Geometric effects in the Dyakonov–Shur theory of Teraherz photodetection

Riccardo Riolo, Marco Polini, Riccardo Mannella, Andrea Tomadin · 2025

Read Geometric effects in the Dyakonov–Shur theory of Teraherz photodetection by Riccardo Riolo, Marco Polini, Riccardo Mannella, Andrea Tomadin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-09696.txt.

Introduction

Riccardo Riolo, Marco Polini, Riccardo Mannella, and Andrea Tomadin published a 2025 physics paper that completes the theory of a terahertz light detector. The Dyakonov-Shur mechanism turns a transistor into a detector: terahertz waves stir plasma waves in the transistor's channel, whose response rectifies them into a measurable signal. It is a leading route to cheap terahertz cameras.

Every earlier theory assumed the transistor's gate spans the whole channel. Real devices use short gates, as they are easier to build, so theory and practice mismatched. The authors build the theory for short gates, and for several gates at once. Gate placement, they find, matters: a poor position costs signal. The completed theory lets designers tune real shapes instead of ideal ones. Terahertz imaging serves security screening, medicine, and communications, all waiting on cheap, sensitive sensors.

Catalog record

CREATORRiccardo Riolo, Marco Polini, Riccardo Mannella, Andrea Tomadin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE44 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-09696

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Geometric effects in the Dyakonov–Shur theory of Teraherz photodetection Riccardo Riolo, Marco Polini, Riccardo Mannella, Andrea Tomadin Abstract Nonlinear resonances of plasma waves in field-effect transitors enable a well-known photodetection mechanism, first introduced by Dyakonov and Shur in the Nineties, especially suited to the Terahertz (THz) frequency range. Theoretical analyses of the mechanism always assume that the gate of the transistor, which is coupled to the antenna receiving the THz signal, is as long as the channel itself, at odds with typical experimental devices, where short gates are usually employed, e.g. due to fabrication constraints. In this work we overcome this limitation and provide a complete theory of Dyakonov–Shur photodetection in the presence of short gates. We develop our theory in such a general fashion that allows us to also treat the case in which multiple gates are present. We find that a sub-optimal positioning of the gates can substantially decrease the detection efficiency of the device and provide a compact analytical formula to quickly estimate the optimal gate position. I Introduction Nonlinear effects in photonics and opto-electronics have been captivating a great deal of attention for decades [ 1 , 2 ] . Typically, these nonlinearities stem from the intrinsic crystalline structure of the material under investigation, which determine its nonlinear polarization functions such as the second- and third-order ones. Another more subtle source of nonlinearity, however, stems from the motion of electrons in a crystal, in the so-called hydrodynamic transport regime [ 3 ] . As in classical fluid dynamics, these hydrodynamic nonlinearities are convection and the coupling of two collective variables, i.e. density and fluid-element velocity. Hydrodynamic effects in a system of electrons roaming in a crystal arise when electron-electron collisions take place on a time scale \tau_{\text{ee}} that is much shorter than the time scale of the evolution of the above mentioned collective quantities. As the electrons move at the Fermi velocity v_{\text{F}} , the mean free path for electron-electron collisions is \ell_{\text{ee}}=v_{\text{F}}\tau_{\text{ee}} . In order to apply the hydrodynamic theory, \ell_{\text{ee}} must be the shortest length scale in the system. In particular \ell_{\text{ee}} must be shorter than the length L of the sample and shorter than the mean free path of electrons, determined by the scattering rate 1/\tau for collisions with phonons, impurities and defects, which do not conserve momentum. In recent years, the realization of the hydrodynamic regime has been demonstrated in pure two-dimensional (2D) electron systems, where the parameters which control the time- and length-scales mentioned above can be easily tuned with the aid of metallic gates. Examples include graphene [ 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 ] , GaAs/AlGaAs quantum wells [ 14 , 13 , 15 ] , and the highly-pure Weyl semimetal \mathrm{WTe}_{2} [ 16 ] . The theory of hydrodynamic transport in graphene [ 17 , 18 , 19 , 20 , 21 , 22 , 23 ] shows some peculiarities due to the linear electron dispersion in the vicinity of the charge neutrality point, and changes qualitatively when the Fermi energy is smaller than the energy scale associated to the electron temperature [ 24 , 25 ] . The scales of interest for the applicability of the hydrodynamic theory strongly depend on the density n and temperature T of the electron liquid. For both GaAs [ 26 ] and single-layer graphene [ 27 , 28 ] , the inverse lifetime of a quasiparticle of energy \varepsilon_{k} , in the limit \lvert\varepsilon_{k}-\varepsilon_{\text{F}}\rvert\ll k_{\text{B}}T\ll\varepsilon_{\text{F}} , \varepsilon_{\text{F}} being the Fermi energy, scales with temperature as 1/\tau_{\text{ee}}\propto-T^{2}\ln T . What are the implications of hydrodynamic electron transport on the fabrication of a device of some use in applications? The first device of practical interest was proposed theoretically by Dyakonov and Shur (DS) in the Nineties [ 29 , 30 , 31 , 32 ] . DS considered a field-effect transistor (FET) where the 2D electron fluid in the channel is in the hydrodynamic transport regime. This device can be used as a THz photodetector [ 31 ] by connecting an antenna to the source and gate contacts. Indeed, a DC potential (the photovoltage ) is established between source and drain when the FET is driven by an AC signal, which is thus rectified . The underlying mechanism involves the generation and propagation of electron density waves ( plasma waves ) [ 29 , 30 , 31 , 32 , 33 , 34 ] , which mix because of the hydrodynamic nonlinearities. For the plasma waves to be described by the hydrodynamic model, however, their frequency \omega must be much shorter than the electron-electron collision rate 1/\tau_{\text{ee}} . This “collisional” regime is the opposite of the usual “collisionless” regime of plasmons [ 26 ] , where electrons can be understood to move freely under the action of a collective field. For GaAs at n=${10}^{11}\text{\,}\mathrm{c}\mathrm{m}^{-2}$ , T=$10\text{\,}\mathrm{K}$ , the scattering length is \ell_{\text{ee}}\sim$200\text{\,}\mathrm{n}\mathrm{m}$ and the maximum frequency of plasma waves f_{\text{max}}\sim$100\text{\,}\mathrm{G}\mathrm{H}\mathrm{z}$ . On the other hand, for single-layer graphene (SLG) at n=${10}^{12}\text{\,}\mathrm{c}\mathrm{m}^{-2}$ , T=$200\text{\,}\mathrm{K}$ , they are \ell_{\text{ee}}\sim$200\text{\,}\mathrm{n}\mathrm{m}$ and f_{\text{max}}\sim$1\text{\,}\mathrm{T}\mathrm{H}\mathrm{z}$ , respectively. DS theory predicts a frequency-resolved detection for a sufficiently clean channel, i.e. when the frequency \omega is larger than the scattering rate 1/\tau . The DS rectification mechanism is of practical importance in the detection of electromagnetic radiation in the THz range because frequencies in this range exceed the cutoff frequency of electronics circuitry, and other devices based on e.g. Schottky barriers, photothermoelectric or photogalvanic effects [ 40 ] offer limited bandwidth or need cryogenic temperatures. The original theory of DS detection dealt with a parabolic-band 2D electron system as the one that can be found in a GaAs electron-doped quantum well but has been extended to devices based on SLG [ 33 ] and bilayer graphene [ 34 ] , where plasma waves at typical electron densities fall naturally in the THz range and ultra-high electron mobilities at room temperature reduce the minimum frequency necessary for frequency-resolved detection. Similarly, experiments in DS photodetection first used FETs based on GaAs heterostructures [ 35 ] and then graphene-based devices [ 36 , 37 , 38 ] . Resonant, frequency-resolved detection was finally measured in bilayer graphene in 2018 in the pioneering experiments by Bandurin et al. [ 39 ] . Notwithstanding this multi-decade-long research activity in DS photodetection, most theoretical works deal with very simple geometries, where a single gate occupies the entire length of the FET channel. However, in a typical experimental setup, the width of the employed gates often tends to be much shorter than the channel length [ 36 , 37 ] due to fabrication constraints. As a consequence, the position of the gate between source and drain becomes a new geometrical parameter that has to be determined before fabrication on the basis of some criterion (for example, in order to maximize the photodetector efficiency). In this work, we systematically address for the first time how the gate length and its precise positioning affect the photovoltage. Moreover, motivated by current fabrication possibilities, we additionally study the effects of multiple gates in a single FET. Our Article is organized as following. In Sec. II , we review the theory of electron hydrodynamics

Read the complete text