Neuromorphic hardware promises to revolutionize information technology with brain-inspired parallel processing, in-memory computing, and energy-efficient implementation of artificial intelligence and machine learning. In particular, two-dimensional (2D) memtransistors enable gate-tunable non-volatile memory, bio-realistic synaptic phenomena, and atomically thin scaling. However, previously reported 2D memtransistors have not achieved low operating voltages without compromising gate-tunability. Here, we overcome this limitation by demonstrating MoS₂ memtransistors with short channel lengths < 400 nm, low operating voltages < 1 V, and high field-effect switching ratios > 10⁴ while concurrently achieving strong memristive responses. This functionality is realized by fabricating back-gated memtransistors using highly polycrystalline monolayer MoS₂ channels on high-κ Al₂O₃ dielectric layers. Finite-element simulations confirm enhanced electrostatic modulation near the channel contacts, which reduces operating voltages without compromising memristive or field-effect switching. Overall, this work demonstrates a pathway for reducing the size and power consumption of 2D memtransistors as is required for ultrahigh-density integration.

CVD MoS₂ Sample Preparation

Continuous polycrystalline films of monolayer MoS₂ were synthesized  by chemical vapor deposition (CVD) using sulfur powder (Millipore-Sigma, 99.98%) and molybdenum trioxide powder (MoO3, 99.98% trace metal, Sigma-Aldrich) following previously a reported method [20]. Growth substrates were c-plane sapphire substrates cleaned by ultra-sonication with acetone and isopropyl alcohol.

Fabrication of MoS₂ Memtransistors

MoS₂ films were transferred to pre-patterned gate electrodes on Si substrates through a wet transfer process [20]. Memtransistor devices were patterned by an electron beam lithography (Raith Voyager 100) process using a PMMA mask followed by metal evaporation (Denton Vacuum Explorer 14) and lift-off in acetone. Dry etching (Samco RIE-10NR) was used to descum and etch MoS₂ to define the channel regions [18]. ALD of Al₂O₃ gate dielectric was deposited at 100°C using trimethylaluminum precursor (Cambridge Nanotech ALD S100).

Material Characterization

CVD films were screened for coverage and monolayer growth quality using optical microscopy, Raman spectroscopy, and photoluminescence spectroscopy (XploRA PLUS). Raman and photoluminescence measurements used a 532 nm laser with 1800 gr/mm under 100X objective. Lateral force microscopy (LFM) measurements were performed under contact mode atomic force microscopy (Asylum Cypher AFM) characterize the film quality and quantify grain size distribution. Soft and thin LFM tips (NanoAndMore PPP-LFMR) were used of dimension 48 µm × 225 µm, ≈ 0.2 N/m force constant, and ≈ 23 kHz resonant frequency. Grain statistics were calculated by measuring flake lateral size using Gwyddion and plotted in Origin.

Electrical Measurements

Electrical measurements were conducted in a vacuum probe station (pressure ≈ 5 x 10^-5 Torr) at room-temperature using LakeShore CRX 4K probe station. Voltage sweep, endurance, retention, and pulse tests were measured using a Keithley 4200A-SCS Parametere Analyser and homebuilt LabVIEW programs.

COMSOL Finite-Element Simulations

A finite-element method is employed to quantify the relationship between charge distribution and potential by solving the Poisson equation, ∇²Φ = -ρ/ε, which calculates the potential at various points within the device. Φ signifies the potential, ρ represents the charge density, and ε is the dielectric constant. In contrast to other software packages, COMSOL Multiphysics® software streamlines the modelling workflow by automatically computing the Schottky barrier height after defining the work function and metal contact-related boundary conditions specified in Table S1 of the Supporting Information document. The Schottky barrier height Φ_𝐵, is calculated as Φ_𝐵 = Φ_M −𝜒, where Φ_𝑀 is the metal contact work function (Au) and 𝜒 is electron affinity (MoS₂). Subsequent calculations assumed the device adheres to a drift-diffusion model without quantum effects to allow model classical charge carrier transport under the contacts and inside the channel.