Column | Why electrolyser ageing matters more than day-one efficiency

Electrolysers are electrochemical systems. From the first operating hour onward, their performance evolves. This evolution affects efficiency, safety and, ultimately, the economics of hydrogen projects, whether in operation or idle.

The good news is that ageing is not a blind spot. Many of the processes that drive performance changes can be influenced – if they are understood early and addressed from the outset. A lifecycle perspective, therefore, does not begin in operation, but with fundamental design and system choices.

I on the need for standardisation in comparing performance on a like-for-like basis, and on the risks created when metrics are defined differently across the market. That challenge remains. But as projects move from design phase into commercial operation, the focus shifts to how performance changes over time, and which challenges may occur during a plant’s operational life, like gas quality decrease or premature stack exchange due to safety limits or leaks.

The key metric is not power consumption in isolation. Instead energy required to produce hydrogen over the system’s lifetime is more significant. As ageing progresses, rising power demand and declining output directly increase the cost per kilogram of hydrogen.

Applied materials, design philosophy and operational profile therefore have a decisive impact on reliability and long-term operational costs – especially at scale.

Ageing is not a single curve

Electrolyser ageing does not follow a simple or linear trajectory. It is shaped by a complex interaction of factors that vary across installations. Load fluctuations, start-stop cycles, temperature, purity of media like water or lye, and operating pressure all influence how quickly and in which way degradation occurs.

Two identical systems operating under different conditions can have markedly different lifetimes. A useful comparison is a 12V vehicle battery. Frequent short cycles, idle periods or extreme ambient conditions accelerate degradation, while stable operation extends life. Electrolysers behave in much the same way.

Stack life, often quoted in years, only becomes meaningful when considered alongside the conditions under which it is achieved. Evaluating an electrolyser based purely on beginning-of-life performance is insufficient. What matters is how that performance evolves, and how predictable that evolution is, not only for a certain short period of time covered by warranties, but also for the overall lifetime of the stack.

The hidden mechanics of degradation

At a high level, ageing appears to be a gradual increase in power demand and a reduction in hydrogen output.

As internal resistance rises, more energy is required to maintain the same current, and therefore the same level of production, until the moment the stacks cannot be economically operated anymore and must be exchanged.

These effects are well known, widely measured and typically reflected in performance discussions. However, they are rarely the factors that ultimately limit operability.

Degradation occurs through multiple mechanisms, some measurable and many others invisible and not considered until failure occurs. Changes in membrane and coating properties, catalyst activity, material integrity, and internal corrosion all contribute to performance decline and, in the worst case, to stack damage and exchange.

Gas behaviour introduces an additional layer of complexity. Hydrogen, as the smallest molecule, diffuses more readily than oxygen across the membrane. Material ageing in both alkaline electrolyser diaphragms and PEM membrane coatings, gradually increases hydrogen crossover into the oxygen stream. Over time, this reduces oxygen purity, lowers hydrogen output and narrows the operational life of the stack.

There are defined thresholds at which gas mixtures become explosive, and systems must remain within safe limits. As purity degrades, the margin for safe operation narrows; if hydrogen in oxygen reaches over 2%, operation must be stopped. What begins as a performance issue can become a safety constraint – often before other performance limits are reached.

Leakage represents another critical dimension. Unlike gradual efficiency losses, leaks often impose potential operational limits. Depending on design philosophy, once detected, it may require an immediate stack exchange. External leaks can release gases or process media into the process building or container, potentially creating safety risks. Internal leaks, particularly in systems with differential pressure, lead to rapid gas mixing, irreversible stack damage and consequently to immediate stack exchange.

Performance warranties and operational reality

Performance warranties typically focus on a limited set of parameters around power consumption and mass flow. While important, they do not capture the full picture.

From a manufacturer’s perspective, warranting a single parameter such as power may be a low‑risk optimum; from an operator’s perspective, however, it is insufficient. Even if the warranted power is achieved, lower hydrogen output or reduced gas purity can mean lower efficiency than assumed in the business case and a shorter stack lifetime than expected.

This creates a growing misalignment between expectations, warranties, and real‑world performance as systems age. Not all ageing effects are equal: some primarily affect economics, while others determine whether the system can be operated safely at all.

From data to predictability

As more electrolysers are deployed, operational data is building a clearer picture of how systems behave over time. A data-driven approach, considering all measurable changes like power, hydrogen output and gas purity, enables more accurate performance modelling.

It allows operators to anticipate how power consumption, hydrogen output and gas purity will evolve under specific conditions, supporting better decision-making. It also identifies the factors that accelerate or mitigate ageing, guiding more effective operation.

Understanding how systems age is only one part of the equation. The other part is designing systems that manage that ageing effectively over the predicted lifetime.

Designing for lifecycle value

Industry experience shows that once a system is in operation, many of the mechanisms that drive degradation can only be influenced within narrow limits. The decisive choices are made much earlier — shaping how cells are designed, how fluids and gases are managed, how pressure, thermal load and current are distributed, and how much safe operating headroom remains as the system ages.

A customer’s choice of the electrolyser system is also closely linked to these early-stage decisions. While all systems aim to convert water and electricity into hydrogen, they differ significantly in how they manage internal conditions – and in the trade-offs they make between them. Each customer should make a deliberate decision on which features and risks are acceptable for their operating profile and business model.

Equally important is how systems are maintained over time. Electrolysers are long-term assets. Stack replacement, refurbishment, and integration of new technologies must also be considered. In a fast evolving market, the ability to access replacement components and benefit from technological improvements is critical to protecting asset value.

A holistic definition of performance

The hydrogen sector is moving from early deployment to industrial scale. With that shift comes a need for greater clarity in how performance is defined and communicated. Ageing is an inherent characteristic of electrochemical systems.

The challenge is not to avoid it, but to understand it, predict it and manage it effectively, which is only possible through elaborate design.

Here’s what customers must ask electrolyser manufacturers when considering building a hydrogen plant:

• How do expected operation profiles and operating conditions affect stack life exchange?
• When is the stack exchange required?
• What is the estimated time to reach this point?
• How can I ensure the life prediction applies realistically to my case situation?
• Are there any situations where the stack needs to be replaced ahead of schedule?