Demand for commercial electrolysis technologies, proton exchange membrane (PEM) and alkaline water electrolysis (AWE), has grown exponentially over the last decade, since these are currently the only commercially available methods with which to produce hydrogen without any direct emission of CO₂. Note that, as with all other technologies, the materials from which they are made contain embodied CO₂ so that they are not yet carbon neutral (Iyer et al., 2024). Strong recent demand has resulted in numerous manufacturers establishing plants with production capacities of > GW/year, using modular systems and to achieve rapid reduction in the cost of electrolysers through learning-by-doing. Nevertheless, even replacing the current grey hydrogen production of 93 Mtpa, would require 940 GW of renewable electrolysis (IEA, 2024). Furthermore, meeting the IEA’s projected demand for hydrogen of ∼550 Mtpa by 2050, would require an additional 200 GW electrolyser capacity per year to reach the ∼5,000 GW required, together with the additional infrastructure for renewable electricity generation. These figures highlight the immediate need to increase production of electrolysers.

Annual global shipment of electrolysers is predicted to exceed 3.6 GW, with the largest increase in demand coming from the Americas and Europe. The dominant class of technology is still alkaline technology, although its share is falling from 89% in 2022 (with 11% PEM) to 71% in 2024 (28% PEM). PEM is gaining popularity in regions other than APAC, especially in the US, where most local manufacturers are supplying PEM. Meanwhile, electrolyser manufacturing capacity is growing faster than the demand. There were more than 100 electrolyser makers in 2023, and their total stack assembly capacity was about ten times as much as the expected shipment. Despite the general overcapacity, investors believe explosive demand rise and therefore more money is flowing into electrolyser manufacturing industry, with China remaining the largest producer of electrolysers globally.

China remains by far the cheapest at producing alkaline electrolysers, selling at approximately 1/3 the price of western countries. This is mostly because of the lower cost of engineering, procurement and construction (EPC) there. Outside of China, the cost of alkaline electrolysers is only marginally less than that for PEM systems.

International politics, together with uncertainty in the quality of low-cost manufacturing and perceived advantages of PEM, are contributing to many regions to source their procurement regionally, which also increases sovereign capability. For example, a 10-fold increase in electrolyser manufacturing capacity is forecast in both Europe and the United States in the coming years. Nevertheless, an analysis of the complaints reported by BNEF shows that failures have little correlation with the nationality of the manufacturer (BloombergNEF, 2024). Instead, data released by ITM Power shows a stronger correlation with the time from the inception of the company. A high failure rate of stacks (25%–40%) occurs during the early years of company inception but decreases (to ∼15%) with experience, indicating a learning curve is present in electrolyser manufacture.

Electrolyser manufacturers are lowering costs and increasing scale by developing standardised modules with a typical capacity of individual stacks of some 10 MW (e.g., Peric and FFI) to 17.5 MW for Siemens (see emerging electrolysis session). Another trend, for both AEM and PEM, is an increase in current density to obtain more yield for same size of stack. For PEM, companies are reducing the amount of iridium used as a catalyst to reduce price; for example, Plug Power has reduced this component to 0.2 kg/MW. Nevertheless, further reduction in Pt or Ir use is not expected to lead to significant savings, with improvements in ionomers and membranes expected to have a greater impact. Electrolysers are also invariably connected “behind the meter”. This is partly because it is often challenging to gain access to the grid if you are a large consumer of energy, and because the power from the grid can sometimes be too unstable (Smith et al., 2022). Nevertheless, most present systems in Germany are on-grid, accessing predominantly wind-based power, but will be connected to users via pipeline, such as that planned at the Port of Rotterdam, owing to the lower cost of transporting hydrogen in a pipeline than expanding grid capacity over such distances (IEA, 2024).

The Siemens technology is based on the Silyzer 300 PEM technology, consisting of 24 stacks within a 17.5 MW module with >75% efficiency producing 333 kg H₂/h. The Siemens Gigafactory in Germany opened in November 2023 with a rated annual capacity of 1 GW. Recent developments have focused on ease of maintenance, while cost reduction is anticipated through continued optimisation and standardisation of the sub-systems (transformer, cooling, compressors), a high-level of automation, ensuring a stable supply chain and reduction in on-site installation processes. The rapid upscaling and roll-out is generating significant learnings about how to continue to lower costs. These include knowledge of how to most effectively meet the range of various industry standards and regulations, which vary at the local, state and national levels (Hablutzel et al., 2023; FCW, 2024).

One example of advances in regulation is in the state of South Australia, where the state government has recently introduced a hydrogen and renewable energy act to legislate the industry under a single act. This aims to streamline investment and make it simpler to zone areas for hydrogen production (GovSA, 2024a). Developing reliable risk management processes (financial, project and safety), and sourcing local skills, together with developing the capability to build, operate and maintain the plant is also important.

Plug Power introduced electrolyser manufacturing into its portfolio to provide hydrogen to its fuel cell product line. They are the largest PEM manufacturer in the world and currently supply >40 tonnes of liquified H₂ per day to their customers. Their 5 and 10 MW electrolyser systems are based on the 1 MW Allagash PEM electrolyser stack. The 1 MW Allagash PEM electrolyser stack can produced up to 0.4 tonnes per day (TPD) of hydrogen (Valdez, 2023).

A current example of a driver for large scale electrolysers is in the state of South Australia, which has committed to building a 250 MW electrolyser coupled to a 200 MW hydrogen power plant and hydrogen storage facility, since awarded to Atco and BOC Linde group, to be operated with PEM technology and hydrogen turbines from GE (Day, 2023). This is driven by a combination of opportunities derived from significant periods of negative power prices, whilst also strengthening grid stability, within a state-grid that has the highest per capita uptake ure 2 of rooftop solar in the world, together with high penetration of wind. The combination of outstanding renewable energy resources and consistent government policy has helped the state transition from 100% fossil fuel derived electricity in 2007 to 70% renewable by 2022 and numerous days with >100% of the energy is derived from a combination of wind and solar (GovSA, 2024b).

The details of the stack specifications are often limited by regulation, (e.g., the need to keep below 1,500 V input to the rectifier). The size of an electrolyser stack is limited by regulation (e.g., the need to keep below the European Union low voltage directive, <1,500 Vdc). For Plug Power, this results in an electrolyser stack up to 200 cells. Rectifier power ratings limit the electrolyser stack cell area to approximately 1,000 cm². They have multiple electrolyser projects greater than 100 MW across Europe and United States, sourced from renewable electricity, including from hydro-electricity from Niagara Falls (Valdez, 2023). In addition to the anticipated role for large, centralised hydrogen production facilities (>10 MW), Neuman and Esser (NEA) group anticipate a need for smaller (≤5 MW) decentralised units. Decentralized plants are anticipated to be sized for individual users, with a specific need for hydrogen or its derivatives (e.g., ammonia). The 5 MW electrolyser module from NEA is sized to fit into a shipping container, simplifying transport, with a voltage set to allow classification low voltage (<1,500 V). These same modules can be configured within large systems, which significantly lowers cost, with capacity to turn off individual modules for maintenance (Viktorov, 2023).

It was also noted that hydrogen production plants have a large footprint, owing to the many ancillary systems. For example, a 100 MW plant producing 50 tpd requires about 3 acres of land. Secure supply chains and balance of plant are also very important, with any disruptions to supply adding to manufacturing costs, while reduction in the price of well-established components, such as valves and large-scale compression, having a big impact (IRENA, 2020). Other opportunities for further cost reduction from technology developments include improved power supply, potential to operate with higher current density and further developments of compressors to accommodate large volume of hydrogen.

The wide range of different regulation and standards around the globe is a further barrier to cost, with any moves to adopting international standards offering significant potential to lower costs further (IRENA, 2020). Current standards also need further development, since many do not address important aspects of product design or output pressure. This results in manufacturers requiring third parties to oversee these aspects to report on their safety. The third parties are often European or US based (due to perception of prestige), expensive and ultimately increases cost.

Figure 2 (Liu et al., 2024) and Table 1 presents a summary of the main types of current and emerging electrolysis technologies, namely, solid-oxide (SOEC) and anion exchange membrane (AEM) electrolysis, together with the difference between them and the PEM and AWE counterparts.

Solar photocatalytic water-splitting has potential to contribute to sustainable and clean hydrogen production owing to its reduced need for infrastructure relative to conventional “green” hydrogen. This technology involves the use of photocatalysts, which are materials that can absorb solar radiation and use this to split water into hydrogen and oxygen. This is achieved by generating an electron-hole pair, which can then react separately with water to produce hydrogen and oxygen at the surface of the catalyst. There are degrees to which the solar absorber and redox reactions are integrated but simply these can be divided into photoelectrochemical (PEC) water splitting devices, which have separate cathode and anodes, and photocatalytic (PC) water splitting devices, in which the chemistry occurs on the same particle. A detailed description of the processes and can be found elsewhere (Kim et al., 2019). Although there are still technical challenges to overcome, such as increasing the efficiency of the photocatalysts and scaling up the technology for practical applications, solar photocatalytic water-splitting holds great promise as a sustainable and carbon-free method for hydrogen production. Several TEAs estimate that photocatalysis have potential for to achieve LCOH of below US$2/kg (Foulkes et al., 2017; Pinaud et al., 2013), if they can achieve a sufficient solar-to-hydrogen efficiency (STH). It is plausible for it to follow a rapid cost reduction, analogous to what occurred with PV (Hisatomi and Domen, 2019). Also important for commercially viability is the need to increase the durability of the photocatalyst, reduce the cost of materials, develop the production processes and optimize the performance of the reactors.

Many architectures for direct solar to H₂ are under development, spanning PV-like approaches, particle based, membrane electrodes and hybrid approaches, with STH efficiency of 20% having already been demonstrated for some of them (Wang et al., 2021). A round table by the DOE on liquid solar fuels in 2019 identified 4 research priorities, which can also be directly applied to identified to solar H₂ production. These are: (i) Understanding mechanisms for durability and performance of materials (e.g., understand and halt photocorrosion processes using ultrafast spectroscopy, applying protective coating); (ii) Control catalyst microenvironment to promote selective and efficiency fuel production; (iii) Bridge the time and length scale (hot carrier effects to generate multiple electron-hole pairs); and (iv) Tailor interactions of complex phenomena to achieve integrated multicomponent systems. Important steps needed to achieve this include benchmarking to demonstrate efficiencies and establishing best practice for scale up of devices and demonstration. International consortiums, such as the Mission Innovation Sunlight to X Innovation Community, are important for this.

The University of Michigan is developing III-V semiconductor materials, because of their long term stability and potential to tune the band-gap across the visible spectrum (Chowdhury et al., 2018). Materials such as InGaN can generate 16.7 mA/cm² photocurrent density, which could correlate with a solar-to-hydrogen (STH) efficiency of 20%. Their tandem photoelectrodes made of InGaN/Si nanowires have been demonstrated to split water with an STH of 9% (Zhou et al., 2023). A key challenge of the large lattice mismatch could potentially be overcome by doping with As or Sb at low levels of even 1%, which has potential to pave the way for a 20% STH photocatalyst.

SunHydrogen is seeking to develop PEC devices using low-cost, earth-abundant materials, and easily scalable approaches. Their latest device is based on fault-tolerant, monolithic photoactive heterostructures integrated with catalysts and protective coatings. It has achieved an active-area STH approaching 10% for devices up to 1,200 cm², and efficiencies over 10% for 100 cm² devices (Mubeen, 2023). Similarly, SoHHytec are progressing a 100 kW scale demonstrator consisting of 5 × 10 m diameter dishes to provide ×1,000 concentration onto a novel photoelectrochemical water-splitting device (Haussener, 2023). Concentration is expected to be crucial for photocatalysis technologies to advance towards commercial reality.

The detailed spectroscopic characterisation of photo-catalysts, particularly in operando conditions, has been used to advance understanding of how photocatalyst performance can be improved. This was used to explain why the performance of photoelectrodes based on GaN/Si improved over 10 h and then stabilised, due to the formation of a protective layer of GaON at grain boundaries (Xiao et al., 2023). This increased stability was confirmed by DFT calculations. It was also used to show that the reaction efficiency of the CO₂ reduction reaction on Cu₂O to produce C₂H₄ can be enhanced using a carbonate electrolyte, through a new species in the XPS O 2p signal that does not occur in different electrolyte (Zhang et al., 2024). It has also been useful in recent efforts to produce a high performing and stable water splitting device using a GaN-based PEC at neutral pH, which achieved a 5% STH with 80 h stability. As part of this work, they proposed a methodology describing best practice for measuring photocatalyst devices entitled “Best practices in PEC: How to measure solar-to-hydrogen efficiency of photocathodes” (Alley et al., 2022).

Together with collaborators, Toyota is developing a water-vapor fed photoelectrode device using ambient air (Zafeiropoulos et al., 2019). Replacing liquid water with vapour has several advantages including simplified reactor design, avoiding the requirement for fresh water and preventing formation of bubbles that can block catalytic sites. They have developed porous photoanodes which could achieve 80% of liquid phase performance at 80% relative humidity (RH) (Zafeiropoulos et al., 2021), and developed a device that achieves 1 mA/cm² at ambient humidity. This was further expanded to photocathodes during the European project (Caretti et al., 2023), Sunlight-To-X. During the project, a TEA of the LCOH showed a small cost advantage could be found by switching from liquid water to a water vapour feedstock, however most of the cost arises from the PEC components and is strongly affected by the STH performance (Savant et al., 2024). The next step of their Sun-to-X project is to scale up to 1 m².

The University of Warwick is developing PEC operation under microgravity environments, which is of interest for operation into space such as on the International Space Station (Ross et al., 2023). Water harvesting on the Moon and Mars would require solar concentrators in the cold environments and could also be useful for terrestrial applications (Ross et al., 2023). Also important in space is device weight and volume, more so than cost, together with recyclability of materials. Direct water splitting in space is also important because the oxygen generator assembly on the ISS consumes 1/3 of the energy for entire life support system. A major challenge of operation under microgravity is the absence of buoyancy, which reduces gas bubbling from the electrodes thus limiting mass transfer (i.e., the bubbles build up as froth), for both H₂ and O₂ although with different effects. The wettability of the electrode is crucial and addition of polystyrene nanoparticles is found improve bubble formation. The use of magnetically-induced buoyancy, in combination with hydrophilic photoelectrodes, was found to be mostly resolve microgravity issue.

The two fundamental drivers for continuing the development of thermochemical technology for hydrogen production from water splitting are, firstly, that the rate of chemical reactions increase exponentially with temperature via the Arrhenius equation and, secondly, that the direct use of heat (e.g., to drive a chemical reaction) offers an increase in efficiency of at least 2, relative to converting a high temperature heat source to hydrogen via electricity. The two main options for net-zero sources of high temperature heat are concentrated solar thermal energy (Steinfeld, 2005) and nuclear energy (Constantin, 2024). While thermochemical water-splitting technologies are still pre-commercial, a complete system for solar liquid fuels via hydrogen has recently been demonstrated at lab-scale (Zoller et al., 2022) and numerous components have been demonstrated at pilot-scale (Thomey et al., 2016; Agrafiotis et al., 2023; Flamant et al., 2023). The recent advances in these technologies that are the focus of the present review are advances in hybridising thermo-chemical technology with electro-chemical routes to lower the reaction temperature, together with those in both materials development and in better understanding of the techno-economic potential of these pathways.

The potential to extend the use of concentrated solar thermal energy to drive chemical reactions seeks to leverage advances in both receiver and heliostat technology, together with those in upscaling and in thermal energy storage, that has enabled lower temperature solar thermal systems to become an important contributor to schedulable electrical power at GW scale (Flamant et al., 2023). A major step in the development of thermo-chemical pathways for water-splitting is the two-step pathway that lowers the temperature of the reaction below the >3,000 K (64% dissociation at 1 bar) required for the direct thermolysis of water and also offers a path to avoid recombination during cooling (Steinfeld, 2005). However, more recent developments offer a further step-change in temperature reduction over the ∼1,800 K needed for two-step reactions. The desire to achieve temperatures below that of two-step reactors is driven by the nonlinear trade-off between the benefits of increased reaction rates that occur with higher temperature and the costs of both achieving safe and reliable operation at higher temperatures and those of increase in the cost of concentration ratio of the heliostat field which, in turn, results from the less-than-parallel nature of the solar insolation. Furthermore, achieving reduction temperatures of ∼1,000 °C also offers the possibility of compatibility with gas-cooled nuclear technology.

Two-step thermo-chemical reactors cycle a catalytic media through two reactions, an endothermic reduction driven with the concentrated solar thermal energy that reduces water to hydrogen and oxygen, and an exothermic oxidation reaction that releases the oxygen to allow the material to be cycled again (Bader and Lipiński, 2017), as follows: Reduction : 1 / δ   M x ‐ δ + H 2 O   →   1 / δ   MO x ‐ δ + H 2 (1) Oxidation : 1 / δ   MO x + →   1 / δ   M   x ‐ δ + 0.5 O 2 (2)

To this end, effort has focussed both on the development of the reactors that are suited both to the harnessing of the concentrated radiation to drive the reduction, followed by its subsequent oxidation, and to the development of the novel catalysts targeted to lower the reaction temperatures. Both the range of reactor types and many of the materials that have been considered previously have been identified in previous reviews (Steinfeld, 2005; Bader and Lipiński, 2017).

An important recent milestone in the development of the reactor technology itself is the recent end-to-end demonstration of the production of synthetic jet fuel from the splitting of both H₂O and CO₂ at ETH Zurich, with a solar to syngas efficiency of 4.1% without the application of heat recovery (Zoller et al., 2022). This level of efficiency has been sufficient to attract ongoing investment in the further development of the technology through company Synhelion. Nevertheless, the reduction reactor operates with a peak reaction temperature of 1,500 °C, which requires specialised optics to achieve the concentration ratio of some 2,500. Furthermore, as it typical of these Red-Ox cyles, the temperature of the oxidation process is lower than that for reduction by some 500 °C, which results in significant exergetic losses in cycling the catalyst through this temperature ranges. Addressing these challenges has therefore become a target for further research.

A wide range of catalysts have been explored to date, particularly targeting perovskite-based and ceria-based media, although many other media including manganese and cobalt-based catalysts have also been explored (Bader and Lipiński, 2017), typically reducing at temperatures above 1,400 °C. In addition to seeking to lower the temperature of reduction, it is also desirable to reduce the temperature differential between the two reactions and to increase the stoichiometric ratio between them, and thereby the amount of oxygen that can be exchanged in each reaction. Suitable catalysts must also be robust and avoid degradation through multiple successive cycles. One option to help to avoid sintering is the cycling between off-stoichiometric conditions, an approach that can also help to increase reaction rates (Wexler et al., 2023) – albeit at the expense of lower conversion per cycle. Another option under development is a program to develop new materials that offer improved performance over currently known materials. Toward this end, simplified models have recently been developed that link crystallographic features to thermo-chemical behaviour (Wexler et al., 2021). These models have also been used with artificial intelligence to greatly accelerate the screening of potential new materials by Witman et al. (2023). This has led to the discovery of new materials (e.g., CCM 2112) with order of magnitude more hydrogen production mol per mol of atom than known previously (Wexler et al., 2023).

Importantly, the temperature needed to achieve a given level of conversion in each reactor also depends on the log of the partial pressure of the product gas, as shown in Equation 3. This shows that the difference between the enthalpy of reaction relative to the reference condition, Δ h o , and that of the entropy, Δ s o , balances the pressure dependent term, where δ is the stoichiometric parameter defined in Equations 1 and 2. For this reason, the reduction reactor has typically been operated either under vacuum or in the presence of a diluent gas, to lower the reactor temperature, while the oxidiser is typically operated at a somewhat higher pressure to reduce the temperature difference between the two reactors. A typical level of pressure swing that has been chosen is to operate at some 10^–5 atm for reduction and atmospheric pressure for oxidation (Panlener et al., 1975). Nevertheless, even with this pressure swing, the temperatures of the reduction and oxidation reactors, respectively, are still some 1,500 °C and 1,000 °C (Zoller et al., 2022).

An important recent development, which offers potential to lower both the temperature of reduction and that differential temperature is the hybridisation of these reactions with electrical voltage, as illustrated by Equation 4. Figure 3 from Perry et al. (2023) shows that it is possible to achieve isothermal cycling between the reduction and oxidation reactors with an applied voltage of 0.5 V at a temperature of some 970 °C (Perry et al., 2023). Even lower temperatures can be achieved with higher voltages. Not only does this significantly lower the challenge of driving the reaction with concentrated solar thermal energy, it also increases the viability of coupling it with thermal energy storage, which will increase the utilisation of the thermo-chemical plant, or of driving the reaction with nuclear energy. Δ h o δ ‐ Δ s o δ T red + R T red ⁡ ln p O 2 / p total 1 / 2 = 0 (3) Δ h o δ ‐ Δ s o δ T red + R T red ⁡ ln p O 2 / p total 1 / 2 + 2 F E red = 0 (4)

A range of other hybrid approaches with electrochemical pathways have also been developed recently. Sandia National laboratories have recently begun thermo-chemical water-splitting with liquid-metal solutions of Zn and Zn-O which allows reduction to occur at even lower temperatures T < 823 K (McDaniel et al., 2022). Another hybrid electro-chemical approach to lower the reduction temperature is the hydrogen sulphur cycle being developed by DLR (Agrafiotis et al., 2019). This approach employs electrolysis of water plus SO₂ to produce sulphuric acid, which can be decomposed thermally at temperatures of around 800 °C. selects an alternative reaction to lower the temperature of the thermal input employs electrochemical energy.

Even without the introduction of an applied voltage, the techno-economic analysis of DLR (Prats-Salvado et al., 2022) has already identified scenarios is which liquid fuels with solar thermal energy is becoming prospective. They project production costs by 2050 of €1.3/L and €1.7/L for gasoline made from hydrogen from solar thermal with CO₂ from industrial sources and direct air capture, respectively, allowing for projected cost reductions in solar thermal technology. This is consistent with the US$3/L of methanol from direct air capture of CO₂ using current costs (Perry et al., 2023). These estimates are broadly consistent with the estimates by Nathan et al. for US$13/kJ of high temperature industrial heat, allowing for similar cost reductions in solar technology (Nathan et al., 2023).

Two other hybrid cycles with solar thermal energy are worth noting, the hybrid sulphur cycle under development by DLR and the hybrid biomass dual fluidised bed gasification cycle for biomass under development by the University of Adelaide. The hybridisation of these processes also allows the temperature at which the energy is added to be below 1,000 °C, increasing the viability of production at scale. One additional driver for the hybrid sulphur cycle is the potential to employ sulphur as a long duration energy storage media, while a key driver to introduce solar thermal into a biomass gasification process is the potential both to increase the net amount of CO₂ that can be captured in the char and that to increase the revenue from high value carbon co-products (Ngo et al., 2024).

In summary, significant developments have occurred recently to demonstrate the viability of thermo-chemical hydrogen production with concentrated solar radiation and to increase the viability of harnessing the increase in reaction rates that occur at elevated temperature both by hybridisation, to lower the reaction temperature, and by the use of modern methods to accelerate the development of improved catalytic materials.

A key recent driver for the ongoing development of technology to utilise biomass resources, using either thermal or biological conversion pathways, to produce hydrogen-rich fuels is the potential to simultaneously achieve low-cost removal of atmospheric CO₂ into the char. Similarly, several key drivers to continue the development of production of hydrogen-rich fuels from refuse-derived fuels are the need to increase circularity (avoiding landfill), to achieve a low-cost, low-carbon fuel by avoided land-fill costs, and because their syngas product is well suited to those applications, such as industrial processes, that do not require high purity hydrogen.

There is growing interest in the thermal decomposition of methane as a complementary route to the other methods and the potential for low-cost, carbon-neutral production in addressing different market niches. Methane pyrolysis is expected to be able to achieve net-zero life cycle emissions if it both utilises a suitable feedstock, such as by blending some 15%–30% bio-gas-derived methane with the natural gas (Diab et al., 2022), together with a mix of sufficiently long-lived carbon-rich products, such as the high-volume materials that are needed by our society, such as construction materials, asphalt, iron carburisation and soil additives, together with high value-low volume materials, such as graphitic battery materials (Parkinson et al., 2018). Furthermore, the energy requirement for methane pyrolysis is less than required for water electrolysis (vide infra) and is also less than kWh requirement for SMR, and requires no water. However, the mass of the carbon co-product is three times higher than that of the hydrogen, and the amount of hydrogen product per unit of methane is half that of methane pyrolysis. Hence its viability, both environmentally and economically, depends on being able to produce the right mix of carbon products with sufficient value and volume. Four alternative classes of methane pyrolysis technology are also under development, notably thermal plasmas, microwave, fluidised bed and molten media catalysis, each of which is best suited to different types of carbon materials. Several good reviews are available on these options (Schneider et al., 2020; Patlolla et al., 2023; Sánchez-Bastardo et al., 2021).

Another opportunity for methane pyrolysis is to capitalise on the availability of existing natural gas distribution systems, and/or relatively cost of its transport, to co-locate the hydrogen production with the site where it is needed (Diab et al., 2022). This reduces the barrier of establishing hydrogen transportation systems. Furthermore the solid carbon products from methane pyrolysis will also be easier to transport than CO₂, providing an advantage over SMR, making it more suited to distributed production (Sánchez-Bastardo et al., 2021).

Methane decomposes in the presence of heat to co-produce hydrogen with solid carbon with an overall reaction (Schneider et al., 2020) expressed as: CH 4   →   C s + 2 H 2   Δ H o = 74.9  kJ / mol (5)

This methane pyrolysis is sometimes referred to as “turquoise hydrogen” and requires high temperatures, typically >1,000 °C, although this can be reduced to ∼800 °C if catalysed. Also, the reaction pathway is extremely complex, poorly understood and difficult to control, being closely aligned to soot formation processes. The reaction enthalpy can potentially be driven with green electricity or concentrated solar thermal as a hearting source, which is another important component of minimising the life-cycle CO₂ emissions to the choice of feed-stock and products, described above. However, more work is needed to better understand life-cycle emissions, given the complexity and diversity in the many potential combinations. With relatively high TRLs (approaching to 8 for some approaches), methane pyrolysis requires only some 25% of the of water electrolysis (see Equation 5). The technology can also use existing infrastructure (natural gas pipe lines), is scalable and can be decentralised for on-site applications. However, more work is needed to better understand the techno-economic viability of the various options, which depends significantly on developing the technology to deliver a range of carbon co-products for various markets and the value of these carbon co-products.

The main common technical challenges being addressed with the various methane pyrolysis approaches are to increase the efficiency of hydrogen production and achieve sufficient quality (e.g., purity) in the carbon product. For those involving a catalyst, namely, the fluidised bed and molten catalyst routes, it is also necessary to address potential contamination by, and loss of, the catalyst loss to the carbon stream, together with deactivation of the catalyst (Patlolla et al., 2023). Further research is also needed to develop the mix of solid carbon products for applications in use such as concrete, agriculture, building materials and, in some cases, as a sequestered product (reverse coal mining). Where biomethane and renewable gas are used as the feedstock, the hydrogen can be carbon-emission-negative, offering further value. Some applications also offer potential to make use of the heat in the process. For example, methane pyrolysis offers the potential to supply hot hydrogen to direct reduction of iron ore (DRI) furnaces. This offers significant value over the currently proposed route for hydrogen-DRI, owing to the high cost of heating hydrogen to some 800 °C due to the low density and high specific heat of hydrogen. Further research and technology development is also needed to address the challenge of fugitive methane emissions. The target for fugitive emissions of <0.4% is a long way from the present value of 1%–2% by the US EPA, although these emissions are dominated by incomplete combustion of flares, which is a solvable problem.

Recent research for the use molten media catalysts, through which methane is bubbled, is targeting the use of low-cost molten salts to achieve temperatures of approximately 900 °C (Patlolla et al., 2023). The use of a molten media is driven by the opportunity to continuously refresh the catalyst surface and facilitate solid carbon removal as it floats. The catalyst can also generate selective PAH products which have greater value than graphite. The UQ team is undertaking fundamental experimental research to identify salt catalysts that are low-cost and non-toxic but that have high conversion rate, hydrogen selectivity, with also consideration of co-products derived from PAHs. Combining a low-value salt with a high value one is a promising option, such as equimolar fractions of MgCl₂ + NaBr, which was proposed based on a systematic study of the impacts of anions and cations (Sheil et al., 2023). It was also found that the Lewis acidity of the transition metal cation can play a large role in the overall hydrogen production.

The microwave pyrolysis technology introduces the microwave radiation into a fluidised bed, in which methane is the fluidising agent for the solid carbon particles approximately 0.5 mm in diameter (Deng et al., 2014). The solid carbon is used to absorb microwave energy and heat the reactor to thermally crack the methane. The technique can achieve a very high hydrogen selectivity at temperatures over 1,000 °C, together with a high conversion rate at 1,200 °C. It also has rapid response to changes in demand (i.e., rapid turn on/turn off), making it compatible with heating with variable renewable electricity and an electricity consumption that is 80% less than electrolysis. The technique produces semi-graphitic solid carbons, which are anticipated be relatively high-value products.

An alternative microwave technology of Sakowin employs microwave plasmalysis of methane without a catalyst, which is presently approaching TRL-7. The technology employs large industrial grade plasma sources, which are commercially available with long lifetime magnetrons (∼7,000 h), targeting a modular standardized unit with a production capacity of 200 kg/day hydrogen, which is stackable to up to 100 MW, which they anticipate will potentially reach around €1/kgH₂ (Gatt, 2023).

The thermal fluidised bed process of Hazer is based on low-cost catalysts of iron ore particles, achieves low temperature of approximately 900 °C, to convert natural gas feedstocks into hydrogen and high-quality graphite. Their technology, which is presently at TRL 6-7, produces up to 95% pure graphitic structure which is encapsulated on nanofragments of metallic iron, making them magnetic. The markets they are targeting include both high volume products such as steel, concrete, carbon black and asphalt, together with high value graphite used for batteries or thermal storage and activated carbon for water treatment. The first fully integrated commercial demonstration plant test program was successfully completed in 2024 achieving over 450 h of continuous operation and conversion consistent with large scale commercial design basis, de-risking the scale-up and commercialisation of the technology. Parallel projects are planned for Canada, Japan and France in next few years for further up-scaling to >2,500 tpa by taking the advantage of existing LNG import terminals or power station sites. They anticipate the lowest delivered cost at North America is ∼US$2/kg H₂ because of the low-cost of gas and power there. As comparison, the cost is∼US$5/kg H₂ in the Asia Pacific region (Forbes, 2023).

Hycamite has developed a technique for thermo-catalytic decomposition (TCD) of methane using proprietary catalysts. Lifecycle emission analysis showed a carbon intensity of 0.84 kgCO₂e/kgH₂ using Finnish grid electricity and LNG, or 0.59 kgCO₂e/kgH₂ with renewable electricity and LNG. Negative carbon emissions is also possible if methane from biogas or synthetic methane made with captured atmospheric carbon is used as feedstocks, which can be as low as −2.75 kgCO₂e/kgH₂ if the carbon is permanently stored. An industrial scale demonstration is expected to be operational soon with a production capacity of 2 kt H₂ and 6 kt solid carbon with a 27,400 tpa CO₂ emission reduction (Rasnasen, 2023).

Natural hydrogen, sometimes referred to as “white” or “gold”, is produced from naturally occurring (typically geological, and in many cases continuous) processes and can accumulate in certain conditions in the Earth’s subsurface (Hand, 2023). Unlike the other processes described above, it therefore does not require manufacturing, giving a potential cost advantage although, as with natural gas, purification may be required. Recent years have witnessed growing scientific and commercial interest in the potential production of naturally occurring hydrogen (Boreham et al., 2021; Stalker et al., 2022), spurred by the broader emergence of the hydrogen economy but also by the discovery of the hydrogen-rich Bourabougou Field in Mali which contains 97.4% H₂ (with minor fractions of He, N₂ and C₁-C₅ hydrocarbons (Prinzhofer et al., 2018); and has been used for local electricity generation since ∼2015. Nevertheless, to-date, this is the only known commercially operating natural H₂ accumulation. Estimates from the US exploration company Koloma, suggest that the life-cycle carbon intensity of natural hydrogen production is ∼0.1 kg CO₂s/kgH₂, which is comparable with those from electrolysis, while also requiring considerably less energy than either that or steam methane reforming (Brandt, 2023). These developments have spurred increasing global activity in natural hydrogen exploration, with recent drilling either planned or occurring in the US, Spain and Australia, and the announcement of a discovery in France in 2023, where the mining law has recently been updated to include natural hydrogen (Hand, 2023). Much of this activity has been funded through private sources, with exploration efforts globally having received relatively little public funding to date.

Trace amounts of natural H₂ have been discovered in many natural gas accumulations in Australia and elsewhere although, until the last few years, the presence or absence of hydrogen has not been routinely tested from drill holes and deep mines (Boreham et al., 2021). Whilst the vast majority (∼80%) of existing measurements indicate <0.01 mol% H₂, a number of accumulations with >10% mol% H₂ have been reported, mostly associated with rocks of Precambrian age. Several notable occurrences have been described in the state of South Australia, including a 1931 well which discovered up to 90% H₂ in the Yorke Peninsula. South Australia was the first jurisdiction in Australia to establish a regulatory framework for natural hydrogen exploration, and as a consequence an initial series of exploration wells were drilled in 2023 by a company called Gold Hydrogen, with their results indicating downhole concentrations of up to 96% H₂ and 36% He (Gold Hydrogen, 2024). It is also known that crystalline salt sequences are likely to play an important role in the trapping of subsurface H₂ accumulations, since such media are among the most effective in trapping such a small molecule (Bradshaw et al., 2023).

Natural hydrogen can be derived from either biogenic (microbial and thermogenic) or abiogenic geological sources, with the latter thought to be far more volumetrically significant. Most geological sources of natural hydrogen involve some degree of fluid-rock interaction, and, whilst over 30 distinct subsurface generation mechanisms have been proposed, the most commonly reported include the degassing of magmas and primordial H₂ from the Earth’s core and mantle, the oxidation of divalent iron (Fe²⁺) rich rocks and minerals through serpentinization, the natural radiolysis of water in iron-rich rocks and thermogenic cracking of organic matter (Boreham et al., 2021). Critically, radiogenic decay can also generate ⁴He which can be used as a tracer for H₂ production and help discriminate between geological sources (Lollar et al., 2014), whilst representing a scarce and valuable resource in its own right (Danabalan et al., 2022). Research by Geoscience Australia has found that natural hydrogen represents an enormous potential resource, with H₂ production rates from water radiolysis and serpentinization can provide an inferred resource potential between ∼1.6 and 58 MMm³ year for onshore Australia down to a depth of 1 km (Boreham et al., 2021). Nevertheless, the scale of these accumulations is more likely to be much smaller than the large (up to multi-TCF) deposits associated with natural gas fields.

Figure 4 illustrates both the similarities to, and differences from, the formation of hydrogen and the well-known natural gas accumulations underground (Jackson et al., 2024). Petroleum exploration typically involves the identification of both a source rock and its genetic link to a trapped resource (Jackson et al., 2024). This typically involves the thermogenic generation of hydrocarbon molecules from an organic-rich sedimentary rock, which then migrate upwards due to buoyancy and accumulate in a high-permeability reservoir rock that is sealed by an overlying low-permeability rock. However, whilst petroleum systems are almost exclusively confined to sedimentary rock sequences, hydrogen systems are typically extra-basinal igneous or metamorphic rocks, so differ fundamentally with respect to the nature and diversity. The ideal temperature range for source mechanism of hydrogen generation by serpentinization, the ideal temperature is between 200 °C–350 °C, with laboratory experiments suggesting H₂ production rates of >200 tonnes per year per km³ of peridotite (Jackson et al., 2024). If produced, hydrogen molecules are expected to migrate into adjacent or overlying sedimentary rocks in a manner similar to hydrocarbon molecules, albeit with greater diffusion rates.

The broad similarity of natural hydrogen to hydrocarbon geology implies that learnings and technologies from the oil and gas industry can be effectively repurposed for natural hydrogen exploration and production. Nevertheless, since the production mechanisms are different, the rates of diffusion are so much higher and research has started much more recently, much less is known about natural hydrogen and substantial research is needed to fill the gaps in knowledge. In particular, the role and availability of water in facilitating natural hydrogen production is poorly understood, particularly in relation to serpentinization processes in high-temperature continental settings. With relatively few operational projects, emissions intensity data for natural hydrogen production are lacking, though the panel discussion noted a recent life-cycle analysis by Brandt (2023) which for a baseline archetypal case estimated a site-boundary greenhouse gas intensity of ∼0.4 kg CO₂eq. GHG per kg of H₂ production, with the largest sources of emissions being fugitive losses from the system and embodied emissions in constructed wellbores and equipment.