Green hydrogen from water electrolysis has ceased to be a topic of strategic presentations — since 2024, the first 100 MW+ projects have launched in Europe, and by 2026, the industry is grappling with two sobering problems: scaling electrolyzer stacks and maintaining hydrogen quality in long supply chains. In both cases, process analytics are stepping in where periodic laboratory samples once sufficed. Raman spectroscopy — operando at electrodes, as product gas monitoring, and as raw material control equipment — is appearing in an increasing number of publications and project specifications.

At Gekko Photonics, we design and manufacture process Raman analyzers in Poland — in inline, at-line/lab, and portable variants. Our largest number of implementations is in process chemistry: resins, cosmetics, fertilizers,, adhesives, and hydrocarbons. In projects related to hydrogen and electrolyzers, we operate in a design phase — we verify on client samples whether Raman is the appropriate method for a specific analyte and matrix before any CAPEX is committed. This article organizes where Raman realistically fits in the hydrogen value chain and what limitations must be understood before discussing feasibility.

Why electrolyzers are a rewarding but demanding area for spectroscopy

In a water electrolyzer, many things happen in a short time and small volume: the hydrogen evolution reaction (HER) occurs at the cathode, the oxygen evolution reaction (OER) at the anode, and between them operates a membrane or diaphragm where ion transport depends on electrolyte purity, polymer composition, and temperature. Each of these domains has its spectral markers — vibrational bands of water molecules adsorbed on the catalyst surface, metal–oxygen bond bands in reorganizing nickel and iron oxyhydroxides, and C-F, C-O-C, and SO₃H bands in polymer membranes. For a team familiar with the typical spectral window of 200–1800 cm⁻¹, electrolysis is a rich area of information.

The problem is that the measurement must be performed under the conditions in which the device operates — under current, in the presence of gas bubbles, sometimes at high pressure and temperatures of 60–90 °C in the case of alkaline electrolyzers. This forces special design compromises. Therefore, most publications from recent months concern laboratory configurations — 3D-printed cells with sapphire windows, SERS probes, coupling with operando electrochemical mode — rather than full-scale industrial installations.

Three areas where Raman enters the hydrogen chain

1. Operando R&D — reaction mechanisms and catalyst degradation

The most widely present application today. Teams developing catalysts for HER and OER use Raman in situ, to track electrode surface reconstruction under polarization. A review in ACS Catalysis from 2023 systematizes how SERS and operando Raman reveal intermediate states of water molecules and metal–hydrogen bonds; newer works show, among others, the behavior of NiMoO₄ in neutral electrolyte or hydrogen spillover on metallic surfaces. Operando Raman is also used for diagnosing the degradation of anion exchange membranes (AEM) — in a publication from JACS 2024, a team described the oxidation products of the ionomer under AEM electrolyzer operating conditions. This is a type of data that post mortem analysis simply will not provide: once the sample is removed from the cell, the signal has already disappeared.

Hardware requirement for R&D: focus the beam on the electrode surface (typically a 20× or 50× objective through a sapphire window), work with 785 nm or 532 nm depending on fluorescence, and have a sufficiently sensitive detector to capture a weak SERS spectrum in the presence of a strong background signal from the electrolyte. These are conditions for a laboratory or microscope spectrometer, not a process probe.

2. Product gas quality control — trace contaminants in H₂

Hydrogen for fuel cells must meet purity requirements at the level of fractions of ppm for key contaminants (CO, S, NH₃, H₂O, hydrocarbons, inert gas). Traditionally, gas chromatography, compact GC-MS, and electrochemical sensors are used here. Gas Raman is slowly entering this niche — as a reagent-free technique, requiring no point calibration, providing simultaneous readout of several components from a single spectrum.

Sensitive gas Raman with a multipass configuration (multiple beam reflections through a gas cell) currently allows detection of hydrogen and its main contaminants in the ppm range without the need for chromatographic separation — this is described in a work in Sensors (MDPI) on trace hydrogen sensing via multipass Raman scattering. For us, this sounds promising as a complement to fast GCs at filling stations and quality control points, but each project requires validation on a real stream — noise from aerosol fluorescence, window cleanliness, and laser stability are real issues. This is not yet a plug-and-play technology.

3. Monitoring of electrolyte and auxiliary media

In alkaline electrolyzers, one typically works with 25–32% KOH; in AEM technology, dilute KOH solutions or pure water + ionomer alone are beginning to appear. Electrolyte concentration, its degradation, accumulation of impurities from raw water, and the presence of metal ions leached from electrodes — these are all parameters that affect stack efficiency and lifetime. Raman in inline mode can be a candidate here for a reference measurement alongside classic conductometry and densitometry, especially when interested in the composition of anionic ions (carbonates, sulfates, silicates) and organic degradation products of the ionomer.

This is the area closest to what process Raman does best — measurement in process liquid, immersion probe, collecting a multicomponent chemometric model. Here, adaptation of the general industrial platform is most straightforward — if the client provides electrolyte samples in the typical operating range.

Limitations to keep in mind before a project

• Matrix fluorescence — some technical electrolytes with organic additives give a strong fluorescence background. 785 nm is usually sufficient, but in some cases, longer excitation (1064 nm) must be used, which has its own compromises (detector sensitivity, cost).
• Gas bubbles in the cell — measurement under current with intense H₂ or O₂ evolution requires a well-thought-out measurement geometry; a thin layer, a back-scatter probe with space for bubble escape, or a narrow time window synchronized with system operation.
• Temperature and pressure — industrial stacks operate at pressures of 30–50 bar, some SOEC projects at temperatures of 700–800 °C. Standard process equipment typically handles temperatures below ~120 °C and pressures of a few bar; direct measurement in an SOEC stack practically does not exist. Sensible measurement points are peripheral streams, condensers, and phase separators.
• Acquisition time vs. dynamics — typical process Raman for liquids collects a spectrum in tens of seconds; for gas in multipass mode, it can take minutes. For most electrolyzer monitoring, this is sufficient, but fast transients (startup, shutdown, load change) require other reference techniques.
• Lack of ready-made chemometric models for electrolyzer electrolytes — unlike cosmetic or resin chemistry, where spectral libraries and PLS models are already stabilized, for hydrogen, a model must be built from scratch on client samples.

Electrolyzer market status in 2026.

From an analytics supplier's perspective, three observations are relevant. First, PEM maintains a strong position in new medium-scale projects (market analyses for 2026 typically give it around 35–40% share in the modern electrolyzer segment alongside mature alkaline technology), and AEM is growing rapidly as a cost compromise between alkaline and PEM. Second, the availability of iridium (a key OER catalyst in PEM) is a real scaling bottleneck — global production at about 7 tons per year, 80% from South Africa, drives intensive work on reducing catalyst loadings and replacing materials. Third, the supply chain is gaining more points where gas quality must be verified: tanker filling, transfer to storage, refueling stations, heavy mobility chains. Each of these points is a potential location for analytical equipment.

Raman spectroscopy — adaptation possibilities for hydrogen and electrolyzers

Our analytical platform is designed as a general industrial family, which we adapt in project mode to specific chemistries. In the context of hydrogen, we see it as follows:

• Spectrally X1 LAB — a benchtop analyzer with a carousel for up to 25 vials, through-package analysis via transparent packaging. A natural starting point for R&D teams developing catalysts, verifying electrolyte sample composition, or monitoring model compositions before moving to inline. It operates with 785 nm, a back-thinned CCD detector with thermoelectric cooling, and chemometric software Spectrally OS. For advanced operando configurations, an external probe can be connected.
• Spectrally X1 INLINE — a process analyzer with an immersion probe mounted in a pipeline or tank nozzle. In hydrogen practice, we see it first in electrolyte monitoring (alkaline KOH, AEM ionomer, feed water), DI water streams, and in some configurations for measuring liquid product after compression. The Spectrally X1 PROBE with the self-cleaning Retractex component is sensible where deposits accumulate — rare in typical electrolytes, but sometimes needed in circuits with iron/nickel admixtures from long-term electrode leaching.
• Spectrally X1 PORTABLE — in a suitcase, with a built-in touchscreen, IP54, for mobile verification of raw materials (e.g., KOH brine, electrolyte additives, cathode materials before assembly) and field quality audits. A practical scenario: a maintenance team travels to a site with the suitcase, performs a quick reference measurement, and returns with data.
• Spectrally OS — a software layer common to the entire X1 family: PLS and CNN models, a library of ~28,000 reference spectra, archiving, integration with DCS/MES. It provides data consistency between the lab and the line — one model validated in LAB can be connected in INLINE.

We do not declare an implementation portfolio in hydrogen today — this is a project area for us. In practice, this means every serious topic begins with a feasibility study on client samples: we collect spectra, check whether sufficient spectral selectivity exists for the analyte of interest, and only then discuss the probe, integration, and CAPEX. See also our decision-maker's guide to Raman in chemical processes — most engineering principles from process chemistry transfer 1:1 to electrolyzers.

FAQ — frequently asked questions

Will Raman replace gas chromatography in hydrogen purity measurement?
In most cases, it will not replace it, but complement it. GC remains the reference for trace contaminants below a few ppm; Raman enters as a fast, reagent-free measurement of several components at once, especially where response time matters (e.g., quick decision on load acceptance).

Which excitation wavelengths make sense for hydrogen and electrolyzers?
Most often 785 nm — a good compromise between sensitivity and background fluorescence. For highly fluorescent matrices, 1064 nm is considered. 532 nm works mainly in SERS work on metallic electrodes under laboratory conditions.

Can measurements be made directly in an electrolyzer stack under pressure?
Direct measurement in a stack operating at 30–50 bar is possible under laboratory conditions with operando cells, but at the industrial level, it is more practical to measure in peripheral streams: phase separators, electrolyte circuits, condensers, product gas lines after pressure reduction. There, standard process equipment has acceptable operating ranges.

Does Gekko have implementations in hydrogen production?
We have the most implementations today in process chemistry — resins, cosmetics, fertilizers, adhesives, hydrocarbons. In hydrogen, we operate project-wise: on client samples, we verify in a feasibility cycle whether Raman is the appropriate method for a given analyte and matrix before the client commits CAPEX. Such a cycle typically takes 2–4 weeks and concludes with a report and recommendation.

What about SOEC electrolyzers operating at 700–800 °C?
At this temperature, Raman measurement in the stack is practically outside the range of standard equipment. Sensible measurement points are streams after heat exchangers and condensers, where the temperature drops to conditions acceptable for the probe. We design the configuration for such a measurement individually.

Let’s talk about your process

At Gekko Photonics, every hydrogen-related project begins with a short technical discussion (typically a 30-minute meeting with an application engineer), during which we determine: the type of electrolyzer, the planned measurement location, available samples, and the expected analytical range. If the topic looks feasible, within 2 weeks we perform a test measurement on your samples in our laboratory in Wrocław — you receive the feasibility report typically within 10 working days from the completion of measurements. All without any obligation on your part.

We invite you to contact us via contact page or directly at spectrally@gekkophotonics.com. The full offer of the Spectrally X1 analyzer family can be found in the product section.