Hand-considered hypotheses, 3501–3600
Adhesives and joining technology: adhesion mechanisms, cure, surface preparation, joint design, durability, and the interaction of bonded joints with fasteners
Written one at a time, not composed. The logical form is printed above every entry because the forms, not the subjects, are what vary: some claim a quantity is misattributed, some assert a bound, some a threshold, some invert a standard practice, several are null or no-go statements that would delete a family of tests if true, and a few use a scaling or limit argument to show that two measurements are one. None has been tested. Numbers followed by “(estimate)” are order-of-magnitude expectations, not measurements; there are no citations here, and nothing in this file is a datum.
Wetting, surface energy, and what a bond is
Misattribution
H3501. Lap-shear strength on a blasted metal is credited to the treatment's chemistry, yet much of the run-to-run scatter tracks the age of the oxide, which hydrates within hours of blasting (estimate). Bonding panels at fixed chemistry across a delay series of hours to days (estimate) should move strength further than swapping primers does.
Hidden term
H3502. Every surface energy quoted from contact angles contains the probe liquid's acid-base term, and in a humid room the dominant acid-base partner is adsorbed water rather than the substrate. Measuring one surface with three probe liquids of different acidity should give three values whose spread is the water term, not the material.
Identity
H3503. Mechanical interlock and dispersive adhesion are taught as competing mechanisms, but on roughness finer than the polymer's entanglement length they are the same contact area counted twice. A roughness series whose wavelength crosses the entanglement length should show the failure locus moving into the adhesive with no change in surface chemistry.
Threshold
H3504. There is a surface-energy difference below which a primer cannot help, because the weak layer sits on the adherend side of the interface and the primer merely bonds to it. Priming substrates of decreasing energy should bring strength down to the unprimed value at a common substrate energy rather than at a common primer.
Inversion of standard practice
H3505. Coupling agents are laid down thicker for "full coverage", yet beyond a monolayer the silane builds a hydrolytically weak interphase, so dry strength rises while hot-wet strength falls. A thickness series from sub-monolayer to tens of nanometres (estimate), tested dry and after boiling, should cross over.
Scale argument
H3506. Adhesion hysteresis and viscoelastic crack blunting are two descriptions of one loss whenever the crack speed is comparable to the polymer's relaxation rate. A peel test over four decades of rate should collapse onto a single curve against reduced speed, which would make a contact-angle loop a mechanical measurement.
Selection effect
H3507. Published joint strengths are drawn disproportionately from specimens that failed cohesively, because an adhesive failure is discarded as a bad bond rather than reported. A laboratory that reports every specimen should show a lower mean and a wider distribution than one that reports only cohesive failures at the same process.
Symmetry claim
H3508. A balanced single-lap joint has a symmetric shear-lag solution, so asymmetry in the measured adherend strains belongs to the fixture rather than to the adhesive. Gauges on both adherends of a nominally aligned joint should differ only when the grips are shimmed.
Conservation claim
H3509. In a roller peel test the supplied energy divides into interface work, plastic bending of the backing, and friction in the roller, and the third term is almost never measured. A roller of known bearing drag, run at two loads, should remove the friction term and leave a peel energy independent of the roller's condition.
Hysteresis
H3510. A surface with large contact-angle hysteresis should not recover its dry strength on redrying, because the pinning that holds the receding contact line also holds water at the interface. Wetting and drying one joint repeatedly should give a strength that falls on each cycle and never returns to its first value.
Trade-off
H3511. Roughness raises joint strength by area and lowers it by concentrating stress at asperity peaks, and the two terms peak at different roughness. A roughness series at fixed adhesive should show an optimum that moves with the adhesive's modulus, which is what two opposing terms look like.
Limit case
H3512. The zero-thickness limit of a bondline is not the strongest joint but a meaningless one, since the useful limit is the thickness at which the adhesive's process zone just fits inside the gap. Below that thickness measured strength should stop depending on the adhesive's toughness, because the joint has become flaw-controlled.
Cure: kinetics, vitrification, inhibition
Misattribution
H3513. The exotherm peak in a thick bondline is read as a kinetic event, but its height and timing are set by heat lost into the adherends. Bonding the same adhesive between aluminium and between a foam of similar thickness should move the peak by more than any single-rate-constant model allows.
Hidden variable
H3514. Conversion at the centre of a bondline differs from the coupon because the adherends act as heat sinks, and the difference is governed by the Biot number rather than by the resin. Curing at two adherend thicknesses and measuring Tg at the mid-plane should separate the two.
Bound
H3515. The residual stress a bondline can develop is bounded by the adhesive's shear modulus at gelation multiplied by its free shrinkage strain, so a stiff adhesive that gels late can develop less stress than a soft one that gels early. Measuring the curvature of thin adherend pairs against gel time should show the ranking by modulus alone failing.
Threshold
H3516. Anaerobic adhesives cure only where the clearance is small enough that oxygen ingress is consumed faster than it arrives, which makes the joint's own fit the cure controller. Curing one grade across a clearance series, with hardness measured at the centre, should show a sharp transition rather than a gradual fall.
Reversal
H3517. Faster heating can leave a diffusion-limited resin more fully cured than slow heating, because the glass transition lags the ramp and mobility persists longer than at equilibrium. Two ramps to the same hold, with Tg and conversion measured afterwards, should invert the usual expectation that slow is more complete.
Identity
H3518. The induction period before a peroxide-cured adhesive gels and the inhibition layer at its open surface are one radical-scavenging process sampled at two depths. A dissolved-oxygen series in a sealed cell should move the induction time and the inhibition thickness together.
Count
H3519. An isothermal DSC cure curve supports at most three independent kinetic parameters before the choice of baseline absorbs the rest, so a five-constant model is fitting the pan. Refitting the same data with three baselines of different curvature should move the rate constants further than their confidence intervals.
Causal ordering
H3520. Loss of tack is used to infer conversion, but tack can vanish before gelation when a low-molecular-weight fraction segregates to the surface and crystallises there. Surface composition and bulk conversion measured on the same cure series should show tack loss preceding gelation by a measurable interval.
Unit trap
H3521. Open time quoted in minutes belongs to a thin film; a thick bead keeps its surface open far longer because the volatile must diffuse out through its own depth. A data sheet open time applied to a bead of another thickness therefore carries an unstated per-thickness normalization, and a thickness series would show the number moving.
Null claim
H3522. Post-cure beyond full conversion should change no mechanical property except through the moisture it drives off, so a post-cured and a fully cured but unpost-cured specimen tested dry ought to be identical. Any difference should vanish when both are dried to the same mass.
Threshold
H3523. UV cure depth has a ceiling set by absorbance and by photoinitiator bleaching, not by lamp power, so dose beyond that ceiling only over-cures the surface. A dose series at fixed thickness should plateau in bottom hardness while surface hardness keeps climbing and yellowing begins.
Reversal
H3524. Adding photoinitiator speeds UV cure until it begins absorbing the light that would have reached the depth, after which through-cure falls while surface cure rises. A concentration series at fixed thickness should show an optimum, and the optimum should move to lower loading as the section thickens.
Artefact mimicking effect
H3525. Cure shrinkage measured by density carries the thermal contraction between cure and room temperature, which for a hot-cured adhesive can be a large share of the total. A buoyancy measurement made at cure temperature should separate shrinkage from contraction.
Hidden term
H3526. The heat a curing adhesive delivers to a bondline is reduced by the latent heat of the water or solvent it desorbs from the adherend, so the measured exotherm is reaction minus desorption. Running the same adhesive on an oven-dry and an equilibrated adherend should shift the exotherm without changing conversion.
Conservation claim
H3527. In a two-part adhesive the mix ratio sets stoichiometry, but the joint's composition also includes the resin left in the static mixer, which is not at the metered ratio. Weighing the mixer's holdup and analysing it should show that short shots through a long mixer are off-ratio even when the meter is correct.
Misattribution
H3528. Off-ratio adhesive is blamed on the meter, but in a cartridge system the first bead is rich in one component because the two streams wet the nozzle differently. Discarding the first few centimetres and testing the next should remove the effect without touching the meter.
Hidden variable
H3529. Moisture-cure polyurethane's rate depends on humidity at the bondline, which inside a closed joint is set by the adherend's moisture content rather than by the room. Bonding oven-dry and equilibrated wood at the same room humidity should give different cure times and different bubble content.
Bound
H3530. The strength of a moisture-cured joint cannot exceed that of the urea-rich skin formed at its exposed surface, because failure starts there. A joint whose exposed fillet is removed before testing should be stronger than one tested with the skin intact.
Threshold
H3531. Cyanoacrylate sets quickly on a neutral surface but is quenched on an acidic one, so the controlling variable is the surface's acid content rather than its roughness. Papers buffered to a range of pH values should show a sharp strength transition at a common pH.
Surface preparation and its limits
Misattribution
H3532. Grit blasting's benefit is credited to surface area, yet most of it comes from removing a weak oxide, so a chemically polished surface of the same roughness should bond as well. Blasted, polished, and blasted-then-polished surfaces matched for roughness should separate area from chemistry.
Hidden variable
H3533. The shelf life of a freshly treated surface is really the adsorption of airborne organics, which depends on the room's ventilation and occupancy more than on the metal. Coupons treated together and exposed in two rooms of equal humidity but different air exchange should give different strengths after the same delay.
Bound
H3534. No preparation can raise joint strength above the cohesive strength of the weakest layer in the stack, and treatment beyond that point only moves the failure locus. A dose series should plateau in strength while the locus travels from interface to cohesive.
Threshold
H3535. Plasma treatment saturates and then over-oxidises, so its dose response has a maximum rather than a plateau, set by the balance of radicals against chain scission. Bonding immediately and after a delay should show the optimum shifting to lower dose as the delay lengthens.
Inversion of standard practice
H3536. Thicker anodic layers are conventionally better for durability, but under peel a thick brittle oxide fractures and hands the crack a sharp interface, so a thin oxide can outperform it. An anodising-time series tested in peel and in shear should rank the two modes oppositely.
Identity
H3537. The water-break test and the contact angle measure the same contamination, and where they disagree the cause is roughness rather than chemistry. A set of surfaces of controlled roughness and controlled contamination should show the two tests diverging only on the rough ones.
Selection effect
H3538. Coupons that pass a water-break test are often the roughest ones, and roughness also bonds well, so the test looks predictive while being confounded. Testing smooth and rough surfaces at matched contamination should dissolve the correlation.
Unit trap
H3539. A surface energy quoted in mN/m from one probe liquid is not the number another probe liquid would give when the acid-base term dominates, so two laboratories can report different values for one surface. Only the probe set reported beside the number makes the values comparable.
Causal ordering
H3540. Primer solvent must leave before the adhesive is applied, and residual solvent rather than primer chemistry sets early strength. A drying-time series at fixed primer and fixed adhesive should show strength rising and then falling as solvent is trapped.
Null claim
H3541. Abrading a thermoplastic adherend should change nothing once the surface is solvent-wiped, because the debris is removed and the chemistry is untouched. A two-by-two of abraded and wiped surfaces should show the abrasion effect disappearing with the wipe.
Hysteresis
H3542. Contamination re-forms faster on a wiped surface than on a freshly blasted one, because the lower-energy wiped surface adsorbs organics more strongly. Cleaning and ageing the same coupon twice should show the second layer growing faster than the first.
Trade-off
H3543. Laser texturing raises bond strength by interlock and simultaneously creates a brittle heat-affected layer, so depth and damage grow together. A pulse-energy series with strength and cross-sectional damage both measured should show an optimum rather than a monotonic gain.
Limit case
H3544. At infinite surface energy the adhesive spreads spontaneously, but spreading is not bonding, and the limit case makes the difference visible: a fully wetting liquid with no cohesion gives a joint of zero strength. Wetting a surface with a series of liquids of rising surface energy and falling molecular weight should show strength peaking before wetting is complete.
Cross-domain analogy
H3545. A primer bath hydrolyses and condenses like a sol, so its age should behave like sol age: viscosity rises, film thickness rises, and bond strength falls after a shelf life. Bath turbidity plotted against joint strength over one bath's life should show the two anti-correlated.
Artefact mimicking effect
H3546. The gain credited to flame treatment of a polymer may be the flame's heat annealing the surface rather than its oxidation, since both change wettability. Flame, hot air, and cold plasma matched for surface oxygen should separate heat from chemistry.
Joint design and the single-lap problem
Misattribution
H3547. The fall of lap-shear strength with overlap length is charged to shear lag, but at long overlaps failure starts at the spew fillet, whose size changes with the squeeze applied during bonding. Joints of equal overlap and different fillet size should differ in a way shear lag cannot predict.
Bound
H3548. A well-designed bonded joint cannot carry more than the adherend's yield load, and a report above that ceiling is measuring strain hardening or grip friction. Joints made with adherends of two yield strengths should track the adherend rather than the adhesive.
Threshold
H3549. Below a critical bondline thickness strength is flaw-controlled and above it void-controlled, so an optimum exists and its position moves with the adhesive's toughness. A thickness series in two adhesives of different toughness should show optima at different thicknesses.
Inversion of standard practice
H3550. Thick bondlines are specified to tolerate poor fit, but in most structural adhesives strength falls with thickness once the process zone no longer spans the gap, so a joint designed around a loose tolerance can be weaker than one designed around a tight one. Bonding the same adherends at two fit tolerances, with thickness measured after cure, should show the loose joint failing first.
Symmetry claim
H3551. The through-thickness stress in a single-lap joint is antisymmetric about the mid-plane, so a symmetric strain reading through the thickness indicates a fixture offset. Gauges on both faces of one adherend should differ in sign and not in magnitude when the load path is aligned.
Scale argument
H3552. Whether a joint is designed by fracture mechanics or by strength depends on the process zone size relative to the bondline, and a thin bondline in a tough adhesive is always strength-controlled. One tough adhesive tested across a thickness series that crosses the process-zone size should show the failure criterion changing.
Count
H3553. Once the adhesive is fixed, a lap joint's capacity has at most two independent geometric parameters, overlap and bondline thickness, so a design of experiments over three is turning a redundant knob. A full factorial should show the third factor's effect vanishing when the first two are held.
Causal ordering
H3554. The spew fillet forms during squeeze and sets peel resistance, so it precedes the overlap in importance for a given adhesive. Trimming the fillet after cure should cost more strength than shortening the overlap by the fillet's length.
Hidden term
H3555. Single-lap shear strength contains a peel component from load-path eccentricity, and shimming the grips to align the paths removes part of what is reported as shear strength. The same joints tested aligned and misaligned should differ by an amount that scales with adherend thickness.
Unit trap
H3556. Strength per bonded area and strength per overlap length differ by the bondline thickness, so a data sheet mixing the two conventions can be wrong by a factor equal to that thickness in millimetres. Converting a quoted value between conventions exposes whether the sheet used a nominal thickness.
No-go claim
H3557. No single-lap specimen can measure adhesive shear strength, because the stress state is mixed and failure begins at the fillet. The honest conclusion is to stop quoting shear strength from single lap; a thick-adherend or napkin-ring specimen reports a different and better-defined number.
Trade-off
H3558. A thicker adherend reduces peel from eccentricity and raises the bending stiffness that prevents the joint from conforming, so a hybrid structure has an optimum adherend thickness. A thickness series at fixed overlap should show strength peaking and then falling.
Threshold
H3559. There is a gap above which joint compliance is dominated by the adhesive and below which it is dominated by the adherends, and the two regimes need different design formulas. Joint stiffness measured across a gap series should show a kink at the crossover rather than a smooth curve.
Reversal
H3560. A scarf joint is meant to eliminate peel by tapering the adherend, but a shallow scarf thins the adhesive into a flaw-sensitive layer, so a steeper scarf can be stronger. A scarf-angle series in two adhesives of different toughness should show the optimum angle moving with toughness.
Conservation claim
H3561. In a tubular joint the torque is carried partly by adhesive shear and partly by friction of the press fit, and a torque-only test cannot separate them. Instrumenting the tube in torsion and again in tension should give two different apparent shear strengths.
Identity
H3562. Cleavage and peel are the same loading at different adherend thickness ratios, so a joint weak in peel should be weak in cleavage once thickness is scaled. A thickness series tested in both modes should collapse onto one curve.
Optimal point
H3563. There is an overlap length beyond which extra length carries no load because the shear-lag length has been exceeded, and that length is set by the adhesive's shear modulus against the adherend's stiffness. Strain gauges along a long overlap should show the load reaching zero before the end.
Durability, water, and time
Misattribution
H3564. Hot-wet strength loss is charged to hydrolysis of the adhesive, but in many joints water displaces the adhesive from the interface and the loss reverses on drying. The same joints tested wet, dried, and re-wet should recover if displacement dominates and not recover if hydrolysis does.
Hidden variable
H3565. A joint's strength follows its humidity history rather than its current humidity, because the diffusion time constant is longer than the test. Joints equilibrated to the same final humidity by different paths should not have the same strength.
Bound
H3566. Strength loss from moisture cannot saturate before the bondline itself does, so a test shorter than the diffusion time reports a rate rather than a durability. Joints aged to different fractions of equilibrium should show the loss still climbing when plotted against square-root time, with no plateau.
Threshold
H3567. Below a stress intensity no subcritical crack growth occurs in the adhesive even at saturation, so a durability test should report a threshold rather than a time to failure. A load series at full saturation should show arrest below a load that the time-to-failure ranking does not reveal.
Reversal
H3568. Cyclic loading can outlive static loading at the same peak stress when each unload lets water redistribute away from the crack tip, so fatigue is not always the more damaging case. Paired static and cyclic tests at matched peak stress and matched saturation should show the cyclic life longer in a water-sensitive adhesive.
Identity
H3569. Creep under load and stress relaxation under fixed displacement are one compliance seen at two boundary conditions, so a joint that creeps is a joint that relaxes. Measuring both on the same specimen should give relaxation times that agree.
Selection effect
H3570. Joints that survive accelerated ageing tend to have thicker bondlines because the diffusion path is longer, so a test that keeps survivors favours the geometry that is weakest in shear. Sectioning the survivors and measuring bondline thickness should make the bias visible.
Hysteresis
H3571. Moisture uptake and loss do not follow the same curve, and the gap is usually called non-Fickian diffusion when it may be damage. Running absorption, desorption, and a second absorption on the same specimen should leave the gap only where cracks formed.
Causal ordering
H3572. In salt-sprayed joints adherend corrosion precedes bond failure, and the adhesive failure seen on the fracture surface is really the corrosion product layer. Cross-sections taken before failure should show corrosion ahead of the crack tip.
Trade-off
H3573. A toughened adhesive gains peel resistance and loses creep resistance, because the rubber particles that blunt cracks also lower the modulus. A series of rubber contents tested in peel and in creep should show the two rankings inverting.
Threshold
H3574. Thermal cycling damages a joint only when the adherend mismatch strain exceeds the adhesive's yield strain, so cycling below that amplitude should do nothing regardless of count. Two amplitudes, one below and one above the estimated threshold, should differ in crack density after the same count.
Limit case
H3575. At a stress well below the adhesive's yield a joint should never fail by creep, yet it does, and the limit case is the flaw that concentrates stress. A specimen with a controlled artificial flaw should fail below the unflawed joint's creep limit, which makes the creep limit a specimen property rather than a material one.
Cross-domain analogy
H3576. A bondline takes up water like a constrained hydrogel rather than a bulk polymer, because the adherends forbid swelling. A thin joint should therefore reach a different equilibrium uptake than a free film of the same adhesive, which bulk diffusion cannot predict.
Artefact mimicking effect
H3577. Strength lost in a hot-water bath includes leaching of unreacted monomer, which changes the adhesive's composition and does not occur in humid air. Ageing joints in water and at the same temperature and water activity in humid air should separate leaching from hydrolysis.
No-go claim
H3578. A single accelerated condition cannot rank two adhesives for durability when their moisture diffusion coefficients differ, because the specimens reach different saturation states at the same exposure time. Ranking at one time and ranking at saturation should disagree, which makes the single-condition ranking meaningless.
Bound
H3579. In a corrosive environment a joint's strength is bounded below by the corroded adherend rather than by the adhesive, once the corrosion layer is thick enough. An exposure series with adherend section loss measured should show failure moving into the metal.
Misattribution
H3580. The knee in a joint's strength-versus-exposure curve is read as the onset of interfacial attack, but it can be the moment a thin specimen's edges dry, since thin specimens equilibrate through their edges. Sealing the edges should move or remove the knee.
Fasteners, hybrids, and the mechanical interface
Misattribution
H3581. Load sharing in a hybrid bolted-and-bonded joint is credited to the adhesive, but much of it passes through friction from clamp load, so the adhesive's arithmetic share overstates its real one. Instrumenting bolts and bondline strains in one joint should expose the friction term.
Hidden term
H3582. Bolt preload in a bonded joint relaxes because the adhesive creeps under the clamp, so the bolt loses load with no embedment and no vibration. Clamp load followed over time in a bonded and an unbonded joint at one temperature should show the bonded one losing faster.
Inversion of standard practice
H3583. Bolts added to a bonded joint are assumed to add safety, but the holes concentrate stress and can initiate cracks in the adhesive, so the hybrid can be less durable than the bond alone. Bonded and bonded-plus-bolted specimens should rank oppositely in fatigue and in static strength.
Threshold
H3584. Above a gap, a bolted joint's clamp load cannot be held by an adhesive shim because the shim creeps; below it, the shim carries the preload and the bolt barely matters. A gap series at fixed torque should show clamp retention falling sharply at a gap set by the shim's creep compliance.
Identity
H3585. Fretting at a bolted interface and fatigue cracking at a bondline edge are one dissipative process at two length scales, both driven by relative slip under a clamp. A joint that frets at its hole should show the same slip-amplitude dependence as a bondline that cracks at its edge.
Count
H3586. A bolted joint's preload has at most two independent loss mechanisms, embedment and thermal mismatch, once the fastener and stack are fixed, so a model with four loss terms is fitting the test. Preload measured at two temperatures and two surface finishes should account for all of it.
Scale argument
H3587. The slip amplitude that drives fretting scales with the fastener's clearance, so a tighter hole should fret less at the same load. Two clearance classes at one torque and one load should differ in damage depth.
Trade-off
H3588. Higher preload raises friction capacity and lowers the fatigue benefit of the clamp, because bolt mean stress rises along with the slip load. A preload series with static capacity and fatigue life both measured should place the two optima at different preloads.
Causal ordering
H3589. Hole quality precedes clamp load in importance for a bolted composite joint, because a damaged hole relieves the clamp and starts the failure. Joints with reamed and with drilled holes at one torque should differ more than joints at two torques with one hole quality.
Unit trap
H3590. Torque becomes preload through a nut factor that hides thread and bearing friction, so a preload quoted from torque is a friction measurement. A strain-gauged bolt recorded alongside torque and lubrication should show the nut factor moving with the lubricant rather than the preload.
Hysteresis
H3591. A bolted joint's load-displacement curve on tightening differs from its curve on loosening, and the gap is the embedment that occurred between them rather than elasticity. Cycling one joint through torque and back should show the gap growing for a few cycles and then saturating.
Reversal
H3592. Interference-fit fasteners are assumed to improve fatigue by preloading the hole, but in a composite the interference can crush fibres and start a delamination, so the fit can shorten life. Two fit classes at one clamp should rank oppositely in metal and in composite.
Artefact mimicking effect
H3593. A bolted connection's measured stiffness includes the test fixture's compliance, and when the fixture is stiffer than the joint the joint looks stiffer than it is. The same joint tested in two fixtures should yield two stiffnesses differing by the fixture term.
Null claim
H3594. Thread-locking compound should leave clamp load at assembly unchanged and act only on retention afterwards, so an immediate torque-tension test should be identical with and without it. Any difference at assembly is the compound behaving as a lubricant.
Artefact mimicking effect
H3595. Bearing failure loads measured in a single-rivet pull test include friction from the grip, which in a tightly clamped specimen can be a large fraction of the total, so the reported bearing strength belongs partly to the fixture. Releasing the clamp, or putting a release film under the washers, should lower the apparent strength.
Optimal point
H3596. A self-piercing rivet has an optimum die depth, since a shallow die leaves the rivet unflared and a deep one thins the lower sheet. A die-depth series with cross-sections and lap-shear tests should show an optimum that moves with sheet thickness.
Identity
H3597. Clinching and self-piercing riveting form the same mechanical interlock with different tools, so their strength should scale with the same geometric parameters. Interlock depths made by both processes should collapse onto one strength curve.
Threshold
H3598. Weld-bonding beats welding alone only above a bondline thickness at which the adhesive can carry load without being squeezed out of the weld zone. A thickness series should show the hybrid benefit appearing abruptly and then falling as the adhesive layer becomes the weak link.
Conservation claim
H3599. In a hybrid joint the load is conserved between adhesive and fasteners at every instant, so the sum of their measured shares must equal the applied load. A test in which the sum falls short is measuring the fixture, and the deficit should appear at the moment of load transfer.
Scale argument
H3600. The distance over which a hybrid joint transfers load is set by its stiffer path, so the bolt dominates near the hole and the adhesive far from it, and the two zones scale with different lengths. Strain gauges along the overlap should show the transfer length changing when the bolts are loosened, which no single-joint model predicts.
Status
One hundred entries, each considered on its own. The labels above the claims are there so that structural repetition is visible: two entries on the same subject should not share a form unless the second is genuinely the same argument applied elsewhere. The subject list was deliberately broad — wetting, cure, preparation, geometry, water, and fasteners — because adhesives fail at whichever of those the specification left unexamined.
Adhesives reward this treatment because so many of their standard numbers are test-geometry artefacts dressed as material properties: a shear strength from a mixed-stress specimen, a surface energy from one probe liquid, a cure schedule from a thin film, a durability ranking from one exposure time. Several entries above attack one of those directly. None has been tested; a hypothesis list is worth the experiments it provokes, and nothing more.
If one of these deserves a bench, it is H3557 — that a single-lap specimen cannot report an adhesive shear strength, which is checkable in a day by testing one adhesive in single lap and in a napkin ring and finding two numbers that cannot both be called shear.
Incoprea Research · hand-considered hypotheses 3501–3600 ·
3 November 2026 · ~/Desktop/Nov3/