Large Transverse Thermoelectric Effect in Weyl Semimetal TaIrTe$_4$ Engineered for Photodetection

TL;DR

This work demonstrates that the anomalous photocurrents observed in TaIrTe$_4$ arise from a large transverse photothermoelectric effect driven by highly anisotropic Seebeck coefficients in a p×n-type Weyl semimetal. By combining scanning photocurrent microscopy, Shockley–Ramo transport theory, and multi-physics simulations, the authors show that edge-local currents can persist even when electrode currents vanish, and that the response extends into the long-wavelength infrared where Weyl-node-related nonlinear effects would be expected to appear but are not observed. They further show that engineering the thermal environment—via substrate steps and regionally varied thermal boundary conductance—can locally enhance and tailor the photocurrent, enabling broadband photodetection schemes with potential for wavefront sensing and edge detection. Overall, the study provides a thermally focused framework for interpreting and optimizing photodetection in anisotropic Weyl semimetals, offering avenues for energy harvesting and thermoelectric cooling technologies without requiring magnetic bias. The results underscore the importance of distinguishing PTE from BPVE in topological materials and highlight thermal landscape engineering as a practical tool for device performance control.

Abstract

Anomalous local photocurrent generation via second-order nonlinear and thermoelectric responses is a signature of many topological semimetals. The emergence of these photocurrents is inherently linked to symmetry breaking and anisotropy of their crystal lattices. Studies of type-II Weyl semimetals of group C$_{2v}$ (WTe$_2$, MoTe$_2$, TaIrTe$_4$) have reported anomalous, nonlocal photocurrents localized to crystals edges or far from electrodes, which are highly dependent on the geometry of the material sample. While originally attributed to a nonlinear charge current response, it was recently shown that these currents could instead be attributed to the anisotropic Seebeck coefficients of the materials. Here, we confirm that anomalous photocurrents observed in TaIrTe$_4$ under either visible or far-infrared far-field illumination originate from the large transverse thermoelectric effect. We engineer the mutual orientation of crystal edges and electrodes as well as the thermal environment of TaIrTe$_4$ to control and amplify its spatial photocurrent response. We show that substrate engineering can locally enhance photocurrent. This framework of thermal device engineering can enable broadband photo detection schemes by leveraging spectral and spatial dependence of photocurrents for applications like wavefront sensing, beam positioning, and edge detection.

Figures (13)

• Figure 1: Experimental overview and PTE mechanisms. (A) Scanning photocurrent microscopy is used to characterize the photocurrent response of TaIrTe$_4$ devices. (B) TaIrTe$_4$ crystallizes in a layered, non-centrosymmetric orthorhombic crystal structure. (C) The XRD spectrum of our TaIrTe$_4$ crystal at ambient conditions shows prominent (00$\ell$) reflections, consistent with the Pmn2$_1$ space group. (D) Second-order NLO photocurrents are only allowed under non-normal incidence illumination according to the NLO tensor $\sigma^{(2)}_{ijk}$. (E) Under the longitudinal PTE, a temperature gradient drives a parallel electric field, (F) which can manifest in SPCM at the electrode-material interface due to partial obstruction of the laser spot, resulting in opposite-sign currents at either electrode. (G) In the transverse PTE, off-diagonal Seebeck tensor elements drive an electric field transverse to the temperature gradient direction. (H) In a p$\times$n-type conductor like TaIrTe$_4$, this can manifest as highly nonlocal edge currents on the off-axis edges. (I) Current at the edge versus electrode as a function of crystal $a$-axis orientation $\alpha$ for $\theta = 45$° and $S_a:S_b = -6:27$. Dashed lines indicate $\alpha = 25.5$° and $\alpha = 154.5$°, where $J_{\text{electrode}} = 0$ while $J_{\text{edge}}$ remains finite. Bottom: schematic of the crystal geometry at $\alpha = 25.5$°, showing the condition for vanishing electrode current with persistent edge currents.
• Figure 2: Longitudinal and transverse PTE in a TaIrTe$_4$ device at 635 nm illumination. (A) A single-crystal TaIrTe$_4$ photodetector with the a-axis aligned along the natural device edge on the right, as identified by (B) angle-resolved polarized Raman spectroscopy. The 147 cm peak has 180° periodicity with maximum intensity along the a-axis. (C) The experimentally measured dielectric function $\varepsilon_{a,b}(\omega)$ of TaIrTe$_4$ in the visible and near-infrared region, approximated as isotropic in-plane. SPCM maps of (D) reflection and photocurrent were measured with 17 µW linearly polarized 635 nm light for (E) $E\parallel a$ and (F) $E\parallel b$, showing no polarization dependence. (G) The weighting field $\nabla\psi$ of the device, as visualized with streamlines, is taken into account along with the anisotropic Seebeck tensor of TaIrTe$_4$ in (H) the simulated photocurrent pattern using Shockley-Ramo theory, which is in good agreement with experiments for both $E\parallel a$ and $E\parallel b$. (I) We simulate the local photocurrent vector field $\vec{J}_{\text{loc}}(\mathbf{r})$ at the off-axis crystal edge, showing that net photocurrent is measured between the electrodes due to a nonzero transverse PTE current component parallel to the weighting field: $\vec{J}_{\text{loc}}(\mathbf{r}) \parallel \nabla\psi$. (J) At the a-axis edge there is zero net current because $\vec{J}_{\text{loc}}(\mathbf{r}) \perp \nabla\psi$ in that region. The chopper frequency roll-off was measured to extract the photocurrent time response at the (K) electrode interface, $\tau = 32 \pm 1.6$ µs, and (L) the off-axis edge, $\tau = 31 \pm 5.0$ µs.
• Figure 3: Long-wave infrared photocurrent response of TaIrTe$_4$. The LWIR response is investigated at (A) the off-axis edge of device A, which is (B) 130 nm thick on a SiO$_2$(285 nm)/Si substrate. The dielectric functions of TaIrTe$_4$ in the LWIR is metallic along the a-axis and dielectric along the b-axis, which is apparent in (C) the calculate E-field intensity in the TaIrTe$_4$ flake for $E \parallel a$ vs. $E \parallel b$ (shown for $\lambda=11$ µm. (D) The modeled absorption into the flake is higher for $E \parallel b$. (E) The thermal response in TaIrTe$_4$ for focused laser illumination in the LWIR is highly dependent on light polarization due to the in-plane optical anisotropy. (F) This is apparent in the temperature $T$ and temperature gradient $\nabla T$ profiles of TaIrTe$_4$ for a focused laser spot, in-plane and out-of-plane. (G) Accounting for the wavelength-dependent optical absorption, the calculated maximum $\nabla T_x$ at the edge is larger for $E\parallel b$. (H) LWIR SPCM photocurrent maps were measured for five wavelengths between 7.7-12 µm for $E||a$ and $E||a$. The results for 11 µm are shown for 285 µW laser power. The black dashed lines correspond to (I) the plotted profile of the edge-current, showing $|I^{edge}_b| >|I^{edge}_a|$. (J) The spectral dependence of the the off-axis edge photocurrent matches the trend predicted by our $\nabla T(\omega)$ calculation. (K) There is a well known plasmonic enhancement when the incident polarization is perpendicular to the electrode interface, shown by (L) FDTD calculated E-field intensity for $E^{90^{\circ}}$ vs. $E^{0^{\circ}}$. (M) In the LWIR region the enhancement is calculated as $\sim2.5\times$.
• Figure 4: Thermally engineered photocurrent response of TaIrTe$_4$. (A) A 200 nm thick TaIrTe$_4$ flake was placed on a substrate, half suspended on top of a 360 nm evaporated SiO$_2$ step. (B) The crystal axes are identified by angle-resolved polarized Raman spectroscopy. (C) The reduced TBC $G_{\text{TaIrTe}_{4}\text{-SiO}_{2}}$ of the evaporated SiO$_2$ is illustrated with a shorter arrow compared to that of the thermally grown SiO$_2$. (D) The device consequently has three regions with different TBC: TaIrTe$_4$ on thermally grown SiO$_2$, a transition region where TaIrTe$_4$ is suspended over air, and TaIrTe$_4$ on evaporated SiO$_2$. (E) The simulated light absorption into the flake is the same on and off the SiO$_2$ step based on the measured $\varepsilon(\omega)$ in the visible. The device was measured with SPCM and the (F) reflection and (G) photocurrent maps at 635 nm illumination (19 µW power) show localization of photocurrent to the device edges in the SiO$_2$ step region of the flake. Accounting for the decreased TBC on the step, (H) the Shockley-Ramo simulated response is in good agreement with the measured response. (I) We predict a slight absorption enhancement for $E \parallel b$, and none for $E \parallel a$ on the SiO$_2$ step in the LWIR. SPCM with (J) $\lambda=12$ µm and (K) $\lambda=7.7$ µm (71 and 230 µW power respectively) for $\angle (\vec{E} , \vec{a})=10^{\circ}$ follow the same photocurrent pattern as the $635$ nm case and (L) the simulated photocurrent pattern accounting for the IR optical response replicates the observed pattern.
• Figure S1: The measured I-V curves of device A ($R=213$ Ω) and and device B ($R=354$ Ω).