Propionic Acid: Purity, Water and Colour—Which Spec Limits Matter for the Process

Propionic acid (CAS 79-09-4, E 280) entering a continuous methyl propionate esterification unit at 99.5 wt% assay may be accepted, while the same ISO tank container is quarantined at a cellulose propionate acylation train if water exceeds 0.05 wt% or Pt-Co colour exceeds 10 APHA. This split is not a contradiction. Assay by acid–base titration reports total acidity, not molecular propionic acid, and cannot distinguish co-titrated acetic, formic, or butyric acid. The purchasing specification is therefore an interaction of three constraints—purity, water, and colour—and the controlling limit changes by unit operation.

Three specification limits are not independent. A high-purity lot by titration can carry water at the upper end of the acceptable range, and a low-water lot can still fail colour if carbonyl condensation has begun. The receiving laboratory should therefore apply all three tests to the same sealed-line composite sample. Split samples, open-drum sampling, and tank-top grab sampling introduce bias; a 0.03 wt% water difference is analytically small but operationally large when the downstream unit is a fixed-bed acylation reactor.

Impurity Fingerprinting Beyond the Total-Acidity Assay

Under the JECFA monograph and FCC provisions for food-additive propionic acid, assay is expressed as 99.5% minimum by weight, with water not more than 0.2%, non-volatile residue not more than 0.01%, and aldehydes not more than 0.2% as propionaldehyde. These limits are identity and food-safety guardrails, not process specifications. A conventional acid–base titration with 0.5 mol/L sodium hydroxide reports total acidity; acetic acid, acrylic acid, and other carboxylic acids co-titrate. A lot that passes the assay by titration may still contain 0.3% acetic acid and 0.1% propionaldehyde. Direct-injection capillary gas chromatography with flame ionisation detection on a nitroterephthalic acid-modified polyethylene glycol column (30 m × 0.32 mm × 0.25 µm) separates propionic acid from acetic acid, propionaldehyde, and acetate esters and is used where speciation matters.

Water is a separate burden. Propionic acid is fully miscible with water; moisture is not visible as a separate layer and must be quantified by coulometric or volumetric Karl Fischer titration per ASTM E203-16 or ISO 760:1978. In continuous esterification, water is a reaction product, and every 1 kg of water in the feed adds removal capacity demand in the azeotrope column or molecular-sieve dryer. A feed water limit of 0.10 wt% is common for technical esterification, while 0.20 wt% is generally limited to feed preservation and simple salt formation where dilution is tolerated.

Propionic acid distillation range is 138.5–142.5 °C under the FCC identity specification. A wider initial boiling point can indicate low-boiling aldehydes or acids. A purity specification alone does not capture this because a sample with 99.5% total acidity and 0.15% propionaldehyde can still exhibit a low initial boiling point. For esterification, the distillation range is a secondary confirmation that light oxygenates are not entering the reaction column and consuming alcohol through acetal formation.

Water is introduced not only by the original producer but also during tanker loading, transit, and drum dispensing. A polymer-grade acid transferred through an unblanketed carbon steel line or a split-tank manifold can gain 0.01–0.03 wt% water in humid conditions, which is enough to shift a borderline esterification lot. Sampling from the top of a tank car without closed-loop circulation can underreport water because propionic acid and water are miscible but not fully homogeneous until mixed; this is an operational source of batch-to-batch variance.

If Feed-Grade Water Creeps Above 0.2%, the Failure Is Not Microbial

In grain and finished-feed preservation, propionic acid is applied as a mould inhibitor, either alone or in blends with formic and acetic acids. The antimicrobial mechanism depends on the concentration of the undissociated acid; propionic acid has a pKa of 4.87, and the undissociated species dominates at low pH. Water in a propionic acid formulation lowers acid concentration and raises equilibrium pH, shifting the dissociation toward the less active propionate anion. A water content of 0.2% by weight is negligible in a dosing tank, but accidental dilution from steam cleaning or tanker washout is not: once a blend falls below approximately 90 wt% total acid, stainless steel 316L tanks and centrifugal pump casings can show localised corrosion, particularly at heat-affected zones. The raw material specification for feed-grade propionic acid therefore retains the JECFA water maximum not because the antimicrobial effect fails at 0.21%, but because water is a de facto marker for poor handling and a predictor of downstream dilution risk.

Feed-grade acceptance also relies on 21 CFR 184.1081, which provides GRAS status for propionic acid in food with good manufacturing practice. The colour limit for feed-grade material is typically relaxed to 20 APHA; colour is not a direct antimicrobial parameter, but colour bodies can signal aldehyde condensation products that produce off-odour in sprayed feed and may deposit in transfer lines.

Neutralisation to sodium propionate or calcium propionate uses 50 wt% or 30 wt% aqueous base in a cooled stirred-tank reactor. In this operation, total acidity is the primary specification because the reaction stoichiometry is set by acid equivalents; however, the difference between propionic acid and acetic acid becomes critical during crystallisation. Acetic acid in the feedstock forms sodium acetate or calcium acetate, both of which have different solubility-temperature profiles than the corresponding propionates. A feed propionic acid with an assay of 99.5% by titration and an acetic acid level of 0.3% can produce a crystallisation mother liquor with acetate accumulation that reduces propionate crystal purity. For spray-dried calcium propionate, residual water in the acid feedstock is a minor heat-load term, but total water in the neutralisation mass balance controls dryer inlet temperature and product moisture; the acid water limit is typically held at 0.2% to avoid excessive evaporation demand.

Polymer-Grade Colour Limits Below 10 APHA Are a Kinetic Boundary, Not an Aesthetic Preference

Cellulose acetate propionate and cellulose propionate production imposes the tightest propionic acid specification. During esterification of cellulose, excess water hydrolyses propionic anhydride to propionic acid, altering the molar ratio of propionyl to acetyl substitution and reducing the degree of polymerisation. A water content above 0.05 wt% in the propionic acid feed is widely rejected for polymer acylation because the resulting substitution drift changes melt viscosity, moulding shrinkage, and impact strength. Colour is evaluated by platinum-cobalt scale per ASTM D1209-05(2019) or ISO 6271-1:2015; polymer-grade certifications observed in supplier documentation commonly list 5–10 APHA maximum, although the exact threshold is application-specific. A Pt-Co colour of 15–20 APHA may be invisible in a drum, but can correspond to yellowing and haze in optical-grade moulded articles. Trace iron is controlled below 0.5 mg/kg for the same reason; iron leached from carbon steel storage lines can catalyse oxidation and colour build-up, and is not captured by a total-acidity assay.

Water determination in propionic acid is not always straightforward by standard volumetric Karl Fischer titration because propionaldehyde and other aldehyde impurities can react with methanol in the KF reagent and release water, causing a drifting endpoint or falsely elevated water result. Coulometric titration with aldehyde-specific reagents or a headspace oven accessory is preferred for polymer-grade samples below 0.05 wt%. A 0.2 wt% water result on a sample with 0.2% propionaldehyde should therefore be verified by the oven method to avoid mischaracterising a dry lot as wet.

For cellulose ester production, the acid feed is often blended with acetic acid and acetic anhydride before entering the acylation reactor. Stoichiometrically, 18 g of water hydrolyses 130.14 g of propionic anhydride; a 0.05 wt% water level in 1,000 kg of propionic acid therefore introduces 0.5 kg of water and destroys approximately 3.6 kg of propionic anhydride. This is not a minor analytical artefact. It increases the anhydride charge required to maintain the target propionyl substitution and shifts the mixed-acid ratio.

Pt-Co measurement is performed on a filtered liquid sample; it does not detect particulate rust or haze. An unfiltered propionic acid lot may contain fine iron oxide particles but read below 10 APHA. For optical-grade cellulose ester processes, the raw acid should be filtered through a 0.45 µm membrane and checked for turbidity in addition to colour; otherwise a clear filtrate with low Pt-Co colour can still contaminate the polymer with insoluble iron species.

For continuous esterification of propionic acid with n-propanol over sulphonic acid-functionalised styrene-divinylbenzene ion-exchange resin, the feed water limit is typically set at 0.10 wt%, not because a 0.20 wt% water level stops the reaction, but because the recycle loop accumulates water and the equilibrium conversion can fall by several percentage points. The resin bed acid capacity is affected less by water than by basic nitrogenous impurities in the acid feedstock; even 50 mg/kg of ammonia or trimethylamine can neutralise surface acid sites and increase the required reactor temperature. Azeotropic water removal using cyclohexane or toluene is standard, and reboiler duty is directly proportional to the sum of water entering with the acid and water produced by esterification. PTFE-lined storage, 316L pumping systems, and dry nitrogen padding are used to preserve the water limit during transfer. Published data for this specific configuration is limited, but process simulations in supplier bulletins commonly show that reducing feed water from 0.2 wt% to 0.1 wt% lowers azeotropic column energy demand by approximately 5–10% per tonne of ester, depending on reflux ratio.

Because the esterification equilibrium constant for propionate esters is moderate, conversion is forced by removing water or using an excess of alcohol. Water in the feed acid is not the only source; alcohol and catalyst streams also carry moisture. A specification of 0.10 wt% on the acid alone is therefore a practical allocation, not an absolute reactor requirement.

Avoid Reading Pt-Co as a Substitute for Aldehyde Speciation

Propionaldehyde is the oxidation/hydroformylation-related impurity most often tied to colour. In propionic acid, aldehyde can undergo aldol condensation to form unsaturated carbonyls that produce chromophores; the resulting Pt-Co colour may remain low while gas chromatography shows 0.15% propionaldehyde. A single Pt-Co number is therefore a screening limit, not a replacement for aldehyde speciation. The JECFA aldehyde ceiling of 0.2% as propionaldehyde is set for food safety and identity, but for esterification it is often too loose because aldehydes form acetals with alcohols, generating off-specification headspace odours in propionate solvents and fragrances. In a 1,000 kg/h n-propyl propionate train, a 0.2% aldehyde load corresponds to 2 kg/h of propionaldehyde entering the reactor; downstream distillation must split this from the product regardless of the acid assay. Colour remains the first visible marker of oxidation during transport, but water and purity remain the quantitative limits.

The use of ASTM D1209 is limited to clear liquids; propionic acid is clear at room temperature. However, the colour test is not a quantitative impurity measurement, and two lots with the same Pt-Co value can have different aldehyde contents because the chromophore response of aldol condensation products is not linear. Gas chromatography with DNPH derivatisation or a specific aldehyde test is used when the downstream chemistry is aldehyde-sensitive.

Pharmaceutical intermediates and agrochemical actives using propionic acid as an acylating agent often apply a tighter internal specification than the food-additive monograph: water is held at 0.05–0.10 wt% by ASTM E203, colour below 10 APHA by ASTM D1209, and propionaldehyde below 0.05% by DNPH derivatisation HPLC. The reason is not that propionic acid is toxic at the food-grade limit, but that trace aldehydes can form hydrazone or oxime adducts with downstream amine intermediates and reduce yield or create genotoxic impurities. No universal pharmacopoeial monograph for propionic acid is applied across all jurisdictions; therefore purchasers rely on supplier certificates of analysis and multi-vendor qualification. The acceptable water limit is determined by the drying capacity of the synthesis vessel and the reaction solvent: for a Grignard or esterification step run at –20 °C to 5 °C, 0.10 wt% water may be sufficient to quench the organometallic reagent, which is why anhydrous acylation routes request 0.05 wt% maximum.

Process-specific limits should be written into the purchase specification as a matrix, not as a single grade. A single tank can meet 99.5% assay, 0.2% water, and 20 APHA colour, and still be unsuitable for a cellulose acylation line requiring 0.05 wt% water and 10 APHA colour. The purchasing document should name the downstream unit and the analytical method for each parameter rather than relying on the supplier's standard commercial certificate.

Propionic acid is corrosive and hygroscopic; specification limits are valid only if the sampling and transfer system excludes atmospheric moisture and iron contamination. ASTM D1209 colour can increase in a tank container after 4–8 weeks of warm storage if the vapour space is not inerted. Stainless steel 316L or polypropylene-lined tanks are used for bulk storage; carbon steel is generally avoided because trace iron accelerates colour formation. The flash point of propionic acid is 52 °C closed cup, so heating for viscosity control is limited and no specification adjustment can substitute for tank inerting. Water ingress through manway gaskets or breather vents is a recurring batch-to-batch variance observed in production-scale operations; a 0.2% water limit may drift to 0.3–0.4% during humid summer transfer if the receiving tank is vented to atmosphere. A 0.03% water bias may not be visible as free water, but it changes esterification column reflux and should trigger re-certification by ISO 760 before the acid is charged to a polymer acylation line.