Epoxy Resin Vendor Qualification Scorecard: 12 Weighted Criteria

The Epoxy Resin Vendor Qualification Scorecard: 12 Weighted Criteria assigns a 100-point weighted total across 12 criteria that cover resin identity, ionic cleanliness, rheological consistency, cure behavior, mechanical response, and documentation integrity. A liquid bisphenol-A diglycidyl ether (DGEBA) resin with an epoxy equivalent weight of 184–190 g/eq and viscosity of 11,000–14,000 mPa·s at 25°C serves as the reference grade in most qualification panels; novolac and cycloaliphatic epoxies require modified windows because of higher functionality and different cure exotherms.

Supplier audits use ISO 9001:2015 clause 8.4.1 as the baseline control requirement, but the scorecard is product-specific. A vendor may pass a paper audit and still fail the weighted scorecard because of a hydrolyzable chloride value above the electronic-grade threshold or because batch-to-batch viscosity drift exceeds the permissible tolerance for a rotary-servo meter-mix-dispense line.

Novolac and cycloaliphatic resins are qualified with the same 12 criteria, but the viscosity method is replaced by softening point or elevated-temperature melt viscosity for semisolid grades. A solid epoxy novolac is not tested at 25°C because it is not a flowable liquid; instead, melt viscosity is measured at 150°C using a cone-and-plate fixture, and the acceptance window is supplier-specific.

What Are the 12 Weighted Criteria in This Supplier Scorecard?

The weighted criteria are grouped into chemical identity, physical consistency, reactivity, and documentation. Each criterion is scored against an acceptance window rather than a preferred value; results outside the window trigger a corrective action request or rejection. The weights are not equal. EEW consistency and viscosity control together account for 27 of 100 points because these two properties drive stoichiometry and dispense ratio in production.

The scorecard is not a pass-fail average. A weighted score is calculated by multiplying each criterion score by its weight and summing the products. A resin that scores zero on hydrolyzable chloride because it exceeds the electronic threshold receives a maximum possible score of 90, but the electronics gate still rejects it. This configuration prevents a high EEW or viscosity score from masking an unacceptable ionic cleanliness profile.

Reactive diluents such as C12–C14 aliphatic glycidyl ethers reduce viscosity but modify EEW and crosslink density. Vendors whose resins contain more than 5 wt% reactive diluent must declare diluent type because flash point, volatility, and sensitization potential change. The volatile content acceptance window is therefore linked to resin composition, not a single generic threshold.

Color is scored by APHA per ASTM D1209. A standard DGEBA resin typically appears water-white to pale yellow; an APHA value above 75 indicates oxidation, contamination, or prolonged storage at elevated temperature. Color does not directly affect mechanical strength in filled applications but serves as a sentinel for process history.

Sampling follows ISO 2859-1 General Inspection Level II with AQL 0.65 for critical resin properties. A minimum sample of 500 g is taken from a homogenized drum after end-over-end rolling for 30 min because DGEBA can develop stratification when reactive diluents are present and storage has been static. The test plan requires splitting the sample into separate airtight vials for moisture, color, and reactivity panels to avoid moisture pickup between replicates.

Rheological Drift and Meter-Mix-Dispense Thresholds

Viscosity acceptance windows are invalid without precise temperature control. A 10°C increase from 25°C lowers the viscosity of a standard DGEBA resin by approximately 50–70%, depending on molecular weight and hydroxyl content. Incoming inspection therefore conditions resin samples in a water bath at 25 ± 0.5°C for 2 h before measurements are taken with a Brookfield RV rotational viscometer using a No. 5 spindle or with cone-and-plate geometry at 20 rpm.

On production lines using rotary-servo meter-mix-dispense equipment with static-mixer outlets, viscosity drift above ±5% from the qualification baseline can alter internal leakage in progressive-cavity pumps and shift the dynamic pressure drop across the mixing head. A resin-side change of 2,000 mPa·s is sufficient to produce off-ratio dispensing in systems without active pressure compensation. Vendor qualification therefore requires viscosity readings at three shear rates or spindle speeds, not simply a single-point value.

EEW control is scored separately. For a DGEBA resin with an EEW of 190 g/eq and a polyamine hardener with an amine hydrogen equivalent weight of 60 g/eq, the stoichiometric mass ratio is 3.17:1. An increase in resin EEW to 192 g/eq shifts the required ratio to 3.20:1; without correction, the cured network carries unreacted epoxy groups or amine-rich domains and exhibits lower glass transition temperature and increased water uptake. Perchloric acid titration in glacial acetic acid with tetraethylammonium bromide is the accepted method for EEW, reported to 0.1 g/eq.

Storage history is also a rheological variable. Epoxy resins advance during storage when exposed to temperatures above 40°C for extended periods. Advancement increases viscosity and EEW; a vendor lot held in an unheated warehouse may show viscosity at the upper acceptance limit but remain within specification at incoming inspection. The scorecard requires storage condition verification on the certificate of analysis and rejects lots with signs of crystallization or phase separation. DGEBA can crystallize below 20°C; re-melting at 50°C for 24 h is permitted only if the vendor verifies that EEW and viscosity return to baseline after cooling to 25°C.

On a filament winding line, a resin lot with lower viscosity than the approved range can drain from vertical mandrels before gelation; a higher-viscosity lot can reduce wet-out of glass roving and increase void content above 2%. These process failures are not captured by standard mechanical tests on clear-cast specimens alone, so the scorecard retains rheological criteria as independent weighted entries.

Hydrolyzable chloride in epoxy resin is released as free chloride during cure or under humid service conditions. In semiconductor packaging and multilayer printed circuit assembly, chloride concentrations above 300 mg/kg correlate with aluminum bond pad corrosion and conductive anodic filament formation. Qualification panels for electronic applications therefore apply a tighter window of ≤ 300 mg/kg, while high-purity resins are commercially available at ≤ 100 mg/kg. Total chlorine, measured separately, includes inactive organically bound chlorine; a value above 900 mg/kg can signal incomplete dehydrohalogenation or residual epichlorohydrin-derived species even when hydrolyzable chloride is low.

Ionic cleanliness is not limited to chloride. Sodium and potassium at single-digit mg/kg levels are critical for electronics because mobile ions accelerate electrochemical migration under bias. The scorecard does not assign a separate weight to cations but uses the hydrolysis and extraction methods in ASTM D1726 and ASTM D1847 as surrogates for total ion cleanliness.

When Hydrolyzable Chloride Exceeds 300 ppm in Electronic Encapsulation

If a candidate resin passes viscosity and EEW but fails hydrolyzable chloride at > 300 mg/kg, the entire lot is disqualified from electronics-grade applications even if the weighted score remains above 70. The requirement derives from observed failure timing: chloride-induced failure occurs after device assembly, not at incoming inspection. A documented production-line failure mode is delayed adhesion loss on copper leadframes after 85°C/85% RH aging; the mechanism is chloride migration to the metal-resin interface.

Qualification includes a 1,000-hour aging study at 85°C/85% RH per IEC 60068-2-78, with shear adhesion tested before and after exposure. Die shear testing follows MIL-STD-883 Method 2019.7, and a post-aging loss greater than 20% relative to time-zero is a rejection condition for semiconductor applications, irrespective of other scorecard results.

Cure behavior is scored by differential scanning calorimetry and gel time. A single heating ramp at 20 K/min from -50°C to 300°C identifies the glass transition of the uncured resin, the onset temperature of cure exotherm, peak exotherm, and total enthalpy. For a DGEBA resin with a dicyandiamide hardener and a substituted urea accelerator, a peak exotherm between 140°C and 160°C is typical; for an ambient-cure aliphatic amine system, gel time measured by ASTM D2471 is a more functional control because the DSC peak is broad and exothermic drift is more difficult to quantify.

Total enthalpy is recorded as a batch-to-batch consistency check. A change in total cure enthalpy of more than 10% from the control is treated as a formulation shift even if the peak temperature remains unchanged, because enthalpy is proportional to the number of reactive epoxy groups available. This check is particularly sensitive to contamination by non-reactive plasticizers or advanced resin fractions.

Glass transition temperature is measured on cured plaques rather than powder. The specimen is first heated to 250°C to erase thermal history, cooled at 20 K/min, and then reheated at 20 K/min. The reported Tg is the midpoint of the second heating curve. For a DGEBA/dicyandiamide system cured at 80°C for 2 h followed by 150°C for 1 h, a Tg lower than the control by more than 5°C places the lot on hold because it indicates incomplete incorporation of the hardener or resin advancement.

Mechanical property verification uses specimens cured in closed steel molds and machined or cast to ASTM D638-14 Type IV dimensions. Tests are conducted at 23 ± 2°C and 50 ± 5% RH after conditioning per ASTM D618. An unfilled DGEBA/amine system passes if tensile strength is between 50 and 70 MPa and elongation at break is at least 4%. Flexural modulus per ASTM D790-17 must fall between 2.5 and 3.2 GPa. Values below this window indicate incomplete cure, plasticization by moisture, or excess reactive diluent; values above the window may indicate over-crosslinking or low residual monomer content.

Batch-to-batch mechanical variance is checked with a minimum of five specimens per batch. A coefficient of variation above 10% for tensile strength triggers a mixing and formulation audit because it may indicate incomplete hardener dispersion or operator-dependent mold release residue. Moisture in resin or hardener is controlled separately; for electronic potting, moisture above 0.05 wt% can create voids in dispensed shots during vacuum degassing, and Karl Fischer titration per ISO 15512 is used with sealed vials because DGEBA is hygroscopic.

Adhesion is not treated as a standalone criterion because adhesion loss in epoxy systems is usually a downstream consequence of chloride contamination, off-ratio mixing, or excessive moisture. Qualification therefore scores adhesion indirectly through tensile, flexural, and ionic cleanliness thresholds.

Batch documentation audits fail when the certificate of analysis omits chlorine speciation

Documentation is a control element rather than a paper formality. The scorecard assigns 10 of 100 points to criterion 12 because a missing chlorine speciation result can conceal an ionic contamination risk. A complete certificate of analysis for epoxy resin must include EEW by ASTM D1652, viscosity by ASTM D2196, hydrolyzable chloride by ASTM D1726, total chlorine by ASTM D1847, moisture by ISO 15512, and color by ASTM D1209. Each value is compared against the qualification window, not against supplier marketing ranges.

Regulatory documentation is evaluated in parallel. Suppliers must provide a REACH SVHC declaration under EC No 1907/2006 Article 33, a RoHS technical file declaration under IEC 63000:2018, and a safety data sheet compliant with REACH Annex II. Food-contact applications require a specific migration declaration under FDA 21 CFR 175.300 or EU Regulation 10/2011 when applicable. The regulatory audit matrix below is used for lot acceptance and is retained for batch traceability.

The scorecard is invalidated if any of the 12 weighted criteria has missing data, because incomplete documentation prevents verification of ionic cleanliness and cure behavior. Vendors are reevaluated at each significant formulation change or at a minimum interval of 12 months, and retained samples from each lot are stored for 24 months for comparative testing.