Third-Party Inspection for Propionic Acid: COA, Retest and Dispute Resolution

Propionic acid (CAS 79-09-4, EINECS 201-176-3) is a C3 monocarboxylic acid with a molecular mass of 74.08 g/mol, a melting point of -20.8 °C, a boiling point of 141.1 °C, and a density of 0.993 g/cm³ at 20 °C. Two principal industrial routes—the Reppe carbonylation of ethylene, carbon monoxide, and water, and the oxidation of propionaldehyde—generate an impure product whose commercial value depends on controlling water, propionaldehyde, formic acid, and color bodies. Food-grade material is subject to 21 CFR 184.1081 and carries E 280 in the European Union; technical-grade material moves into calcium and sodium propionates, cellulose acetate propionate production, and herbicide intermediates. Because the acid is corrosive to aluminum, galvanized steel, and unlined carbon steel, production-scale storage and transfer rely on 316L stainless steel or lined carbon steel, with nitrogen blanketing to limit moisture pickup and headspace oxidation. Under these constraints, third-party inspection for propionic acid functions as a documentary and physical verification checkpoint, not as a substitute for a supplier’s process analytical data.

Impurity profiles in Reppe-produced propionic acid are route-specific. Carbonylation of ethylene under nickel carbonyl or palladium catalysis can leave trace nickel carbonyl decomposition products and formic acid; oxidation of propionaldehyde tends to leave unreacted propionaldehyde and acetic acid. These differences matter in a dispute because a GC-FID method that was calibrated only against acetic acid and propionic acid may not resolve ethylene glycol or heavy condensation products that appear in a specific production route. Published data for complete impurity profiles across different production units is limited; buyers with sensitive downstream catalysis should request a method-specific chromatogram and a quantified unknown peak policy rather than relying only on the sum-of-impurities calculation.

What Does a Propionic Acid COA Actually Quantify?

A certificate of analysis for propionic acid should specify whether the purity value is chromatographic purity, acidimetric assay, or both. GC-FID with a polyethylene glycol capillary column, typically a 30 m × 0.25 mm × 0.25 µm film, is used to separate acetic acid, propionic acid, propionaldehyde, and traces of unsaturated compounds. The acidimetric assay, based on titration with 0.5 mol/L sodium hydroxide to a phenolphthalein endpoint, gives total acidity and will count formic acid and other short-chain acids as propionic acid; therefore a COA that reports only acidimetric assay can mask a formic acid shift. The property table below lists representative limits and methods for a commercial propionic acid lot. Supplier specifications vary by grade, salt production, and downstream polymer tolerance, so the COA limits in a purchase contract should be read against the buyer’s actual unit operation.

Each of these values carries a defined measurement uncertainty. In an ISO/IEC 17025:2017 laboratory, a GC-FID purity result near 99.5 wt% may have an expanded uncertainty (k=2) of ±0.1 wt% to ±0.2 wt% depending on calibration strategy and internal standard purity. This uncertainty becomes decisive when a referee test falls on a specification boundary. Acidimetric assay uncertainty is usually narrower for total acidity, but its specificity is lower; the method cannot distinguish propionic acid from formic acid. Therefore the COA must identify the exact test designation and not use the word “purity” as a synonym for “assay.” For color, ASTM D1209 uses platinum-cobalt standards and a visual or instrumental comparator; the repeatability of visual APHA judgments is poorer than instrument-based readings, which is why some contracts require instrumental color only.

Certificate authentication includes confirming that the COA is issued from the supplier’s enterprise resource planning system, that the batch number on the drum matches the COA, and that calibration dates for the laboratory instruments are not expired. In one common inspection finding, a COA from a distributor may list a supplier’s original lot number but the drums contain re-drummed material from a different batch; because propionic acid is re-drummed to smaller packages for the feed and food sectors, the original COA may not reflect the actual product. Third-party inspection under ISO/IEC 17020:2012 should therefore compare the seal numbers, gross weights, and batch markings before sample collection.

Bulk liquid sampling under ISO/IEC 17020:2012 should begin with vessel recirculation or rollover. Propionic acid is a homogeneous liquid at ambient temperature, but dead legs in loading arms, bottom valves, and pump suction lines can hold stale material from a previous lot or cleaning solvent. A third-party inspector should document the loading line configuration, confirm that the line has been purged with the same product, and collect upper, middle, and lower samples through a PTFE-lined zone sampler if circulation cannot be performed. The primary sample is split into three retained portions under nitrogen: one for the supplier, one for the buyer, and one sealed referee portion held for the dispute window. ASTM E300 is commonly cited as the sampling practice for industrial chemicals; however, because the flash point of propionic acid is approximately 52 °C, inspectors must exclude open flames, hot surfaces, and non-bonded handheld equipment. Conductivity-bonded polyethylene or PTFE sample containers with vapor-tight caps prevent spillage and reduce moisture ingress during transit.

When Retest Triggers Arise From Water or Aldehyde Drift

Water is the most frequent retest trigger after drumming. Open-head steel drums with clamp-ring closures are common for propionic acid, but the gasket seat is a moisture-access point if the ring is not re-torqued after temperature cycling. A drum stored outdoors in a region with relative humidity above 60% can absorb enough atmospheric water to shift the water result from 0.12 wt% to 0.28 wt% over several weeks if the bung gasket is damaged. Aldehyde drift is less common in sealed containers but can occur when the headspace contains oxygen and the product is exposed to iron contamination from corroded steel. A buyer’s incoming QC result should not automatically overturn a COA. Before a retest, the laboratory should review the chain of custody, the sampling port condition, the container type, and the Karl Fischer coulometer’s drift rate. If the buyer’s sample was drawn from a transfer line that had been flushed with water, the retest must use a reserved split sample or a fresh sample pulled by a third-party inspector from the unmoved drum set. Published data for production-scale moisture uptake rates of propionic acid in damaged open-head drums is limited; field failure investigations in feed-grade preservation supply chains identify closure failure and water-flushed transfer lines as the two dominant root causes, not true assay loss.

Retesting should follow the same primary method stated on the COA. A buyer’s incoming GC result from a non-validated in-house method is not a valid basis for rejection if the contract designates the supplier’s ISO/IEC 17025:2017 method. For water, both laboratories should run duplicate coulometric Karl Fischer determinations on fresh sample bottles, with oven-assisted extraction if the acid shows hydration. For acidity/titration, the sodium hydroxide titrant should be standardized daily, and the phenolphthalein endpoint should be matched to a pH 8.1 blank. For aldehyde content, the hydroxylamine hydrochloride titration should be run within 24 h of sample collection because aldehyde values can drift with repeated septum punctures.

Contractual retest clauses for propionic acid are not harmonized across jurisdictions. Buyers should avoid accepting a retest right that excludes water or aldehyde parameters after a specified period. Because propionic acid is hygroscopic, a COA result for water in an unopened drum is generally valid only if the drum seal is intact; once the seal is broken, the supplier usually disclaims water-related retest obligations. Incoming QC should therefore measure water before any drum is opened for production, not after partial use.

Referee Analysis, Rounding Rules, and Contractual Retain Policies

Dispute resolution for a propionic acid COA should be structured as a three-stage sequence: sampling record reconciliation, referee split-sample analysis, and application of the method’s reproducibility limit. The contract should require the referee laboratory to be independent, to hold ISO/IEC 17025:2017 accreditation for the specific methods, and to test the sealed retain sample within 90 days of the original COA issuance. If the retained sample is stored at 5 °C to 25 °C under nitrogen and the container closure is intact, propionic acid assay and water values are generally stable during this window; published formal stability studies for propionic acid retain samples are limited, so the 90-day window should be accepted only if both parties agree in writing.

ASTM E29 governs rounding of the reported value. A result of 99.45 wt% is rounded to 99.5 wt% under the standard’s common rounding rule for a specification limit that is expressed to one decimal place. If the buyer rejects a lot because the gas chromatographic purity is 99.43 wt%, the referee result and its measurement uncertainty should be evaluated against the specification minimum, not against the original COA value alone. When the test method’s reproducibility limit is known, the difference between the buyer’s laboratory and the referee laboratory can be judged statistically; ISO 4259-1:2017 may be adopted for this purpose by contract, though it was originally developed for petroleum-related products. If the referee value differs from the original COA by more than the expanded uncertainty of the referee method and falls below the specification limit, the lot is non-conforming. If the values agree within the stated uncertainty, the original COA result stands and the buyer bears the referee costs unless the contract contains a cost-allocation clause.

In production-scale salt conversion, propionic acid is often metered through positive-displacement pumps into aqueous sodium hydroxide or calcium hydroxide reactors. The acid’s corrosive nature at temperatures above 60 °C increases the corrosion rate on 316L stainless steel; chlorine-containing deposits and stagnant aqueous phases can initiate pitting. Therefore third-party inspectors should not limit their review to COA parameters; a pre-loading survey should also document whether the storage tank has a bottom water leg where water and acid can separate after prolonged settling. Although propionic acid is fully miscible with water, a water layer can form when ambient temperature cycling causes condensation inside a vented tank. The presence of a separate aqueous phase may not alter the COA, but it can generate localized corrosion and contamination that later appears as iron or chromium in the product. For this reason, inspection reports should record tank nozzle orientation, heating coil condition, nitrogen blanket pressure, and the presence of any visible scale or sediment. Propionic acid is incompatible with strong bases, amines, oxidizing agents, and reactive metals; neutralization with sodium hydroxide is intentionally performed in salt converters but must be conducted with cooling because the acid-base reaction liberates heat.