Commercial acrylonitrile leaving a propylene ammoxidation unit is rarely considered a single-component product. Even after extractive distillation and finishing, the monomer retains a distribution of low-molecular-weight compounds introduced in the reactor, quench, or storage. The acrylonitrile impurities that matter are water, acetaldehyde, hydrogen cyanide, and peroxides. These four classes interact with polymerisation chemistry in different ways. Water changes phase balance and solvent quality; acetaldehyde participates in chain transfer and colour-body formation; hydrogen cyanide is both an acute toxicological hazard and a redox-poisoning species; peroxides reduce induction periods and can trigger exothermic polymerisation. A monomer assay of 99.5 wt% does not by itself eliminate these effects if the polar impurities are distributed in the remaining 0.5 wt%.

How Much Water Is Too Much in Acrylonitrile Before Suspension Polymerisation?

Water enters acrylonitrile in the quench water absorber and, if azeotropic drying is incomplete, remains after the finishing column. At ambient temperature, acrylonitrile and water are partially miscible. Polymer-grade material is commonly specified at water contents below 0.5 wt%; however, fibre and electron-grade users often set internal limits below 0.05 wt% because residual water in solution polymerisation depresses the solvent quality of dimethylformamide dope and promotes nitrile hydrolysis. In suspension polymerisation, water is less detrimental to phase continuity, but it shifts the monomer-to-water ratio and affects particle porosity. A 10-tray azeotropic distillation column with a water-decanter removes the acrylonitrile-water azeotrope; the overhead temperature is near 71 °C. The accepted referee method for water is ASTM E203-24 volumetric Karl Fischer titration, using a formamide-methanol solvent to dissolve the sample without interference from nitrile side reactions. Oven-assisted coulometric Karl Fischer methods per ASTM E1064-24 provide lower limits of detection for moisture levels below 20 mg/kg. Production-scale batch data show that water content can vary by 0.02–0.10 wt% between top and bottom tank sampling points if the storage tank is not recirculated. This stratification is important when a fibre spinning line draws from the bottom nozzle and then receives a water-rich heel after tank switching.

Hydrogen cyanide is generated in the ammoxidation reactor and leaves the purification train as a light ends impurity. Its vapour pressure at 20 °C is approximately 85 kPa, so any residual HCN partitions rapidly into the tank headspace. That creates an exposure and corrosion burden in conservation vents and carbon adsorbers. The ACGIH ceiling limit for HCN is 4.7 ppm, and the OSHA PEL is 10 ppm as an 8-h time-weighted average. Beyond toxicity, HCN is a strong complexing agent for the ferrous iron in redox initiation systems. In nitrile-butadiene rubber emulsion copolymerisation, a residual HCN concentration above 1 mg/kg can delay initiation and broaden the molecular weight distribution, because cyanide ion binds the iron centre and reduces radical generation. Monomer specifications therefore test HCN by headspace gas chromatography with a nitrogen-phosphorus detector; the limit of quantification is typically 0.1 mg/kg. Caustic scrubbers on storage tank vents are operated at pH >11 to strip HCN from displaced vapour and convert it to sodium cyanide. Published data for the exact loss rate of HCN from uninhibited acrylonitrile during ambient storage is limited, but industrial tank-farm experience indicates that HCN concentration is highest in the first liquid withdrawal after railcar unloading because vapour-liquid partitioning concentrates it in the liquid layer. This operational boundary is why sampling valves are kept in heated enclosures and lines are blown with nitrogen before grab sampling.

Acetaldehyde’s Chain-Transfer Coefficient and Colour Formation in PAN Precursor Fibre

Acetaldehyde is a low-boiling aldehyde with a normal boiling point of 20.2 °C. It appears in acrylonitrile from incomplete oxidation of propylene and from oxidative degradation in reboilers when oxygen leaks into the finishing train. Because of its aldehyde hydrogen, acetaldehyde acts as a chain-transfer agent in free-radical polymerisation; a concentration of 20 mg/kg in monomer can measurably lower number-average molecular weight if polymerisation is not re-optimised. In polyacrylonitrile precursor fibre, acetaldehyde and its aldol condensation products cause yellowness index increases even when the residual aldehyde is not detected by routine purity gas chromatography. The analytical method typically uses a 30 m × 0.25 mm low-bleed cyanopropylphenyl dimethylpolysiloxane column with a flame ionisation detector and a split ratio of 50:1. The limit of quantification is near 5 mg/kg. A production-scale continuous polymerisation line processing acrylonitrile at multi-tonne hourly rates must therefore maintain acetaldehyde below 50 mg/kg to avoid batch-to-batch variation in fibre tenacity. In ABS emulsion polymerisation, acetaldehyde can also destabilise the latex by reacting with emulsifier amines, but the emulsion shear and stabiliser package exerts a stronger influence than trace aldehyde. Because published data for exact chain-transfer constants in acrylonitrile-acetaldehyde systems is limited, operators rely on molecular weight trending rather than a single kinetic constant.

Peroxide impurities in acrylonitrile arise from the slow autoxidation of the vinyl group when dissolved oxygen is present. This reaction is intentionally controlled but not eliminated when inhibitor systems based on hydroquinone monomethyl ether are used. MEHQ is typically added at 35–50 mg/kg in commercial acrylonitrile. The inhibitor requires a small concentration of dissolved oxygen to regenerate its inhibiting quinone form; therefore an oxygen-free tank is not always a stable tank. A dissolved oxygen concentration of 10–25 mg/kg is common in storage stabilisation, while the peroxide value measured as active oxygen is maintained below 1 mg/kg. The referee method for organic peroxides is ASTM E298-17a. When peroxide value exceeds 1 mg/kg, the induction period for uncontrolled thermal polymerisation shortens. This is especially dangerous in monomer recycle streams, where distillation can concentrate peroxides in the reboiler and deadlegs. Production-scale tank farms therefore apply recirculation pumps and avoid stagnant tees in piping that connects the storage sphere to the day tank. If a railcar arrives with a peroxide value above the specification, it is typically rejected or re-inhibited with an additional 10–15 mg/kg MEHQ under nitrogen. Field measurements indicate that peroxide formation is highest in half-empty storage spheres during summer months, because vapour-phase oxygen diffusion is faster at higher diurnal temperatures. Thermocouple arrays positioned at 2 m vertical intervals in a 500 m³ storage sphere have recorded a temperature difference of 3–6 °C between the top vapour space and the bottom liquid layer; that gradient is sufficient to induce convective circulation that moves oxygen-saturated liquid into the monomer phase.

When Peroxide Value Above 1 mg/kg Reaches a Nitrile-Butadiene Rubber Emulsion Line

Cold NBR recipes use a redox initiation system of ferrous sulfate, sodium formaldehyde sulfoxylate, and an organic hydroperoxide. If acrylonitrile feed already contains peroxides, the effective initiator concentration is shifted upward. A peroxide value of 1–2 mg/kg in the monomer feed is often enough to reduce induction time by 15–30 s in a 10 L pilot reactor and to increase the early exotherm. In production-scale trains fitted with external cooling and a recirculation loop, that shift appears as higher agitator torque during the first 10 min of polymerisation. The corrective action is not to reduce oxygen below the inhibitor threshold but to add an additional shortstop before monomer recovery. Shortstop systems based on sodium dimethyldithiocarbamate and hydroxylamine neutralise the extra radical flux. When modelling the process, the peroxide contribution is treated as an incremental radical input rather than as an impurity that can be ignored in the mass balance. Published data for this specific configuration is limited, but the qualitative response is reproducible in production records.

Distillation Train Configurations for Water, HCN, and Acetaldehyde Partitioning

Water forms a minimum-boiling azeotrope with acrylonitrile at approximately 71 °C and 88 wt% acrylonitrile in the organic phase. This means that a simple distillation cannot reduce water below the azeotropic value; an additional decanter or molecular sieve drying step is required for water below 50 mg/kg. Acetaldehyde and HCN are both low-boiling, but acetaldehyde reacts with HCN under alkaline conditions to form acetocyanohydrin. If the pH of a caustic wash exceeds 10, the cyanohydrin remains in the heavy phase and can degrade back to acetaldehyde and HCN in downstream reboilers. Therefore finishing units keep the wash pH in the range 8.5–9.5 and operate the light-ends column overhead at a reflux ratio of 3:1 to depress acetaldehyde below 20 mg/kg without sending excess acrylonitrile overhead. Data from distillation simulations show that the relative volatility of acetaldehyde to acrylonitrile is greater than 10 at 40 °C, so the separation is feasible; the practical problem is not volatility but reactive cyanohydrin formation.

Because the impurities require orthogonal analytical methods, a polymer-grade release usually combines Karl Fischer titration, gas chromatography, headspace analysis, and iodometric titration. Table 1 summarises the control levels and analytical methods applied at a continuous polyacrylonitrile precursor line. The table is not a specification but a cross-section of release and trending limits used in production.

Acrylonitrile containing these impurities cannot be dried by arbitrary solid adsorbents because acetaldehyde and HCN will compete for adsorption sites and may displace water. The recommended order of operations is to remove water by azeotropic distillation or molecular sieves first, then strip light acetaldehyde and HCN in a low-pressure column, and finally monitor peroxide value after inhibitor addition. Peroxide-forming monomer should not be stored in contact with copper alloys; production transfer lines are specified as 316L stainless steel with internal roughness below 0.8 µm Ra to avoid catalytic peroxide decomposition. Agitated storage tanks with nitrogen blanketing and continuous MEHQ injection are used when peroxide value is near 1 mg/kg active oxygen.