Researchers at Monash University in Australia have conducted a lifecycle analysis and net energy analysis (LCA/NEA) of a hypothetical large-scale solar-electrolysis plant for the production of green hydrogen. The open-access paper on the study is published in the RSC journal Energy & Environmental Science.
An important consideration of solar-electrolysis in the context of climate mitigation is the enormity of upscaling required—both at the global scale with respect to the investment, land area, materials, and embodied energy; and at the project scale with respect to the potential localised impacts of gigawatt scale plants. According to the IEA, less than 0.1% of hydrogen is currently produced via water electrolysis and only a fraction of this production is powered by renewable energy.
Taking IRENA’s REmap scenario as a reference, renewable hydrogen could deliver 5% of total final energy demand in 2050. Assuming that a half of this demand is met by solar-electrolysis, 3,100 GW of solar would need to be dedicated to hydrogen production in 2050, or around four-times the current world solar PV installed capacity (based on the current world capacity factor for solar PV of 14%). These projections would imply that perhaps hundreds of gigawatt-scale plants will need to be in operation by 2050. Targets for hydrogen demand beyond 2050 are much greater.
Along with hydrogen production and use, an energy transition will involve the synchronous upscaling of renewable electricity, batteries, and energy use technologies. The materials and metals demanded by a low-carbon economy are projected to be immense and create challenges along the full supply pathway. Some of the most critical materials include cobalt, lithium, nickel, indium, silver, and tellurium. For hydrogen, the platinum group metals used in PEM electrolysers and fuel cells are critical. There is no immediate concern for copper resources, but the average ore grade is declining as higher grade deposits become exhausted.
Energy extraction and processing costs increase super-linearly with declining ore grade, and therefore will tend to worsen EROI. In light of the sheer scale of the hydrogen challenge, several questions demand close consideration. For instance, what will the costs be for water electrolysis powered by solar PV, in energy and material terms; what are the trade-offs between hydrogen and electric transition pathways; and how will energetic, financial, social and other constraints determine or shape future hydrogen production pathways?
—Palmer et al.
They calculated hydrogen production as the ratio of annual solar farm electricity output (minus transmission losses =and balance-of-plant loads), and electrolyzer efficiency. Solar farm electricity generation was determined using average global horizontal irradiance (GHI) for Learmonth, Western Australia of 2,200 kWh m-2 yr-1. They assumed electrolyzer efficiency of 55 kWh kg-1 H2, with sensitivity of 50 and 60 kWh kg-1 H2.
They ran two baseline scenarios:
The solar-battery scenario assumed that on-site battery storage powers standby operation with no grid connection, and depending on electrolyzer turndown, maintains minimum electrolyzer operation during periods of low solar electricity.
The solar-grid scenario assumed that the hydrogen site is connected to the regional grid—the North West Interconnected System (NWIS) of Western Australia. The grid connection enables grid imports to support load balancing of variable solar supply; and grid exports during surplus solar generation (i.e. solar generation exceeding electrolyzer capacity). The NWIS comprises mostly gas-fired generation and has a GHG emission factor of 620 g CO2-e kWh-1. This is higher than the global average of 510 g CO2-e kWh-1.
Key results from the study were:
For both baseline scenarios, the GHG emission intensity is around a quarter that of H2 produced from SMR. However, under reasonably anticipated conditions with grid buffering, the GHG emissions may be comparable to SMR.
For both baseline scenarios, the EROI is less than comparable estimates for fossil fuels, and under some conditions, much less.
For the solar-battery scenario, the solar modules were the most significant component that influenced EROI and GHG emissions.
For the solar-grid scenario, the solar modules were important for EROI, but grid emissions were equally important for GHG emissions.
Electrolyzer turndown is a key sensitivity. The baseline turndown was set at 95%, but at 80 to 90% turndown, the EROI and GHG are adversely impacted. For the solar-battery configuration, a lower turndown (higher minimum electrolyzer load) may be infeasible due to impractically large battery capacity.
The emissions factor of electricity for the solar module supply chain is a critical sensitivity. Production in a lower emission region would significantly decrease the GHG emission intensity.
Operational lifetime is important for both EROI and GHG intensity. Early decommissioning due to obsolescence, accelerated degradation, or premature failure would worsen both metrics.
We found that EROI was lower than the reference EROI for fossil fuels under baseline scenarios, and under some 20 conditions with multiple sensitivities, much lower. Ongoing efficiency improvements in solar module manufacturing will drive improvements in EROI, while resource constraints will worsen EROI.
Further work needs to be undertaken to ascertain turndown and ramp rates of electrolysers and post-electrolysis processes at scale, the impact on control, safety, degradation and performance, and integration with hydrogen liquefaction or ammonia plants. Design for sustainability implies that life cycle parameters need to be treated as objective functions in plant optimisation. We recommend that LCA and NEA are integrated with project planning to inform decision making to ensure that hydrogen meets the goals of sustainable production.
—Palmer et al.