Work-function devices: MIS, MIM, Schottky diodes & GasFETs

Summary

A catalytic Pd or Pt film (5–30 nm) splits H2 into atoms, which diffuse to a buried metal/insulator or metal/semiconductor interface and form a dipole layer. That dipole shifts the film’s work function, read out as a threshold-voltage shift (MOSFET/MIS capacitor), a barrier-height change (Schottky diode), or a two-dimensional-electron-gas density change (HEMT). Same chemistry, different circuit: MOSFET/MISFET, MOS/MIS capacitor, Schottky diode, GasFET, HEMT.

The physics is fifty years old and does not need oxygen, an advantage in buried, oxygen-poor settings. The same catalytic surface that gives it sensitivity is poisoned by CO, H2S, SO2, and NOx. H2S is the most likely field failure mode, since it co-occurs with natural H2 seeps.

At a glance — published performance envelope

Detection limit

0.005–40 ppm

Range

0.005–50,000 ppm

Response t₉₀

1.5–172 s

Recovery

2–240 s

Selectivity

low

Drift

high

Operating T

−40 … 800 °C

Humidity

0–95 % RH

Power

1–500 mW

Unit cost

$5–$2k

TRL

7

Main interferents: H₂S, CO, CH₄, NH₃ · poisons: CO, H₂S, SO₂. Ranges span the published devices/variants covered below — per-claim citations are in the body; data: src/data/modalities/work-function-devices.yaml.

How it works

Catalytic transition-metal films (Pd, Pt, Ir, or Pd/Pt alloys) share one chemistry step and differ only in how the resulting signal is read out electrically.

Dissociation and diffusion. H2 splits into atomic H on the catalytic film with near-zero activation barrier on Pd and Pt [1][2]. Atomic H diffuses through the thin (5–30 nm) film, bulk diffusion through Pd, grain-boundary diffusion through Pt, and accumulates at the buried interface underneath [3][4][5].

Dipole and work-function shift. The adsorbed H atoms polarize at that interface into an oriented dipole layer [3][6][7], lowering the effective work function seen by the semiconductor or counter-electrode. Device stacks report shifts up to ≈0.5 eV at high H2 partial pressure; bare single-crystal Pd under UHV shows smaller values, 0.18–0.36 eV by crystal face, since it measures the surface dipole directly, without a sensor’s amplifying interface geometry [3][8][9][10]. Clean Pd’s work function is ≈5.2 eV [7]; every device below converts the shift, not the absolute value, into a signal. The Helmholtz dipole-sheet relation is

$\Delta\Phi = \frac{q\,N_H\,p}{\varepsilon_0}$

linear at low coverage.

Temkin isotherm. Dipole-dipole repulsion bends that relation at higher coverage: adsorption heat falls roughly linearly with coverage $\theta$,

$\theta = \frac{RT}{b_T}\ln(A\,p)$

giving the log(H2) response reported across MOS, MIS and Schottky devices [13][14]. Eriksson, Lundström and Ekedahl derived this from an electrostatic adsorbate-interaction model for the Pd–SiO2 interface, matching measured coverage over three decades of $p_{H_2}$ and predicting saturation near $1\times10^{18}\,\mathrm{m^{-2}}$ [11][12].

MIS/MOSFET readout. The work-function change appears as a rigid shift in flat-band voltage (capacitor) or threshold voltage (transistor):

$\Delta V_{FB} = \Delta V_T = \frac{\Delta\Phi_M}{q}$

A Pd-gate MOS capacitor’s C–V curve translates laterally on H2 exposure without changing shape (≈0.17 V at 91 ppm H2, one plasma-oxide device [15]); a MISFET reads the same shift as a $V_T$-compensating gate voltage, the mode used since Lundström’s original device [16].

Schottky-diode readout. The same interfacial dipole modulates the barrier height $\Phi_B$ instead of a MOSFET threshold, giving an exponential rather than linear response:

$J = A^{*}T^{2}\exp\!\left(-\frac{q\Phi_B}{kT}\right)\left[\exp\!\left(\frac{qV}{nkT}\right)-1\right]$

Modest barrier changes (tens of meV) can swing forward current by orders of magnitude [17][18][19].

MIM. Two mechanisms share the name. A static MIM capacitor or tunnel diode (metal/thin insulator/metal, ≲4 nm) is the electrical analogue of the MIS flat-band shift, with a metal rather than semiconductor counter-electrode [20][21]. A dynamic “chemicurrent” MIM or MIS diode instead uses the exothermicity of H2 dissociation and recombination to excite hot electron-hole pairs directly; if the film is thin enough (≲10 nm) for hot carriers to reach the interface ballistically, a DC current flows at zero bias, driven by reaction rate, not coverage. This has stayed a UHV research tool, not a field sensor [22][23].

Key numbers for this family:

Quantity	Value	Ref.
Pd clean-surface work function	≈5.2 eV	[7]
Max. reported ΔΦ from H2 dipole layer (device)	up to ≈0.5 eV	[15][16]
ΔΦ on Pd(100)/(110)/(111) (UHV, bare surface)	0.20 / 0.36 / 0.18 eV	[9]
Max. interfacial H areal density, Pd–SiO2	~1×10¹⁸ m⁻²	[11][12]
H2 activation energy, absorption into bulk Pd	60–80 kJ/mol	[2]
Original Lundström device: gate stack	10 nm SiO2 / 10 nm Pd	[24]
Original Lundström device: LOD / temperature / response	40 ppm H2 in air at 150 °C, t < 2 min	[24]

Device classes

All rows share the dissociation-diffusion-dipole chemistry above; they differ in gate/junction architecture and readout.

Class	Readout	Note	Ref
Pd/Pt-gate MOSFET (“Lundström FET”)	$\Delta V_T$ at fixed $I_D$, or the compensating $V_{GS}$	First demonstrated 1975	[24]
Pd/Pt-gate MOS capacitor	C–V flat-band shift; PLL readout avoids a full field sweep	No channel	[25][26]
Alternative-dielectric MIS (Ta2O5, La2O3, HfO2, Al2O3, ZnO)	as above	Lower operating T, CMOS-compatible; more leakage/noise	[13][27][28][29]
SiC/GaN “high-temperature MIS”	as above	Semiconductor itself, not just the film, survives 200–600 °C	[30][31][26]
Static MIM capacitor/tunnel diode	Tunneling I–V or capacitance change	Metal-only analogue of the MOS capacitor	[20][21]
Chemicurrent/hot-electron MIM & MIS diode	Zero-bias DC current from adsorption exothermicity	Measures reaction rate; UHV research tool	[22][23]
Plasmonic “MIM” waveguide (different family)	Optical resonance shift on Pd-hydride formation	Not electronic despite the shared acronym; see Plasmonic metamaterials	[32]
Pd/Pt–Si Schottky diode (incl. Pd/SiO2/Si)	Forward current at fixed bias	One of the earliest Pd-semiconductor H2 sensors	[17]
Pd/III-V Schottky (GaAs, InP, AlGaAs)	as above	Faster, different barrier heights than Si	—
Pd/Pt–TiO2 Schottky	as above	Barrier shift also involves O reduction to water	[17][18]
Pt/GaN, Pd/GaN, Pt-AlGaN/GaN Schottky	as above	GaN survives to 800 °C; same mechanism as Si	[33][34]
Pd/Pt–SiC Schottky	as above	30-day stable at 500 °C	[19]
Pd/Pt-gated AlGaN/GaN HEMT	Dipole modulates 2DEG density	Sub-ppm LOD, operation to hundreds of °C	[34][35]
Graphene/2D-material-gate MIS-HEMT	as above	Active frontier variant	[34]
Suspended-gate FET (SGFET)	Capacitive coupling across an air gap	Extends range past the classic Lundström FET	—
Kelvin-probe (vibrating capacitor)	Non-contact work-function measurement	Lab tool, not field-deployable	—
Siemens/AppliedSensor “GasFET”	Capacitive coupling into a FET channel	Commercial descendant of Lundström/SGFET	—

Performance envelope

How to read this table

Every row is sourced to a specific device or paper; treat the ranges as “what has been demonstrated across the literature/commercial space,” not a guarantee for any given part number.

Device type	LOD (H2)	Useful range	Response/recovery	Typical op. temp	Notes / ref.
Original Pd-gate MOSFET (Lundström 1975)	40 ppm (stated)	not fully characterized upward	t90 < 2 min	150 °C	10 nm SiO2 / 10 nm Pd gate [24]
Pd-gate FET (annealed Pd, later study)	~1 ppm (tested 1–10,000 ppm)	1–10,000 ppm	~1 min / 2–4 min recovery (to 1 ppm)	100 °C	Baseline drift <30 mV/month over 3 months [36]
Pd/Ta2O5-SiO2-Si MISFET (long-term study)	ppm-level	multi-decade	seconds–minutes / minutes	ambient/elevated	ΔV_CM fell 0.46→0.36 V (−21.7%), sensitivity fell up to 47.8% over 200 cycles/8 weeks; zero-line drift 50 mV → <5 mV/4 weeks [16]
MOS capacitor, plasma SiO2, Pd gate	sub-100 ppm demonstrated	tested to low-ppm	not stated	not stated	ΔV_FB ≈0.17 V at 91 ppm H2 [15]
Pd/TiO2/Si MIS capacitor	0.1–2 ppm (tested)	0.1–2 ppm and up	not stated	varied	[15]
SiC MOS capacitor / MOSFET-capacitor (Pt or Pd/Ta2O5/SiC)	ppm-level, exhaust-relevant	H2, CO, hydrocarbons simultaneously	fast (SiC > Si rate)	to 200 °C, 45 vol% H2O tolerant	[25][31]
Pd/Ta2O5/SiC Schottky diode	tested to 5,000 ppm	up to 5,000 ppm	rise 7.1 s / decay 18 s at optimum T	~300 °C optimum	SiC outperforms Si via larger bandgap (3.26 eV) [18]
Pd- and Pt-SiC Schottky diodes (high-T)	not given as ppm LOD	H2 and CH4 co-detection	fast (transient ΔI–t method)	stable 30 days at 500 °C	[19]
Pd/InP MOS Schottky diode	<4.3 ppm (RT, one study); 8 ppm (another)	to ~1% and beyond	~1.5 s	RT to elevated	Current changes >10⁶× at 0.1% H2/N2 in one study [34]
Pt/GaN Schottky diode	~1 ppm (best reported)	to 1% H2/air	not stated precisely	300–850 K (to ~577 °C) explored	Sensitivity Factor 14,685 / SR 44,636 at 9,970 ppm in one variant (unverified) [34]
Pt-AlGaN/GaN Schottky diode	not given as ppm	qualitative H2 response	not stated	operated at 800 °C	Forward current rise via barrier-height reduction [33][34]
Pd/AlGaN/GaN HEMT (ppb-class)	50 ppb (one device, unverified); 658% response to 4% H2 (WO3-selective variant)	ppb to several %	2.3–10 s device-dependent; 2.6 s at 200 ppm	RT to 200 °C	[34]
Floating-gate FET (CNT channel + Pd NP + Y2O3 gate)	90 ppb (RT); 5 ppb (~100 °C)	ppb to %	not fully stated	RT–100 °C	Work-function/gate-coupling, not resistive [37]
MIM tunnel-diode hydrogen sensor (design study)	not stated as ppm LOD	v/v-% level	172 s (example)	not stated	Sensitivity up to 267 pm/(v/v-% H2) in one optimization (unverified) [21]
MIM/MIS “chemicurrent” hot-electron diode	N/A: measures reaction flux, not steady concentration	UHV atomic-beam fluxes	sub-second (transient)	UHV/lab, not field-rated	Detects atomic-H adsorption events directly, not a calibrated ppm sensor [22][23]
Commercial FE (AppliedSensor/ams HLS-440 family)	field-rated for leak detection, not ppm-trace	0–10% H2 (HLS-440P)	fast (vendor-stated)	wide, IP6K9, ATEX Zone 2	FE signal + thermal-conductivity channel; crash-tested in FCEVs [38]
H2scan solid-state (MIS + Pd-alloy resistor hybrid)	25 ppm (GRIDSCAN 5000, transformer-oil DGA)	25–5,000 ppm (GRIDSCAN); 4,000 ppm–5% (HY-ALERTA)	t90 ≤ 60 s (HY-ALERTA)	ambient–industrial; no recalibration claimed for 10 yr	Hybrid MIS + Pd-alloy thin-film resistor, not a pure work-function device [39]

The bare catalytic gate is inherently low-to-medium selectivity: anything that dissociates on Pd/Pt and produces mobile interfacial H registers, including H2S and, weakly, some hydrocarbons at high temperature. Overlayers, alternate catalytic metals, WO3 capping, dual-FET designs, and temperature tuning push effective selectivity to medium-to-high for specific interferents [6][40].

Drift is a first-order limitation: MISFETs show zero-line drift and sensitivity loss with accumulated dose, tens of percent over weeks of cycling [16]; the best commercial devices claim <30 mV/month drift [36] or multi-year calibration-free operation (H2scan’s 10-year claim [39]).

Elevated-temperature devices (100–300 °C Si-based, 500–800 °C SiC/GaN) need heater power, tens to a few hundred mW for small MEMS heaters; room-temperature nanostructured variants cut this substantially but remain lab-stage.

Figures of merit

Specializations of the universal set defined on the Figures of merit page; units follow Concentration units.

Figures-of-merit table for this modality (12 rows)

Figure of merit	Symbol / formula	Value for this family	How to read it
Work-function shift	$\Delta\Phi = \dfrac{q\,N_H\,p}{\varepsilon_0}$	Up to ≈0.5 eV at high $p_{H_2}$ in device stacks; 0.18–0.36 eV on bare single-crystal Pd under UHV [9][15][16]	What every readout ($V_T$, $V_{FB}$, $\Phi_B$) converts into a signal; compare against clean Pd’s ≈5.2 eV baseline [7]
Threshold-/flat-band-voltage shift	$\Delta V_T=\Delta V_{FB}=\Delta\Phi_M/q$	ΔV_FB ≈0.17 V at 91 ppm H2, plasma-oxide Pd-MOS capacitor [15]; ΔV_CM fell 0.46 to 0.36 V with accumulated dose, Pd-Ta2O5-SiO2-Si MISFET [16]	What a field circuit digitizes: a rigid voltage-axis shift; a shape change signals trap generation, not H2
Temkin log-linear slope	$\theta=\dfrac{RT}{b_T}\ln(Ap)$, so $\Delta\Phi\propto\ln p_{H_2}$	Matches measured Pd–SiO2 coverage over 3+ decades of $p_{H_2}$ [11][12]; no universal slope value reported (unverified)	Plot on log concentration axes; a log-linear device looks flat at low ppm on a linear plot
Saturation ΔΦ (ceiling)	$\Delta\Phi\to\Delta\Phi_{max}\approx0.5\text{ eV}$	≈0.5 eV, reached near/above percent-level H2 in Pd-gate stacks [15][16]	Above this knee sensitivity collapses; treat the device as saturated near the LEL-adjacent end
Barrier-height sensitivity (Schottky)	$J=A^{*}T^{2}\exp(-q\Phi_B/kT)[\exp(qV/nkT)-1]$, $\Delta\Phi_B\approx-\Delta\Phi_M$	Current changes >10⁶× at 0.1% H2/N2, Pd/InP diode; rise 7.1 s / decay 18 s, Pd/Ta2O5/SiC at ~300 °C optimum [18][34]	A small barrier swing gives a large current swing; why Schottky variants suit passive downhole readout
Limit of detection (LOD)	$\mathrm{LOD}=k\,\sigma_{blank}/S$	90 ppb (RT) / 5 ppb (~100 °C), floating-gate FET [37]; 50 ppb, Pd/AlGaN/GaN HEMT (unverified) [34]; 40 ppm, 1975 Lundström FET [24]	Single-point stated LODs, not IUPAC 3σ figures at a stated averaging time; treat as optimistic
Selectivity / cross-sensitivity	$K_{ij}=S_j/S_i$; overall $=1/\max_j(K_{ij})$	Low-to-medium intrinsically; NREL/JRC ISO 26142 testing found SO2/H2S/NO2/HMDS shifts usually did not persist [41]; a WO3-capped Pd-AlGaN/GaN HEMT reached ≥8× discrimination vs. CH4/CO2/NH3/H2S/NO2 at 150–250 °C [34]	H2S is the worst case for Hestia: co-evolved with the target gas
Response / recovery time	$t_{90}=2.303\,\tau$	$t_{90}$ <2 min, Lundström 1975 [24]; ~1 min / 2–4 min recovery, annealed Pd-gate FET [36]; rise 7.1 s / decay 18 s, Pd/Ta2O5/SiC [18]	H diffusion through Pd/Pt is thermally activated (60–80 kJ/mol); recovery is usually slower
Dynamic range	$N_{\text{decades}}=\log_{10}(C_{max}/C_{min})$	~4 decades, 1–10,000 ppm, annealed Pd-gate FET [36]; ~5–6 decades, ppb-class HEMT/floating-gate devices [34][37]	Log-linear (Temkin), not linear; no single constant $S$ spans the range
Drift / stability	$R_0(t)=R_0(0)+d_Zt$, $S(t)=S(0)(1+d_St)$	ΔV_CM fell 0.46 to 0.36 V, sensitivity fell up to 47.8% over 200 cycles/8 weeks, Pd-Ta2O5-SiO2-Si MISFET [16]; <30 mV/month claimed, annealed Pd-gate FET [36]; H2scan claims >10-yr calibration-free [39]	Dose-dependent, not just time; budget recalibration against dose
Power	Duty-cycled vs. continuous heater draw	Tens to a few hundred mW, MEMS heaters, 100–300 °C Si devices; more for 500–800 °C SiC/GaN; RT nanostructured devices cut this substantially (unverified)	Passive Schottky/MIM diodes need no heater, only a bias/readout circuit
Operating envelope	$C_{\text{reported}}=C_{\text{raw}}\cdot f_T(T)\cdot f_{RH}(RH)\cdot f_P(P)$	RT (nanostructured HEMT/floating-gate) to 800 °C (Pt-AlGaN/GaN Schottky) [33][34]; O2 not required [42]; no downhole pressure data found	Don’t assume ambient-pressure curves hold downhole

What dominates

Selectivity (H2S resistance) and dose-dependent drift decide usability for Hestia. The same catalytic gate that swings up to ≈0.5 eV per decade of $p_{H_2}$ exposes the sites H2S deactivates, so a sub-ppb LOD means little until H2S resistance and post-dose drift are qualified over months, not the short ISO 26142 windows in the literature.

Secondary trade-off: room-temperature ppb-class devices remain lab-stage, while the SiC/GaN devices proven for downhole duty (500–800 °C, 30-day stability) trade that survivability for heater power and a less mature ppm-level LOD. No part wins on all three axes at once.

Cross-sensitivity & environment

Catalytic poisoning of the Pd/Pt surface is the dominant environmental failure mode; H2S is the most likely one at Hestia’s target sites.

Factor	Effect	Design response	Ref
CO	Chemisorbs on Pd, blocks H2 dissociation sites; sensor can underestimate H2 if exposure is too short for steady state	—	[2][5]
H2S	Same as CO at low concentration; at high concentration forms a surface sulfide that shuts down H penetration	Protective membrane, scrubber, or periodic recalibration	[2][5]
NOx, SO2	Compete with H2 for adsorption sites on Pd	—	[2][5]
H2O	Longer response, lower amplitude in most formulations; some SiC-MOS designs tolerate ~45 vol% water vapor while retaining H2/CO/hydrocarbon sensitivity	Device-dependent; verify per part	[5][31]
HMDS/siloxanes	Classic poison for combustion-type sensors; the ISO 26142 10 ppm test may understate real-world siloxane exposure	—	[41]
Pd α↔β hysteresis	Phase transition near 20–30 mbar H2 at room temperature makes the reading history-dependent (absorption vs. desorption) near that range	PdAu alloying (~25 at% Au); PdCu alloying (~5 at% Cu, also fixes CO poisoning to 5,000 ppm); Pd particle size <10 nm, hysteresis eliminated below ~3 nm at 30 °C	[2]
CH4	Not perfectly excluded at elevated temperature; SiC Schottky diodes detect H2 and CH4 simultaneously rather than excluding one	WO3-capped Pd-AlGaN/GaN HEMT reached ≥8× discrimination vs. CH4 (and CO2, NH3, H2S, NO2) at 150–250 °C	[19][34]
CO2	Minimal interferent in the selectivity-engineered HEMT study; not exhaustively characterized elsewhere	—	—
He	No dissociative chemistry on Pd/Pt expected; not reported as an interferent in any source reviewed, but not explicitly ruled out either	—	—
Temperature	H2 absorption into bulk Pd has a 60–80 kJ/mol activation energy, so kinetics and baseline drift with ambient swings unless compensated	On-chip heater plus thermosensor feedback loops	[2][16]
Humidity	No device type reviewed reported acceptable performance at 95% RH across the Pd-nanostructure literature	Certain SiC-MOS designs and AppliedSensor’s HLS-440P report better tolerance	[2][5][31][38]
Pressure	No dedicated high-pressure (multi-bar to hundreds-of-bar) characterization study was found; downhole sensor-electronics work typically compensates rather than relying on pressure-immune physics	Needs a dedicated bench qualification before downhole use	—
Oxygen	Not required: H2 dissociates directly on bare Pd/Pt. Some studies show larger, faster signal in N2 than in air because O2 competes for adsorption sites	Air- and N2/anoxic-calibrated baselines differ and must be tracked separately	[42]

H2S plays a double role: the strongest catalytic poison, and at natural seeps a co-evolved gas Hestia may want to detect. NREL/JRC benchmarking under ISO 26142 found poisoning resistance generally good, with SO2/H2S/NO2/HMDS baseline shifts usually not persisting after removal, but the tests were short ISO exposures, not multi-year H2S-rich field deployment [41].

Use for geologic H₂

Field verdict

The physics is mature and genuinely O2-independent, an advantage downhole. SiC/GaN Schottky and HEMT variants are the best match for high-temperature borehole work (30 days at 500 °C; demonstrated to 800 °C), and a passive Schottky diode needs only two conductors to surface. H2S poisoning, common at Hestia’s target seeps, is the most likely field failure mode, and no source characterizes performance at downhole pressure. Do not deploy bare Pd/Pt-gate devices unattended in H2S-bearing settings without a protective membrane, a cross-validating sensor, or a recalibration schedule.

Soil-gas surveys. Not today’s default: electrochemical cells and portable analyzers such as the GA5000 anchor current field protocols [43]. Work-function devices suit fixed, unattended stations instead: no O2 dependence suits anoxic soil pockets, and the best lab devices (floating-gate FET, 5–90 ppb; nanostructured HEMTs, 50 ppb) cover the ~1 ppm exploration-anomaly floor against the ≈0.53 ppm background. H2S cross-poisoning is the limitation: budget a protective membrane, periodic recalibration, or an orthogonal cross-check (thermal-conductivity cell, electrochemical cell, or GC).

Downhole logging. The strongest use case for the SiC/GaN Schottky/HEMT branch: Pd/Pt-SiC diodes hold 30 days at 500 °C, and Pt-AlGaN/GaN diodes have run at 800 °C, temperatures no Si-based MOSFET survives. A passive Schottky diode or MIM/MIS capacitor needs only two conductors to surface, unlike a full MOSFET or a heated device needing local power. Downhole-electronics packaging work on AlGaN/GaN, Ga2O3/GaN, and SiC/GaN heterostructure FETs (not H2-specific) targets 250–400 °C and may be directly reusable here [44]. No source reviewed reports hydrostatic-pressure characterization at downhole pressures for this family; treat that as untested, not fine.

Continuous monitoring stations. The best-matched commercial use case today: AppliedSensor/ams HLS-440-family FE sensors and H2scan’s MIS-hybrid GRIDSCAN/HY-ALERTA lines are field-proven, ATEX/IP-rated, and already used for transformer health, FCEV safety, and process monitoring. Their native ranges (H2scan 25 ppm–5%; AppliedSensor 0–10% H2) are tuned to the 4 vol% LEL, not to a sub-10-ppm anomaly against a 0.53 ppm background; the ppb-class lab devices that could resolve one are not yet packaged commercial products.

Bottom line. MIS/MOSFET devices suit surface stations; SiC/GaN Schottky/HEMT devices suit high-temperature downhole work; static MIM devices are valid but commercially immature; chemicurrent MIM devices are UHV research tools, not exploration-ready. H2S co-poisoning must be engineered around, not assumed away. See Why H2 is hard and Geologic context.

Vendors & products

Vendor	Product line	Device physics	Key specs	Price class	Source
AppliedSensor (Reutlingen, Germany; acquired by ams AG, 2014)	HLS-440 / HLS-440P / HLS-442 “Field Effect (FE)” hydrogen leak modules	Heated FET with catalytic-metal gate; FE signal + thermal-conductivity channel for selectivity	HLS-440P: 0–10% H2, IP6K9, ATEX Zone 2, CAN interface; crash-tested on FCEV SUVs under a DOT/NHTSA program	Automotive-module class (no public price)	[38]
ams AG / ams-OSRAM	Continued HLS-44x line post-acquisition	Same FE/thermal-conductivity hybrid	HLS-442: raw FE signal at 10-bit resolution	—	[38]
H2scan	GRIDSCAN 5000/5000 HP (transformer DGA); HY-ALERTA 1600/5020/5021 (safety monitors); HY-OPTIMA 700B/2700/5000 (process analyzers)	Hybrid: MIS interfacial-charge device plus Pd-Ni thin-film resistor (Pd-Ni chosen to avoid the α↔β transition at high H2 overpressure)	GRIDSCAN 5000: 25–5,000 ppm, ±15 ppm or ±10% (HP); HY-ALERTA: 4,000 ppm–5% H2, t90 ≤60 s; HY-OPTIMA 2700: explosion-proof, to 95% RH, 60 °C; claimed >10-yr no-recalibration	Industrial instrument class	[39]
SenSiC AB (Kista, Sweden; Linköping SiC-FET spin-off)	iCSiC (combustion control), iQSiC (CO+O2+NOx), EmiSiC (marine NH3 slip)	SiC-FET/SiC-MOS chip placed in-situ directly in hot exhaust gas	Stated window 175–450 °C; in-situ placement eliminates sample-cooling lag	Industrial process-control class	[30]
NevadaNano	MPS™ Hydrogen & H2 Gas Sensor, MPS 3.0–6.0 ULP	Not a work-function device: MEMS membrane, Joule heater/RTD reads bulk thermal-conductivity/heat-capacity signature. Listed for contrast; see Thermal conductivity & acoustic sensors	0–100% LEL, 0.1% LEL resolution, T90 <20 s, −40–75 °C, 0–100% RH, 1.35–29 mW; claimed poisoning-immune, no field calibration	OEM module class	[45]
Linköping University (academic, not a vendor)	Original Lundström Pd-gate MOSFET (1975); 25-yr SiC-FET program feeding SenSiC	MOSFET, MOS capacitor, Schottky, SiC-FET: the academic lineage underlying most of this page	—	N/A	[24][30]
Amphenol Advanced Sensors / SGX Sensortech	EC4 series, SGX-BLD1/SGX-BLD1A	FET technology indicated for this space, but exact device physics per SKU not independently confirmed; do not cite externally without verifying the primary datasheet	SGX-BLD1: automotive/industrial; SGX-BLD1A-Compact: battery-failure detection	Automotive/industrial class	(unverified)
Bosch	Fuel-cell H2 leak sensors + “virtual hydrogen sensor”	MEMS leak sensors (principle not detailed in sources); the “virtual sensor” is a software construct, not a physical device	Integrated into Bosch FCEV powertrain sensor portfolio	Automotive OEM class	(unverified)
Academic/national-lab groups (Linköping; Georgia Tech/Ren-Pearton; Korean/Chinese groups)	Pd/Pt-SiC and Pt-AlGaN/GaN high-temperature Schottky diodes; ppb-class floating-gate FET and HEMT devices	See Device classes and Performance envelope, above	See tables above	Research-stage, TRL 3–5	[11][17][19][18]

Open questions

1. No downhole hydrostatic-pressure characterization exists for this family. Commission bench data before specifying anything downhole.
2. NREL/JRC’s H2S data [41] used ISO 26142 concentrations set for fuel-cell/industrial safety, not seep-relevant H2:H2S ratios. A dedicated months-long poisoning study is needed.
3. No field-deployed static-MIM product was identified. Confirm whether this is a real commercialization gap or such products exist under other names.
4. SGX Sensortech/Amphenol (EC4, SGX-BLD1) device physics is not independently confirmed. Verify against primary datasheets before citing these as work-function devices.
5. Investigate whether the AlGaN/GaN, Ga2O3/GaN, SiC/GaN downhole-electronics packaging effort [44] can pair with the HEMT/Schottky physics above.
6. The “no O2 required” advantage is qualitative only (larger, faster signal in N2 than air [42]); a controlled air-vs-anoxic comparison is needed.
7. Cost data is essentially absent from the public literature. Treat estimates as order-of-magnitude until a vendor quote is obtained.
8. Ga2O3-based catalytic-gate sensors have not been demonstrated (only Ga2O3 for general extreme-environment electronics); worth checking given its wider bandgap than GaN/SiC.
9. Ref-36 drift claim, resolved: the <30 mV/month figure [36] is confirmed to Morita, Nakamura, Kim, Sensors and Actuators B 33 (1996); no further verification needed.

Related pages

• Why H2 is hard
• Geologic context
• MOX chemiresistors
• Palladium resistive sensors
• Catalytic pellistors
• Thermal conductivity & acoustic sensors
• Electrochemical cells
• Plasmonic metamaterials

References

Full reference list (45 entries)

1. General palladium hydrogen-absorption/solubility background. US Patent 9,618,465, https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/9618465.
2. “High-Performance Nanostructured Palladium-Based Hydrogen Sensors — Current Limitations and Strategies for Their Mitigation,” ACS Sensors, DOI 10.1021/acssensors.0c02019, https://pmc.ncbi.nlm.nih.gov/articles/PMC7735785/.
3. “Field-Effect Transistors for Gas Sensing,” IntechOpen, https://www.intechopen.com/chapters/54938.
4. Palmisano, V. et al., “Selectivity and resistance to poisons of commercial hydrogen sensors,” https://www.sciencedirect.com/science/article/pii/S0360319915005224.
5. See ref. 2.
6. Lundström, I. et al., “Twenty-five years of field effect gas sensor research in Linköping,” Sensors and Actuators B 121(1), 247–262 (2007), DOI 10.1016/j.snb.2006.09.046.
7. “Hydrogen gas sensitive semiconductor sensor,” US Patent, https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/7622080.
8. “Sensing properties of palladium-gate MOS (Pd-MOS) hydrogen sensor based on plasma grown silicon dioxide,” https://www.sciencedirect.com/science/article/abs/pii/S0925400599000696.
9. Work-function change on Pd(100)/(110)/(111) single crystals upon H2 adsorption, surface-science literature (unverified — no specific paper identified; the cited 0.20/0.36/0.18 eV values are consistent with the published single-crystal literature, but this entry lacks a checkable author/title/venue).
10. “Temperature-dependent work function shifts of hydrogenated/deuteriated palladium,” https://www.sciencedirect.com/science/article/abs/pii/S0039602804002262.
11. Fogelberg, Eriksson, Dannetun, Petersson, J. Appl. Phys. 78, 988 (1995).
12. Eriksson, M., Lundström, I., Ekedahl, L.-G., “A model of the Temkin isotherm behavior for hydrogen adsorption at Pd–SiO2 interfaces,” J. Appl. Phys. 82(6), 3143–3146 (1997), https://pubs.aip.org/aip/jap/article-abstract/82/6/3143/178806.
13. “Structure and Technological Parameters’ Effect on MISFET-Based Hydrogen Sensors’ Characteristics,” MDPI, DOI 10.3390/s23063273.
14. General Temkin isotherm formulation, adsorption-science literature (unverified — no specific paper identified).
15. See ref. 8; “Investigation of Pd/TiO2/Si MIS capacitor as hydrogen sensor,” https://www.researchgate.net/publication/271208400_Investigation_of_PdTiO2Si_MIS_capacitor_as_hydrogen_sensor.
16. “Performance Degradations of MISFET-Based Hydrogen Sensors with a Pd-Ta2O5-SiO2-Si Structure During Long-Term Operation,” Sensors 19(8), 1855 (2019), DOI 10.3390/s19081855.
17. “Reactions of hydrogen at the interface of palladium-titanium dioxide Schottky diodes as hydrogen sensors,” https://www.sciencedirect.com/science/article/abs/pii/0039602894900809.
18. “Highly Fast Response of Pd/Ta2O5/SiC and Pd/Ta2O5/Si Schottky Diode-Based Hydrogen Sensors,” PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC7913499/.
19. Baranzahi, A. et al., “Pd- and Pt-SiC Schottky diodes for detection of H2 and CH4 at high temperature,” https://www.sciencedirect.com/science/article/abs/pii/S0925400501007250.
20. “Metal-insulator-metal sensors monitoring charge flow during thermal desorption,” https://www.sciencedirect.com/science/article/abs/pii/S0039602818301134.
21. “Hydrogen detection method using gas sensor having a metal oxide layer,” US Patent, https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/10591432. (does not support the claim)
22. Nienhaus, H., Gergen, B., Weinberg, W.H., McFarland, E.W. et al., “Detection of chemically induced hot charge carriers with ultrathin metal film Schottky contacts,” Surface Science, https://www.sciencedirect.com/science/article/abs/pii/S0039602802016254.
23. US Patent 7,274,082 B2, “Chemical sensor using chemically induced electron-hole production at a Schottky barrier,” https://patents.google.com/patent/US7274082.
24. Lundström, I., Shivaraman, S., Svensson, C., Lundkvist, L., “A hydrogen-sensitive MOS field-effect transistor,” Applied Physics Letters 26, 55–57 (1975), DOI 10.1063/1.88053.
25. “PLL-Based Readout Circuit for SiC-MOS Capacitor Hydrogen Sensors in Industrial Environments,” PMC, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8879939/.
26. “MOSFE-Capacitor Silicon Carbide-Based Hydrogen Gas Sensors,” PMC, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10098966/.
27. “CMOS Compatible Hydrogen Sensor Using Platinum Gate and ALD–Aluminum Oxide,” Sensors 24(10), 3020 (2024), DOI 10.3390/s24103020.
28. “A study on MIS Schottky diode based hydrogen sensor using La2O3 as gate insulator,” https://www.sciencedirect.com/science/article/abs/pii/S0026271412001060.
29. “Design and fabrication of MOSFET type hydrogen gas sensor using MEMS process,” https://www.researchgate.net/publication/280926583.
30. Lloyd Spetz, A., “SiC Based Field Effect Gas Sensors for Industrial Applications,” physica status solidi (a) 185(1), 15 (2001); SenSiC, https://sensic.se/technology/.
31. SiC MOS capacitor water-vapor tolerance data (see refs. 25–26).
32. “Fast hydrogen detection via plasmonic metal-insulator-metal technology,” Applied Optics, https://opg.optica.org/ao/abstract.cfm?uri=ao-57-34-10032.
33. Song, J., Lu, W., Flynn, J.S., Brandes, G.R., “Pt−AlGaN/GaN Schottky diodes operated at 800 °C for hydrogen sensing,” Applied Physics Letters 87, 133501 (2005), DOI 10.1063/1.2058227.
34. Pt/GaN, Pt-AlGaN/GaN, and Pd/AlGaN/GaN HEMT hydrogen-sensing performance data: PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC10099067; MDPI, https://www.mdpi.com/1424-8220/23/7/3465. (does not support the “Sensitivity Factor 14,685” or “50 ppb” claims attributed to it elsewhere on this page)
35. See ref. 33; Ren/Pearton-group HEMT lineage.
36. Morita, M., Nakamura, K., Kim, J.-H., “Langmuir analysis on hydrogen gas response of palladium-gate FET,” Sensors and Actuators B 33 (1996), https://www.sciencedirect.com/science/article/abs/pii/0925400596019569.
37. “Sub-10 parts per billion detection of hydrogen with floating gate transistors built on semiconducting carbon nanotube film,” https://www.sciencedirect.com/science/article/abs/pii/S0008622321004656.
38. AppliedSensor/ams HLS-440 family: Tracxn, https://tracxn.com/d/companies/appliedsensor/ (link dead); datasheet, https://datasheets.globalspec.com/ds/ams/hls-440p-a/; NREL FCEV crash-test report, https://docs.nrel.gov/docs/fy12osti/56177.pdf.
39. H2scan product data: https://h2scan.com/product-types/safety-monitors/.
40. Lundström et al. (see ref. 6).
41. Palmisano, V. et al. (see ref. 4); ISO 26142:2010, https://www.iso.org/standard/52319.html.
42. “Oxygen effect on the operation of the MOS-structure-based hydrogen sensor,” https://www.sciencedirect.com/science/article/abs/pii/092540059487039X.
43. “Exploring natural hydrogen hotspots: a review and soil-gas survey design for identifying seepage,” Geoenergy, https://www.lyellcollection.org/doi/full/10.1144/geoenergy2023-014.
44. “Sensing Electronics for Downhole Drilling in Extreme Environments,” https://labpartnering.org/technologies/3f72efda-2142-450a-aa56-547d82b5ed4b; US Patent 11,680,476, https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/11680476.
45. NevadaNano MPS datasheet: https://nevadanano.com/mps-hydrogen-gas-sensor/.