Propylene Oxide Factory Audit Checklist: 20 Points for an Epoxide Plant

Propylene oxide (PO, CAS 75-56-9) is a low-boiling epoxide with a normal boiling point of 34 °C, a vapor pressure of 59 kPa at 20 °C, and flammable limits in air of 1.9 vol% to 36 vol%. Because PO is miscible with water and common organic solvents, the same volatile mass shifts between liquid, vapor, and by-product polymer depending on temperature and catalytic metal surfaces. The audit checklist is organized by process flow, from propylene feed through purification and storage, and is intended for commissioning, insurer, and plant technical authority walkdowns. It does not replace the site permission-to-operate procedure.

Thermal Degradation and Peroxide Accumulation in Distillate Receivers

Two thermal degradation mechanisms govern the audit of PO distillation and storage equipment: vapor-phase autoxidation to propylene hydroperoxide and acid-catalysed ring-opening to propylene glycol and polyether oligomers. The autoxidation rate is not linear with temperature; local wall temperatures above 40 °C on reboiler tubes can produce peroxide concentrations that are not detected by routine product acidity. In the rectifying section, iron oxide scale from carbon steel piping catalyses hydroperoxide decomposition. The audit therefore requires infrared thermography of the reboiler shell during operation, not only a review of the DCS tube-side outlet temperature. A hot spot with a temperature difference greater than 5 °C from the adjacent tube row is recorded as a fouling or staging anomaly. The reflux drum is checked for oxygen ingress using a paramagnetic oxygen analyzer span of 5 vol% O2 in nitrogen. Calibration records are reviewed against ASTM E681 flammability data for PO-air mixtures. Published data for specific peroxide accumulation rates in this equipment configuration is limited, but the audit checklist treats every stagnant vapor space below a rupture disk as a peroxide formation zone until nitrogen blanketing is confirmed.

What Limits Water Content in Polymer-Grade PO?

Water entry into the purification train is a primary audit point because PO reacts slowly with water to propylene glycol, contributing to acidity and increasing reboiler fouling. The water content in commercial PO is typically controlled below 50 ppm when measured by Karl Fischer titration according to ASTM E203. Audit checks include desiccant bed differential pressure, moisture analyzer drift at the product rundown, and the integrity of heat exchanger tube bundles that could introduce cooling water through pinhole leaks. A shell-and-tube overhead condenser with tube-side cooling water is treated as a high-risk interface; a pressure test frequency of 36 months or less is common. Published data for this specific configuration is limited, and the auditor verifies that field sample lines are heat-traced and flushed because cold spots can trap condensed moisture and produce non-representative readings.

In the HPPO route, the epoxidation reactor feed is a mixture of propylene, methanol, hydrogen peroxide, and water over a titanium silicalite-1 catalyst. The ratio of propylene to hydrogen peroxide is sampled at the mixer outlet rather than inferred from feed meters. A local ratio below 1.0 mol/mol creates free hydrogen peroxide that can decompose on the catalyst, releasing oxygen and generating temperature excursions above 85 °C. Ratios above 3.0 mol/mol increase the vapor load to the propylene recycle compressor and lower the partial pressure of methanol, which assists in wetting the catalyst surface. The reactor is operated in the liquid phase at 1.5 MPa to 3.0 MPa and 30 °C to 80 °C. The step-response curve of reactor outlet temperature against hydrogen peroxide feed rate is checked for dead time longer than 120 s, because excessive dead time indicates bypassing or channeling in the fixed bed.

Chlorohydrin plants require a different emphasis on the saponifier and brine system. The saponifier outlet pH and the excess lime ratio are recorded against the mass balance of propylene chlorohydrin isomers. If the brine leaving the saponifier contains more than 0.1 wt% organic carbon, the downstream epoxide recovery is penalized because organic chlorides form azeotropes and complex with water. The audit evaluates the saponifier agitator seal flush and lime slaker temperature because uneven slaking changes the available calcium hydroxide and displaces the saponification endpoint. Published data for this specific configuration is limited, but the relationship between pH and chlorohydrin conversion is well established in plant trials.

C3 Splitter Fouling and Pressure Drop Signatures

Fouling in the C3 splitter is evaluated by comparing first-stage pressure drop against the commissioning baseline at the same reflux ratio and propylene feed rate. The column may be packed with high-performance structured packing or sieve trays, and the acceptable pressure drop is site-specific. A rise of 0.02 MPa above the design value while reboiler steam flow remains constant is treated as a distributor or packing fouling event. Overhead propylene purity is measured by gas chromatography according to ASTM D2163; a value below 99.0 mol% with propane in specification indicates tray maldistribution or entrainment rather than insufficient reboil. The bottom stream is checked for propylene carryover above 0.5 mol%. Additionally, the feed coalescer differential pressure is checked, because free water in propylene feed leads to hydrate formation in expander or column internals and biases the C3 splitter mass balance.

When a Storage Tank Relief Valve Is Out of Service

When a storage tank relief valve is out of service, the audit verifies that the isolation is administratively controlled and that the remaining relief path is sized for the fire case with the out-of-service valve isolated. PO storage vessels are pressure-rated or low-pressure tanks; the normal operating pressure is maintained with a nitrogen pad at 3.5 kPa to 7.0 kPa gauge. Relief valve sizing follows API 2000 and the vessel code data plate under ASME BPVC Section VIII Division 1. The audit requires a recent capacity calculation, not merely a tag number, and checks for blanketing valve hysteresis by trending tank pressure during a simulated liquid withdrawal. Vacuum relief is checked independently, because PO vapor condensed by a sudden rainstorm can create a vacuum that collapses the tank. Tank liquid temperature is read from a bottom-mounted RTD and is maintained at or below 25 °C. Fire zone drainage under NFPA 30 is verified by ensuring that the dike valve is normally closed and that its position is visibly tagged.

Reviewing the 20-Point Audit Matrix

Seal pot temperature measurement often reveals latent hazards that are not visible in product specification. Insulation or steam tracing can maintain a PO seal flush line above 30 °C even when the storage tank RTD reads 20 °C. An auditor removes a small insulation section at the seal flush return and measures surface temperature with a calibrated infrared pyrometer. A deviation greater than 2 °C between the local gauge and the pyrometer indicates either an immersion-depth error or a failed remote transmitter. The same check is applied to sample line dead-legs because trapped PO has been found to contain peroxide species and iron oxide after restart.