Frictional Anisotropy of CVD Bi-Layer Graphene Correlated with Surface Corrugated Structures

Abstract

Atomically-thin 2D nanomaterials can be easily deformed and have surface corrugations which can influence the frictional characteristics of the 2D nanomaterials. Chemical vapor deposition (CVD) graphene can be grown in a wafer scale, which is suitable as a large-area surface coating film. The CVD growth involves cooling process to room temperature, and the thermal expansion coefficients mismatch between graphene and the metallic substrate induces a compressive strain in graphene, resulting in the surface corrugations such as wrinkles and atomic ripples. Such corrugations can induce the friction anisotropy of graphene, and therefore, accurate imaging of the surface corrugation is significant for better understanding about the friction anisotropy of CVD graphene. In this work, the combinatorial analysis using friction force microscopy (FFM) and transverse shear microscopy (TSM) was implemented to unveil the friction anisotropy of CVD bi-layer graphene. The periodic friction anisotropy of the wrinkles was measured following a sinusoidal curve depending on the angles between the wrinkles and the scanning tip, and the two domains were observed to have the different friction signals due to the different directions of the atomic ripples, which was confirmed by the high-resolution FFM and TSM imaging. In addition, we revealed that the atomic ripples can be easily suppressed by ironing the surface during AFM scans with an appropriate normal force. This work demonstrates that the friction anisotropy of CVD bilayer graphene is well-correlated with the corrugated structures and the local friction anisotropy induced by the atomic ripples can be controllably removed by simple AFM scans.

1. Introduction

Two-dimensional (2D) layered nanomaterials such as graphene, molybdenum disulfide (MoS₂), or tungsten disulfide (WS₂) offer a promising solution in areas of solid-state tribology, due to their excellent mechanical properties as well as low friction. Among a variety of 2D nanomaterials, graphene has been extensively studied as a surface protective coating film and a solid-state lubricant because it is atomically thin and strong, chemically stable under severe environmental conditions [1-3]. Mechanical exfoliated graphene can be fabricated by a simple adhesive taping method and has the outstanding crystal quality with very low structural defects, making it an appropriate candidate for the fundamental study on inherent mechanical surface properties of graphene. Friction force microscopy (FFM) was used to investigate the surface frictional characteristics of graphene [4]. While mechanically exfoliated graphene has atomically flat surface, its atomically-thin layer is susceptible to the out-of-plane deformation by the residual stress during being transferred onto a substrate [5-7], resulting in the formation of the surface corrugations. The recent FFM measurements on the mechanical exfoliated graphene revealed that graphene exhibits the periodic friction anisotropy following the sinusoidal function at the periodicity of 180° depending on the relative scan direction of the FFM tip over graphene surface. The resulting friction anisotropy was found to be due to the presence of atomic ripples.

While mechanically exfoliated graphene has been extensively explored as a model material in order to achieve the fundamental understandings about nanoscale friction of graphene, studies on the frictional characteristics of chemical vapor deposited (CVD) graphene, which can be synthesized in a wafer scale essential for practical applications, is relatively sparse. The friction anisotropy of CVD graphene which was induced by the nanoscale wrinkles has been reported [8], however, due to the complicated surface structures of the CVD graphene, the friction anisotropy originated from the atomic scale ripples has not been fully understood yet. CVD graphene can have various surface corrugation structures due to the compressive strain induced by thermal expansion coefficient (TEC) mismatch between graphene sheet and the growth substrate [9-10]. In addition, surface corrugation can result from the transfer process of CVD graphene to a target substrate such as a silicon substrate with an oxide layer. The complicated surface structures on CVD graphene can influence the frictional characteristics of graphene [8,11]. Moreover, surface corrugations on CVD graphene surfaces can modify the electronic properties of the graphene, the performance of graphene-based electronic device can be affected [12]. Therefore, it is essential to understand the surface structure-property relation of CVD graphene, which can be achieved by accurate surface imaging of structures and properties.

In this study, two types of surface corrugation structures, apparent nanoscale wrinkles and atomic ripples, over CVD bi-layer graphene were investigated using friction force microscopy (FFM) and transverse shear microscopy (TSM). A single crystalline domain of the bi-layer graphene was considered to exclude the effect of crystallographic orientations of graphene when TSM imaging. We conducted the FFM measurements by rotating the sample at a fixed FFM tip scan direction, and the periodic friction anisotropy of the apparent nanoscale wrinkles was unambiguously observed to follow the sinusoidal function at the periodicity of 180°. Moreover, when considering a relatively large area of the bi-layer graphene, two definite domains were observed to have the different friction signals, which was revealed due to the presence of the atomic ripples over seemingly flat graphene areas. High resolution FFM and TSM imaging confirmed that the atomic ripples are formed in the two domains at the two different directions, and the atomic ripple domains on the CVD bi-layer graphene which can dominantly affect the friction anisotropy at a large area can be removed by the repeated AFM scans at contact with increasing the normal load. This work suggests that the surface corrugations at multiple scales such as apparent nanoscale wrinkles and atomic scale ripples can result in the local friction anisotropy of CVD graphene. However, when considering a relatively large area of graphene at contact with asperities of a counter surface in friction, the complicated surface structures can nullify such local friction anisotropy and especially, the repeated friction can lead to the suppression of the atomic ripples making the flat surface of graphene whose friction might be no long anisotropic, but isotropic.

2. Experimental Method

2-1. Sample preparation

Graphene was synthesized by atmospheric pressure chemical vapor deposition (APCVD) method on a copper foil, and CVD-grown hexagonal boron nitride (h-BN) sample was purchased from the commercial vendors, Graphenea Inc. (USA). The graphene sample was transferred onto the SiO₂/Si substrate by a h-BN assisted wet transfer method to obtain a clean graphene surface without polymeric residue for accurate imaging of surface structures[13]. A poly methyl methacrylate (PMMA)/ h-BN layer was used to support the graphene film during the transfer process so that graphene film is not directly exposed to the PMMA.

2-2. AFM measurements

To explore the friction anisotropy of graphene, we conducted the FFM and TSM measurements using an atomic force microscope (AFM) (XE-7, Park’s system Inc., Korea) with a Si tip (PPP-LFMR, Park’s system Inc., Korea). All the measurements were conducted in ambient conditions (20-30º C, 30~40% relative humidity) and with a constant scan rate of 1 Hz. The FFM and TSM images were obtained by subtracting the forward and backward directions scans and dividing by two to extract the friction and shear signals. To observe the friction anisotropy of graphene, both the FFM and TSM measurements were conducted by rotating the graphene sample in a clockwise and the rotation of the sample started from when the directional wrinkles, formed during the CVD growth process of graphene, are parallel to the FFM tip scan direction which is in a fixed direction along the x-axis. Due to the fixed FFM tip scan direction, as the sample is rotated, the angle between the directional wrinkle and the FFM tip scan direction changes. To describe the angle at which the sample is rotated, the “relative angle” is defined as the angle between the FFM tip scan direction and the directional wrinkle.

3. Results

Figure 1 shows the FFM measurement results measured at a nominal normal load of 6 nN, demonstrating the local friction anisotropy of apparent nanoscale wrinkles. Figures 1(a) and 1(b) show the friction images of the single-layer graphene and bilayer graphene surfaces for the relative angle of 0° and 90º , respectively. The layer number was confirmed by Raman spectroscopy (Fig. 1(c)).

Fig. 1. FFM images of the single-layer and bi-layer graphene surfaces in a single crystalline graphene island at the relative angles of (a) 0° and (b) 90°. (c) Raman spectra of the single-layer and bi-layer graphene areas. (d) The plot of the friction signals for the directional wrinkles indicated in the inset of the FFM image, depending on the relative angle from 0° to 180°, which shows the periodic friction anisotropy.

The Raman spectrum of the single-layer graphene area shows sharp 2D peak with the I_2d/I_G ratio of ~2.9, indicating the single-layer graphene, whereas that of the bi-layer graphene area shows the broad 2D peak with the I_2d/I_G ratio of ~1.03. The difference in the friction signals between single-layer and bi-layer graphene can be observed due to their different bending stiffness finally resulting in the different resistance against atomic puckering by FFM tip scans at contact [14]. At the relative angle of 0°, the directional wrinkles cannot be imaged since the FFM tip scans on the top of the wrinkles and no difference in the friction signals occurs between the directional wrinkles and the flat surface of graphene. In contrast, the FFM image at 90° clearly shows the directional wrinkles. In this case, the FFM tip scans in perpendicular to the wrinkles and the tip collides with the side of the wrinkle structure, resulting in a large contact area and high frictional signal, which is the mechanism to generate the friction anisotropy of the wrinkles depending on the relative angles between the tip scan direction and the wrinkles[8]. To investigate the friction anisotropy of these apparent directional wrinkles, the sample was rotated from 0° to 180° with the interval of about 15° . The friction anisotropy of the wrinkles follows the sinusoidal function, |sin θ|, with the periodicity of 180° . That is, the magnitude of the friction signal is maximum at the relative angle of 90º , and it is minimum at 0° and 180° [5-7]. Figure 1(d) shows the results of the FFM measurements, and it is confirmed that the friction anisotropy of the directional wrinkles exhibits the 180° periodicity following the sinusoidal curve as expected.

To investigate the friction anisotropy in a large domain of the bi-layer graphene, the repeated FFM measurements were conducted by rotating the sample in a clockwise at the interval of about 15° . Fig. 2 shows the FFM images measured at the different relative angle with a constant normal load of 6 nN. Depending on the relative angle, the specific domain in the bi-layer graphene region (yellow dashed area) shows the different friction signal from the surrounding bi-layer area. At the relative angles of 0° and 90º , no observable difference in the friction signals was found over the bi-layer region (Figure 2(a) and 2(c)). However, at the relative angles of 27° and 137° , the yellow dashed area shows the different friction signal from the surrounding bi-layer area where the lower friction signal was measured at the relative angle of 27° and the higher friction signal was measured at the relative angle of 137° . This result suggests that the two domains (yellow dashed area vs. its surrounding) can have the different atomic ripple directions.

Fig. 2. The FFM images of the single and bi-layer graphene surfaces at the different relative angle of the sample to the FFM scan direction, which highlights the presence of the domain friction anisotropy on the bi-layer graphene area.

To confirm the different atomic ripple directions of the two domains, the high-resolution FFM and TSM measurements were conducted on two spots, the boundaries of the two domains which show the different friction signals, as shown in Fig. 3. TSM enables to image atomic ripple domains as well as crystallographic directions of materials due to its higher sensitivity than FFM, by measuring the torsional deflection of the cantilever tip when the tip scans the surface in the direction parallel to the cantilever axis. The TSM measurements can provide the direction of the cantilever tip torsional deflection depending on atomic ripple orientations, which is a useful information, in addition to the magnitude of the deflection, while FFM imaging can provide only the information for the magnitude of the lateral deflection[5,15,16]. Fig. 3(a) and 3(b) are the FFM and TSM images of the two different spots, respectively, at the relative angle of 0° . The FFM images show no observable differences between the two domains, but the two domains can be clearly identified in the TSM images and the sign of the TSM signals in each domain is opposite. This observation suggests that the two domains have different atomic ripple orientations, but at the same angle relative to the FFM scan direction[6]. The positive sign of the TSM signal in a green dashed area indicates that the ripples are aligned in the direction passing through the 2nd and 4th quadrants, and the other domain has the ripples aligned in the direction across the 1st and 3rd quadrants as indicated by the negative sign of the TSM signal.

Fig. 3. High-resolution FFM (left) and TSM (right) images on the two spots, (a) spot 1 and (b) spot 2, at the boundaries of the two domains having different directions of the atomic ripples in Figure 3.

Even though the domain friction anisotropy in a large area of the bi-layer graphene can result from the atomic ripples aligned in the different directions, it can be easily disappeared by mechanically suppressing the atomic ripples simply using contact-mode AFM scans at the sufficient normal load. Figures 4(a) and 4(b) show the change of the FFM and TSM images by applying the different normal loads of the AFM tip when scanning, respectively. At the normal load of 0.1 nN, the spatial variation of the FFM signals can be clearly observed, but when increasing the normal load, the FFM signals become uniform over the graphene surface. The corresponding TSM images in Figure 4(b) confirm this observation more clearly. The TSM images show the atomic ripple domains with the different directions at the load of 0.1 nN, and similar to the FFM images, the variation of the TSM signals over the scanned area is reduced when increasing the normal load. To provide the change of the TSM signals more unambiguously, the line profiles of the TSM signals on the areas indicated as the two green dashed lines (denoted as profiles ‘1’ and ‘2’) are plotted in Figures 4(c) and 4(d), respectively. In the TSM image measured at the normal load of 0.1nN, along the green dashed line profile, the TSM signal shows the visible contrast (high TSM signal and low TSM signal region). We defined the high and low TSM signal regions as “A” and “B” denoted in Figures 4(c) and 4(d), respectively, and the clear TSM signal difference across the regions of A and B can be observed in both line profiles. But the TSM signals become uniform as the normal load increases, indicating that the atomic ripples are suppressed by the AFM tip scans at the sufficiently high normal loads. Through the change of the FFM and TSM signals depending on the normal load, we can conclude that the domain friction anisotropy of graphene can be eliminated by ironing the atomic ripples during continuous frictional contacts.

Fig. 4. (a) FFM and (b) TSM images measured by varying the normal load of the tip scanning on the graphene surface from 0.1 nN to 5 nN, and the corresponding TSM signals plotted following the two green dashed lines, (c) profile 1 and (d) profile 2.

4. Conclusion

In conclusion, we unveiled the friction anisotropy of the CVD bi-layer graphene having complicated surface corrugated structures, such as apparent nanoscale wrinkles and atomic ripples, by the combinatorial analysis of the FFM and TSM imaging. The repeated FFM and TSM measurements by rotating the graphene sample revealed the local friction anisotropy of CVD bi-layer graphene by the apparent nanoscale wrinkle structures and by the presence of atomic ripple domains aligned in the different directions over the seemingly flat graphene surfaces. In addition, we experimentally verified that the atomic ripples can be easily and selectively suppressed simply by AFM tip scans at the contact mode with the sufficient normal load applied. This work provides the robust understanding about the friction anisotropy of CVD bi-layer graphene at the multiple scales correlated with the surface corrugated structures, which can be foundational to explore a wafer-scale CVD graphene as a tribological coating film for realizing superlubricity.