Low-carbon hydrogen production routes

How do three routes to hydrogen from water compare on efficiency or energy use, longest demonstrated operation, largest scale built and main limitation?

This table sets out three routes to hydrogen from water on four dimensions, using only figures from the sources listed. Most figures come from laboratory, prototype or demonstration studies, and empty cells mark where no reliable public figure was found.

Low-carbon hydrogen production routes. Each cell shows a value, how well it is evidenced, and its sources.
Option	Efficiency or energy useSolar-to-hydrogen efficiency (%), or cell voltage (V) at a stated current density (A per cm2). Lab figures unless stated.	Longest demonstrated operationLongest reported test duration in hours, with the type of test stated.	Largest scale builtLargest device, stack or plant reported. Units are kW for stacks, cm2 of module or electrode area, or GW of capacity.	Main limitationThe main constraint that the cited sources report for this route.
Freshwater electrolysis	No reliable public figure	No reliable public figure	No reliable public figure	Installed electrolyser capacity and low-emissions hydrogen volumes remain low. The IEA says developers wait for clarity on government support before making investments (IEA, October 2024).Reported, unverified[1]
Direct seawater electrolysis	About 2.0 V at 0.5 A per cm2 in a 1 kW-scale stack, alkaline seawater, laboratory test (Sun et al., 2024).Demonstrated[4]	Over 3,200 hours at 250 mA per cm2 in a membrane-based demonstration system, without failure (Xie et al., 2022). No independent replication in the sources read.Demonstrated[3]	1 kW-scale stack at about 0.5 A per cm2, laboratory result (Sun et al., 2024). This is the largest seawater device among the sources read.Demonstrated[4]	Seawater components cause side reactions and corrosion at the electrodes. The authors report a design that avoids these (Xie et al., 2022).Reported, unverified[3]
Artificial photosynthesis	11.2% unbiased solar-to-hydrogen, 16 cm2 module, one-sun illumination, laboratory (Hansora et al., 2025).Demonstrated[5]	Photoanode stable for over 520 hours under simulated sunlight, a component test (Liu et al., 2025). The abstract read gives no full device lifetime.Demonstrated[6]	441 cm2 leaf reached 2.7% solar-to-hydrogen in natural sunlight (Liu et al., 2025).Demonstrated[6]	The absorber is lead-based (chlorine-doped FAPbI3). Stability was reported for 140 hours on 1 cm2 photoelectrodes, and the abstract calls the 16 cm2 module stable without giving its duration (Hansora et al., 2025).Demonstrated[5]

Low-carbon hydrogen production routes. Each cell shows a value, how well it is evidenced, and its sources.

Option

Efficiency or energy useSolar-to-hydrogen efficiency (%), or cell voltage (V) at a stated current density (A per cm2). Lab figures unless stated.

Longest demonstrated operationLongest reported test duration in hours, with the type of test stated.

Largest scale builtLargest device, stack or plant reported. Units are kW for stacks, cm2 of module or electrode area, or GW of capacity.

Main limitationThe main constraint that the cited sources report for this route.

Freshwater electrolysis

No reliable public figure

Installed electrolyser capacity and low-emissions hydrogen volumes remain low. The IEA says developers wait for clarity on government support before making investments (IEA, October 2024).Reported, unverified[1]

Direct seawater electrolysis

About 2.0 V at 0.5 A per cm2 in a 1 kW-scale stack, alkaline seawater, laboratory test (Sun et al., 2024).Demonstrated[4]

Over 3,200 hours at 250 mA per cm2 in a membrane-based demonstration system, without failure (Xie et al., 2022). No independent replication in the sources read.Demonstrated[3]

1 kW-scale stack at about 0.5 A per cm2, laboratory result (Sun et al., 2024). This is the largest seawater device among the sources read.Demonstrated[4]

Seawater components cause side reactions and corrosion at the electrodes. The authors report a design that avoids these (Xie et al., 2022).Reported, unverified[3]

Artificial photosynthesis

11.2% unbiased solar-to-hydrogen, 16 cm2 module, one-sun illumination, laboratory (Hansora et al., 2025).Demonstrated[5]

Photoanode stable for over 520 hours under simulated sunlight, a component test (Liu et al., 2025). The abstract read gives no full device lifetime.Demonstrated[6]

441 cm2 leaf reached 2.7% solar-to-hydrogen in natural sunlight (Liu et al., 2025).Demonstrated[6]

The absorber is lead-based (chlorine-doped FAPbI3). Stability was reported for 140 hours on 1 cm2 photoelectrodes, and the abstract calls the 16 cm2 module stable without giving its duration (Hansora et al., 2025).Demonstrated[5]

What the table shows

The freshwater row has no scale figure in this table. The IEA reports 20 GW of electrolyser capacity reached final investment decision globally by October 2024, which is committed rather than built capacity.
The seawater row's longest test is over 3,200 hours at 250 mA per cm2 (Xie et al., 2022), and its largest stack is 1 kW-scale (Sun et al., 2024). Each comes from one laboratory study.
The artificial photosynthesis row's 11.2% figure is for a 16 cm2 module, and its 441 cm2 leaf reached 2.7% in natural sunlight. These are different devices under different conditions.
No row gives a figure for operation at commercial scale in the sources read, and the freshwater row has no efficiency or operating-time figure.

Read with care

The laboratory, prototype and demonstration figures come from different studies that use different materials, areas and test conditions, so the cells are not a controlled comparison.
The measures are not like-for-like. The seawater cell gives a voltage at a stated current density, and the artificial photosynthesis cell gives a solar-to-hydrogen efficiency. Neither is the energy use of a complete plant.
The freshwater row has no scale figure. The IEA's 20 GW is a global total of committed capacity at final investment decision, not installed capacity. The IEA's 2026 installed capacity chart was not read, and its values are marked as estimates.
Most seawater and artificial photosynthesis figures are single-study results, and the sources read do not include independent replication.

Sources

other
Low-emissions hydrogen projects grow as policy support races to catch up
International Energy Agency, 2 Oct 2024. Accessed 21 Oct 2026.
government report
Hydrogen - Production and electrolyser installed capacity
International Energy Agency, Jun 2026. Accessed 21 Oct 2026.
peer reviewed paper
A membrane-based seawater electrolyser for hydrogen generation
Heping Xie, Zhiyu Zhao, Tao Liu et al.. Nature, 30 Nov 2022. Accessed 21 Oct 2026.
peer reviewed paper
Corrosion-resistant NiFe anode towards kilowatt-scale alkaline seawater electrolysis
Xiaogang Sun, Wei Shen, Hao Liu et al.. Nature Communications, 28 Nov 2024. Accessed 21 Oct 2026.
peer reviewed paper
Scalable and durable module-sized artificial leaf with a solar-to-hydrogen efficiency over 10%
Dharmesh Hansora, Rashmi Mehrotra, Eunseo Noh et al.. Nature Communications, 6 May 2025. Accessed 21 Oct 2026.
peer reviewed paper
A standalone bismuth vanadate-silicon artificial leaf achieving 8.4% efficiency for hydrogen production
Boyan Liu, Xin Wang, Yingjuan Zhang et al.. Nature Communications, 21 Mar 2025. Accessed 21 Oct 2026.