Following the discovery of graphene prepared by the micromechanical exfoliation technique [
1], much research intension has been paid on the two-dimensional (2D) [
2–
7] material due to its superior characteristic properties, including excellent physical [
8,
9], mechanical [
10,
11], chemical [
12,
13], electronic and optoelectronic properties [
14–
17]. One of the intriguing advantages of the 2D layered material is that their isotropic properties can be converted to be anisotropic, which results from the breaking symmetry owing to their intrinsic puckering behaviors. The puckering effect of 2D materials is due to the out-of-plane elastic deformation and the bending stiffness, and have attracted wide research interest [
18,
19]. The anisotropies of 2D material exerted powerful influences on the ultrathin flexible nano-electromechanical and thermal devices [
20].
Progress in the general understanding of isotropic to anisotropic relationships in 2D ultra-thin films requires experimental tools capable of imaging the details. Nanome-chanical deformation has been extensively investigated by the atomic force microscopy (AFM) indentation experiments on suspended films [
21–
23]. Based on AFM platform [
24–
27], friction force microscopy (FFM) and transverse shear microscopy (TSM), as the derivative mode of contact AFM [
28–
32], now have been widely used to study the elastic deformation [
19,
31], crystal structure [
28], surficial ripples of 2D materials [
8,
34,
35] and so on. Such studies have revealed an important phenomenon in which nanomechanical contact behavior could be dependent on the number of layers and substrates. While this layer-dependent transient phenomena in 2D materials on different supporting substrates has not been experimentally investigated.
Here, we focused experimental investigation on the surface layer-dependent puckering behavior of 2D materials through the extensive TSM characterization for their crystallographic orientation imaging. First, we directly verify that the puckering effect mainly occurs on the top layer of 2D systems through the exclusive crystallographic orientations of the top layer of MoS2 on the SiO2/Si substrate. Second, the growth position of the adlayer graphene was definitely determined below the first layer during the CVD growth process by a combination of TSM and scanning Kelvin probe microscopy (SKPM) measurements. Finally, the TSM method was further generalized to realize the crystallographic orientation imaging of MoS2 monolayer grown on the crystallized Al2O3 substrate.
All of the AFM measurements were conducted using an Asylum research AFM system (MFP-3D infinity). Both FFM and TSM are the derivative modes of contact AFM mode, as illustrated in Fig. 1. For the 2D thin films, during the contact AFM scanning, the film shows puckering effect around the moving AFM tip, as shown in Fig. 1(b). The puckered geometry induces the film to vertical relax at the front edge of the moving tip and simultaneously stretch the rear region of the moving tip. For FFM, the scan direction of the AFM tip is perpendicular to the long axis of the AFM cantilever, and then the friction property of the sample was obtained by detecting the torsion signal of the cantilever. For TSM, the scan direction of the AFM tip is parallel to the long axis of the cantilever, and then the shear property of the sample was obtained by detecting the torsion signal of the cantilever. For the 2D materials, the FFM measurements have shown that the friction force increases with the decreased thin film thickness due to the enhanced puckering effect. This puckering effect also enabled the TSM method to successfully image the orientation of the crystalline domains in star-shape MoS
2 monolayer flakes in our previously reported work [
27]. While until now, it is still not clear about the toplayer-dependent puckering effect, and more experimental work should be performed to deeply understand this unique and typical nanomechanical behavior of 2D materials.
First, the atomically thin CVD grown MoS
2 monolayer flakes were characterized by FFM and TSM imaging, as shown in Fig. 1. No crystalline differentiated contrast was observed in both optical and AFM topography images of MoS
2 monolayer flakes on the SiO
2/Si substrate [
7]. Strain can induce the transformation from isotropy to anisotropy of both in-plane normal tension stress and shear stiffness, but shear stiffness exhibits much greater anisotropy [
31]. Here, even enhanced by the puckering effect (strain applied by probe), the FFM signals still do not reveal any significant contrast dependence on the crystallographic orientations on the grains due to their almost isotropic stress-strain mechanical behavior. While the clearly crystalline grains were clearly resolved in the TSM images through the anisotropic shear-strain mechanical behavior. The reason for the difference of imaging contrast between FFM and TSM was discussed in our previous work [
31].
In order to investigate the toplayer-dependent puckering effect, we further perform the systematic AFM characterization on the bilayer MoS2 flake made of two bottom layer grains and two top layer grains, as shown in Fig. 2. Due to the van der Walls (vdW) interactions between the top and bottom layer, the adhesion and puckering effect will most occur between the AFM tip and the top layer when the AFM tip approach, contact and move on the top layer [Figs. 2(a) and (b)]. Figure 2(c) and (d) show the optical and AFM topography images of the bilayer MoS2 flakes. The bilayer areas can be clearly resolved and were marked by the white dashed lines.
Figure 2(e) shows the FFM images of this bilayer MoS
2 flake. No crystalline differentiated contrast was observed on both top and bottom layer grains. While it is noted that the friction signal on the bilayer area is slightly lower on the monolayer area, as shown in Fig. 2(g), which is in consistent with the previously reports in 2D materials [
30,
38–
41]. Figure 2(f) shows the TSM images of this bilayer MoS
2 flake. Both the top layer grains (P
21 and P
22) and the bottom layer grains (P
11 and P
12) were clearly resolved in the TSM image. The P
21 and P
11 grains are in a near-pristine 2H stacking. The top layer grain of P
22 shows almost the identical shear signal with the bottom layer grain of P
12 due to their pristine 2H stacking with the same but inversion-asymmetry crystallographic orientation. Since the vdW interaction between the top and the bottom MoS
2 layer is less than the bonding between bottom MoS
2 layer and substrate, the difference between shear signals (the trace and retrace) of top layer is larger than bottom layer [Fig. 2(h)]. It is noted that the top layer grain of P
21 (and P
22) is stacked over both the bottom layer grains of P
11(P
12), while no shear contrast was observed within the grain of P
21 and P
22. This result directly demonstrates that the puckering effect mostly happened on the top layer, then the corresponding puckering-induced shear signal in the bottom grains should be very small in contrast to the top layer grain.
It is necessary to further investigate whether the anisotropic domains of the top layer grains depend on the crystallographic orientation in the TSM images. In Figs. 3(a) and (b), the crystallographic orientation (armchair and zigzag) of evert grain could be determined according to their primitive growth zigzag edges. By systematic TSM characterization with a series of clockwise rotations Fig. 3(c)], the periods of 60◦of the shear signals with the rotation angle are presented [Fig. 3(d)], which is consistent with the monolayer MoS
2 in our previous work [
31]. There is an interesting phenomenon that a singularity (a jump in value) exists in the zz orientation [~60◦, 120◦and 180◦]. This phenomenon can be explained by shear stiffness anisotropy along zz orientation and AC orientation. Under the same transverse deformation, a greater restoring force results in increasing lateral torsion of the cantilever near the zz direction. The amorphous SiO
2/Si does not show any obvious dependency on the rotation angle. These results further confirm our ideas on the toplayer-pendent crystallographic orientation imaging, also demonstrate the general application of TSM to determine the crystallographic orientations of 2D materials, especially in the twisted bilayer or few-layer systems.
The toplayer-dependent TSM characterization can also be used to determine the special growth mode of bilayer graphene in combination with the SKPM measurements, as shown in Fig. 4. Figures 4(a) and (b) schematically show that the adhesion/puckering effect between the static/moving AFM tip and 2D layers are mostly emerge in the contacted top layer graphene. Figures 4(c) and (f) shows the AFM topography images of the bilayer graphene films transferred on the SiO
2/Si substrates (See details in supporting information). It is clearly that the bilayer graphene regions were not resolved in the topography images. Figures 4(d) and (g) show the corresponding surface potential of the bilayer graphene films obtained by the SKPM measurements. The bilayer and monolayer graphene regions were clearly distinguished by their different surface potentials and in good agreement with previously reported work [
36]. To further determine whether the second growth layer is underlying the first layer or not, we perform the TSM characterization of the graphene films, as shown in Figs. 4(e) and (h). The crystalline domains of the top layer graphene were clearly observed, while no bilayer regions marked by the white dashed lines were resolved. We can directly conclude that the second graphene layer is underlying the first layer, confirm the underlying growth model of bilayer graphene [
36].
The previous TSM measurements of 2D materials were performed on the amorphous SiO2/Si substrate. We found that amorphous substrate does not contribute any shear signals during the TSM imaging. While for the 2D materials on the crystalline substrates, whether the TSM method is still applicable or not, should be further investigated. The MoS2 monolayer films directly grown on the crystalline Al2O3 substrate were further characterized by the TSM measurements, as shown in Fig. 5. The substrate terraces with straight edges were resolved in the AFM deflection images, indicating the crystalline nature of the Al2O3 substrate. No crystalline MoS2 domains were resolved in the AFM topography and deflection images, but in the TSM images. Then we can conclude that the underlying crystalline substrate does not apparently contribute to the TSM shear signal comparing to the 2D layers with the puckering effect, and the TSM characterization can be generalized to 2D materials on the crystalline substrates.
In summary, we have identified the toplayer-dependent crystallographic orientation imaging of 2D materials, which enable the crystallographic orientation imaging of bilayer films with transverse shear microscopy. We experimentally demonstrate that only the top layer MoS2 was crystallographic orientation imaged with TSM, which will be useful in the twist bilayer systems. The unique intercalation growth mode of the second layer graphene during the CVD growth process was confirmed with the toplayer-dependent TSM characterization and SKPM measurements. The crystallographic orientation imaging of MoS 2 monolayer on the crystalline Al2O3 substrate was further realized and demonstrated the general applications of the TSM method. Our research will be beneficial in understanding the nanomechanical behaviors of 2D systems and provides a convenient and powerful approach to facilitate the nondestructive crystallographic orientation characterization of various 2D atomic crystal systems.