seen that the GNPs have a relatively layered structure of platelets, demonstrating wrinkled surface texture. Nanoplatelets are stacked on top of each other, which can lead to stacks or aggregates. The shape and microstructures are typical structure of the GNPs, as reported elsewhere [5, 9, 10]. Figure 2 shows the AFM measurements of the GNPs. It is obvious that the GNPs are well exfoliated, corresponding to a particle size range of 1–5 µm and a thickness around 3.5 nm. The large aspect ratio of GNPs is an important factor in enhancing contact area with matrix.

image

Figure 1. SEM image of as-received GNPs.

Download figure to PowerPoint

image

Figure 2. AFM tapping-mode image (a) of GNPs and their size distribution (b).

Download figure to PowerPoint

Nanofiller dispersion is an important issue since GNPs have an inherent tendency to form agglomerates due to strong van der Waals attraction, large surface areas and π–π interaction [18], and therefore a good dispersion is an important factor for mechanical reinforcement. From the high magnification images of the ball-milled powders shown in Fig. 3, it is clear that the mixed powders with 8 vol.% GNPs [Fig. 3(a)] give homogeneously dispersed GNPs and most of GNPs are embedded into the Cu matrix powder rather than on the surface. Nearly no GNP agglomerates can be found. This indicates that the ball-milling process applied in the present work is effective to achieve the uniform mixture of the GNPs and Cu powders. Nonetheless, with increase in GNP concentration the dispersion becomes more challenging, as indicated in Fig. 3(b) for 12 vol.% GNP powders, in which the GNP agglomerates are visible in some regions.

image

Figure 3. SEM images of the microcosmic morphology of ball-milled (a) 8 vol.% GNP/Cu powders and (b) 12 vol.% GNP/Cu powders.

Download figure to PowerPoint

Figure 4 shows the morphology of the ball-milled powders with various GNP contents. It can be seen in Fig. 4(a) that the Cu powders become irregular in shape after ball milling process, combined with increasing in particle size as well. This is due to the overwhelming ductility of Cu powders and possible dynamic recovery processes occurring, causing the Cu powders first crushed by plastic deformation under the impact of the balls and then subsequently cold-welded to form larger particles having rough surfaces [19]. However, unlike the pure Cu powders, GNP–Cu mixture powders show a noticeable size reduction, and this size reduction is enhanced with increasing GNP content, as shown in Fig. 4(b–d). This means that the GNPs can play a beneficial role as grinding aids due to small size and wrinkle structure of GNPs, which can effectively prevent the agglomeration of mixture powders. Such prevention of the particle agglomeration facilitates the homogeneous distribution of the GNPs within the Cu matrix. Figure 4(e) shows the SEM image for the polished surface of the 8 vol.% GNP/Cu composites. The white dot phases represent GNPs as determined by EDS shown in Fig. 4(f). It can be found that the GNPs are still homogeneously dispersed in Cu matrix, which suggests that the uniform dispersion of GNPs in ball-milled powders can be well retained in the post-sintering process.

image

Figure 4. SEM images of the macroscopical morphology (inset shows the particle size distributions) of ball-milled (a) Cu powders; (b) 5 vol.% GNP/Cu powders; (c) 8 vol.% GNP/Cu powders; (d) 12 vol.% GNP/Cu powders; (e) consolidated 8 vol.% GNP/Cu composites. (f) EDS element analysis of the white dot phase marked in (e).

Download figure to PowerPoint

Raman spectroscopy is a powerful nondestructive tool to characterize carbonaceous materials, particularly for distinguishing ordered and disordered crystal structures of carbon. Figure 5 shows the Raman spectra of the pristine GNPs, milled GNP-Cu mixed powders and sintered bulk GNP/Cu composites. As seen, the Raman spectrum of pristine GNPs displays clearly a G peak at 1584 cm−1, a D peak at 1343 cm−1 and a 2D peak at 2695 cm−1, which are the signatures of the graphene like structure [20]. In addition, the Raman spectrum of pristine GNPs shows a broad 2D peak, displaying the feature of multi-layered GNPs while the Raman spectrum of few-layered or bi-layered GNPs shows a sharp 2D peak [20]. This is in line with AFM measurements shown in Fig. 2. In contrast, the Raman spectrum of the GNPs in milled powders and sintered GNP/Cu composites show no noticeable change, indicating that the intrinsic structure of GNP can be well maintained in the processing. From Fig. 5, the ratios of the peak intensity between D band and G band, ID/IG, are calculated, which is used to qualitatively characterize the change of defects in the GNPs [21]. It is observed that the ID/IG ratio of milled GNPs is increased compared to that of pristine GNPs. This is mostly related to the repeated deformation, cold welding and fracturing actions of the ball milling process, which can cause some damage to the surfaces and edges of the GNPs. Nevertheless, almost no further increased GNP defects are detected after sintering process.

image

Figure 5. Raman spectra of the pristine GNPs (bottom), mixed powder of milled GNP/Cu powders (center), and sintered GNP/Cu composites (top).

Download figure to PowerPoint

Figure 6 shows the XRD patterns of the GNP, Cu and GNP/Cu composites. For GNP, the diffraction peaks at 2θ = ∼24.3° and 43.6° are attributed to the graphite-like structure (002) and (100), respectively. For the GNP/Cu composites, only the peaks belonged to Cu are observed, while no GNP peaks are detected. This can be explained by the small loading of GNPs in the sample and the low scattering length of carbon compared to metal atoms [22]. In addition, the GNP/Cu composites reveal a peak broadening upon milling, which suggests a reduction in matrix grain size. Furthermore, no additional peaks for the GNP/Cu composites are identified, indicating that there is no oxides/carbides formation in our GNP/Cu composites.

image

Figure 6. XRD patterns of the GNP, Cu and sintered GNP/Cu composites.

Download figure to PowerPoint

Figure 7 shows the yield strength and Young's modulus of GNP/Cu composites as a function of GNP content. It is obvious that the addition of GNPs causes a significant improvement on the mechanical properties of the Cu-matrix and this improvement is enhanced with increasing GNP content. For instance, a low level of 8 vol.% GNPs added is found to result in a remarkable increase in yield strength and Young's modulus to 114 and 37%, respectively. The significant increase in mechanical properties of GNP/Cu composites is originally attributed to the ultra-high intrinsic strength (∼125 GPa) and elastic modulus (∼1.0 TPa) of the GNP, which is realized by the homogeneous dispersion of GNPs in the Cu matrix. In addition, refining grains may effectively strengthen a material in terms of well-known Hall–Petch relation [23], which shows that the materials can be strengthened by decreasing the grain size. The grain structure observation shown in Fig. 8 displays an average grain size of approximately ∼10 µm for the Cu matrix and ∼4 µm for 8 vol.% GNP/Cu composites. This observation is consistence with previous XRD analysis (Fig. 6), which reveals a clear increase in peak width for the composites or the grain refinement. Similar to CNT/metal composites [24], it is thought that the much finer grain size in GNP/Cu composites is attributed to the effective pinning effect of the nano-sized GNPs on the grain boundaries, in which the dislocation motion could be blocked at the sites of GNPs. Consequently, the dislocation accumulation due to the addition of GNP hinders the growth of the recrystallized grains during the processing, which contributes to the high strength of GNP/Cu composites.

image

Figure 7. Yield strength and Young's modulus of GNP/Cu composites versus GNP volume fraction.

Download figure to PowerPoint

image

Figure 8. Grain structures of (a) Cu-matrix and (b) 8 vol.% GNP/Cu composites.

Download figure to PowerPoint

However, as seen in Fig. 7, with further increasing GNP content up to 12 vol.%, the increments of the yield strength and Young's modulus dramatically reduce to 46 and 24%, respectively. The less effective enhancement in 12 vol.% GNP/Cu composites mainly arises from the GNP aggregations in the ball-milled powders as shown in Fig. 3(b). The GNP agglomerates would form steric obstacles in composites consolidation, restricting matrix materials to flow into the agglomerates and resulting in the formation of pores in the composites. This can be confirmed by the relative density measurements as shown in Fig. 9, in which the relative density of 12 vol.% GNP/Cu composites is 96.4%, showing a large amount of porosity. Subsequently, these pores act as stress concentration sites for plastic instability, weakening the efficiency of the mechanical improvement by incorporation of GNPs.

image

Figure 9. Relative density of GNP/Cu composites versus GNP volume fraction.

Download figure to PowerPoint

In order to profoundly understand the strengthening behavior of GNP/Cu composites, it is necessary to compare the experimental results with theoretical predictions. The Halpin–Tsai model [25, 26] is frequently applied to estimating the modulus of the composites in which the reinforcement spatial distribution and aspect ratio are included. Considering the random orientation or unidirectional distribution of GNPs in the polymer matrix, the Halpin–Tsai equation is given by

  • display math(1)
  • display math(2)
  • display math(3)

where E is the Young's moduli, the subscripts g, c, || refer to the GNP, the composites with randomly oriented and unidirectionally distributed reinforcements, respectively. p and f are the aspect ratio and volume fraction of GNPs in the composites, respectively, in which p is calculated by dividing the GNP length by its GNP thickness. In the calculations, the Young's modulus of the Cu matrix is 76 GPa obtained from the measurement, and that of the GNP is 1.0 TPa provided from the reported results [27]. The average length and thickness of GNPs are 2 µm and 3.5 nm, respectively, obtained from AFM images (Fig. 2). Substituting these parameters into Eqs. (1) and (2), the comparison between our experiments and theoretical calculations are shown in Fig. 10. Although the incorporation of GNP can provide a marked improvement in mechanical performance of the composites, it is obvious that the measurements are evidently lower than the predictions. This implies that there is still some room for further enhancement in mechanical performance of GNP/Cu composites by improvement of processing and quality of GNP. The gap between the predictions and experimental results is attributed to the following reasons.

image

Figure 10. Comparison between experimental data and theoretical calculations of Young's moduli Ec (Eq. (1)) and E|| (Eq. (2)) for GNP/Cu composites as a function of GNP volume fraction.

Download figure to PowerPoint

Reinforcement structural defects are often generated during the ball-milling process, as shown in Fig. 5, which can result in the loss of intrinsic stiffness of the GNPs [28]. Thus the structural defects in GNP possibly reduce the effective stiffness of the GNPs and further decrease the modulus of the composites. Aside from the issue with the defects of GNPs, the reinforcing effectiveness observed from GNP/Cu composites may be limited by problems with interfacial adhesion of GNP-matrix and spatial distribution of GNP. In general, the composite interfacial bonding can be divided into four types of mechanical bonding, physical bonding (van der Waals interactions), diffusion bonding, and reaction bonding, with bonding strength increasing in the same order [29]. In this context, the adhesion between Cu-matrix and GNP is attributed to the type of mechanical bonding or van der Waals interactions due to non-wetting of the GNP and Cu, resulting in an insufficient interfacial bonding. Moreover, it is known that the GNP has ultra-high strength along the in-plane direction but low strength in through-thickness direction, the random distribution of GNPs with various orientations would disturb this unidirectional load transfer mechanism and reduce the strengthen efficiency of the GNPs. It is shown in Fig. 10 that the measurements are significantly lower than the predictions with considering well-aligned GNPs in the composites. Accordingly, further improvements of the mechanical properties could potentially be made through the combined ways including optimizing the milling process to allow a lesser damaging to GNPs, establishing a high interfacial strength between the GNPs and Cu matrix by the addition of matrix-alloying elements to ensure a certain reaction (i.e., carbide formation) at the interface, and conducting the post-processing such as rolling or extrusion to realize the alignment of GNPs.

4 Conclusions

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

Cu-matrix composites reinforced with the GNPs were fabricated by a powder metallurgy route. The 3–8 vol.% GNPs were well dispersed in the Cu matrix by ball milling, and dense composites were also obtained by hot-sintering. The GNP agglomerations were presented in the powder mixture with high GNP content of 12 vol.%, which further inhibited the densification of the composites. Raman study revealed no change in GNP structure but an increase of defects in ball-milled GNPs. XRD analysis indicated a reduction in matrix grain size and no oxides/carbides formation in the composites. Tensile test results showed a maximum 114 and 37% increases in yield strength and Young's modulus with 8 vol.% GNP content, respectively. The microstructure examination indicated that the extraordinary GNP reinforcement of Cu-matrix composites is attributed to the homogeneous dispersion of GNPs and grain refinement. GNP agglomerations induce the presence of large porosity in the composites, which greatly weaken the mechanical improvement by incorporation of GNPs. The obtained mechanical measurements were obviously below the theoretical predictions, which were proposed to be related to the following three factors: the reduced stiffness of GNPs due to the defects introduced in ball milling process, insufficient interfacial bonding due to the non-wetting of GNP and Cu, and disturbed unidirectional load transfer mechanism by random oriented GNPs.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

This study was financially supported by Doctoral Start-up Scientific Research Fund (2011010431).

References

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References