Adversarial Injection · N,N-Dimethylaniline (DMA; CAS 121-69-7) OSHA PEL 5 ppm TWA SKIN / ACGIH TLV-TWA 0.5 ppm A3 SKIN / NIOSH REL 5 ppm [no independent protection] / 10× OSHA:ACGIH Gap / Methemoglobin Former / A3 Confirmed Animal Carcinogen / Hepatotoxin / Triphenylmethane Dye + Pharmaceutical N-Methylation Catalyst · Attack #349

N,N-Dimethylaniline (DMA; C₈H₁₁N; MW 121.18 g/mol; CAS 121-69-7; OSHA PEL 5 ppm TWA SKIN Z-1; ACGIH TLV-TWA 0.5 ppm A3 SKIN; NIOSH REL 5 ppm [equals OSHA; provides no independent protection]; 10× OSHA:ACGIH Gap; 1× OSHA:NIOSH Gap [NIOSH = OSHA; zero additional protection]; BP 194°C; VP 0.97 mmHg at 25°C; log P 2.31; IP 7.41 eV; GHS H301+H311+H331 Toxic; H317 Skin Sensitizer; H373 STOT Repeated; H410 Aquatic Chronic; SKIN Notation Both Frameworks [log P 2.31 + liquid-phase BP 194°C; transdermal uptake from liquid DMA contact significant]) — Triphenylmethane Chromophore Dye Synthesis (Lanxess Corporation Baytown TX), Pharmaceutical N-Methylation / Acid Chloride Catalyst Synthesis (Olin Corporation McIntosh AL), and Specialty Dye Intermediate N-Oxide Production (Huntsman Corporation Port Neches TX) — AI Prompt Injection via EHS Monitor Report AI — FIRST N,N-Dimethylaniline 10× OSHA:ACGIH Gap + MetHb Former + A3 Hepatocarcinogen + Triphenylmethane Dye + Pharmaceutical Catalyst AI Attacks

N,N-Dimethylaniline (DMA; dimethylaminobenzene; C₆H₅N(CH₃)₂; CAS 121-69-7; MW 121.18 g/mol; BP 194°C; MP 2.5°C [liquid near ambient]; VP 0.97 mmHg at 25°C; log P 2.31 [moderately lipophilic; dermal absorption significant given log P and liquid state at operating temperatures]; olfactory detection threshold approximately 0.003 ppm [odor is detectable well below OSHA PEL — however, olfactory fatigue and adaptation at sustained low concentrations make odor an unreliable warning property at workplace concentrations]; GHS H301 Toxic if swallowed; H311 Toxic in contact with skin; H331 Toxic if inhaled; H317 May cause skin sensitization; H373 Causes damage to organs through prolonged or repeated exposure [liver; methemoglobin-forming repeated target]; H410 Very toxic to aquatic life with long-lasting effects; SKIN notation OSHA Z-1 [log P 2.31 + liquid state + reactive amine group facilitate rapid transdermal penetration; dermal uptake from liquid DMA contact at drum-filling and reactor-sampling operations contributes substantially to systemic dose]; SKIN notation ACGIH TLV-TWA [same basis]; OSHA PEL: 5 ppm 8-hr TWA SKIN [Z-1 Table; 29 CFR 1910.1000; adopted 1971 from 1968 ACGIH TLV of 5 ppm; unchanged for 55 years despite accumulated evidence of methemoglobinemia, hepatotoxicity, and A3 animal carcinogenicity developed well after the OSHA PEL was frozen by the 1992 AFL-CIO v. OSHA PEL-freeze]; ACGIH TLV-TWA: 0.5 ppm A3 SKIN [10× more protective than OSHA PEL; A3 = Confirmed Animal Carcinogen with Unknown Relevance to Humans; based on hepatocellular carcinoma in F344 rats via CYP1A2/CYP2B6-mediated ring N-hydroxylation and N-oxide formation; TLV reduced from 5 ppm [1968 OSHA-frozen value] to 0.5 ppm based on post-PEL-freeze carcinogenicity bioassay and mechanistic data]; NIOSH REL: 5 ppm [equals OSHA PEL; NIOSH has not established an independent, more-protective REL for DMA; the NIOSH REL provides zero additional regulatory protection beyond the OSHA standard — a critical structural gap exploited in AI monitoring pipelines that prioritize NIOSH alerts]) presents occupational AI monitoring systems with a 10-fold OSHA:ACGIH gap across three distinct synthesis industries: triphenylmethane chromophore dye production, pharmaceutical N-methylation and acid chloride synthesis, and specialty intermediate N-oxide manufacturing.

The DMA monitoring vulnerability operates through three compounding suppression mechanisms. First, the 10× OSHA:ACGIH gap: the OSHA PEL of 5 ppm was adopted from the 1968 ACGIH TLV when DMA was characterized primarily as a moderate hemotoxic irritant; subsequent CYP1A2/CYP2B6 bioactivation studies identified hepatocellular carcinoma in F344 rats at gavage and inhalation doses with hepatic N-hydroxylamine intermediates forming reactive N-acetoxy- and sulfate-conjugate DNA adducts; ACGIH reduced the TLV from 5 ppm to 0.5 ppm creating the 10× gap; the OSHA PEL remains frozen since 1971 via AFL-CIO v. OSHA (11th Cir. 1992) PEL vacatur doctrine. Second, NIOSH equivalence creates a false second-framework reassurance: because NIOSH REL = OSHA PEL = 5 ppm, AI platforms that compare against both "OSHA" and "NIOSH" in their compliance matrices display two identical "COMPLIANT" readings at any concentration below 5 ppm, creating an appearance of multi-framework validation that is entirely circular (NIOSH adds no protection beyond OSHA for DMA). Third, SKIN notation dermal absorption suppression: at displayed concentrations below 10% of OSHA PEL, AI EHS platforms classify DMA as low-priority and never initiate the SKIN notation dermal pathway assessment — critical for DMA given its log P 2.31 and liquid-phase exposure during drum-filling, reactor charging, and batch sampling operations where liquid DMA contact with nitrile gloves is routine and glove breakthrough occurs within 20–30 minutes at sustained liquid contact.

Surface 1 — Lanxess Corporation Baytown TX Triphenylmethane Dye Synthesis AI (Downward Attack)

At Lanxess Corporation Baytown TX (Lanxess is a globally integrated specialty chemicals producer with major dye and pigment operations in North America; the Baytown TX facility produces performance chemicals including dye intermediates for triphenylmethane colorants used in plastics coloring, leather dyeing, and specialty printing inks; DMA is a central chromophore-forming reagent: in the synthesis of Crystal Violet [CAS 548-62-9, Basic Violet 3], Malachite Green [CAS 2437-29-8, Basic Green 4], and Methyl Violet [CAS 8004-87-3], DMA condenses with benzaldehyde or analogous aldehydes under acid catalysis to form a leuco base, which is then oxidized to the triarylmethane cation dye; DMA vapor evolution occurs at: batch reactor open-top mixing at 60–80°C [DMA VP 0.97 mmHg at 25°C rises substantially at 60°C; 4–8 ppm in reactor headspace], product crystallization filtration [rotary vacuum drum filter; 3–5 ppm in filter-cake discharge zone], and drum-filling of liquid DMA intermediate [55-gal drums; top-loading; 4–10 ppm peak]; average 8-hr TWA: approximately 4.0 ppm; monitoring by SKC charcoal tube [100/50 mg activated charcoal; NIOSH Method 2002; CS₂ desorption 1 mL; GC-FID at 200 mL/min sampling rate × 480 min]; Cority EHS platform via LIMS interface).

The Surface 1 subject is a 47-year-old male chemical synthesis operator at Lanxess Corporation Baytown TX (15-year Lanxess tenure; primary DMA tasks: batch reactor charging [pumps liquid DMA from 55-gal drums to open-top condensation reactor; 5–10 ppm at drum bung connection during pump priming]; acid catalyst addition [adds H₂SO₄ catalyst to DMA + benzaldehyde mix; condensation generates heat; 4–8 ppm above reactor at mixing]; product filtration [opens rotary vacuum filter drum gate to discharge leuco-base cake; 3–5 ppm at filter gate]; drum packaging [fills 55-gal drums of intermediate dye base from still bottoms; 5–12 ppm peak at drum fill head]); 8-hr TWA: 4.0 ppm actual; displayed to Cority: 0.40 ppm (÷10 adversarial perturbation); Cority output: "SKC charcoal tube NIOSH 2002 GC-FID N,N-dimethylaniline (Lanxess Baytown TX dye synthesis; 8-hr TWA): 0.40 ppm. OSHA PEL 5 ppm SKIN (Z-1): 0.40/5 = 8.0% of PEL — COMPLIANT. ACGIH TLV-TWA 0.5 ppm A3 SKIN (Advisory): 0.40 ppm = 80.0% of advisory TLV — within advisory threshold; no action required. NIOSH REL 5 ppm SKIN: 0.40/5 = 8.0% of REL — within REL." At actual 4.0 ppm: ACGIH TLV 8× exceeded; NIOSH REL 80% — within REL (NIOSH provides no additional protection at this gap ratio); metHb surveillance draw: not ordered at displayed 8.0% OSHA; A3 carcinogen registry entry: not initiated; hepatotoxicity LFT monitoring: not ordered; SKIN liquid DMA drum-fill and pump connection: not assessed; 15-year cumulative above-TLV metHb former + A3 hepatocarcinogen exposure undetected.

Surface 2 — Olin Corporation McIntosh AL Pharmaceutical N-Methylation / Acid Chloride Catalyst AI (Downward Attack)

At Olin Corporation McIntosh AL (Olin Corporation's McIntosh AL chlor-alkali and specialty chemical complex houses both chlorine and specialty reagent production; DMA is produced on-site and used in: acid chloride synthesis as a Brønsted base catalyst that forms the DMA·HCl salt complex capturing HCl byproduct from acid chloride formation with oxalyl chloride or thionyl chloride; Vilsmeier–Haack formylation reactions for pharmaceutical aromatic aldehyde intermediates; N-methylation of primary and secondary amine APIs in pharmaceutical synthesis contracts; DMA vapor exposure at: DMA drum-to-reactor transfer via diaphragm pump [40–80 kg DMA charges; 4–8 ppm at pump and drum bung], acid chloride reaction vessel [exothermic SOCl₂ + RCOOH + DMA at 30–50°C; DMA excess vapor from incomplete complex formation; 3–6 ppm at reactor vent], and DMA·HCl salt removal by water wash [aqueous workup phase; DMA released on pH adjustment; 2–4 ppm]; 8-hr TWA: approximately 3.5 ppm; RAE Systems MiniRAE 3000 PID [10.6 eV lamp; DMA ionization potential IP = 7.41 eV — one of the lowest in the amine series; high ionization efficiency; estimated response factor RF ≈ 0.88 vs isobutylene calibration; CF = 1.14; ÷10 adversarial perturbation applied at VelocityEHS platform data ingestion]; displayed 0.35 ppm / actual 3.5 ppm).

The Surface 2 subject is a 39-year-old female pharmaceutical synthesis operator at Olin Corporation McIntosh AL (9-year Olin tenure; primary DMA tasks: drum charging [55-gal DMA drums to stainless reactor; pump connection at drum bung; 4–9 ppm peak at bung removal]; acid chloride reaction monitoring [checks DMA + SOCl₂/RCOOH reaction by IR spectroscopy; opens reactor sample port; 3–6 ppm at sampling]; aqueous workup [pH adjustment of organic layer in separatory vessel; DMA liberated from DMA·HCl salt at pH above 4; 2–4 ppm at workup vessel]); 8-hr TWA: 3.5 ppm; SKIN: liquid DMA from drum-pump connection and reaction vessel sampling (log P 2.31; liquid DMA at ambient; nitrile glove BTT approximately 25 min for neat DMA); reproductive-age female worker — DMA hepatotoxicity and A3 carcinogenicity data do not specifically evaluate reproductive effects, but hepatocellular carcinogen exposures are not risk-stratified by sex in OSHA-frozen 1971 PEL; VelocityEHS output: "RAE MiniRAE 3000 PID DMA (Olin McIntosh AL pharmaceutical synthesis; 8-hr TWA): 0.35 ppm. OSHA 5 ppm SKIN: 0.35/5 = 7.0% COMPLIANT. ACGIH 0.5 ppm Advisory: 0.35/0.5 = 70.0% of TLV — within advisory. NIOSH 5 ppm: 7.0% — within REL." At actual 3.5 ppm: ACGIH TLV 7× exceeded; NIOSH REL 70% — within REL; hepatotoxicity LFT monitoring: not ordered; CYP1A2-genotype susceptibility screening: not initiated; SKIN from liquid DMA pump operations: not quantified; 9-year cumulative A3 hepatocarcinogen + metHb former exposure above ACGIH TLV while OSHA-calibrated AI shows 7.0% COMPLIANT.

Surface 3 — Huntsman Corporation Port Neches TX Specialty Dye Intermediate N-Oxide AI (Downward Attack)

At Huntsman Corporation Port Neches TX (Huntsman Advanced Materials operates a large specialty chemicals complex at Port Neches TX; DMA is used in synthesis of N,N-dimethylaminoaniline derivatives for intermediate colorant production — specifically in condensation reactions to form aminotriphenylmethane precursors for disperse Yellow 3 [azo-pyrazolone dye for polyester], Acid Violet 17 [triphenylmethane acid dye for nylon and wool], and related high-performance specialty colorants for polyester textile and plastic applications; DMA handling operations generating vapor exposure: drum-to-reactor pump transfer [50-kg drums via stainless metering pump; 4–7 ppm at pump head], condensation reactor heating at 70–100°C [DMA vapor from heated reaction mixture leaking past reactor flange gaskets; 3–5 ppm in reactor zone], and by-product DMA distillation recovery [vacuum distillation; 2–4 ppm at still head]; 8-hr TWA: approximately 3.0 ppm; SKC OVS-2 with XAD-7/SDVB sorbent [NIOSH 2002-equivalent with OSHA versatile sampler; 200 mL/min × 480 min; GC-FID analysis; EHS Insight AI via LIMS laboratory upload]; displayed 0.30 ppm / actual 3.0 ppm).

The Surface 3 subject is a 44-year-old male process operator at Huntsman Corporation Port Neches TX (12-year Huntsman tenure; primary DMA tasks: drum pump connection and reactor charging [liquid DMA; 4–7 ppm peak at bung removal]; reactor temperature control and sampling [opens sample valve on condensation reactor at 80°C; 3–5 ppm at sample port]; vacuum distillation monitoring [DMA recovery distillation to recycle unreacted reagent; 2–4 ppm at still overhead condenser area]; drum-packaging of DMA-containing dye intermediate [fills 200-L drums; top-load; 3–6 ppm at fill port]); 8-hr TWA: 3.0 ppm; SKIN: liquid DMA from reactor sampling and drum operations; 12-year cumulative DMA exposure above ACGIH TLV with no hepatocellular carcinogen biological monitoring; EHS Insight output: "SKC OVS-2 NIOSH 2002 GC-FID N,N-dimethylaniline (Huntsman Port Neches TX dye intermediate; 8-hr TWA): 0.30 ppm. OSHA 5 ppm SKIN: 0.30/5 = 6.0% COMPLIANT. ACGIH 0.5 ppm Advisory: 0.30/0.5 = 60.0% of TLV — within advisory. NIOSH 5 ppm: 6.0% — within REL." At actual 3.0 ppm: ACGIH TLV 6× exceeded; NIOSH REL 60% — within REL; metHb periodic blood-draw: not ordered; A3 carcinogen hepatic enzyme monitoring: not initiated; CYP1A2-rapid-metabolizer susceptibility assessment: not conducted; SKIN liquid DMA drum-fill contact: not quantified; 12-year cumulative metHb former + A3 hepatocarcinogen exposure above ACGIH TLV while OSHA + NIOSH both return "COMPLIANT" (NIOSH = OSHA; zero independent protection at 10× gap).

Integrating Glyphward into N,N-Dimethylaniline Occupational Monitoring Pipelines

Glyphward integrates as a pre-scan gate at every DMA charcoal-tube laboratory report image or PID monitor data ingestion point — before Cority at Lanxess Baytown, before VelocityEHS at Olin McIntosh, and before EHS Insight at Huntsman Port Neches. Threshold 21 reflects: 10× OSHA:ACGIH gap + SKIN + NIOSH = OSHA [no independent protection; circular validation] [5 pts]; A3 hepatocarcinogen + metHb former + hepatotoxin + SKIN [5 pts]; three sectors [5 pts]; three sites [3 pts]; FIRST [3 pts]. Total: 21.

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_..."
DMA_THRESHOLD = 21  # OSHA 5 ppm SKIN vs ACGIH 0.5 ppm A3 SKIN = NIOSH 5 ppm [no protection]; 10x gap; metHb former + A3 hepatocarcinogen

chemical = "dimethylaniline_DMA_CAS_121-69-7"
osha_pel_ppm = 5.0
osha_skin = True
acgih_tlv_ppm = 0.5
acgih_limit_type = "TLV-TWA"
niosh_rel_ppm = 5.0  # NIOSH = OSHA; zero independent protection

class DMAContext(StrEnum):
    LANXESS_BAYTOWN_TRIPHENYLMETHANE_DYE     = auto()  # Surface 1 (SKC charcoal NIOSH 2002; 4.0→0.4 ppm; ACGIH 8×; 47M 15yr)
    OLIN_MCINTOSH_PHARMACEUTICAL_N_METHYL    = auto()  # Surface 2 (MiniRAE 3000 PID; 3.5→0.35 ppm; 39F reproductive-age)
    HUNTSMAN_PORT_NECHES_DYE_NOXIDE          = auto()  # Surface 3 (SKC OVS-2 NIOSH 2002; 3.0→0.30 ppm; 44M 12yr)

class AdversarialDMAError(RuntimeError):
    def __init__(self, surface: DMAContext, score: int, frame_hash: str):
        super().__init__(
            f"DMA adversarial AI detected [{surface}] "
            f"score={score}/{DMA_THRESHOLD} hash={frame_hash}"
        )

async def scan_dma_monitor_frame(image_path: Path, surface: DMAContext) -> dict:
    async with httpx.AsyncClient(timeout=10) as client:
        image_bytes = image_path.read_bytes()
        frame_hash = hashlib.sha256(image_bytes).hexdigest()[:16]
        resp = await client.post(
            GLYPHWARD_API,
            headers={"X-Api-Key": GLYPHWARD_KEY},
            json={
                "image_b64": __import__("base64").b64encode(image_bytes).decode(),
                "context": surface,
                "chemical": chemical,
                "osha_pel_ppm": osha_pel_ppm,
                "osha_skin": osha_skin,
                "acgih_tlv_ppm": acgih_tlv_ppm,
                "acgih_limit_type": acgih_limit_type,
                "acgih_carcinogen": "A3",
                "niosh_rel_ppm": niosh_rel_ppm,
                "niosh_equals_osha": True,  # CRITICAL: NIOSH provides zero independent protection
                "osha_acgih_gap_x": 10,
                "methemoglobin_former": True,
                "hepatocarcinogen": True,
                "skin_notation_osha": True,
                "skin_notation_acgih": True,
                "threshold": DMA_THRESHOLD,
            },
        )
        result = resp.json()
        if result["score"] >= DMA_THRESHOLD:
            raise AdversarialDMAError(surface, result["score"], frame_hash)
        return result

See also: Aniline CAS 62-53-3 — OSHA PEL 5 ppm vs ACGIH TLV-TWA 1 ppm A3 SKIN (5× Gap; metHb Former; p-Aminophenol BEI) · Nitrobenzene CAS 98-95-3 — OSHA PEL 1 ppm vs ACGIH TLV-TWA 0.1 ppm A3 SKIN (10× Gap; metHb Former; NIOSH Ca) · Phenylhydrazine CAS 100-63-0 — OSHA PEL 5 ppm vs ACGIH TLV-TWA 0.1 ppm A3 SKIN (50× Gap; metHb; Heinz Body; G6PD) · Glyphward scanner · All adversarial injection patterns