In immersion stripping of amine-cured epoxy coatings (200–500 µm dry film thickness) deposited on SAE 1020 carbon steel process vessels, a dibasic ester/benzyl alcohol gel formulated with hydrophobic fumed silica at 4.0 wt% and a thixotropic index of 3.5 (measured per ISO 3219:1994 at 0.1 s⁻¹ and 10 s⁻¹) delivers a steady-state penetration rate of 8–12 µm/min at a bath temperature of 25°C, enabling complete removal within 90–120 min immersion, a significant reduction from the 6–8 h typical of hot alkaline permanganate processes. Exceeding 35°C bath temperature triggers evaporative skinning due to preferential loss of the lower-boiling benzyl alcohol component (boiling point 205°C), forming a surface film that reduces effective solvent activity at the coating interface and lowers penetration to 4–6 µm/min within 30 min of exposure; simultaneously, process environments with relative humidity above 70% induce hygroscopic uptake of atmospheric moisture into the gel layer, diluting the continuous solvent phase and collapsing the low-shear viscosity from a nominal 8,000 mPa·s to below 2,000 mPa·s, which results in gravity-driven sag on vertical surfaces and a penetration decay to ≤5 µm/min. The immersion system is operated in a polypropylene tank with external cartridge filtration to 50 µm and low-shear recirculation at 40 L/min through a peristaltic pump to avoid coating particle size reduction that would alter subsequent waste classification.
| Standard Designation | Title / Parameter Measured | Application Scenario |
|---|---|---|
| ASTM D6189-97 (2014) | Standard Practice for Evaluating the Efficiency of Chemical Removers for Organic Coatings | Epoxy on steel, powder coating on aluminum |
| ISO 3219:1994 | Plastics — Polymers/resins in the liquid state or as emulsions or dispersions — Determination of viscosity using a rotational viscometer with defined shear rate | All gel systems |
| ISO 2555:2018 | Plastics — Resins in the liquid state or as emulsions or dispersions — Determination of Brookfield RV viscosity | Gel rheology control |
| SSPC-PA 2 | Procedure for Determining Conformance to Dry Coating Thickness Requirements | Coating thickness verification |
| ASTM D3359-17 | Standard Test Methods for Rating Adhesion by Tape Test | Residual coating adhesion after stripping |
| ASTM D2344/D2344M-16 | Standard Test Method for Short-Beam Strength of Polymer Matrix Composite Materials and Their Laminates | Composite substrate integrity (polyurethane on CFRP) |
| SAE AMS-C-83363 | Coating, Polyurethane, Aircraft Exterior, Chemical Agent Resistant | Aerospace composites |
| ISO 4287:1997 | Geometrical Product Specifications (GPS) — Surface texture: Profile method — Terms, definitions and surface texture parameters | Aluminum surface roughness after acid stripping |
| ISO 22088-3:2006 | Plastics — Determination of resistance to environmental stress cracking (ESC) — Part 3: Bent strip method | ABS automotive trim |
| ISO 4624:2016 | Paints and varnishes — Pull-off test for adhesion | Adhesion verification post-stripping |
The gel penetration mechanism in this configuration is governed by solvent diffusion through the semi-gelled boundary layer adjacent to the coating surface, a process that can be approximated by a Fickian model with an apparent diffusion coefficient on the order of 5×10⁻⁷ cm²/s for the benzyl alcohol/dibasic ester blend at 25°C, rising to 1.2×10⁻⁶ cm²/s at 30°C. The fumed silica network (BET 200 m²/g) forms a three-dimensional percolating structure with a storage modulus G′ of 400 Pa at 1 Hz and 1% strain, sufficient to immobilize the liquid phase yet yielding under the low-shear conditions of immersion to allow convective transport of solvated coating fragments from the diffusion front. Once the amine-cured bisphenol-A epoxy matrix absorbs 8–12 wt% solvent, its glass transition temperature is depressed from an initial 95°C (measured by differential scanning calorimetry at 10°C/min) to below the bath temperature, inducing a rubber-to-fluid transition that manifests as a sharp increase in gravimetric mass loss rate. This threshold mass uptake corresponds to a penetration front progression that can be tracked by sequential cross-sectional microscopy of coupons extracted at 15 min intervals, showing a distinct swollen interphase layer of 30–50 µm ahead of the physical erosion front. The practical consequence is that the stripping process exhibits a lag phase of 20–30 min during which no visible mass loss occurs, followed by a linear stripping regime that ceases when the residual backing layer falls below 10 µm, at which point the mechanical agitation from recirculation is insufficient to dislodge the last adhered fragments and a short post-rinse with a high-flash-point ketone is required.
An operational boundary arises from the gel’s sensitivity to both evaporative skinning and hygroscopic dilution. In tank configurations open to ambient air, the evaporation rate of benzyl alcohol at 35°C and 0.5 m/s air velocity is approximately 0.8 g/m²/min, calculated from the Antoine equation vapor pressure of 0.013 kPa, leading to a surface layer enrichment of dibasic ester and concomitant viscosity increase exceeding 50,000 mPa·s within 60 min of uninterrupted exposure. This skin inhibits further solvent exchange and reduces penetration to <3 µm/min. To mitigate, immersion tanks are fitted with floating polypropylene cover balls to reduce the exposed surface area by 90%, and the bath is re-dosed with benzyl alcohol at a rate of 0.2 vol% per 8 h shift based on gas chromatographic headspace analysis. Hygroscopic dilution is countered by maintaining the gel’s initial water content below 1.5 wt% (Karl Fischer titration) and installing a desiccant breather on the tank vent. Formulations containing ≥6 wt% silica are more resistant to viscosity collapse but exhibit poor initial wetting of the coating interface, requiring a pre-wetting step with a low-viscosity surfactant solution (0.1% sodium dioctyl sulfosuccinate in DBE) to reduce the contact angle on epoxy from 60° to below 20°.
For the immersion stripping of two-component polyurethane topcoats (75–150 µm DFT) applied over epoxy-based carbon fiber composite aircraft fairings, a gel formulated with γ-butyrolactone (BLO) as the primary active solvent, gelled with 5.0 wt% organically modified bentonite (C10-C16 alkyl quaternary ammonium intercalated montmorillonite), and activated with a non-ionic ethoxylated alcohol surfactant (2 wt%, HLB 12.5), achieves a penetration rate of 4.2 µm/min at 20°C bath temperature, accelerating to 6.5 µm/min when the process temperature is raised to 25°C. However, concurrent solvent uptake by the epoxy laminate matrix exceeds 2.0% by mass after 4 h immersion at 25°C, as measured by thermogravimetric analysis of 2 mm thick laminate coupons exposed to the gel on one face, and this threshold corresponds to a 12% decline in interlaminar shear strength (ILSS) when tested per ASTM D2344 on 20-ply unidirectional prepreg laminates with a 0°/90° layup. The safe immersion window is thereby constrained to ≤3 h at a maximum bath temperature of 22°C to maintain laminate ILSS above 90% of the initial 75 MPa value, aligning with the acceptance criteria of SAE AMS-C-83363 for chemical agent resistant coatings on aircraft structures.
A multi-factor interaction emerges because the same temperature rise that linearly accelerates the stripping rate through both Arrhenius-driven solvent diffusion and reduced gel viscosity simultaneously lowers the activation energy for solvent ingress into the epoxy matrix between the carbon fibers. The apparent activation energy for BLO diffusion through the polyurethane coating is determined to be 28 kJ/mol by tracking penetration front advancement at 15°C, 20°C, and 25°C under isothermal conditions, while the activation energy for BLO permeation into the cured epoxy laminate is estimated at 25 kJ/mol from mass uptake curves, values sufficiently close that a thermal bias cannot selectively enhance coating penetration without proportionally increasing the driving force for matrix swelling. Moreover, at 25°C, the organoclay gel network begins to exhibit partial exfoliation under the osmotic stress generated by the concentration gradient of dissolved polyurethane oligomers, leading to a time-dependent reduction in the gel’s yield stress from an initial 120 Pa (controlled stress ramp from 0.1 Pa to 500 Pa over 600 s) to below 40 Pa after 180 min, which results in slumping on vertical surfaces and inadequate film thickness to sustain a uniform solvent reservoir. This gel degradation is measurable in situ via a drop in the viscoelastic storage modulus from 1,200 Pa to 350 Pa at 1 Hz, monitored every 30 min on a stress-controlled rheometer fitted with a solvent trap. The contradictory requirement — high temperature to accelerate stripping versus low temperature to protect composite structure — forces a compromise process point of 22°C ±1°C with an immersion time limit of 165 min, verified by ILSS testing on witness coupons co-immersed with production parts. Any excursion above 23°C triggers automatic cooling water injection into the tank jacket and immediate removal of parts for an intermediate rinse and inspection.
The immersion vessel is constructed of stainless steel 316L with a controlled immersion depth of 1.5 m and is designed for batch processing of up to 12 fairing segments simultaneously. Each part is supported on a polypropylene grid that maintains a minimum 100 mm clearance from the tank floor to prevent accumulation of dislodged coating sludge on the composite surface. A low-velocity (0.1 m/s linear) pumped recirculation loop equipped with a 25 µm woven polypropylene bag filter removes particulate before it can redeposit. The gel film thickness on each part is maintained at 2.0–2.5 mm by an automated spray application onto the dry part prior to immersion; this thickness is critical because below 1.5 mm, solvent depletion during the 3 h cycle leads to a cessation of penetration and the formation of a semi-dried interface that requires mechanical brushing to remove. Gel penetration is quantified destructively on sacrificial witness panels of identical coating/composite construction, with cross-sections microtomed at -40°C and examined by confocal fluorescence microscopy after staining the polyurethane layer with a solvatochromic dye incorporated in the gel at 0.01 wt%. The depth of fluorescence penetration correlates with the gravimetric mass loss with a coefficient of determination R² of 0.97.
A further operational boundary is the susceptibility of the BLO-based gel to hydrolysis at elevated temperature and neutral pH, with an observed degradation rate of 0.5% solvent loss per day at 22°C forming γ-hydroxybutyric acid, which can catalyze epoxy matrix degradation. To suppress this, the gel is buffered with 0.5 wt% triphenyl phosphate acting as an acid scavenger, and the bath is continuously sparged with dry nitrogen to maintain dissolved oxygen below 1 ppm. Parts exhibiting copper mesh lightning strike protection are excluded from this process because BLO can selectively dissolve the epoxy film adhesive bonding the copper mesh to the laminate, causing disbanding detectable by tap test per ASTM D4580.
Immersion stripping of thermosetting polyester/triglycidyl isocyanurate (TGIC) powder coatings (60–120 µm DFT) from architectural 6063-T6 aluminum extrusions is accomplished using a gel based on dihydrolevoglucosenone (Cyrene™) as the primary penetrating solvent, thickened with 6.0 wt% modified hydrogenated castor oil, and activated with 5–10 wt% aqueous formic acid (85% concentration). At a bath temperature of 30°C, this formulation achieves a penetration rate of 20–30 µm/min, reducing the total immersion cycle to 15–25 min for complete removal. However, the inclusion of formic acid drives the gel pH to values as low as 2.8, at which point the aluminum substrate undergoes uniform etching with a mass loss rate exceeding 0.5 mg/cm²/h and an increase in surface roughness Ra of 0.8 µm per ISO 4287 measured over a 4 mm evaluation length with a 2 µm radius diamond stylus. The resulting surface pitting concentrates stress under subsequent anodizing layers and violates the 0.3 µm Ra maximum specified by the Qualicoat 16th Edition specification for architectural powder coating pretreatment. Consequently, the gel requires buffering with 2.0 wt% ammonium bicarbonate to maintain a steady pH of 4.0–4.5 throughout the 30 min bath life, reducing aluminum mass loss to less than 0.05 mg/cm²/h while only moderately suppressing the penetration rate to 18–25 µm/min.
| Formulation Code | Formic Acid (wt%) | Temperature (°C) | Gel pH | Penetration Rate (µm/min) | Al Mass Loss (mg/cm²/h) | Ra Increase (µm) |
|---|---|---|---|---|---|---|
| A1 | 5 | 30 | 3.8 | 20 | 0.15 | 0.19 |
| A2 | 10 | 30 | 2.8 | 30 | 0.62 | 0.82 |
| A3 | 10 + buffer | 30 | 4.2 | 25 | 0.04 | 0.06 |
| B1 | 5 | 40 | 3.7 | 38 | 0.28 | 0.35 |
| B2 | 10 | 40 | 2.7 | 52 | 1.10 | 1.45 |
| B3 | 10 + buffer | 40 | 4.1 | 45 | 0.09 | 0.12 |
The penetration mechanism for polyester/TGIC powder coatings under acidic gel conditions diverges from simple solvent diffusion; it proceeds primarily by acid-catalyzed hydrolysis of ester linkages in the polyester backbone, swelling the crosslinked network to a critical degree where chain scission causes a transition from a ductile solid to a friable, fragmented state. The Cyrene™ solvent functions both as a solubilizing medium for the acid and as a swelling agent that transports hydrolytic fragments away from the reaction front, with a synergistic effect observed when the Cyrene™/water ratio is maintained near 85:15 v/v, corresponding to the azeotropic composition that maximizes acid dissociation while preventing aluminum passivation by aluminum formate precipitation. The buffering system using ammonium bicarbonate introduces ammonium cations that form a transient protective film on the aluminum surface at pH >4.0, confirmed by X-ray photoelectron spectroscopy detection of nitrogen at 399.5 eV binding energy on the substrate after rinsing. Stripping tanks for this application are constructed of unfilled polypropylene, rated for continuous service at 80°C, and fitted with external immersion heaters sheathed in PTFE to avoid fluoride leaching from alternative sheath materials that could locally etch the aluminum. The bath is agitated by low-pressure air bubbling through a 10 µm porous polyethylene sparger tube placed at the tank bottom, generating a rolling motion sufficient to dislodge loosened coating platelets without creating foam, which would inhibit gel contact with the aluminum surface. A critical process control point is the monitoring of dissolved aluminum concentration in the gel bath by inductively coupled plasma optical emission spectroscopy; when aluminum content exceeds 500 ppm, the bath life is considered exhausted because the accumulated aluminum formate complexes begin to buffer the gel upward to pH >5.5, dramatically slowing penetration to <10 µm/min. At this exhaustion point, the batch is discarded for disposal as acidic hazardous waste per RCRA requirements.
Operational limits also include a strict prohibition against processing extrusions with anodized or chromate-conversion coated surfaces that have been partially exposed through coating damage, because the formic acid gel will attack the conversion coating leading to localized pitting at a rate 3× greater than on bare aluminum. Additionally, gel application thickness below 1.5 mm leads to rapid acid consumption and a pH rise that terminates the hydrolysis reaction before the full coating thickness is penetrated, leaving a partially degraded layer that requires re-immersion or manual abrasive removal. The process is validated by comparing the penetration rate on production extrusions to that on standard reference panels with 80 µm DFT of a commercial polyester/TGIC powder coating (Tiger Drylac Series 49), with acceptance criteria requiring ≥95% removal within 25 min at 30°C.
Stripping of sequentially applied acrylic basecoat/clearcoat or acrylic/2K urethane systems (60–120 µm total DFT) from injection-molded ABS (acrylonitrile butadiene styrene, grade Lustran 348) interior trim components requires a gel system based on propylene carbonate (60 vol%) and ethanol (40 vol%), thickened with 3.5 wt% hydroxypropyl methylcellulose (viscosity grade 4000 mPa·s at 2% aqueous). At an immersion bath temperature of 22°C, the gel achieves a penetration rate of only 1.5–2.8 µm/min due to the relatively low solvent power of the blend toward the crosslinked urethane network and the necessity to limit ethanol content — the primary active solvent for acrylic layers — to a maximum of 25 vol% to avoid environmental stress cracking (ESC) of the ABS substrate. When ethanol exceeds this threshold, the critical strain for ESC, measured per ISO 22088-3 using a 0.5% strain bent-strip jig for 24 h exposure, falls below 0.5%, resulting in visible craze formation at gate vestige areas and other stress concentrations inherent to the molded part geometry. The penetration process is therefore operated near the solubility limit of the coating but below the ESC threshold of the substrate, a narrow composition window that demands online monitoring of ethanol concentration by near-infrared spectroscopy with a measurement frequency of 15 min and automatic replenishment through a metering pump triggered at −1.5 vol% deviation from the set point.
The kinetic limitation arises from a diffusion-controlled mechanism in which the gel forms a semi-permeable skin at the coating interface due to the rapid evaporation of ethanol from the thin immersion bath surface (0.5 m² exposed area) even at low ambient temperatures. Gel film thickness on the part is maintained at 2.0–3.0 mm, and when it falls below 2 mm, solvent exhaustion during the extended 8–12 h cycle leads to a plateau in penetration depth at approximately 65–80% of the total coating thickness, leaving a swollen but adherent residue that must be mechanically wiped. To mitigate solvent losses, the immersion vessel is a fully enclosed stainless steel tank with a nitrogen-blanketed headspace and a chilled condenser loop (−5°C) to recover evaporating ethanol, achieving a solvent recovery efficiency of 92%. The gel is formulated with 0.5 wt% butylated hydroxytoluene (BHT) to suppress oxidative degradation of propylene carbonate under the slightly elevated temperature and extended duration, as monitored by peroxide value titration (target <5 meq/kg). A significant incompatibility exists with parts containing integrally molded polycarbonate accent overlays adhered to the ABS; the propylene carbonate selectively diffuses into the polycarbonate at 4× the rate into the coating, causing interlayer delamination and optical hazing, so these assemblies are excluded from immersion and are stripped manually with a gel poultice applied only to the painted surface with a precision dispensing robot.
The stripping bath temperature is held at 22°C ±1°C by a closed-loop chiller circulating propylene glycol through the tank jacket, because excursions to 25°C accelerate ethanol evaporation and shift the gel composition into the ESC-prone region even if the bulk ethanol content initially meets specifications. The penetration rate is measured on sacrificial tabs cut from the same production lot of ABS panels, coated and cured under identical conditions, by using micrometry of the swollen layer every 60 min and verifying complete removal by methyl ethyl ketone rub test per ASTM D5402 for acrylics. A bath life of 40 hours is typical before the accumulation of dissolved acrylic and urethane residues raises the gel viscosity above 25,000 mPa·s (Brookfield RV spindle #7, 20 rpm) and necessitates batch replacement. Process exhaust is monitored for ethanol vapor to maintain workplace exposure below the OSHA permissible exposure limit of 1,000 ppm (8-h TWA), with a fixed photoionization detector interlocked to the tank lid mechanism.