Theresia Knobloch Dipl.-Ing. Publications

Research into two-dimensional materials has seen an exponential increase during the last decade, triggered by the discovery of the electric field effect in graphene in 2004. However, graphene lacks a band gap which leads to a very low current on/off ratio, making it unsuitable for digital circuits. Attention has therefore gradually shifted to other two-dimensional materials possessing a band gap, such as molybdenum disulphide (MoS₂), a member of the large group of transition metal dichalcogenides.
Since the first practical realization of a field-effect transistor (FET), based on one atomic monolayer of MoS₂ in 2011, many research groups have succeeded in fabricating these devices. Although these devices showed promising properties, they have still not met the high expectations. Apart from a carrier mobility that does not exceed the order of 100 cm²V^-1s^-1, widely known reliability issues, such as hysteresis commonly observed in gate-transfer characteristics, have severely inhibited industrial applications of these novel devices. We have therefore developed a physical modeling approach for describing hysteresis in MoS₂-based FETs. In this way, we have arrived at an enhanced understanding of this phenomenon, which may help to mitigate the hysteresis problem in the future.
Our approach consists of a drift-diffusion-based TCAD model coupled to a full four-state non-radiative multiphonon (NMP) model, used to describe charge capture and emission events in the gate dielectric. The drift-diffusion equations are computationally very efficient and entirely sufficient for describing charge transport through the channel of large-area MoS₂ FETs. The four-state NMP model was previously developed for describing charge capture and emission events in silicon dioxide (SiO₂), which causes bias temperature instabilities (BTI) in silicon-based devices.
This model has now been applied and calibrated to novel MoS₂ devices. The degradation in these devices is governed by a defect band located at approximately 2.9 eV below the conduction band edge of SiO₂, which is illustrated in Fig. 1. Simulations performed with this setup could accurately capture hysteresis in gate-transfer characteristics (see Fig. 2). This is a strong indication that charge capture and emission processes are in fact one of the main reasons for the hysteresis observed in MoS₂ FETs.

Fig. 1: Band alignment of the band gap in silicon to the band gap in one monoatomic layer of MoS₂ and to the defect bands in SiO₂. The upper band dominates the degradation in MoS₂ FETs, while the lower defect band is responsible for BTI in silicon-based devices.

Fig. 2: Comparison of measured and simulated hysteresis in the gate-transfer characteristics of one device. The characteristics were captured using a sweep time of T = 1/f = 80 s.