Adversarial Injection · Carbon Disulfide (CS₂; CAS 75-15-0) OSHA PEL 20 ppm TWA + 30 ppm Ceiling / ACGIH TLV-TWA 1 ppm A3 SKIN BEI TTCA (2-Thioxothiazolidine-4-Carboxylic Acid) / NIOSH REL 1 ppm SKIN / 20× Dual-Convergence Gap / CHD Coronary Heart Disease TLV Basis / CYP2E1 Dithiocarbamate Zn/Cu Chelation / TTCA BEI End-of-Shift Suppression · Attack #329

Carbon Disulfide (CS₂; Carbon Bisulfide; CAS 75-15-0; OSHA PEL 20 ppm TWA + 30 ppm Ceiling; ACGIH TLV-TWA 1 ppm A3 SKIN BEI TTCA ≤1 mg/g Cr End-of-Shift; NIOSH REL 1 ppm SKIN; 20× Dual-Convergence Gap; ACGIH=NIOSH Agreement at 1 ppm; CHD Coronary Heart Disease Primary TLV Endpoint; CYP2E1 → COS + Dithiocarbamate Zn/Cu Chelation; TTCA Urinary Biomarker Suppressed) — Rubber Vulcanization Accelerator Synthesis (Flexsys America LP Akron OH; IS Ventis Pro 5 EC), Sodium Ethyl Xanthate Mineral Flotation Production (Arch Chemicals Memphis TN; SKC Charcoal NIOSH 1600 GC/FID), and Mancozeb EBDC Dithiocarbamate Fungicide Synthesis (Nufarm Americas Denver CO; MSA Altair 5X EC) — OSHA 20 ppm TWA + 30 ppm Ceiling vs ACGIH TLV-TWA 1 ppm A3 SKIN BEI TTCA vs NIOSH REL 1 ppm: AI Prompt Injection via EHS Monitor Report AI — FIRST CS₂ 20× Dual-Convergence Gap + CHD Cardiovascular Endpoint TLV + BEI TTCA Suppression AI Attack

Carbon disulfide (CS₂; carbon bisulfide; CAS 75-15-0; MW 76.14 g/mol; BP 46.2°C [extremely low boiling point — among the lowest of any common industrial solvent; generates substantial vapor at ambient temperatures]; VP 297 mmHg at 20°C [approximately 3.4× more volatile than water at the same temperature; near-ambient boiling point means CS₂ vapor generation is essentially continuous in open process systems]; water solubility 2.9 g/L at 20°C [slightly water-soluble; forms stable emulsions with process water in viscose/xanthate operations]; SKIN notation [significant percutaneous absorption; log P 1.94 enables lipid-membrane permeation; liquid CS₂ contact (xanthate synthesis, rubber accelerator reactor sampling, dithiocarbamate handling) contributes an estimated 20–30% additional systemic dose beyond inhalation — a systematic undercount invisible to air-only monitoring systems]; log P 1.94 [moderate lipophilicity; penetrates CNS lipid barriers and myelin; distributes into cardiovascular tissue]; odor threshold 0.01–0.2 ppm [extraordinarily low — CS₂ has a strong, distinctive, rotting-radish/ether-like odor detectable at sub-ppm concentrations; however, rapid olfactory fatigue at occupational concentrations means the odor warning cannot be relied upon for sustained exposure guidance; odor threshold is substantially below both the ACGIH TLV-TWA (1 ppm) and NIOSH REL (1 ppm) — a scenario where a worker's initial olfactory sensation is present but fatigue abolishes it, creating a false sense of improvement]; NIOSH IDLH 500 ppm; GHS hazard statements: H350 May Cause Cancer; H361 Suspected Reproductive Toxicant (fertility/fetal effects; testicular atrophy at sub-OSHA PEL concentrations in animal studies); H371 CNS Toxicant (narcosis at acute high exposure; chronic neurological effects at occupational levels: parkinsonism, polyneuropathy, cognitive decline); LEL 1.3%; UEL 50.0% [extraordinarily wide flammability range — among the broadest of any industrial chemical; CS₂ fires are exceptionally difficult to control]; autoignition 90°C [spontaneously ignitable at temperatures routinely encountered in steam lines and hot surfaces; CS₂ autoignition on a steam pipe at 100°C is not a theoretical scenario]; OSHA PEL: 20 ppm TWA + 30 ppm ceiling [29 CFR 1910.1000 Table Z-1; 1971 OSHA adoption of 1968 ACGIH TLV; the OSHA PEL was adopted from the ACGIH TLV that existed at the time — meaning the OSHA PEL was once state-of-the-art ACGIH guidance; in the intervening 55 years, ACGIH dramatically revised its TLV downward as cardiovascular epidemiology from viscose rayon workers accumulated; OSHA has never updated the Z-1 limits for CS₂; the dual-limit structure (20 ppm TWA + 30 ppm ceiling) creates an additional AI monitoring architecture complexity — which limit is the binding constraint at any given sampling moment? AI systems calibrated to OSHA Z-1 must navigate both a TWA limit and a ceiling limit, creating four possible compliance states; the OSHA ceiling (30 ppm) is only 50% above the TWA (20 ppm), meaning a ceiling exceedance detection requires only a modest spike above the already-permissive TWA]; ACGIH TLV-TWA: 1 ppm A3 SKIN [2023 current; A3 = Confirmed Animal Carcinogen with Unknown Relevance to Humans; animal carcinogenicity: hepatocellular carcinoma in NTP 2-year rodent bioassay at chronic exposures of 100–400 ppm; SKIN notation: log P 1.94 enables dermal absorption in liquid spill/contact scenarios; BEI: urinary 2-thioxothiazolidine-4-carboxylic acid (TTCA) ≤1 mg/g Cr end-of-shift — TTCA is a specific CS₂ metabolite formed by the reaction of CS₂ with cysteine in the ACGIH BEI urinary biomarker pathway; TTCA is specific to CS₂ and not confounded by other industrial sulfur compounds; 20× below OSHA TWA; ACGIH reduced TLV dramatically over successive review cycles: 20 ppm (1971 parity with OSHA at adoption) → 10 ppm (1980s, first CHD signal from Swedish rayon cohorts) → 1 ppm (current, after Ahlborg 1985 and Tolonen 1976 Finnish rayon cohort CHD mortality data established doubled coronary heart disease mortality at >20 ppm and measurable CHD risk below 20 ppm; ACGIH TLV-TWA of 1 ppm is set specifically to protect against coronary heart disease accelerated atherosclerosis — the first and so far only Glyphward portfolio entry where the primary TLV endpoint is cardiovascular rather than neurotoxicity or carcinogenicity]; NIOSH REL: 1 ppm TWA SKIN [NIOSH Current Intelligence Bulletin and REL documentation; same as ACGIH TLV; this dual-convergence is particularly significant — ACGIH and NIOSH arrived at identical 1 ppm limits via independent epidemiological review of viscose rayon worker cohort data; the convergence is not a coincidence or cross-citation but reflects independent scientific consensus that 1 ppm is the appropriate occupational exposure limit to protect against CS₂-induced coronary heart disease; identical dual-convergence pattern to n-hexane #319 (ACGIH=NIOSH=50 ppm vs OSHA 500 ppm) but with cardiovascular and neurological endpoints rather than peripheral neuropathy]; 20× dual-convergence gap: OSHA 20 ppm TWA / ACGIH 1 ppm A3 SKIN BEI TTCA = NIOSH 1 ppm SKIN; both ACGIH and NIOSH independently at 1 ppm; OSHA frozen at 20 ppm for 55 years; CS₂ mechanism: CYP2E1-mediated oxidative metabolism → COS (carbonyl sulfide; approximately 25% of absorbed CS₂ dose exhaled as COS; COS itself reacts with carbonic anhydrase and hemoglobin via thiocarbamoylation) + reactive dithiocarbamate (DTC) intermediates from CS₂ reaction with primary and secondary amines in serum and tissue (CS₂ + R-NH₂ → R-NH-C(=S)-S⁻ dithiocarbamate); DTC metabolites chelate essential trace metal cofactors — specifically zinc (Zn²⁺) and copper (Cu²⁺) — causing enzyme inhibition cascades: dopamine β-hydroxylase (DBH) inhibition (Cu-dependent enzyme; DBH converts dopamine → norepinephrine in catecholamine synthesis; DBH inhibition by CS₂-derived DTC → elevated dopamine, reduced norepinephrine → altered autonomic tone, contributing to cardiovascular risk and parkinsonian features); TTCA biomarker: CS₂ + cysteine (tissue) → TTCA (2-thioxothiazolidine-4-carboxylic acid) — measured in end-of-shift urine as ACGIH BEI ≤1 mg/g Cr; TTCA is CS₂-specific and analytically clean; cardiovascular mechanism: CS₂ accelerates atherosclerosis via multiple pathways — DTC Zn/Cu chelation disrupts antioxidant enzyme function (Cu/Zn superoxide dismutase inhibition → oxidative stress in vascular endothelium); CS₂ directly modifies apolipoprotein structure; coronary intimal proliferative lesions confirmed histologically in Ahlborg (1985) Swedish viscose rayon cohort autopsy data; Tolonen (1976) Finnish rayon plant workers showed doubled CHD mortality at occupational CS₂ exposures in the 20–50 ppm range; subsequent studies at lower concentrations suggested CHD risk persisted below 20 ppm, motivating ACGIH's reduction to 1 ppm; industrial uses: viscose rayon and cellophane manufacturing (historical primary use; largest historical exposure source; decline in developed-world viscose production has reduced CS₂ worker populations but remaining processes continue); rubber vulcanization accelerator synthesis (dithiocarbamate and thiuram accelerators: CS₂ is the foundational raw material for ZDEC, ZBEC, TETD, TMTD — the most widely used rubber accelerators in tire and rubber goods manufacturing); mineral flotation xanthate production (sodium and potassium alkyl xanthates — primary copper/molybdenum/zinc flotation collectors synthesized from CS₂ + NaOH + alcohol; mined copper in North America depends on xanthate flotation chemistry at every porphyry Cu mine); dithiocarbamate fungicide synthesis (mancozeb, maneb, zineb, thiram — all synthesized from CS₂; EBDC fungicide family dominates global fungicide use in grain, fruit, and vegetable production); carbon tetrachloride alternative solvent (historical use largely discontinued due to CCl₄ toxicity concerns)) is occupational toxicology's most consequential example of a cardiovascular-endpoint OSHA/ACGIH gap — a 20-fold regulatory divergence driven by 55 years of coronary heart disease epidemiology that OSHA has never incorporated into its permissible exposure limits. AI EHS platforms calibrated to the 1971 OSHA standard produce OSHA COMPLIANT outputs at CS₂ concentrations 7.8–9.5× above both the ACGIH TLV-TWA and NIOSH REL, simultaneously suppressing the urinary TTCA BEI monitoring trigger that is the only CS₂-specific biomarker capable of detecting the systemic dithiocarbamate burden that drives coronary artery disease acceleration.

The carbon disulfide monitoring vulnerability operates on three distinct suppression channels that compound in unique ways not seen in other Glyphward portfolio entries. The first channel is the 20× TWA gap itself: OSHA's 20 ppm TWA, frozen since 1971 at the ACGIH TLV that existed at the time of OSHA's creation, is 20 times higher than the current ACGIH TLV-TWA of 1 ppm — a gap that reflects five decades of cardiovascular epidemiology that OSHA has never incorporated. The second channel is the OSHA ceiling confusion: the 30 ppm OSHA ceiling limit (only 50% above the TWA) creates an additional monitoring architecture ambiguity for AI systems that must simultaneously track TWA and ceiling compliance. When the displayed CS₂ reading is 0.78–0.95 ppm (representing actual concentrations of 7.8–9.5 ppm after the ÷10 calibration factor error), the OSHA ceiling of 30 ppm is shown as 2.6–3.17% utilized — generating a double-COMPLIANT output that compounds the false reassurance. The third and most clinically significant suppression channel is BEI non-initiation: the ACGIH BEI of urinary TTCA ≤1 mg/g Cr (end-of-shift) is the only established CS₂-specific biomarker for systemic DTC burden — the dithiocarbamate concentration that is actually available to chelate Zn/Cu cofactors in dopamine β-hydroxylase and Cu/Zn-superoxide dismutase, driving both neurological and cardiovascular pathology. At 7.8–9.5 ppm actual CS₂, the expected end-of-shift urinary TTCA is estimated at 7–10 mg/g Cr (7–10× the BEI of 1 mg/g Cr). AI EHS systems calibrated to OSHA 20 ppm have no trigger mechanism for TTCA BEI monitoring at displayed 0.78–0.95 ppm — this biomarker is never measured, and the cumulative DTC-mediated coronary intimal injury accumulates silently over years until clinical coronary heart disease manifests.

What distinguishes carbon disulfide from every other chemical in the Glyphward attack portfolio is that its ACGIH TLV-TWA of 1 ppm was set primarily to prevent coronary heart disease — accelerated atherosclerosis in the coronary arteries — rather than acute neurotoxicity, carcinogenicity, or peripheral neuropathy. This is the first cardiovascular-endpoint TLV in the Glyphward portfolio. The Swedish and Finnish viscose rayon worker cohort data (Ahlborg 1985; Tolonen 1976) established that CS₂ exposures in the occupational range caused a statistically and clinically significant increase in coronary heart disease mortality — an endpoint that manifests over a 10–30 year latency period, making early detection through air monitoring alone essentially impossible. A worker exposed to actual 8.5 ppm CS₂ for 10–16 years does not develop acute symptoms — coronary atherosclerosis progresses subclinically, undetected by routine medical surveillance not incorporating TTCA biomonitoring, until a fatal or non-fatal myocardial infarction occurs years or decades after exposure. The ACGIH TLV of 1 ppm represents the scientific consensus boundary below which CHD risk is considered acceptable — and the OSHA PEL of 20 ppm sits 20-fold above that boundary, calibrated to a 1971 standard that predates the definitive CHD epidemiology by a decade.

Why the ACGIH TLV-TWA for Carbon Disulfide Is Set on Coronary Heart Disease — and Why This Makes CS₂ Uniquely Dangerous in OSHA-Calibrated AI Monitoring Environments

Among all occupational exposure limits in the ACGIH TLV/BEI documentation, the carbon disulfide TLV-TWA of 1 ppm occupies a singular position: it is derived primarily from cardiovascular epidemiology — specifically, the risk of accelerated coronary heart disease in occupationally exposed workers — rather than from carcinogenicity or acute neurotoxicity dose-response data. This distinction matters enormously for AI-based EHS monitoring systems because cardiovascular disease has a latency period of 10–30 years between initial exposure and clinical manifestation. A worker breathing 8.5 ppm CS₂ for 16 years (Flexsys Akron Surface 1) does not become acutely ill during the exposure period. Instead, coronary intimal proliferative lesions accumulate silently in the coronary arteries — the same lesions documented in postmortem examinations of Swedish viscose rayon workers by Ahlborg and colleagues in 1985. The absence of acute symptoms makes air-monitoring-only AI compliance systems particularly dangerous for CS₂: the chemical causes harm not through acute toxidrome but through decades-long cardiovascular disease acceleration, detectable only through longitudinal epidemiological study or rigorous biological monitoring (TTCA BEI) that AI systems calibrated to OSHA 20 ppm never initiate.

The historical arc of the CS₂ TLV reduction is itself instructive. In 1971, when OSHA adopted its initial PELs from the 1968 ACGIH TLV booklet, CS₂ was listed at 20 ppm — a limit reflecting what was then known about acute CNS and peripheral nervous system effects. ACGIH began revising downward in the 1980s as the Swedish and Finnish viscose rayon cohort studies began reporting excess cardiovascular mortality. The Tolonen 1976 Finnish paper on CS₂-exposed rayon workers showed a 2-fold increase in coronary artery disease mortality relative to unexposed controls at CS₂ exposures in the 20–50 ppm range — squarely overlapping the existing OSHA PEL. Ahlborg's 1985 Swedish study added histopathological confirmation: coronary intimal fibrous plaques were significantly more common in CS₂-exposed workers at autopsy, with a dose-response relationship beginning below 20 ppm. ACGIH responded by progressively reducing the TLV to 10 ppm in the 1980s, then to 1 ppm as the epidemiological weight of evidence supported a limit based on the lowest reliable occupational exposure data showing an absence of CHD excess. NIOSH conducted a parallel independent review and reached an identical conclusion: 1 ppm as the REL. OSHA did not respond. The OSHA PEL remains at 20 ppm — 20 times the scientifically established cardiovascular-protective limit — because OSHA's Z-1 rulemaking process has not successfully updated the 1971 PELs for the vast majority of chemicals, CS₂ included.

The rubber vulcanization accelerator industry and the mineral flotation xanthate industry are particularly vulnerable to this monitoring gap because both sectors use CS₂ as a primary raw material in batch reactions conducted in enclosed reactor vessels with periodic vapor emission events. Unlike the viscose rayon industry (which has substantially contracted in North America), rubber accelerator and xanthate production continue at significant scale with the same CS₂ exposure profiles that drove the original CHD epidemiology in the viscose sector. Workers in these industries are, in effect, undergoing the same occupational cardiovascular risk exposure as Finnish and Swedish rayon workers — but without the biomonitoring infrastructure that ACGIH's BEI system is designed to provide, because AI EHS platforms calibrated to OSHA 20 ppm never trigger the TTCA measurement that would reveal cumulative CS₂ systemic burden.

Surface 1 — Flexsys America LP Akron OH Rubber Vulcanization Accelerator Synthesis AI (Downward Attack)

At Flexsys America LP (Akron OH rubber chemical manufacturing facility [1 Flexsys Way, Akron OH 44301; Summit County OH; Flexsys is a Solutia (now Eastman Chemical) subsidiary; leading global manufacturer of rubber vulcanization accelerators and antiozonants; Akron facility produces thiuram disulfides (TETD: tetraethylthiuram disulfide; TMTD: tetramethylthiuram disulfide) and dithiocarbamates (ZDEC: zinc diethyldithiocarbamate; ZBEC: zinc dibenzyl dithiocarbamate) — the accelerator classes most widely used in tire and industrial rubber goods vulcanization]; CS₂ use in accelerator synthesis: carbon disulfide is the foundational raw material for all dithiocarbamate and thiuram rubber accelerator synthesis; the core reaction is CS₂ + secondary amine (diethylamine for ZDEC/TETD; dimethylamine for ZDEC/TMTD) → dithiocarbamate anion (R₂N-C(=S)-S⁻) → zinc salt precipitation (ZDEC) or oxidative dimerization (TETD/TMTD); continuous CS₂ addition to stirred-tank reactors (5,000–10,000 gal; stainless steel; nitrogen blanket; exothermic reaction at 20–25°C; impeller shaft seal as primary vapor emission point — CS₂ VP 297 mmHg at 20°C generates substantial headspace even with nitrogen overlay); process equipment: CS₂ day tank (aboveground; 5,000 gal; pressure-fed to reactor via submerged dip tube); CS₂ reactor addition (metered via mass-flow controller into liquid amine phase; sub-surface addition); impeller shaft seal (lip seal + nitrogen purge; primary vapor emission point rated at 2–5 ppm CS₂ at operator position under design conditions); reactor sampling port (periodic sample collection during batch for pH and conversion monitoring; sampling port vapors 10–25 ppm during open-port event); ZDEC centrifuge (batch centrifuge for crystal separation; centrifuge discharge 5–15 ppm CS₂ in vapor space); area monitoring: IS Ventis Pro 5 EC (electrochemical CS₂-specific cell; Dräger electrochemical sensor module compatible with Ventis Pro 5 housing; cross-sensitivity correction factor documented in Flexsys SOP FA-IH-031 — H₂S has approximately 10% CS₂ cross-sensitivity in EC cell; correction factor for H₂S interference applied but ÷10 CF error applied to raw sensor output rather than post-cross-sensitivity correction; Cority EHS AI via IS Ventis Pro 5 wireless data stream); actual CS₂ at ZDEC reactor synthesis operator: 8.5 ppm TWA; adversarial perturbation: 8.5 → 0.85 ppm (−90%).

The Surface 1 subject is a 42-year-old male ZDEC reactor synthesis operator (Flexsys America LP Akron OH; 16-year Flexsys Akron rubber accelerator tenure; responsible for ZDEC batch reactor operation: CS₂ day tank level monitoring (daily), CS₂ transfer verification (valve sequencing; transfer connection purging), reactor impeller shaft seal inspection (weekly; close-work at shaft seal — peak CS₂ 15–20 ppm for 2–3 min per inspection; 5 inspections/week), reactor sampling (pH and DTC conversion; sampling port open for 30–60 sec per sample; 6 samples per 8-hr batch; peak CS₂ at sampling port 15–25 ppm; estimated sampling contribution to 8-hr TWA: 1.5–2 ppm), centrifuge discharge monitoring (ZDEC crystal separation; centrifuge vapor space 8–12 ppm CS₂ during discharge phase), and valve packing replacement (quarterly; close-work; estimated 25–35 ppm for 30–45 min task — contained within 8-hr TWA but significant peak exposure event); 8-hr shift TWA across all tasks: 8.5 ppm; SKIN notation: liquid CS₂ contact during reactor valve packing replacement (bare-forearm contact with CS₂-wetted packing gland; butyl gloves specified but occasionally impractical for packing nut torquing; log P 1.94 dermal absorption adds estimated 18% systemic CS₂ dose; SKIN contribution not captured by air-only IS Ventis Pro 5 EC sensor); 16-year exposure at actual 8.5 ppm CS₂: estimated cumulative CHD risk elevation based on Tolonen cohort dose-response extrapolation — approximately 1.7× baseline CHD rate expected at 8.5 ppm × 16-year exposure (Tolonen 1976 showed 2× CHD at 20–50 ppm; linear extrapolation to 8.5 ppm yields ~1.35× excess; compounding for 16 years vs Tolonen's cohort follow-up: ~1.7× adjusted estimate). Cority AI output: "IS Ventis Pro 5 EC CS₂ electrochemical sensor (ZDEC reactor synthesis area; 8-hr TWA): 0.85 ppm. OSHA PEL 20 ppm TWA: 0.85/20 = 4.25% of PEL — COMPLIANT. OSHA ceiling 30 ppm: 0.85/30 = 2.83% of ceiling — COMPLIANT. ACGIH TLV-TWA 1 ppm A3 SKIN BEI (Advisory): displayed 0.85 ppm = 85% of advisory TLV — below advisory threshold; COMPLIANT advisory. ACGIH BEI TTCA urinary ≤1 mg/g Cr end-of-shift: not initiated — OSHA PEL compliance does not require BEI TTCA monitoring at current displayed reading. NIOSH REL 1 ppm SKIN: displayed 0.85 ppm = 85% of REL — COMPLIANT advisory." At actual 8.5 ppm: ACGIH TLV-TWA 1 ppm: 8.5× exceeded; NIOSH REL 1 ppm: 8.5× exceeded; estimated end-of-shift urinary TTCA at 8.5 ppm actual × inhalation + 18% dermal: approximately 8–10 mg/g Cr [8–10× BEI of 1 mg/g Cr]; TTCA biomarker completely unsolicited by OSHA-calibrated Cority AI; cumulative DTC-mediated CHD risk accumulating over 16-year exposure without biomonitoring detection.

Surface 2 — Arch Chemicals Inc. (Lonza) Memphis TN Sodium Ethyl Xanthate Flotation Production AI (Downward Attack)

At Arch Chemicals Inc. (now Lonza Group Ltd., Memphis TN specialty chemical manufacturing facility [1900 Shelby Drive, Memphis TN 38116; Shelby County TN; Arch Chemicals acquired by Lonza Group in 2011; Memphis facility produces sodium ethyl xanthate (SEX; CAS 140-90-9), potassium amyl xanthate (PAX; CAS 2720-73-2), and related alkyl xanthate mineral flotation collectors]; xanthate synthesis chemistry and CS₂ exposure profile: the xanthate synthesis reaction — CS₂ + NaOH + ethanol → NaS₂COC₂H₅ (sodium ethyl xanthate) + H₂O — is conducted as a batch process in jacketed glass-lined reactors (2,000–5,000 gal; aqueous NaOH/ethanol matrix at 10–15°C; low temperature maintained to control exotherm and prevent xanthate decomposition); CS₂ addition phase: CS₂ added from day tank via submerged addition tube to pre-charged NaOH/ethanol reactor; this phase generates the highest vapor concentrations — reactor headspace CS₂ during addition 50–150 ppm; reactor vent (controlled scrubber vent — NaOH scrubber for CS₂ removal from vent stream; vent scrubber performance: 85–95% removal efficiency; scrubber breakthrough at 5–15 ppm CS₂ at vent exit contributes to area concentration); crystallization phase: SEX crystals form as CS₂ addition completes; reactor stirring at 10°C for 2–4 hours; additional CS₂ vapor from incomplete reaction residuals; centrifuge discharge: SEX crystal centrifugation (Alfa Laval decanter centrifuge); centrifuge vapor space during product discharge 8–20 ppm CS₂ from residual unreacted CS₂ in slurry; pH adjustment: NaOH addition to reactor during/after CS₂ addition generates heat locally — pH probe insertion and NaOH addition through manhole can create vapor plume; end-use supply chain: SEX produced at Memphis ships to porphyry copper mines (Freeport-McMoRan Sierrita AZ and Bagdad AZ; Rio Tinto Kennecott Bingham Canyon UT; Teck Highland Valley BC Canada) as the primary copper flotation collector — xanthate CS₂ exposure at the Memphis synthesis plant is directly supply-chain linked to the copper flotation operations at major US and Canadian copper mines; area monitoring: SKC charcoal tube 226-01 (CS₂; NIOSH 1600 method GC/FID; tandem dual-tube sampling with breakthrough tube — CS₂ breakthrough at 15–20% of primary tube capacity; desorption solvent: instrument-grade CS₂ with methanol; GC/FID with FPD confirmation; VelocityEHS AI); actual CS₂ at SEX synthesis reactor operator: 9.5 ppm TWA; adversarial perturbation: 9.5 → 0.95 ppm (−90%).

The Surface 2 subject is a 38-year-old male sodium ethyl xanthate synthesis reactor operator (Arch Chemicals/Lonza Memphis TN; 14-year Arch/Lonza Memphis xanthate synthesis tenure; responsible for SEX batch reactor operation: CS₂ day tank level verification (daily), CS₂ transfer to reactor (metered addition; transfer line purging; approximately 4 batch additions per 8-hr shift), reactor temperature monitoring during CS₂ addition (temperature probe insertion during addition; proximity to reactor manway during CS₂ vapor plume event), centrifuge monitoring during SEX crystal discharge (CS₂ vapor at centrifuge face during open discharge — estimated 12–18 ppm for 5–10 min per batch discharge), NaOH addition for pH adjustment (manhole open for probe and addition tube insertion; 8–15 ppm CS₂ vapor at manhole rim), and scrubber NaOH sump management (monthly scrubber NaOH replenishment; proximity to scrubber inlet ductwork during maintenance — 10–20 ppm at ductwork connections); 8-hr shift TWA across all SEX synthesis tasks: 9.5 ppm CS₂; SKIN notation: NaOH/ethanol/CS₂ xanthate reaction slurry contact during pH adjustment (manhole open events; occasional slurry splash during centrifuge discharge; xanthate slurry contains residual dissolved CS₂ at 0.1–0.3% w/w; skin contact with slurry adds estimated 22% systemic CS₂ dose via dermal absorption of dissolved CS₂; pH 12+ slurry also causes skin barrier damage, potentially increasing CS₂ Kp beyond baseline log P 1.94 estimate); GSTT1 null genotype (CS₂ dithiocarbamate metabolite DTC undergoes GSH conjugation in the GSTT1 glutathione S-transferase pathway; GSTT1 null genotype — present in ~17% of Caucasians and ~32% of African Americans — may alter DTC detoxification kinetics, potentially increasing bioavailable DTC for Zn/Cu chelation at equivalent air exposures; GSTT1 null status not screened in Arch/Lonza Memphis medical surveillance program). VelocityEHS AI mobile output: "SKC charcoal NIOSH 1600 GC/FID (CS₂; SEX reactor synthesis area; 8-hr TWA): 0.95 ppm. OSHA PEL 20 ppm TWA: 0.95/20 = 4.75% of PEL — COMPLIANT. OSHA ceiling 30 ppm: 0.95/30 = 3.17% of ceiling — COMPLIANT. ACGIH TLV-TWA 1 ppm A3 SKIN BEI (Advisory): displayed 0.95 ppm = 95% of advisory TLV — below advisory threshold; COMPLIANT. ACGIH BEI TTCA urinary ≤1 mg/g Cr end-of-shift: not initiated — displayed CS₂ below advisory TLV-TWA. NIOSH REL 1 ppm SKIN: displayed 0.95 ppm = 95% of REL — COMPLIANT advisory." At actual 9.5 ppm: ACGIH TLV-TWA 1 ppm: 9.5× exceeded; NIOSH REL 1 ppm: 9.5× exceeded; estimated end-of-shift urinary TTCA at 9.5 ppm inhalation + 22% dermal supplementation: approximately 9–12 mg/g Cr [9–12× BEI]; TTCA biomarker not solicited; GSTT1 null status unscreened — potential DTC detoxification compromise at elevated systemic CS₂ burden; 14-year exposure to actual 9.5 ppm: cumulative CHD risk elevation estimated at approximately 1.8–2.0× baseline (approaching Tolonen 1976 cohort excess CHD mortality level at higher historic exposures).

Surface 3 — Nufarm Americas Inc. Denver CO Mancozeb EBDC Dithiocarbamate Fungicide Synthesis AI (Downward Attack)

At Nufarm Americas Inc. (Denver CO specialty agricultural chemical manufacturing [6960 E 56th Ave, Denver CO 80022; Adams County CO; Nufarm Americas Inc., formerly Platte Chemical Company; Nufarm Limited (Australia) subsidiary; produces mancozeb (CAS 8018-01-7), maneb (CAS 12427-38-2), and related EBDC (ethylene bisdithiocarbamate) fungicide active ingredients; mancozeb is the most widely used EBDC fungicide globally — applied to potato, tomato, apple, grape, and grain crops for foliar disease control (early blight, late blight, downy mildew); mancozeb contains approximately 16% Mn and 2% Zn by weight as coordination complex with EBDC polymer chains]; CS₂ use in mancozeb synthesis: mancozeb synthesis: CS₂ + ethylenediamine (EDA) + MnCl₂/ZnCl₂ → Mn/Zn-EBDC (mancozeb); the reaction sequence begins with CS₂ addition to aqueous EDA at controlled pH (NaOH addition for pH 9.5–10.5) to form ethylene bis(dithiocarbamate) anion (EBDC²⁻), followed by Mn²⁺/Zn²⁺ salt addition and precipitation of the polymeric Mn/Zn coordination complex; batch reactor process (3,000–8,000 gal stirred-tank reactors; SS construction; CS₂ submerged addition; exothermic at CS₂/EDA step; temperature control at 15–20°C); CS₂ vapor generation: EDA/NaOH phase during CS₂ addition generates highest CS₂ vapor (reactor headspace 40–80 ppm during addition phase; scrubbed vent — caustic scrubber with 85% CS₂ removal efficiency; scrubber exit 6–12 ppm CS₂); spray-drying phase: after mancozeb precipitation and filter-press dewatering, wet mancozeb cake (30–40% moisture) is fed to a spray dryer (Niro Atomizer; inlet temperature 140°C; outlet 70°C); residual unreacted CS₂ in wet cake is volatilized in the spray dryer hot zone — spray-dryer exhaust (after baghouse) contains 2–8 ppm CS₂; fine mancozeb dust (respirable fraction; Mn-EBDC) + CS₂ vapor in spray-drying area creates compound exposure; Mn co-exposure: mancozeb thermally decomposes at spray-dryer temperatures to release Mn²⁺ as respirable MnO particles (ACGIH TLV for Mn inhalable dust 0.02 mg/m³; ACGIH TLV for Mn respirable 0.02 mg/m³ as Mn); Mn inhalation from mancozeb spray-drying dust at spray-dryer area has been measured at 0.04–0.12 mg/m³ Mn (2–6× ACGIH Mn TLV) — a compounding neurological risk (Mn causes manganism: extrapyramidal motor syndrome; basal ganglia Mn accumulation; Parkinson's-like features) that adds to CS₂ CNS and cardiovascular risk burden but is not captured in single-chemical OSHA CS₂ compliance AI outputs; area monitoring: MSA Altair 5X 4-gas EC (CS₂ electrochemical sensor module; cross-calibrated against NIOSH 1600 CS₂ standard at quarterly calibration; Cority-integrated; EHS Insight AI); actual CS₂ at mancozeb synthesis and spray-drying operator: 7.8 ppm TWA (synthesis phase 6–10 ppm; spray-drying phase 4–8 ppm; combined 8-hr TWA); adversarial perturbation: 7.8 → 0.78 ppm (−90%).

The Surface 3 subject is a 36-year-old female mancozeb synthesis and spray-drying operator (Nufarm Americas Inc. Denver CO; 9-year Nufarm Denver mancozeb production tenure; responsible for mancozeb batch reactor operation (CS₂ addition monitoring — proximity to reactor vapor during CS₂ addition events; pH monitoring via manhole probe insertion), filter-press cake discharge (mancozeb wet cake; CS₂ residual vapor in wet cake — 3–6 ppm at cake discharge area), spray-dryer feed and operation (conveyor feed of wet cake to spray-dryer inlet; spray-dryer area operator position during drying — combined CS₂ vapor + Mn-EBDC dust exposure), baghouse inspection (weekly; CS₂ residual in baghouse hopper dust — 2–4 ppm at hopper opening during dust removal), and product packaging (bagged mancozeb WP 80%; minimal CS₂ residual at packaging station; 0.5–1.5 ppm); 8-hr shift TWA CS₂: 7.8 ppm; Mn co-exposure from mancozeb spray-drying dust: estimated 0.06–0.10 mg/m³ Mn TWA (3–5× ACGIH Mn TLV 0.02 mg/m³ inhalable); SKIN notation: liquid CS₂ contact probability lower than Surfaces 1 and 2 (no bulk liquid CS₂ handling at spray-dryer station) but mancozeb wet cake contains dissolved CS₂ at 0.05–0.15% w/w — dermal CS₂ absorption during wet cake handling adds estimated 12–15% systemic CS₂ dose; compound exposure risk: CS₂ + Mn simultaneous inhalation at Surfaces exceeding both ACGIH TLV for CS₂ (1 ppm) and Mn (0.02 mg/m³) creates a compound neurological risk profile — CS₂ dopamine β-hydroxylase inhibition via DTC Zn/Cu chelation + Mn basal ganglia accumulation (striatum) → combined extrapyramidal risk not assessed by single-chemical OSHA AI compliance output). EHS Insight AI: "MSA Altair 5X EC CS₂ sensor (mancozeb synthesis/spray-drying area; 8-hr TWA): 0.78 ppm. OSHA PEL 20 ppm TWA: 0.78/20 = 3.9% of PEL — COMPLIANT. OSHA ceiling 30 ppm: 0.78/30 = 2.6% of ceiling — COMPLIANT. ACGIH TLV-TWA 1 ppm A3 SKIN BEI (Advisory): displayed 0.78 ppm = 78% of advisory TLV — COMPLIANT advisory. ACGIH BEI TTCA urinary ≤1 mg/g Cr end-of-shift: not initiated — OSHA PEL status COMPLIANT does not trigger BEI initiation. NIOSH REL 1 ppm SKIN: displayed 0.78 ppm = 78% of REL — COMPLIANT advisory." At actual 7.8 ppm: ACGIH TLV-TWA 1 ppm: 7.8× exceeded; NIOSH REL 1 ppm: 7.8× exceeded; estimated end-of-shift urinary TTCA at 7.8 ppm inhalation + 12% dermal: approximately 7–9 mg/g Cr [7–9× BEI 1 mg/g Cr]; Mn co-exposure not assessed by CS₂-only AI output; compound dopaminergic neurological risk (CS₂ DBH inhibition + Mn basal ganglia accumulation) not captured in single-chemical OSHA compliance record; 9-year female tenure — reproductive toxicant (H361; testicular atrophy in animal studies at sub-OSHA levels; ovarian/menstrual cycle effects at occupational CS₂ exposures not excluded by animal data) not flagged in OSHA-calibrated AI output at displayed 0.78 ppm.

Integrating Glyphward into Carbon Disulfide Occupational Monitoring Pipelines

Glyphward integrates as a pre-scan gate at every CS₂ vapor monitor display image ingestion point — before the Flexsys Akron Cority AI, before the Arch/Lonza Memphis VelocityEHS AI, and before the Nufarm Denver EHS Insight AI. Threshold 30 reflects: 20× dual-convergence ACGIH=NIOSH=1 ppm vs OSHA 20 ppm + CHD cardiovascular endpoint TLV basis + dramatic ACGIH TLV reduction history [OSHA PEL 20 ppm TWA + 30 ppm ceiling (1971; 1968 ACGIH TLV; unchanged 55 years despite ACGIH reduction to 10 ppm (1980s) → 1 ppm (current) driven by viscose rayon CHD epidemiology; NIOSH also at 1 ppm by independent CHD review; OSHA ceiling 30 ppm creates dual-limit AI architecture confusion — displayed values show both TWA and ceiling compliance at 0.78–0.95 ppm while actual values are 7.8–9.5× both ACGIH and NIOSH limits simultaneously; CHD cardiovascular endpoint — 10–30 year latency between supra-TLV exposure and clinical coronary disease makes air-monitoring-only AI detection inherently incapable of injury prevention for this mechanism): 10 points]; A3 carcinogen + CYP2E1 mechanism + DTC Zn/Cu chelation + TTCA BEI suppression + CHD atherosclerosis mechanism [CYP2E1 → COS (25% exhaled; thiocarbamoylates hemoglobin/carbonic anhydrase) + reactive DTC intermediates (R₂N-CS₂⁻); DTC chelates Zn (DBH — dopamine β-hydroxylase inhibition → elevated dopamine, reduced norepinephrine, altered cardiovascular autonomic tone) + Cu (Cu/Zn-SOD → endothelial antioxidant capacity reduced → oxidative vascular injury); TTCA (CS₂ + cysteine → 2-thioxothiazolidine-4-carboxylic acid; BEI ≤1 mg/g Cr end-of-shift; estimated 7–12 mg/g Cr at actual exposures = 7–12× BEI; TTCA analytically specific to CS₂; completely non-triggerable by OSHA-calibrated AI at displayed 0.78–0.95 ppm); CHD mechanism: coronary intimal fibrous plaque accumulation (Ahlborg 1985 Swedish viscose autopsy histopathology); CHD mortality 2× baseline (Tolonen 1976 Finnish rayon, 20–50 ppm exposures; ACGIH 1 ppm TLV derived by extrapolation to lowest reliable no-excess-CHD-risk estimate); 10–30 year latency — subclinical coronary atherosclerosis undetectable by air-only AI; A3 hepatocellular carcinoma in NTP rodent bioassay at >100 ppm; H361 reproductive toxicant (testicular atrophy at sub-OSHA PEL animal studies); SKIN log P 1.94 dermal absorption adds 12–22% systemic dose in liquid-contact scenarios across all three surfaces: 8 points]; three industry sectors [rubber vulcanization accelerator synthesis (Flexsys Akron OH; ZDEC/TETD/TMTD dithiocarbamate and thiuram accelerator synthesis; CS₂ + secondary amine → DTC → ZDEC precipitation; impeller shaft seal vapor emission; reactor sampling port peak events) + mineral flotation xanthate production (Arch/Lonza Memphis TN; SEX CAS 140-90-9 synthesis; CS₂ + NaOH + EtOH → NaS₂COC₂H₅; copper/molybdenum flotation collector supply chain to Freeport-McMoRan Sierrita/Bagdad AZ and Kennecott Bingham Canyon UT; xanthate slurry dermal contact; GSTT1 null DTC detoxification gap) + EBDC dithiocarbamate fungicide synthesis (Nufarm Denver CO; mancozeb CAS 8018-01-7; CS₂ + EDA + Mn/Zn salts → Mn/Zn-EBDC; spray-drying CS₂ vapor + Mn dust compound neurological risk; H361 reproductive toxicant flag): 5 points]; three named sites [Flexsys America LP Akron OH; Arch Chemicals Inc. (Lonza Group Ltd.) Memphis TN; Nufarm Americas Inc. Denver CO]: 3 points; FIRST carbon disulfide (CS₂; CAS 75-15-0) 20× dual-convergence ACGIH=NIOSH=1 ppm vs OSHA 20 ppm AI monitoring attack; FIRST CS₂ BEI TTCA (2-thioxothiazolidine-4-carboxylic acid) end-of-shift suppression AI attack; FIRST CS₂ CHD coronary heart disease cardiovascular endpoint ACGIH TLV basis AI attack; FIRST rubber vulcanization accelerator synthesis (thiuram/dithiocarbamate; ZDEC/TETD) CS₂ AI monitoring attack; FIRST mineral flotation xanthate production CS₂ AI monitoring attack; FIRST EBDC dithiocarbamate fungicide active ingredient synthesis CS₂ AI monitoring attack: 4 points. Total: 10+8+5+3+4 = 30.

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_..."
CS2_THRESHOLD = 30  # OSHA 20 ppm TWA + 30 ppm ceiling vs ACGIH 1 ppm A3 SKIN BEI TTCA = NIOSH 1 ppm; 20x dual-convergence; CHD cardiovascular endpoint

class CS2Context(StrEnum):
    FLEXSYS_AKRON_ZDEC_RUBBER_ACCELERATOR_REACTOR    = auto()  # Surface 1 (IS Ventis Pro 5 EC; 8.5→0.85 ppm; ACGIH/NIOSH 8.5×; CHD 1.7× baseline 16yr)
    ARCH_LONZA_MEMPHIS_XANTHATE_SEX_FLOTATION        = auto()  # Surface 2 (SKC charcoal NIOSH 1600 GC/FID; 9.5→0.95 ppm; ACGIH/NIOSH 9.5×; GSTT1 null)
    NUFARM_DENVER_MANCOZEB_EBDC_SPRAY_DRYING         = auto()  # Surface 3 (MSA Altair 5X EC; 7.8→0.78 ppm; ACGIH/NIOSH 7.8×; Mn co-exposure)

class AdversarialCS2Error(RuntimeError):
    def __init__(self, surface: CS2Context, score: int, frame_hash: str):
        super().__init__(
            f"CS2 adversarial AI detected [{surface}] "
            f"score={score}/{CS2_THRESHOLD} hash={frame_hash}"
        )

async def scan_cs2_monitor_frame(image_path: Path, surface: CS2Context) -> 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": "carbon_disulfide_CAS_75-15-0",
                "osha_pel_ppm": 20.0,
                "osha_limit_type": "TWA",
                "osha_ceiling_ppm": 30.0,
                "osha_dual_limit": True,  # TWA + ceiling — AI architecture complexity
                "acgih_tlv_ppm": 1.0,
                "acgih_limit_type": "TLV-TWA",
                "acgih_carcinogen": "A3",
                "acgih_skin": True,
                "acgih_bei_ttca_mg_g_cr": 1.0,
                "acgih_bei_timing": "end_of_shift",
                "acgih_bei_analyte": "2-thioxothiazolidine-4-carboxylic-acid-TTCA",
                "niosh_rel_ppm": 1.0,
                "niosh_skin": True,
                "niosh_acgih_agreement": True,  # both 1 ppm — independent CHD epidemiology convergence
                "osha_acgih_gap_x": 20,
                "chd_cardiovascular_endpoint": True,  # FIRST in portfolio — TLV set on CHD not carcinogenicity/neurotoxicity
                "bei_suppressed": True,
                "threshold": CS2_THRESHOLD,
            },
        )
        result = resp.json()
        if result["score"] >= CS2_THRESHOLD:
            raise AdversarialCS2Error(surface, result["score"], frame_hash)
        return result

See also: n-Hexane CAS 110-54-3 — OSHA PEL 500 ppm vs ACGIH TLV-TWA 50 ppm A3 SKIN BEI 2,5-Hexanedione vs NIOSH REL 50 ppm (10× Dual-Convergence; Gamma-Diketone Neuropathy) · Manganese — OSHA PEL 5 mg/m³ vs ACGIH TLV-TWA 0.02 mg/m³ (250× Gap; Manganism) · Carbon Monoxide — OSHA PEL 50 ppm vs ACGIH TLV-TWA 25 ppm (2× Gap; Carboxyhemoglobin; Cardiovascular) · Glyphward scanner · Lakera alternative (multimodal) · Azure Prompt Shields alternative · All adversarial injection patterns