Epichlorohydrin for Epoxy Resins: Grade Selection for BPA-Type and Novolac Systems

Epichlorohydrin (ECH, 1-chloro-2,3-epoxypropane) is the primary glycidylating reagent for liquid and solid bisphenol A-type epoxy resins and for epoxidized phenol-novolac resins. The molecular mass is 92.53 g/mol, the boiling point at atmospheric pressure is 116.1 °C at 101.3 kPa, and the flash point is approximately 31 °C. ECH produced by allyl chloride hypochlorination or glycerol dichlorination is delivered with trace water, iron, chlorinated acyclic by-products, and colour precursors. These impurities are not inert in the alkaline dehydrochlorination stage. For BPA-type systems, the main commercial target is the diglycidyl ether of bisphenol A with an epoxide equivalent weight near 180–190 g/eq; for novolac systems, the target is a multifunctional glycidyl ether with epoxide equivalent weight commonly in the range 165–180 g/eq. Grade selection therefore differs in the water, iron, and acidity limits that protect each polymer chain architecture.

Resin producers rarely rely on a supplier certificate alone. The quality gate is a laboratory cook test, in which the incoming ECH is reacted under a fixed ECH-to-phenol ratio and a standardized caustic feed profile. The resulting resin is tested for epoxide equivalent weight by ASTM D1652, colour by ASTM D1544, hydrolyzable chloride by ASTM D1726, and apparent viscosity by ASTM D2196. The test exposes interactions that isolated ECH specifications cannot show: wetness consumes caustic, iron catalyses colour formation, and acidity shifts the pH profile. A lot with high assay but elevated water can therefore fail the resin-side chloride limit even though the ECH certificate appears acceptable.

What Purity Ceilings Are Required for BPA-Based Liquid Epoxy Resin Manufacture?

Bisphenol A is converted to the liquid epoxy resin through a two-stage etherification–dehydrochlorination sequence. The stoichiometric demand is 2 mol ECH per 1 mol bisphenol A. Commercial liquid resin processes operate with excess ECH at a molar ratio from 2.5:1 to 10.0:1; the purpose of the excess is to suppress oligomer formation and maintain the n = 0 diglycidyl fraction. Under these conditions, water in ECH is the highest-sensitivity impurity. At water levels above 150 mg/kg, the oxirane ring opens to 3-chloro-1,2-propanediol and hydrogen chloride. The acid attacks the caustic charge, and the diol terminates the growing epoxy chain as a bound chlorohydrin. The final resin then shows increased epoxy equivalent weight, increased hydrolyzable chloride, and poorer filtration. Consequently, ECH for standard liquid BPA resin is commonly controlled at water below 100 mg/kg, colour below 10 Pt-Co by ASTM D1209, acidity as HCl below 20 mg/kg by ASTM D1613, and iron below 0.5 mg/kg by inductively coupled plasma optical emission spectrometry.

The iron limit matters because ferric species carried through the etherification step catalyse oxidative colour development during caustic addition. In a 20 m³ glass-lined batch reactor, an iron excursion to 1.5 mg/kg in ECH has been observed to move final resin colour from 20 APHA toward 60 APHA when the dehydrochlorination exotherm exceeds 62 °C. Such vessels are often operated at an agitator tip speed of 2.5–4.0 m/s, with the caustic feed introduced by a dip pipe below the liquid surface. For laminating-grade BPA resin, the ECH specification is tightened further: water below 50 mg/kg and iron below 0.2 mg/kg are typical, because residual chloride and iron can reduce insulation resistance after damp heat ageing. The resin-side electrical performance is evaluated by boiling-water extraction, not by ECH purity alone.

When the same merchant ECH is used for solid BPA-type epoxy resin with an EEW range of 400–1400 g/eq, the impurity effect changes. Solid resin manufacture intentionally uses a lower ECH excess to promote chain extension. A water level of 120 mg/kg in ECH can consume enough reactive epoxide equivalents to broaden the molecular weight distribution and reduce batch-to-batch melt viscosity reproducibility. Plants that finish solid resin in a 44:1 L/D twin-screw devolatilizer must then remove residual ECH along with the glycerin-derived chloropropanediol fraction. The devolatilizer ordinarily operates at 5–20 kPa absolute and a melt seal temperature not exceeding 180 °C. A broader molecular weight distribution increases the melt viscosity near the steam-side seal, slows volatile removal, and raises the residual ECH tail. Published data for this specific configuration are limited, but production troubleshooting identifies upstream ECH water as the first factor to verify when devolatilizer vacuum demand increases.

When Novolac Functionality Demands Tighter Epoxide Homolog Distribution

Epoxidized phenol-novolac resins are prepared from phenol-formaldehyde novolac backbones with phenolic hydroxyl functionality typically between 2.2 and 3.6 groups per molecule. Each phenolic hydroxyl requires 1 mol ECH and 1 mol sodium hydroxide. The reaction is run at 60–80 °C with an ECH-to-phenolic-OH molar ratio of 2.0:1 to 6.0:1. The higher viscosity of the novolac backbone reduces heat transfer and makes the dehydrochlorination stage more sensitive to localized alkali excursions. In novolac systems, the ECH assay can be slightly lower than for electronic BPA resins, but the wetness and iron ceilings remain strict. Water below 150 mg/kg is generally acceptable for non-electronic novolac production. Water above 250 mg/kg creates glycol ether structures that lower the effective glycidyl functionality and push the resin EEW above 210 g/eq. This is not a cosmetic shift; it changes the stoichiometric hardener demand and reduces the cured glass transition temperature in high-heat applications.

Because novolac epoxy resins contain multiple glycidyl groups per molecule, propagation during synthesis is amplified relative to difunctional BPA resin. A targeted EEW of 165–180 g/eq and a melt viscosity at 150 °C below 4.0 Pa·s may fall outside the acceptable injection-transfer window when the ECH feed contains excess water or iron. The iron limit is commonly held below 1.0 mg/kg; colour is relaxed to 15 Pt-Co because the novolac backbone itself contributes colour. The assay floor for novolac grade ECH is typically 99.85 wt% by gas chromatography using an internal standard. The specification still prioritizes low water and low iron over absolute assay, because the polymer chain has more reactive sites that can terminate as chloropropanediol groups.

Incoming ECH tanks at epoxy resin plants are often sampled at the unloading manifold for density by ASTM D4052, water by ASTM E203, colour by ASTM D1209, and acidity by ASTM D1613. The density range of 1.180–1.185 g/cm³ at 20 °C is a useful quick check for gross contamination by lower-density chlorinated solvents or higher-density heavy ends. The difference between the BPA-type and novolac column in the table is not a quality hierarchy; it is a consequence of different resin polymerization kinetics and downstream purification. A novolac producer may accept 150 mg/kg water because the reaction uses additional caustic and washing, while a laminating-grade BPA producer cannot tolerate the same wetness because of ionic cleanliness targets.

Impurity Transfer and Processing Window Limits in Caustic Dehydrochlorination

The dehydrochlorination step is the narrowest processing window for both resin families. For BPA-type liquid resin, the reaction is controlled at 52–58 °C. Below 50 °C the conversion rate per hour declines and cycle time increases; above 62 °C the rate of hydrolysis and epoxide homopolymerization increases sharply. This is a control cliff, not a linear drift. ECH water content moderates local caustic concentration because water enters the organic phase and dilutes the sodium hydroxide film around dispersed caustic droplets. Raising ECH water from 80 mg/kg to 160 mg/kg may require increasing the caustic concentration from 48 wt% to 50 wt% or reducing the caustic feed rate by approximately 3–5 % to avoid a batch temperature spike above 65 °C. Without this adjustment, gel particles can form at the caustic addition point and reduce filtration rates.

In novolac epoxidation, the temperature is kept at 60–80 °C, with the upper boundary set by gel nucleation near pH 14.0 at the alkali inlet. High-iron ECH worsens the control problem because iron chlorides can buffer the local pH and produce a lag between thermocouple readings and reaction progress. ECH storage is therefore specified with nitrogen blanketing when ambient humidity exceeds 60 %. If moisture ingress drives ECH water above 200 mg/kg, the tank should be isolated and predried, and the receiving lot reassigned to less moisture-sensitive resin production. ECH is incompatible with concentrated acids, strong bases in poorly mixed zones, aluminium, zinc, and amine-based additives, so the grade-transfer line must be cleaned and inerted before switching from commodity to electronic resin service.

These boundaries are empirical lot-transfer guides, not universal thresholds. The response of any specific reactor depends on the heat-transfer coefficient of the vessel, the caustic injection geometry, and the resin finishing train. For this reason, grade selection at the receiving laboratory should include a small-scale caustic addition experiment with temperature recording at 1 s intervals. A lot that produces a second temperature peak after caustic addition is often carrying a higher concentration of chlorinated acyclic intermediates, even when the gas chromatographic assay remains within specification.

Regulatory classification also shapes the grade-selection logic. ECH is a hazardous chemical with harmonized classification under CLP; it is inventory-controlled, and closed-loop unloading is standard. The resin producer’s operating permit often requires vapour-monitoring at the unloading station and restricted storage in tanks with a nitrogen pad. These restrictions do not change the resin chemistry, but they can make it preferable to select a higher-purity, low-water ECH that can be transferred without tank drying. Finished resins derived from ECH are not treated as ECH; however, residual free ECH is controlled by final stripping. Liquid BPA resins commonly leave the evaporator with free ECH below 10 mg/kg, and electronic grades are typically below 5 mg/kg. The finishing vacuum is coupled to the incoming ECH quality because water-derived chloropropanediols require additional stripping energy and wash water.

Incoming ECH specification used in resin manufacture includes gas-chromatographic assay, water by ASTM E203, colour by ASTM D1209, acidity by ASTM D1613, density by ASTM D4052, and iron by ICP-OES. For BPA-type liquid resin, the highest sensitivity belongs to water and iron; for novolac resin, the highest sensitivity belongs to water and acidity. For a receiving lot where ECH water is 180 mg/kg and acidity is 25 mg/kg, the production control chemist may choose to reduce the first caustic charge by 2 % and increase the post-wash water volume by 10 % to keep the finished resin hydrolyzable chloride below 300 mg/kg. This type of lot-based compensation is documented only when the plant operates with a validated cook-test protocol and a closed-loop temperature log.