An arXiv v1 preprint dated 20 August 2026 reports that 1T-TaS2 crystals changed between insulating and metallic states by different routes: during cooling, the inferred share of material in each phase shifted in discrete, avalanche-like steps, while during heating it changed continuously.
The study combined transport, calorimetry and Raman measurements to characterize the transition. It used high-quality single crystals with reported purity above 99.995%, as well as a 30 nm thin-layer transport sample in which the asymmetric kinetics were pronounced.
The transition split into two patterns
A higher temperature-ramp rate obscured individual transformation events, with 6 K/min given as an example. Raman tracking showed a nearly single-step cooling transition within approximately 1 K.
The sharper cooling pattern was not confined to one type of Raman response: both modes that hardened and modes that softened changed more sharply during cooling than during heating.
Pulses exposed the asymmetry
The researchers then used isothermal phase-ordering runs, with each run including a complete thermal cycle, approximately five minutes of equilibration, 500 current pulses and a further five minutes of relaxation.
The two branches responded differently to those pulses. During heating, the response moved in one direction toward the metallic phase; during cooling, it switched stochastically back and forth without complete accumulation.
The net perturbation-induced transformation was nearly twice as large during heating as during cooling. That larger heating response persisted across perturbation strengths and when the response was extracted directly from resistance rather than from the model used to infer phase fraction.
What the authors think is happening
The authors interpret the split as evidence that the cooling and heating branches have different degrees of athermality. In their framework, cooling is strongly athermal, while heating is less athermal than cooling, helping account for the avalanche-like and continuous pathways, respectively.
A Landau ϕ6 free-energy model with time-dependent Ginzburg–Landau dynamics was reported to capture the essential qualitative asymmetry, representing heating as less athermal than cooling. The model is phenomenological, so its behavior is not an independent experimental validation of the microscopic mechanism.
Thermal conductivity—the ability of a material to carry heat—was higher in the insulating phase than in the metallic phase, and bulk values were larger than those reported for thin layers. The authors point to this phase-dependent transport as a possible correlate of the branch-specific response, but the microscopic connection remains unresolved.
The evidence has limits
The evidence is limited to laboratory measurements and phenomenological modeling of 1T-TaS2 crystals under the reported thermal-ramp and pulsed-heating protocols. It does not establish that unequal athermality is the sole microscopic source of the asymmetry or that the same mechanism applies across first-order transitions.
The report does not state how many crystals or independent measurements were included, nor how many repeated cycles were run. It also gives no inferential test or replicate-based variance summary for the nearly twofold pulse comparison.
The inferred insulating-phase fraction used a 2D percolation model with t = 1.3 and ϕc = 0.5; changing ϕc was reported not to alter the qualitative behavior or main conclusions. The nearly twofold comparison therefore remains a reported experimental result rather than a universal estimate.
Paper data and sources
Original title: Fluctuation-Controlled Asymmetric Kinetics in Metal-Insulator Transitions
Authors: Tapas Bar, David Pesquera, Timm Swoboda et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-20
DOI: Not available
Original paper · Full text