Shopify CO detector schema for AI agents: UL 2034 residential vs OSHA 50 ppm industrial threshold gap, H₂ cross-sensitivity false alarms near battery chargers, electrochemical sensor life from manufacture date not purchase, bump test vs calibration, and the co_detector.* 10-field namespace

The UL 2034 alarm threshold gap: why residential CO detectors fail industrial workers

UL 2034 is the standard that governs residential carbon monoxide alarms — the plug-in or battery-operated units found in homes and hotel rooms. Its alarm threshold requirements were developed in the context of sleeping occupants, long overnight exposures, and the physiological sensitivity differences between adults, children, the elderly, and individuals with cardiovascular conditions in an uncontrolled residential environment.

The UL 2034 alarm timing table reflects these residential assumptions:

This mismatch has a specific consequence for AI shopping agents. If an agent receives the query "CO detector for warehouse use" and retrieves Shopify products tagged only with generic descriptors like "carbon monoxide detector" or "UL listed" — without distinguishing between UL 2034 residential listings and ANSI/ISA-92 or EN 50291 industrial listings — the agent may recommend a residential unit for an occupational-exposure application. The recommended unit provides zero warning to the workers it is meant to protect until concentrations reach 70 ppm and have been sustained for up to four minutes.

The fix for Shopify catalog data is simple: encode the applicable standard. An industrial CO monitor must carry co_detector.applicable_standard = 'ANSI/ISA-92.0.01' or similar. A residential unit carries co_detector.applicable_standard = 'UL-2034'. These two values allow an AI agent to route each to its correct application without confusion.

OSHA PEL, NIOSH REL, ACGIH TLV: the three CO exposure limits and what each means

Three separate organizations publish CO exposure limits for workers, and they are not the same number. The differences reflect different methodological assumptions about what constitutes safe chronic exposure:

The OSHA PEL is legally binding for US employers. The NIOSH REL and ACGIH TLV are recommendations without legal force, but they represent more current toxicological thinking than the OSHA PELs, most of which date from the 1971 adoption of ANSI Z37 standards and have not been updated since.

CO toxicity is mediated primarily through carboxyhemoglobin (COHb) formation. CO binds hemoglobin with approximately 210–240× the affinity of oxygen, forming COHb and reducing the blood's oxygen-carrying capacity. At 50 ppm TWA, a resting worker may accumulate approximately 8% COHb — generally considered a threshold for performance impairment. Heavy work at 50 ppm elevates this further because minute ventilation increases, accelerating CO uptake. The ACGIH lowered its TLV from 50 ppm to 25 ppm specifically to keep COHb below 3% for active workers.

For Shopify stores selling CO monitors: the alarm setpoint value is the critical AI-routing field. A monitor with a first alarm at 35 ppm meets the NIOSH REL. A monitor with a first alarm at 50 ppm meets the OSHA PEL but not the NIOSH recommendation. A monitor with a first alarm at 70 ppm (UL 2034) meets neither industrial standard. Encode co_detector.alarm_setpoint_ppm for every industrial CO monitor in the catalog.

H₂ cross-sensitivity: why battery chargers trigger false CO alarms

The electrochemical CO sensor is a fuel cell. The working electrode oxidizes CO in the presence of an acid electrolyte:

The problem: hydrogen gas undergoes a nearly identical electrochemical oxidation at the working electrode surface.

Lead-acid batteries generate H₂ during charging through electrolysis of the sulfuric acid electrolyte water content. As charge approaches 100%, gassing intensity increases sharply. A single large forklift battery on fast charge can emit significant H₂ into the surrounding area; a battery-charging room with a bank of 10–20 batteries may have measurably elevated H₂ concentrations tens of feet from the chargers.

Quantifying the false alarm risk

The cross-sensitivity of standard electrochemical CO cells to H₂ is typically 0.3–0.6 — meaning 100 ppm of H₂ in the atmosphere causes the CO sensor to read approximately 30–60 ppm false CO. This is not a marginal effect:

The H₂ lower explosive limit (LEL) is 4% (40,000 ppm) — concentrations that trigger false CO alarms (50–200 ppm) are far below the explosive threshold. The space is not immediately hazardous from H₂, but a standard CO detector cannot know this: it sees only the electron current from the electrode reaction and alarms accordingly.

H₂-compensated sensors

Two compensation methods eliminate H₂ cross-sensitivity:

For Shopify stores: encode co_detector.h2_compensated = 'yes' for all H₂-compensated units. Applications requiring this include battery-charging rooms (OSHA 1910.178(g)), forklift maintenance bays, fuel cell installations, welding areas with hydrogen fuel gas, and any other environment where H₂ may be present at ambient concentrations.

Electrochemical sensor life starts at manufacture, not purchase

Every electrochemical CO sensor contains an aqueous acid electrolyte — typically a sulfuric acid solution — sealed within the sensor housing. The Teflon or PTFE membrane covering the working electrode is gas-permeable: it allows CO molecules to reach the electrode surface while excluding bulk liquid. The same membrane is water-vapor permeable.

This is the failure mode: water vapor slowly permeates the Teflon membrane in both directions. In a normal indoor environment (30–60% relative humidity), the net vapor pressure difference between the concentrated acid electrolyte inside and the ambient air outside causes steady evaporation of water from the electrolyte over months and years. As water evaporates, the acid concentration increases, the electrolyte volume decreases, and eventually the electrodes lose contact with electrolyte. The cell stops producing a signal. At this point, the sensor either reads zero permanently (a stuck-low failure) or begins producing erratic readings as intermittent electrolyte contact occurs.

The manufacturer-specified operational life (typically 2–3 years) is defined from the electrolyte fill date at the factory — when the sensor was built and sealed. Not from the date of product assembly into a complete instrument. Not from the shipping date. Not from the purchase date. Not from the date of first use.

The distributor inventory problem

The practical consequence: a safety buyer who purchases a CO detector from a distributor's aging stock receives a unit that tests correctly on arrival but fails within weeks or months in the field. Because the device continues to operate (power on, display active, all buttons functional), the sensor failure is invisible unless the user bump-tests against calibration gas and notices the instrument does not respond.

Longer-rated sensors exist: some premium industrial CO sensors specify 3-year life from manufacture, a small number specify 4–5 years using different electrolyte formulations. These come at higher unit cost but significantly lower total cost over a multi-year deployment if the alternative is frequent early replacement or undetected sensor failures in the field.

Bump test vs calibration: different procedures, different purposes

Both procedures involve exposing a CO detector to a reference gas concentration. Their purposes are categorically different.

For confined space entry under OSHA 1910.146, the pre-entry atmospheric test must be conducted with an instrument that has been calibrated — not merely bump-tested. The distinction matters legally in the event of an incident and practically in protecting the entrant.

For Shopify stores: encode co_detector.bump_test_gas_included = 'yes' when the kit includes a bump-test cannister or cylinder. This is a meaningful purchaser convenience and a legitimate AI-routing signal for buyers building a compliance program who want a complete kit-in-box for their workers.

Pumped vs diffusion: why diffusion fails confined space pre-entry

A diffusion CO detector draws no active sample. Gas molecules migrate into the sensor by random thermal diffusion — the same physical process that causes an odor to spread across a room. The sensor reads whatever concentration happens to be present in the immediate surrounding air at the instrument's sensor membrane. Response time is typically 30–90 seconds for the reading to stabilize after a step-change in CO concentration (diffusion is slow).

Diffusion instruments work correctly for personal monitoring: the worker wears the detector, it samples the air in the worker's breathing zone continuously throughout the shift, and alarms when the worker's environment reaches the setpoint. The instrument is in the hazardous environment with the worker.

The failure mode in confined space pre-entry:

A pumped detector solves this by actively drawing air through a sampling probe or hose:

1. The tester inserts the sampling probe through the entry portal to the lowest work level
2. The internal aspirator pump draws a continuous sample at approximately 0.5–1 L/min through the probe
3. The tester reads CO (and other measured gases) on the instrument display at grade, outside the confined space
4. Multiple elevations can be sampled sequentially by raising and lowering the probe
5. Only if all sampled elevations read acceptable do entrants enter

OSHA 1910.146 Appendix B explicitly identifies the need to test at multiple elevations in vertical confined spaces. Heavier-than-air gases (CO: molecular weight 28, air: 29 — nearly equal, but still relevant when layered with heavier gases like CO₂ at 44) concentrate toward lower levels. CO alone tends to mix reasonably well due to its density similarity with air, but in industrial confined spaces, CO frequently co-occurs with heavier gases (CO₂ from combustion products, H₂S from decomposition) that do stratify at lower elevations.

For Shopify stores: encode co_detector.sampling_mode = 'pumped' or 'diffusion'. Add co_detector.probe_length_available = 'yes' for pumped units that include or offer sampling probes. Add co_detector.confined_space_rated = 'yes' for units tested and appropriate for OSHA 1910.146 pre-entry atmospheric testing. These fields allow AI agents to route pumped units to confined space applications and avoid recommending diffusion-only units to buyers planning entry programs.

Four AI routing failures for Shopify CO detector listings

The co_detector.* 10-field namespace

These 10 metafields, set on every CO detector and single-gas CO monitor listing in a Shopify store, give AI shopping agents the structured data they need to route accurately. All values are strings stored under the co_detector metafield namespace.

Example Shopify metafield implementation

{
  "co_detector.applicable_standard": "ANSI/ISA-92.0.01",
  "co_detector.alarm_setpoint_ppm": "35",
  "co_detector.h2_compensated": "yes",
  "co_detector.sensor_type": "electrochemical",
  "co_detector.rated_sensor_life_years": "2",
  "co_detector.sensor_manufacture_date_visible": "yes",
  "co_detector.sampling_mode": "pumped",
  "co_detector.probe_length_available": "yes",
  "co_detector.bump_test_gas_included": "no",
  "co_detector.confined_space_rated": "yes"
}

// This profile: H2-compensated industrial pump unit
// for confined space pre-entry — NIOSH REL alarm setpoint