Propylene oxide (PO, CAS 75-56-9, molecular weight 58.08 g/mol, normal boiling point 34.2 °C at 101.3 kPa) is a pressure-liquefied epoxide handled as a closed-loop feedstock rather than a solvent-like intermediate. Its water solubility near 40.5 wt% at 20 °C means moisture ingress does not remain as a separate aqueous phase but enters the bulk liquid. Propylene Oxide Water Content & Acidity Control: Why ppm Matters in Epoxide Handling is fundamentally a stoichiometric and catalytic-control problem. Each 100 ppm (mg/kg) water in PO can hydrolyze 322 ppm PO to propylene glycol under acidic or prolonged storage conditions, while acid at 20 ppm as acetic acid can consume basic catalyst or accelerate epoxide ring opening. Specifications for polyether-grade PO commonly require water content below 100 ppm and acidity below 20 ppm as acetic acid, with analytical control anchored to ASTM E203 for Karl Fischer water and ASTM D1613 for acidity in volatile organic liquids.

Downstream unit operations—KOH-catalyzed polyether polyol reactors, double metal cyanide (DMC) catalyzed alkoxylation, propylene glycol ether synthesis, and propylene carbonate reactors—interpret water not as an inert diluent but as the smallest difunctional initiator available. In storage, water also creates a solvent-rich boundary layer at tank heels where slow hydrolysis can accumulate propylene glycol. This shifts hydroxyl number, broadens molecular weight distribution, and changes the mass balance of both PO consumption and product assay.

How Does Water Content at 100 ppm Change Polyether Polyol Functionality and Molecular Weight Distribution?

In KOH-catalyzed polyether polyol production, a designed initiator such as glycerine or sorbitol provides the functional core; water entering with PO acts as an unplanned difunctional initiator. The reaction C₃H₆O + H₂O → C₃H₈O₂ consumes 58.08 g PO per 18.02 g water and forms 76.09 g propylene glycol. Therefore, 100 ppm water in PO corresponds to 0.1 kg water per metric ton, which consumes 0.322 kg PO and generates 0.423 kg propylene glycol. In a 25 m³ feed tank containing approximately 20.75 t PO at 20 °C and 0.830 kg/L, a 100 ppm water contamination level represents 2.08 kg water, consuming 6.69 kg PO and forming 8.77 kg propylene glycol. These masses appear small, but they appear at the low-molecular-weight tail of the polyol distribution.

KOH processes running at 105–120 °C and 0.4–0.6 MPa are sensitive to this water because propylene glycol contains two hydroxyl groups and therefore forms a residual diol backbone that cannot be removed by ordinary vacuum stripping. A triol product with target nominal functionality 3.0 and hydroxyl number 56 mg KOH/g by ASTM D4274 may show a broader molecular weight distribution when water-derived diol is present. The effect is measured not only by hydroxyl number drift but by extractable diol content and by the ratio of primary to secondary hydroxyl end groups. DMC-catalyzed alkoxylation can tolerate some water, but water still must be controlled because it limits the molecular weight ceiling of the polyol and alters the Brookfield viscosity at 25 °C. Published kinetic data for water initiation in a specific DMC catalyst formulation is limited, but the stoichiometric mass balance remains fixed.

Acidity in PO is reported as acetic acid because the titration endpoint of ASTM D1613 uses alcoholic potassium hydroxide and the result is expressed as mg/kg acetic acid. A 25 m³ railcar at 20 ppm acidity carries approximately 0.42 kg acid as acetic acid. In acid-catalyzed epoxide chemistry, even 20 ppm acidity is sufficient to promote ring opening and isomerization, particularly when water is present at the same low-ppm level. The protonation of the epoxide oxygen generates an activated intermediate that reacts with water to form propylene glycol; the same intermediate can rearrange to carbonyl byproducts such as propionaldehyde. These secondary products affect both downstream titration profiles and the odor of stored PO.

Storage and transfer equipment documents acid attack as a corrosion variable, but the larger process risk is catalytic. In KOH-catalyzed polyol lines, acetic acid neutralizes potassium hydroxide on a 1:1 molar basis. For 20 ppm acidity in 20 t PO, the acid load is 6.67 mol acetic acid, which consumes 0.374 kg KOH and generates 0.654 kg potassium acetate. That salt remains in the polyol phase and can influence conductivity, filtration, or downstream isocyanate compatibility in polyurethane formulations. In DMC-catalyzed processes, acidic species are a broader catalyst-poisoning class because residual acid groups can coordinate to active metal sites; published data for specific DMC catalyst activity loss at exactly 20 ppm acidity is limited, but incoming specification limits are set below this level to avoid non-linear productivity loss.

When Incoming PO Acidity Exceeds 20 ppm by ASTM D1613

Receiving a railcar or iso container with acidity above 20 ppm should trigger quarantine rather than immediate blending. Closed-loop sampling from a bottom sample assembly with a 316L stainless steel cylinder, purged with nitrogen at 30–50 kPa, provides the sample for ASTM D1613. If the titration value exceeds specification, the acid species are already circulatable through the tank farm, reactor feed lines, and eventually the catalyst bed or initiator solution. Blending with lower-acidity PO is not always allowed because acid distribution in a 25 m³ ISO container is not guaranteed uniform, and stratification near water-rich heels can create localized acid concentration above the analytical value.

Acid excursions above 20 ppm in PO feed can be managed operationally by reducing storage temperature below 10 °C to slow hydrolysis and by shifting feed to processes whose catalyst can accept a higher acid load. For KOH-catalyzed polyols, an additional KOH charge can be calculated from the acid value; however, the resulting potassium salt load must be checked against filtration and final conductivity specifications. For 20 ppm acidity in 1 t PO, the neutralizing KOH demand is 18.7 g KOH per metric ton, with 32.7 g potassium acetate produced. For DMC-catalyzed alkoxylation, such neutralization is not straightforward because the catalyst is not a soluble strong base and acid contact can cause permanent activity loss. In such cases, rerouting the contaminated PO to a distillation column with low-residence-time reboilers is the only robust corrective action. Published data for specific DMC catalyst poisoning by acetic acid at 20–30 ppm is limited, but the conservative specification exists because recovery is not assured.

Closed-Loop Water Analysis at 1 ppm Resolution in PO Transfer Lines

Water determination in PO uses sealed Karl Fischer titration cells because open sample handling introduces atmospheric moisture faster than the analytical signal. ASTM E203 provides volumetric Karl Fischer titration for water in liquids; ISO 760 is an equivalent framework. A sample loop constructed of 316L tubing with internal electropolished surface roughness below 0.8 µm Ra and dead volume below 50 mL minimizes hold-up and moisture adsorption. Coulometric Karl Fischer instruments can resolve water below 10 ppm, but sampling loss or air ingress typically dominates the measurement uncertainty at 1–10 ppm.

Online process analyzers require calibration against primary laboratory KF data. A dryer or molecular sieve trap on the nitrogen purge line must be monitored for breakthrough because ambient humidity at 60% RH can add moisture to a partially dried sample system at a rate that exceeds the specification limit of 100 ppm within minutes if seals leak. Closed-loop sampling is therefore not a convenience but a measurement boundary condition. For PO transfer lines, the sample point should be located after the pump and before the reactor inlet, with a return line to the storage tank to avoid releasing PO vapor to atmosphere.

Propylene carbonate routes using PO and carbon dioxide operate with water-sensitive mass balances because water consumes the epoxide to propylene glycol before carbonate insertion. In a catalytic carbonation reactor at 150–200 °C and 7–10 MPa, 100 ppm water in PO feed reduces selectivity to propylene carbonate by the same stoichiometric pathway observed in storage. Published data for specific catalyst systems is limited, but the byproduct propylene glycol accumulates in the carbonate purification train and must be removed by fractional distillation. In glycol ether production, water acts as a competing nucleophile: methoxide or ethoxide addition to PO is the target, while water addition produces propylene glycol. The resulting propylene glycol raises the bottom temperature of the downstream separation column and increases the load on the vacuum system.

Direct Acidity Neutralization Options and Their Limits in PO Feedstock

Aqueous caustic washing is not a straightforward in-line correction for acidic PO because the epoxide ring opens under basic conditions and the contactor can generate propylene glycol while creating an aqueous waste stream. Solid-bed acid scavengers such as activated alumina or amine-functional resins can remove some acid species, but they create a potential for localized exotherm if the bed channeling permits unreacted epoxide to contact acidic sites at high concentration. Published data for PO-specific adsorption capacity of these beds is limited. The industrial correction is therefore distillation or rerouting to an existing purification train with a structured packing column, low-residence-time reboiler, and overhead decanter designed for PO-water separation.

For acid levels just above specification, reducing the storage temperature to 5–10 °C and minimizing water content can slow acid-catalyzed hydrolysis long enough to use the material in non-catalyst-sensitive applications, but this does not remove the acid. Acidic species must be accounted for in any KOH-catalyzed alkoxylation charge by mass balance, and the resulting potassium acetate must be included in final polyol ash, color, and conductivity assessments. In DMC-catalyzed systems, the operational boundary is stricter: acid exposure is best avoided because catalyst activity recovery is not guaranteed by additional temperature or reaction time.

Managing Water Content Across Railcar Unloading and Tank Farm Headspace

Unloading PO through flexible hoses introduces moisture unless the line is dried with nitrogen and pressure-tested before transfer. Seal-less pumps, such as magnetically driven or canned-motor designs, eliminate mechanical-seal flush water from the product side. Nitrogen padding of storage tanks at 30–80 kPa with a supply dew point of -40 °C or lower limits headspace moisture ingress; however, breathing cycles during tank filling and emptying can still pull humid air if the pad system is undersized. A 25 m³ storage tank at 20 °C and 50% RH headspace holds only a few grams of water vapor, but repeated loading cycles without dried padding can accumulate water in the liquid phase over months.

Water can also concentrate in tank heels because propylene glycol-water mixtures are denser than PO. Draining and sampling the heel after each campaign is therefore part of water control, but heel drainage must be conducted in closed systems because PO is flammable and pressure-liquefied. A tank heel sample drawn from the bottom valve after a 24 h settling period can reveal water or glycol accumulation even when the bulk tank sample remains below 100 ppm. For polyether reactor feed lines, a continuous dry nitrogen sweep on the sample cabinet and daily verification against ASTM E203 are typical control actions.