Two papers published this year sense hydrogen with palladium in two different devices. One, in Materials & Design, co-sputters a Pd-Ag-Ni alloy film and reads the change in its resistance at room temperature. The other, in Microsystems & Nanoengineering, puts a palladium gate on an AlGaN/GaN transistor and reads its threshold shift down to 0.1 ppm. The two are not interchangeable. Anyone about to buy Pd alloy online monitoring hardware is choosing between two ways of reading one reaction.
Two ways to read one reaction
Palladium absorbs hydrogen into its own lattice, and both families exploit what that absorption changes. The choice is which change gets instrumented: a bulk property of the metal, or an interface property of a junction. They diverge first on stability: palladium hydride changes phase as loading rises, and repeated cycling embrittles a pure film. The 2026 ternary work answers that — a Pd50.1Ag22.4Ni22.5 composition was selected for room-temperature response with resistance to hydrogen embrittlement.
The resistive thin film: a resistor whose value follows hydrogen
A palladium alloy film sits between electrodes on an insulating substrate, and absorbed hydrogen raises its resistance. The circuit matters as much as the film: a two-wire reading includes leads and contacts that drift with temperature, so a four-wire connection senses the film alone. The film’s condition matters just as much. A 2026 study in Sensors and Actuators B engineered nanocracks about 10 nm wide into a 15 nm palladium film, keeping conductive paths partly connected so resistance modulates continuously rather than switching abruptly when a gap closes. It detected 100 ppb to 1 % hydrogen in ambient air and tolerates the cycling a transformer imposes.
The MOS element: a transistor whose threshold moves
In the second family the palladium is a gate rather than a resistor. Hydrogen diffuses through it and adsorbs at the metal-semiconductor interface, forming a dipole layer that lowers the effective work function — as though a small positive bias had been applied, increasing the electron gas under the gate and lowering channel resistance.
The transistor amplifies its own signal, which is why this family reaches very low concentrations. The 2026 AlGaN/GaN design pairs palladium-gated sensing HEMTs with passivated reference HEMTs in a Wheatstone bridge, which rejects the common-mode drift and bias-temperature instability that make a single-ended MIS or Schottky device wander once warm. Reported: a 0.1 ppm detection limit, monotonic from 0.1 to 3,000 ppm. PAS DGA’s other platform, photoacoustic spectroscopy, reads the same gases optically.
Which principle fits which duty
Neither element is better; the installation assigns them.
| Property | Resistive Pd alloy thin film | MOS / FET element |
|---|---|---|
| Quantity measured | Film resistance, four-wire | Threshold voltage shift |
| Low-end reach | Set by resistance resolution | Set by the junction |
| Range | Wide and continuous | Narrower, steeper near threshold |
| Drift source | Film and contact temperature | Interface charge, bias history |
The duty cycle decides: resolving a slow rise over months, in oil near either temperature extreme, is a different job from catching fast excursions at the lowest concentrations. PAS DGA builds both. The DGA-300 uses an MOS-type palladium alloy element rated 2–2,000 ppm, 2 ppm limit of detection, ±20 % or ±2 ppm, 0–60 °C; the DGA-500 uses a four-wire palladium alloy film rated 5–5,000 ppm, 5 ppm limit of detection, ±15 % or ±5 ppm, −40 to 105 °C. Both are IP67; figures are manufacturer datasheet values.
What should I ask before buying Pd alloy online monitoring hardware?
Four questions separate a defensible specification from a weak one.
- Which quantity, over what range? A limit of detection without a stated range is not a specification. Ask for both, and for the accuracy basis: percentage of reading, or percentage plus a floor in ppm.
- What happens when hydrogen is removed? Recovery, not response, is where hysteresis shows.
- What is the oil temperature rating? A film characterised at ambient is not the same instrument at 105 °C.
- What else in the oil can reach the element? Palladium responds to hydrogen; anything else dissolved in the oil is a cross-sensitivity question.
Does a MOS element replace a resistive one?
No, and the two 2026 papers are not competing designs. The transistor family wins where the need is the lowest detectable concentration over a short diffusion path in a device that can be heated and biased under control; the resistive family wins where a wide continuous range must come from a circuit trusted for years.
PAS DGA supplies both; selection follows from oil temperature and the concentration range a fleet shows. Send us the fleet list with unit ratings and DGA history, and we will say which element belongs where before you buy Pd alloy online monitoring hardware for any of them.
The hydrogen sensor range lists the element in each model, palladium thin-film sensing describes it more closely, and online versus offline DGA covers continuous data against laboratory sampling.
Sources
- Fan, X. et al. “High-throughput prepared Pd-Ag-Ni ternary alloy thin films for room-temperature highly sensitive hydrogen sensing.” Materials & Design 262 (2026).
- Hong, J. et al. “Crack-Engineered Palladium Nanocrack Networks for Trace-to-Leak Hydrogen Monitoring under Ambient Conditions.” Sensors and Actuators B: Chemical (2026).
- “AlGaN/GaN HEMT H2 sensor with integrated Wheatstone bridge and on-chip microheater for 0.1-ppm detection.” Microsystems & Nanoengineering (2026).
- DGA-300 / DGA-500 figures: manufacturer datasheet (vendor data).