This assessment, China Epoxy Resin Capacity & Utilisation: 2026 Half-Year Review, examines operational capacity in Chinese epoxy resin production lines rather than trade-level consumption forecasts. The scope covers liquid diglycidyl ether of bisphenol A (DGEBA), semi-solid and solid DGEBA, brominated epoxy resin, and high-purity electronics grades. Capacity is evaluated on a dry-resin basis after water and solvent removal. Because provincial statistical verification for the first half of 2026 remains incomplete at the cut-off date, the review uses the most recent validated monthly run-rate data and labels all extrapolations as provisional. No nominal project announcement is treated as operational capacity without supporting steam, epichlorohydrin recovery, and wastewater chloride data.

Nameplate capacity in Chinese epoxy resin production is not a single fixed value. A reactor line designated at 50,000 t/year may be constrained by the salt filtration and epichlorohydrin recovery train to 43,000–46,000 t/year when high-purity epichlorohydrin or low-salt feedstock is not available. Utilisation is therefore calculated against the effective bottleneck capacity, not the reactor nameplate alone.

What Actually Limits Half-Year Operating Rates in Integrated and Non-Integrated Lines?

The operating rate is calculated as actual dry resin output divided by the bottleneck-adjusted nameplate capacity, normalised to the half-year reference period. Most liquid epoxy resin plants in eastern China use an annual streaming-hour convention of 8,000 h. After statutory maintenance, reactor cleaning, and salt filter replacement, the technical maximum is commonly 92–95% of nameplate. Three sections impose the binding constraints: epichlorohydrin recovery, brine washing, and vacuum stripping. The reactor itself is seldom the bottleneck. When throughput is raised, the wiped-film evaporator must remove residual epichlorohydrin and water without degrading the resin. Incomplete devolatilisation raises residual solvent and hydrolyzable chloride, which forces a quality downgrade.

Production-scale experience in Chinese plants indicates that pressure drop across the plate-and-frame salt filter increases more sharply than the throughput ratio. When the filter is operated beyond its rated cake capacity, brine washing becomes incomplete and sodium chloride carryover enters the resin phase. The subsequent vacuum stripping stage cannot compensate for sodium chloride carryover because salt remains insoluble in the resin matrix. This is a hard utilisation boundary, not a market signal.

Jiangsu, Shandong, and Zhejiang coastal corridors contain the largest concentration of Chinese liquid epoxy resin capacity. These clusters share chlor-alkali and phenol-acetone infrastructure. Units without captive bisphenol A exhibit higher operating-rate sensitivity during phenol-acetone price shocks. Public environmental compliance records from previous heating seasons show that temporary boiler emission controls in eastern China can reduce steam pressure below 0.8 MPa at the reboiler of a wiped-film evaporator. When steam pressure falls, distillation rate falls even if the reactor and salt filter are clear. The 2026 half-year review treats this as an external utilisation risk specific to centralised steam supply.

Shandong assets with captive chlor-alkali and epichlorohydrin feedstocks are better positioned for sustained run rates. Jiangsu units face tighter wastewater chloride limits and are more likely to run campaign-mode production, with line stops between grades. Campaign-mode operation reduces half-year utilisation by the same amount as a mechanical outage, even when no equipment failure occurs.

When ECH and BPA Prices Do Not Move Together, Run-Rate Decisions Follow Integration Depth

The variable cost structure of a non-integrated liquid DGEBA line is dominated by epichlorohydrin and bisphenol A purchases. Public process licensor data give typical consumption of 0.55–0.60 t epichlorohydrin and 0.65–0.68 t bisphenol A per tonne of liquid resin. Caustic soda and salt disposal add approximately 8–12% to variable cost depending on local brine treatment and wastewater chloride limits. When merchant epichlorohydrin prices rise faster than liquid resin netback, a non-integrated plant loses variable margin at run rates that would still be positive for an integrated producer. The review separates utilisation into integrated and non-integrated cohorts precisely because the same operating rate can have different economic meaning.

Published unit-level breakeven utilisation for Chinese non-integrated epoxy resin plants is limited. Trade-press surveys have placed the cash-cost breakeven above 70% during oversupply periods, but this varies with coastal logistics and mechanical vapour recompression brine recovery investment. Inland units face an additional 2–4% variable cost from liquid epichlorohydrin and bisphenol A transport, reducing run-rate flexibility. Brine recovery by mechanical vapour recompression lowers solid salt disposal cost, but the associated capital expenditure raises fixed cost and does not directly improve capacity utilisation.

EEW and Hydrolyzable Chloride Limits During High-Run-Rate Campaigns

High utilisation campaigns are audited against epoxy equivalent weight, rotational viscosity, hydrolyzable chloride, and moisture. ASTM D1652-11 is used for epoxy equivalent weight; ASTM D2196-20 for rotational viscosity; ASTM D1726-11 for hydrolyzable chloride. For standard liquid DGEBA, the epoxy equivalent weight control band is usually 180–190 g/eq, viscosity at 25 °C is 11,000–15,000 mPa·s, and moisture is kept below 1,000 ppm by Karl Fischer titration. When a line exceeds design stripping capacity, residual epichlorohydrin and saponifiable chlorine increase. Hydrolyzable chloride above 500 ppm disqualifies the batch from general liquid resin applications; above 300 ppm, it is excluded from most electronics laminate grades.

Solid DGEBA flake operations require pre-drying when ambient relative humidity exceeds 60%. Moisture absorption above 0.1% by mass shortens storage stability and reduces coatings clarity. Packing lines must be purged with desiccated air; otherwise moisture batch-to-batch variance becomes the effective utilisation limiter rather than reactor throughput.

Production of solid DGEBA by fusion advancement uses liquid DGEBA and bisphenol A in a nitrogen-sparged reactor at 160–180 °C. Target epoxy equivalent weight is typically 450–550 g/eq for standard solid resin and 800–950 g/eq for high-molecular-weight coatings grades. The reaction is monitored by viscosity and epoxy equivalent weight; overcooking causes gel particles and filter blockage. Batch cycle time is controlled by the heat-transfer coefficient of the reactor jacket, not by the epoxy addition rate. During high utilisation campaigns, solid resin units often accept longer batch times if the cooling water return temperature exceeds 35 °C, because the exotherm cannot be removed fast enough to avoid thermal runaway.

If brine washing is incomplete, solid salt carryover limits the next campaign. Sodium chloride particles act as nucleation sites in solid flake and can raise haze in solvent-borne coating formulations. Operators therefore monitor filter pressure drop and wash-water conductivity after each dehydrohalogenation batch. Wash-water conductivity above 5,000 µS/cm indicates insufficient salt displacement. At that point, the line must slow production regardless of order backlog. This conductivity threshold is a practical utilisation governor in Chinese liquid epoxy resin plants processing high-salt caustic streams.

Wind blade laminating resin qualification imposes additional constraints on mixed viscosity and gel time under vacuum infusion. Fabricators using glass-fibre-reinforced epoxy laminates typically require mixed viscosity at 23 °C below 250 mPa·s for infusion, with gel time between 60 and 120 min for full shell moulds. These parameters are checked with rotational viscometry and cup gel timers, not by the neat resin sales specification alone. During the first half of 2026, wind-grade liquid epoxy resin utilisation is assessed against narrow epoxy equivalent weight control around 186 g/eq rather than against reactor capacity in some coastal units. Published producer-specific validation data for this configuration is limited; most wind blade resin qualifications remain confidential between formulator and fabricator.

Electronics-grade epoxy resin for copper-clad laminates is purchased against tighter maximum sodium and chloride limits. Hydrolyzable chloride is typically specified at or below 300 ppm, and sodium ion content below 1 ppm for certain FR-4 applications. The 2026 half-year review treats this segment as separate capacity because reactors must be dedicated or thoroughly cleaned between general and electronics-grade campaigns. Cleaning cycles of 24–48 h between grades reduce effective half-year capacity even when reactor nameplate is unchanged. Published production data for individual electronics-grade epoxy resin units during the first half of 2026 is limited because laminate qualification confidentiality prevents line-by-line utilisation tracking.

Brominated epoxy resin utilisation is tied to copper-clad laminate and printed circuit board demand. The process charges tetrabromobisphenol A and requires strict control of epoxy equivalent weight and free bromide. Bromide salts catalyse dehydrohalogenation and can shorten reactor cycle time, but they also raise corrosion risk in vacuum systems. Operators measure corrosion coupons in the vapour line and track chloride and bromide release during each campaign. When gas-phase corrosion rates exceed 0.1 mm/year, turnaround intervals shorten and half-year utilisation falls below the mechanical limit. This operational boundary is included in the review as a materials-dependent constraint rather than a feedstock logistics constraint.