Thin-film field-effect transistors in hydrogenated amorphous silicon are notoriously unstable due to the formation of silicon dangling bond trapping states in the accumulated channel region during operation. Here, we show that by using a source-gated transistor a major improvement in stability is obtained. This occurs because the electron quasi-Fermi level is pinned near the center of the band in the active source region of the device and strong accumulation of electrons is prevented. Thin-film field-effect transistors (FETs) made using hydrogenated amorphous silicon ͑a-Si: H͒ as the semiconductor are notoriously unstable. This instability, characterized by a large shift in threshold voltage and a reduction in drain current during operation, 1 severely limits their use in both analog and digital circuits for displays and large area electronics, in general. As such, the most important application is as a simple pixel switch in actively addressed liquid-crystal display arrays with anything more demanding such as peripheral or pixel circuitry being difficult or impossible.Silicon dangling bond defects are formed over a wide range of energies within the band gap of amorphous silicon in regions where the electron quasi-Fermi level moves from its equilibrium position determined during the growth of the a-Si: H.1,2 This position, close to the center of the band in undoped material, determines the energy distribution and number of the different bonding configurations between silicon and hydrogen when in chemical equilibrium. For a FET, the electron quasi-Fermi level in the channel region moves toward the conduction band as the transistor is biased into accumulation. 3,4 This is accompanied by an increase in the electron concentration. The microscopic mechanisms for the creation of dangling bond defect states are complex 3 as the material strives to achieve an equilibrium between the weak silicon-silicon bonds, the dangling bond defects and the electron concentration.4 Defect states trap electrons and there is a shift in the threshold voltage of the FET.Since the defect generation mechanism is fundamental, there is very little that can be done about it in FETs as they rely on electron accumulation in a channel region for their operation. At a given gate voltage, drift and instability are independent of current and are particularly severe when operating well above threshold.Source-gated transistors (SGTs) 5 form a class of thinfilm transistors in which the current is controlled entirely by the source. 6 Their most important advantage compared with a FET is the much smaller saturation voltage V SAT . An example of the characteristics of a SGT made using a Schottky barrier source is shown in Fig. 1 together with a schematic showing its structure. V SAT at V G =20 is Ͻ2.5 V. This compares with 17 V ͑V G − V T ͒ for a FET made with the same layers. The SGT also has a large output impedance. 5,6 The main difference with a FET is the provision of a source barrier rather than an ohmic contact and a gate located below it that con...