A general approach to modeling graphene field-effect transistors for photothermoelectric detection

Owing to the unique linear Dirac-cone dispersion of graphene, graphene field-effect transistors (GFETs) exhibit advantages including high carrier mobility and ambipolar transport. Leveraging these properties, GFETs provide a promising route for broadband response from terahertz (THz) to ultraviolet wavelengths and self-powered photodetection at room temperature based on photothermoelectric (PTE) effect. However, the absence of a unified modeling framework has hindered systematic device optimization and circuit-level implementation. This work presents a general modeling methodology that couples dark state transport, PTE conversion, and intrinsic noise into an integrated framework. Three sub-models are developed for the drain-to-source current and transconductance: a uniform drift model with minimum complexity, a high carrier density model for the high carrier density regime, and a low carrier density model applicable near the charge neutrality point. A virtual optical power port is introduced to represent optical excitation within circuit simulations, enabling co-design with readout electronics. Illuminated state behavior is described by the Seebeck coefficient derived from Mott's formula combined with a hot electron temperature profile. Noise is modeled using Johnson–Nyquist and 1/f components. Experimental validation on CVD-grown GFETs under 0.288 THz illumination demonstrates strong agreement between predictions and measurements. This work establishes a systematic modeling framework for GFET PTE detectors by integrating I–V characteristics, PTE response as well as noise behavior into a unified scheme. The framework provides a reliable foundation for device performance optimization, readout circuit design, thereby accelerating the transition of GFET PTE detectors from laboratory prototypes to practical optoelectronic systems.