An investigation into the use of pulsed power for efficient hydrogen production MonkNigel A. 2022 <div>Hydrogen can potentially become a major energy vector and/or energy store in many future scenarios but is constrained by relatively poor energy-storage-use conversion efficiencies and an uncertain ‘green’ credibility. Identification of methods for increasing hydrogen production efficiency would have significant benefits.</div><div>It is known that commencement of electrolysis occurs at optimum efficiency; this is quickly subdued by the onset of loss mechanisms. Also, operating at high voltage would be advantageous because the consequence is higher electrode surface current density or smaller electrode area, either of which directly reduces hydrogen production cost due to plant amortisation; however, increasing voltage increasingly leads to heightened losses. This thesis hypothesises that it may be possible to operate at high voltage whilst retaining the optimum efficiency level by operating intermittently, i.e. by using pulsed power, thereby pre-empting and minimising loss onset. That is, pulsed power will not enhance efficiency per se, it will recuperate or pre-empt heightened losses incurred when attempting to operate at DC voltages substantially above conventional values.</div><div>In this way, this thesis addresses the question, “can pulsed power be used to enhance aqueous electrolysis hydrogen production efficiency?” The topic has not hitherto been comprehensively and systematically studied. This thesis reviews previous and current hydrogen research but finds a lack of relevant activity. </div><div>This thesis reviews analytical electrochemical models but finds none that are capable of direct application to the pulsed power situation. Conventional models are substantially empirical and embody statistically averaging thermodynamics. This thesis reviews atomic level simulation modelling, as an approach to gain insight into pulsed power electrolysis, and finds that physically accurate simulations can accommodate as few as 64 atoms over an extremely short time span, but can provide useful insights. There is therefore no prospect of creating an accurate model of a practical scale electrolyser covering a useful time period using this approach.</div><div>A set of equipment was designed, constructed and tested, comprising voltagemultiplier-based variable power supply, variable timing switch, mixed-gas low-flow volume measurement meter and variable geometry electrolysis cell. The power supply was found to require careful impedance matching to the load while the pulse generator required additional flexibility over frequency of operation. It was determined the overall electrolyser design restricted maximum operating efficiencies. </div><div>The real and reactive behaviour of the cell was tested at frequencies between 20 Hz and 200 kHz using Electrochemical Impedance Spectroscopy methodology via an LCR bridge meter. The essentially capacitive nature of an aqueous parallel plate electrolyser was examined and capacitance was found to approach zero at higher frequencies. Voltage-current phase was found to change significantly between zero and -90° against frequency; the frequency of the transition varied with electrolyte conductivity and electrode spacing. Impedance was found to approach a low value at high frequency regardless of conductivity or separation but some ‘high frequencies’ at some conditions were beyond the range of the LCR bridge so behaviour was inferred to be consistent with other conditions. Re-stated: the impedance of low conductivity solutions at high frequency was found to become as low as that of high conductivity solutions at high frequency, which was generally lower than that at low frequency.</div><div>The EIS method was assessed to determine whether it could provide a useful predictor of electrolyser operation at high voltage by comparing low voltage AC impedance from the EIS analysis against high voltage impedance calculated using a Fast Fourier Transform analysis of the voltage and current oscillograms recorded during efficiency testing. It was found that the effect of temperature variation during efficiency testing swamped the data such that the assessment was inconclusive.</div><div>The basic energy conversion efficiency was measured for a selection of electrolyte concentrations, electrode separations and current densities and directly compared for pulsed and continuous DC operation. Performance was not found to differ significantly between pulsed power compared to DC power. It is proposed that this is because the operating conditions achieved represented simple ‘interrupted-DC’ electrolysis; the operating regime where high voltage pulsed electrolysis might transition to an enhanced efficiency mode was not achieved. A DC characterisation provided I-V curves which confirmed the expected transition was not surpassed. A method to mitigate the losses attributed to high voltage electrolysis by pulsed application was not discovered within the test conditions achieved but the high frequency reduction in impedance demonstrates this may be a viable method to pursue.</div>