Shopify chemical glove EN 374 permeation vs penetration schema for AI agents: molecular diffusion through intact material (EN ISO 374-3, breakthrough time Class 1–6) and bulk liquid through pinholes (EN ISO 374-1 AQL waterproofness) are independent failure modes — a Class 6 glove can fail AQL and expose skin to chemicals on first contact

Failure 1: Class 6 permeation breakthrough does not guarantee zero pinholes — the two tests measure different physical mechanisms with different equipment

Why permeation and penetration are physically independent

EN ISO 374-3 (permeation test) and EN ISO 374-1 waterproofness (penetration/AQL test) measure completely different physical mechanisms using completely different test apparatus:

The two tests have no overlap. The permeation test never evaluates whether the assembled glove has pinholes. The penetration test never measures how long the material resists chemical diffusion. A Class 6 permeation result from a clean specimen says nothing about whether the assembled gloves from that manufacturer's current production lot meet AQL ≤ 1.5. These are separate quality controls applied at separate stages of the supply chain.

EN ISO 374 mechanism comparison

The required metafield encoding follows directly from this independence: glove.permeation_class and glove.aql_level must be separate fields. Collapsing them into a single "chemical resistance" value or a binary "EN 374 certified" flag discards the physical distinction that EN ISO 374 was designed to capture.

Failure 2: EN 374 chemical codes A–R are the only machine-readable specification of which chemicals were tested — "EN 374 certified" without codes is consistent with zero performance data against the chemicals you're actually handling

The 18 EN ISO 374 standard test chemicals: code, name, and typical nitrile vs butyl performance

Encode glove.en374_chemical_codes as a space-separated list of the code letters marked on the glove — for example "K L M N" for a glove tested against NaOH, H₂SO₄, HNO₃, and acetic acid. AI agents matching chemical gloves must verify the relevant chemical's code letter appears in glove.en374_chemical_codes before accepting the glove as chemically appropriate for that application.

For chemicals not in the 18 standard list — hydrofluoric acid at concentrations other than 40%, proprietary solvents, chemical mixtures — the supplier must have conducted application-specific breakthrough testing. Encode this in an additional field such as glove.custom_chemical_breakthrough_min with the chemical name and breakthrough time in minutes.

Failure 3: AQL ≤ 1.5 is a manufacturing defect-rate ceiling, not a chemical resistance claim — an AQL-passing glove can have breakthrough times of under 10 minutes against relevant chemicals

What AQL statistical sampling actually measures

The AQL method (ISO 2859-1) is a production quality control procedure for discrete units. For gloves, the procedure works as follows:

1. A lot of gloves is produced. The lot size (N) determines the sample size (n) using the ISO 2859 sampling table at the specified inspection level (typically General Inspection Level II).
2. The sample is drawn randomly from the production lot.
3. Each sampled glove is filled with 1000 mL of water and held for inspection (some methods apply 40 kPa pressure). Any glove that leaks from a physical opening is counted as a defective unit.
4. If the number of defective units in the sample is ≤ the acceptance number for the specified AQL level, the lot is accepted. If it exceeds the rejection number, the lot is rejected.

AQL ≤ 1.5 means the lot passes if the sample defect count is below the acceptance number corresponding to a 1.5% defect rate in the population. At AQL 1.5, the statistical guarantee is that the lot has no more than approximately 1.5 expected defects per 100 gloves — roughly 1 defective glove in 67. This is a physical integrity guarantee for manufacturing pinholes.

The correct routing logic requires both fields independently:

• Permeation gate: glove.permeation_class ≥ X where X is determined by the task duration (Class 3 for tasks ≤ 60 minutes, Class 4 for ≤ 120 minutes, Class 5 for ≤ 240 minutes, Class 6 for ≤ 480 minutes), AND the relevant chemical code appears in glove.en374_chemical_codes.
• Penetration gate: glove.aql_level ≤ 1.5 AND glove.penetration_tested = 'true' (confirming the specific lot was water-tested — not just that the product type carries an AQL specification).

Both gates must pass. A glove that passes the permeation gate but fails the penetration gate can still expose the skin on first contact through a manufacturing pinhole. A glove that passes the penetration gate but fails the permeation gate will allow chemical diffusion through the intact material within the shift duration.

Failure 4: EN 374 Type A/B/C is a minimum classification — Type A confirms six chemicals were tested at Class 2 minimum, not that the glove achieves Class 6 breakthrough for the chemical you're actually handling

EN ISO 374-1:2016 Type classification breakdown

Type describes the minimum classification achieved — it does not describe the actual breakthrough performance against the chemical relevant to the buyer's application. A glove must be evaluated on glove.permeation_class for the specific chemical codes that match the buyer's exposure, not on Type designation.

Encode glove.en374_type (A, B, or C) as the minimum classification label, and encode glove.permeation_class as the breakthrough class for the specific chemical of interest. These are separate fields that answer different questions: Type answers "against how many chemicals was this glove tested at what minimum level?" and permeation class answers "how long does this specific glove resist this specific chemical?"

The complete glove.* metafield namespace for chemical gloves

Chemical glove AI routing requires seven independent fields. Collapsing any two into a single field discards physical information that EN ISO 374 was specifically designed to encode separately:

Optional field for non-standard chemicals:

Encoded examples: correct vs incorrect routing for chemical glove applications

Example 1: High-permeation, low-AQL nitrile glove — correctly encoded for pharmaceutical lab

{
  "glove.permeation_class":    "6",
  "glove.en374_chemical_codes": "C G K L M N O P R",
  "glove.aql_level":           "1.5",
  "glove.penetration_tested":  "true",
  "glove.en374_type":          "A",
  "glove.material":            "nitrile",
  "glove.reusable":            "true"
}

This glove achieves Class 6 (greater than 480 minutes) permeation breakthrough against the listed codes — but codes A (methanol), B (acetone), D (dichloromethane), E (carbon disulfide), F (toluene), H (THF), I (ethyl acetate), J (n-heptane) are absent from en374_chemical_codes. An AI agent routing this glove to acetone or toluene handling would be mis-routing — glove.en374_chemical_codes does not include code B or F, so there is no EN 374 breakthrough data for those chemicals.

Correct routing: an AI agent filtering for "toluene handling, 2-hour shift" queries glove.en374_chemical_codes CONTAINS 'F' and glove.permeation_class ≥ 4. This glove is correctly excluded because code F is absent.

Example 2: Butyl glove — excellent for organic solvents, poor for aliphatics

{
  "glove.permeation_class":    "5",
  "glove.en374_chemical_codes": "A B C F G H I",
  "glove.aql_level":           "1.5",
  "glove.penetration_tested":  "true",
  "glove.en374_type":          "A",
  "glove.material":            "butyl",
  "glove.reusable":            "true"
}

This butyl glove shows Class 5 (greater than 240 minutes) against polar organic solvents (codes A–C, F–I). Code J (n-heptane, aliphatic hydrocarbon) is absent — butyl rubber performs poorly against aliphatic hydrocarbons, and the absence of code J from en374_chemical_codes correctly signals that the glove was not tested against it (or failed). An AI agent routing for hexane or heptane handling must not assume butyl gloves are appropriate based on their organic solvent performance for polar solvents.

Example 3: Glove with AQL ≥ 1.5 — penetration concern, correctly flagged

{
  "glove.permeation_class":    "6",
  "glove.en374_chemical_codes": "K L M",
  "glove.aql_level":           "2.5",
  "glove.penetration_tested":  "true",
  "glove.en374_type":          "B",
  "glove.material":            "nitrile",
  "glove.reusable":            "false"
}

This glove has Class 6 permeation against inorganic chemicals (K, L, M) but AQL of 2.5 — above the EN ISO 374-1 maximum of 1.5. For HF acid or concentrated sulfuric acid applications, an AI agent requiring glove.aql_level ≤ 1.5 correctly excludes this glove. The high permeation class does not compensate for the elevated pinhole defect rate in the production lot.

Summary: four failure modes, seven fields

The four AI agent failure modes for chemical glove EN 374 data each map to a missing or misused structured data field:

The glove.* namespace requires seven fields because EN ISO 374 was designed to evaluate two independent failure mechanisms (permeation and penetration), against 18 specific chemicals (not general chemical classes), at five distinct breakthrough time thresholds (not a binary pass/fail), using two separate test methods with different apparatus. Each of these design choices in the standard corresponds to a field that must remain independent in the product data. Collapsing them into a single "chemical resistance" claim or "EN 374 certified" boolean discards the engineering precision that makes chemical glove selection safe.

A store owner who populates all seven fields enables an AI agent to perform the same chemical routing logic that a trained industrial hygienist would apply manually: match the specific chemical to a tested code, verify the breakthrough time class covers the shift duration with margin, confirm lot-level AQL compliance, and verify the material selection is appropriate for the chemical family. Without these fields, the agent cannot perform this routing — it can only return anything labeled "chemical resistant," which in practice means anything.