Applied Science and Convergence Technology 2021; 30(2): 45-49
Published online March 31, 2021
Copyright © The Korean Vacuum Society.
Soo Ho Choia, Ji Hoon Choib, Chang Seok Ohb, Gyeongtak Hanb, Hu Young Jeongc, Young-Min Kima,b, Soo Min Kimd,∗, and Ki Kang Kima,b,∗
aCenter for Integrated Nanostructure Physics (CINAP), Institute for Basic Science (IBS), Sungkyunkwan University, Suwon 16419, Republic of Korea
bDepartment of Energy Science, Sungkyunkwan University, Suwon 16419, Republic of Korea
cUNIST Central Research Facilities and Department of Materials Science and Engineering, UNIST, Ulsan 44919, Republic of Korea
dDepartment of Chemistry, Sookmyoung Women’s University, Seoul 14072, Republic of Korea
Herein, we report on a novel method for transferring two-dimensional (2D) materials grown on Au substrates using sulfur intercalation between the 2D materials and the Au surfaces. The strong nature of the S–Au bond allows intercalation of sulfur atoms into their interface, under a sulfur-rich atmosphere, at 600 °C. The relaxed interfacial interaction achieved via intercalation is carefully confirmed by recovering phonon mode and work function of tungsten disulfide (WS2) in Raman spectra and Kelvin probe force microscopy, and, more importantly, by observing the expansion of the interfacial distance, from 0.24 to 0.44 nm, using cross-sectional transmission electron microscopy. The released interactions facilitate delamination of WS2 from the Au surface, using an electrochemical bubbling method. The resultant Au foil then is reused for repeated WS2 growth. The successful transfer of other 2D materials, including molybdenum disulfide and hexagonal boron nitride, is also demonstrated. Our strategy advances the use of Au substrates for growing wafer-scale 2D monolayers.
Keywords: Two-dimensional materials, Transition metal dichalcogenides, Chemical vapor deposition, Gold, Intercalation
Interest in two-dimensional (2D), van der Waals (vdW) layered materials, including graphene (Gr), transition metal dichalcogenides (TMDC), and hexagonal boron nitride (hBN), has recently increased, owing to their unusual physical and chemical properties [1–6]. To uti-lize those emerging materials into industrial applications, wafer-scaled growth of 2D materials is highly desired. Among various growth tech-niques including molecular beam epitaxy, atomic layer deposition, puls -ed layer deposition, and chemical vapor deposition (CVD) [7–10], CVD method has been mostly employed, due to its several advantages compared to other growth techniques, which include mass produc-tion, cost-effectiveness, and low temperature operation.
The substrate is most important key ingredient for the CVD growth of 2D materials to control morphology, orientation, and growth modes [10–22]. While chemically stable insulating substrates such as SiO2, Al2O3, glass, and quartz enable to grow the various poly-crystalline TMDCs film, the lack of catalytic activity impedes the synthesis of Gr and hBN film. On the other hand, the catalytic metal substrates such as Cu, Ni, and Fe have been widely exploited for the growth of Gr and hBN, but the high chemical reactivity with chalcogens and transition metals to form alloy and compound, prohibits to use as the growth substrate for TMDCs. To date, only Au metal can serve as univer-sal growth substrate for most of 2D materials owing to possessing the unique characteristics of high chemical inertness (that is, no alloy and compound formation), scarce B, N, S, and transition metal solubilities in bulk Au, and its ability to act as a catalyst for activating precursors.
Recently, wafer-scale, single-crystal (SC) 2D films such as hBN, WS2, MoS2, WSe2, MoSe2/WSe2 lateral heterostructure, and W1−xMox-S2 alloy have been successfully synthesized on Au foil, using the self-collimation and epitaxial growth methods [16, 20]. When using SC 2D materials in practical device applications, it inevitably becomes necessary to transfer 2D film onto other substrates. CVD-grown 2D materials are typically transferred after wet-etching the growth sub-strate. For example, Cu foils used for graphene growth have typically been etched away in etchant solution , while for more costly Au foils, non-destructive methods, such as electrochemical bubble trans-fers, have been more popular than wet-etching process. A few studies have reported transferring 2D material grown on Au foils via electro-chemical bubbling [16,18], although it was found that the strong Au–S interaction reduced the efficiency with which 2D materials could be detached from the Au surface during electrochemical bubbling, lead-ing to poor transfer yields . Despite this, it appears that relaxing this strong interfacial interaction to attain a higher transfer yield has not been studied.
Here, we report a universal transfer method for 2D material grown on Au foil, using S-intercalation into the interface between the 2D ma-terial and the Au surface. The S-intercalation was carried out by an-nealing in an S-rich atmosphere, at 600 °C. The interfacial distance between WS2 and the Au surface had increased, from 0.24 to 0.44 nm, reducing the interfacial interaction. Such relaxed interaction restored the WS2 electronic structure toward its undoped state, and, more im-portantly, allowed easy WS2 film delamination from the Au surface, using electrochemical bubbling. Eventually, the cm-scale WS2 film was successfully transferred onto a SiO2 substrate. Importantly, we have also shown that the Au foils could be reused for repeated WS2 growth cycles, and have demonstrated the effective transfers of mono-layer MoS2 and hBN grown on Au substrates.
1 × 1 cm2, high-purity Au foils (0.2 mm thick, 99.99 %, iNexus Inc.) were cleaned using ultra-sonication in acetone and isopropyl alcohol respectively, for 30 s. The surface impurities on Au substrates were then removed by dipping into Au etchant (GE-8111, Transene), for 10 min. The etchant residue was rinsed away using deionized (DI) water, and the Au foils were then further annealed, at 1000 °C for 1 h, under Ar and H2 atmospheres with flow rates of 1000 and 50 sccm, respec-tively. To obtain smooth surfaces, the Au foils were melted onto a W foil (0.1-mm thick, 99.95 %, Alfa Aesar), at 1080 °C for 20 min, under the same conditions. All the experiments in this work were conducted at atmospheric pressure.
WS2 monolayer growth was carried out through sulfurization of tungsten precursor-coated Au substrates . The W precursor was prepared by dissolving 2 wt% sodium tungstate dihydrate in acetylace-tone. The prepared solution was spun onto an Au substrate, at 2500 rpm for 60 s, and an ammonium sulfide [(NH4)2S] solution, acting as a sulfur precursor, was supplied, using a bubbler system . The precursor-coated Au substrate was then loaded into a 2-inch quartz tube—which had been purged for 15 min using high-purity (99.9999 %) Ar, with a flow rate of 350 sccm—and placed into a furnace. The temperature of the furnace containing the quartz tube was increased to 800 °C, over 10 min, and was then maintained at this temperature for 15 min, for WS2 growth. This growth took place under an H2 and (NH4)2S atmosphere, supplied at flow rates of 5 and 20 sccm, respec-tively. After the growth process, the quartz tube was naturally cooled to room temperature. Monolayer MoS2 flakes were grown under con-ditions similar to those applied for WS2 growth, except that sodium molybdate was used as the Mo precursor.
The hBN monolayer growth method was described in our previous report . Briefly, Au foil stacked on W foil was mounted into a 1-inch quartz tube in a furnace. The temperature was elevated to 1100 °C, under Ar and H2 with flow rates of 500 and 40 sccm, to establish liquid Au on the W foil, and then borazine was supplied (as the hBN precursor) for 10 min, at the flow rate of 0.4 sccm. After hBN grains had been grown, the furnace was rapidly cooled to room temperature.
A two-zone furnace system was used for the sulfur intercalation process. The upstream and downstream zones were used to vaporize sulfur powder, and to provide the thermal energy for sulfur interca-lation, respectively. Then, 1 g of sulfur powder, contained in a quartz boat, and as-grown 2D materials on Au substrate, were loaded into the center of each zone. To avoid any damage to the 2D materials during S-intercalation, the quartz tube was completely purged initially, by supplying Ar gas with a flow rate of 500 sccm, for 30 min. The tem-peratures in the two zones were increased to 340 and 600 °C for 15 min, and then maintained at these levels for 30 min. After finishing S-intercalation, the temperatures were allowed to lower naturally, to room temperature. The entire process was conducted under Ar and H2 atmospheres, at flow rates of 300 and 10 sccm, respectively, and at atmospheric pressure.
The poly(methyl methacrylate) (A9 PMMA, MicroChem) layer was spun as the supporting layer onto the 2D material, at 4000 rpm for 60 s, and then dried in a 200 °C oven for 5 min. For an electrochemical transfer, the sample and Pt foil were connected to cathode and anode, respectively, in a power supply . The PMMA/2D material layer was delaminated from the Au surface using generated H2 bubbles un-der an applied voltage of 3–10 V, in 1 M NaOH electrolyte. NaOH residues underneath the PMMA/2D material layer were removed by floating on DI water several times, before, the PMMA/2D material layer was transferred onto a SiO2/Si substrate. To reuse Au substrates, they were cleaned—by dipping in piranha solution for 3–5 h, 10 % nitric acid for 30 min, and buffer oxide etchant for 30 min—prior to being reused.
Overall morphologies of the TMDC and hBN monolayers on Au foil were characterized by field emission scanning electron microscopy (FE-SEM, JSM-7100F, JEOL) and optical microscopy (Eclipse LV150, Nikon). The cross-sectional structures and interlayer distance between WS2 and Au surface were analyzed using an aberration-corrected scan-ning transmission electron microscopy (STEM, JEM-ARM 200CF, JEOL), on the samples prepared by a focused ion beam (FIB, FEI He-lios NanoLab 450) milling and lift-off process. Modulations of the in-terfacial interaction and the electronic structure by the S-intercalation were confirmed using Raman spectroscopy with a laser wavelength of 532 nm (XperRAM 100, Nanobase), and a Kelvin probe force micro-scope (KPFM, N8-NEOS, Bruker), equipped with Pt-coated atomic force microscopy (AFM) probes (Multi-75E, Budget Sensors).
To intercalated S atoms, the annealing of as-grown WS2 on Au foil was conducted in an S-rich atmosphere, at 600 °C, using CVD to evaporate the S atoms [Fig. 1(a)]. To study the effect of S-intercalation on WS2, samples were characterized using Raman spectroscopy. While the confocal Raman mapping image of as-grown WS2 samples showed the solid intensity at the triangular WS2 grain edges [Fig. 1(b)], strong Raman intensities were detected at both the inner and edge regions of the WS2 grains after S-intercalation [Fig. 1(c)]. The representative center region Raman spectra [I and II in Fig. 1(d)] showed significant Raman intensity enhancement, whereas only marginal change was ob-served at the edge regions [III and IV in Fig. 1(e)]. Such strong Raman intensity was attributed to the relaxed interfacial interaction between WS2 and the Au surface. We noted that the optimized intercalation temperature ranged from 500 to 600 °C (not shown here). WS2 grains were found to be damaged at 700 °C, whereas the intercalation was found to be inefficient at 400 °C. We also found that the evaporation of S powders was more effective than H2S or (NH4)2S (not shown here).
To evaluate the relaxed interfacial interaction, the interfacial dis-tance was measured directly, using cross-sectional, annular dark-field, scanning transmission electron microscopy (ADF-STEM). In the cross-sectional ADF-STEM image, we were able to distinguish the WS2 layer from the Au surface of as-grown WS2 quite clearly, as shown in Fig. 2(a). After S-intercalation, the interfacial distance had increased, from 0.24 to 0.44 nm [Fig. 2(b)]. This distance was further confirmed using the energy-dispersive X-ray spectroscopy (EDX) mapping technique. ADF-STEM and corresponding EDX mapping images of WS2 samples before and after S-intercalation showed the presence of W and Au el-ements [Figs. 2(c)–(f)]. We found it difficult to identify the presence of elemental S in the EDX mapping imagery, due to the Au Mα and S Kα energy overlaps near 2.2 eV. We did find, however, that the EDX line profiles along the white arrows in Figs. 2(e) and 2(f) confirmed that the interlayer distances between W and Au atoms were similar to those directly measured in the cross-sectional ADF-STEM images.
To investigate the electronic structure modulation achieved by the S-intercalation, samples were further characterized using KPFM. The KPFM image and its line profile at the center region of a WS2 grain after S-intercalation, yielded a contact potential difference (ΔVCPD) that was lower, by 76 mV, than that of the Au surface (Figs. 3(c) and 3(d)), whereas ΔVCPD was higher, by 32 mV, in as-grown WS₂ samples (Figs. 3(a) and 3(b)). The lower WS2 ΔVCPD value indicated a higher work function for WS2 than that seen for Au.
A WS2/Au energy band diagram has been presented, as Fig. 3(e), to explain the WS2 electronic structure modification—achieved by S-intercalation—which we observed in the KPFM results. Work func-tions of 4.82 eV (for Au) and 4.9 eV (for WS2) were selected from the previous literatures [24, 25]. At the WS2 contact with Au, in an as-grown sample, electrons were preferentially transferred from Au to WS2, due to the lower Au work function, with the result that n-doping into WS2 occurred. After S-intercalation, however, electrons could not be transferred to WS2, due to the increased interfacial distance. It is acknowledged that the electronic structure of intercalated S layer has not been revealed so far. Nevertheless, our de-doping hypothesis was supported by the KPFM results.
Finally, electrochemical bubbling transfer was conducted, after S-intercalation. The optical and corresponding AFM images of WS2 grains transferred onto the SiO2/Si substrate show the complete trans-fer of WS2 grains over the whole region [Figs. 4(a) and 4(b)], regard-less of growth conditions. The characteristic Raman phonon mode and strong negative trion (X−, at ~1.96 eV) emissions can be seen clearly in Figs. 4(c) and 4(d) . Furthermore, 1 × 1 cm2-sized WS2 films were successfully transferred onto a SiO2 surface (Fig. 4(e) and inset). Interestingly, only edge regions of triangular WS2 grains were transferred without S-intercalation, as seen in Figs. 4(f) and 4(g), with the center regions remaining on the Au surface, as seen in the inset of Fig. 4(f). These results strongly supported the concept that weaken-ing the interfacial interaction was a crucial factor in achieving a high transfer yield. Furthermore, the confocal Raman mapping images in Figs. 1(b) and 1(c) can be used as one of important indicators in deter-mining the transfer yield—that is, the regions where the WS2 Raman intensity was strong, were easily delaminated from the Au surface. In addition, after the WS2 had been transferred, the Au substrate was reused for further growth processes, as shown in Fig. 5. The surface morphology remained virtually identical to a pristine Au surface and was readily available for further WS2 growth.
To demonstrate the universal transfer of 2D materials grown on Au surfaces, molybdenum disulfide (MoS2) and hBN were also trans-ferred, after S-intercalation. The enhancements of E12g and A1g phonon modes for MoS2 after S-intercalation can be seen quite clearly in the MoS2 confocal Raman mapping images shown in Figs. 6(a)–(c), in consistent with the change of Raman intensity seen for WS2. AFM image of MoS2 flakes transferred onto a SiO2/Si substrate shows the wrinkles on the MoS2 flakes, demonstrating successful MoS2 trans-fer, as seen in Fig. 6(d). Extracted height profiles, combined with the phonon energy differences (ω) of ~19 cm−1 in the Raman spectrum, were further evidence of the successful transfer of a MoS2 flake mono-layer [Figs. 6(e) and 6(f)] .
For hBN, the optical image of as-grown hBN on Au foil shows no noticeable hBN flakes, while they are precisely visible after the S-intercalation process [Figs. 7(a) and 7(b)]. The surface potential dif-ference between as-grown hBN flakes and Au was found to be small (~ 18 mV) in a pristine sample. This became larger (~ 70 mV) after S-intercalation [Figs. 7(c) and 7(e)]. This variation might be ascribed to the change in charge transfer behavior between hBN and the Au surface, brought about by the presence of an interfacial S layer. Af-ter a transfer of hBN flakes, the monolayer thickness of hBN and its phonon mode of E2g near 1372 cm−1 were characterized, with the re-sults shown in Figs. 7(f) and 7(g) .
In summary, we have successfully developed a method for the universal transfer of 2D materials grown on Au foil, using S-intercalation at the interface between the 2D materials and the Au surface. The S-intercalation extended the interfacial distance, as was conclusively confirmed using cross-sectional TEM. Raman spectroscopy, and KPFM provided further evidence that the S layer blocked charge transfers be-tween the 2D materials and the Au surface. This releasing interaction enabled the successful transfer of cm-scale, monolayer WS2 film. We anticipate that our novel transfer method will contribute to the use of SC 2D monolayers grown on Au foil in industrial applications.
This research was supported by the Institute for Basic Science (IBS-R011-D1), and by the Basic Research Program of the National Re-search Foundation of Korea (NRF), funded by the Ministry of Science, ICT & Future Planning (Grant Nos 2018R1A2B2002302, 2020R1A2-C1006207, and 2020R1A4A3079710). H. Y. J. acknowledges support from the Creative Materials Discovery Program through the National Research Foundation of Korea (NRF-2016M3D1A1900035). S. M. K. acknowledges support by Samsung Research Funding Incubation Cen-ter of Samsung Electronics under Project Number SRFC-MA1901-04.