Adhesive Bonding in Manufacturing: The Complete Engineering Guide to Structural Adhesives, Joint Design, Surface Science, and Failure Analysis

A dense equation-driven engineering deep-dive into structural adhesive bonding — covering thermodynamic work of adhesion, Volkersen shear-lag, Goland-Reissner…

·

Adhesive bonding is the oldest joining technology known to humanity — birch-bark tar was used to haft stone tools 200,000 years ago — and simultaneously the most modern: the Boeing 787 Dreamliner contains over 60 kilometres of bonded joints replacing nearly 1.5 million mechanical fasteners. Between these extremes lies a discipline governed by interfacial thermodynamics, polymer physics, continuum fracture mechanics, and electrochemical surface science.

This guide covers everything from the thermodynamic origin of adhesion to the practical selection of adhesives for manufacturing. Whether you are bonding 3D-printed PLA brackets, designing a bonded aluminium spaceframe, or specifying a structural adhesive for a medical device, the principles are the same — and the equations matter.


1. The Thermodynamic Origin of Adhesion

Adhesion begins with wetting. An adhesive cannot form a load-bearing interface with a substrate it does not intimately contact. The thermodynamics of wetting is governed by the Young equation, which balances the interfacial tensions at the three-phase contact line:

where \gamma_{SV} is the solid-vapour surface energy, \gamma_{SL} the solid-liquid interfacial energy, \gamma_{LV} the liquid-vapour surface tension, and \theta the contact angle. Complete wetting occurs when \theta \to 0 — the adhesive spreads spontaneously across the substrate.

The thermodynamic work of adhesion W_A — the reversible work required to separate unit area of interface — is given by the Young-Dupré equation:

For a typical epoxy on aluminium ( \theta \approx 5^\circ , \gamma_{LV} \approx 45 \text{ mJ/m}^2 ), this gives:

This ~90 mJ/m² is two orders of magnitude smaller than the practical fracture energy of a structural adhesive joint (~1–10 kJ/m²). The enormous discrepancy proves that thermodynamic adhesion is necessary but not sufficient — the dominant contributions to practical joint strength come from energy dissipation mechanisms (plastic deformation, fibril formation, crack bridging) that thermodynamics alone does not capture.

Surface Energy Components: Owens-Wendt Model

Real surfaces have both dispersive (van der Waals) and polar (hydrogen bonding, acid-base) contributions. The Owens-Wendt geometric mean model decomposes surface energy:

where superscripts d and p denote dispersive and polar components. By measuring contact angles with two probe liquids of known \gamma_L^d, \gamma_L^p (typically water and diiodomethane), one solves for \gamma_S^d and \gamma_S^p of the unknown solid.

Material · \gamma_S^d (mJ/m²) · \gamma_S^p (mJ/m²) · Total \gamma_S

PTFE (Teflon) · 18.6 · 0.5 · 19.1

Polypropylene · 29.0 · 1.0 · 30.0

Polycarbonate · 30.5 · 8.4 · 39.0

Epoxy (cured) · 35.0 · 8.0 · 43.0

Aluminium (degreased) · 35.0 · 4.0 · 39.0

Aluminium (acid-etched) · 38.0 · 12.0 · 50.0

Glass (clean) · 30.0 · 34.0 · 64.0

The practical rule: A substrate with \gamma_S > 40 \text{ mJ/m}^2 is generally bondable; below 30 mJ/m², surface treatment (plasma, corona, primer) is mandatory. PTFE at 19 mJ/m² must be chemically etched with sodium naphthalenide to achieve any useful bond strength.


2. Mechanics of Adhesive Joints

2.1 The Single-Lap Joint and Volkersen's Shear-Lag Analysis

The workhorse configuration in adhesive testing and design is the single-lap joint. Two adherends of thickness t , overlapped by length L , bonded by an adhesive layer of thickness h and shear modulus G_a . Under remote tensile stress \sigma_0 , the differential straining of the two adherends loads the adhesive predominantly in shear — but the shear stress is highly non-uniform along the overlap.

Volkersen's shear-lag solution (1938) gives the shear stress distribution:

where the dimensionless parameter \omega (the shear-lag coefficient) is:

Here E is the adherend Young's modulus (assumed equal for both — the solution generalises for dissimilar adherends). The stress concentration factor at the ends of the overlap is:

For \omega > 3 (a long, stiff overlap), \coth{\omega} \approx 1 , so \text{SCF} \approx \omega — the peak stress scales nearly linearly with overlap length. This is why doubling the overlap length does not double the load capacity — the additional length carries progressively less stress.

Numerical example: Aluminium adherends, t = 2 \text{ mm} , E = 70 \text{ GPa} , bonded with a toughened epoxy ( G_a \approx 500 \text{ MPa} ), overlap L = 25 \text{ mm} , bond line h = 0.2 \text{ mm} :

The peak shear stress at the ends is 2.41× the average — a significant concentration that governs failure initiation.

2.2 Peel Stresses: The Goland-Reissner Factor

Volkersen neglected the bending moment induced by the load path eccentricity in a single-lap joint. In reality, the offset between the two adherend centre-lines produces a bending moment that generates peel stresses (through-thickness tensile stress) at the overlap ends. These peel stresses are the primary cause of failure in single-lap joints because adhesives are typically 10–100× weaker in tension than in shear.

Goland and Reissner (1944) introduced the moment factor k that captures the reduction in edge bending moment due to the overlap constraint:

where:

The maximum peel stress at the overlap end is:

where \lambda_p = \sqrt[4]{E_a / (4 D h)} is the peel wavelength parameter, D the adherend flexural rigidity, and E_a the adhesive tensile modulus. The function f(\lambda_p L) approaches 1 for long overlaps.

The takeaway: Peel stresses are minimised by:

  1. Reducing adherend thickness t — less bending eccentricity
  2. Increasing overlap length L — the constraint increases, reducing k
  3. Using a tapered or scarfed adherend at the overlap ends
  4. Selecting a lower-modulus adhesive — distributes the peel over a larger zone

2.3 Bond Line Thickness Optimisation

Contrary to intuition, a thinner bond line generally produces a stronger joint — but with a minimum practical limit. The shear strain in the adhesive under a given stress is:

For a fixed joint displacement, a thinner bond line concentrates the strain into a smaller volume, requiring higher stress to produce the same deformation. In practice:

The strength reduction with increasing bond line is partly explained by the maximum flaw size scaling: according to linear elastic fracture mechanics, the failure stress for a bond line containing intrinsic flaws of characteristic size a_0 \propto h scales as:


3. Adhesive Chemistries: A Systematic Comparison

3.1 Epoxy Adhesives

Chemistry: Epoxide ring-opening polymerisation with amine, anhydride, or mercaptan hardeners. The reaction proceeds via step-growth — one epoxide group reacts with one amine hydrogen.

Key Properties:

Selection equation — DiBenedetto's model for T_g vs. conversion:

where T_{g0} is the glass transition of the unreacted mixture, T_{g\infty} that of the fully cured network, \alpha the degree of cure, and \lambda \approx 0.5\text{–}0.8 an empirical fitting parameter. A partially cured epoxy may have a T_g 30–50°C below its fully post-cured value — critical for service temperature rating.

Best for: Structural metal and composite bonding, aerospace (film adhesives like 3M AF 163-2, Henkel LOCTITE EA 9696), high-temperature applications.

3.2 Polyurethane Adhesives

Chemistry: Isocyanate ( -\text{N}=\text{C}=\text{O} ) reaction with polyols to form urethane linkages. Moisture-cure variants use atmospheric water to generate amines that further react.

Key Properties:

The peel advantage: Polyurethanes owe their peel resistance to the two-phase morphology — hard segments (urethane/urea domains, ~5–10 nm) act as physical crosslinks and energy-dissipating filler particles within a soft polyether/polyester matrix. During peel loading, the hard domains cavitate and undergo plastic deformation, absorbing orders of magnitude more energy than the brittle fracture characteristic of unfilled epoxies.

Best for: Dissimilar material bonding (metal-plastic, plastic-glass), applications requiring vibration damping or thermal expansion accommodation, automotive windscreen bonding (Sikaflex, Betaseal).

3.3 Acrylic Adhesives (Toughened / "Second Generation")

Chemistry: Free-radical polymerisation of methyl methacrylate (MMA) or related monomers, initiated by a peroxide-amine redox system. "Toughened" variants incorporate elastomeric domains (chlorosulfonated polyethylene, core-shell rubber particles) that phase-separate during cure.

Key Properties:

Best for: High-speed production lines where surface preparation is minimal, bonding of thermoplastics and composites, magnet bonding in electric motors.

3.4 Cyanoacrylates (Instant Adhesives)

Chemistry: Anionic polymerisation of alkyl 2-cyanoacrylate initiated by trace surface moisture (weak base catalysis). The electron-withdrawing cyano ( -\text{CN} ) and ester ( -\text{COOR} ) groups make the monomer extraordinarily electrophilic — polymerisation is essentially instantaneous on alkaline surfaces.

Key Properties:

The gap limitation is geometric, not chemical: Cyanoacrylate polymerisation is so rapid that the monomer cannot flow into gaps larger than ~0.2 mm before polymerising. The exothermic reaction in a thick bond line also generates temperatures exceeding 100°C locally, causing thermal degradation and bubble formation.

Best for: Small-part assembly, wire tacking, temporary fixturing, medical device bonding (butyl and octyl cyanoacrylates are biocompatible).

3.5 Anaerobic Adhesives

Chemistry: Methacrylate monomers stabilised by dissolved oxygen. In the presence of metal ions (Fe, Cu) and absence of oxygen — i.e., inside a threaded joint — the peroxide initiator decomposes to free radicals and polymerisation proceeds.

Key Properties:

The cure mechanism dictates the application: Oxygen inhibition means the adhesive stays liquid indefinitely in the bottle (the bottle is oxygen-permeable LDPE) and cures only when confined in a metal joint. This self-regulating cure is what makes threadlockers and retaining compounds so reliable — they cure only where you want them to.

Best for: Threadlocking, cylindrical assembly (bearings, bushings), flange sealing, gasketing (form-in-place).

3.6 Silicone Adhesives

Chemistry: Condensation or addition-cure polydimethylsiloxane (PDMS). Addition-cure (platinum-catalysed hydrosilylation) produces no byproducts; condensation-cure (acetoxy, alkoxy) releases acetic acid or alcohol.

Key Properties:

Why silicones are weak but indispensable: The siloxane backbone is exceptionally flexible (rotational barrier ~0.8 kJ/mol vs. ~12 kJ/mol for C-C), giving silicones a bulk modulus that is roughly temperature-invariant across a 350°C range. They are selected not for load-bearing capacity but for sealing, gap-filling, and thermal/electrical insulation across extreme temperature ranges that would embrittle any organic polymer.

Best for: High-temperature sealing, electronic encapsulation, medical devices, glazing and curtain-wall structural silicone (SSG) in building facades.

3.7 Adhesive Selection Matrix

Property · Epoxy · PU · Toughened Acrylic · Cyanoacrylate · Anaerobic · Silicone

Lap shear (MPa) · 15–35 · 5–15 · 15–30 · 10–25 · N/A (thread) · 1–3

Peel (N/mm) · 1–15 · 5–20 · 5–12 · < 1 · < 1 · 3–10

Max service T (°C) · 220 · 120 · 130 · 80–120 · 230 · 300

Gap fill (mm) · 0.05–5 · 0.1–10 · 0.1–4 · 0.05–0.2 · 0.05–0.25 · 0.1–10

Surface tolerance · Poor · Fair · Excellent · Fair · Fair · Good

Cure time · Min–days · Min–hours · 30 s–5 min · 2–60 s · 5 min–24 h · Min–hours

₹/kg (approx.) · 800–3,000 · 500–1,500 · 1,000–3,000 · 5,000–15,000 · 2,000–5,000 · 300–2,000


4. Surface Preparation: The Four-Level Hierarchy

Adhesive joint failure is overwhelmingly interfacial — the adhesive separates cleanly from the substrate. In over 80% of bond failure investigations, the root cause traces not to the adhesive but to inadequate surface preparation. The hierarchy of surface treatments goes from simply cleaning to generating a new surface chemistry.

Level 1: Degreasing and Cleaning

Removes organic contaminants (oils, greases, release agents) that form a weak boundary layer.

Water-break test is the universal cleanliness verification: DI water sprayed onto the surface should form a continuous sheet, not discrete droplets. Any beading = residual hydrophobic contamination.

Level 2: Mechanical Abrasion

Increases surface area and removes weak oxide layers.

Surface roughness and wetting — the Wenzel equation:

where r is the roughness factor (actual surface area / projected area, always ≥ 1). For a hydrophilic surface ( \theta_Y < 90^\circ ), roughness improves wetting ( \theta_W < \theta_Y ). For a hydrophobic surface ( \theta_Y > 90^\circ ), roughness makes it worse.

Level 3: Chemical Etching

Generates a micro-rough oxide morphology optimised for mechanical interlocking.

Level 4: Primers, Coupling Agents, and Plasma

The practical rule for production: Level 1 + Level 2 is sufficient for 80% of non-structural applications. Level 3 or 4 is mandatory for primary-structure aerospace, medical implant, and safety-critical automotive bonding.


5. Durability and Environmental Degradation

5.1 Moisture Attack — The Dominant Degradation Mechanism

Water attacks bonded joints through three simultaneous mechanisms:

  1. Plasticisation: Water molecules (kinetic diameter 0.27 nm) diffuse into the polymer network, increasing free volume and reducing T_g . The Fox equation estimates the plasticised T_g :

where w_p, w_w are weight fractions of polymer and water, and T_{g,w} \approx -137^\circ\text{C} for pure water. Even 2 wt% water uptake can depress epoxy T_g by 20–30°C, dropping the service ceiling below the operating temperature — and the joint fails.

  1. Interfacial displacement: Water molecules compete with adhesive-substrate bonds at the interface. The thermodynamic work of adhesion in the presence of water is:

If W_{A,\text{wet}} < 0 , the interface is thermodynamically unstable in the presence of water — spontaneous debonding will occur given sufficient time. Epoxy-aluminium interfaces are particularly susceptible; silane primers improve the wet durability by providing covalent siloxane ( \text{Si-O-Al} ) bonds that resist displacement.

  1. Substrate corrosion: On steel and aluminium, water permeating through the adhesive causes underfilm corrosion. The corrosion products (hydrated oxides) occupy 2–4× the volume of the parent metal, generating wedging stresses that propagate interfacial debonding.

5.2 Accelerated Ageing and Lifetime Prediction

The Arrhenius model relates degradation rate k to temperature:

where E_a is the activation energy for the degradation process (typically 50–100 kJ/mol for moisture-driven epoxy degradation). By ageing joints at elevated temperatures (40°C, 60°C, 80°C, 95% RH) and measuring strength retention over time, one extrapolates to service conditions. The time-temperature superposition principle — embodied in the WLF equation for viscoelastic materials — allows prediction of decades-long lifetimes from months of accelerated data.

where a_T is the shift factor, C_1 \approx 17.4 and C_2 \approx 51.6 K for many polymers when T_{\text{ref}} = T_g .

Practical reality: A bonded aluminium joint exposed to a tropical environment (35°C, 90% RH) for 5 years typically retains only 40–60% of its initial strength if unprotected. A properly primed joint with a durable epoxy retains > 85%.

5.3 Galvanic Corrosion in Bonded Dissimilar-Metal Joints

Bonding a carbon-fibre composite ( E_{\text{corr}} \approx +0.1 \text{ V}_{\text{SHE}} ) to aluminium ( E_{\text{corr}} \approx -0.85 \text{ V}_{\text{SHE}} ) creates a galvanic couple with a driving voltage of ~1 V. If moisture bridges the bond line, the aluminium corrodes sacrificially. Mitigation strategies:


6. Testing and Failure Analysis

6.1 Mechanical Test Geometries

Test · Standard · What It Measures · Failure Mode Insight

Single-lap shear · ASTM D1002 / ISO 4587 · Apparent shear strength · Overlap length must be standardised — results are NOT transferable across geometries

Double-lap shear · ASTM D3528 · Shear with reduced peel · Minimises bending, closer to true shear

T-peel · ASTM D1876 · Peel strength · Sensitive to adhesive ductility

Floating roller peel · ASTM D3167 · Controlled peel angle · Most reproducible peel test for metal bonding

DCB (double cantilever beam) · ASTM D5528 / ISO 25217 · Mode-I fracture toughness G_{Ic} · The definitive test for adhesive fracture mechanics

ENF (end-notched flexure) · ASTM D7905 · Mode-II fracture toughness G_{IIc} · Pure shear fracture

Mixed-mode bending (MMB) · ASTM D6671 · Mixed-mode G_I/G_{II} envelope · Defines the fracture locus for design

6.2 Interpreting Failure Modes

Post-fracture examination of the failure surface is the single most important diagnostic in adhesion science. Identify the mode:

6.3 Nondestructive Testing (NDT)


7. Joint Design Rules

7.1 The Five Commandments of Adhesive Joint Design

  1. Load the adhesive in shear, not peel. Adhesives typically have peel strengths 5–50× lower than their shear strengths. If peel cannot be eliminated, use a flexible (low-modulus) adhesive and increase the bond area to reduce peel stress intensity.
  1. Maximise bond area in the loading direction. Overlap length governs shear area. A 12.5 mm overlap on a 25 mm wide joint gives 312 mm² of bond area — double the width to 50 mm and you double the load capacity (unlike doubling overlap length, which hits the Volkersen ceiling).
  1. Design for uniform stress distribution. The Volkersen/Goland-Reissner analyses show that shear stress peaks at overlap ends. Mitigation: taper the adherend ends, use a spew fillet (the adhesive squeezed out at the overlap end forms a natural fillet that reduces stress concentration by 20–40%), or design a double-strap joint that eliminates the load-path eccentricity.
  1. Control bond line thickness. Use spacer beads (glass microspheres of controlled diameter, typically 0.1–0.25 mm) in the adhesive to maintain consistent thickness. Wire spacers achieve the same goal in flat joints.
  1. Design for inspection and repair. Include witness holes, allow access for NDT probes, and avoid completely enclosed bond cavities that trap volatiles. A joint that cannot be inspected must not be used in a safety-critical application.

7.2 Common Joint Configurations

Configuration · Stress State · Efficiency · Notes

Single lap · Shear + peel · Low-medium · Simplest, but peel limits load capacity

Double lap · Near-pure shear · High · Symmetric — eliminates bending

Scarf joint · Shear + tension · Very high (can approach 100% of parent strength) · Requires precise machining; bond line is angled to the load

Step lap · Shear + compression · High (80–95%) · Practical compromise — machined steps rather than continuous taper

T-joint · Peel-dominant · Poor · Avoid if possible; use co-curing or mechanical fasteners as backup

Tubular lap (shaft-collar) · Torsional shear · Excellent · Adhesive loaded in pure shear around the circumference


8. Application Case Studies

8.1 Boeing 787 Composite Bonding

The 787's fuselage is a co-cured carbon-fibre/epoxy monocoque — the stringers (longitudinal stiffeners) are bonded to the skin in a single autoclave cycle, eliminating approximately 50,000 fasteners per aircraft compared to an equivalent aluminium barrel. The co-curing process means there is no discrete "adhesive" — the matrix resin of the prepreg serves as both composite matrix and structural adhesive, forming a continuous crosslinked network across the interface.

The principal challenge is lightning strike protection: carbon fibre has electrical resistivity \rho \approx 10^{-5} \text{ }\Omega\cdot\text{m} — six orders of magnitude higher than aluminium. A lightning strike on an all-composite fuselage can generate enough Joule heating ( I^2R ) to vaporise resin and delaminate the bond line. The solution is an expanded copper foil mesh co-cured into the outer ply — it carries the lightning current (200 kA peak, 500 MJ/Ω total action integral) to the airframe ground network without damaging the structural bond.

8.2 Automotive Structural Bonding — Mixed Joining

Modern vehicle bodies (BMW 7 Series, Audi A8, Tesla Model S/Y) use weld-bonding — a combination of structural adhesive and resistance spot welds. The adhesive provides continuous stiffness (eliminating the "oil-canning" effect of discrete spot welds), corrosion sealing, and ~30% improvement in fatigue life, while the spot welds provide fixturing during adhesive cure and act as crack arrestors — a propagating bond-line crack hits a weld nugget and stops.

A typical automotive structural adhesive is a one-part, heat-curing toughened epoxy ( T_{\text{cure}} = 160\text{–}180^\circ\text{C} , matching the electrocoat bake oven cycle). Lap shear strength on electrogalvanised steel: 25–30 MPa. The bond line thickness is controlled by the metal surface topography (typically 0.15–0.3 mm).

Fatigue life improvement from weld-bonding vs. spot welding alone, for a lap joint tested at 30% of ultimate load:

The adhesive redistributes stress away from the spot-weld notch, delaying crack initiation.

8.3 Consumer Electronics — Pressure-Sensitive Adhesives and Die Attach

Smartphones contain 10–20 different adhesive applications: display lamination (optically clear adhesive, OCA), battery fixation (stretch-release PSA strips), die attach (silver-filled epoxy or sintered Ag for power semiconductors), underfill (capillary-flow epoxy for flip-chip BGA reliability), and enclosure sealing (foam-in-place polyurethane gasketing for IP68 water resistance).

The capillary underfill process is a manufacturing marvel: a low-viscosity ( \eta \approx 0.5\text{–}5 \text{ Pa·s} ) epoxy is dispensed along one or two edges of the flip-chip die. Capillary action — driven by the surface energy of the solder bumps and substrate — draws the liquid into the 50–100 μm gap. The Washburn equation governs the flow:

where x is the flow front position, R the effective capillary radius (determined by bump pitch and standoff height), and t the time. For a 10 × 10 mm die with 150 μm bump pitch, underfill time is typically 30–120 seconds — after which the assembly enters a snap-cure oven (165°C, 5 minutes) to gel the epoxy and lock in the stress distribution that compensates for the CTE mismatch between silicon ( \alpha \approx 2.6 \text{ ppm/K} ) and FR-4 substrate ( \alpha \approx 14\text{–}18 \text{ ppm/K} ).

8.4 DIY and 3D Printing — Adhesive Bonding of FDM Parts

FDM-printed PLA and PETG parts present unique bonding challenges: layer lines act as stress concentrators, porosity provides paths for adhesive wicking (and, conversely, solvent attack), and the semi-crystalline nature of many filaments resists solvent welding.

Empirical lap-shear data (FDM PLA, 3 mm thick, 25 mm overlap, room temperature cure):

Adhesive · Lap Shear (MPa) · Failure Mode · Notes

Cyanoacrylate (ethyl) · 3–5 · Substrate — interlayer delamination · Fast, but brittle; fails at the weakest layer, not the bond line

Two-part epoxy (5 min) · 5–8 · Substrate — interlayer delamination · More tolerant of surface texture; 5-min epoxy is weaker than 24-hr

Two-part epoxy (24 hr) · 8–12 · Mixed cohesive/substrate · Best strength, but the printed part becomes the weak link

"3D Gloop" (solvent cement, MEK/dichloromethane blend) · 6–9 · Substrate (solvent-softened zone) · Chemically welds rather than bonds — the interface disappears

Hot-melt (EVA) · 1–2 · Cohesive in adhesive · Too weak for structural use; good for temporary fixturing

Polyurethane (Gorilla Glue) · 2–4 · Cohesive/adhesive mixed · Foaming can fill gaps but reduces strength; requires clamping

The key insight: For FDM parts, the substrate is almost always weaker than a properly selected adhesive. Optimise print orientation so the bond line is parallel to layer planes — placing the interlayer weakness in shear rather than tension/peel. A joint loaded in pure shear perpendicular to the build direction will typically fail at 10–15 MPa (the interlayer adhesion limit of PLA) regardless of the adhesive's nominal 25 MPa capability.

For PETG and ABS, solvent welding with the appropriate solvent (dichloromethane/MEK for ABS, cyclohexanone for PETG) produces a bond that approaches 100% of the bulk material strength because the solvent dissolves and re-fuses the polymer chains across the interface — there is no discrete bond line to fail.


9. Manufacturing Process Design

9.1 Adhesive Dispensing Methods

Method · Accuracy · Throughput · Adhesive Types · ₹ Equipment Cost

Manual (cartridge + static mixer) · ±0.5 g · Low · 2-part epoxies, PU, acrylic · ₹5,000–25,000 (gun)

Pneumatic time-pressure · ±5–10% · Medium · Most · ₹1–5 lakh

Positive-displacement (auger, piston) · ±1–3% · Medium-high · Pastes, filled epoxies · ₹5–15 lakh

Jet dispensing (piezo/pneumatic) · ±2% per dot · Very high (200+ dots/s) · Low-viscosity (< 100 Pa·s) · ₹15–50 lakh

Robotic bead (6-axis) · ±0.1 mm path · High · Most · ₹10–50 lakh

9.2 Cure Process Control

For heat-curing adhesives, the cure schedule is a time-temperature path that determines the final network architecture:

9.3 Quality Assurance

Production bonding requires:


10. Summary: The Adhesive Selection Flowchart

Ask these questions in order:

  1. What are the adherends? Metal-metal → epoxy or acrylic. Plastic-plastic → acrylic, PU, or solvent weld. Metal-plastic → PU or acrylic. Composite-composite → epoxy (same chemistry as matrix). Glass/ceramic → epoxy with silane primer.
  1. What is the service temperature? < 80°C → most chemistries work. 80–150°C → epoxy or high-temperature acrylic. 150–250°C → high-Tg epoxy, BMI (bismaleimide) adhesive. > 250°C → silicone, ceramic adhesive, or mechanical fastening.
  1. Is peel a concern? Yes → toughened epoxy, PU, or acrylic. Design to load in shear if possible.
  1. What is the production speed? < 5 s → cyanoacrylate. < 5 min → acrylic. < 30 min → fast epoxy, PU. Hours → standard epoxy, moisture-cure PU.
  1. What is the gap tolerance? < 0.1 mm → cyanoacrylate, anaerobic. 0.1–1 mm → epoxy, acrylic. > 1 mm → PU, epoxy with thick bond line formulation, or a shim/mechanical spacer.
  1. What surface preparation is feasible? Minimal → acrylic (bonds through light oil). Standard (degrease + abrade) → epoxy, PU. Maximum (etch + prime) → aerospace epoxy film.
  1. What is the environmental exposure? Hot-wet → epoxy with corrosion-inhibiting primer. UV + weather → silicone or PU (aliphatic). Chemical immersion → epoxy (high crosslink density). Thermal cycling → PU or flexible epoxy (low T_g accommodates CTE mismatch).

The bottom line: There is no universal "best" adhesive. The optimal choice is a multi-variable optimisation problem constrained by substrates, service environment, production economics, and safety. But with the thermodynamic, mechanical, and chemical principles in this guide, you have the analytical framework to make that choice with confidence — and to troubleshoot it when it fails.


Further Reading:

More FabFlow blog posts