Cyclohexanone for Caprolactam & Nylon 6: Feedstock Guide

Cyclohexanone entering the caprolactam chain is not a commodity ketone but a monomer-precursor stream in which trace carbonyl, alcohol, and unsaturated impurities propagate into caprolactam and then into polyamide 6. Incoming inspection for caprolactam-grade material is typically conducted by ASTM D2192 gas chromatography, Karl Fischer titration per ISO 760, and Pt–Co colour measurement per ASTM D1209. A commercial supply specification of 99.8 % minimum ketone purity is common; the remaining 0.2 % is dominated by cyclohexanol, water, and acidic species. Variations in this residual fraction alter the pH profile of oximation, the UV transmittance of caprolactam, and the relative viscosity stability of the final polymer. For this reason, feedstock qualification is normally based not only on total ketone content but on impurity classes that survive oximation and Beckmann rearrangement.

Cyclohexanone is received in stainless steel or lined carbon steel tanks under nitrogen blanketing. Trace iron pickup above 0.05 mg/kg during storage can accelerate colour formation because the ketone undergoes acid-catalyzed aldol condensation. Storage temperature is limited to 30 °C in long-term tanks and 45 °C in short-term day tanks. A nitrogen pad purity of 99.5 % or higher is specified to avoid peroxide formation, because peroxides react during oximation and increase the consumption of hydroxylamine. Material handling limits are defined by a closed-cup flash point of 44 °C per ASTM D93. Process areas require electrical classification and nitrogen purging of storage tanks. The lower explosive limit is approximately 1.1 % by volume at ambient pressure; this boundary is not specific to the feedstock route but controls the design of oxidation and distillation vents.

Phenol Hydrogenation Route: Distillation Bottlenecks and Ketone Quality

In the phenol hydrogenation pathway, phenol is reduced to cyclohexanol over supported nickel catalysts at 120–180 °C and 0.1–0.5 MPa hydrogen pressure. The hydrogenation step is highly exothermic and is carried out in fixed-bed or slurry reactors with hydrogen recycle to control temperature. The intermediate is then dehydrogenated over copper-zinc or copper-chromium catalysts in a fixed-bed reactor operating at 220–280 °C. Equilibrium limits single-pass cyclohexanol conversion to approximately 80–90 %; unreacted alcohol is recovered and recycled. The ketone purification train typically comprises a rough vacuum column for low boilers, a product column, and a heavies stripper.

Because cyclohexanol and cyclohexanone form a close-boiling mixture, high-purity ketone requires a column of 40–60 theoretical stages and a reflux ratio above 5:1. Production-scale units often observe pressure drop excursions in the product column when traces of phenol form high-boiling condensation products. The overhead ketone is then polished by acid washing and water removal to achieve the caprolactam-grade specification. In some plants, the phenol route is preferred when low-ring-impurity ketone is required, but it transfers the purity burden to hydrogenation by-product control. Residual phenol in the cyclohexanone feed should be below 50 mg/kg because phenolic species produce colour-forming compounds in downstream caprolactam.

Dehydrogenation catalyst deactivation is observed as a progressive drop in cyclohexanol conversion at constant reactor inlet temperature. The usual cause is fouling by high-boiling condensation products formed from trace phenol and oxygen ingress. Operators compensate by raising the inlet temperature within the 220–280 °C window; beyond 300 °C, dehydrogenation selectivity falls because the ketone undergoes ring-condensation. The product column reboiler also accumulates heavy organics and must be designed with spare heat-transfer surface area of approximately 20 % to maintain throughput between cleaning intervals. Published data for exact fouling rates across licensees is limited, so the spare duty is set from campaign operating data rather than from first principles.

What Limits Cyclohexane Oxidation Selectivity in the Ketone-Alcohol Feed?

When cyclohexane liquid-phase air oxidation is used to generate the ketone-alcohol feed, the process operates intentionally at low hydrocarbon conversion. Conversion is held at 4–8 % in uncatalyzed or cobalt-catalyzed systems because cyclohexanone and cyclohexanol are more readily oxidized than cyclohexane. Selectivity to ketone-alcohol oil under these conditions ranges from 80–90 %. In boric acid–modified oxidation at 160–180 °C, the borate ester of cyclohexanol protects the hydroxyl group and permits conversion near 10–15 %, with selectivity above 90 %. The resulting mixed oil is then dehydrogenated over zinc oxide–calcium oxide catalysts at 350–450 °C to convert cyclohexanol to cyclohexanone.

A persistent production bottleneck in continuous oxidation trains is the formation of di-acids and hydroxy-acids that deposit on heat-exchange surfaces and reduce cooling efficiency. That fouling imposes a cleaning interval and must be controlled by limiting per-pass oxygen feed and by maintaining alkali neutralization of acid by-products. Cyclohexyl hydroperoxide is the primary oxidation intermediate, and its thermal decomposition determines the ketone-to-alcohol ratio. Ketone-rich feeds are preferred for caprolactam because they minimize the cyclohexanol load sent to the dehydrogenation reactor, but excessive residence time increases alpha-oxidation products that cannot be removed by distillation and that later appear as caprolactam UV impurities.

Oxygen mass transfer is another constraint. Commercial reactors use sparged air or oxygen-lean gas with oxygen partial pressure below the flammability envelope of cyclohexane/oxygen mixtures. Off-gas oxygen concentration is kept below 2 % to avoid explosive conditions. Reactor residence times are typically 30–60 min, but the selectivity loss can increase by 1–2 % for every 10 °C temperature rise above 170 °C in uncatalyzed oxidation. The ketone-to-alcohol ratio is often controlled between 1:1 and 2:1 to avoid excessive alcohol load in the subsequent dehydrogenation section.

The availability of cyclohexanone oxime governs the yield of caprolactam; therefore the oxime formation step is operated as a high-conversion liquid-phase reaction. In conventional hydroxylamine sulfate oximation, cyclohexanone is contacted with hydroxylamine sulfate and ammonia at 80–90 °C. Turbine-agitated reactors with external cooling loops are specified because the oximation reaction is exothermic and pH-sensitive. The endpoint is maintained in a weakly acidic range near pH 5–6, where residual free hydroxylamine is minimized. Residual ketone in the organic phase must be kept below 0.1 % by mass because unreacted ketone enters the Beckmann section and forms caprolactam impurities that reduce the permanganate absorption number.

In the alternative ammoximation route, cyclohexanone, ammonia, and hydrogen peroxide react over a titanium silicalite-1 catalyst at 70–90 °C and 0.1–0.3 MPa, achieving ketone conversion above 99 %. That route greatly reduces ammonium sulfate co-production but increases sensitivity to trace sodium and heavy-metal ions that shorten catalyst life. Feedstock sodium and potassium above 0.1 mg/kg are often specified as exclusion limits for ammoximation units. When conventional oximation is used, ammonium sulfate output creates an economic and material balance constraint; the salt is either sold as fertilizer or routed to sulfuric acid regeneration, but its market value is frequently insufficient to offset disposal logistics.

Phase separation in the oximation reactor is influenced by residual cyclohexanone concentration. The oxime forms a low-density organic phase that is continuously withdrawn; aqueous ammonium sulfate solution is sent to crystallizers. Because oxime is thermally sensitive, steam stripping is used to remove dissolved ketone from the aqueous phase before salt crystallization. Excessive ketone carryover into the crystallizer causes colour development in the ammonium sulfate by-product. In plants that recycle the aqueous phase, the accumulation of sodium and nitrate ions reduces oximation selectivity and can cause nucleation of oxime crystals in transfer lines.

When Oleum Strength Deviates During Beckmann Rearrangement

The rearrangement of cyclohexanone oxime to ε-caprolactam is conducted in oleum at temperatures between 80 °C and 120 °C. Oleum strength is maintained in the 20–30 % free SO3 range because lower concentrations slow the exothermic rearrangement and higher concentrations increase sulfonation side reactions. The oxime must be added gradually to avoid thermal runaway; reaction trains use loop reactors or cascaded stirred reactors with emergency quench systems. The product is a caprolactam sulfate solution that is neutralized with ammonia, generating 2.8–4.2 kg of ammonium sulfate per kilogram of caprolactam.

In the Beckmann step, the concentration of free SO3 is a critical control parameter. A deviation below the target window leaves unreacted oxime and produces nitrile by-products, while a deviation above it darkens the crude caprolactam and increases sulfate ester formation. On-line density and conductivity measurements are used to infer free SO3, with laboratory acidimetric titration as the reference method. In gas-phase Beckmann rearrangement over high-silica zeolites, published data show reduced ammonium sulfate formation, but industrial operation is less tolerant of high-boiling organic feed impurities that coke the catalyst. The conventional liquid-phase route remains the dominant process for existing caprolactam plants, and its control strategy must accommodate feed ketone variability.

The neutralization of caprolactam sulfate is another localized heat source. Ammonia injection is staged to keep the pH below 7 in the first stage and below 9 in the final stage; high pH accelerates hydrolysis of caprolactam to ε-aminocaproic acid, which increases losses. Ammonium sulfate crystal size is controlled by residence time and agitation because fine crystals carry mother liquor into the caprolactam extraction section. Poor wash efficiency leaves sulfate in crude caprolactam, and residual sulfate above 5 mg/kg in the refined lactam interferes with polymerization by causing haze and corrosion in VK tube feed systems.

Crude caprolactam from neutralization contains water, ammonium sulfate, and organic impurities. The refining sequence includes extraction with benzene or toluene, water stripping, oxidation with potassium permanganate or hydrogenation over supported nickel, and vacuum distillation. For nylon-6 polymerization, the caprolactam specification is tightly controlled. The permanganate absorption number according to ISO 8660 is typically required below 5, and UV absorbance at 290 nm by ISO 7059 is held at or below 0.05 AU. Hazen colour of a 50 % aqueous solution by ISO 8112 is maintained at 5 or lower. These values are directly affected by cyclohexanone feed impurities: unconverted ketone, cyclohexanol, and aldol condensation products formed during ketone purification can survive oximation and rearrangement as amino compounds or unsaturated chromophores.

The ketone feed specification is only the first boundary. In a multi-train caprolactam plant, identical cyclohexanone purity can produce different caprolactam UV numbers if the impurity composition differs, because gas chromatography purity does not distinguish between colour-neutral and chromophore-forming species. Ketone-derived by-products such as cyclohexylidene cyclohexanone and 2-cyclohexylidene cyclohexanone are suspected chromophore precursors; published data for their exact formation rates in commercial oximation feed systems is limited. Process laboratories therefore use UV absorbance sweeps from 260 nm to 320 nm on the cyclohexanone feed as an additional screen.

Hydrogenation polishing removes residual unsaturation. The unit typically uses a fixed-bed nickel catalyst at 60–100 °C and 1–2 MPa hydrogen pressure. This step reduces the permanganate number, but it cannot remove saturated amine impurities; therefore distillation cut points must be controlled at the front end. If cyclohexanone feed contains cyclohexylamine or aniline precursors, these species survive as volatile bases and are measured in caprolactam by ISO 8111. A volatile base value above 1 mg/kg can prolong the polymerization induction period and produce a resin with unstable melt viscosity.

Hydrolytic Polymerization in VK Tube Reactors: Feedstock-Derived Impurity Effects

Caprolactam is polymerized in continuous VK tube reactors with L/D ratios commonly specified from 20:1 to 40:1, operating at atmospheric pressure or slight overpressure and a temperature profile from 220 °C at the top to 260–280 °C in the lower zones. The reaction is initiated by 2–4 % water; ring-opening is followed by polyaddition, and the equilibrium monomer content at the outlet is 8–10 % by mass. The melt is discharged through spinning pumps and the monomer is recovered by vacuum stripping. Viscosity number is measured in 96 % sulfuric acid according to ISO 307 and is controlled by the addition of chain regulators such as acetic acid or hexamethylenediamine.

Feedstock-derived amine and unsaturation impurities function as monofunctional chain terminators. Their effect is seen as a reduction in the number-average molecular weight at constant regulator feed, increased melt flow variability under ISO 1133-1:2022 conditions, and lower extractable-free stability. Production-scale extrusion of polyamide 6 from such resin on twin-screw extruders with L/D ratios of 40:1 and vacuum devolatilization can still retain residual caprolactam above 0.3 % when the caprolactam feed did not meet the permanganate number specification. Therefore the cyclohexanone-to-caprolactam contract is not only a purity specification but a chain-growth stoichiometry limit.

Impurity effects in VK reactors are sometimes masked by the batching interval of caprolactam silos. A single out-of-specification caprolactam lot can be diluted with on-spec material, but the resulting melt flow shift appears only after the silo turnover. This lag complicates root-cause analysis and requires that feed tanks be sampled after every 24 h of operation. The molecular weight distribution also responds to water concentration; if water in caprolactam exceeds 0.15 %, the polymerization rate increases in the upper tube zones and can lead to unstable melt plug flow. Conversely, water below 0.05 % slows ring-opening and increases residual monomer in the unpolymerized stream. Both conditions are detectable by ISO 307 viscosity number and ISO 1133-1:2022 melt flow tests on dried resin.