Silicon nanowires are structures made from silicon with at least one spatial dimension in the nanometer regime (1-100 nm). From these nanowires, silicon nanowire field-effect transistors can be constructed. Since their introduction in 2001 silicon nanowire field-effect transistors have been studied because of their promising application as selective sensors for biological and chemical species. Their large surface-to-volume ratio promises an increased sensitivity compared to conventional, planar field-effect transistors. Selectivity can be added by smart modification of the surface of the nanowire. The application of nanowire field-effect transistors as chemical sensors for ions is studied in this thesis. After briefly discussing the working principle of field-effect transistors, an extensive review on reported state-of-the-art surface modification techniques is presented. By far, most of this work covers the covalent attachment of molecules to the silicon oxide layer that is typically present on as-prepared silicon nanowires. In addition, some examples on non-covalent approaches have been reviewed. Alternatively, the oxide layer can be removed to attach molecules directly onto silicon. Nanowire field-effect transistors are prepared using top-down fabrication techniques. Nanowires with a width of 20 to 2000 nanometer, a length of 3 to 5 micrometer and a height of 50 to 100 nanometer are etched from silicon-on-insulator wafers. Using a newly developed complementary metal-oxide-semiconductor compatible process, nanowire field-effect transistors are obtained. This process allows for the use of a broad variety of front oxides, which enables selective surface modification of the nanowires and the implementation of materials with different dielectric constants. Furthermore, the presence of a silicon nitride passivation layer on the area around the nanowire provides the possibility to modify only the nanowire surface. Chips typically consist of 28 individual nanowire field-effect transistors. The devices are characterized in air and the variation of the threshold voltage over the wafer is mapped. It is shown that the electrical characteristics make the devices suitable for sensing experiments. The nanowire field-effect transistors are electrically characterized while exposed to water using a newly developed flow cell. A pulsed gate potential is applied as method for stable characterization. While most methods reported in literature use a reference electrode, the pulsed method can be applied without it. Using the sensitivity towards protons of the Si-OH groups at the SiO2 surface, the pH of aqueous solutions is determined using this pulsed gate potential method. It is found that upon increasing pH the threshold voltage increases, which is as expected. The nanowire field-effect transistors are studied using two types of gating: back gating and front gating through the liquid via an Ag/AgCl electrode. It is found that both methods can be used to gate the device, and that in general smaller potentials are needed for front gating compared to back gating. Using this front-gate method, the influence of the non-aqueous solvent 1,4-dioxane on the device characteristics is studied by exposing the devices to different water-1,4-dioxane mixtures. The dependency of the threshold voltage on the mixture composition is found to be related to the decreased dissociation of the surface silanol groups and the electrical conductivity of the mixture used. In the final experiments, the nanowire field-effect transistors are covered with an ionophore-containing polymer membrane via drop casting. This membrane consists of a silicon rubber polymer (Siloprene), embedding the ionophore valinomycin, which has a K+/Na+ selectivity of ~10^5. Using a setup with liquid gating and an Ag/AgCl reference electrode, accurate potassium ion concentrations are determined. Upon exposure of the membrane modified device to sodium ions at a fixed potassium ion concentration, the potential of the device only changes at very high sodium ion concentrations. This confirms the high ion selectivity of the membrane modified device. These results are comparable in line with those obtained using a conventional ion-selective electrode setup. In conclusion, the experiments presented in this PhD thesis lead to an increased understanding of the electrical behavior of nanowire field-effect transistors under different circumstances. This will support the construction and operation of nanowire field-effect sensors, facilitating the further development of advanced sensors.