Defect chemistry in semiconductor-based amperometric electro- and photoelectrochemical devices

Abstract

Surface and small bulk modifications of thin layers greatly affect their properties, e.g., the translation of voltage input to current output. The current intensity is a simple and processable signal, which is exploited by several devices. Some like field-effect transistors have the potential to imitate the human olfactory system; however, more generalised, to give rise to selective and sensitive detection devices for biomolecules. Nevertheless, some other devices like photoelectrochemical cells have the potential to render energy production sustainable.On the one hand, this work focusses on simple sensing such as live pH detection with a LGFET and aims at controlling the functionalisation of GO used as channel for enhanced device sensitivities. On the other hand, the work also concentrates on the co-doping as valid synergistic approach for the performance improvement of hematite as PEC photoanodic template material.Thermal treatments in different atmospheres lead to a less current-resistant GO which is referred to as rGO. According to the atmosphere, GO can deliver higher sensitivities to pH variations. Upon heating in protected atmosphere by means of a custom-made setup, the current intensity flowing through GO is monitored, and the arising current transients can be followed live. First, it is important that the heat is well conducted from the gas in the setup chamber, hence we see highest resistance and lowest performance upon usage of vacuum atmosphere. Upon comparison of Ar and 2.6% H2/Ar mixture, we see that H2 does not affect chemically GO, though the slightly different behaviour is assigned to a much higher thermal conductivity than Ar only. By contrast, NH3 affects chemically GO towards the lowest resistance and achieving remarkable pH sensitivities. These are comparable with reported rGO pH sensors, where the reducing agent was the reliable hydrazine hydrate.The temperature ramp turns out crucial for effective defunctionalisation of GO and proper functionalisation of rGO. In fact, by applying a different synthesis method with higher ramp, NH3 leads to a record pH sensitivity hovering around 60 mA pH-1. NH3 might exploit its nucleophilic properties to enter the surface and the C-network of GO; while, contemporaneously, it might also exploit its great solubility and dissolve in the intercalated water that GO carries.As earlier cited, the second main topic treats co-doping, and this is accomplished in an easy-synthesised photoanode template, i.e., a hematite spin-coated thin film. It is the first time that Sb, Li co-doping is realised, and it is also the first time that direct Sb5+-doping is reported. The addition of the dopants is modest as high as a nominal at. 1%. The Sb5+ oxidation state maintains in the mono- and co-doped samples and addition of Li+ exhibits a synergistic effect on the performance, which is already greatly expanded by Sb5+. At 1.4 V vs RHE, the current increase is 67-fold compared to standard hematite thin film. The benefits might lie predominantly in the higher carrier concentration and a lower surface resistance. The J-V curves do not suggest catalysed surface reaction. Of interest would be the application of the Fe(III)-OEC and reaction with hydrogen peroxide in the electrolyte.Furthermore, the presence and management of VO in Li+-doped hematite shows curious first results, as if there would be a competition of the two species, in terms of space or photoelectrochemical contribute.For both projects, I-V or J-V curve characterisation is performed on samples where single variables are sequentially altered. In the rGO FET project, GO current increase is live-monitored, IdsVg measurements are completing the final pH sensing results (Idst). In the co-doping project, J-V curves are recorded without illumination, and additionally by chopping the illumination to extract more information. Results are also backed by comprehensive XPS, XRD, Raman characterisation, and moreover, by supplementary SEM and UV-Vis analysis. More characterisation and few extra sample batches are anyway needed to further strengthen what are still hypotheses.