There is growing evidence that our increasing consumption of fossil fuels is leading to a change in climate. Such predictions have brought new urgency to the development of clean, renewable sources of energy that will permit the current level of world economic growth to continue without damage to our ecosystem. Photovoltaic cells based on organic or organic/inorganic hybrid materials have shown rapid improvements over the past decade, comparing favourably with existing inorganic semiconductor technology on energy, scalability and cost associated with manufacture. The most promising materials for organic or hybrid photovoltaics are based on blends of two components at whose interface light-generated excitations dissociate into charges contributing to a photocurrent. Blend morphology on the meso-scale plays a crucial role in these systems, with efficient photovoltaic operation requiring both large interfacial area and existence of carrier percolation paths to the electrodes. The proposed work will establish how both aims can be achieved, using a powerful new combination of non-contact femtosecond time-resolved techniques to examine a range of novel mesoscopic blends. This methodology will allow the simultaneous examination of exciton diffusion and dissociation, charge-carrier generation, recombination and conductivity, providing direct clues to the optimisation of materials for photovoltaics. Collaborations with researchers working on making photovoltaic devices will ensure that knowledge gained from these non-contact material probes will directly feed into enhancing device performance. This combined approach will allow the UK's exceptionally high expertise in the area of organic electronics to contribute effectively to its current goal of reducing harmful greenhouse gas emission.

Wearable electronics like smart watches are popular gadgets, their value as biosensor is limited however as they are not directly interfaced with the skin. While body temperature, heart rate and blood oxygen content and important factors to evaluate the overall health of an individual, these parameters lack the information contained in biological metabolites excreted through sweat. The presence or fluctuation in concentration of many metabolites can be directly related to a medical condition. High glucose levels could be indicative of diabetes, whilst the presence of lactate would indicate fatigue and high cortisol, a steroid hormone, levels often pointing towards increased stress levels. This fellowship extension will focus on the integration of flexible and self-repairing organic semiconductors into skin-wearable biosensors for non-invasive health monitoring. The wearable sensors will be fully conformal, similar to a standard band aid, and robust enough to be worn directly on the skin. By carefully functionalising the active material, the organic semiconductor, we aim to achieve a high selectivity and sensitivity, able to record the minute changes of biological metabolites excreted through sweat on the skin. The realisation of such light weight and robust wearable sensors directly applied to the human skin, would make the continous monitoring of relevant biological metabolites a reality, opening new avenues for more cost-effective preventive healthcare and patient-centred care.

Our overall aim is to make fundamental step-changes in understanding of seafloor processes and hazards by developing and demonstrating novel sensor systems, which can form widespread and long-term listening networks. These low-cost and energy-efficient sensors comprise hydrophones (acoustic noise in water column) and geophones (ground shaking). Data will be returned via pop-up floats and satellite links, as has been pioneered by the highly successful Argo Project for water-column profile. This type of low-cost network could have unusually widespread applications for warning against threats to valuable seabed infrastructure, monitoring leaks from CCS facilities or gas pipelines, or for tsunami warning systems. Here we aim to answer fundamental questions about how submarine mass-flows (turbidity currents and landslides) are triggered, and then behave. These hazardous and often powerful (2-20 m/s) submarine events form the largest sediment accumulations, deepest canyons, and longest channel systems on our planet. Turbidity currents can runout for hundreds to thousands of kilometres, to break seabed cable networks that carry >95% of global data traffic, including the internet and financial markets, or strategic oil and gas pipelines. These flows play a globally important role in organic carbon and nutrient transfer to the deep ocean, and geochemical cycles; whilst their deposits host valuable oil and gas reserves worldwide. Submarine mass flows are notoriously difficult to measure in action, and there are very few measurements compared to their subaerial cousins. This means there are fundamental gaps in basic understanding about how submarine mass flows are triggered, their frequency and runout, and how they behave. Recent monitoring has made advances using power-hungry (active source) sensors, such as acoustic Doppler current profilers (ADCPs). But active-source sensors have major disadvantages, and cannot be deployed globally. They can only measure for short periods, are located on moorings anchored inside these powerful flows (which often carry the expensive mooring and sensors away), and they need multiple periods of expensive research vessels to be both deployed and recovered. We will therefore design, build and test passive sensors that can be deployed over widespread areas at far lower cost. These novel sensors will record mass-flow timing and triggers; and changes in front speed (from transit times), and flow power (via strength of acoustic or vibration signal). We will first determine how submarine mass flows are best recorded by hydrophones and geophones, and how that record varies with flow speed and type, or distance to sensor. Our preliminary work at three sites already shows that hydrophone and geophones do record mass-flows. Here we will determine the best way to capture that mass-flow signal, and to distinguish it from other processes. This work will form the basis for designing a new generation of low-cost (< £5k) smart sensors that return data without expensive surface vessels; via pop-up floats and satellite links. Advances in technology make this project timely, as they allow on-board data processing by smart hydrophones or geophones to reduce data volumes, which can be triggered to record for short periods at much higher frequency. We will field-test the new smart sensors, and thus demonstrate how they can answer major science questions. We seek to understand what triggers submarine flows, and how this initial trigger mechanism affects flow behaviour. In particular, how are submarine flows linked to hazardous river floods, storms or earthquakes, and hence how do they record those hazards? Do submarine flows in diverse settings show consistent modes of behaviour, and if not, what causes those differences? To do this, we will deploy these new sensors along the Congo Canyon (dilute river, passive margin, no cyclones) offshore Taiwan.

We are interested in the incorporation of nitrogen into semiconductors such as GaAs, InAs and GaSb. This is important because the band gap of the parent III/V semiconductor is substantially reduced by the incorporation of very small amounts of nitrogen. These so-called dilute nitrides show promise for use in tailoring the wavelength and efficiency of novel semiconductor lasers and other optoelectronic devices. Although GaAsN and InGaAsN are currently being studied mainly for their applications in photodetectors and lasers in the 1.3 to 1.55 um telecomms wavelength range there is far less research into dilute nitride compounds for the mid-infrared (2-5 um) spectral range which is rich in applications. However, there are problems associated with incorporation of N and degradation of the crystalline quality and especially as nitrogen content in the material is increased beyond 1%. This project seeks to investigate the growth of dilute nitrides for the mid-infrared spectral range using growth from the liquid phase rather than from the gas phase.One key advantage of this approach is that we do not need any N plasma to introduce the nitrogen atoms and so we can avoid all the damage from the energetic N ion species generated as a by-product from the plasma source normally used in vapour phase growth. Liquid phase epitaxy (LPE) is well known to produce material of excellent crystalline perfection. The proposed project seeks to build on our existing expertise in LPE growth and mid-infrared optoelectronics at Lancaster and study the resulting material properties of GaAsN, InAsN, GaSbN with a view towards evaluating their potential for use in mid-infrared optoelectronic devices. We aim to investigate both bulk materials and also corresponding dilute N nanostructures. The preparation of dilute N III-V alloys with high quantum efficiency would be a real breakthrough, particularly for use within mid-infrared light sources and detectors for which there are many practical applications. Moreover, if the approach proves successful it can be readily extended to other technologically important alloys such as InGaAsN and GaAsPN.

Autonomous driving (AD) has a huge market and IS receiving enormous attention in both academia and industry. To deal with complex scenarios, autonomous vehicles (AVs) will use reinforcement learning (RL) to design high-level planners in the functional layer but always suffer from safety issues during sim-to-real transfer. One of the main challenges is that the current practice of functional-layer design does not sufficiently consider the uncertainty in the architecture layer, e.g., the software layer and hardware layer. This open challenge will be tackled in this project by a comprehensive study of the interaction between RL and architecture-layer uncertainty. Specifically, we will build virtual AD scenarios on the simulation platform with formal modeling of architecture-layer uncertainty based on real-world data (WP1). The impact of uncertainties on RL will be discussed via the design of cross-layer uncertainty-aware RL (WP2). Inversely, we will also study the robustness of an RL with respect to cross-layer uncertainty by computing the Pareto front of the largest software/hardware uncertainty patterns that a given RL is robust to (WP3). Extensive analysis including verification (WP2, WP3), simulation (WP2, WP3), and real-world experiments (WP4) will be carried out. The success of this project will greatly improve the practicability of RL in AD with a broader impact on other robotics applications.