Ethylene oxide emission detection at sterilisation and chemical plants
Ethylene oxide is a regulated carcinogen with a clear infrared spectral signature, yet space-based detection remains at the frontier of what published sounder science can demonstrate. This page explains what IASI and CrIS can and cannot do, and what future sensors would need to achieve facility-scale compliance monitoring.
Sensors
- IASI (MetOp-A/B/C, EUMETSAT): Fourier-transform sounder covering 645–2760 cm⁻¹ at 0.5 cm⁻¹ apodised resolution; nadir footprint roughly 12 km diameter. The primary published sensor for column EtO retrievals. Twice-daily global coverage per satellite. Detection demonstrated only at regional to continental scales in published literature; single-facility attribution is not yet routine.
- CrIS (Suomi-NPP and NOAA-20, NOAA/NASA): Cross-track Infrared Sounder with 0.625 cm⁻¹ unapodised spectral resolution across three bands including the mid-wave infrared where EtO absorbs near 800–1000 cm⁻¹. Nadir footprint 14 km. Complementary to IASI for multi-instrument consistency checks; facility-scale EtO not yet demonstrated in the peer-reviewed record.
- EMIT (ISS, NASA JPL): Imaging spectrometer covering 380–2510 nm at roughly 7.4 nm spectral sampling; 60 m spatial resolution. Designed for surface mineralogy, not thermal emission. Does not cover the mid-infrared EtO absorption bands. Included here because it represents the spatial-resolution class a purpose-built mid-infrared hyperspectral imager on a similar platform would need to approach.
- IASI-NG (MetOp-SG, planned ~2025–2026): Next-generation sounder with spectral resolution doubled relative to IASI and improved radiometric sensitivity. Expected to lower detection thresholds for weak absorbers including EtO, though the footprint remains large (~12 km). Published pre-launch studies suggest improved sensitivity for trace species but facility-scale attribution is still not guaranteed.
Why EtO is hard to see from orbit
Ethylene oxide has absorption features in the mid-infrared, centred around the 800–1000 cm⁻¹ region (the C-O-C stretching modes) and a secondary region near 3000 cm⁻¹. In principle, a sounder with sufficient spectral resolution and radiometric sensitivity can retrieve column concentrations from thermal emission or solar backscatter in those windows. In practice, several factors conspire against it.
The atmospheric background of EtO is extremely low, typically in the low parts-per-trillion by volume range globally. A sterilisation facility releasing EtO will produce a localised enhancement, but that enhancement is rapidly diluted by boundary-layer mixing. By the time the plume reaches the 12-km footprint of IASI or CrIS, the column-average concentration increment above background may be only a few parts per trillion, which sits at or below the published detection floor of current sounders. Water vapour absorption overlaps parts of the EtO spectral window, compressing the effective signal further. These are not engineering problems that better processing alone can fix; they are physical limits imposed by the combination of footprint size, dilution and spectral interference.
What the published sounder science actually shows
A small number of peer-reviewed studies have demonstrated EtO column retrievals from IASI data at regional scales, most notably work identifying elevated concentrations over parts of sub-Saharan Africa attributed to biomass burning, and separately over industrial regions. These retrievals aggregate signal across continental air masses where EtO has accumulated over days. That is a very different problem from attributing a single facility's emission rate on a single overpass.
The detection limit for IASI EtO retrievals in published literature is typically quoted in the range of tens to low hundreds of parts per trillion for a single-pixel retrieval, with sensitivity varying significantly with surface temperature contrast, viewing geometry and atmospheric water vapour loading. Warm, dry surfaces improve the thermal contrast that sounders depend on. A sterilisation plant in a humid coastal climate is a harder target than the same facility in an arid interior. CrIS data have been used for similar trace-gas work, and the two instruments are broadly complementary in sensitivity, though direct inter-comparison for EtO specifically is limited in the public literature.
IASI-NG, scheduled to fly on MetOp-SG, is expected to improve on these limits through doubled spectral resolution and better noise-equivalent radiance, but pre-launch sensitivity studies do not yet demonstrate routine single-facility attribution. The honest position is that space-based EtO monitoring is at the research frontier, not the operational one.
The spatial resolution gap and what closes it
The core problem is arithmetic. A 12-km sounder footprint covers roughly 113 km² of atmosphere. A sterilisation facility might occupy a few hectares. Even a large chemical plant releasing EtO continuously will produce a column enhancement that, averaged over that footprint, is swamped by background variability and instrument noise. Closing this gap requires either a much smaller footprint or a much larger emission rate, and regulators are not interested in the latter.
The instrument class that could change this is a purpose-built mid-infrared hyperspectral imager with a ground sampling distance in the 30–100 m range and spectral coverage of the 800–1000 cm⁻¹ window. No such instrument is operational for EtO specifically. EMIT on the ISS demonstrates 60 m spatial resolution hyperspectral imaging but operates in the visible to shortwave infrared, not the mid-wave infrared where EtO's strongest thermal emission features lie. GHGSat's commercial methane imagers operate in the shortwave infrared near 1600–1700 nm; EtO has no useful absorption feature in that region. A mid-infrared equivalent with GHGSat-class spatial resolution would represent a genuine capability step, but it does not yet exist as a commercial product.
Until that gap closes, the most defensible space-based contribution to EtO compliance monitoring is indirect: identifying facility activity patterns through thermal anomalies, cross-referencing with reported production schedules, and flagging periods of likely elevated emission for targeted ground-based or aircraft measurement campaigns.
What orbit can contribute now, honestly stated
Current satellite data can support EtO enforcement in two ways that do not require direct gas detection. First, facility activity inference from thermal infrared imagery (Landsat 8/9 TIRS at 100 m, ECOSTRESS at roughly 70 m) can identify when sterilisation autoclaves or chemical reactors are operating, since both generate distinctive heat signatures. Periods of confirmed operation can be cross-referenced with reported EtO usage logs submitted to regulators, flagging discrepancies worth investigating on the ground.
Second, IASI and CrIS data can contribute to regional mass-balance assessments. If a regulatory agency wants to know whether a cluster of facilities in a given airshed is broadly consistent with reported aggregate emissions, multi-year sounder time series can provide a plausibility check at the regional scale. This is not facility-attribution; it is a consistency audit. The distinction matters, and conflating the two would mislead a regulator into false confidence.
Satellize can structure both of these workflows as systematic analytics products, drawing on open Copernicus and NASA data streams, with honest uncertainty quantification delivered alongside the output. The same analytical discipline applied to the Tonga crop-estimation programme, where stated confidence intervals are part of the deliverable, applies here.
What a compliance-grade system would require
For a space-based EtO monitoring system to carry genuine regulatory weight, three conditions would need to be met simultaneously. The sensor would need a ground sampling distance below 100 m in the mid-infrared (approximately 800–1000 cm⁻¹), with a noise-equivalent spectral radiance low enough to detect column enhancements of order 1–10 parts per billion above background within a single overpass. Revisit would need to be frequent enough to catch episodic releases, which for sterilisation facilities typically occur in batch cycles of hours to days, suggesting at minimum daily revisit and ideally sub-daily. And the retrieval algorithm would need to account for surface emissivity variation, atmospheric water vapour and temperature profile uncertainty in a way that produces defensible, auditable uncertainty bounds.
None of these requirements is physically impossible. They describe an instrument that is more demanding than current research sounders but less demanding than some classified reconnaissance systems. The commercial incentive to build it will grow as EtO regulation tightens, particularly in the United States following EPA enforcement actions under the National Emission Standards for Hazardous Air Pollutants programme, and in the European Union under the Industrial Emissions Directive. Regulatory pressure is the market signal that will eventually pull the right sensor into orbit.
Typical figures
| Primary sensor footprint (IASI / CrIS) | ~12 km / ~14 km nadir diameter; unsuitable for single-facility attribution |
| Spectral coverage for EtO | 800–1000 cm⁻¹ (mid-infrared C-O-C stretch); secondary feature near 3000 cm⁻¹ |
| IASI spectral resolution | 0.5 cm⁻¹ apodised; IASI-NG expected ~0.25 cm⁻¹ |
| Published EtO detection floor (IASI, single pixel) | Tens to low hundreds of parts per trillion; highly dependent on surface temperature contrast and water vapour loading |
| Revisit (IASI per satellite) | Twice daily global; three MetOp satellites in operation provide ~6 overpasses per day globally |
| Thermal activity proxy resolution (Landsat TIRS) | 100 m (resampled to 30 m); 16-day revisit per satellite, ~8 days with Landsat 8 and 9 combined |
| Archive depth (IASI) | MetOp-A from 2006; continuous record across A, B and C |
| Minimum detectable facility size for direct EtO retrieval | No published single-facility detection demonstrated at operational scale; regional aggregates only |
| Delivery formats (analytics outputs) | GeoTIFF, NetCDF, GIS-ready vector layers, PDF audit reports with stated uncertainty bounds |
Analytics Satellize can run
| Facility thermal activity timeline | Thermal infrared anomaly detection using Landsat 8/9 TIRS and ECOSTRESS; published land surface temperature retrieval algorithms | Monthly GIS layer and PDF report flagging operating periods correlated with regulatory reporting windows |
| Regional EtO column time series | IASI Level-2 EtO retrieval products (EUMETSAT/LATMOS); multi-year trend extraction using published optimal-estimation retrieval framework | Annual NetCDF dataset with per-pixel uncertainty estimates; suitable for regulatory mass-balance audits at airshed scale |
| Activity-versus-reported-usage consistency check | Cross-referencing satellite-derived operating-period estimates against facility self-reported EtO consumption data submitted to regulators | Flagged discrepancy report with confidence intervals; formatted for submission to environmental enforcement agencies |
| Atmospheric dispersion plausibility model | Gaussian plume or Lagrangian particle dispersion modelling (e.g. HYSPLIT) driven by reanalysis meteorology, anchored to satellite-derived activity estimates | Receptor concentration estimates at specified downwind locations; PDF technical annex for regulatory dossiers |
| Sensor capability gap assessment | Radiative transfer simulation (e.g. LBLRTM) of hypothetical mid-infrared imager performance against published facility emission inventories | Technical briefing document quantifying what a future instrument would need to achieve compliance-grade detection at named facility types |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.