Sulfur-deficient polycrystalline two-dimensional molybdenum disulfide (MoS₂) transistors exhibit gate-tunable memristive properties which show promise for complex neuromorphic learning and high-performance logic and memory operations.^1,2 The current understanding of the dominant switching mechanism in these ‘memtransistors’ is that the bias-induced movement of sulfur vacancies (V_s), likely along grain boundaries (GB), changes the height of the Schottky barrier at the current-limiting contact through image-force lowering.^3-5 The requirement for GB within the channel places a lower limit on the channel length (L_c), on the order of the grain size in the MoS₂. However, the source-drain voltage needed to switch between resistance states in the best-performing memtransistors⁴ (L_c ≈ 5 μm) is currently too large for practical memory applications. One approach to facilitate the resistance switching, without reducing L_c, would be through acceleration of the defect migration achieved through post-fabrication surface treatments applied to the channel region. The ‘all-surface’ nature of two-dimensional transistors makes surface modifications such as vapor adsorption a well-known and powerful approach for changing the electronic transport properties of MoS₂,^6-8 but little work has been done to explore the influence that surface treatments may have on the kinetics of the V_s movement. We have investigated two-terminal back-gated MoS₂ memtransistors on 300-nm SiO₂ using variable-temperature drain voltage pulse trains, complemented with in situ Raman spectroscopy in a controlled atmosphere environment, to demonstrate the effect of adsorbed molecular species on the kinetics of the resistive switching. We also examined how the electronic trap-filling processes concomitant with the resistive switching are modified by the adsorption of molecular species and how these modifications alter the kinetics of the resistive switching in turn. Atmospheric water was found to remain adsorbed to the MoS₂ channel even after application of vacuum (P = 1 x 10^-5 Torr) overnight, and was only successfully removed in vacuo after either heating to 100ºC or subjecting the channel to ~50 cyclic drain voltage sweeps (-30 V < V_ds < +30 V). ‘Bunched’ pulse train measurements with 0 V gate bias performed before and after desorbing water reveal that two timescales are associated with the drain current evolution in the pristine (i.e., water-desorbed) film: a fast process which relaxes quickly, superimposed on a slower persistent change in resistance. The presence of water adsorbed to the MoS₂ channel suppresses the fast process, which is tentatively assigned to electronic trap filling, while the slower persistent process corresponding to the long-lived resistive memory is maintained. However, the fast trap-filling process is observed for both the water-adsorbed and pristine condition when a subthreshold bias (-70 V) is applied to the back gate, suggesting that the energy of the trap state has been shifted significantly by the presence of the water. Further pulsing experiments on the pristine state using different pulse-bunch groupings indicate that when the pulses are bunched such that they do not fill the traps, the persistent memory switching is accelerated. Finally, experiments using ethanol and isopropyl alcohol vapors to identify whether tuning properties of the molecular adsorbate can potentially accelerate memtransistor switching kinetics in the technologically-promising subthreshold region are also presented.

This work was funded by the Laboratory Directed Research and Development Program and performed, in part, at the Center for Integrated Nanotechnologies, an Office of Science User Facility operated for the US Department of Energy (DOE) Office of Science. Sandia National Laboratories is a multi-mission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the US Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525. The views expressed in this article do not necessarily represent the views of the US Department of Energy or the United States Government.

1. Strukov, et al. Nature 2008, 453, 80–83.
2. Yang et al. Nat. Nanotechnol. 2013, 8, 13–24.
3. Sangwan et al. Nat. Nanotechnol. 2015, 10, 403-406.
4. Sangwan et al. Nature 2018, 554, 500-504.
5. Li et al. ACS Nano 2018, 12, 9240-9252.
6. Chen et al. J. Vac. Sci. Technol., B 2014, 32, 06FF02.
7. Rao et al. 2D Mater. 2019, 6, 045031.
8. Gustafson et al. J. Phys. Chem. C 2021, 125, 8712-8718.