Adhesion and Self-Healing between Monolayer Molybdenum Disulfide and Silicon Oxide

The adhesion interactions of two-dimensional (2D) materials are of importance in developing flexible electronic devices due to relatively large surface forces. Here, we investigated the adhesion properties of large-area monolayer MoS2 grown on silicon oxide by using chemical vapor deposition. Fracture mechanics concepts using double cantilever beam configuration were used to characterize the adhesion interaction between MoS2 and silicon oxide. While the interface between MoS2 and silicon oxide was fractured under displacement control, force-displacement response was recorded. The separation energy, adhesion strength and range of the interactions between MoS2 and silicon oxide were characterized by analytical and numerical analyses. In addition to the fundamental adhesion properties of MoS2, we found that MoS2 monolayers on silicon oxide had self-healing properties, meaning that when the separated MoS2 and silicon oxide were brought into contact, the interface healed. The self-healing property of MoS2 is potentially applicable to the development of new composites or devices using 2D materials.


Monolayer MoS 2 on SiO 2
High-quality monolayer MoS 2 was grown on a SiO 2 strip (8 × 40 mm) by using chemical vapor deposition (CVD). The presence and uniformity of the monolayer MoS 2 was confirmed with photoluminescence (PL) and Raman spectroscopy. Figure S1 shows the PL spectrum of the MoS2 layer on SiO 2 . Clear PL peaks are observed at 1.89 and 2.03 eV, which closely match the peak positions of A 1 (trion) and B1 (exciton) of monolayer MoS2 described in the literature [1]. Figure S1. The photoluminescence spectrum of monolayer MoS2.
In order to prove the homogeneity and uniformity of the CVD-grown MoS 2 sample, we obtained Raman spectra at different positions along the longitudinal direction of the strip ( Figure S2a). As shown in Figure S2b, the distances between E 1 2g and A 1g peaks are 20.

Preparation of the double cantilever beam fracture samples
Once high-quality and monolayer MoS2 was successfully synthesized on a silicon oxide strip, the other silicon oxide strip was partially coated with uncured epoxy (EP30, Master Bond, Inc.) in order to fabricate the double cantilever beam (DCB) specimen. The epoxy coated silicon oxide strip was placed on the MoS 2 surface ( Figure S3a). The DCB specimen (cross section is Si/SiO 2 /MoS 2 /epoxy/SiO 2 /Si) was cured at 100 ºC for two hours ( Figure S3b). The epoxy terminus creates the initial crack of length ( 0 a ) which ranged from 1 cm to 2 cm. Finally, aluminum loading fixtures were bonded to DCB specimens

Monolayer MoS 2 transferred on epoxy
As long as the silicon strip with MoS2 was properly bonded to the second silicon strip with a thin epoxy layer, the MoS 2 was completely transferred to the epoxy layer. In order to confirm the complete transfer of MoS2 onto the epoxy, the transferred MoS2 on the epoxy side was checked with Raman spectroscopy after the fracture test ( Figure S4a). Figure S4b shows the Raman spectra obtained from the growth substrate S1 and the transferred substrate S2. Two peaks for the E 1 2g and A1g are present at 383 and 404 cm -1 for both substrates, respectively. The presence of the E 1 2g and A 1g peaks for the substrate S2 explicitly proves that monolayer of MoS2 on the silicon oxide was successfully transferred to the epoxy surface. Furthermore, there is no strain effect on the MoS 2 layer due to the fact that the characteristic peaks were identical before and after transfer [2].
To check the uniformity of the transferred MoS2, several areas of the substrates S1 and S2 were characterized by Raman mapping of the peak intensity of the A 1g . After transfer, Raman peak intensity (A1g) of the MoS2 at the region (3) of the substrate S1 has almost a trivial value as shown in Figure S4c, which indicates that the MoS 2 was completely detached from SiO 2 . On the contrary, the epoxy regions (5) and (6) of the substrate S2 present clear presence of Raman peak intensity (A 1g ) of the MoS 2 film as shown in Figure S4d. These results prove that the MoS2 was perfectly transferred onto the epoxy during the fracture experiment. positions of the substrates S1 and S2 described in Figure S4a.

Applied displacement profiles
In all the experiments that were conducted, the separation of the silicon strips was controlled by specifying the displacements applied at the point of application of the load. The time history of a general case is shown schematically in Figure S5a. The initial loading is shown in blue, where the applied displacement rate was constant and sufficient to produce separation of the intrinsic interface. Once steady state growth of the crack had occurred, the applied displacement was returned to zero at the same rate as the loading portion. In some cases the loading was reapplied (red) at the same rate but to a higher level, resulting in the separation of any healed region from the previous unloading as well as crack growth along a new portion of the intrinsic interface, whereupon the specimen was again unloaded. As many as four cycles were applied, with increasingly larger maximum applied displacements. The circled numbers correspond to different regions of the interface that were healed or separated, making reference to discussion related to Figure 4.

Traction-separation relations
Interactions during separation are characterized by their fracture energy, strength, and interaction range.
Although simple beam theory is sufficient for characterizing the fracture energy, further details of the interaction must be determined by finite element analysis (FEA). The continuum description of the adhesion interaction between two surfaces is often expressed as the traction-separation relation. Although many functional forms are possible, one of the simplest, the bilinear form ( Figure S6) was selected for this study.  ) is the area underneath the traction-separation relation shown in Figure S6. Thus, this model allowed initiation and growth of interfacial fracture to simulate the peeling of the MoS 2 layer from the silicon strip. Finally, it should be noted that the simulation of unloading was unable to account for the healing that was observed in the experiments. This was due to unexpected complications that remain unresolved at this time. Figure S6. Schematic of a bilinear traction-separation relation.

Roughness of MoS2
The atomic force microscope images (Figures S7a and S7b)

Multi-cycle loads
The initial observation of healing (Figure 3) motivated the examination of the effect of repeated healing and separation using multi-cycle loads (Figures 5 and S8). The force-displacement response in the left column represents the response of i-th load cycle (i =1, 2, 3, and 4), with its corresponding resistance curve in the right column. No matter which cycle load was selected, its fracture resistance curve exhibited the characteristic response shown in Figure 4. Figure S8. The force-displacement and its fracture resistance plots during cyclic tests.