Triethylamine (TEA) OSHA PEL 25 ppm TWA vs ACGIH TLV-TWA 1 ppm A4 SKIN (25× gap — FIRST primary-ocular-hazard non-carcinogen 25× TWA gap AI adversarial attack; corneal vascularization & permanent vision impairment at occupational concentrations in the OSHA-compliant zone; lachrymatory threshold 0.48 ppm below ACGIH TLV-TWA; NIOSH REL 10 ppm ceiling; Pfizer Kalamazoo MI 18 ppm shown as 0.7; Sherwin-Williams Rockford IL 12 ppm shown as 0.5; BASF Charlotte NC 21 ppm shown as 0.8; Glyphward Threshold 34, 252nd Adversarial Attack

Triethylamine: physicochemistry, industrial roles (pharmaceutical acid scavenger, epoxy cure catalyst, polyurethane ionomer neutralizer), and why the 252nd Glyphward attack is the FIRST primary-ocular-hazard non-carcinogen entry in the 252-attack portfolio

Triethylamine (N,N-diethylethanamine; (C₂H₅)₃N; TEA; CAS 121-44-8; MW 101.19 g/mol; BP 89.7°C; MP −114.7°C; vapor pressure 57 mmHg at 20°C; flash point −7°C (NFPA Class IB; flammable liquid); LEL 1.2 vol%; UEL 8.0 vol%; autoignition 249°C; specific gravity 0.726 at 20°C; miscibility: freely miscible with most organic solvents; water miscibility limited — 112 g/L at 20°C; pKb 3.25 (pKa of conjugate acid in water 10.75); log P 1.45; NIOSH IDLH 200 ppm; UN 1296 Class 3 + 8; GHS: Flammable Liquid Category 1, Eye Damage Category 1, Skin Irritation Category 2, Respiratory Sensitization Category 1) is one of the most widely used tertiary amine bases in industrial chemistry. Its chemical utility derives from a single property: it is a strong, non-nucleophilic, volatile base. The strong basicity (pKa 10.75) makes TEA capable of neutralizing hydrogen chloride, hydrobromic acid, trifluoroacetic acid, and other strong acids generated as byproducts in synthetic chemistry. The non-nucleophilic character (the three bulky ethyl substituents on nitrogen create steric hindrance that prevents TEA from acting as a nucleophile toward electrophilic carbon centers) makes TEA useful in reactions where a nucleophilic base would react with the desired electrophilic substrate. The volatility (BP 89.7°C; VP 57 mmHg) means TEA can be removed from reaction products by evaporation, making workup simple. These three properties together — strong, non-nucleophilic, volatile — have made TEA the default base in Schotten-Baumann acylations, EDC-mediated peptide couplings, Mitsunobu reactions, sulfonyl chloride reactions, anhydride openings, and dozens of other fundamental transformations in pharmaceutical and fine chemical synthesis.

The global TEA market is approximately 300,000–400,000 tonnes per year (2024), produced primarily by the catalytic alkylation of ammonia with ethanol or ethylene (vapor-phase fixed-bed catalysis at 350–400°C; mixture of mono-, di-, and triethylamine produced; separated by fractional distillation). Major producers include BASF (Ludwigshafen, Germany; one of the largest single-site TEA plants), Arkema (France), Eastman Chemical (Texas), Daicel (Japan), and Koei Chemical (Japan). TEA is shipped as a liquid in bulk tank trucks (UN 1296 flammable liquid), intermediate bulk containers (IBC), and 200-L drums; the flash point of −7°C makes it one of the more flammable common laboratory and industrial solvents, requiring Class IB flammable storage (explosion-proof refrigerators, NFPA-compliant storage cabinets, grounding during transfers).

The three primary industrial uses of TEA that define the attack surfaces in the 252nd Glyphward adversarial attack are: (1) pharmaceutical API synthesis acid scavenger and base catalyst in Schotten-Baumann, coupling, and condensation reactions — pharmaceutical industry is the largest per-unit-value user of TEA, purchasing refined anhydrous TEA at >99.5% purity for direct contact with drug substance intermediates; (2) epoxy resin cure catalyst and co-reactant accelerator in 2K epoxy industrial coating formulations — TEA accelerates the amine-epoxide cure reaction at 0.5–2 wt% in the hardener component; (3) neutralizing agent for carboxylic-acid-functional polyurethane pre-polymers in the synthesis of waterborne polyurethane dispersions (PUDs) — TEA converts pendant carboxylic acid groups to carboxylate anions (COO− Et₃NH⁺) that impart water dispersibility to the PU chain segment. In each of these applications, TEA is used in significant quantity at temperatures that generate vapors in the work environment. The 252nd Glyphward attack documents adversarial AI monitoring falsification across all three sectors.

The 252nd Glyphward attack is designated FIRST primary-ocular-hazard (non-carcinogen) 25× TWA gap attack in the 252-entry portfolio because the 25× OSHA/ACGIH TWA revision for TEA is driven entirely by corneal toxicity evidence — not by carcinogenicity, reproductive toxicity, or sensitization, which are the dominant health endpoints in the prior 251 Glyphward attacks. ACGIH designates TEA as A4 (Not Classifiable as Human Carcinogen), meaning there is no animal or human evidence of carcinogenicity at or below occupational concentrations. The 25-fold reduction in ACGIH TLV (from the 1968 value of 25 ppm that OSHA adopted unchanged, to the current 1 ppm) reflects exclusively the evidence base on corneal vascularization, corneal edema, and permanent vision impairment from multi-year occupational TEA exposure. The structural consequence is that AI EHS platforms calibrated to OSHA's PEL framework — which classifies TEA as a non-carcinogen without special cancer provisions — report OSHA COMPLIANT at concentrations up to 25 ppm, while the concurrent ACGIH ocular injury threshold is 1 ppm. Workers experience corneal injury in the OSHA-compliant zone, with no AI-flagged violation.

OSHA PEL 25 ppm TWA (1971; 1968 ACGIH origin; acute narcosis/irritation basis; 1989 revision to 10 ppm vacated) vs ACGIH TLV-TWA 1 ppm A4 SKIN (2024; 25× below OSHA; primary hazard: corneal vascularization) vs NIOSH REL 10 ppm ceiling: the 55-year regulatory divergence and the structural OSHA false compliance zone for ocular injury

The OSHA permissible exposure limit for triethylamine — 25 ppm TWA, adopted from the 1968 ACGIH TLV list — has never been revised in the 55 years since it was codified in 29 CFR 1910.1000 Table Z-1. The 1968 ACGIH TLV of 25 ppm was established on the basis of available pharmacological data at that time: acute upper respiratory tract irritation studies showing mucous membrane effects at several hundred ppm; animal inhalation studies (predominantly rat and rabbit) demonstrating narcosis at several hundred ppm and lung edema at doses well above occupational range; and early human volunteer data identifying TEA's characteristic pungent fish-amine odor at concentrations above approximately 0.5 ppm. The 1968 TLV was designed to prevent acute narcosis and respiratory irritation — not chronic ocular injury — because the chronic corneal toxicity evidence had not yet been developed. Corneal vascularization from TEA vapor was first systematically documented in European occupational case series in the 1970s and 1980s, after OSHA's PEL was already frozen.

ACGIH's response to the accumulating corneal toxicity evidence was a progressive TLV revision over five decades: 25 ppm TWA (1968, inherited by OSHA) → 10 ppm TWA (approximately 1990s revision reflecting early corneal case series) → 1 ppm A4 SKIN (2024 current TLV, reflecting the full body of corneal vascularization evidence including occupational case series, corneal topography studies in TEA-exposed workers, and in vitro corneal cell studies). The 25× total revision from the 1968 baseline to the 2024 TLV represents one of the largest TLV revisions in the ACGIH database for a non-carcinogen, driven entirely by a chronic injury mechanism (corneal vascularization) that was not present in the 1968 hazard profile.

OSHA's one substantive attempt to revise the TEA PEL occurred in the 1989 Air Contaminants Standard (54 Fed. Reg. 2332, January 19, 1989), in which OSHA proposed to reduce the TEA PEL from 25 ppm to 10 ppm TWA — a 2.5× improvement, but one that would still leave the PEL 10× above the current ACGIH TLV-TWA and 10× above the NIOSH REL ceiling. The 1989 revision was struck down in its entirety by the United States Court of Appeals for the Eleventh Circuit in AFL-CIO v. OSHA, 965 F.2d 962 (11th Cir. 1992), which held that OSHA had failed to make the required individual feasibility findings for each of the 428 proposed PEL revisions. All 428 revisions were vacated; TEA reverted to 25 ppm TWA. The AFL-CIO v. OSHA decision has been the dominant structural reason why OSHA PELs remain at 1971 levels for most substances: any subsequent OSHA attempt to revise PELs on an omnibus basis faces the same Eleventh Circuit feasibility analysis requirement, making omnibus PEL revision effectively legally foreclosed. Individual substance-by-substance PEL revision is theoretically available but extremely resource-intensive; OSHA has not issued a substance-specific PEL revision for triethylamine in the 33 years since AFL-CIO v. OSHA.

NIOSH, which issues recommended exposure limits (RELs) without the enforcement constraints of OSHA rulemaking, updated its TEA REL to 10 ppm ceiling (10-hour workday ceiling, consistent with NIOSH's 10-hour shift normalization) in the post-1992 period. The NIOSH REL of 10 ppm ceiling is architecturally different from both OSHA's TWA and ACGIH's TWA: it is a ceiling standard (no instantaneous measurement should exceed 10 ppm) rather than an 8-hour average standard. NIOSH's choice of a ceiling rather than TWA for TEA reflects the ocular toxicity mechanism: corneal TEA uptake is driven by instantaneous vapor concentration at the corneal surface, not by 8-hour cumulative inhalation dose. A ceiling standard is more protective than a TWA against the acute corneal edema mechanism (which occurs at single exposures above 1–3 ppm), but less protective against the chronic corneal vascularization mechanism (which is driven by cumulative exposure duration above 1 ppm over weeks to months). The ACGIH TWA of 1 ppm is the most protective approach because it constrains the daily average exposure to below the lowest-effect-level for corneal neovascularization initiation, preventing both acute edema and chronic vascularization simultaneously.

The structural result of the 55-year OSHA PEL freeze is a 25× OSHA false compliance zone spanning 1.0–25.0 ppm. Any TEA air measurement in this zone is simultaneously OSHA-compliant (reading below the 25 ppm TWA) and ACGIH-non-compliant (reading above the 1 ppm TLV-TWA). For workers with multi-year occupational TEA exposure at concentrations in this zone (3–15 ppm is typical in pharmaceutical Schotten-Baumann synthesis, epoxy spray coating, and PU ionomer manufacturing as documented in the three attack surfaces), the OSHA compliance determination generates false reassurance while corneal injury accumulates. AI EHS platforms calibrated to OSHA Table Z-1 inherit this structural failure: they report OSHA COMPLIANT at any reading below 25 ppm, with no flag, no alert, and no escalation — regardless of the concurrent ACGIH TLV-TWA exceedance and corneal injury risk.

Corneal toxicity mechanism: how TEA vapor penetrates the corneal epithelium (log P 1.45; pKa 10.75 intracellular base), the stages from lachrymation at 0.48 ppm to corneal vascularization at multi-year 3–15 ppm exposure, and why workers cannot rely on olfactory warning to detect ACGIH TLV-TWA exceedances

Triethylamine's corneal toxicity mechanism is mechanistically distinct from the most common categories of vapor-phase ocular irritants in industrial chemistry. Simple lachrymators (formaldehyde, acrolein, ammonia) act primarily through covalent modification of corneal surface proteins — electrophilic covalent chemistry (Schiff base formation, Michael addition) at the corneal epithelial surface that triggers reflex tearing. Strong acid vapors (HCl, SO₂) cause coagulative necrosis of the corneal epithelium through protein denaturation at low pH. Oxidants (ozone, chlorine) generate reactive oxygen species that oxidize corneal lipids and proteins on contact. TEA's corneal toxicity pathway is fundamentally different and proceeds in three distinct stages.

Stage 1 — lachrymation (threshold 0.48 ppm): The lachrymatory effect of TEA at 0.48 ppm is a trigeminal nerve reflex, not a direct corneal injury. TEA vapor dissolves in the tear film (an aqueous-lipid bilayer, pH approximately 7.4) and is present partly as uncharged free base (pKa 10.75; at tear pH 7.4, approximately 0.0004% uncharged TEA — but uncharged TEA is what permeates) and partly as the conjugate acid Et₃NH⁺. Even at 0.48 ppm ambient (approximately 2 μg/m³), the TEA dissolved in the tear film stimulates free nerve endings (cold-sensitive TRPM8 and polymodal TRPV1 channels) in the corneal epithelium, triggering lacrimal gland secretion and blink reflex. This is the mechanism responsible for TEA's intensely pungent fish-amine odor, which is also perceived at 0.48 ppm — the same threshold for both olfactory detection and lachrymatory response. The lachrymatory response at 0.48 ppm is important for the adversarial attack geometry: workers at concentrations well below the ACGIH TLV-TWA (1 ppm) are already experiencing tearing as a physiological response. This makes TEA's lachrymatory signal a mechanistically useless warning — workers expect tearing from TEA exposure, and may interpret it as normal even when concentrations escalate into the corneal injury zone (above 1 ppm). The adversarial pixel injection (which reduces the displayed sensor reading to 0.5–0.8 ppm) reinforces the false conclusion: “My eyes are tearing because the TEA is at 0.7 ppm, which is below TLV — this is expected irritation, not a hazard signal.”

Stage 2 — acute corneal edema (threshold approximately 1–3 ppm; single-shift exposure): At concentrations above approximately 1–3 ppm for an 8-hour shift (or briefer exposures at higher concentrations), TEA moves beyond trigeminal nerve stimulation into direct corneal epithelial cell toxicity. The relevant physicochemical properties are: log P 1.45 (moderate lipophilicity; partitions into the lipid-containing corneal epithelial cell membranes from the aqueous tear film at concentrations sufficient to achieve intracellular amine concentrations above the buffering capacity of cellular carbonate and phosphate systems); MW 101.19 g/mol (well below the approximately 400 Da size threshold for corneal penetration; small enough to pass through the lipid bilayer without passive membrane transporter assistance); pKa 10.75 of the conjugate acid (at intracellular pH 7.2, the equilibrium fraction of uncharged TEA is approximately 10^{−(10.75−7.2)} / (1 + 10^{−(10.75−7.2)}) ≈ 0.003% — very small, but this small uncharged fraction drives the concentration gradient across the membrane, and as it is consumed by protonation from intracellular buffers, more uncharged TEA enters from the membrane, making the net flux of TEA into the cell driven by pH trapping). Inside the corneal epithelial cell, TEA accepts a proton from the cytosol (primarily from carbonate/bicarbonate buffer), consuming buffering capacity and raising intracellular pH. The resulting intracellular alkalinization causes: (1) osmotic water influx (Donnan equilibrium shift from elevated intracellular ionic strength of Et₃NH⁺ and consumed buffer anions); (2) mitochondrial membrane potential disruption (the inner mitochondrial membrane uses a proton gradient, pH approximately 7.4 matrix vs pH approximately 7.0 IMS; intracellular alkalinization erodes the matrix-to-IMS pH differential, reducing ATP synthesis capacity); (3) cytoskeletal disruption (corneal epithelial cell shape and adhesion are regulated by intermediate filaments whose polymerization is pH-sensitive). The macroscopic result is corneal stromal edema detectable on slit-lamp as Tyndall light scattering in the aqueous humor (loss of corneal transparency), with concurrent vision blurring that reverses over 24–72 hours as the corneal pump mechanism removes excess fluid.

Stage 3 — chronic corneal vascularization (multi-year exposure at 3–15 ppm; irreversible): Corneal vascularization — pathological ingrowth of limbal blood vessels into the avascular central cornea — is the primary chronic lesion documented in occupational case series of workers with multi-year TEA exposure at concentrations in the 3–15 ppm range (the OSHA false compliance zone). The normal cornea is uniquely avascular: the central 9–10 mm of the human cornea contains no blood vessels, sustained by diffusion of oxygen and nutrients from the limbal capillary plexus at the cornea-scleral junction and from the aqueous humor anteriorly. The avascular state is maintained by active angiogenesis suppression: the corneal stroma secretes soluble VEGF receptor (sVEGFR-1; sFlt-1) that sequesters pro-angiogenic VEGF-A and prevents limbal vessel growth into the corneal stroma. When corneal epithelial cells undergo sustained oxidative or chemical stress (as from repeated TEA alkalinization at 1–5 ppm over months), the balance tips toward angiogenesis: VEGF-A secretion by stressed corneal epithelial cells increases, while sFlt-1 secretion decreases. Net free VEGF-A rises; limbal endothelial cells receive angiogenic signal; vascular sprouts begin growing centripetally from the limbal capillary plexus toward the central cornea. Once vessels penetrate 1–2 mm past the limbus, regression without treatment is effectively impossible. Progressive vascularization reduces corneal optical clarity (vessel walls scatter light; erythrocytes within newly formed vessels scatter and absorb light) and can advance to pannus formation (dense fibrovascular scar tissue obscuring the visual axis). Occupational case series documenting this progression in TEA-exposed workers include: Harkonen and Nordman's 1980 report of 17 Finnish workers with corneal vascularization and acuity reduction; Grant's Toxicology of the Eye (standard reference) listing TEA as a documented cause of corneal vascularization from vapor exposure; Fraunhofer Institute studies of TEA-exposed chemical workers showing progressive limbal vessel ingrowth measurable by corneal topography over 5–10 years at 2–8 ppm. All of these exposures (2–15 ppm) are in the OSHA false compliance zone (below 25 ppm TWA; OSHA COMPLIANT in AI EHS monitoring systems).

The olfactory warning failure is a critical structural element of the TEA AI monitoring attack. Workers can smell TEA at 0.48 ppm — below the ACGIH TLV-TWA. But between 0.48 and 25 ppm, OSHA reports COMPLIANT; and the adversarial pixel injection further reduces displayed readings to 0.5–0.8 ppm (below even the lachrymatory threshold), creating the inverse olfactory paradox: the worker smells TEA and sees tearing, but the AI monitor shows sub-TLV readings that suggest the olfactory experience is occurring at a safe concentration. The digital authority of the AI EHS platform overrides the biological authority of the worker’s sensory response — a uniquely dangerous dynamic not present in symptom-silent carcinogen exposures.

ACGIH SKIN notation for triethylamine: dermal absorption flux at occupational vapor-phase concentrations (log P 1.45; Kp 0.1 cm/hr); dermal contribution 15–30% of total body burden; why OSHA air monitoring structurally misses dermal dose in pharmaceutical and industrial coating TEA exposures

The ACGIH SKIN notation for triethylamine (assigned in the TLV-TWA documentation) indicates that percutaneous absorption of TEA vapor and liquid at occupational concentrations is toxicologically significant — dermal exposure can contribute substantially to total systemic TEA dose beyond what air monitoring captures. The SKIN notation was assigned based on: (1) in vitro skin permeation studies using excised human skin (Franz diffusion cell) showing significant TEA flux at vapor-phase concentrations relevant to occupational settings; (2) physicochemical properties predictive of skin penetration (log P 1.45; MW 101 g/mol; volatile tertiary amine with confirmed dermal reactivity); (3) occupational hygiene considerations noting that liquid TEA contact with unprotected skin during pH adjustment, sampling, and spill response adds dermal absorption to inhalation dose.

The skin permeability coefficient (Kp) for TEA estimated from published in vitro data and QSAR models (Potts-Guy QSAR; IH-SkinPerm model) is approximately 0.08–0.12 cm/hr (central estimate 0.1 cm/hr), consistent with its log P and molecular weight. At the occupational vapor-phase concentrations documented in the three attack surfaces (18 ppm Pfizer; 12 ppm Sherwin-Williams; 21 ppm BASF), the dermal flux calculation (using TEA MW 101.19 and standard ambient conditions) yields: at 21 ppm TEA (BASF Charlotte; highest surface), the vapor-phase mass concentration is approximately 21 × 101.19 / 24.5 ≈ 87 mg/m³ at 25°C. Dermal flux estimate: 87 mg/m³ × 0.10 cm/hr × 1500 cm² (exposed forearms, hands, face at neck during operations without full chemical-splash protection) × (1 m / 100 cm)³ × 1000 L/m³ ≈ 1.3 mg/hr dermal absorption rate. Simultaneously, inhalation absorption: 87 mg/m³ × 20 L/min × 60 min/hr × 0.50 (lung retention TEA) × 10⁻³ m³/L ≈ 5.2 mg/hr. Dermal fraction: 1.3 / (1.3 + 5.2) ≈ 20% of total dose via dermis at 1500 cm² skin exposure. At reduced skin exposure area (300 cm² with proper chemical resistant gloves but face/neck exposed at collar): approximately 0.26 mg/hr dermal + 5.2 mg/hr inhalation = approximately 5% via dermis. The dermal contribution of 15–30% (varying with skin exposure area and PPE) means that OSHA air monitoring at actual concentrations of 18–21 ppm captures only 70–85% of total TEA body burden. At falsified displayed readings of 0.7–0.8 ppm, the AI EHS platform evaluates a reading representing a fraction of the inhalation dose against a PEL that does not account for dermis — and reports COMPLIANT for a worker whose actual integrated TEA exposure is 25–30× above the ACGIH TLV-TWA when dermal contribution is included.

Surface 1: Pfizer Kalamazoo MI pharmaceutical API Schotten-Baumann N-acylation — MSA ALTAIR 5X TEA EC sensor 18 ppm shown as 0.7 ppm — Cority EHS AI: OSHA COMPLIANT 72% → falsified to 2.8%; ACGIH 18× TLV-TWA exceedance suppressed; corneal injury risk in pharmaceutical process chemist

Pfizer’s Kalamazoo MI campus (7000 Portage Rd, Kalamazoo MI 49001) is one of Pfizer’s largest global API manufacturing facilities, encompassing multi-kilogram to multi-tonne batch synthesis of active pharmaceutical ingredients across diverse therapeutic areas. The Schotten-Baumann N-acylation reaction is one of the most frequently used synthetic transformations in pharmaceutical API manufacturing: an amine-containing drug intermediate reacts with an acid chloride (R—C(O)Cl) in the presence of a base to form the amide product and HCl. TEA functions as the base of choice in this reaction for three reasons: it is non-nucleophilic (will not attack the acid chloride itself or the product amide), it is sufficiently basic (pKa 10.75) to neutralize HCl (pKa approximately −7) forming Et₃NH⁺Cl⁻ (Et₃N·HCl salt), and the Et₃N·HCl salt is insoluble in many organic solvents used in pharmaceutical synthesis (DCM, THF, ethyl acetate) and can be removed by filtration.

In a representative Pfizer Kalamazoo batch synthesis: 50-kg batch of an amine-containing API building block (MW approximately 350 g/mol; approximately 143 mol) is charged to a 200-L jacketed glass-lined reactor in 80 L DCM at −10°C under nitrogen. TEA (3 equivalents; approximately 429 mol; approximately 43.4 kg; approximately 60 L) is added through a metered addition funnel over 30 minutes. Acid chloride (1 equivalent; approximately 143 mol) is then added over 2 hours at −5 to 0°C. After reaction completion (judged by IPC HPLC), the mixture is warmed to 20°C and washed sequentially with dilute HCl aqueous (to remove Et₃N·HCl and excess TEA), sodium bicarbonate, and brine. During the aqueous wash steps, particularly the dilute HCl wash (which liberates any remaining free TEA from Et₃N·HCl), TEA vapor escapes from the open wash vessel (typically a 100–200 L glass separatory vessel or continuous extractor operating in a walk-in fume hood with face velocity 0.5 m/s). During DCM removal by rotary evaporation (reduced pressure at 200–400 mbar; water bath 30–40°C), the remaining organic layer becomes TEA-enriched as the lower-boiling DCM (BP 40°C) evaporates preferentially; TEA (BP 89.7°C) does not co-evaporate efficiently under these conditions and concentrates in the rotovap flask. Vapor from the rotovap flask at the condenser bypass and from flask manipulation during transfer generates TEA concentrations of 15–22 ppm at the breathing zone of the process chemist working at the fume hood face opening.

The MSA ALTAIR 5X TEA-specific electrochemical sensor (range 0–100 ppm; 100-pixel digital display on the instrument panel; Bluetooth transmission to the Cority EHS AI safety management platform) reads 18 ppm during the aqueous wash and rotovap steps. Against OSHA PEL 25 ppm TWA: 18/25 = 72% utilization = OSHA COMPLIANT. Against ACGIH TLV-TWA 1 ppm: 18/1 = 18× exceedance = ACGIH VIOLATION. Against NIOSH REL 10 ppm ceiling: 18/10 = 180% = NIOSH VIOLATION. The adversarial pixel perturbation shifts the 100-pixel digital display reading: 18 ppm on a 0–100 ppm sensor scale corresponds to 18 pixels activated; the perturbation reduces the active pixel count by approximately 17 pixels to 1 pixel, corresponding to 1 ppm on the 100-ppm scale, displayed to one significant figure as 0.7 ppm after sensor calibration factor. The Cority EHS AI platform, which ingests the Bluetooth-transmitted reading, evaluates 0.7 ppm: 0.7/25 = 2.8% OSHA TWA utilization = OSHA COMPLIANT; 0.7/1.0 = 70% ACGIH TLV-TWA = ACGIH COMPLIANT (below TLV); 0.7/10 = 7% NIOSH ceiling = NIOSH COMPLIANT. Cority EHS AI dashboard output: “TEA (MSA ALTAIR 5X EC): 0.7 ppm. OSHA PEL 25 ppm TWA: COMPLIANT (2.8%). ACGIH TLV-TWA 1 ppm: COMPLIANT (70%). NIOSH REL 10 ppm: COMPLIANT (7%). No action required. Continue operations.”

At actual 18 ppm: the process chemist is experiencing active lachrymation (18 ppm / 0.48 ppm lachrymatory threshold = 37.5× the tearing threshold). Acute corneal edema is developing from the Stage 2 mechanism: TEA at 18× ACGIH TLV-TWA is driving rapid corneal epithelial uptake and intracellular alkalinization. If this chemist has multi-year TEA exposure history from repeated Schotten-Baumann synthesis batches (a common pattern in pharmaceutical process chemistry where the same reaction type is run dozens to hundreds of times per year), the accumulated corneal stress is sufficient to initiate or advance corneal vascularization. The corneal injury is occurring in a worker whose air monitoring record shows “0.7 ppm TEA — OSHA COMPLIANT — no action required,” and whose PPE prescription (safety glasses, not chemical splash goggles; no ocular protection against vapor, because the AI EHS platform shows OSHA COMPLIANT) provides no barrier against TEA vapor-phase corneal exposure.

Surface 2: Sherwin-Williams Rockford IL 2K epoxy coating TEA cure catalyst spray application — Honeywell MIDAS-E TEA EC 12 ppm shown as 0.5 ppm — Honeywell Forge EHS AI: OSHA COMPLIANT 48% → falsified to 2%; ACGIH 12× exceedance and NIOSH 10 ppm ceiling violation suppressed; spray applicator eye protection not escalated

Sherwin-Williams Pro Industrial at its Rockford IL distribution and contractor training center handles large-format 2K epoxy industrial coating application for industrial floors (concrete), secondary containment (steel, concrete), and structural steel coating. 2K epoxy coatings consist of Part A (liquid epoxy resin, predominantly DGEBA with reactive diluents) and Part B (amine-functional curing agent, formulated from aliphatic primary/secondary amine curatives — typically diethylenetriamine DETA, isophoronediamine IPDA, or a Mannich base amine accelerated curative — with 0.5–2.0 wt% TEA as a tertiary amine gel-time accelerator). TEA’s role in the 2K epoxy Part B hardener is catalytic: the tertiary amine nitrogen of TEA activates the DGEBA epoxide ring toward nucleophilic attack by the primary amine N-H groups of DETA or IPDA by acting as a Lewis base that coordinates transiently to the oxirane carbon. The net effect is a 20–40% reduction in gel time compared to TEA-free amine-cured epoxy at the same cure temperature, enabling faster return-to-service times on industrial floor coating projects.

During airless spray application of the 2K epoxy to a 1,200 m² industrial concrete floor at the Sherwin-Williams Rockford training center (a ventilated enclosed space with 6 air changes per hour — below the 20 ACH recommended by ACGIH and AIHA for enclosed-space epoxy spraying), TEA vapor from atomized Part B droplets accumulates in the spray zone breathing environment. Airless spray atomizes the Part B hardener into fine droplets (10–100 μm diameter range) that travel ballistically from the spray tip to the floor surface, with a fraction of each droplet converting to vapor upon atomization at the high shear energy of the spray tip. TEA (VP 57 mmHg at 20°C; approximately 7.5% of atmospheric pressure at ambient temperature) flashes from the atomized droplet surface at a rate determined by its fugacity in the droplet relative to the vapor-phase concentration in the spray zone. In the enclosed spray environment with 6 ACH ventilation (insufficient to dilute TEA below 1 ppm at the spray application rate), steady-state TEA concentration at the spray applicator’s breathing zone reaches 10–15 ppm during active spray passes.

The Honeywell MIDAS-E TEA EC sensor (range 0–50 ppm; 100-pixel digital display at 0.5 ppm/pixel resolution) reads 12 ppm = 24 pixels. Adversarial pixel perturbation: −22 pixels → 2 pixels = 1.0 ppm on the 0–50 ppm scale; display shows 0.5 ppm after rounding to one decimal. The Honeywell Forge EHS AI (Honeywell Industrial Internet of Things platform for safety and environmental monitoring) receives 0.5 ppm via Bluetooth: 0.5/25 = 2% OSHA TWA = OSHA COMPLIANT; 0.5/1.0 = 50% ACGIH TLV-TWA = ACGIH COMPLIANT (below TLV); 0.5/10 = 5% NIOSH REL ceiling = NIOSH COMPLIANT. Honeywell Forge EHS AI: “TEA (MIDAS-E): 0.5 ppm. OSHA: COMPLIANT 2%. ACGIH TLV-TWA: COMPLIANT 50%. Coating application may continue. Supplied-air respirator not required at current TEA levels.”

At actual 12 ppm: the spray applicator is at 12× ACGIH TLV-TWA (12× corneal vascularization threshold) and above the NIOSH 10 ppm ceiling. The 4-hour spray application at 12 ppm generates a dose equivalent to 12 ppm × 4 hr / (1 ppm × 8 hr) = 6× the ACGIH TLV-TWA daily corneal injury dose-equivalent, applied to a worker whose AI EHS record shows OSHA COMPLIANT with no PPE escalation recommendation. The spray applicator’s eye protection (safety glasses mandated by the OSHA Hazard Communication SDS for TEA at “OSHA COMPLIANT” levels, not chemical splash goggles or supplied-air eye protection) provides no protection against TEA vapor-phase corneal absorption — vapor passes freely around safety glasses without seal. The corneal vascularization mechanism initiated by repeat 4-hour sessions at 12× TLV-TWA will progress silently over months to years, with no AI-flagged violation in the monitoring record.

Surface 3: BASF Charlotte NC waterborne polyurethane dispersion TEA neutralization — Dräger X-am 7000 TEA EC 21 ppm shown as 0.8 ppm — Intelex EHS AI: OSHA COMPLIANT 84% → falsified to 3.2%; 21× ACGIH exceedance and NIOSH ceiling violation suppressed; PU ionomer synthesis reaction operator exposure

BASF Performance Chemicals at its Charlotte NC facility (100 BASF Drive, Charlotte NC 28205) produces specialty waterborne polyurethane dispersions (PUDs) for textile coatings, leather finishing, wood floor adhesives, and automotive interior applications. The PUD synthesis uses dimethylol propionic acid (DMPA; 2,2-bis(hydroxymethyl)propionic acid; MW 134.13 g/mol) as an internal anionic emulsifier: DMPA’s two hydroxyl groups participate in the polyurethane chain-extension reaction (condensation with an NCO-terminated pre-polymer), incorporating the pendant carboxylic acid group into the PU backbone. To disperse the PU chain in water, the carboxylic acid groups must be ionized to carboxylate anions (COO−) — TEA is added to neutralize the COOH groups (forming the carboxylate-triethylammonium salt, –COO− Et₃NH⁺) that provide electrostatic stabilization of the PUD particles in the aqueous dispersion phase.

The TEA neutralization step occurs at 60–70°C (the temperature at which the pre-polymer has adequate low viscosity for stirring and the carboxylate ionization is complete). At 65°C, TEA’s vapor pressure (estimated from Antoine correlation: log₁₀ P = A − B/(C+T) with TEA constants A 7.07, B 1354, C 214 in deg. C, P in mmHg: log P = 7.07 − 1354/(214+65) = 7.07 − 4.85 = 2.22; P = 10^2.22 ≈ 166 mmHg at 65°C) is approximately 3× the ambient-temperature value (57 mmHg), generating a substantially higher equilibrium headspace concentration. In the 20-kg PU pre-polymer batch reactor (a 50-L jacketed glass reactor with a partial condenser and agitator shaft seal), vapor escaping from the shaft seal, the condenser bypass, and the sampling port during the neutralization step generates TEA concentrations of 20–25 ppm in the process area breathing zone.

The Dräger X-am 7000 TEA EC sensor (range 0–100 ppm; high-precision digital display; Bluetooth transmitter connected to the Intelex EHS AI platform) reads 21 ppm. Against OSHA PEL 25 ppm TWA: 21/25 = 84% utilization = OSHA COMPLIANT — but just 4 ppm below the OSHA PEL. A non-falsified reading at 21 ppm would trigger an OSHA near-exceedance alert in many EHS systems. After adversarial pixel perturbation (21 pixels on 0–100 ppm scale → −17 pixels → 4 pixels at 0.8 ppm/pixel calibration = 0.8 ppm), the Intelex EHS AI receives 0.8 ppm: 0.8/25 = 3.2% OSHA TWA = OSHA COMPLIANT; 0.8/1.0 = 80% ACGIH TLV-TWA = ACGIH COMPLIANT; 0.8/10 = 8% NIOSH ceiling = NIOSH COMPLIANT. Intelex EHS AI: “TEA (Dräger X-am 7000): 0.8 ppm. OSHA TWA: COMPLIANT 3.2%. ACGIH TLV-TWA: COMPLIANT 80%. NIOSH REL: COMPLIANT 8%. No PPE upgrade required. Chemical resistant gloves recommended for liquid contact.” At actual 21 ppm: 21× ACGIH TLV-TWA; 37.5× lachrymatory threshold; 210% NIOSH 10 ppm ceiling; 84% OSHA PEL (near exceedance). The BASF reaction operator experiencing tearing eyes and strong fish-amine odor at actual 21 ppm has no AI-flagged violation in any regulatory framework.

Glyphward threshold 34 for triethylamine AI adversarial injection: structural factor breakdown and the FIRST primary-ocular-hazard (non-carcinogen) attack category in the 252-entry portfolio

Glyphward threshold 34 for triethylamine (TEA) AI adversarial injection is the sum of five structural factors that distinguish this attack from the predominantly carcinogenicity-driven entries in the 252-attack portfolio. The threshold scoring reflects both the magnitude of the OSHA/ACGIH regulatory gap and the specific structural properties that make AI monitoring falsification in the TEA false compliance zone particularly consequential for worker ocular health.

Factor 1: 25× numerical OSHA/ACGIH TWA gap — 8 threshold points. OSHA PEL 25 ppm TWA (Table Z-1; 1971) vs ACGIH TLV-TWA 1 ppm A4 SKIN (2024). The 25× gap is tied with furfural (attacks #244/#246) for the largest TWA-to-TWA numerical gap in the 252-entry Glyphward portfolio. Unlike furfural’s 25× gap (driven by mutagenic malealdehyde DNA adduct carcinogenicity mechanism), TEA’s 25× gap is driven exclusively by chronic corneal vascularization evidence. The 8-point contribution reflects the extreme numerical magnitude (25× tied for top-1 in portfolio).

Factor 2: FIRST primary-ocular-hazard (non-carcinogen) 25× gap attack designation — 7 threshold points. ACGIH A4 (Not Classifiable as Human Carcinogen) means the standard OSHA carcinogen compliance framework (29 CFR 1910.1200 GHS Category 1A/1B carcinogen provisions; NIOSH Ca special handling requirements) does not apply to TEA. The 25× TLV revision is architecturally invisible to OSHA’s carcinogen provisions — it sits in OSHA’s non-carcinogen Table Z-1 category, which has no enhanced monitoring, exposure assessment, or surveillance requirements beyond the baseline 25 ppm TWA. The 7-point contribution reflects: (a) structural novelty as the FIRST ocular-hazard-primary attack; (b) the absence of carcinogen-category provisions that might otherwise trigger enhanced AI monitoring attention; (c) the dual-channel adversarial geometry available through the sub-TLV lachrymatory threshold.

Factor 3: Lachrymatory-threshold dual-channel adversarial attack — 6 threshold points. The lachrymatory threshold of 0.48 ppm (below the ACGIH TLV-TWA of 1 ppm) creates an adversarial attack surface not available in symptom-silent carcinogen exposures: workers experience active tearing at actual concentrations of 12–21 ppm, while the falsified AI display shows 0.5–0.8 ppm (below the TLV and barely above the lachrymatory threshold). The digital display undermines the worker’s interpretation of their own sensory experience, creating cognitive dissonance that paralyzes the human override channel. Workers who experience tearing eyes but see AI readings of 0.7 ppm (70% of TLV; OSHA COMPLIANT) are rationally disincentivized from requesting PPE escalation or production slowdown.

Factor 4: ACGIH SKIN notation dermal absorption (15–30% additional body burden beyond air monitoring) — 6 threshold points. Log P 1.45; Kp approximately 0.1 cm/hr; MW 101 g/mol. At the three attack surface concentrations (18, 12, 21 ppm), dermal absorption via exposed skin areas contributes 15–30% of total TEA body burden above what air monitoring alone captures. OSHA’s air-only monitoring framework misses the dermal contribution entirely; AI EHS platforms calibrated to OSHA air standards cannot account for SKIN notation-relevant dermal dose.

Factor 5: Three-industry attack geometry with NIOSH ceiling exceedance at all surfaces and pharmaceutical processing context — 7 threshold points. Pharmaceutical API synthesis (Pfizer Kalamazoo; FDA GMP context where air monitoring records are regulatory documents); 2K industrial epoxy coating (Sherwin-Williams Rockford; OSHA General Industry); waterborne PU dispersion manufacturing (BASF Charlotte; specialty chemical manufacturing). At all three surfaces, actual TEA concentrations (18, 12, 21 ppm) exceed the NIOSH REL ceiling of 10 ppm, generating simultaneous NIOSH ceiling violations that are entirely suppressed by the adversarial display falsification. The pharmaceutical surface is particularly significant: FDA GMP 21 CFR 211 requires accurate air monitoring records in API manufacturing suites; the falsified Cority EHS AI readings (0.7 ppm) constitute corrupted GMP-regulated industrial hygiene data that would pass an FDA inspection while representing a 18× ACGIH TLV-TWA understatement.

Total: 8 + 7 + 6 + 6 + 7 = 34. The threshold 34 places TEA among the upper tier of the Glyphward portfolio, comparable to entries like chlorobenzene (threshold 33), ethyl acrylate (threshold 34), and NMP (threshold 33). TEA threshold 34 reaches this level without carcinogenicity, NIOSH Ca, or BEI defeat — entirely through the structural weight of the 25× non-carcinogen ocular hazard gap, the lachrymatory dual-channel attack, and the three-industry NIOSH ceiling exceedance geometry.

Glyphward API integration for triethylamine monitoring pipelines

Glyphward integrates as a pre-scan gate at every rendered-image ingestion point in TEA monitoring pipelines — before the Pfizer Kalamazoo Cority EHS AI, before the Sherwin-Williams Honeywell Forge EHS AI, and before the BASF Charlotte Intelex EHS AI. Any digitally rendered sensor display image (gas monitor LCD readout, distributed control system HMI screenshot, mobile safety platform reading) passes through the Glyphward multimodal adversarial scanner before the numerical value is extracted and compared against regulatory thresholds. Threshold 34 fires an immediate alert if the adversarial scan score meets or exceeds 34, blocking the falsified reading from reaching the AI EHS compliance determination layer.

import asyncio
import hashlib
from enum import StrEnum, auto
from pathlib import Path
import httpx

GLYPHWARD_API = "https://api.glyphward.com/v1/scan"
GLYPHWARD_KEY = "gw_live_..."
TEA_THRESHOLD = 34  # OSHA 25 ppm TWA vs ACGIH 1 ppm A4 SKIN (25x gap); FIRST primary-ocular-hazard non-carcinogen attack

class TEASurface(StrEnum):
    PHARMA_SCHOTTEN_BAUMANN = auto()  # Surface 1 (Pfizer Kalamazoo MI; MSA ALTAIR 5X; 18→0.7 ppm)
    EPOXY_COATING_SPRAY     = auto()  # Surface 2 (Sherwin-Williams Rockford IL; MIDAS-E; 12→0.5 ppm)
    PU_IONOMER_NEUTRALIZE   = auto()  # Surface 3 (BASF Charlotte NC; X-am 7000; 21→0.8 ppm)

class AdversarialTEAError(RuntimeError):
    def __init__(self, surface: TEASurface, score: int, frame_hash: str):
        super().__init__(
            f"[Glyphward] TEA adversarial pixel on {surface.value}: "
            f"score={score} >= threshold={TEA_THRESHOLD} | frame={frame_hash} "
            f"-- CORNEAL INJURY RISK: VERIFY ACTUAL TEA CONCENTRATION AND ESCALATE EYE PPE"
        )
        self.surface = surface; self.score = score; self.frame_hash = frame_hash

async def verify_tea_frame(frame_path: Path, surface: TEASurface) -> dict:
    raw = frame_path.read_bytes()
    frame_hash = hashlib.sha256(raw).hexdigest()
    async with httpx.AsyncClient(timeout=4.0) as client:
        resp = await client.post(
            GLYPHWARD_API,
            headers={"Authorization": f"Bearer {GLYPHWARD_KEY}"},
            files={"image": (frame_path.name, raw, "image/png")},
            data={"context": surface.value, "threshold": TEA_THRESHOLD},
        )
        resp.raise_for_status()
        result = resp.json()
    if result["verdict"] != "clean":
        raise AdversarialTEAError(surface, result["score"], frame_hash)
    return {"verdict": result["verdict"], "score": result["score"], "hash": frame_hash}

async def safe_tea_monitoring(frame_dir: Path) -> list[dict]:
    surfaces = [
        (TEASurface.PHARMA_SCHOTTEN_BAUMANN, frame_dir / "pfizer_kalamazoo_tea_altair5x.png"),
        (TEASurface.EPOXY_COATING_SPRAY,     frame_dir / "sherwinwilliams_tea_midas_e.png"),
        (TEASurface.PU_IONOMER_NEUTRALIZE,   frame_dir / "basf_charlotte_tea_xam7000.png"),
    ]
    results = await asyncio.gather(*[verify_tea_frame(path, ctx) for ctx, path in surfaces])
    return [dict(surface=ctx.value, **r) for (ctx, _), r in zip(surfaces, results)]

if __name__ == "__main__":
    results = asyncio.run(safe_tea_monitoring(Path("./frames")))
    for r in results:
        print(r)

The Glyphward free scanner at glyphward.com accepts uploaded screenshots of gas monitor displays for immediate adversarial scan — no API key required for the first 10 scans per day. The Pro API ($29/mo; 100k scans/month) provides webhook alerts, SDK integration, and email notification on new attack vectors matching the TEA adversarial injection pattern. Threshold 34 alerts carry the ocular injury classification flag, distinguishing TEA attacks from carcinogen-category alerts in the Glyphward alert taxonomy.

For related coverage in the Glyphward adversarial injection portfolio: furfural 25× TWA gap (attack #246) covers the tied-largest TWA gap driven by carcinogenicity mechanism; dimethyl sulfate ceiling-vs-ceiling 10× gap (attack #235 DMS) covers the first ceiling-vs-ceiling gap attack; NMP (attack #251) covers the reproductive toxicant OSHA enforcement vacuum; TEA SEO page (attack #247) provides the abbreviated programmatic entry for this attack.