The industrial synthesis of the diglycidyl ether of bisphenol A (DGEBA) n=0 fraction is controlled at the point where the sodium salt of bisphenol A reacts with epichlorohydrin in a two-phase reaction mass. The n=0 molecule, 2,2-bis(4-glycidyloxyphenyl)propane, has a theoretical molar mass of 340.41 g/mol and an epoxide equivalent weight of 170.2 g/eq; the corresponding commercial liquid resins are not pure n=0 but rather oligomer distributions with epoxy equivalent weight specifications commonly spanning 182 g/eq to 192 g/eq and n=0 contents of 75–85% by HPLC area normalization. The excess of epichlorohydrin over bisphenol A performs two functions: it ensures that the phenoxide chain ends are capped as glycidyl ethers before they can attack existing epoxide groups, and it dilutes the organic phase so that the condensation product remains fluid enough to be washed and stripped. However, excess epichlorohydrin also creates a recycle burden, a hydrolysis sink, and a brine contamination pathway. The control problem is therefore not solved by a single molar ratio; it is solved by maintaining a stable ECH/BPA ratio at the reacting interface, by preventing accumulation of water and 3-chloro-1,2-propanediol in the recycle, by dosing caustic at a rate that neutralizes the liberated hydrochloric acid without opening the oxirane ring, and by stripping the recovered monomer below the temperature at which thermal advancement regenerates n≥1 homologues. Published process descriptions for the taffy route commonly place the ECH/BPA molar ratio between 5:1 and 15:1, with the lower boundary fixed by the onset of oligomerization and the upper boundary fixed by recovery energy and aqueous waste volume. The analytical requirements for n=0 isolation include epoxy equivalent weight titration according to ISO 3001:1999 or ASTM D1652-11, reversed-phase HPLC calibrated against a certified BADGE reference, hydrolyzable chloride determination, and ionic chloride measurement after water extraction. Each method captures a different consequence of excess ECH control: titration detects the average epoxide concentration, HPLC resolves the n=0 peak from n=1 and n=2 homologues, and chloride measurements detect the residues of epichlorohydrin hydrolysis and sodium chloride entrainment that reduce the electrical and curing performance of the isolated monomer.

The stoichiometric path to the n=0 adduct requires 2 mol of epichlorohydrin per 1 mol of bisphenol A, with 2 mol of sodium hydroxide consumed to dehydrochlorinate the chlorohydrin intermediate. In practice, the epichlorohydrin that is hydrolyzed to 3-chloro-1,2-propanediol does not dehydrochlorinate cleanly under the same conditions; the resulting diol can remain as a bound impurity or react further to glycidol, which can initiate polymerization or generate epoxy alcohols with lower epoxy values than the target. The excess ECH therefore must be high enough to compensate for hydrolysis, but the water introduced with caustic and the water generated in the condensation are the main sources of that hydrolysis. Industrial reactors operating with aqueous caustic and a separate organic epichlorohydrin phase are mass-transfer-limited at the interface, and the local water concentration at the interface is not represented by the bulk water content of the reactor. This interfacial complexity explains why two batches with identical bulk ECH/BPA molar ratios can produce different n=0 contents if the agitation rate, caustic concentration, or phase-transfer catalyst concentration is changed. A reactor equipped with a decanter and reflux splitter returns the upper organic or lower organic phase depending on density; the selection affects whether water-saturated epichlorohydrin re-enters the reaction zone, and this recycle composition drift is often the primary source of batch-to-batch n=0 variation on production lines.

What Limits the n=0 Fraction When Epichlorohydrin Is Recycled as a Water-Saturated Azeotrope?

Epichlorohydrin recovery from the condensation reactor is normally conducted by atmospheric or reduced-pressure distillation, and the recovered distillate is returned to the reaction after phase separation. Because epichlorohydrin and water form a minimum-boiling azeotrope, the distillate is not anhydrous epichlorohydrin; it is a two-phase mixture whose organic layer remains saturated with water at the condenser temperature. The water content of that organic layer is not inert. It participates directly in oxirane ring opening, forming 3-chloro-1,2-propanediol, and it reduces the rate of the desired Williamson etherification by diluting the caustic and altering the ionic strength of the aqueous phase. When the recycled epichlorohydrin is returned without a drying step, the water concentration in the organic feed rises until the reactor reaches a steady state in which the rate of hydrolysis balances the rate of water removal in the brine and the vent. This steady-state water concentration may be high enough to suppress the n=0 fraction even though the bulk ECH/BPA molar ratio is maintained at a high value. The effect is most pronounced in campaigns that recycle epichlorohydrin from a common storage tank without batch-wise compositional analysis. The water content of the recovered organic layer should therefore be measured by Karl Fischer titration, and if the value exceeds the specified threshold for the process, the recycle stream should be dried over molecular sieves or passed through a distillation column that rejects water as a separate phase. Published plant data for this specific configuration is limited, but the direction of the effect is consistent with the known hydrolysis sensitivity of epichlorohydrin under alkaline aqueous conditions.

In addition to water, the recovered epichlorohydrin can contain 3-chloro-1,2-propanediol and glycidol, both of which have boiling points or azeotropic behavior that make separation from epichlorohydrin incomplete in simple batch distillation. These hydroxy-functional impurities are partly soluble in the organic phase and can react with bisphenol A or with the growing DGEBA chain, creating mono-epoxide or higher-chlorine species that lower the n=0 purity. The recirculation of these impurities means that the true molar excess of active epichlorohydrin is lower than the gross ECH/BPA ratio calculated from the raw feed mass. A mass balance around the recovery loop, with HPLC or gas chromatography for oxygenated impurities, is therefore required before a plant can claim that a specific ECH/BPA ratio corresponds to a reproducible n=0 yield. The control action is to purge a portion of the recovered epichlorohydrin or to distil it through a side-stream that removes the intermediate-boiling hydrolysis products, while keeping the water concentration below the threshold at which the n=0 peak begins to decline. The ratio of ECH to BPA can then be expressed as the active ECH/BPA ratio, not the nominal ratio, and this active ratio is the variable that correlates with the n=0 fraction in HPLC data.

In the caustic feed step, the rate and concentration of aqueous sodium hydroxide introduced into the epichlorohydrin-rich organic phase determine whether the n=0 product remains monomeric or advances to higher molecular weight during the final washing and stripping steps. A caustic feed that is too concentrated, such as 50 wt% sodium hydroxide added rapidly at the start of the reaction, can hydrolyze epichlorohydrin at the liquid–liquid interface before the phenoxide intermediate has formed, producing a mixture of chlorohydrin and glycidol that reduces the effective epoxy functionality. A dilute caustic feed, such as 20 wt% sodium hydroxide metered over several hours, lowers the local hydroxyl ion concentration but increases the total water load, which must be removed later. The preferred profile in many taffy-process operations is a staged caustic addition in which the first portion is added after the bisphenol A and epichlorohydrin have formed a stable emulsion, and the remaining caustic is added incrementally while the reactor is held at a temperature that balances dehydrochlorination rate against oxirane hydrolysis. Temperature control in this stage is a threshold parameter because the rate of epichlorohydrin hydrolysis increases rapidly above 50°C under alkaline conditions, while the dehydrochlorination of the chlorohydrin intermediate is too slow below 40°C. The operating band of 40–50°C during the first condensation phase is therefore a narrow process window, and excursions above this range are typically visible in the HPLC chromatogram as a reduced n=0 peak and an enlarged n=1 peak after stripping. The use of phase-transfer catalysts such as benzyltriethylammonium chloride or tetrabutylammonium bromide can accelerate the interfacial reaction at lower temperatures, but the catalyst cation must be selected for caustic stability; quaternary ammonium salts degrade by Hofmann elimination at high temperatures and high hydroxide concentrations, and the decomposition products can contaminate the resin with amines that catalyze epoxy advancement during storage. Production-scale reactors with low-shear agitation may require a higher catalyst dose or longer residence time, and the resulting n=0 content may vary between identical vessels if the impeller tip speed and the emulsion droplet size distribution are not matched. This is a batch-to-batch variance source that cannot be corrected by adjusting the ECH/BPA ratio alone.

When Epichlorohydrin Is Stripped Below the Thermal Advancement Threshold

After the condensation and washing steps, the reaction mass contains excess epichlorohydrin, water, sodium chloride, and a mixture of DGEBA oligomers. The excess epichlorohydrin must be removed to meet residual monomer and volatile limits, but the stripping operation is also where the n=0 fraction can be lost by thermal advancement. If the resin is exposed to temperatures above approximately 140–150°C for extended periods, the free phenolic hydroxyl groups in the n=1 and higher oligomers can add to the oxirane rings of the n=0 monomer, forming additional n=1 and n=2 species and reducing the n=0 fraction. The stripping temperature and pressure are therefore coupled: epichlorohydrin has a normal boiling point of approximately 115–117°C, but the final traces require reduced pressure because the resin phase imposes mass-transfer limitations. A wiped-film or thin-film evaporator operated at 140–160°C and 1–5 kPa absolute pressure can remove residual epichlorohydrin while limiting residence time to minutes, but the wall film must be uniform to avoid hot spots. Rotor fouling occurs when sodium chloride fines or gel particles adhere to the heated wall, and these fouling regions can expose the liquid to much higher local temperatures than the bulk set point. The result is a decrease in n=0 purity despite a constant feed ratio. The downstream analytical consequence is an increase in epoxy equivalent weight from 170.2 g/eq toward 180 g/eq or higher, even though the pre-strip HPLC sample showed a satisfactory n=0 distribution. The process limit is therefore not the average evaporator temperature but the maximum film temperature and the residence time distribution in the hot zone. A short-path distillation unit with internal condenser and a feed degassing stage is often preferred when the target is a crystalline n=0 fraction, because it reduces the concentration of alkaline residues before the heat exposure step. If the resin contains free sodium hydroxide above 50 mg/kg or hydrolyzable chloride above 300 mg/kg, thermal advancement is accelerated, and the n=0 yield after stripping may fall below the crystallizer feed specification.

The pressure control system must also prevent air leakage; oxygen ingress at high stripping temperatures discolors the resin and can generate bound carbonyl species that affect the curing stoichiometry. In plants that recover epichlorohydrin by steam distillation, the water vapor raises the local water concentration and hydrolyzes the oxirane ring; the n=0 fraction declines while the diol content increases. Steam stripping is therefore unsuitable for high-purity n=0 production unless the condensed ECH-water mixture is dried before recycle and the resin is subsequently washed with acidified water to remove sodium ions. The preferred sequence is to strip the bulk of excess epichlorohydrin at moderate vacuum, then wash the resin with water or dilute phosphoric acid to neutralize caustic and extract sodium chloride, then perform a final devolatilization in a wiped-film evaporator under high vacuum. The order of operations matters because washing before the final strip removes salts that would otherwise catalyze advancement in the hot film, while washing after the final strip can leave water in the product and trigger haze or crystallization of the n=0 monomer in storage.

When the stripped resin is transferred to a crystallization vessel, the isolation of the n=0 fraction from the residual liquid depends on the selectivity of the solvent, the cooling rate, and the chloride content of the feed. The purified n=0 material has a published melting range of 42–44°C, which is sufficiently low to require controlled cooling but high enough to permit solid handling at ambient temperatures in most plants. Methanol and methyl ethyl ketone are cited in preparative procedures, but the exact yield and mother-liquor recycle behavior depend on the oligomer distribution and the residual chloride content. When the stripped resin contains a high n=0 fraction, the crystallization yield is determined mainly by the cooling rate and the seed crystal morphology; rapid cooling produces fine needles that occlude mother liquor and require longer filtration times, while slow cooling produces larger crystals but increases batch cycle time. The mother liquor from the crystallizer is enriched in n=1, n=2, and higher oligomers, and it can be blended into standard liquid epoxy resin only if the hydrolyzable chloride and sodium ion content are within the specification for that lower-value product. The crystallizer feed must be free of water and free of residual epichlorohydrin above the limit specified by local air and hazardous substance regulations, because these volatile impurities can plasticize the crystal surface and lower the melting point. The filtration step after crystallization is often the production bottleneck for high-purity n=0; the slurry viscosity, the crystal size distribution, and the residual solvent content determine whether a centrifuge or a pressure filter is acceptable. A nitrogen-blanketed pressure filter or basket centrifuge is used in some low-chloride operations because the n=0 cake is oxygen-sensitive and because the solid can melt if the jacket temperature exceeds 35°C. Published plant-scale data for this specific equipment configuration is limited, but the boundary conditions are set by the melting point and by the explosion limits of the solvent.

Catalyst Selection, Brine Filtration, and Hydrolyzable Chloride Limits

The selection of the phase-transfer catalyst and the removal of brine before the final stripping step interact with the ECH excess control because any ionic residue in the organic phase can catalyze epoxide–phenol advancement during thermal processing. Quaternary ammonium salts that are soluble in both phases accelerate the condensation of bisphenol A with epichlorohydrin, but they also increase the solubility of water in the organic phase and can be difficult to remove from the final resin. Benzyltriethylammonium chloride and tetrabutylammonium bromide are common, but their decomposition products can include tributylamine or benzyl chloride species that affect the color and the storage stability of the isolated n=0. The catalyst is usually added at a mole fraction below 1 mol% relative to bisphenol A, and the exact amount is adjusted according to the emulsion quality and the caustic concentration. Excess catalyst does not increase the n=0 yield once the interfacial area is saturated; it increases the residual ionic content and the difficulty of the subsequent water wash. Brine filtration and coalescer operation are therefore part of the excess ECH control loop, because the aqueous phase contains sodium chloride and sodium hydroxide that must be separated from the organic phase before the heat-sensitive n=0 monomer is stripped. A decanter with a residence time that is too short produces a hazy organic phase with dispersed brine droplets; those droplets carry sodium ions into the vacuum evaporator and create hot spots that advance the resin. Coalescers with hydrophobic fluoropolymer media or salt-tolerant hydrophilic media are used to break the emulsion, but the media must be compatible with epichlorohydrin and with the traces of caustic. The pressure drop across the coalescer should be monitored because gel particles from epichlorohydrin polymerization can blind the media, and a blinded coalescer allows salt carryover to the stripper. This is a production-scale failure mode that often appears as a sudden increase in ionic chloride and a simultaneous decrease in n=0 content after stripping.

The hydrolyzable chloride content of the isolated n=0 fraction is a critical specification for electrical and adhesive applications because residual chlorohydrin groups can release hydrogen chloride during cure or during long-term aging, causing corrosion of metallic contacts and degradation of adjacent polymers. High-purity n=0 grades for electronic encapsulation may require hydrolyzable chloride below 50 mg/kg and total chloride below 300 mg/kg, while standard liquid resins commonly allow hydrolyzable chloride up to 500 mg/kg or more. These limits constrain the epichlorohydrin excess control strategy because every unit of excess epichlorohydrin that is hydrolyzed becomes a source of chlorohydrin impurity, and the later dehydrochlorination to glycidol consumes additional caustic. The process must therefore minimize hydrolysis in the reactor, maximize the removal of sodium chloride in the water wash, and avoid introducing chlorinated solvents or chlorinated catalyst residues into the final product. The measurement of hydrolyzable chloride is typically performed by titration of the chloride liberated by reaction with alcoholic potassium hydroxide; total chloride is measured after combustion or after reaction with metallic sodium. The methods should follow recognized standard procedures where available, and the results must be reported with the specific method because the numerical values are method-dependent. The n=0 fraction should also be tested for sodium and potassium ions by atomic absorption or ion chromatography, because these alkali metal ions can catalyze the polymerization of the oxirane ring during long-term storage and reduce the shelf life below the expected value.

For specification testing, analytical monitoring of the n=0 fraction requires the parallel use of titration, chromatography, and rheological methods, because no single test can distinguish the n=0 molecule from its oligomers. The following table summarizes the relevant parameters and the standard methods most commonly referenced for liquid and solid DGEBA n=0 materials. The ranges in the table are specification envelopes derived from commercial technical data sheets and preparative literature, not from a single universal standard; a producer must validate the exact limits against the intended application and the regulatory framework.

If the isolated n=0 crystals are not protected from moisture during storage, the surface opens to hydrolysis and the epoxy value drifts downward even though the crystallizer discharge was within specification. Residual epichlorohydrin above 10 mg/kg in the solid phase can exude during storage and create a flammable vapor space, while moisture above 0.1 wt% can promote formation of white surface haze and reduce the epoxy value. Storage in sealed aluminum-lined bags under dry nitrogen is used when the n=0 material is destined for electronic-grade applications; the warehouse temperature should not exceed 30°C because partial melting at the crystal surface can fuse particles and create caking that complicates downstream weighing and dissolution. If the warehouse relative humidity exceeds 60%, the bags should be kept in a desiccated enclosure or the material should be pre-dried before use. The reheating of n=0 crystals for formulation must be performed slowly because localized heating above 50°C can initiate advancement through phenolic impurities, shifting the epoxy equivalent weight upward even though the product was originally in specification. The n=0 crystals should not be stored near amine curing agents or strong acids because exothermic curing or oxirane ring opening can occur even in the solid state if vapors migrate across the storage area. These handling limits are the final constraints on the excess epichlorohydrin control strategy: the monomer can be produced, stripped, crystallized, and filtered to a high n=0 purity, but the downstream value is lost if the storage environment allows water, residual alkali, or excessive heat to convert the n=0 fraction back into oligomers or chlorohydrin species.