As green hydrogen production expands, oxygen should no longer be treated simply as a by-product.

It could become a useful industrial feedstock in its own right. That creates opportunities, but it also puts new demands on gas purity measurement and process safety.

Most discussions about water electrolysis understandably concentrate on hydrogen. Hydrogen is the energy carrier, reducing agent and potential replacement for fossil-derived feedstocks. Oxygen tends to appear on the other side of the equation and receive rather less attention.

A recent ChemSusChem perspective by Martin Held and Jan Backmann makes an interesting case for what the authors call “green oxygen”: oxygen produced alongside hydrogen by electrolysis powered by renewable electricity. Their argument is that increasing electrolyser capacity could produce enough oxygen for it to become part of a wider industrial decarbonisation infrastructure rather than simply a secondary stream to be vented.

The numbers make the point rather well. Water electrolysis produces approximately eight kilograms of oxygen for every kilogram of hydrogen. Citing the IEA Global Hydrogen Review 2025, the authors consider a scenario of 28 million tonnes per year of electrolytic hydrogen production by 2030. That would correspond to roughly 220 million tonnes of oxygen per year. At that scale, oxygen stops looking like an inconvenient by-product.

From by-product to industrial resource

There is nothing new about industrial oxygen. It is already widely used in metals, chemicals, combustion, glass manufacture, pulp and paper, wastewater treatment, healthcare and many other applications.

Linde, for example, already offers oxygen produced using certified carbon-free electricity. It points to applications ranging from furnace enhancement and oxidative bleaching to wastewater treatment and oxygenation. The basic commercial idea is straightforward: where an industrial process already consumes oxygen, changing the way that oxygen is produced can contribute to reducing the associated carbon footprint.

What is more interesting is what happens when large hydrogen projects begin producing substantial quantities of oxygen at the same industrial sites where other decarbonisation processes are being developed.

The ChemSusChem paper suggests that this could lead to integrated networks involving three gases: hydrogen, oxygen and carbon dioxide. Hydrogen provides reducing power and energy. CO₂ becomes something to capture, transport, store or convert. Oxygen can support processes that make that capture or conversion more efficient.

This is quite close to how successful industrial complexes have always developed. One plant’s secondary stream becomes another plant’s feedstock. Nothing particularly glamorous about it, but industrial integration has been saving energy and money for decades.

Oxy-fuel combustion is an obvious opportunity

One particularly interesting application is oxy-fuel combustion. Conventional combustion uses air, meaning roughly four volumes of nitrogen accompany each volume of oxygen into the process. Much of that nitrogen eventually leaves with the flue gas. If CO₂ subsequently needs to be captured, it must therefore be separated from a considerably diluted gas stream.

Using oxygen or oxygen-enriched gas instead can produce a much more concentrated CO₂-containing exhaust. The ChemSusChem authors argue that oxygen generated by electrolysers could therefore make oxy-fuel combustion and subsequent carbon capture more attractive, particularly where hydrogen production, industrial combustion and CCS or CCU facilities are geographically integrated.

There are other possibilities. Supplying oxygen rather than compressed air can improve oxygen transfer in wastewater treatment, where aeration represents a substantial part of the plant’s electricity consumption. The paper also discusses aquaculture and even large-scale oxygenation of depleted natural waters.

Not every application will make economic sense. Oxygen is relatively inexpensive compared with hydrogen, while compression, purification, storage and transport certainly are not. This strongly favours local industrial integration rather than moving oxygen over unnecessary distances.

There is another side to the oxygen opportunity

For process engineers, however, commercial utilisation is only half the story. An electrolyser deliberately produces two gases that should remain separated: hydrogen and oxygen.

That means oxygen purity is not merely a product-quality issue. The presence of hydrogen in an oxygen stream can become a process-safety issue, just as oxygen contamination of the hydrogen stream can be.

The requirement becomes more important during dynamic operation. Renewable-powered electrolysers may experience changing loads, start-ups, shutdowns and pressure variations rather than sitting indefinitely at one comfortable steady-state operating point.

If oxygen is going to move from a vented secondary stream to a commercially useful product, confidence in its composition becomes still more important.

Before oxygen can be compressed, stored, piped to another process or supplied as a product, the operator needs to know what is actually in it.

And what happens when that oxygen meets CO₂?

The green-oxygen concept also creates an interesting measurement problem further downstream. If green oxygen supports oxy-fuel combustion and carbon capture, the resulting process chain eventually produces CO₂-rich streams that must themselves meet increasingly demanding specifications.

Oxygen can enter CO₂ streams through compressor seals, valves, shutdown operations or other points of atmospheric ingress. Depending on the application, oxygen may need to be controlled at ppm concentrations.

This is already relevant to CO₂ pipeline transport, EOR and CCUS. Modcon’s work on CO₂-rich process streams identifies continuous trace oxygen measurement in the 0–100 ppm range as relevant to pipeline specification and process protection.

This is a rather good example of why the emerging hydrogen–oxygen–CO₂ economy cannot be considered as three separate gas industries. They increasingly meet each other.

Why trace oxygen in CO₂ is not an easy measurement

Measuring tens of ppm of oxygen in a predominantly CO₂ stream sounds simple until the complete measurement system is considered.

At these concentrations, a small amount of atmospheric air entering an extractive sample system through a fitting or connection can become significant. Sample tubing, calibration-gas quality and fitting integrity can contribute more uncertainty than the sensing principle itself.

Pressure adds another complication. Optical fluorescence-quenching measurement responds fundamentally to oxygen partial pressure. The sensing element contains a luminescent material excited by light. Molecular oxygen quenches the fluorescence, changing its lifetime; that change can be related to oxygen concentration through the Stern–Volmer relationship.

If total process pressure changes substantially, compensation is therefore required if the analyzer is to report concentration correctly.

This becomes particularly relevant as CO₂ moves from relatively low-pressure process gas through dehydration and compression towards dense-phase pipeline conditions.

Bringing the measurement to the process

One way of dealing with these problems is to avoid creating an elaborate sample system in the first place.

The MOD-1040 optical oxygen analyzer uses fluorescence-quenching technology for direct oxygen measurement. Its lifetime-based phase-shift technique determines oxygen from the change in fluorescence lifetime rather than simply measuring optical intensity. The instrument incorporates process pressure and temperature compensation so that changes in operating conditions can be accounted for in the reported oxygen concentration.

For CO₂ applications, the important engineering point is the possibility of in-situ installation directly on the process.

This can eliminate or substantially reduce sample extraction, long tubing runs, pressure reduction and the associated opportunity for atmospheric oxygen ingress. In high-pressure applications, the MOD-1040 can be installed directly at the process rather than first reducing the gas to analyser pressure.

The same principle can be applied at different points through a CO₂ processing train. The application arrangements developed by Modcon cover wet CO₂ upstream of dehydration, dry gas downstream of dehydration and high-pressure dense-phase or supercritical CO₂ at custody-transfer conditions. The installation details change with pressure, moisture and temperature, but the measurement objective remains the same: measure oxygen as close as reasonably possible to the process that matters.

Green oxygen needs good measurement

The interesting message from the emerging green-oxygen discussion is therefore broader than finding another market for electrolyser oxygen. As hydrogen, oxygen and captured CO₂ become increasingly connected, gas composition becomes part of the infrastructure.

Hydrogen producers need to know that their product remains within specification. Oxygen intended for industrial use needs appropriate purity and safety monitoring. CO₂ moving towards utilisation, sequestration or pipeline transport needs its own impurity measurements. The more these streams are compressed, traded and interconnected, the less satisfactory occasional laboratory verification becomes.

The ChemSusChem paper makes a convincing argument that oxygen could become an important part of future integrated low-carbon industrial networks. Linde’s commercial approach already demonstrates that lower-carbon oxygen is moving beyond the theoretical stage. There is a useful engineering conclusion behind all of this.

The hydrogen economy may produce far more than hydrogen. But turning its other gas streams into useful products will depend on knowing their composition continuously, reliably and at the point where the measurement actually matters.

That is where process analysis earns its keep.