BPA-Free Epoxy Alternatives: Bisphenol F, Novolac & Bio-Based Compared

Bisphenol A diglycidyl ether (DGEBA) remains the benchmark for ambient-cure and heat-cure epoxy systems, but formulators in can coating, tank lining, flooring, and structural adhesive segments now require BPA-free alternatives where migration limits, workplace exposure controls, or restricted substance lists apply. Comparative selection is performed against test standards rather than nominal substitution. The three candidate backbones are bisphenol F diglycidyl ether (BPF), epoxy phenol novolac (EPN) and epoxy cresol novolac (ECN) resins, and bio-based epoxy intermediates derived from cardanol, sorbitol, isosorbide, lignin, and epoxidized vegetable oils.

The evaluation uses epoxy equivalent weight (EEW) per ASTM D1652, dynamic viscosity per ISO 3219 and ASTM D2393, glass transition temperature after cure per ISO 11357-2, water absorption after 24 h immersion per ASTM D570, tensile properties per ISO 527-2 or ASTM D638-14, and chemical immersion per ASTM D543. Cured network crosslink density is reported from rubbery plateau storage modulus in dynamic mechanical analysis at 1 Hz using a dual-cantilever clamp.

Regulatory pressure for BPA replacement is not uniform across applications. Commission Regulation (EU) 2018/213 amending Regulation (EU) 10/2011 sets a specific migration limit for bisphenol A of 0.05 mg/kg in plastic food-contact materials and prohibits BPA in polycarbonate infant feeding bottles. BPF does not currently carry a harmonized SML under Regulation (EU) 10/2011; compliance for food-contact use therefore requires member-state authorization or end-testing under EN 1186 and EN 13130. In thermal paper, REACH Annex XVII entry 66 restricts BPA to less than 0.02% by weight. BPA-free status is not automatically equivalent to zero endocrine activity; BPF has been reported in receptor assays to exhibit estrogenic activity, so replacement qualification must include endocrine activity review where applicable. For metal can linings in the United States, FDA 21 CFR 175.300 provides positive listing but does not certify formulated systems; migration testing under 21 CFR 175.300 conditions of use remains mandatory.

What Makes Bisphenol F Diglycidyl Ether a Low-Viscosity Replacement for DGEBA?

Bisphenol F diglycidyl ether is manufactured by glycidylation of a mixed-isomer formaldehyde–phenol condensate. The resulting resin contains 2,2′-, 2,4′-, and 4,4′-methylene diphenol diglycidyl ether isomers under CAS 2095-03-6. Commercial grades are liquids at 25 °C with EEW typically in the range 160–175 g/eq, compared with 182–192 g/eq for standard DGEBA. Dynamic viscosity at 25 °C is commonly 2.0–5.0 Pa·s, which is approximately one-third to one-half the viscosity of a standard liquid DGEBA resin at 11–14 Pa·s.

The lower viscosity originates from the shorter methylene bridge and absence of the bulky isopropylidene group, which reduces molecular volume and enhances segmental mobility. This permits solvent-free formulation at application temperatures of 40–60 °C in plural-component airless spray units with 24:1 to 45:1 ratio pumps and internally heated static mixers. In ambient-cure industrial flooring, BPF-based primers frequently meet the viscosity threshold for direct application at 80% volume solids without reactive diluents, whereas DGEBA requires a solvent or reactive diluent to achieve equivalent wetting on concrete at 10 °C.

The cured network exhibits higher crosslink density potential because of the lower molecular mass between epoxy groups, but the aromatic ring content is lower than in DGEBA and water absorption after 24 h immersion per ASTM D570 is generally reported at 0.30–0.80% for amine-cured BPF versus 0.10–0.30% for DGEBA. This moisture uptake differential becomes significant for marine topcoats and immersion linings. In high-humidity application environments above RH 60%, pre-drying of aggregate and substrate moisture control below 4% by conductance is required to avoid amine blush and intercoat adhesion loss under ASTM D3359 crosshatch testing.

BPF-based epoxy systems have been applied in metal can linings where lower processing viscosity is needed for high-speed coil coating. Coating weight control at 8–12 g/m² dry film and bake schedules of 3–5 min at 200–220 °C are specified. After curing, methyl ethyl ketone double rub resistance per ASTM D5402 exceeds 50 double rubs in fully crosslinked films, but overbake above 230 °C generates yellowing and adhesion loss under ASTM D3359.

Solvent-free formulations based on BPF with cycloaliphatic amine adducts can generate shorter gel times than DGEBA at equal stoichiometric ratio, especially in mass above 200 g. Uncontrolled exotherm in a 500 mL cup may exceed 180 °C; therefore meter-mix equipment with chilled static mixers at 15–20 °C is used for high-volume dispensing to avoid thermal runaway. The low viscosity also permits higher filler loading up to 55 wt% barium sulfate or 30 wt% wollastonite, but filler loading above these ranges raises mix viscosity above 20 Pa·s and reduces spray atomization with conventional 0.019–0.023 inch spray tips. In filled systems, moisture absorbed by the filler must be controlled to less than 0.2 wt% before addition, because residual water reacts with oxirane and reduces crosslink density in the final network.

Novolac Epoxy Resins and High Crosslink Density Under Acidic Immersion

Epoxy phenol novolac and epoxy cresol novolac resins are produced by epoxidizing phenol- or cresol-formaldehyde novolac oligomers. The average functionality of commercial grades ranges from 2.5 to 6 oxirane groups per molecule, which creates a three-dimensional network with higher crosslink density than difunctional DGEBA or BPF. EEW values commonly fall between 170 g/eq and 190 g/eq for EPN and 195 g/eq and 230 g/eq for many ECN grades. These resins are semi-solid to solid at 25 °C and require heated transfer at 60–80 °C or dissolution in a low-boiling ketone solvent for application.

After cure with aromatic amines or anhydrides, glass transition temperatures of 180–220 °C are reported by differential scanning calorimetry at a heating rate of 10 K/min under nitrogen per ISO 11357-2. Chemical immersion under ASTM D543 in 10–25% sulfuric acid at 23 °C and 50 °C shows lower mass change and fewer surface defects than DGEBA and BPF networks. The high aromatic ring density reduces solvent ingress and water absorption after 24 h immersion to 0.05–0.20% by mass, depending on hardener and post-cure schedule.

The trade-off is mechanical brittleness. Tensile elongation at break for a 2,4-diaminotoluene-cured EPN system is typically below 5%, and pull-off adhesion to blast-cleaned steel may drop below 2 MPa on thermal cycling if the coating modulus exceeds 4 GPa and the substrate expands at a different rate. For tank linings subjected to acid splash or fuming, a 25–40% volume solids post-cure at 150 °C for 2 h is commonly specified to achieve full oxirane conversion; insufficient post-cure leaves unreacted epoxide at 5–10% and creates sites for acid penetration.

On production lines, EPN novolac coatings are applied with a heated plural-component airless spray system maintaining A-side and B-side reservoirs at 70 °C and 60 °C respectively. Static mixers with 24–32 elements provide low shear; pot life after mixing can be 15–20 min at 70 °C, requiring machine flush intervals below 10 min to prevent hard pack in the mix manifold. Vacuum de-aerating at 25–50 mbar during inline mixing reduces pinholes that otherwise become rapid diffusion paths in the immersion lining. Filler loading with 10–20 wt% micaceous iron oxide improves barrier properties but raises viscosity to 2,000–4,000 mPa·s at 70 °C; therefore spray tip size is increased to 0.023–0.027 inch and airless pressure maintained at 180–250 bar.

Sorbitol diglycidyl ether raises crosslink density through a bicyclic isosorbide core without bisphenol-derived aromatic rings

Bio-based epoxy resins are not a single chemical family. Epoxidized soybean oil and epoxidized linseed oil have oxirane oxygen contents of 6.0–7.0% and 8.0–9.0% respectively, corresponding to EEW values above 200 g/eq in many commercial grades. They are aliphatic, flexible, and hydrophobic, but their cured networks with amine hardeners typically exhibit glass transition temperatures below 40 °C unless blended with rigid aromatic or cycloaliphatic epoxies. Cardanol-based glycidyl ethers contain a C₁₅ unsaturated aliphatic side chain; viscosity at 25 °C is often below 0.30 Pa·s, which makes them suitable as reactive diluents, but the low aromatic content caps Tg at 40–80 °C with standard cycloaliphatic amines.

Sorbitol and isosorbide glycidyl ethers have higher oxirane functionality and a heterocyclic bicyclic structure. Commercial isosorbide diglycidyl ether EEW is commonly 150–180 g/eq, with viscosity of 5–20 Pa·s at 25 °C. When cured with anhydrides or aromatic amines, Tg values of 120–160 °C are reported, which places these systems closer to DGEBA than to vegetable-oil epoxies. The main obstacles are high water affinity, dark color, and batch-to-batch variation in pH, residual hydroxyl content, and oligomer distribution. Bio-based carbon content can be measured by ASTM D6866; values above 80% bio-based carbon are common for cardanol and sorbitol systems, while epoxidized soybean oil approaches 95–100% bio-based carbon.

Water absorption after 24 h immersion per ASTM D570 for isosorbide-based networks frequently falls in the range 0.50–2.00%, which is higher than EPN and DGEBA. In a high-humidity floor coating applied at 20 °C and 65% RH, this produces visible blush at the surface if amine hardeners are used without a blush-resistant formulation or infrared flash-off. Adhesion to steel in immersion service may decline from initial pull-off values above 3 MPa to below 1.5 MPa after 500 h of 40 °C distilled water immersion; published data for this specific configuration is limited, and each supplier lot must be re-qualified because the hydroxyl equivalent weight and residual alkali catalyst have been observed to shift by ±10% between production campaigns.

When Bio-Based Viscosity Exceeds 10 Pa·s at 25 °C, Meter-Mix Dispensing Requires Heated Reservoirs

High-viscosity bio-based glycidyl ethers such as certain sorbitol grades cannot be processed in standard diaphragm pumps or unheated static mixers without cavitation and ratio drift. Meter-mix equipment must maintain resin reservoirs at 50–70 °C to reduce viscosity below 2 Pa·s at the pump inlet. The hardener side may need to remain at 25–35 °C; differential temperature between A and B components larger than 15 °C creates viscosity mismatch and segregation in low-shear static mixers. Gear pumps with hardened steel internals are preferred over diaphragm pumps because the semi-solid resin at 25 °C can stall air-driven double-diaphragm units at cycle rates above 20 strokes/min.

Degassing is performed at 25–50 mbar for 10–15 min after preheating; moisture ingress above 500 ppm in the resin storage vessel extends the induction period and reduces glass transition by acting as a chain-transfer agent. At high filler loadings, a planetary mixer with 40–50 mbar vacuum and a butterfly stirrer running at 20–40 rpm prevents air entrapment that lowers dielectric strength in electrical potting applications. Gel time at 60 °C for a stoichiometric isosorbide–anhydride system may be as short as 15–25 min; production units therefore use static mixers of 12–18 elements and active cooling on the mix block to prevent exothermic advance in the disposable mixer.

In a solvent-free immersion lining formulated with an EPN novolac binder and 15 wt% micaceous iron oxide, application at a substrate temperature of 20 °C requires the A-side to be held at 70 °C and the B-side at 50 °C. The mixed material is applied by plural-component airless spray at a wet film thickness of 300–500 µm and back-rolled with a short-nap roller to release entrapped air. After a 7-day ambient cure and a 2 h post-cure at 150 °C, pull-off adhesion per ASTM D4541 exceeds 4 MPa on SA 2.5 blast-cleaned carbon steel, and dry film hardness reaches Shore D 85 per ASTM D2240. Acid immersion testing per ASTM D543 in 10% sulfuric acid at 40 °C produces a mass change below 0.5% after 1,000 h. The same formulation applied without post-cure undergoes blushing and interlayer delamination after 168 h of immersion, indicating that high-Tg novolac performance is post-cure dependent rather than intrinsic.