Shell Phenol

Produced via the air oxidation of cumene to cumene hydroperoxide followed by acid-catalyzed cleavage in the presence of sulfuric acid, Shell phenol is recovered through a sequence of vacuum distillation stages operating at pressures between 10 kPa and 25 kPa and bottoms temperatures not exceeding 185 °C, a thermal ceiling imposed to prevent exothermic decomposition into quinone methide intermediates and high‑molecular‑weight tars that irreversibly foul structured packing in the refining column. The cumene oxidation itself is conducted in a cascaded bubble‑column reactor train maintaining dissolved oxygen below 100 ppm in the liquid phase to stay outside the flammability envelope dictated by the ternary phase diagram of cumene‑air‑water at 90–110 °C and 0.5–0.8 MPa. The crude phenol stream—roughly 97 wt% phenol, with acetol, mesityl oxide, α‑methylstyrene, and cumene as key organic impurities—undergoes an extractive distillation with a water‑entrainer system to knock down acetol below 50 ppm before high‑purity recovery. This front‑end distillation design, typical of world‑scale plants with nameplate capacities exceeding 400,000 t/a, fundamentally determines the baseline contaminant fingerprint that propagates into every downstream derivative, particularly those requiring optical clarity or food‑contact compliance.

The correlation between phenolic grade and permissible impurity thresholds is codified through a matrix of specification standards that govern industrial, resin‑grade, and polycarbonate‑grade phenol. While ASTM D2439‑20 defines three grades (A, B, and C) primarily by crystallisation point (minimum 40.6 °C for Grade A) and residue on evaporation, actual polycarbonate‑grade phenol traded between producers and bisphenol‑A plants often invokes supplementary limits for carbonyl compounds, iron, and sulphur content because even single‑digit‑ppm levels of isopropyl benzene hydroperoxide can poison the acidic ion‑exchange catalysts used in bisphenol‑A synthesis. A representative segment of these specification layers, matched with their respective analytical test methods, is provided below.

Failure to meet the carbonyl or iron ceilings in the polycarbonate column reliably manifests as elevated colour in molten bisphenol‑A and, downstream, as yellow‑index drift in optical‑disc‑grade polycarbonate extruded sheet. Storage of high‑purity phenol in carbon‑steel tanks, even when nitrogen‑blanketed, is limited to a maximum turnover interval of 72 hours because the diffusion of oxygen through the headspace nitrogen layer at 25 °C can generate 0.5‑2 ppm per day of additional carbonyl impurities via autoxidation, a phenomenon extensively documented in bulk phenol handling guidelines published by the Chemical Industries Association.

How Does the Phenol‑to‑Acetone Feed Ratio Govern Bisphenol‑A Isomer Distribution?

In the acid‑catalysed condensation of phenol with acetone to produce bisphenol‑A (4,4′‑isopropylidenediphenol), the molar ratio of phenol to acetone at the reactor inlet exercises a dominant influence over the thermodynamic and kinetic competition between the desired para,para‑isomer and the unwanted ortho,para‑isomer (2,4′‑BPA), while also dictating the propensity of acetone to undergo self‑aldol condensation to mesityl oxide and subsequent cyclisation products that accelerate catalyst deactivation. Stoichiometric theory demands 2 moles of phenol per mole of acetone, yet industrial fixed‑bed reactors employing a macroreticular sulfonated styrene‑divinylbenzene ion‑exchange resin as catalyst and promoter—typically crosslinked at 4–8 % divinylbenzene and operated with an exchange capacity of 4.8–5.2 meq/g dry resin—are routinely fed at a phenol‑to‑acetone molar ratio between 8:1 and 14:1. This enormous stoichiometric excess is not a kinetic necessity for the main condensation; rather, it serves to suppress the formation of Dianin’s compound (a chromane‑type spiroketal adduct) and to keep acetone solvated in a phenol‑rich phase that minimises unimolecular dehydration of the intermediate carbinol. At ratios below 6:1, the reactor effluent has been observed on commercial‑scale units to contain Dianin’s compound at concentrations exceeding 1,200 ppm, concentrations at which downstream falling‑film crystallisers suffer heel fouling because the adduct co‑crystallises with the 1:1 phenol‑BPA adduct and shifts the solid‑liquid equilibrium temperature at the scraper wall by more than 3 °C.

The liquid hourly space velocity (LHSV) through the fixed bed is maintained between 0.8 h⁻¹ and 1.5 h⁻¹, a window bounded on the low side by the onset of external mass‑transfer limitations (detectable as a flattening of the Arrhenius plot below 55 °C) and on the high side by mechanical attrition of resin beads at bed pressure drops exceeding 0.15 MPa/m. Temperature control is exercised via shell‑side tempered‑water circulation, with the jacket set point held at 65–75 °C to avoid irreversible resin sulfonic‑acid‑group cleavage, which accelerates at localised hot spots above 95 °C. Hot spots themselves arise because the heat of reaction—approximately ‑45 kJ/mol of acetone converted—is released in a narrow band at the reactor inlet, sometimes requiring a graded catalyst loading where the top 20 cm of the bed is diluted with inert ceramic balls to distribute the adiabatic temperature rise of 4–7 °C across a longer axial distance. Real‑time monitoring of the catalyst health is achieved by tracking the acetone breakthrough curve; a shift of the breakthrough front to less than 70 % of bed utilisation typically triggers an in‑situ regeneration cycle using a 5 wt% aqueous phenol‑sulfuric acid stream at 80 °C, which hydrolyses precipitated alkylphenyl ethers that otherwise block micropores below 20 nm diameter as confirmed by nitrogen physisorption (BET) analysis.

A comparative overview of catalyst systems encountered in legacy and modern BPA trains is set out below, where the 4,4′‑BPA selectivity is defined as the mass fraction of total BPA isomers isolated after the first crystallisation stage.

The shift from homogeneous HCl to heterogeneous resin catalysts eliminates chloride‑induced stress‑corrosion cracking in downstream stainless‑steel distillation columns (particularly at the reboiler crevices where chloride concentrates to 50‑200 ppm), but introduces an operational dependency on feedstock phenol purity that the HCl route could tolerate. Specifically, unreacted cumene hydroperoxide carried over from phenol purification—measured as “active oxygen” by an iodometric titration method per ASTM E298—attacks the resin matrix divinylbenzene crosslinkages at a rate proportional to the peroxide concentration raised to the 1.3 power, according to kinetic profiles published in ion‑exchange degradation studies. This imposes a strict incoming phenol specification of active oxygen below 1 mg/kg, routinely verified by an online amperometric probe installed downstream of the phenol feed pump.

In the production of glass‑fibre‑reinforced phenolic moulding compounds intended for commutator insulation in automotive starter motors, novolac resin is first prepared by the condensation of molten phenol with 37‑wt% formalin in the presence of oxalic acid dihydrate (0.3–0.8 phr) as catalyst. The reaction is conducted in a jacketed, anchor‑agitated reactor under total reflux until the free‑formaldehyde content, determined by hydroxylamine hydrochloride titration (ISO 11402), falls below 0.1 wt%. At this point, water and unreacted phenol are stripped under vacuum that is gradually reduced to 5 kPa absolute while the batch temperature is raised to 170 °C; the endpoint is controlled by the melt viscosity of a cooled sample measured on an ICI cone‑and‑plate viscometer at 150 °C, targeting 15–25 Pa·s for high‑flow‑length grades. The dehydrated novolac melt is then discharged onto a cooled flaker belt and crushed to flakes with a residual phenol content of 3–6 wt%, a range that plasticises the compound during later injection moulding but must not exceed 7 wt% because higher levels plasticise the cured network to the extent that the glass transition temperature, measured by differential scanning calorimetry at 10 K/min ramp, drops below 180 °C and fails the heat‑deflection requirement of ASTM D5948‑20 for type MPF‑I materials.

Compounding of the novolac flake with hexamethylenetetramine (HEXA) as hardener (typically 12–16 parts per hundred resin), chopped glass fibre (6‑12 mm strand length, 30–40 wt% loading), calcium stearate lubricant, and magnesium oxide acid‑scavenger is executed either on a heated two‑roll mill with friction ratio 1.1:1 at 90–110 °C front roll temperature or in a co‑rotating, intermeshing twin‑screw extruder with L/D of 40:1 and a barrel temperature profile ramping from 70 °C at the feed throat to 95 °C at the die. The critical processing hazard during compounding is the premature initiation of HEXA decomposition and phenol‑formaldehyde crosslinking if screw‑fillage in the extruder exceeds 60 % and local melt temperature exceeds 120 °C, a threshold that, if breached for more than 30 seconds residence time, produces incipient gel particles visible as specks in the final moulded part. Operators mitigate this by controlling the barrel cooling with water‑circulated jackets set to a maximum inlet water temperature of 85 °C and by maintaining screw speed below 200 rpm for screw diameters above 50 mm.

Injection moulding of the granulated compound occurs at barrel temperatures of 75–95 °C (feed to nozzle), mould temperatures of 165–185 °C, injection pressure of 100–140 MPa, and back pressure below 0.5 MPa to avoid fibre breakage. Under these conditions the HEXA releases ammonia and formaldehyde, which crosslink the novolac into a network with a crosslink density, νₑ, of 2.5–4.5 mol/m³ as determined by dynamic mechanical analysis (DMA) rubber‑plateau modulus at 200 °C according to ASTM D7028. The runner‑and‑gate system must be designed for cold‑slug‑well capture because the melt exhibits a spiral‑flow length of only 35–50 cm when injected into a standard 2‑mm‑thick spiral mould, a value that drops by 15 % if the dwell time in the barrel at steady‑state exceeds 5 minutes due to gradual advancement of the B‑stage conversion. This progressive reactivity is why frequent purging with a high‑melt‑index polystyrene grade is recommended during manufacturing interruptions longer than 10 minutes.

Liquid Epoxy Resin Synthesis and the Taffy Process Residue Dilemma

Liquid epoxy resin (LER) derived from Shell phenol travels through the bisphenol‑A adduct to a condensation reaction with epichlorohydrin (ECH) under alkaline conditions, but the choice between the conventional ‘taffy’ process and the epichlorohydrin‑excess ‘advancement’ process carries significant consequences for polymer‑bound chloride content and resin colour. In the taffy route, bisphenol‑A and excess ECH (molar ratio 1:6–1:10) are charged into a kettle along with a 20‑50 wt% aqueous sodium hydroxide solution dosed at a rate to maintain the pH of the aqueous phase between 9.5 and 10.5, preventing hydrolysis of oxirane rings into chlorohydrins. The reaction mass, which transitions from a two‑liquid‑phase system to a highly viscous, pale‑yellow semi‑solid (the taffy), is sheared by heavy‑duty Z‑blade kneaders with a torque overload protection set at 8 kN·m and jacket‑circulated water at 60 °C. At completion of dehydrochlorination, excess ECH is recovered by vacuum distillation down to 1–3 kPa at 130 °C, after which the crude resin is dissolved in toluene or methyl isobutyl ketone, washed with water to remove sodium chloride, and filtered through a 5‑µm ceramic membrane to achieve a residual hydrolyzable chloride content below 300 mg/kg when measured by potentiometric titration per ISO 4583. Residual chloride above 500 mg/kg acts as a latent catalyst poison in amine‑cured formulations, retarding gel time by up to 30 % and reducing crosslink density because sodium chloride crystals, precipitated during cure, nucleate stress‑concentration sites that reduce tensile elongation by 20–40 % relative to low‑chloride analogues.

The advancement process, in contrast, reacts a low‑molecular‑weight liquid epoxy resin (EEW 180‑190 g/eq) with additional bisphenol‑A in the presence of a phosphonium catalyst, typically ethyltriphenylphosphonium acetate at 200‑500 ppm, in a wiped‑film evaporator operating at 160‑180 °C under 0.5–2 kPa absolute pressure. The continuous removal of water drives the polyaddition to the target epoxide equivalent weight, for example 450‑500 g/eq for powder‑coating solid resins. Procurement specifications for Shell phenol intended for the advancement route must guarantee a total carbonyl number below 10 mg KOH/g because carbonyl‑capped phenol molecules terminate the growing oxazolidinone chain when reacted with isocyanate‑based curing agents in hybrid epoxy‑urethane systems.

Alkylation of phenol with branched nonene (propylene trimer) to produce technical nonylphenol under catalysis by a macroporous sulfonic acid ion‑exchange resin at 90–120 °C and a phenol‑to‑olefin molar ratio of 3:1–5:1 yields a positional isomer mixture where the nonyl group attaches predominantly at the para position (85–92 %) but with sufficient ortho isomer (8–15 %) that the alkylate pour point, measured by ASTM D97, remains below ‑25 °C—a requirement for its subsequent ethoxylation to nonionic surfactants that must remain pumpable in outdoor storage tanks at sub‑zero temperatures. The continuous stirred‑tank reactor used for this exothermic alkylation (∆H ≈ ‑65 kJ/mol) is designed with an external pumped recirculation loop through a shell‑and‑tube heat exchanger capable of removing 150 kW per tonne of olefin fed, and the loop return nozzle is positioned tangentially to prevent channelling of the recycle stream across the top of the resin bed. Environmental permitting of this operation within the European Union now requires monitoring of free nonylphenol content in the distillate overheads to demonstrate compliance with the restriction limit of 0.1 wt% (per REACH Annex XVII, entry 46), and the analyser of choice is an online liquid chromatograph with fluorescence detection (excitation at 230 nm, emission at 310 nm) calibrated against a certified reference material. Any batch found to contain nonylphenol ethoxylate (NPEO) above 100 mg/kg is diverted to a thin‑film evaporator where the ethoxylate oligomers are stripped at 220 °C and 0.1 kPa until the UV absorption at 277 nm of the bottoms product reaches baseline.

When Optical‑Grade Polycarbonate Demands Sub‑ppm Naphthalene Content

Pre‑polymer from the interfacial phosgenation of bisphenol‑A incorporates the entire impurity profile inherited from the phenol precursor, and naphthalene—introduced as a trace contaminant during cumene oxidation if cumene feedstock is not sufficiently refined—is one of the most destructive species because it co‑condenses with bisphenol‑A during chain extension to form naphthalene‑terminated end‑groups that absorb at 375‑420 nm and impart a measurable yellow hue to the injection‑moulded polycarbonate disc even at concentrations below 1 mg/kg in the polymer. Industrial fibre‑optic‑grade polycarbonate (e.g., for DVD‑ROM substrates) forces suppliers to test every cargo of Shell phenol for naphthalene by gas chromatography‑mass spectrometry (SIM mode, quantifying the ion at m/z 128 against an internal standard of d₈‑naphthalene) with a detection limit of 0.02 mg/kg and a shipping‑window specification of ≤0.2 mg/kg. The interfacial polymerisation plant itself is laid out with dedicated storage tanks, piping, and pump seals that exclude graphitic packing material (which liberates polycyclic aromatics under friction heat) and any thread‑sealing compounds containing naphthalene sulfonate plasticisers. A parallel constraint applies to iron; the ferric ion content in the recirculated aqueous sodium hydroxide brine must be held below 0.3 mg/L because iron‑phenol complexes that form in the interface during phosgenation produce a greenish‑brown discolouration in the polymer granules that cannot be fully masked by toners without elevating the haze value above 0.5 % as per ASTM D1003‑21.

Cyclohexanone Intermediate from Phenol Hydrogenation: Catalyst Poisoning Traps

The route from Shell phenol to caprolactam proceeds through liquid‑phase hydrogenation of phenol to cyclohexanol, catalysed by a supported nickel‑chromium or Raney‑nickel catalyst in a slurry‑bed reactor at 140–170 °C and hydrogen partial pressure of 2–4 MPa, followed by vapour‑phase dehydrogenation of the cyclohexanol to cyclohexanone over a copper‑zinc catalyst at 220–260 °C. Published data for this specific configuration indicates that the phenol hydrogenation rate drops exponentially as the sulphur content of the phenol feed rises from 1 mg/kg to 5 mg/kg, with the deactivation following a power‑law dependence of reaction order 0.6 based on exit‑gas hydrogen sulphide measurement. Consequently, a guard bed of ZnO extrudates (binder‑free, BET surface area > 25 m²/g) is inserted upstream of the hydrogenation reactor and changed out after the cumulative phenol throughput reaches 40,000 kg per kilogram of bed weight, unless an online sulfur‑chemiluminescence detector (ASTM D5453) triggers an alarm at the 3 mg/kg S breakthrough concentration. Side reactions producing dicyclohexyl ether become detectable in the cyclohexanone distillation train when the cyclohexanol conversion in the dehydrogenation unit exceeds 70 % per pass, because the copper catalyst active sites catalyse bimolecular ether formation from two cyclohexanol molecules at a rate proportional to the mole fraction of unconverted alcohol, hence a per‑pass conversion of 65–68 % is deliberately maintained by adjusting the air‑to‑feed ratio in the partial‑oxidation reboiler.