Abstract

Electronic devices deployed in industrial manufacturing environments are frequently exposed to metalworking cutting fluids during enclosure machining and subsequent assembly operations. Conventional encapsulation adhesives, while effective against moisture and moderate thermal stress, suffer progressive degradation when immersed in water-miscible cutting fluid emulsions containing surfactants, extreme-pressure additives, and corrosion inhibitors. This work presents the design, synthesis, and characterization of a fluorine-silicon co-modified epoxy-polyurethane hybrid adhesive engineered to resist cutting-fluid penetration while maintaining long-term waterproof integrity for electronic packaging. The adhesive system incorporates a perfluoropolyether-modified urethane prepolymer covalently bonded into an epoxy network, reinforced with silane-functionalized hydrophobic nano-silica to create a tortuous diffusion path and a low-surface-energy interface. We investigate the influence of fluorine content, silicon segment length, and nanofiller loading on water uptake, contact angle, lap-shear strength retention, and interfacial adhesion after accelerated immersion in a semi-synthetic cutting fluid at elevated temperature. Microstructural analysis via Fourier-transform infrared spectroscopy, scanning electron microscopy, and dynamic mechanical analysis reveals that nanoscale phase separation of fluorine-rich domains, combined with a highly crosslinked interpenetrating network, is the principal mechanism underlying fluid resistance. A long-term ageing model based on Fickian diffusion coupled with interfacial debonding kinetics is proposed to predict service lifetime under combined cutting-fluid and thermal exposure. The results indicate that the optimized adhesive retains more than 80% of its initial lap-shear strength after 1000 hours of cutting-fluid immersion at 60 degrees Celsius, outperforming commercial epoxy and silicone reference adhesives. These findings provide a materials design framework for electronic protection adhesives intended for harsh manufacturing and service environments.

1. Introduction

a Cutting-Fluid-Resistant Waterproof Adhesive

Electronic devices are increasingly integrated into industrial manufacturing systems, where they must operate reliably in the presence of lubricants, coolants, and process chemicals. Among these environmental stressors, metalworking cutting fluids represent a particularly aggressive medium because they are formulated as complex emulsions combining mineral or vegetable oils, nonionic and anionic surfactants, extreme-pressure additives, corrosion inhibitors, and biocides in an aqueous continuous phase [1,2]. During the machining of electronic device enclosures, mounting brackets, and heat sinks, cutting fluids can come into direct contact with adhesive-sealed seams, connector gaskets, and encapsulation boundaries. Subsequent cleaning steps may not fully remove fluid residues trapped in micro-gaps, leading to prolonged chemical exposure that can compromise the protective function of the adhesive over the device lifetime.

Polymeric adhesives and sealants are the primary materials used to protect electronic assemblies from environmental ingress. Epoxy resins dominate structural and encapsulation applications due to their excellent adhesion to diverse substrates, high mechanical strength, and good chemical resistance [3,4]. Silicone elastomers are widely employed for their low modulus, thermal stability, and inherent water repellency [5,6]. Polyurethane systems offer tunable flexibility and impact resistance [7]. However, each of these conventional chemistries exhibits well-documented limitations when exposed to cutting-fluid environments. Epoxies absorb significant moisture through their hydrophilic hydroxyl and ether groups, leading to plasticization, reduced glass transition temperature, and interfacial delamination at the adhesive-substrate boundary [8,9]. Silicones, while hydrophobic, swell in hydrocarbon-based fluids and suffer from relatively low cohesive strength and poor adhesion to low-surface-energy substrates [10]. Polyurethanes are susceptible to hydrolysis of urethane linkages under acidic or basic conditions, which are common in cutting-fluid formulations [11]. The surfactants in cutting fluids also act as wetting agents that accelerate fluid penetration along adhesive-substrate interfaces, a degradation pathway distinct from simple bulk water absorption [12].

The objective of this work is to develop an adhesive system that simultaneously resists cutting-fluid penetration, maintains long-term waterproof integrity, and preserves sufficient mechanical and interfacial properties for electronic packaging applications. The central hypothesis is that a synergistic combination of fluorinated low-surface-energy segments, silicon-containing flexible chains, and hydrophobic nanofillers, covalently integrated into a highly crosslinked epoxy-polyurethane interpenetrating network, can address the multiple degradation mechanisms active in cutting-fluid environments. Fluorinated segments, particularly perfluoropolyether chains, provide exceptionally low surface energy and resistance to both polar and non-polar solvents [13,14]. Silicon-containing segments, such as polydimethylsiloxane, contribute low-temperature flexibility, low water absorption, and improved dielectric properties [15,16]. Hydrophobic nano-silica particles, when properly dispersed and surface-functionalized, create a tortuous path for fluid diffusion and can enhance interfacial adhesion through mechanical interlocking and chemical coupling [17,18]. The epoxy-polyurethane hybrid network combines the high crosslink density and adhesion of epoxy with the toughness and segmental mobility of polyurethane, while the interpenetrating structure restricts macroscopic phase separation [19,20].

Three principal challenges must be overcome to realize this design. First, fluorinated and silicon-containing segments are inherently incompatible with hydrocarbon epoxy matrices due to large differences in solubility parameters, which can lead to macroscopic phase separation, optical haze, and degraded mechanical properties if not controlled at the nanoscale [21,22]. Second, low-surface-energy fluorinated groups tend to migrate to the adhesive-air surface during curing, which is beneficial for surface hydrophobicity but can deplete fluorine from the adhesive-substrate interface where fluid ingress is most critical [23]. Third, the incorporation of nanofillers into a high-viscosity hybrid resin system presents dispersion challenges; agglomerated particles act as defect sites that can accelerate rather than impede fluid penetration [24,25].

The proposed solution addresses these challenges through three integrated design modules. First, a perfluoropolyether-diol is reacted with an aliphatic diisocyanate to form a fluorinated urethane prepolymer whose terminal isocyanate groups can covalently bond into the epoxy network through reaction with hydroxyl groups generated during epoxy curing, thereby anchoring the fluorinated segments and preventing macroscopic phase separation. Second, a hydroxyl-terminated polydimethylsiloxane is incorporated as a co-soft segment in the urethane prepolymer, providing silicon-rich domains that complement the fluorine-rich domains in creating a multi-scale low-energy barrier. Third, fumed nano-silica particles are surface-modified with a fluorinated alkylsilane coupling agent to render them hydrophobic and compatible with the fluorine-silicon rich domains, while their residual silanol groups can participate in the epoxy curing reaction to create covalent filler-matrix bonds. A two-stage curing protocol is employed: a low-temperature pre-gel stage that allows controlled nanoscale phase separation and fluorine migration to both free and substrate interfaces, followed by a high-temperature post-cure that locks in the morphology through full network formation.

This work makes three contributions. We demonstrate a fluorine-silicon co-modified epoxy-polyurethane hybrid adhesive architecture that resists cutting-fluid immersion while maintaining long-term waterproof performance, with systematic characterization of composition-property relationships. We elucidate the degradation mechanisms of conventional and proposed adhesives under cutting-fluid exposure, distinguishing bulk fluid absorption, interfacial debonding, and surfactant-assisted wetting through gravimetric, mechanical, spectroscopic, and microscopic analyses. We also propose a coupled diffusion-debonding ageing model that predicts long-term adhesive performance under combined cutting-fluid and thermal stress, providing a quantitative tool for electronic packaging reliability assessment. The remainder of this paper is organized as follows: Section 2 describes the materials, synthesis procedures, and characterization methods; Section 3 presents and discusses the experimental results; Section 4 concludes the work and outlines future directions.

2. Experimental

2.1 Materials

Bisphenol-A diglycidyl ether epoxy resin (DGEBA, epoxy equivalent weight approximately 185-195 g/eq) was used as the base matrix resin. A methylhexahydrophthalic anhydride curing agent was selected for its low viscosity, good electrical properties, and reduced moisture sensitivity compared to amine curing systems. 1-methylimidazole was employed as a cure accelerator.

For the fluorinated component, a hydroxyl-terminated perfluoropolyether (PFPE-diol, average molecular weight approximately 2000 g/mol, fluorine content approximately 60 wt%) was used. The PFPE-diol was reacted with isophorone diisocyanate (IPDI) to form a fluorinated urethane prepolymer. For the silicon component, a hydroxyl-terminated polydimethylsiloxane (PDMS-diol, average molecular weight approximately 1000 g/mol) was incorporated as a co-soft segment.

Dibutyltin dilaurate (DBTDL) was used as a catalyst for the urethane prepolymer synthesis. Hydrophobic fumed nano-silica (primary particle size approximately 7-12 nm, specific surface area approximately 200 m2/g) was surface-modified with tridecafluorooctyltriethoxysilane (FAS-13) to improve compatibility with the fluorinated matrix. A commercial semi-synthetic water-miscible cutting fluid, diluted to 5% concentration in deionized water (representative of typical industrial working concentrations), was used for all immersion tests.

2.2 Synthesis of Fluorine-Silicon Modified Urethane Prepolymer

The fluorine-silicon co-modified urethane prepolymer was synthesized via a two-step addition reaction. In the first step, IPDI was charged into a four-necked glass reactor equipped with a mechanical stirrer, nitrogen inlet, and thermometer, and heated to 60 degrees Celsius. A mixture of PFPE-diol and PDMS-diol at predetermined molar ratios (with total diol to IPDI molar ratio fixed at 1:2) was added dropwise over a period of 60 minutes, with DBTDL catalyst at 0.05 wt% relative to total solids. The reaction was maintained at 70 degrees Celsius for 3 hours until the isocyanate content, determined by dibutylamine back-titration, reached the theoretical value. The resulting NCO-terminated fluorine-silicon urethane prepolymer was cooled to room temperature and stored under nitrogen.

Three prepolymer formulations were prepared with varying PFPE:PDMS molar ratios: FPU-1 (100:0, PFPE only), FPU-2 (70:30), and FPU-3 (50:50). A control prepolymer (PU-0) was synthesized using poly(tetramethylene ether) glycol (PTMEG, Mn 1000) instead of fluorinated and siliconated diols, for comparison.

2.3 Preparation of Hybrid Adhesive Formulations

The hybrid adhesive was prepared by blending the NCO-terminated urethane prepolymer with DGEBA epoxy resin at varying weight ratios (prepolymer:epoxy = 10:90, 20:80, 30:70, 40:60). The mixture was stirred at 60 degrees Celsius for 30 minutes to ensure homogeneity. The anhydride curing agent was then added at a stoichiometric ratio of 0.85:1 (anhydride:epoxy), followed by 1-methylimidazole accelerator at 1.0 phr relative to epoxy. For nanofiller-reinforced formulations, FAS-13 modified nano-silica was added at loadings of 1, 3, 5, and 7 wt% relative to total resin, and dispersed using a high-shear mixer at 3000 rpm for 15 minutes, followed by sonication in an ice-water bath for 30 minutes to break up agglomerates.

The adhesive formulations were degassed under vacuum (less than 5 mbar) at 50 degrees Celsius for 20 minutes before casting. Curing was performed in a two-stage protocol: first at 80 degrees Celsius for 2 hours (pre-gel stage allowing controlled phase separation), then at 120 degrees Celsius for 4 hours (post-cure for full network development), followed by slow cooling to room temperature at 1 degree Celsius per minute.

2.4 Characterization

Fourier-transform infrared spectroscopy (FTIR) was performed on cured adhesive films using an attenuated total reflection (ATR) accessory, with a resolution of 4 cm-1 and 32 scans, over the range 4000-600 cm-1. The consumption of isocyanate groups (2270 cm-1) and the formation of urethane carbonyl (1720 cm-1) and C-F bonds (1100-1300 cm-1) were monitored.

Contact angle measurements were conducted using the sessile drop method with deionized water and diiodomethane as probe liquids, at room temperature. At least five measurements were taken at different locations on each sample, and the surface energy was calculated using the Owens-Wendt method.

Water and cutting-fluid absorption were measured gravimetrically. Cured adhesive specimens (dimensions 20 mm x 20 mm x 1 mm) were immersed in deionized water or 5% cutting-fluid solution at 25, 40, and 60 degrees Celsius. At predetermined intervals, specimens were removed, blotted dry with lint-free tissue, and weighed to an accuracy of 0.01 mg. The percentage fluid uptake was calculated as:

$$M_t = \frac{W_t – W_0}{W_0} \times 100\$$

where $$W_$$ is the initial dry weight and $$W_$$ is the weight at time $$$$. The diffusion coefficient was determined from the initial linear region of the $$M_$$ versus $$\sqrt{t$$ curve using:

$$D = \pi \left(\frac{h}{4 M_\infty}\right)^2 \left(\frac{M_t}{\sqrt{t}}\right)^$$

where $$$$ is the specimen thickness and $$M_\inft$$ is the equilibrium fluid uptake.

Lap-shear strength was measured according to ASTM D1002 using aluminum alloy (6061-T6) adherends (101.6 mm x 25.4 mm x 1.6 mm), with a bond area of 25.4 mm x 12.7 mm and an adhesive layer thickness of approximately 0.2 mm controlled by glass bead spacers. Tests were conducted at a crosshead speed of 1.3 mm/min using a universal testing machine. At least five specimens were tested for each formulation and ageing condition. Strength retention was calculated as the ratio of aged strength to initial strength.

Dynamic mechanical analysis (DMA) was performed on cured rectangular specimens (30 mm x 10 mm x 1 mm) in tension mode at a frequency of 1 Hz, a heating rate of 3 degrees Celsius per minute, from -80 to 200 degrees Celsius. The storage modulus, loss modulus, and tan delta were recorded, and the glass transition temperature was taken as the peak of the tan delta curve.

Scanning electron microscopy (SEM) was used to examine the fracture surfaces of lap-shear specimens and the cross-sectional morphology of cryo-fractured adhesive films. Samples were sputter-coated with gold-palladium prior to observation. Energy-dispersive X-ray spectroscopy (EDS) was used to map the spatial distribution of fluorine and silicon elements.

Interfacial adhesion assessment was performed using a tapered double cantilever beam (TDCB) test configuration to measure the critical strain energy release rate ($$G_{IC$$) of the adhesive-aluminum interface, both before and after cutting-fluid ageing. The test followed the principles described in the literature for evaluating subcritical interfacial debonding under chemical exposure [26].

2.5 Accelerated Ageing Protocol

Accelerated ageing was conducted by fully immersing cured adhesive specimens and bonded lap-shear joints in 5% semi-synthetic cutting-fluid solution at 60 degrees Celsius for up to 1000 hours. Control specimens were aged in deionized water at the same temperature for comparison. At intervals of 168, 336, 504, 750, and 1000 hours, specimens were removed for mechanical testing, weight measurement, and microstructural analysis. Additional ageing at 40 and 25 degrees Celsius was conducted for selected formulations to enable Arrhenius extrapolation of long-term performance.

3. Results and Discussion

3.1 Chemical Structure and Network Formation

FTIR spectroscopy confirmed the successful synthesis of the fluorine-silicon urethane prepolymer and its subsequent integration into the epoxy hybrid network. The prepolymer spectra exhibited characteristic absorption bands at 2270 cm-1 (NCO stretching), 1720 cm-1 (urethane carbonyl), 1540 cm-1 (N-H bending), and a broad multiplet in the 1100-1300 cm-1 region corresponding to C-F stretching vibrations of the PFPE segments. The Si-CH3 symmetric deformation at 1260 cm-1 and Si-O-Si asymmetric stretching at 1000-1100 cm-1 confirmed the presence of PDMS segments. After blending with epoxy and curing, the NCO peak at 2270 cm-1 disappeared, indicating complete consumption of terminal isocyanate groups through reaction with hydroxyl groups generated during epoxy ring-opening polymerization. The persistence of C-F and Si-O-Si absorption bands in the cured network confirmed that the fluorinated and siliconated segments were retained without chemical degradation during the high-temperature cure.

The curing reaction involves the formation of an interpenetrating polymer network (IPN) in which the epoxy-anhydride copolymerization and the urethane chain extension occur simultaneously. The NCO-terminated prepolymer acts as a reactive modifier: its terminal isocyanate groups react with hydroxyl groups generated during epoxy curing, covalently grafting the fluorine-silicon soft segments onto the epoxy network. This covalent integration is critical for preventing macroscopic phase separation and for ensuring that the low-surface-energy segments remain anchored within the network rather than blooming to the surface or leaching out during fluid exposure.

3.2 Surface Properties and Wettability

The water contact angle of the cured adhesive films increased systematically with fluorine content. The control epoxy-polyurethane adhesive (PU-0, no fluorine or silicon) exhibited a water contact angle of approximately 72 degrees, consistent with the hydrophilic nature of unmodified epoxy-anhydride networks. The incorporation of PFPE segments raised the contact angle to approximately 105 degrees at 20 wt% prepolymer loading, and the co-incorporation of PDMS (FPU-2, 70:30 PFPE:PDMS) further increased the contact angle to approximately 112 degrees. The higher contact angle in the co-modified system can be attributed to the combined surface enrichment of both CF2/CF3 groups (from PFPE) and Si-CH3 groups (from PDMS), which have lower surface energies than hydrocarbon groups [27,28].

The diiodomethane contact angle, which probes the dispersive component of surface energy, also increased with fluorine and silicon content, indicating a reduction in the overall surface energy. Calculated surface energies decreased from approximately 42 mN/m for the control to approximately 18 mN/m for the FPU-2 formulation at 30 wt% prepolymer loading. This low surface energy is a key factor in resisting the wetting and penetration of cutting-fluid surfactants, which typically reduce the surface tension of the aqueous phase to 30-35 mN/m [29]. An adhesive surface energy below that of the cutting fluid creates a thermodynamic barrier to spontaneous wetting, slowing the initial ingress of fluid along the adhesive-substrate interface.

The addition of FAS-13 modified nano-silica further increased the water contact angle by 5-8 degrees, depending on loading. This enhancement arises from the combination of the hydrophobic fluorinated silane coating on the particle surface and the micro-nano hierarchical roughness created by well-dispersed nanoparticles at the film surface [30,31]. At loadings above 5 wt%, however, the contact angle increase leveled off, and at 7 wt% a slight decrease was observed, likely due to particle agglomeration that disrupted the uniform surface roughness.

3.3 Fluid Absorption and Diffusion Behavior

The equilibrium fluid uptake and diffusion coefficients for selected formulations immersed in deionized water and 5% cutting-fluid solution at 60 degrees Celsius are summarized in Table 1.

Table 1. Equilibrium fluid uptake and diffusion coefficients of adhesive formulations at 60 degrees Celsius.

The control epoxy adhesive showed the highest fluid uptake in both media; cutting-fluid absorption (3.61 wt%) exceeded water absorption (2.84 wt%). This difference reflects the surfactant and additive components of the cutting fluid, which partition into the epoxy matrix in addition to water, increasing the total mass uptake [12,32]. The non-fluorinated polyurethane prepolymer (PU-0) reduced fluid uptake moderately, primarily through increased crosslink density and reduced free volume of the hybrid network. The polyurethane segments themselves, however, contain hydrophilic urethane and ether linkages that limit the improvement in water resistance.

The fluorinated formulations (FPU-1 and FPU-2) showed markedly reduced fluid uptake. At 20 wt% prepolymer loading, FPU-2 exhibited equilibrium water uptake of only 0.63 wt% and cutting-fluid uptake of 0.79 wt%, representing 78% reductions in both media relative to the control epoxy. This improvement arises from three synergistic mechanisms: (1) the low polarizability and strong C-F bonds of PFPE segments create a hydrophobic environment that excludes water molecules; (2) the PDMS segments contribute additional hydrophobicity and low water solubility; and (3) the nanoscale phase separation of fluorine-silicon rich domains creates a dispersed low-permeability phase that increases the tortuosity of the diffusion path [33,34].

The addition of 3 wt% FAS-13 modified nano-silica to the FPU-2 formulation further reduced equilibrium fluid uptake to 0.41 wt% (water) and 0.52 wt% (cutting fluid). The nanofillers act as impermeable obstacles that force diffusing fluid molecules to follow a more tortuous path through the matrix, effectively increasing the diffusion path length [35,36]. The covalent bonding between the silane-modified filler surface and the epoxy network also reduces the interfacial free volume that could otherwise serve as preferential diffusion pathways.

The diffusion coefficients followed the same trend as equilibrium uptake, with the FPU-2 + 3 wt% nano-SiO2 formulation exhibiting the lowest values. The ratio of cutting-fluid to water diffusion coefficients was approximately 1.2-1.3 across all formulations, indicating that while cutting fluids penetrate somewhat faster than pure water due to surfactant-assisted wetting, the fundamental diffusion mechanism remains similar and is governed primarily by the polymer network structure.

The temperature dependence of fluid absorption followed an Arrhenius relationship, with activation energies for diffusion ranging from approximately 35 kJ/mol for the control epoxy to approximately 55 kJ/mol for the FPU-2 + nano-SiO2 formulation. The higher activation energy for the fluorinated system reflects the greater energy barrier for fluid molecules to penetrate through the tightly crosslinked, low-free-volume network containing fluorine-rich domains.

3.4 Mechanical Properties and Ageing Performance

The initial lap-shear strength of the adhesive formulations on aluminum substrates is presented in Table 2. The control epoxy exhibited the highest initial strength (approximately 24 MPa), consistent with its high crosslink density and strong polar adhesion to the oxidized aluminum surface. The incorporation of polyurethane prepolymer reduced the initial strength moderately, as the soft segments increased the overall network flexibility and reduced the crosslink density. However, the fluorinated formulations maintained good initial strength: FPU-2 at 20 wt% loading exhibited approximately 19 MPa, and the addition of 3 wt% nano-silica restored the strength to approximately 21 MPa through particle reinforcement and improved load transfer at the filler-matrix interface [37,38].

Table 2. Initial lap-shear strength and strength retention after 1000 hours of cutting-fluid immersion at 60 degrees Celsius.

After 1000 hours of cutting-fluid immersion at 60 degrees Celsius, the control epoxy retained only 28% of its initial lap-shear strength, with failure occurring predominantly at the adhesive-aluminum interface. This severe degradation is consistent with the high fluid uptake observed for the control epoxy: absorbed water and cutting-fluid components plasticize the matrix, reduce the glass transition temperature, and hydrolyze the adhesive-substrate bonds, leading to interfacial debonding [9,39]. The commercial silicone reference adhesive, despite its hydrophobicity, retained only 33% of its initial strength due to its inherently low cohesive strength and swelling in the hydrocarbon components of the cutting fluid.

The fluorinated formulations showed markedly improved ageing performance. FPU-2 at 20 wt% loading retained 83% of its initial strength after 1000 hours, and the failure mode remained predominantly cohesive within the adhesive layer rather than interfacial. This retention of cohesive failure indicates that the adhesive-substrate interface remained intact, and that the residual strength was limited by the bulk adhesive properties rather than by interfacial degradation. The addition of nano-silica at 1-3 wt% further improved both the absolute aged strength and the retention percentage, with the 3 wt% formulation achieving 17.6 MPa after ageing (84% retention). At 5 wt% loading, however, performance declined slightly due to particle agglomeration, which created defect sites that facilitated fluid ingress and crack initiation.

The time-dependent strength retention curves for selected formulations showed that the control epoxy experienced a rapid initial strength loss (approximately 40% loss within the first 168 hours), followed by a more gradual decline. This two-stage behavior is characteristic of adhesive joint degradation in aqueous environments: the initial rapid loss corresponds to interfacial weakening as water and surfactants penetrate along the adhesive-substrate interface, while the subsequent gradual decline reflects bulk matrix plasticization and slow hydrolysis of chemical bonds [12,40]. In contrast, the FPU-2 + nano-SiO2 formulation exhibited a much slower and more linear strength decline, with no abrupt initial drop, indicating that the low-surface-energy interface and high crosslink density effectively suppressed the rapid interfacial wetting and debonding mechanism.

The critical strain energy release rate ($$G_{IC$$) of the adhesive-aluminum interface, measured by TDCB tests, confirmed these observations. The initial $$G_{IC$$ values were similar across formulations (approximately 350-450 J/m2), but after 500 hours of cutting-fluid immersion, the control epoxy showed a reduction to less than 80 J/m2 (interfacial failure), while the FPU-2 + nano-SiO2 formulation maintained $$G_{IC$$ above 300 J/m2 with cohesive failure. This indicates that the fluorine-silicon modified adhesive preserves interfacial toughness under chemical exposure, a property that is essential for long-term electronic packaging reliability where the adhesive-seal interface is the primary barrier against environmental ingress.

3.5 Thermomechanical Properties and Network Structure

DMA analysis revealed the influence of fluorine-silicon modification on the viscoelastic behavior and network structure of the hybrid adhesives. The control epoxy exhibited a single tan delta peak at approximately 135 degrees Celsius, corresponding to the glass transition of the homogeneous epoxy-anhydride network. The PU-0 formulation (non-fluorinated polyurethane hybrid) showed a broadened tan delta peak with a slight shoulder at lower temperature, indicating partial microphase separation between the polyurethane soft segments and the epoxy hard phase.

The fluorinated formulations (FPU-1 and FPU-2) exhibited two distinct tan delta peaks: a low-temperature peak at approximately -60 to -40 degrees Celsius corresponding to the glass transition of the fluorine-silicon rich soft domains, and a high-temperature peak at approximately 110-125 degrees Celsius corresponding to the epoxy-rich hard phase. The presence of two distinct Tg values confirms nanoscale phase separation, which is desirable for achieving a combination of low-temperature flexibility (from the soft domains) and high-temperature mechanical integrity (from the hard phase) [41,42]. The separation between the two Tg values was larger for FPU-2 (PFPE:PDMS = 70:30) than for FPU-1 (PFPE only), suggesting that the co-presence of PFPE and PDMS segments promotes more pronounced and better-defined phase separation.

The storage modulus in the rubbery plateau region (above the high-temperature Tg) is proportional to the crosslink density of the network. The FPU-2 formulation exhibited a higher rubbery plateau modulus than the PU-0 control at the same prepolymer loading, indicating that the fluorine-silicon segments, despite their soft nature, do not reduce the effective crosslink density because they are covalently anchored into the network and form a reinforcing dispersed phase. The addition of nano-silica further increased the rubbery plateau modulus, consistent with the hydrodynamic reinforcement effect of rigid fillers and the additional crosslink sites provided by the silane-functionalized particle surface.

The crosslink density, calculated from the rubbery plateau modulus using rubber elasticity theory, increased from approximately 1.8 x 10^3 mol/m3 for the control epoxy to approximately 2.5 x 10^3 mol/m3 for FPU-2 + 3 wt% nano-SiO2. Despite the addition of soft segments, this increase is explained by the dual-cure network structure: the NCO-terminated prepolymer not only grafts onto the epoxy network but also undergoes chain extension and crosslinking through allophanate and biuret formation at the elevated cure temperature, creating additional crosslink points that compensate for the dilution effect of the soft segments.

3.6 Microstructural Analysis and Degradation Mechanisms

SEM examination of cryo-fractured surfaces revealed the morphology of the phase-separated networks. The control epoxy exhibited a smooth, featureless fracture surface characteristic of a homogeneous brittle thermoset. The PU-0 formulation showed a slightly rougher surface with small (approximately 50-100 nm) domains, indicating incipient phase separation. The FPU-2 formulation exhibited a well-defined sea-island morphology with dispersed domains of approximately 100-200 nm in diameter, uniformly distributed throughout the epoxy matrix. EDS mapping confirmed that these dispersed domains were enriched in fluorine and silicon, while the continuous matrix was predominantly hydrocarbon epoxy. This nanoscale phase separation is the optimal morphology: it is fine enough to maintain optical clarity and mechanical integrity, yet sufficiently pronounced to create discrete low-permeability domains that impede fluid diffusion [43,44].

At higher prepolymer loadings (above 30 wt%), the domain size increased to approximately 300-500 nm, and some degree of co-continuous morphology was observed. This larger-scale phase separation correlated with reduced mechanical properties and slightly higher fluid uptake, as the larger domains created longer continuous low-permeability pathways but also introduced more domain-boundary free volume that could facilitate diffusion. The two-stage curing protocol was critical for controlling the phase separation scale: the low-temperature pre-gel stage allowed sufficient mobility for the fluorine-silicon segments to aggregate into nanoscale domains before the gel point, while the subsequent high-temperature cure locked in this morphology without allowing further coarsening.

SEM analysis of the fracture surfaces of aged lap-shear specimens provided direct evidence of the degradation mechanisms. For the control epoxy after cutting-fluid immersion, the fracture surface was smooth and exhibited clear interfacial debonding, with residual aluminum oxide visible on the adhesive side, indicating that failure occurred at or very near the adhesive-substrate interface. EDS analysis of the debonded surface showed the presence of sodium, sulfur, and phosphorus, which are characteristic components of cutting-fluid additives (surfactants, extreme-pressure agents, corrosion inhibitors), confirming that cutting-fluid components had penetrated to and accumulated at the interface [1,29].

For the FPU-2 + nano-SiO2 formulation after the same ageing period, the fracture surface was rough and exhibited ductile tearing features characteristic of cohesive failure within the adhesive. No exposed aluminum substrate was visible, and EDS mapping showed only trace amounts of cutting-fluid elements, uniformly distributed throughout the adhesive rather than concentrated at the interface. This confirms that the fluorine-silicon modified adhesive effectively prevented cutting-fluid components from reaching the adhesive-substrate interface, and that the residual strength was governed by the bulk adhesive properties rather than interfacial degradation.

FTIR analysis of aged adhesive specimens provided further insight into chemical degradation. The control epoxy showed a noticeable increase in the hydroxyl absorption band (3200-3600 cm-1) and the appearance of new carbonyl absorption at approximately 1710 cm-1 after cutting-fluid immersion, consistent with hydrolysis of ester bonds in the anhydride-cured epoxy network and oxidation of the polymer matrix. The FPU-2 formulation showed minimal changes in its FTIR spectrum after ageing, with no significant increase in hydroxyl or new carbonyl absorption, indicating that the fluorine-silicon network was chemically stable under the cutting-fluid exposure conditions. The stability of the C-F and Si-O-Si absorption bands confirmed that neither the fluorinated nor the siliconated segments underwent chemical degradation during immersion.

3.7 Long-Term Ageing Model and Lifetime Prediction

A coupled diffusion-debonding model was developed to predict the long-term performance of adhesive joints under cutting-fluid exposure. The model consists of two interconnected components:

Bulk diffusion component: Fluid penetration into the adhesive follows Fickian diffusion, with the concentration profile $$C(x,t$$ governed by:

$$\frac{\partial C}{\partial t} = D \frac{\partial^2 C}{\partial x^2$$

where $$$$ is the effective diffusion coefficient, which is composition-dependent and increases with local fluid concentration due to plasticization.

Interfacial debonding component: The degradation of interfacial adhesion is modeled as a stress-assisted chemical reaction in which fluid molecules at the interface hydrolyze adhesive-substrate bonds. The interfacial bond density $$\sigma(t$$ evolves according to:

$$\frac{d\sigma}{dt} = -k_0 \exp\left(-\frac{E_a}{RT}\right) C_{int}(t) \sigma(t$$

where $$k_$$ is the pre-exponential factor, $$E_$$ is the activation energy for interfacial bond hydrolysis, and $$C_{int}(t$$ is the time-dependent fluid concentration at the interface, obtained from the diffusion solution.

The residual lap-shear strength is then expressed as a function of both the bulk matrix properties (degraded by plasticization, proportional to $$C_{avg}(t$$) and the interfacial bond density:

$$\tau(t) = \tau_0 \left[1 – \alpha C_{avg}(t)\right] \frac{\sigma(t)}{\sigma_0$$

where $$\tau_$$ is the initial strength, $$\alph$$ is the plasticization coefficient, and $$\sigma_$$ is the initial interfacial bond density.

The model parameters were calibrated using the experimental data from the 60 degrees Celsius immersion tests, and validated against the 40 and 25 degrees Celsius data. The model accurately reproduced the two-stage strength degradation behavior of the control epoxy (rapid interfacial debonding followed by slow bulk degradation) and the more gradual, near-linear decline of the fluorinated formulation. The key difference between the formulations was captured by the parameter $$C_{int}(t$$: for the control epoxy, the low contact angle and high surface energy allowed rapid fluid accumulation at the interface, leading to a high $$C_{int$$ and fast debonding; for the FPU-2 + nano-SiO2 formulation, the low surface energy and high diffusion barrier kept $$C_{int$$ at a very low level, effectively suppressing the interfacial debonding mechanism.

Using the calibrated model and Arrhenius extrapolation, the predicted time to reach 50% strength retention (a common reliability threshold for electronic packaging adhesives) at 25 degrees Celsius was approximately 1.5 years for the control epoxy, compared to more than 15 years for the FPU-2 + 3 wt% nano-SiO2 formulation. This order-of-magnitude improvement in predicted service lifetime demonstrates the potential of the fluorine-silicon co-modified adhesive for long-term electronic device protection in cutting-fluid-exposed environments. These predictions are based on accelerated immersion testing and do not account for additional stressors such as thermal cycling, mechanical vibration, or electrical bias, which may act synergistically with chemical exposure in real-world applications.

3.8 Discussion of Composition-Property Relationships

The systematic variation of formulation parameters reveals several important composition-property relationships that guide the design of cutting-fluid-resistant adhesives.

Fluorine content: Increasing the PFPE content from 0 to 20 wt% (relative to total resin) produced the largest improvements in contact angle, fluid resistance, and ageing performance. Beyond 20 wt%, the marginal improvement in fluid resistance diminished, while the initial mechanical strength and adhesion to substrates began to decline due to the increasing proportion of low-surface-energy, low-cohesive-strength fluorinated domains. The optimal PFPE content was therefore in the range of 15-25 wt%, where the benefits of fluorination were maximized without excessive sacrifice of mechanical properties.

PFPE:PDMS ratio: The co-modified system (70:30 PFPE:PDMS) outperformed both the PFPE-only (100:0) and PDMS-rich (50:50) systems. The PFPE-only system, while providing excellent chemical resistance, suffered from relatively poor low-temperature flexibility and higher modulus, which could lead to stress cracking under thermal cycling. The PDMS-rich system exhibited good flexibility but lower chemical resistance due to the susceptibility of PDMS to swelling in hydrocarbon fluids. The 70:30 ratio achieved an optimal balance: PFPE provided the primary chemical resistance and low surface energy, while PDMS contributed flexibility, additional hydrophobicity, and promoted more uniform nanoscale phase separation.

Nanofiller loading: The optimal nano-silica loading was 3 wt%. At this loading, the particles were well-dispersed and provided maximum tortuosity enhancement and mechanical reinforcement without creating agglomeration defects. Below 3 wt%, the reinforcing and barrier effects were insufficient to fully offset the strength reduction caused by the soft fluorine-silicon segments. Above 3 wt%, particle agglomeration became increasingly prevalent, creating voids and defect sites that compromised both mechanical and barrier properties. The surface modification of nanoparticles with FAS-13 was essential: unmodified hydrophilic nano-silica showed poor dispersion in the fluorinated matrix and actually increased fluid uptake by creating hydrophilic particle-matrix interfaces.

Curing protocol: The two-stage curing protocol (80 degrees Celsius pre-gel followed by 120 degrees Celsius post-cure) was critical for achieving the optimal nanoscale phase-separated morphology. A single-stage high-temperature cure (directly at 120 degrees Celsius) resulted in faster gelation that trapped the fluorine-silicon segments in a more homogeneous state with less pronounced phase separation, leading to higher fluid uptake and lower contact angle. A very low-temperature cure (below 70 degrees Celsius) allowed excessive phase separation and domain coarsening before gelation, resulting in larger domains and reduced mechanical properties. The 80 degrees Celsius pre-gel temperature provided the optimal balance of mobility for controlled phase separation and sufficient reaction rate to reach the gel point before domain coarsening.

4. Conclusion

This work presents a fluorine-silicon co-modified epoxy-polyurethane hybrid adhesive engineered for resistance to cutting-fluid penetration and long-term waterproof protection of electronic devices. Through systematic formulation design and characterization, the following conclusions can be drawn:

First, the covalent integration of perfluoropolyether and polydimethylsiloxane segments into an epoxy-polyurethane interpenetrating network, combined with silane-functionalized hydrophobic nano-silica, creates a multi-scale barrier that resists cutting-fluid penetration through three synergistic mechanisms: low surface energy that suppresses surfactant-assisted interfacial wetting, nanoscale phase-separated fluorine-silicon domains that increase diffusion tortuosity, and a highly crosslinked network that restricts molecular mobility and fluid uptake. The optimized formulation (20 wt% FPU-2 prepolymer with 70:30 PFPE:PDMS ratio, 3 wt% FAS-13 modified nano-silica, two-stage cure) achieved equilibrium cutting-fluid uptake of only 0.52 wt% at 60 degrees Celsius, compared to 3.61 wt% for a control epoxy.

Second, the fluorine-silicon modified adhesive retained 84% of its initial lap-shear strength after 1000 hours of cutting-fluid immersion at 60 degrees Celsius, with failure remaining cohesive within the adhesive rather than interfacial. This represents a substantial improvement over the control epoxy (28% retention, interfacial failure) and commercial silicone (33% retention). Microstructural and spectroscopic analyses confirmed that the improved performance stems from the suppression of interfacial debonding: cutting-fluid components were prevented from reaching and accumulating at the adhesive-substrate interface, and the fluorine-silicon network showed no evidence of chemical degradation during immersion.

Third, a coupled diffusion-debonding ageing model was developed that accurately predicts the time-dependent strength degradation of adhesive joints under cutting-fluid exposure. The model captures the distinction between rapid interfacial debonding (dominant in conventional adhesives) and slow bulk degradation (dominant in the fluorine-silicon modified adhesive). Arrhenius extrapolation predicts more than a tenfold improvement in service lifetime at room temperature, demonstrating the potential of this materials design approach for long-term electronic reliability in harsh manufacturing environments.

Future work should extend the evaluation to include combined stressors such as thermal cycling, mechanical vibration, and electrical bias, which may act synergistically with chemical exposure in real-world electronic packaging applications. The adhesion performance on additional substrate materials relevant to electronic packaging (copper, FR-4, polyimide, and liquid crystal polymer) should also be evaluated. The environmental and health implications of fluorinated adhesive systems, including potential release of per- and polyfluoroalkyl substances during manufacturing and end-of-life, warrant careful consideration and may motivate the exploration of fluorine-free alternatives such as long-chain alkyl siloxanes or bio-based hydrophobic modifiers.