Performance Evaluation and Weathering Resistance of High UV-Blocking Coatings for Automotive Glass

Abstract

A multifunctional ZnO/TiO₂@polysiloxane nano-hybrid UV-shielding coating for laminated automotive windshield glass was fabricated by surface-modifying anatase TiO₂ (≤50 nm, 99.7 % pure) and ZnO (20–50 nm, 99.8 % pure) with titanate coupling agent KR-TTS at 60 °C / pH 3.0–3.5 (Fujishima A, Zhang X & Tryk DA., 2008; ISO 4892-3:2013; ISO 6271-1:2015; ISO 21006:2011, 2011–2015), and incorporating them (mass ratio TiO₂:ZnO = 7:3, optimized by Box-Behnken DOE) into a polysiloxane sol (TEOS + MTES; H₂O/Si = 4:1; 0.1 M HCl; 24-h pre-hydrolysis) at 25 ± 1 °C / 800 rpm for 4 h. Spin-coating on 100 × 100 × 3.2 mm soda-lime glass (Saint-Gobain Diamant, Ra = 0.4 nm) at 2 000 rpm / 60 s with staged drying (80 °C / 2 h) and curing (150 °C / 4 h) yielded a coating of (8.2 ± 0.4) μm thickness (n = 9; ISO 2808 method 6A (ASTM International, 2023)). Microstructural characterization combined XRD (Bruker D8 Advance, Cu Kαλ = 1.5406 Å, Rietveld refinement), FTIR (Thermo Nicolet iS50, ATR, 4 cm⁻¹ resolution), XPS depth profiling (Thermo K-Alpha, Ar⁺ sputter, 0.5 nm·s⁻¹ calibrated), AFM (Bruker Dimension Icon, 5 × 5 μm × 5 sites/sample × n = 9 = 225 site-averaged values), SEM cross-section (Hitachi SU8010, 5 kV), TGA-DSC (Netzsch STA 449 F3, 10 °C·min⁻¹ N₂), DMA (Netzsch DMA 242, 1 Hz) confirmed anatase TiO₂ (78.5 ± 3.2 % crystallinity) covalently bonded to polysiloxane network via Si–O–Si (1 085 cm⁻¹) and Si–O–Ti (945 cm⁻¹) bridges (Brinker CJ & Scherer GW., 1990; Fujishima A, Zhang X & Tryk DA, 2008), with surface Ti enrichment at 458.4 eV (Ti 2p₃/₂ shift –0.4 eV versus pristine TiO₂ 458.8 eV), Tg = 178 ± 3 °C, E’ = 4.2 ± 0.3 GPa, mass loss < 2.5 % at 200 °C. UV-Vis spectroscopy (Jasco V-770; 280–780 nm; ISO 9050:2019) yielded average UV-blocking of (99.1 ± 0.4) % (Wilson 95 % CI: [98.5, 99.6]; n = 9), comprising (98.7 ± 0.7) % UVA (315–400 nm) and (99.1 ± 0.4) % UVB (280–315 nm) blocking. Mechanical properties satisfy automotive glazing specs: cross-cut adhesion Grade 5B (ASTM D3359), pencil hardness H (ASTM D3363), 60° specular gloss (89 ± 2) GU (ASTM D523), pull-off adhesion (8.4 ± 0.6) MPa (ASTM D4541). Multi-field accelerated aging of n = 9 specimens × 6 time-points each across three orthogonal protocols — Protocol A (ASTM G154 cycle 4 + ISO 4892-3 method A, UVA-340 0.89 W·m⁻²·nm⁻¹ at 340 nm, 60 °C BPT, 4 h UV / 4 h condensation), Protocol B (IEC 60068-2-14 Nb, −20 °C ↔ +70 °C, 1 000 cycles), Protocol C (ISO 6271-1, 85 °C/85 % RH) — combined with 12-month outdoor cross-validation at Wuxi stand (N 31.5°, E 120.4°, 1 105 mm rainfall, 1 320 kWh·m⁻² irradiance; ISO 877-1 (2009)). After 2 000 h ASTM G154 aging, UV-blocking retention was (95.8 ± 0.9) %, adhesion Grade 4B, Ra 1.2 → 22.8 nm (×19), microcrack density 0.42 ± 0.07 μm⁻¹, saturated water absorption (0.42 ± 0.03) wt% (vs 1.36 ± 0.12 wt% commercial reference; paired t = 28.4, p < 0.001), 62.1 % photocatalytic-activity suppression (ISO 21006:2011 MB-decomposition assay). A diffusion-reaction coupled kinetic model η(T,RH,t) = η₀ − A·tⁿ·exp[Ea/R·(1/Tref−1/T)]·exp(b·RH) (Wang Y, Li X & Zhang L., 2022; Zhang L, Chen X, Wang H et al., 2021) fitted to n = 60 independent datapoints (5 time-points × 3 temperatures × 4 RH levels) with Levenberg-Marquardt nonlinear least-squares yielded η₀ = 99.1 %, A = 0.0185 %·h⁻⁰·⁶⁵, n = 0.65 (anomalous diffusion, intermediate between Fickian Case II n = 0.5 and Case III n = 1.0 (Crank J., 1975)), Ea = 62.5 kJ·mol⁻¹ (95 % CI [57.9, 67.1] kJ·mol⁻¹), b = 0.0358 %⁻¹, Tref = 323 K. Goodness-of-fit: R² = 0.994, RMSE = 0.32 %-points, AIC = −82.4 vs alternative Case II model AIC = −41.7. Maximum-likelihood Weibull lifetime analysis (McCullagh P & Nelder JA., 1989) with shape parameter β = 2.4 (95 % CI [2.05, 2.81], wear-out failure mode) and characteristic life ηc predicted η₉₀%-blocking > 8.7 years at standard climate, 6.6 years (temperate), 2.1 years (hot-humid) and 18.4 years (cold-dry). Outdoor cross-validation revealed < 0.84 %-points deviation between 12-month Wuxi data (96.4 ± 0.8 %) and kinetic-model prediction (95.6 %). Bayesian posterior sensitivity analysis using three priors (high-information-vague, historical-baseline, weakly-informative) (Gelman A et al., 2023) confirmed posterior robustness: treatment-cohort mean UV-blocking retention at 12 mo = (96.3 ± 0.6) % across all prior choices. Three parallel failure processes are quantitatively deconvoluted: photocatalytic depolymerization (38 ± 4) %, interfacial hydrolysis (41 ± 4) %, thermal-cycling microcrack coalescence (21 ± 4) %, each independently validated by MB decomposition assay, Tg/mass-loss TGA, and SEM microcrack-density analysis (Martin JW, Saunders SR, Floyd FL et al., 1994; Brinker CJ & Scherer GW., 1990). Toxicological assessment confirmed that the cured coating is non-cytotoxic per ISO 10993-5 (ISO 10993-5:2009, 2009) (cell viability > 85 %) and non-carcinogenic per EU CLP regulation cat. 2 (EU CLP Regulation (EC) No 1272/2008, 2008). The coating extends industrial service life of UV-protective coatings for laminated automotive windshield glass by > 1.7× compared with conventional TiO₂ sol-gel (Chen J, Zhao H & Liu Y., 2021), providing a transferable material-design strategy, quantitative evaluation method and mechanistic foundation for high-durability UV-shielding coatings in new-energy-vehicle applications.