A laboratory study reports evidence that two electrical controls can separate how much Li⁺ is present at a van der Waals interface from how quickly the ions move. That is the central test of a double-gated device: whether Li⁺ occupancy and transport rate can be varied independently.
The devices combined mechanically exfoliated few-layer graphene (FLG) or MoS₂ on hBN with a LiTFSI/PEG electrolyte and carbon-cloth gate electrodes. Ionic and electronic currents were measured simultaneously while custom code controlled the two gate-voltage combinations independently.
One voltage sum, one voltage difference
The researchers tracked the sum of the gate voltages, ΣV = Vb + Vt, and their difference, ΔV = Vb − Vt. The authors proposed that ΣV would set ionic occupancy, while ΔV would set the voltage drop associated with transport.
When ΔV was varied while the device remained on fixed ionic-density plateaus, ionic current rose linearly. The fitted slopes were 1.75 ± 0.05 nA/V for Stage 1, 4.4 ± 0.2 nA/V for Stage 2 and 6.4 ± 0.2 nA/V for Stage 3. The measured slopes differed across the density stages.
A signal that held up under repeated switching
The ionic response was hysteretic, meaning its path differed as the voltage was raised and lowered. Its ON-OFF ratio was about 100, and the devices sustained more than 1,000 ON-OFF cycles without degradation.
Reference tests produced different results. Blocked-channel devices showed no measurable current; alternative-ion or hole-doping conditions were associated with irreversible leakage and device failure. Single-gated devices lost hysteresis and showed leakage after roughly 10 ON-OFF cycles.
Charge accounting also followed the switching steps: the charge measured during intercalation approximately matched the combined charge of three deintercalation peaks, and the charge in each event was independent of ΔV. The estimated Li⁺ density was about 6 × 10¹³ cm⁻² at the highest plateau, falling by about 2 × 10¹³ cm⁻² per step.
Heat, logic and memory
Temperature tests showed a clear difference between the two states. From 20 to 80 °C, OFF-state current stayed at background while ON-state plateau current increased exponentially. An Arrhenius analysis, which relates current to temperature, gave activation-energy estimates of 0.6 ± 0.2 eV for FLG and 0.7 ± 0.1 eV for MoS₂.
The same two controls were assigned different jobs in an ionic-transistor demonstration. ΣV served as a digital WRITE input for ionic ON/OFF switching, while ΔV served as an analogue READ input for current magnitude. In a short-pulse protocol, a WRITE pulse above the intercalation threshold was associated with an ON current 100 times background; a second, below-threshold pulse was associated with OFF.
Both FLG and MoS₂ showed non-volatile ionic memory retention in the demonstrated protocol. The study does not state how long that retention lasted.
What the numbers establish
The study reports reproducible hysteretic features across 14 different devices. Its threshold-potential characterization included 11 FLG devices and three MoS₂ devices.
The effective Li⁺ mobility was estimated at about 1 × 10⁻⁴ cm² V⁻¹ s⁻¹ for all plateaus. The paper describes this as an order-of-magnitude, model-based estimate, so it is a scale for the transport rather than a precision direct measurement.
The work is labeled a preprint. Its reported measurements concern Li⁺ transport and electronic current in double-gated FLG/MoS₂-hBN interfaces.
Paper data and sources
Original title: Electrostatic control of Li+ density and transport rate in double-gated van der Waals devices
Authors: E. Hoenig, X. Zhang, C. Li et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-24
DOI: Not available
Original paper · Full text