A specification sheet that cannot be read literally
Online dissolved gas analysis (DGA) monitors are sold on gas count: three gases, five gases, seven, nine. The implicit promise is that every gas in the list is measured the same way, to the same standard. For one of them — hydrogen — that promise is physically impossible to keep in a photoacoustic instrument. This is not a shortcoming of any particular vendor. It is molecular physics, and understanding it is the fastest way to tell a credible multi-gas specification from a glossed one.
Why hydrogen is invisible to photoacoustic spectroscopy
Photoacoustic spectroscopy (PAS) works by absorbing light. A pulsed or modulated light source — a laser in laser photoacoustic spectroscopy (L-PAS), or a broadband lamp with filters in simpler designs — illuminates a gas sample. A molecule absorbs a photon at a wavelength matching one of its vibrational transitions, the absorbed energy becomes local heat, the heat expands the gas, and a microphone hears the resulting pressure wave.
No absorption, no signal. And absorption requires a changing dipole moment during the vibration. Hydrogen is a homonuclear diatomic molecule: two identical atoms sharing a perfectly symmetric bond. Its vibration produces no change in dipole moment, so its fundamental vibrational transitions are not infrared-active. The same reasoning applies to nitrogen and oxygen. It is why these gases are described as infrared-inactive, and it is why the same blindness affects every infrared absorption technique — not only PAS, but non-dispersive infrared (NDIR) and Fourier-transform infrared (FTIR) as well.
This has a direct commercial consequence: every photoacoustic DGA monitor on the market measures hydrogen with a separate sensor. The nine-gas headline describes the optical path plus a hydrogen channel that works on an entirely different principle. ABB’s own patent for an FTIR-based analyser states it plainly, specifying that O2, H2 and N2 “may be performed optically and/or with non-optical sensors” — resistive, capacitive or thermo-conductive — and naming a paramagnetic analyser for oxygen (US 10,832,854 B2).
Why this matters more than the gas count
If hydrogen were the least important gas, the blind spot would be an academic footnote. It is the opposite. Hydrogen is the first fault gas to appear in most incipient conditions: it is generated by partial discharge, by arcing, by thermal decomposition of oil, and — inconveniently — by stray gassing in healthy new units. It is the gas that most early-warning thresholds are built around, and it is the one whose rate of rise drives the decision to escalate.
So the two specifications are not equally important. A monitor’s optical path determines how well it resolves methane, ethylene and acetylene. Its hydrogen channel determines whether the earliest alarm is trustworthy at all. A datasheet that leads with the gas count and hides the hydrogen channel’s principle and drift behavior has hidden the part that matters most.
How the hydrogen channel is actually built
Four sensing principles dominate. They are not equivalent in selectivity or long-term stability.
| Principle | How it detects H2 | Practical character |
|---|---|---|
| Metal-oxide semiconductor | Resistance shifts as hydrogen adsorbs on a heated oxide surface | Widely used; sensitive, but responds to other reducing gases and drifts with humidity |
| Thermal conductivity | Hydrogen conducts heat far better than the surrounding gas mixture | Simple and robust, but reads total thermal conductivity, not hydrogen specifically |
| Palladium-alloy thin film | Resistance changes as hydrogen absorbs into the palladium lattice | Selective by physical mechanism rather than by filtering; stable baseline |
| Electrochemical cell | Current from hydrogen oxidation at an electrode | Common in portable instruments; consumable electrolyte |
Selectivity is the dividing line. Thermal conductivity and metal-oxide sensors respond to whatever else changes the ambient conditions; a palladium-alloy film responds to hydrogen because hydrogen is the only gas that readily dissolves into its lattice. That difference shows up not on the day of commissioning but in the second and third year, as baseline drift and cross-sensitivity accumulate.
The detection-limit ladder
Hydrogen is abundant in a transformer, so its detection limit is rarely the constraint — acetylene is. But the two are specified together in any honest comparison, and the published figures span an order of magnitude. As vendor-published data: L-PAS instruments reach C2H2 ≤ 0.1 ppm, and the commercial L-PAS platform is specified at 0.05 ppm; Camlin TOTUS G9 at 0.1 ppm; Vaisala OPT100 at ±0.5 ppm; Qualitrol Serveron TM8 at 1 ppm. Treat all of these as vendor data, not independent verification — and note that a hydrogen channel’s useful figure is its drift specification over years, which datasheets rarely state.
Sources
- Infrared selection rule: hydrogen, nitrogen and oxygen are homonuclear diatomic molecules with no permanent dipole moment, so their fundamental vibrational transitions are not infrared-active. Standard molecular spectroscopy; the same blindness affects PAS, NDIR and FTIR alike.
- US Patent 10,832,854 B2 (ABB Schweiz AG), Dissolved gas analysis devices, systems, and methods — non-optical measurement of O2, H2 and N2.
- Hübert, T., Boon-Brett, L., Black, G. & Banach, U., “Hydrogen sensors – A review,” Sensors and Actuators B: Chemical 157(2):329–352, 2011 — hydrogen sensing principles, selectivity and operating lifetime.
- Detection-limit figures for L-PAS, Camlin TOTUS G9, Vaisala OPT100 and Qualitrol Serveron TM8 are vendor-published data, not independently verified.
What to ask before you buy
- Which gas is measured optically, and which by a dedicated sensor?
- What is the hydrogen channel’s principle — and its stated cross-sensitivity to moisture and other reducing gases?
- What is the hydrogen baseline drift per year, and is recalibration field-performable?
- If the hydrogen channel is disabled, what does the monitor still report — and would your early-warning logic still function?
The last question is the sharp one. On many assets the alarm philosophy rests almost entirely on hydrogen and its rate of change — so a fleet manager who knows this reads a nine-gas datasheet very differently from one who does not.
PAS DGA: photoacoustic gases, a purpose-built hydrogen channel
We build both halves of the instrument deliberately. The DGA-900 measures nine gases plus moisture with laser photoacoustic spectroscopy, and pairs that optical path with a dedicated hydrogen channel — because photoacoustic spectroscopy cannot see hydrogen and no honest instrument pretends otherwise. For assets where hydrogen and moisture are the whole requirement, the DGA-500 is built on a palladium-alloy thin-film hydrogen sensor, selected for the reason above: selectivity by physical mechanism, not by filtering. The mechanism itself — Sieverts’ law, the palladium α-to-β transition, and why nickel is alloyed in — is set out in in-situ hydrogen sensing with a palladium-nickel thin film. Read the PAS technology overview for the optical side, browse the hydrogen sensor range, or contact PAS DGA and ask us directly which gas in our specification is measured how.