Indentation-debonding of an Adhered Surface Layer

In many technological applications (e.g. circuitboards), relatively thin, soft polymer layers are bonded to a harder (metal) surface. A new mechanical test for the adhesive bond strength between these layers using sharp, needle-like indenters, has been developed. This test and its experimental-analytical features are described in this paper. In the central indented region of the bonded polymer, compressive deformations of the top layer and subsurface take place. At the rim of the central indented area, however, a tensile stress is induced in the interfacial bond between the top layer and the substrate. Debonding occurs when this stress exceeds the bond strength. A debonded polymer surface layer displays Newton's rings because it diffracts light. Observations indicate that the debonded segment of the surface layer behaves as an elastic plate between the rim of indentation and the unfailed bonded region. The “peeling moment” capacity of adhesive at the bonded edge limits the bond resistance. Considering the latter as a constant parameter (a characteristic of the bond) a relation predicting the debonded radius in terms of the indenting force is derived.     

Effects of adding nano-clay (montmorillonite) on performance of polyvinyl acetate (PVAc) and urea-formaldehyde (UF) adhesives in Carapa guianensis,a tropical species

The aim of this study was to improve the bond strength resistance of polyvinyl acetate (PVAc) and urea-formaldehyde (UF) adhesives modified with nano-clay (montmorillonite) with a tropical species of wood known to exhibit adhesion related problems. These adhesives were evaluated with 1.0 and 1.5 wt% nano-clay concentrations with lap shear strength (SS), and the percentage of wood failure (PWF) in dry and wet conditions being evaluated. An additional aim of this study was to observe the presence of nano-clay within both adhesive types using Atomic Force Microscopy (AFM) and the Transmission Electron Microscopy (TEM). Color, viscosity and the thermostability of these adhesives with nano-clay were also evaluated. First, AFM and TEM studies showed adequate dispersion and impregnation of nano-clay. The viscosity of PVAc adhesive was not affected by the incorporation of nano-clay, whereas the UF adhesive was. With both PVAc and UF adhesives, the presence of nano-clay increased the L^⁎ and b^⁎ color parameters, especially when 1.5 wt% nano-clay was used. The incorporation of the nano-clay improved thermostability, as determined by thermogravimetric analysis (TGA). Finally, it was shown that the nano-clay incorporation improved SS and PWF. The highest values of SS were obtained when nano-clay was added at 1.5 wt% concentration in the PVAc adhesive under dry conditions. SS was not affected by nano-clay addition in the UF adhesive under dry conditions. However, under wet conditions, both 1.0 and 1.5 wt% loadings of nano-clay increased SS with both adhesive types. The addition of nano-clay in both proportions increased PWF by approximately 15% and between 20–30% in dry and wet conditions, respectively, for the PVAc adhesive. For the UF adhesive, PWF increased by approximately 10% under dry conditions and 25–50% in wet conditions.     

Development of a simplified etch-and-rinse adhesive containing niobiophosphate bioactive glass

PurposeThe aim of this study was to evaluate radiopacity, degree of conversion (DC), Knoop hardness (KHN), ultimate tensile strength (UTS) and microtensile bond strength (µTBS) to dentin of an experimental adhesive containing micro-filler of niobium–phosphate bioactive glass (NPG).Materials and methodsThe NPG glass was produced by fusion of NbO₅, Na₂CO₃, CaO, (NH₄)₂HPO₄ at 1400 °C. After cooling, the glass was ground to a mean particle size<25 µm, and either added (40 wt%) to an experimental adhesive resin mix containing monomers and solvent, or not. The DC of the adhesives was evaluated by Fourier transform infrared spectroscopy. Flat dentin surfaces were obtained from 16 molar teeth, and prepared for use to evaluate µTBS (n=8). An hourglass-shaped matrix (UTS and KHN) or disk-shaped matrix (radiopacity) was filled with adhesive and light-polymerized. The data from each test were analyzed by appropriate statistical methods.ResultsThe presence of glass particles made the adhesive system radiopaque. Addition of bioactive NPG glass particles to the adhesive system prevented decreases in bond strength; reduced the UTS and increased DC and KHN. All groups showed predominance of adhesive failure mode.ConclusionAddition of 40% NPG glass may be an alternative to obtain an adhesive system with adequate mechanical and bond strength to dentin properties.     

The role of patch hybridization on tensile response of cracked panel repaired with hybrid composite patch: experimental and numerical investigation

ABSTRACT

This article presents a distinct perspective of structural repair by bonding the hybrid composite patch. A novel hybrid composite patch is prepared from the carbon and glass fibers to repair the cracked panel. Different volume fractions of constituents are maintained to prepare the composite patch with varying stiffness. The elastic constant of the composite patch is derived by applying the rule of hybrid mixture and modified Halpin Tsai equation. The stress intensity factor in the panel and interfacial stresses in the adhesive layer are evaluated to assess repair efficiency and repair durability. Effects of the elastic modulus of the adhesive on the performance of composite patch repair are demonstrated. The load carrying capacity and failure strength are examined for variation in patch stiffness. The disbonded surface morphology is investigated through scanning electron microscopy after failure. The results reveal that the hybrid composite patch provided sufficient reinforcement to reduce the stress intensity and interfacial stresses. Patch hybridization has offered a pragmatic solution and proposed as an alternative patch material to repair the cracked structure.     

The effect of different bonding protocols on the repair microshear bond strength of water-stored CAD/CAM resin composite

This study evaluated the repair microshear bond strength (μSBS) of water stored CAD/CAM resin composite under eight different surface treatments using a silane-containing universal adhesive in etch-and-rinse and self-etching modes. In total, 48 CAD/CAM resin composite slices were prepared from Lava Ultimate CAD/CAM blocks and stored in water for 6 months. The slices were assigned into 8 main groups, according to surface treatments (no treatment, no-treatment/silane, surface grinding, surface grinding/silane, sandblasting, sandblasting/silane, silica coating and silica coating/silane). Each main group was divided according to the universal adhesive application mode (either the etch-and-rinse mode or the self-etch mode). Each slice received 6 resin composite micro-cylinders (0.8 × 1 mm). Micro-shear bond strength was run at 0.5 mm/min crosshead speed until failure. Treated surfaces were examined using SEM. Bond strength data were statistically analyzed using Two-Way ANOVA/Tukey HSD post hoc test. Only ‘surface treatment’ significantly affected the repair μSBS (p ? 0.001). Parameters ‘Adhesive application mode’ and ‘surface treatment × adhesive mode’ showed no significant effect on μSBS (p = 0.458 and p = 0.286 respectively). Regardless of the adhesive application mode, silica coating showed the highest μSBS (21.6 ± 6.8 MPa), while sandblasting/silane showed the lowest μSBS (13.0 ± 6.1 MPa). Regardless of adhesive application mode, the use of silica coating to treat the water-stored CAD/CAM resin composite surfaces is crucial to improve the repair bond strength.     

Strap repairs using embedded patches: numerical analysis and experimental results

Adhesively bonded repairs offer an attractive option for repair of aluminium structures, compared to more traditional methods such as fastening or welding. The single-strap (SS) and double-strap (DS) repairs are very straightforward to execute but stresses in the adhesive layer peak at the overlap ends. The DS repair requires both sides of the damaged structures to be reachable for repair, which is often not possible. In strap repairs, with the patches bonded at the outer surfaces, some limitations emerge such as the weight, aerodynamics and aesthetics. To minimize these effects, SS and DS repairs with embedded patches were evaluated in this work, such that the patches are flush with the adherends. For this purpose, in this work standard SS and DS repairs, and also with the patches embedded in the adherends, were tested under tension to allow the optimization of some repair variables such as the overlap length (L_O) and type of adhesive, thus allowing the maximization of the repair strength. The effect of embedding the patch/patches on the fracture modes and failure loads was compared with finite elements (FE) analysis. The FE analysis was performed in ABAQUS® and cohesive zone modelling was used for the simulation of damage onset and growth in the adhesive layer. The comparison with the test data revealed an accurate prediction for all kinds of joints and provided some principles regarding this technique.     

Effect of surface pre-treatment on surface characteristics and adhesive bond strength of aluminium alloy

In this work aluminium alloy surfaces have been subjected to three different methods of surface pre-treatments such as solvent degreasing, FPL (Forest Products Laboratory) etching and priming using an epoxy based primer. The treated surfaces were evaluated for surface energy, contact angle, surface topography, surface roughness and adhesive strength characteristics. The influence of surface pre-treatments on the variation of polar, dispersive and total surface energy of the surfaces is addressed. A wettability test was performed on the surfaces using an epoxy adhesive in order to assess the influence of the pre-treatment techniques on substrate/adhesive interaction. Theoretical work of adhesion values for the various pre-treated surfaces were calculated using the contact angle data and further tested experimentally by adhesive bond strength evaluation by tensile testing of a single lap aluminium-epoxy-aluminium assembly. The method of surface pre-treatment showed a profound effect on the surface topography and roughness by AFM. This study reveals that a combination of high surface energy and high surface roughness of the substrate along with good wettability of the adhesive contributed to the highest joint strength for the aluminium alloy through the FPL etching pre-treatment.     

Strength model of adhesive bonded composite pipe joints under tension

A theoretical model is developed to predict the strain of the pipe, coupling, and adhesive under tensile loading of an adhesive bonded joint. The model is found to be within 10 percent of the experimental pipe and coupling strain. Based on the model, several failure modes and their locations are defined and related to the measured data. In this investigation, delamination is the dominating mode of failure. The delamination stress for each test sample is within 7 percent of the average theoretical delamination stress. In addition, the effect of the coupling length, coupling Young's modulus, adhesive shear modulus, and adhesive thickness on the delamination failure are investigated. The model shows that decreasing the modulus of the coupling improves the delamination failure load; however, the coupling strain at the middle of the joint is increased by this variation. Increasing the shear modulus of the adhesive provides the most significant improvement of the joint delamination failure load. Two geometric factors, the joint length and the adhesive thickness also affect the joint failure load. The joint delamination failure load can only be significantly improved by increasing the bonding length up to a certain limit. Increasing the adhesive thickness increases the delamination failure load, however, a large gap between the pipe and coupling may contribute to misalignment during installation which may result in imposed moments under tensile loading. This study can supply the manufacturers with the appropriate design parameters to improve the joint performance significantly under tensile loading.     

Adhesion properties and fracture mechanism of plasma sprayed NiCrAl/Al2O3–13wt.%TiO2 coatings by post annealing

Plasma-sprayed NiCrAl/Al₂O₃–13wt.%TiO₂ coatings (AT13) deposited on mild steel substrate were annealed with varying temperatures in air. The adhesion of the coating was evaluated by tensile adhesive strength test. The microstructure and the fracture mechanism were studied using optical microscopy, X-ray diffraction, and scanning electron spectroscopy/energy dispersive spectroscopy. It was found that the tensile bond strength of the coatings increased with increasing of annealing temperature at first and then decreased with increasing of annealing temperature further. The as-sprayed coating fractured at the interfaces of substrate/bond layer and bond layer/ceramic coating with a brittle–ductile mixed fracture. The measured strength expressed the adhesive strength and internal adhesive strength of the coating. The failure of the coating annealed at 300, 400, and 500 °C took place at the interface of substrate/bond layer and had a mixed fracture surface of transgranular cleavage fracture and localized ductile fracture. The strength obtained is the adhesive strength between the coating and the steel substrate. The coating annealed at 400 °C had a maximum strength of 42.9 MPa. When the temperature is above 600 °C, the bonding strength would be damaged. Therefore, there is a proper annealing temperature which can significantly improve the bond strength of the coating.     

Microstructural analysis of the adhesion mechanism between old concrete substrate and UHPFC

The performance of any repaired concrete structure, and thus its service life, depends on the quality of the interfacial transition zone of the composite system formed by the repair material and the existing concrete substrate. In this work, the properties of the interfacial transition zone between normal concrete (NC) substrate as an old concrete and ultra-high performance fiber-reinforced concrete (UHPFC) as a repair material was investigated. Pull-off and splitting cylinder tensile tests were performed to quantify the bond strength in direct and indirect tensions, respectively. The microstructure of the interfacial transition zone was also studied using scanning electron microscopy and energy dispersive X-ray spectroscopy (SEM/EDS). Different types of NC substrate surface preparation methods were used. An optical three-dimensional surface metrology device was used to estimate the substrate roughness parameters. Based on the results, high interfacial bond strength was achieved on the 3rd, 7th, and 28th days. The pull-off test results revealed that all failures occurred in the substrate, regardless of the substrate surface roughness. The majority of failures in the split tensile test also occurred in the substrate. SEM/EDS proved that the use of UHPFC as a repair material chemically, physically, and mechanically improved the repaired interfacial transition zone to become stronger and denser, as well as more uniform, and durable. Moreover, the use of UHPFC increased the service life of repaired structures and minimized the number and extent of interventions to the lowest possible level.     

Compatibility of silorane composite repaired with methacrylate-based restorative systems

The aim of this study was to evaluate the repair bond strength of silorane composites using either the silorane or methacrylate-based restorative systems. Expired silorane composite was used as the substrate material in all experimental groups. Silorane blocks (5?×?5?×?4?mm) were fabricated and stored at 37?°C for 24?h. Six experimental groups were developed according to the repair: I-silorane composite (no intermediary); II-P90 Bond/Silorane; III-P90 Adhesive System (primer/bond)/silorane; IV-P90 bond/Scotchbond Universal/methacrylate composite (Filtek P60); V-Scotchbond Universal/methacrylate; and VI-silane/Adper Single Bond 2/methacrylate. The repaired blocks were stored for 24?h at 37?°C, and then sectioned, yielding stick-shaped specimens (0.5?mm²) that were tested in tensile (0.5?mm/min). The results were analyzed using ANOVA/Tukey test (α?=?0.05). The interfacial micromorphology and nanoleakage were also analyzed under SEM. Scotchbond Universal/methacrylate composite, either associated with the P90 bond or not, exhibited similar bond strength to that of P90 Adhesive System/silorane composite. Scotchbond Universal either associated with the P90 Bond or not to repair the silorane allowed no pre-testing failures. The worst scenarios were repairing the silorane with no intermediary (G-I) or combination silane/Adper Single Bond 2/methacrylate composite (G-VI) that presented significantly lower bond strengths and higher incidences of pre-failure testing. The importance of the silane was not confirmed. Characteristic micromorphology and no signs of nanoleakage were identified in all experimental groups. The silane-containing, phosphorylated methacrylate-based adhesive associated with a methacrylate composite was proven to reliably repair the expired composite in a similar way to that of the application of dedicated silorane adhesive.     

Efficiency of different repair kits on bonding to aged dental resin composite substrates

ObjectiveTo assess the efficiency of intraoral repair kits on the tensile bond strength (TBS) of resin composites (RCs) to aged RC substrates.Methods840 aged (six months, 37 °C, distilled water) RC substrates (Tetric EvoCeram) were air-abraded (CoJet) with and without following phosphoric acid contamination or treated with silicon carbide (SiC) grinding paper. Seven repair kits were used as intermediate agents (Embrace First-Coat, CLEARFIL CERAMIC PRIMER, Tokuso Ceramic Primer, Monobond Plus+Heliobond; Scotchbond Universal, One Coat Bond and visio.link) for conditioning. Specimens were repaired using two direct RCs (Clearfil Majesty ES2 and Clearfil Majesty Posterior), stored in distilled water (37 °C, 24 h) and thermal aged (5 °C/55 °C, 10,000 cycles). The cohesive strength of the repair RCs (N=40) served as control and was determined by applying the RCs on the fresh polymerized substrates, followed by thermal-aging procedure. TBS and failure types were determined and evaluated with three-/one-way ANOVA, and chi-square test (p<0.05).ResultsThe highest influence on the TBS was exerted by the intermediate agent (repair kit) (partial eta squared η_P²=0.320, p<0.001), while the impacts of the repair RC (η_P²=0.017, p<0.001) and surface pre-treatment (η_P²=0.015, p=0.003) were significant but low. Except for Embrace First Coat and Tokuso Ceramic Primer, phosphoric acid contamination after air-abrasion maintains the TBS.ConclusionsAir-abrasion induced superior TBS compared with grinding the surface with SiC paper prior to repair. Tested universal adhesives as well as the combination between a universal primer and an adhesive were in-vitro efficient intermediate agents for repairing aged RCs.     

Bio-adhesion evaluation of a chitosan-based bone bio-adhesive

Conventional treatment of complex fractures includes the use of plates and nails, which may compromise the affected limb's functionality. Previous studies have demonstrated promising results through chemical, mechanical, and cytotoxicity tests of a chitosan-based adhesive—proposed as a new method to bond high energy fractures—in dry environments with adequate adhesion, malleability, and biocompatibility. In this study, we focused on performing an evaluation of bio-adhesives’ mechanical properties and bone-adhesive joint using two chitosan-based formulations (with and without a cross-linking agent). The texture profile analysis determined adhesive properties, such as cohesiveness, adhesiveness, hardness, and resilience at different cure times. Bone-adhesive joint was evaluated according to the tensile bond strength test and shear bond strength test. Fracture toughness and cohesive strength were calculated through a rigid double cantilever beam test at mode I failure. Bone-adhesive joints were tested in two environments: dry and submersed in water at 37 °C for 1, 6, and 24 h (curing time), an approximation of surgery conditions. The experimental results showed an incremental of adhesiveness and hardness of the cross-linked adhesive during the first 15 min, which was determined as the usage time to spread on the bone fracture. The joint interaction between the adhesive and bone surfaces was studied; chitosan-based formulations showed an adhesive joint failure under dry conditions in most of the cases. However, this behavior changed under aqueous conditions, presenting cohesive failures. Under aqueous conditions, cross-linked bone-adhesive presented an augmented tensile bond strength up to 0.024 ± 0.0036 MPa, a shear bond strength up to 0.031 ± 0.0069 MPa, and fracture toughness of 2.38 ± 0.54 J/m² was observed with a cure time of 24 h. Finally, the presence of the cross-linking agent in the cross-linked bio-adhesive reduced the sensitivity of the adhesive to water; a promising finding that should be explored in future studies.     

The effect of mechanical roughening and chemical treatment on shear bond strength of urethane dimethacrylate denture base resin

Relining of ill-fitting denture is often required to establish the fit of denture base, hence better retention and stability of the prostheses. However clinical success depends on the ability of reline resin to bond with denture base. The effect of surface preparations of urethane dimethacrylate (UDMA) denture base resin (Eclipse) on the shear bond strength (SBS) to auto-polymerizing polyethyl methacrylate reline material was evaluated. Eclipse specimens were mechanically prepared using two different tungsten carbide burs and submitted to chemical treatments either with dichloromethane (Secure adhesive) or methyl acetate (Eclipse Bonding Agent). Reline resin was then applied to the prepared surface and shear bond strength was tested after 24 h. Data was analyzed using two-way ANOVA and post-hoc Tukey HSD test at p=0.05. The morphological changes of Eclipse surfaces after preparations were also observed under SEM. The results showed that SBS was significantly affected by mechanical roughening, chemical treatment and their interactions. Higher reline SBS values were observed for Eclipse specimens without mechanical roughening compared to those with roughening. Both chemical agents improved reline SBS with the highest bond strength shown when chemically treated using Secure adhesive. For mechanically roughened specimens, Eclipse Bonding Agent (BA) resulted in significantly higher reline bond strength than Secure adhesive. SEM showed different surface appearance of Eclipse resin with various mechanical and chemical preparations.     

Thermoplastic bonding to metals via injection molding for macro-composite manufacture

In this work, we explore a new method of in-situ joining of polymers to metals in injection molding to allow direct bonding between thermoplastic and metal parts. Such a method can integrate several downstream steps in product manufacture, allow optimal design of products and joints, and avoid adhesive application, assembly, and associated difficlties. A variety of process parameters and their effects upon the interface tensile strengths were examined. A full factorial experiment was conducted involving four of the critical process parameters identified. The effects upon tensile strength at break of the following process parameters were studied: (1) adherend surface temperature, (2) screw linear velocity, (3) bondline thickness, and (4) pack and hold pressure. The fracture surfaces and the thermoplastic metal interfaces were analyzed. The bonds fabricated with higher adherend surface temperatures have increased mean tensile strength and less adhesive failure. This increase in mean bond tensile strength and less adhesive failure was due to increased polymer penetration of the adherend surface roughness, at the micrometer level, as shown in the analysis of the polymer-metal interface by a scanning electron microscope (SEM).     

Adhesion of substrate–adherent combinations for early composite repairs: Effect of intermediate adhesive resin application

This study evaluated the effect of intermediate adhesive resin application (IAR) on tensile bond strength (TBS) for early composite repairs in situations where substrate and repair composite bonded together were once of the same kind with the substrate (similar) and once other than the substrate material (dissimilar). Specimens from three types of composites (TPH Spectrum (TPH), Charisma (CHA) and Filtek Z250 (Z250)) were fabricated. The specimens in each composite group (n=72) were randomly divided into six subgroups (n=12). In each composite group, the similar and two dissimilar composites were bonded onto the substrates once using an IAR (Adper Single Bond Plus) and once without. After water storage for 1 week at 37 °C, substrate–adherent combinations were submitted to tensile test. Data were analyzed with three-way ANOVA and Tukey's tests (α=0.05). The substrate–adherent combination (p=0.0001), adherent (repair) composite (p=0.0001), and application of IAR (p=0.0001) significantly affected the results. Utilization of IAR improved the repair bond strength for all composite combinations.     

Efficacy of different surface treatments and universal adhesives on the microtensile bond strength of bulk-fill composite repair

Abstract

The purpose of this in vitro study was to evaluate the influence of different surface treatments and aging on the microtensile bond strength (μTBS) of bulk-fill composite resins. Bulk-fill composites (Filtek One; 3M ESPE) randomly received five different surface treatments: (1) no treatment, control, (2) 37% phosphoric acid etching (PA), (3) 9% hydrofluoric acid etching (HF), (4) air-borne particle abrasion with 50-μm alumina particles (Al₂O₃), (5) tribochemical silica coating (CoJet). Following, the specimens were divided into three subgroups according to universal adhesive applied: Clearfil Universal Bond (CU; Kuraray), Prime&Bond Universal (PBU; Dentsply Sirona), or Single Bond Universal (SBU; 3M ESPE). A nanofill composite (Filtek Ultimate; 3M ESPE) was employed as a repair. Bonded specimens were stored in water for 24?h at 37?°C or thermal aged, then subjected to the μTBS test. Additionally, specimens were analyzed with a contact profilometer and were evaluated with scanning electron microscopy. Control and PA treatments were showed the lowest µTBS (p?<?0.05), and there was no significant difference between these two groups (p?>?0.05). Al₂O₃ and CoJet treatments generally exhibited a similar influence on µTBS values. In addition, a correlation was found between surface roughness and bond strength (r?=?0.831). CoJet resulted in significantly higher repair µTBS values when compared to the other surface treatments. In addition, the use of silane-containing universal adhesive was increased the cohesive failure rate and maintained the repair µTBS values after thermocycling.     

Advances in the proof test for certification of bonded repairs – Increasing the Technology Readiness Level

The availability of an efficient, cost-effective repair technology is an important maintenance requirement to restore structural integrity to metallic and composite airframe structures damaged in service. Generally repair involves attachment of a reinforcing structural element or patch to replace the damaged load path. Traditionally, the reinforcements are attached to the structure with rivets or bolts; however, attachment by adhesive bonding offers many structural and cosmetic advantages.However, bonded repairs of primary structure are very difficult to certify this is because available non-destructive procedures, such as ultrasonics or thermography are unable to detect weak adhesive bonds. In view of the limitation of non-destructive inspection an alternative approach is to directly apply stress to the actual repair bond region or to a very close simulation of the region.In this paper, further work is documented on a proof test of bonded repair coupons (BRCs) that are bonded to the parent structure at the same time as bonding of the repair patch. Therefore, the BRCs are close representation of the actual repair bond strength. To assess the bond strength, immediately after patch application and also possibly through the life of the repair, the BRCs are subject to a previously determined proof load in torsion.The aim of the study is to improve the Technical Readiness Level of the test when applied to various parent-structure/patch-repair systems, including carbon-epoxy/carbon-epoxy; aluminium/boron-epoxy and aluminium/aluminium. Improved BRC application methods were developed to increase the reliability and consistency of the results, and sensitivity to cure condition, surface treatment, contamination, and fatigue damage were evaluated.A detailed finite element (FE) study was undertaken to: a) simulate stresses in the BRC, adhesive and parent structure during the proof test, b) compare the stresses in the patch and BRC when the parent material is under stress and c) investigate the influence of BRC proximity to the patch tip when the parent material is under stress.A conclusion from the FE analysis and fatigue study was that a BRC with the appropriate ply configuration could represent the bondline stresses experienced at the patch tip, and hence could also be used to monitor fatigue damage.     

Strength of adhesively bonded joints under mixed axial and shear loading

This study aims at optimising adhesive properties in an aluminium/structural epoxy assembly for different conditions of surface pre-treatment. We consider the mechanical behaviour and failure under proportional, multi-axial loading using an instrumented, Arcan-type test. Values of fracture strength were found to be dispersed (even for a given surface treatment). Typically dispersion was of the order of 15%. This statistical behaviour, also observed with a simple tensile test, seems to be related to the heterogeneous nature of the microstructure of the adhesive bond, which contains voids, as well as mineral particles for reinforcement. A statistical analysis is suggested for use in conjunction with a strength envelope in practical design, for cases when the stress distribution is significantly heterogeneous. It is believed that this approach may be developed in order to understand the well-known scatter of adhesion strength results, and thus contribute to better reliability assessment.