A Study of Metal Interactions with Transition Metal Dichalcogenides
Restricted (Penn State Only)
- Author:
- Agyapong, Ama Duffie
- Graduate Program:
- Materials Science and Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- June 14, 2023
- Committee Members:
- John Mauro, Program Head/Chair
Robert Rioux, Outside Field Member
Suzanne Mohney, Chair & Dissertation Advisor
Susan Sinnott, Major Field Member
Saptarshi Das, Outside Unit & Field Member - Keywords:
- TMDS
physical vapor deposition
2D materials
MoS2
WS2
contact metals
Raman spectroscopy
intermetallics
alloys
interfacial reactions
monolayer - Abstract:
- As field effect transistors (FETs) continue to shrink towards the sub-nanometer regime, interest in two-dimensional transition metal dichalcogenides (2D TMDs) such as molybdenum disulfide (MoS2) and tungsten disulfide (WS¬2) grows. Unlike three-dimensional semiconductors, many 2D TMDs remain semiconducting at sub-nanometer thicknesses, which is favorable for the electrostatics of scaled devices. Critical to the successful integration of TMDs in FETs is reducing contact resistance at the metal/semiconductor interface and understanding how metal contacts form on the TMD van der Waals surface. Beyond electronics, there are other fields in which the interactions of metals on TMDs are important, such as catalysis. The goal of this dissertation is to better understand metal/TMD interfaces by examining interfacial reactions, epitaxy of ordered and disordered intermetallic alloys, and the impact of different metal deposition conditions on the structure and doping of TMD monolayers. The first study aims to investigate the reactivity between contact metals (Ti, Al, Cu, Pd, and Au) and 1L WS2. Studies of this buried interface often require destructive techniques, such as using a focused ion beam to access the buried metal/TMD interface for microscopy, which is time-consuming and could potentially modify the interface. The goal of this study was to non-destructively detect interfacial reactions at this interface using Raman spectroscopy as a screening tool to distinguish reactive metals quickly and effectively from non-reactive metals. Our approach involved acquiring Raman spectra from the backside of the metal/WS2 stack through a single-crystal transparent sapphire growth substrate, which allowed us to probe the interface between a coalesced metal film and 1L WS2. The disappearance of the Raman-active phonon modes for WS2 suggested the consumption of WS2 through reactions with the continuous metal film, as observed completely for Ti upon deposition and nearly completely for Al after annealing at and above 100 °C. On the other hand, the persistence of multiple Raman-active phonon modes for WS2 confirms that Au, Cu, and Pd are unreactive with WS2 upon deposition and after cumulatively annealing for 1 h at 100, 200, and 300 °C, even though unreactive metal overlayers could shift some of the peaks in the spectrum. Results matched thermodynamic predictions of metals on bulk WS2. A subsequent study focused on applying criteria for room- and low-temperature quasi-van der Waals epitaxy for elements on MoS2 and WSe2 to the synthesis of highly-oriented single-crystal intermetallic phases on MoS2 growth templates. In addition to thermodynamic stability between the film and substrate, we also considered crystallographic symmetry and atomic mobilities in our selection of intermetallic catalysts to synthesize (PdIn, PdCu, PtAl2, PtCu, PtBi, and PtSn). The phases were directly cosputtered onto exfoliated MoS2 on TEM grids by DC magnetron sputtering with the exception of PtBi where the Pt layer was sputtered, and the Bi layer was evaporated by electron beam evaporation. The Bi/Pt layers were then annealed on a hotplate at 200 °C for 60 min in an ambient environment. We collected plan-view transmission electron microscopy (TEM) images and selected-area electron diffraction (SAED) patterns to determine the phase, crystallinity, and orientation of the intermetallics on MoS2. The resulting PdIn film was highly oriented or textured for as-deposited and annealed conditions with an orientation relationship of PdIn(111)∥MoS_2 (0001); PdIn[11 ̅0]∥MoS_2 [11 ̅00]. We were able to form epitaxial PdCu film at room temperature and after annealing, although they were disordered, obtaining the orientational relationship PdCu(111)∥MoS_2 (0001); PdCu[2 ̅11]∥MoS_2 [11 ̅00]. PtCu was epitaxial at room temperature and after annealing; however, it formed a disordered phase on MoS2 at room temperature with the orientation relationship of PtCu(111)∥MoS_2 (0001);PtCu[2 ̅11]∥MoS_2 [11 ̅00] and an ordered phase on MoS2 after annealing with the orientation relationship of PtCu(0001)∥MoS_2 (0001);PtCu[11 ̅00]∥MoS_2 [11 ̅00]. We observed regions of annealed PtAl2 films in which the (100) PtAl2 plane aligned with the basal plane of MoS2. However, we were unsuccessful in depositing epitaxial PtSn and PtBi on MoS2, possibly due to reaction between S-Sn and incomplete intermixing, respectively. In the final part of this dissertation, we examined the effects of physical vapor deposition of contact metals on 1L MoS2 using the non-destructive backside-illumination Raman spectroscopy method. Metal films (Cu, Pd, Pt, Bi, Sn, and In) were deposited via electron beam (e-beam) evaporation and DC magnetron sputtering at a nominal thickness of 100 nm, except for evaporated Pt (70 nm), which was made thinner due to heat generated during its evaporation. We were able to observe the persistence of the characteristic MoS2 first-order Raman modes after the evaporation of all metals; however, some sputtered metals (Pd, Pt, Sn, and In) showed evidence of damaged 1L MoS2, which resulted in the disappearance of the MoS2 phonon vibrational modes. The degree of damage induced by the sputtering process could be related to the kinetic energies of the atomic species as they arrive at the MoS2 surface. We were able to correlate the disappearance of the first-order Raman modes of the MoS2 beneath the sputtered metals with the cohesive energies of the metals, which correlates with experimentally measured impinging forces of sputtered species from the literature. Our results also indicated that metals with higher cohesive and adsorption energies were more likely to damage monolayer TMDs during sputter deposition. In some cases, we observed the first-order out-of-plane MoS2 vibrational mode redshift or else become two distinct peaks, which tells us about electron doping of MoS2 by metals with modest cohesive and adsorption energies (Bi, In, Sn, and Au). The appearance of the two peaks seems to be linked to metal clusters forming during thin-film deposition and/or bilayer regions on monolayer MoS2, which may result in inhomogeneous doping. The studies presented in this dissertation expand our fundamental understanding of 3D metallic films on 2D semiconductors. They demonstrate and apply a powerful screening tool that provides a non-destructive approach to observing changes at the buried metal/TMD interface. They furthermore contribute to our understanding of the growth of metal and alloy films prepared by physical vapor deposition on transition metal dichalcogenides.