Polycarbonate (PC) is an amorphous bisphenol-A-based engineering thermoplastic with density 1.20 g/cm³ (ISO 1183-1), glass transition temperature approximately 147 °C (ISO 11357-2), tensile stress at yield 60–70 MPa (ISO 527-2), tensile modulus 2.2–2.4 GPa, and notched Izod impact strength above 600 J/m (ASTM D256) for 3.2 mm specimens. These properties make adhesive joining attractive for transparent housings, glazing, medical devices, and lighting lenses, but the amorphous structure that provides impact resistance also makes PC susceptible to environmental stress crazing. The central selection problem is Gluing Polycarbonate: Which Adhesives Are Safe and Which Ones Craze the Surface. A solvent-free adhesive may produce an acceptable lap-shear value, yet still craze a stressed molded part if its monomer package or cure residues attack the polycarbonate surface. Qualification therefore requires evaluating both mechanical shear strength and the tendency of the uncured and cured adhesive to initiate microvoided craze zones.

Crazing is not straightforward dissolution. It is a localized tensile-yield phenomenon in which solvent or monomer penetration reduces chain entanglement density at a surface flaw, machined edge, gate vestige, or weld line. The craze consists of oriented polymer fibrils separated by microvoids; it scatters light and reduces load-bearing cross-section before a true crack forms. Adhesives containing ketones, aromatic hydrocarbons, chlorinated hydrocarbons, or low-molecular-weight esters can initiate this response at outer-fiber strains below 1.0% when molded-in stress is present. The relevant test is ISO 22088-3, in which a pre-strained PC strip is exposed to the adhesive or its volatile components and examined for craze formation.

Does Solvent Solubility Parameter Predict Adhesive-Induced Crazing in Polycarbonate?

The Hildebrand solubility parameter of PC is approximately 19.8 MPa^1/2. Solvents with δ values between 18 and 21 MPa^1/2 and small molar volumes penetrate the surface readily: methylene chloride is approximately 20.2 MPa^1/2, acetone 19.9 MPa^1/2, methyl ethyl ketone 19.0 MPa^1/2, and toluene 18.2 MPa^1/2. Adhesives formulated with these solvents, or that release them as cure by-products, are high risk for polycarbonate. Aliphatic hydrocarbons with δ below 16 MPa^1/2 are weaker swelling agents, but they may still carry residues that interfere with adhesion. The predictive value of Hildebrand δ alone is limited because hydrogen-bonding and polar contributions also control diffusion into the polar carbonate linkage. For adhesive qualification, ASTM D543 immersion and ISO 22088-3 constant-strain exposure with the actual cured and uncured adhesive are required.

On injection molding lines, PC melts are processed at barrel temperatures of 280–320 °C and mold temperatures of 70–95 °C to minimize molded-in stress; clamp forces from 50 to 200 tons are common for lighting and medical housings. Adhesive joints near hot-runner gates, weld lines, or metal inserts carry residual stress concentrations that are not visible until a solvent or aggressive monomer penetrates. Production records from PC glazing fabricators show that craze failures often localize at gate vestiges where photoelastic stress exceeds 20 MPa, rather than on unstressed flat areas. This distribution is a practical warning: adhesive compatibility must be tested on as-molded parts, not on annealed laboratory coupons, because annealing can erase the stress state that causes field crazing.

Adhesive Classes That Maintain PC Clarity and Impact Resistance

100% solids two-part epoxy and two-part polyurethane adhesives are the most widely specified for PC because the absence of volatile solvent removes the main crazing agent. Epoxy formulations based on bisphenol-A diglycidyl ether and cycloaliphatic amines produce room-temperature lap shear strengths of 8–18 MPa on PC when tested according to ISO 4587; failure is frequently cohesive in the adhesive or substrate. The cure exotherm of a fast epoxy mixed in 100 g mass can exceed 120 °C; for thin PC sections this thermal spike can introduce additional stress. Slow hardeners or mix volumes below 50 mL are therefore specified. Two-part polyurethanes with aliphatic isocyanate prepolymers give lower modulus, lap shear typically 4–10 MPa, and better peel flexibility. They are preferred for PC parts subject to impact or thermal expansion mismatch because the cured adhesive can retain elongation above 50%.

Neutral-cure one-part RTV silicones of the alkoxy type are safe on PC and are specified to ASTM C920 for a movement capability of Class 25; tensile adhesion on PC after water immersion is typically 0.3–1.0 MPa. Acetoxy-cure silicones, which release acetic acid, should not be used on polycarbonate that contains residual stress or is enclosed. Silane-terminated polyether, often designated MS polymer, is another solvent-free moisture-cure option with lap shear of 2–4 MPa and high elongation. UV-curable acrylic adhesives can be used provided the formulation is low in free monofunctional acrylate monomer and cure depth is confirmed at 365 nm with 3–5 J/cm²; uncured monomer migrating from shadowed areas can craze PC over time.

When Solvent-Free UV Acrylics Are Used on Clear Polycarbonate, Cure Depth Must Be Measured

UV-curable acrylic adhesives transmitted through PC require attention to the UV absorption of the substrate. Standard PC grades absorb more than 90% of UV below 380 nm, so through-cure is limited to sheet thicknesses below 3–4 mm unless a visible-blue photoinitiator is used. Equipment with 365 nm LED or medium-pressure mercury lamps should deliver 3–5 J/cm² to the bondline; the adhesive supplier’s stated cure depth is often measured on glass, not on polycarbonate. When cure is incomplete, residual acrylate monomer acts as a crazing agent. The joint should be checked by ISO 4587 lap shear after 24 h at 23 °C and after 7 days at 70 °C/62% RH to expose delayed monomer attack.

Cyanoacrylate adhesives require separate qualification because the ethyl ester monomer is a potential stress-cracking agent for polycarbonate. Standard ethyl cyanoacrylate can craze PC parts under tensile or flexural load, especially when the adhesive is applied as a thick bead and unreacted monomer remains at the fillet. Higher-molecular-weight butyl or octyl cyanoacrylates and non-blooming grades show less tendency to craze and are generally preferred for low-clearance PC bonding below 0.25 mm. Lap shear values on PC are typically 4–8 MPa (ISO 4587), but peel strength is low. Medical device housing joints should be tested under ISO 10993-5 after adhesive cure because residual cyanoacrylate monomer can be cytotoxic if not fully reacted.

Before adhesive application, the PC surface is cleaned only with a mild aqueous detergent solution or aliphatic naphtha; ketones, esters, aromatic hydrocarbons, and chlorinated solvents are excluded. If isopropanol is specified for degreasing, it should be a 50% aqueous solution applied with a cleanroom wipe and immediately followed by dry air; even then, isopropanol can craze PC under high molded-in stress. Abrasion is generally not used because PC scratches readily and the scratches can act as craze initiators. Corona or atmospheric plasma treatment at 50–100 mm/s can raise surface free energy from approximately 38 mN/m to above 60 mN/m and improve wetting of 100% solids adhesives, but it does not remove surface contamination; cleaning must precede treatment.

Bond-Line Testing on Molded PC Housings Under Humid Ageing

Qualification of an adhesive for PC should include lap shear specimens cut from production plaques with the same mold temperature, gate geometry, and orientation as the final part. ISO 4587 specimens are 25 mm wide with a 12.5 mm overlap and are pulled at 1 mm/min; ASTM D3163 is used for rigid plastic lap-shear joints. A suitable test matrix includes as-molded, annealed, and 85 °C/85% RH-conditioned specimens to capture hydrolytic degradation and residual-stress release. For PC, the failure mode should remain cohesive in the adhesive or cohesive in the substrate; interfacial failure indicates inadequate surface preparation or an aggressive solvent residue. ASTM D950 impact and ISO 11339 peel tests add information on impact and flexible-joint performance but are secondary to lap shear for rigid structural joints.

Where a solvent-borne adhesive cannot be eliminated, the part is annealed below the glass transition temperature, typically at 120 °C for 2 h, to reduce molded-in stress before bonding. If total volatile organic compounds exceed 5% by mass, the adhesive is not used on stressed PC unless full ISO 22088-3 testing is completed. Polychloroprene contact adhesives supplied in toluene or methyl ethyl ketone are not acceptable for PC glazing because the solvent remains trapped under the bonded overlap and produces delayed craze arrays. Structural acrylics based on methyl methacrylate are also high risk unless the manufacturer has published PC-specific environmental stress cracking data; several formulations contain 30–50% by mass methacrylate monomers that can solvate the surface. Adhesives that cure by moisture, UV, or two-component addition are preferred. On stressed production parts, the only defensible qualification is a combination of ISO 22088-3 constant-strain exposure, ISO 4587 lap shear after conditioning, and visual inspection under polarized light at 10×–20× magnification for craze initiation.