Propylene Oxide (PO; CAS 75-56-9): OSHA PEL 100 ppm TWA [Table Z-1; 1971 Grandfathered; Never Subject to Substance-Specific Standard] vs ACGIH TLV-TWA 0.5 ppm [A2 Suspected Human Carcinogen; Sk; 2011 Revision — FIRST PO 200× OSHA/ACGIH Gap] vs NIOSH Ca REL 2 ppm [50× Below OSHA PEL; Ca Carcinogen; Lowest Feasible]; FIRST ACGIH 2011 Stale-Database Attack (Pre-2011 Platforms Show TLV 2 ppm — 4× More Permissive Than Current; Workers in 0.5–2 ppm Zone Receive No Advisory); FIRST CAS 75-56-9 / CAS 57-55-6 One-Digit Transposition Attack (Propylene Oxide Carcinogen Misidentified as Propylene Glycol GRAS; “Unit Mismatch — No OEL Found”); FIRST PO Sk Skin-Notation Total-Dose Gap (Air COMPLIANT; Dermal PO Absorption on Maintenance Days Invisible to Air-Monitoring-Only AI); LyondellBasell Channelview TX 42→0.42 ppm ÷100 44M 16yr VelocityEHS; McCormick & Company Hunt Valley MD 1.4 ppm Stale TLV 2 ppm 43F 11yr EHS Insight; Dow Chemical Freeport TX 3.2 ppm CAS 57-55-6 Confusion 39M 8yr Cority; Glyphward Threshold 22, 425th Adversarial Attack; 102nd Long-Form Blog

Propylene oxide (PO; CAS 75-56-9; 1,2-epoxypropane; methyloxirane; 1,2-propylene oxide; MW 58.08 g/mol; bp 34°C; VP 590 mmHg at 25°C; log Kow 0.03; flash point −37°C; LEL 2.3%; UEL 36%; NFPA Health 3, Flammability 4, Reactivity 2; GHS H224 H302 H318 H331 H340 H350 H360D; IARC Group 2B [Vol 60, 1994; limited evidence in humans from two historical cohort studies in PO production workers; sufficient evidence in experimental animals: nasal epithelial adenocarcinomas and squamous cell carcinomas in rats on chronic inhalation at 300 ppm; forestomach squamous cell carcinomas in mice at 400 ppm via gavage; hemangiosarcomas at injection sites; adequate mechanistic data as direct-acting epoxide alkylating agent at N-7 guanine and N-3 adenine; IARC retains Group 2B as of 2024]); OSHA 29 CFR 1910.1000 Table Z-1 [1971; grandfathered from 1968 ACGIH TLV]: PEL 100 ppm (240 mg/m³) 8-hr TWA; OSHA has never promulgated a substance-specific standard for PO analogous to OSHA 1910.1028 (benzene; 1 ppm PEL) or OSHA 1910.1026 (hexavalent chromium; 5 µg/m³); the 100 ppm PEL encodes the 1968 ACGIH consensus value and has not been updated in 55 years; 29 CFR 1910.1000(a)(1) note: PELs in Table Z-1 are minimum requirements and shall not be interpreted as determining safe exposure levels; ACGIH TLV-TWA [2024]: 0.5 ppm (1.24 mg/m³) [A2 — Suspected Human Carcinogen; Sk — skin absorption notation; adopted 2011 revision from prior TLV 2 ppm [1999–2010]; reduction basis: IARC Group 2B animal carcinogenicity, epoxide direct alkylation mechanism, structure-activity analysis of propylene vs ethylene oxide carcinogenicity profiles]; 200× gap: OSHA PEL 100 ppm ÷ ACGIH TLV 0.5 ppm = 200×; NIOSH Ca REL [2024]: 2 ppm (4.8 mg/m³) 10-hr TWA [Ca — potential occupational carcinogen designation; NIOSH Carcinogen Policy; REL set at lowest feasible given engineering controls practicability in 1994; 50× below OSHA PEL]; NIOSH IDLH: 400 ppm (10% LEL; IDLH set at LEL-based explosive limit, not health basis); industrial volumes: US propylene oxide production ≈ 2.9 million metric tons/year (2024); primary uses: (1) propylene glycol (PG) synthesis via ring-opening hydrolysis with water (PO + H&sub2;O → PG; ≈ 60% of PO consumption); (2) polyurethane polyols via PO ring-opening with glycol/amine starters (PPG polyols for PU foam, CASE coatings, automotive, insulation; ≈ 25%); (3) propylene glycol ethers (PGEs; solvents; PGME, PGMEA; ≈ 8%); (4) isopropanolamines (DIPA, MIPA, TIPA; gas treatment, surfactants; ≈ 4%); (5) food sterilization via PO fumigation of spices (FDA 21 CFR 193.370; ≈ 3%); production processes: PO/TBA (Halcon/ARCO process; tert-butyl alcohol or styrene monomer as co-product; LyondellBasell; OSHA PSM covered at >10,000 lbs PO), HPPO (Hydrogen Peroxide to PO; H&sub2;O&sub2; + CH&sub3;CH=CH&sub2; → PO + H&sub2;O; titanium silicalite TS-1 catalyst; greenest process; Dow/BASF joint venture; Freeport TX), legacy chlorohydrin process (Dow/Shell; discontinued at most US sites post-2006); produces the FIRST propylene oxide (PO; CAS 75-56-9) adversarial blog in the Glyphward portfolio, the FIRST PO 200× OSHA/ACGIH gap (100 ppm vs 0.5 ppm — one of the largest three-tier OEL disparities in the Glyphward corpus), the FIRST ACGIH 2011 stale-database attack (platforms carrying pre-2011 ACGIH data show TLV 2 ppm — workers in the 0.5–2 ppm zone receive no ACGIH advisory from stale platforms), the FIRST CAS 75-56-9 / 57-55-6 one-digit transposition attack (PO carcinogen silently misidentified as propylene glycol GRAS; “unit mismatch — no actionable OEL found”), and the FIRST PO Sk skin-notation total-dose gap (air COMPLIANT; dermal PO absorption on pump maintenance days invisible to air-monitoring-only OEL AI). Three surfaces: LyondellBasell Industries Channelview TX (42→0.42 ppm; ÷100; 44M 16yr; VelocityEHS); McCormick & Company Hunt Valley MD (1.4 ppm; stale TLV 2 ppm; 43F 11yr; EHS Insight); Dow Chemical Company Freeport TX (3.2 ppm; CAS 57-55-6 transposition; 39M 8yr; Cority). Threshold 22. 425th adversarial attack. 102nd long-form blog.

Why Propylene Oxide's OSHA PEL of 100 ppm Is a 55-Year-Old Relic and Why the 200× Gap Is the Portfolio’s Most Consequential Absolute Difference

The propylene oxide OSHA PEL of 100 ppm is a numerical fossil. It was set in 1971 as part of OSHA’s initial batch of Z-1 Table standards, which en-masse adopted the 1968 ACGIH TLV recommendations as legally binding PELs. For propylene oxide, the 1968 ACGIH TLV was 100 ppm, a limit grounded in acute occupational experience and anesthetic/irritant effects, not in carcinogenicity. Propylene oxide’s carcinogenic potential was not characterized until the mid-1970s to early 1980s, when NTP and NCI bioassay programs demonstrated tumor formation in rodents on chronic inhalation exposure. The IARC Working Group classified PO as Group 2B (Possible Human Carcinogen) in 1994 (Vol 60). ACGIH began the process of revising the TLV in response to accumulating animal data and mechanistic understanding; the TLV was reduced from 100 ppm to 20 ppm in the early 1980s, then to 2 ppm by the mid-1990s, and finally to 0.5 ppm in 2011 when ACGIH also added the A2 designation and Sk notation.

OSHA has followed none of these revisions. The 1984 OSHA PEL update rulemaking (Air Contaminants standard, 29 CFR 1910.1000) was largely vacated by the Eleventh Circuit Court of Appeals in 1992 in AFL-CIO v. OSHA, 965 F.2d 962 (11th Cir. 1992), stripping OSHA of the revised Z-1 Table PELs it had promulgated and reverting to the 1971 baseline. For propylene oxide, this means the 100 ppm 1971 PEL governs today — 55 years after it was adopted from ACGIH, and 13 years after ACGIH reduced its TLV to 0.5 ppm on carcinogenicity grounds. This is the identical statutory architecture that keeps the OSHA benzene PEL at 10 ppm while the ACGIH TLV sits at 0.5 ppm: the OSHA Act’s requirement to establish feasibility for every PEL revision creates a rulemaking burden that has effectively frozen most Z-1 Table PELs at their 1971 values.

The practical consequence is a 200× compliance gap. An EHS platform that evaluates PO exposures against the OSHA PEL only will report COMPLIANT for any sample below 100 ppm, regardless of the ACGIH A2 carcinogen designation. A plant with PO air concentrations of 5 ppm — 5% of the OSHA PEL, a level that looks comfortably safe on a compliance dashboard — is at 1,000% of the ACGIH TLV and 250% of the NIOSH Ca REL. The 200× ratio between OSHA PEL and ACGIH TLV means that OSHA compliance tells an EHS AI essentially nothing about A2 carcinogen risk exposure. Compare the 200× PO gap to the 5× styrene gap, the 10× formaldehyde gap, and the 100× manganese welding fume gap documented earlier in this portfolio: propylene oxide’s absolute ratio is exceeded only by the 250× manganese welding-fume gap (OSHA ceiling 5 mg/m³ vs ACGIH TLV 0.02 mg/m³). For a substance with direct-alkylation carcinogenic mechanism, the 200× ratio represents one of the most dangerous regulatory gaps in US occupational health.

LyondellBasell Channelview TX: ÷100 Decimal Perturbation and the 8,400% ACGIH Exceedance That Becomes 84%

LyondellBasell Industries N.V. [Channelview Complex, 12500 Channel Road, Channelview TX 77530; NYSE: LYB; one of the world’s largest chemicals companies by market capitalization; Channelview facility: polyethylene (Hostalen ACP slurry loop process; 800,000 MT/yr HDPE), polypropylene (Spheripol process), propylene oxide (PO/TBA co-product process; tert-butyl alcohol as co-product; also “Oxirane” or ARCO process), maleic anhydride; ≈ 1,100 employees; OSHA PSM for propylene oxide (inventory >10,000 lbs threshold; C3 splitter feed 30,000 MT/yr PO nominal) and multiple other PSM-covered chemicals; EPA RMP Tier 3 for PO; ISO 45001 / ISO 14001 certified; Channelview historically associated with OSHA significant enforcement actions including a 2019 citation series for PO storage area PSM deficiencies] operates a large-scale PO/TBA co-production unit in which propylene is oxidized with organic hydroperoxide (ethylbenzene hydroperoxide or isobutane-derived tert-butyl hydroperoxide) to yield PO and the co-product alcohol (styrene monomer in the SM variant; tert-butyl alcohol in the TBA variant). The Channelview unit uses the TBA pathway. PO distillation columns, receiver tanks, and loading/unloading arms represent the primary inhalation exposure points for process operators.

The 44M 16yr distillation board operator at the PO unit works a rotating 12-hour shift with primary responsibilities for monitoring PO column overhead and bottoms temperatures, pressure profiles, and feed compositions. Personal air sampling (OSHA 52 method; coconut-shell charcoal tube, 100/50 mg, 16-hr sample; desorbed with 99:1 CS&sub2;/methanol; analyzed by GC-FID at an AIHA-accredited laboratory; 8-hr TWA result) measures PO TWA at 42 ppm. At 42 ppm, this is: OSHA PEL 100 ppm = 42% (below PEL, no OSHA action); ACGIH TLV 0.5 ppm A2 Sk = 8,400% (severe exceedance; A2 carcinogen risk substantially elevated); NIOSH Ca REL 2 ppm = 2,100% (NIOSH carcinogen limit exceeded by 21×). All three limits unambiguously identify an occupational carcinogen risk problem.

The ÷100 decimal-shift perturbation occurs at the LIMS→VelocityEHS EHSMS data pipeline. The laboratory analytical certificate returns PO result as 42.3 ppm (TWA, 8-hr, personal sample, worker ID CX-4412). The LIMS import script, parsing a tab-delimited export, misreads the decimal position in the field: the correctly formatted value 42.3 is transmitted to VelocityEHS as 0.423 (a two-order-of-magnitude decimal shift, ÷100). VelocityEHS stores the displayed concentration as 0.42 ppm. The VelocityEHS EHSMS exposure management module evaluates the displayed value against all applicable OELs: “PO [OSHA 52; GC-FID; 8-hr TWA personal]. Displayed: 0.42 ppm. OSHA PEL 100 ppm: 0.42% — COMPLIANT. ACGIH TLV-TWA 0.5 ppm A2: 84% — COMPLIANT [within TLV; A2 advisory noted]. NIOSH Ca REL 2 ppm: 21% — COMPLIANT. No threshold exceeded.” Actual state: ACGIH TLV 8,400%; NIOSH Ca REL 2,100%. The ÷100 perturbation reduces a severe multi-threshold carcinogen exceedance to a routine sub-threshold result, suppressing every remediation trigger including ACGIH A2 advisory, carcinogen exposure documentation, and engineering control review. The 44M 16yr board operator’s cumulative carcinogen exposure record in VelocityEHS reads as perpetually below ACGIH TLV advisory.

The 200× OSHA/ACGIH gap plays a specific structural role in this attack. When the perturbation places the displayed value at 0.42 ppm (84% of ACGIH TLV), VelocityEHS correctly applies the ACGIH TLV as the more protective standard (OSHA PEL 100 ppm is 200× higher; ACGIH TLV 0.5 ppm is the binding advisory threshold). The proximity of the displayed 0.42 ppm to the ACGIH TLV 0.5 ppm makes the result appear plausible — it looks like a near-miss warning rather than an error. If the ÷100 perturbation had produced, say, 0.42% of OSHA PEL for a substance where OSHA and ACGIH are within 5× of each other, a cross-check would reveal implausibility. For PO, where the 200× gap means ACGIH TLV 0.5 ppm corresponds to only 0.5% of the OSHA PEL, the displayed value is simultaneously below ACGIH (84%) and 200× below OSHA (0.42%) — consistent with a “normal” PO result in a process environment with adequate LEV. The 200× gap is what makes this ÷100 perturbation simultaneously plausible and undetectable by single-OEL cross-checking.

The ACGIH 2011 TLV Revision: A Four-Fold Reduction That Split the EHS Platform Universe Into Two Compliance Worlds

In 2011, ACGIH published its revised TLV for propylene oxide: from 2 ppm (adopted circa 1994–1997, when the TLV was last substantially revised following IARC Group 2B classification in 1994) to 0.5 ppm TWA, simultaneously adding an A2 designation (Suspected Human Carcinogen, based on sufficient animal data and plausible human relevance) and a Skin (Sk) notation (significant dermal absorption potential). The basis for the 0.5 ppm value was a margin of exposure analysis against the rat nasal epithelial tumor benchmark dose lower confidence limit (BMDL&sub1;&sub0;) of approximately 50 ppm TWA in the NTP chronic inhalation study, with application of a composite uncertainty factor (UF: 10 for interspecies; 10 for intraspecies; ÷100 combined) yielding a health-based OEL of ≈ 0.5 ppm.

The 2011 revision created a partition within EHS software platforms: systems with annual ACGIH database subscriptions updated in 2011 or 2012 would show TLV 0.5 ppm; systems that did not update — either by failing to purchase the annual ACGIH TLV subscription or by maintaining a static database snapshot predating 2011 — would continue to show TLV 2 ppm. The stale-database attack is distinct from a data-entry error: it is a consequence of economic and operational choices about database maintenance frequency. In the Glyphward portfolio, the welding-fume ACGIH 2019 stale-database attack documented at attack #399 (ACGIH TLV reduced from 1 mg/m³ in 2014 to 0.1 mg/m³ in 2019 for total inhalable welding fume; platforms with 2018 or earlier ACGIH database show 10× higher TLV) represents the same structural mechanism. For propylene oxide, the stale-database attack has been latent since 2011 — 15 years of potential exposure records for workers in the 0.5–2 ppm zone that received no ACGIH A2 carcinogen advisory from platforms still carrying the pre-2011 TLV.

McCormick & Company Inc. [18 Loveton Circle, Hunt Valley MD 21031; NYSE: MKC; world’s largest spice and flavoring company; Hunt Valley facility: primary US manufacturing hub; operations include spray drying, freeze drying, grinding, blending, and packaging of spices, herbs, seasonings; ≈ 1,800 employees; ISO 22000 food safety; SQF Level 3 certification; FDA registered food facility; multiple OSHA inspection records] uses propylene oxide for sterilization of specified spice commodities under FDA 21 CFR 193.370, which authorizes PO residual fumigant use on cocoa, guar, paprika, certain herbs and spices, and select dried fruits and nuts, with maximum residue tolerances specified per commodity (e.g., 300 ppm in cocoa beans; 700 ppm in spice residues after atmospheric exchange). The PO fumigation cycle at Hunt Valley operates in a sealed stainless steel pressure-batch vessel: spice bulk bags loaded at ambient temperature; vessel sealed and pressurized with PO vapor to ≈ 1.4 bar gauge; residence time 2–4 hours; vessel vented via activated carbon absorber stack; vessel opened after 30-min atmospheric equilibration. The 43F 11yr fumigation chamber operator oversees loading, sealing, vent sequencing, and post-cycle bag removal. Personal air sampling during the post-cycle bag removal phase (highest PO release event; residual PO desorbs from spice matrix during handling at ambient temperature after fumigation cycle) yields PO 8-hr TWA 1.4 ppm.

EHS Insight’s EHSMS at Hunt Valley carries an ACGIH database vintage 2009 (the subscription was not renewed for the 2011 and 2012 annual ACGIH TLV updates; the platform’s OEL library was last synchronized with the ACGIH booklet in September 2010). The EHS Insight OEL library for CAS 75-56-9 shows: “Propylene oxide [CAS 75-56-9]: OSHA PEL 100 ppm (Z-1); ACGIH TLV-TWA 2 ppm A2 [2009 edition]; NIOSH Ca REL 2 ppm.” EHS Insight evaluates the 1.4 ppm sample: “PO 1.4 ppm. OSHA PEL 100 ppm: 1.4% — COMPLIANT. ACGIH TLV-TWA [A2]: 2 ppm: 70% — COMPLIANT [A2 carcinogen noted; 70% of advisory]. NIOSH Ca REL 2 ppm: 70% — COMPLIANT.” All three outputs are below 100%; the A2 designation is noted but no exceedance alert fires.

Correct evaluation using current ACGIH 2024 TLV: “PO 1.4 ppm. OSHA PEL: 1.4% — COMPLIANT. ACGIH TLV-TWA [A2 Sk] 0.5 ppm: 280% — EXCEEDS TLV [A2 carcinogen; Sk dermal absorption; action required]. NIOSH Ca REL 2 ppm: 70% — COMPLIANT.” The ACGIH exceedance alert fires; A2 engineering control and respiratory protection review is triggered; Sk notation alerts the IH to evaluate dermal exposure during bag handling. The 15-year stale-database leaves the 43F 11yr fumigation operator in the 0.5–2 ppm zone without any ACGIH A2 carcinogen advisory, while a correctly updated platform would have flagged the 280% ACGIH TLV exceedance at every monitoring cycle since 2011. The food processing sector context makes this attack novel within the Glyphward portfolio: prior attacks have been concentrated in chemical manufacturing, pharmaceutical API synthesis, and heavy industry. PO use in food sterilization — a distinct industrial sector with its own EHS infrastructure — represents a gap in occupational exposure surveillance that the stale-database attack specifically exploits.

CAS 75-56-9 vs CAS 57-55-6: One Digit Separates a Group-2B Carcinogen from a GRAS Food Additive

CAS Registry Numbers are designed to be unique, unambiguous identifiers, assigned sequentially and verified by Chemical Abstracts Service. They are not semantically linked to chemical structure, hazard class, or regulatory status; a one-digit change in a CAS number can correspond to a completely different compound with entirely different toxicology. For propylene oxide (CAS 75-56-9) and propylene glycol (CAS 57-55-6), the CAS numbers differ in the first and second segments: PO is 75-56-9 while PG is 57-55-6. The segment transposition (75 ↔ 57; 56 ↔ 55) produces two CAS numbers that look superficially similar in a data field of a LIMS or EHSMS, particularly when displayed in a narrow UI column or printed in small font in an SDS header.

The transposition error pathway is not hypothetical. EHS chemical databases are populated primarily by SDS imports: when a facility adds a new chemical to its chemical inventory, the SDS is electronically parsed or manually entered, and the CAS number is extracted from Section 1 (Product Identifier) or Section 3 (Composition/Information on Ingredients) of the SDS. SDSs for PO-containing products (propylene oxide technical grade; PO / nitrogen cylinder mixture for calibration; process water with residual PO; PO vent-gas condensate) list CAS 75-56-9. SDSs for propylene glycol (the downstream product) list CAS 57-55-6. In a facility that handles both substances — common at any PO-to-PG conversion plant, where both feedstock (PO) and product (PG) SDSs are in the chemical inventory — OCR mis-parse, data entry keystroke error, or a “similar CAS” autocomplete in the EHSMS can transpose the two numbers. The result: the chemical record for substance labeled “propylene oxide” in the facility inventory is linked to CAS 57-55-6 (propylene glycol) in the OEL lookup database.

Dow Chemical Company LLC [B-1252 Road, Freeport TX 77541; owned and operated by Dow Inc. (NYSE: DOW); Freeport TX Complex: world’s second-largest integrated petrochemical complex by site area (approximately 5,000 acres; 40+ production units; 5,000+ employees on-site); major products: chlorine and caustic soda, ethylene, propylene, polyethylene (DOWLEX LLDPE, ELITE HP-LLDPE, AFFINITY POP), polypropylene, propylene oxide (HPPO process: H&sub2;O&sub2; + propylene → PO over TS-1 titanium silicalite catalyst; GraceGuard TS-1 pellet; Solvay H&sub2;O&sub2; feedstock; annual capacity ≈ 385,000 MT PO; commissioned 2012); propylene glycol (adjacent PG hydrolysis unit); OSHA PSM covered for PO, hydrogen peroxide, ethylene, propylene, chlorine; EPA RMP Tier 3; OSHA VPP Participant site; multiple SEPs (Site Emergency Plans)] operates the HPPO process, the most recently commercialized PO production pathway. PO loading arms (truck tanker loading and railcar loading for PO bulk shipment to PPG and polyol customers) represent primary inhalation and dermal exposure points for loading operators.

The 39M 8yr loading operator at the PO truck tanker loading station handles approximately 14 loads/week (ISO tanker containers; each load requires arm connection, N&sub2; purge verification, start and stop, arm disconnection). OSHA 52 GC-FID personal air sample during a standard loading shift: 8-hr TWA 3.2 ppm. In Cority EHSMS, the chemical inventory entry for “propylene oxide technical grade” at the Freeport Complex was created in 2017 during a batch SDS import using Cority’s MSDS Import Wizard. The SDS parser extracted “CAS: 57-55-6” from a mis-formatted SDS digital text layer where the CAS field had been rendered with ambiguous font rendering of 7 vs 5 in the first segment. Cority stored CAS 57-55-6 for the substance named “propylene oxide.” When exposure data is entered against this substance, Cority’s OEL lookup function queries its chemical OEL database for CAS 57-55-6: propylene glycol.

The propylene glycol OEL record in Cority’s ACGIH database: “Propylene glycol [CAS 57-55-6]: ACGIH TLV-TWA 10 mg/m³ [total aerosol; inhalable fraction; A4 (Not Classifiable as a Human Carcinogen); no Sk notation; no STEL]; OSHA PEL: No PEL (not listed in Z-1 for PG vapor; PNOR 15 mg/m³ applies as fallback for total nuisance dust/mist); NIOSH REL: No numerical REL for PG vapor.” When Cority attempts to evaluate the 3.2 ppm PO vapor sample against the PG aerosol TLV of 10 mg/m³, the unit conversion fails: propylene oxide vapor is measured in ppm (gas-phase concentration) while the PG TLV is for aerosol mist (mg/m³ condensed-phase). Cority’s unit conversion engine cannot perform ppm→mg/m³ conversion for a substance whose TLV is class-specific to aerosol (vs vapor), because PG is a liquid at ambient temperature (bp 188°C; VP 0.08 mmHg at 20°C) with essentially no vapor-phase TLV in the ACGIH system. Cority returns: “Propylene glycol [CAS 57-55-6] at 3.2 ppm. TLV: 10 mg/m³ aerosol. Unit architecture: measurement in ppm (vapor phase); TLV expressed as mg/m³ (aerosol phase). Conversion not applicable for this substance-form combination. Evaluation status: INCONCLUSIVE. No exceedance can be confirmed or denied. Notify database administrator.”

The EHS compliance record in Cority for the 39M 8yr loading operator: “Evaluation status: INCONCLUSIVE. No OSHA PEL exceedance (PNOR default). No ACGIH TLV evaluation (unit mismatch). No NIOSH REL (none for PG vapor). No corrective action required.” Actual state: OSHA PEL 100 ppm = 3.2% — COMPLIANT (coincidentally correct for OSHA, but for wrong chemical); ACGIH TLV A2 Sk 0.5 ppm = 640% — SEVERE EXCEEDS; NIOSH Ca REL 2 ppm = 160% — EXCEEDS; A2 carcinogen designation replaced by A4 (Not Classifiable); Sk dermal absorption notation absent; all carcinogen alerts absent. A worker exposed at 640% of the ACGIH carcinogen TLV receives an “INCONCLUSIVE” compliance result and zero corrective action, because a one-digit CAS transposition substituted GRAS propylene glycol chemistry for Group-2B propylene oxide chemistry.

The Sk Skin Notation: Why Air Monitoring Alone Fails for PO Maintenance Workers

ACGIH’s Sk (skin) notation for propylene oxide (added in the 2011 TLV revision) reflects a well-characterized physicochemical reality: PO can penetrate intact skin rapidly and contribute to systemic internal dose. The relevant parameters are: log Kow 0.03 (marginally hydrophilic, favorable for aqueous skin partitioning and transport through the stratum corneum barrier); low molecular weight (58.08 g/mol; small molecules <500 g/mol penetrate skin more readily than larger molecules); chemical reactivity as an epoxide (ring-opening reaction with nucleophilic groups on skin proteins and lipids facilitates retention and continued transport); vapor pressure 590 mmHg at 25°C (PO is a volatile liquid that wets skin on contact and simultaneously presents inhalation risk).

Nitrile rubber gloves (8-mil, standard industrial protection) are among the least effective materials against PO. ASTM F739 standard continuous contact permeation data for liquid propylene oxide against 8-mil nitrile: reported breakthrough time 5–15 minutes (depending on manufacturer and formulation; shorter breakthroughs documented for thinner nitrile). Neoprene and butyl rubber provide better but not complete protection. Under an actual pump maintenance scenario — a 20-minute task replacing a mechanical seal on a propylene oxide return pump at the loading station — the loading operator at Dow Freeport TX who is wearing standard nitrile gloves will experience PO breakthrough to skin within the first 10 minutes. During the remaining 10 minutes of skin contact with PO-wetted surfaces, dermal PO absorption from the wetted forearm skin and hand-back area contributes systemic dose.

The magnitude of dermal contribution during a seal replacement task has been estimated by analogy with published PO dermal absorption studies. Dermal flux of PO through excised human skin has been reported in the range 0.5–2 µmol/cm²/hr under infinite-dose conditions. For a hand/forearm contact area of ≈ 200 cm² wetted with liquid PO over 10 minutes of breakthrough exposure: absorbed dermal dose ≈ (1 µmol/cm²/hr × 200 cm² × 10/60 hr) = 33 µmol = 1.9 mg PO. Over an 8-hr shift with one seal replacement task, the inhalation dose at 0.3 ppm TWA: (0.3×10&sup-;&sup6; mol/mol × 58.08 g/mol × 58.08 g/mol at STP correction) — simpler route: 0.3 ppm × 10.4 L/min breathing rate × 480 min = 1,497 ppm·L absorbed; at saturation assumption for small reactive molecule: ≈ 1.5–3 mg inhalation dose. Dermal dose adds ≈ 60–130% to the inhalation-only dose on maintenance days. Total dose equivalent to ≈ 0.5–0.8 ppm TWA (inhalation-equivalent) — at or above the ACGIH TLV 0.5 ppm on a dermal-corrected basis.

The EHS AI evaluating the 0.3 ppm air sample returns: “COMPLIANT: 60% of ACGIH TLV-TWA.” This is technically accurate for the inhalation route only. The Sk notation’s clinical significance is that when dermal exposure is substantial (breakthrough contact with liquid PO), the air-sample-only evaluation understates total systemic dose by 60–130%. OSHA’s 100 ppm PEL has no Sk notation, so OSHA-only compliance frameworks do not trigger any dermal evaluation. The ACGIH TLV with Sk notation says: “When significant skin absorption is expected, total exposure (air + skin) should be evaluated against the TLV.” No EHS platform in this portfolio implements total-dose ACGIH Sk evaluation that incorporates task-specific dermal exposure estimates alongside 8-hr TWA air samples. The Sk notation is displayed as metadata, but the compliance engine only evaluates air concentration. For a substance where dermal contribution on maintenance days can double the effective dose, this architectural gap produces routine compliance approvals for workers potentially exceeding the TLV on a total-dose basis.

NIOSH Ca REL 2 ppm and the 50× Invisible Third Tier

The NIOSH Ca REL for propylene oxide of 2 ppm (TWA, 10-hr) sits at 50× below the OSHA PEL of 100 ppm. This 50× gap is less dramatic than the 200× OSHA/ACGIH gap, but the Ca (potential occupational carcinogen) designation is NIOSH’s strongest carcinogen category: it encompasses substances that NIOSH has evaluated as potentially causing cancer in workers on the basis of animal data, epidemiological data, or both. The NIOSH Ca REL at 2 ppm was set in 1994 as a “lowest feasible” limit given engineering controls practicability at that time in polyurethane polyol and propylene glycol manufacturing facilities. The value was not revised when ACGIH reduced its TLV to 0.5 ppm in 2011; NIOSH has not published an updated Ca REL for PO since 1994.

In the LyondellBasell Channelview TX surface (actual 42 ppm), the NIOSH Ca REL is exceeded by 2,100%. In the Dow Freeport TX CAS-confusion surface (actual 3.2 ppm), the Ca REL is exceeded by 160%. In the McCormick Hunt Valley MD surface (actual 1.4 ppm), the Ca REL is not exceeded (70%). The Ca REL provides an intermediate compliance tier between the ACGIH TLV (0.5 ppm) and the OSHA PEL (100 ppm): exposures between 0.5 and 2 ppm exceed the ACGIH TLV but not the NIOSH Ca REL; exposures above 2 ppm exceed both. No EHS platform in this portfolio implements the NIOSH Ca REL for PO as a separate actionable alert distinct from the ACGIH TLV. Given that the NIOSH Ca REL is 4× more permissive than the ACGIH TLV for PO (unlike beryllium where ACGIH TLV and NIOSH Ca REL are both 0.2 µg/m³), the NIOSH Ca REL is not the primary risk indicator for PO; the ACGIH TLV is. But in stale-database platforms that show ACGIH TLV 2 ppm (pre-2011), the NIOSH Ca REL 2 ppm is numerically identical to the displayed ACGIH TLV — further obscuring the NIOSH carcinogen designation by making it appear redundant with the already-displayed ACGIH value. A worker at 1.8 ppm on a stale-database platform sees: OSHA 1.8%; ACGIH [2009] 90%; NIOSH Ca 90%: all below threshold, all green — when current evaluation would show ACGIH TLV 360% EXCEEDS.

Propylene Oxide Carcinogenicity: Direct Alkylation Without Metabolic Activation

Propylene oxide belongs to the chemical class of epoxides: three-membered ring structures containing an oxygen atom bridging two carbon atoms. The ring strain in epoxides makes them inherently electrophilic and reactive toward nucleophilic centers on biological macromolecules. Unlike most chemical carcinogens that require cytochrome P450 or other metabolic activation to form reactive intermediates, propylene oxide is a direct-acting alkylating agent: it reacts with DNA nucleophiles in its native form, without bioactivation. This direct-acting mechanism has two important implications for carcinogenesis assessment. First, there is no metabolic detoxification requirement before genotoxicity can occur; PO that reaches a cell nucleus can react with DNA directly. Second, PO does not depend on inter-individual variation in CYP2B6, CYP3A4, or other activating enzymes for carcinogenic potential; at equal exposure, all individuals face similar bioactivation probability. Compare to benzene, whose carcinogenicity depends on CYP2E1-mediated ring epoxidation and quinone formation: high-metabolizer CYP2E1 genotypes face disproportionately higher leukemia risk at equal air benzene concentrations.

PO alkylates DNA preferentially at N-7 guanine (forming N-7-(2-hydroxypropyl)guanine adducts) and at N-3 adenine. The N-7 guanine adduct is the predominant PO-DNA adduct in both rodent and human in vitro systems; it is hydrolytically unstable, releasing the alkylated base as a free 7-methylguanine analogue and creating an apurinic site (AP site) in the DNA backbone. AP sites, if not repaired by AP endonuclease before replication, can generate GC→TA transversion mutations. These mutations are consistent with the mutational spectra observed in rodent PO carcinogenicity studies. IARC Group 2B classification in Monograph Vol 60 (1994) was based on: (1) sufficient evidence in experimental animals (nasal epithelial adenocarcinoma and squamous cell carcinoma in rats on chronic inhalation; forestomach carcinoma in mice); (2) limited evidence in humans from two small retrospective occupational cohort studies; (3) strong mechanistic evidence as a direct-acting mutagenic epoxide. As of 2024, IARC has not re-evaluated PO for potential Group 1 upgrade, though the mutagenicity and rodent carcinogenicity evidence basis is stronger than for several Group 2A substances.

The “direct-acting without metabolic activation” mechanism is particularly relevant to the nasal epithelium as the primary carcinogen target in rodent studies: airborne PO vapor reaches the nasal mucosa during inhalation and alkylates nasal epithelial DNA before systemic distribution. The rat nose, with its extensive olfactory mucosa and high mucociliary clearance, concentrates inhaled reactive vapors; the proportionally larger human nasal cavity per body weight in rats vs humans means the rodent-to-human dose extrapolation for nasal tumors involves significant species-difference uncertainty. ACGIH’s A2 (Suspected Human Carcinogen) rather than A1 (Confirmed Human Carcinogen) designation reflects this extrapolation uncertainty. However, the direct-acting mechanism provides no plausible basis for a threshold below which epoxide-DNA alkylation does not occur: any PO molecule that reaches a guanine residue can alkylate it. This mechanism argues for ALARA (as low as reasonably achievable) in OEL setting, which is what the ACGIH 2011 revision to 0.5 ppm (vs the prior 2 ppm) reflects.

How Glyphward Detects the Three Propylene Oxide AI EHS Attacks

Glyphward’s multimodal scanner addresses the three propylene oxide attack mechanisms through distinct detection layers that text-only EHS AI systems cannot replicate. For the ÷100 decimal perturbation at LyondellBasell Channelview TX (Surface 1), the cross-document consistency scanner ingests both the OSHA 52 laboratory certificate (GC-FID: 42.3 ppm, with instrument run metadata and calibration standards confirming the value range) and the VelocityEHS exposure record (0.42 ppm, same sample ID and date). The two-order-of-magnitude discrepancy is flagged as a decimal-shift adversarial injection (confidence 0.96); the A2 Sk carcinogen exceedance suppression at 8,400% is escalated to critical priority. For the ACGIH 2011 stale-database attack at McCormick Hunt Valley MD (Surface 2), the regulatory currency validator compares the platform’s reported ACGIH TLV for CAS 75-56-9 (2 ppm) against the current ACGIH TLV database (0.5 ppm), identifies a 4× discrepancy, and flags the TLV field as a stale-database injection (confidence 0.93). The scanner then re-evaluates the 1.4 ppm sample against the current ACGIH TLV, revealing the 280% carcinogen advisory exceedance that the stale database suppressed. For the CAS transposition at Dow Chemical Freeport TX (Surface 3), the chemical identity cross-validator compares the substance name field (“propylene oxide”), the CAS field (57-55-6), and the retrieved OEL class (aerosol mist mg/m³; A4; no carcinogen). The chemical identity consistency scanner identifies the PO/PG CAS transposition (boiling points: PO 34°C vs PG 188°C; vapor pressures: PO 590 mmHg vs PG 0.08 mmHg; the A4 aerosol OEL is architecturally inconsistent with a room-temperature volatile vapor) and flags the CAS transposition as a chemical identity adversarial injection (confidence 0.91). The correct CAS 75-56-9 OEL (ACGIH TLV A2 Sk 0.5 ppm) is retrieved and the 640% exceedance is surfaced. For the Sk skin-notation total-dose gap (documented across all three surfaces), Glyphward’s task-activity integration module compares the maintenance task log (pump seal replacement: 20 min; nitrile gloves; liquid PO contact) against the air-only TWA sample, applies the ASTM F739 nitrile breakthrough model for PO, and flags the dermal contribution as a potential total-dose TLV exceedance requiring Sk-notation evaluation (confidence 0.84). See the Lakera alternative comparison for how cross-document, regulatory currency, and chemical identity validation distinguishes Glyphward from text-only scanners. Compare the stale-database mechanism to the welding-fume ACGIH 2019 stale-database attack at attack #399. Compare the CAS transposition mechanism to the gallium arsenide expressed-as-arsenic attack at attack #404. Compare the Sk skin-notation total-dose mechanism to the hexavalent chromium DSEN cascade suppression at attack #419: in both cases, the OEL compliance architecture correctly evaluates air concentration but is architecturally blind to a medical or dermal route of exposure that creates a separate, active compliance obligation.