Protocols for the Measurement of Adhesive Fracture Toughness by Peel Tests

This article reports on the work of the European Structural Integrity Society Technical Committee 4 (ESIS TC4) and its activities in the development of test protocols for peel fracture. Thirteen laboratories have been working on peel test methods in ESIS TC4 since 1997 and their activities are ongoing.

The aim of the work is to develop robust and credible test methods for the determination of adhesive fracture toughness by peel tests. Several geometric configurations have been used, namely, multi-angle fixed arm peel, T-peel, and roller assisted peel in the form of a mandrel test.

The starting point of their work is an established analysis of a peel method that is often developed from a global energy approach. The adopted analysis is combined with an experimental approach in order to resolve ambiguities in the determination of adhesive fracture toughness (G_A). The test methods involve the measurement of peel strength in order to calculate the total input energy for peel (G) and the calculation of the plastic bending energy (G_P) during peel. The latter is often obtained from a measurement of the tensile behaviour of the peel arm. Adhesive fracture toughness is then G - G_P.

Four ESIS TC4 projects are described. The first relates to fixed arm peel whilst the second and third involve both fixed arm and T-peel. The fourth project combines mandrel peel and fixed arm peel. Each project uses different types of polymeric adhesives in the form of quite different laminate systems. The selection of the laminate system enables all characteristics of laminate property to be embraced, for example, thin and thick adhesive layers, polymeric, and metallic peel arms and a range of flexibility in the laminates.

The development of the enabling science required to establish the test protocols is described and software for conducting all calculations is referenced.     

THE DEVELOPMENT OF A MANDREL PEEL TEST FOR THE MEASUREMENT OF ADHESIVE FRACTURE TOUGHNESS OF EPOXY–METAL LAMINATES

A mandrel peel test is established for measuring the adhesive fracture toughness of a metal/rubber-toughened epoxy laminate system. By adopting an energy balance analysis it is possible to determine directly both adhesive fracture toughness and plastic work in bending the peel arm around the mandrel. The suitability of the procedure is examined for various types of metal peel arms, which are classified in terms of their ability to deform plastically during the test. The plastic work is also predicted theoretically, and comparisons are made between the measured and calculated values. The fracture energies determined from the mandrel tests are compared with those obtained from 90° fixed-arm peel tests. For the calculations of plastic work in bending in the fixed arm test, various options are used when modelling the tensile stress-strain behaviour of the peel arm material. In addition, the adhesive layer thickness is considered in terms of its influence on the calculation of adhesive fracture toughness.     

THE DEVELOPMENT OF A MANDREL PEEL TEST FOR THE MEASUREMENT OF ADHESIVE FRACTURE TOUGHNESS OF EPOXY-METAL LAMINATES

A mandrel peel test is established for measuring the adhesive fracture toughness of a metal/rubber-toughened epoxy laminate system. By adopting an energy balance analysis it is possible to determine directly both adhesive fracture toughness and plastic work in bending the peel arm around the mandrel. The suitability of the procedure is examined for various types of metal peel arms, which are classified in terms of their ability to deform plastically during the test. The plastic work is also predicted theoretically, and comparisons are made between the measured and calculated values. The fracture energies determined from the mandrel tests are compared with those obtained from 90° fixed-arm peel tests. For the calculations of plastic work in bending in the fixed arm test, various options are used when modelling the tensile stress-strain behaviour of the peel arm material. In addition, the adhesive layer thickness is considered in terms of its influence on the calculation of adhesive fracture toughness.     

Comparison of Peel Tests for Metal–Polymer Laminates for Aerospace Applications

Standard peel tests for aerospace laminates based on metal–polymer systems, namely floating-roller and climbing-drum peel methods, have been accommodated in a unified theory of peeling. This theory also accommodates more basic peel tests such as T-peel and fixed-arm peel and also newer methods such as mandrel peel. These five methods have been applied to two aerospace laminate systems to critically examine their use in the determination of adhesive strength. The theory has been used to unify the outputs from the tests in terms of adhesive fracture toughness. In this way, the comparative merits of the methods can be commented on.

The validity of the standard methods has been put in doubt because of the absence of a correction for plastic bending energy and also because of the poor conformance of the peel arm to the roller system used in these methods. The unified theory and some measurements of peel-arm curvature help but not completely overcome some of these difficulties.

A further complication that arises in peel is a change in the plane of fracture. This reflects a transition from cohesive fracture in the adhesive to an adhesive fracture at the interfaces among adhesive, primer, and substrate. It is likely that such plane-of-fracture phenomena are intrinsic to evaluation of the laminate and that contemplation of cohesive fracture toughness for the adhesive cannot accommodate such events.     

Comparison of Peel Tests for Metal-Polymer Laminates for Aerospace Applications

Standard peel tests for aerospace laminates based on metal-polymer systems, namely floating-roller and climbing-drum peel methods, have been accommodated in a unified theory of peeling. This theory also accommodates more basic peel tests such as T-peel and fixed-arm peel and also newer methods such as mandrel peel. These five methods have been applied to two aerospace laminate systems to critically examine their use in the determination of adhesive strength. The theory has been used to unify the outputs from the tests in terms of adhesive fracture toughness. In this way, the comparative merits of the methods can be commented on.

The validity of the standard methods has been put in doubt because of the absence of a correction for plastic bending energy and also because of the poor conformance of the peel arm to the roller system used in these methods. The unified theory and some measurements of peel-arm curvature help but not completely overcome some of these difficulties.

A further complication that arises in peel is a change in the plane of fracture. This reflects a transition from cohesive fracture in the adhesive to an adhesive fracture at the interfaces among adhesive, primer, and substrate. It is likely that such plane-of-fracture phenomena are intrinsic to evaluation of the laminate and that contemplation of cohesive fracture toughness for the adhesive cannot accommodate such events.     

Mode-I adhesive fracture energy of carbon fibre composite joints with nanoreinforced epoxy adhesives

The effect of the addition of carbon nanoreinforcements to an epoxy adhesive on the strength and toughness of carbon fibre/epoxy composite joints was studied. The laminate surfaces, treated with peel ply, were characterised by profilometry, image analysis and wettability. The mechanical properties of the joints were determined by lap shear testing and double cantilever beam testing. The fracture mechanisms were studied by scanning electron microscopy.The addition of carbon nanofibres and carbon nanotubes caused an increase in the mode-I adhesive fracture energy, G_IC, of the joints while their lap shear strengths remained approximately constant. This improvement in the fracture behaviour was attributed to the occurrence of toughening mechanisms when carbon nanoreinforcements were added to the epoxy adhesive. Additionally, the use of carbon nanotubes improved the interfacial strength between the adhesive and the substrate, changing the crack growth behaviour and the macroscopic failure mode.     

Quantitative Detection of Peel Ply Derived Contaminants via Wettability Measurements

Abstract

The theoretical relationship between surface energies and adhesion are briefly reviewed. Surface energies obtained from contact angles measurements can be extremely useful but difficult to obtain in manufacturing and field environments. A novel technique for probing surface energies via liquid drops, created using ballistic deposition, was investigated. The use of this technique as a quality assurance tool for detecting peel ply derived siloxane contaminants is discussed. It is shown that a readily detectable threshold amount of contamination is required to affect fracture toughness, and that the amount is similar for several distinct adhesive systems.     

Comparison of Analytical,Numerical, and Experimental Methods in Deriving Fracture Toughness Properties of Adhesives Using Bonded Double Lap Joint Specimens

Stress and fracture analysis of bonded double lap joint (DLJ) specimens have been investigated in this paper. Numerical and analytical methods have been used to obtain shear- and peel-stress distributions in the DLJ. The generalized analytical solution for the peel stress was calculated for various forms of the DLJ geometry and, by using crack closure integral (CCI) and by means of the J-integral approach, the analytical strain energy-release rate, G, was calculated. Experimental fracture tests have also been conducted to validate the results. The specimens were made of steel substrates bonded by an adhesive and loaded under tension. Specimens with cracks on both sides and at either end of the DLJ interface were tested to compare the fracture behavior for the two crack positions where tensile and compressive peel stresses exist. Tests confirmed that the substrates essentially behave elastically. Therefore, a linear elastic solution for the bonded region of the DLJ was developed. The fracture energy parameter, G, calculated from the elastic experimental compliance for different crack lengths, was compared with numerical and analytical calculations using the experimental fracture loads. The stresses from analytical analysis were also compared with those from the finite element results. The strain energy-release rate for fracture, G _f , for the adhesive has been shown to have no R-curve resistance, was relatively independent of crack length, and compared well with those obtained from numerical and analytical solutions. However, it was found that fracture energy for the crack starter in the position where the peel stress was tensile was about 20% lower than where the crack was positioned at the side, where the peel stress was found to be compressive.     

Substrate constraint and adhesive thickness effects on fracture toughness of adhesive joints

The adhesive thickness effect on fracture behaviour of adhesive joints has been studied using the boundary effect model recently developed for specimen size effect on fracture properties of concrete, and the essential work of fracture model for ligament (uncracked region) effect on largescale yield of bulk metals and polymers. The leading common mechanism responsible for the nonlinear elastic fracture mechanics behaviours, such as adhesive thickness effect of adhesive joints, specimen size effect of brittle heterogeneous materials and notch dependence of deeply notched metal and polymer specimens, is discussed. These two fracture mechanics models show that the height variation of a fracture process zone (FPZ) or a plastic zone is directly responsible for any change in fracture energy measurements such as the specific fracture energy G _f and the critical strain energy release rate G _c. Both models show that G _f is rapidly reduced when the crack-tip approaches the back-face boundary of a specimen because only a limited FPZ or plastic zone height h _FPZ can be developed in the boundary region. In the case of a thin adhesive joint, the development of a plastic zone height is limited by the thickness of the adhesive sandwiched between the upper and lower adherends or substrates. Consequently, a linear relationship between the adhesive joint toughness and adhesive thickness is established. Test results on adhesive joints from the literature are analysed and compared with the new adhesive joint failure model based on the two well-established fracture mechanics models developed for other material systems.     

The Mechanical Behaviour of Polyimide/Copper Laminates Part 2: Peel Energy Measurements

The mechanical peel behaviour of laminates consisting of polyimide films adhered to copper foil using a modified acrylic adhesive has been studied over a wide range of test rates and temperatures. The laminates were prepared from polyimide films which had been subjected to either a “high-thermal history” or a “low-thermal history” treatment during the production of the film. The measured peel energies of the laminates could be superimposed to give a master curve of peel energy versus the reduced rate of peel test, Ra_T , where R is the rate of peel test and a_T is the time-temperature shift factor. The appropriate shift factors were a function of the test temperature and were mainly deduced from tensile tests conducted on the bulk adhesive. The “high-thermal history” laminates gave higher peel energies and the locus of failure of the laminates was mainly by cohesive fracture through the adhesive layer. At low values of log₁₀ Ra_T , i.e. Low rates of peel and high test temperatures, the “low-thermal history” laminates also failed in the adhesive layer and possessed similar peel energies to those measured for the “high-thermal history” laminates. However, at high log₁₀ Ra_T values, the peel energies measured for the “low-thermal history” laminates were lower and showed a wider scatter. These arose from a different locus of failure occurring in these “low-thermal history” laminates when tested under these conditions. Namely, it was found that most of these laminates failed in a weak boundary layer in the outer regions of the “low-thermal history” polyimide film.     

Delamination failure mechanisms in microlayers of polycarbonate and poly(styrene-co-acrylonitrile)

The peel strength and delamination failure mode of coextruded microlayer sheets consisting of alternating layers of polycarbonate (PC) and poly(styrene-co-acrylonitrile) (SAN) were studied with the T-peel test. Four delamination modes were observed: two modes where the crack propagated along the PC–SAN interface and two other modes where the crack propagated through crazes in the SAN. The SAN layer thickness determined whether crack propagation was interfacial or through crazes. Crazing and crack propagation through crazes were observed only if the SAN layer was thicker than 1.5 μm. As the thickness of the SAN layer increased, the amount of crazing in front of the crack tip and the amount of craze fracture gradually increased; the peel strength increased accordingly. If the SAN layers were thinner than 1.5 μm and the PC layers were relatively thick, the crack propagated along a single interface. The peel strength for this delamination mode was the lowest and equal to about 90 J/m², independent of layer thicknesses. This delamination mode came closest to providing a ”real” measure of the adhesive toughness of PC to SAN. With both interfacial and craze delamination, the crack could move from layer to layer if the PC was thin enough. Tearing of the relatively thin PC layers increased the peel strength of the multiple-layer delamination modes. © 1998 John Wiley & Sons, Inc. J Appl Polym Sci 68:793–805, 1998     

Evolution of crack path and fracture surface with degradation in rubber-toughened epoxy adhesive joints: Application to open-faced specimens

The crack path and fracture surface in the mixed-mode fracture of two different rubber-toughened epoxy adhesives were evaluated using double-layered open-faced double cantilever beam (ODCB) specimens in which the primary adhesive layer had been environmentally aged. The crack path in the mixed-mode fracture of unaged ODCB specimens was unexpectedly in the secondary adhesive layer, and several hypotheses were examined to explain this. It was concluded that a reduced residual stress in the secondary adhesive layer produced stable crack growth in the secondary layer instead of the expected path in the primary layer. The average crack path depth, fracture surface roughness and maximum elevation in the fracture surface profiles were then measured using optical profilometry as a function of the degree of aging. The results showed a strong relationship between all these parameters and the critical strain energy release rate, G_cs, irrespective of the type of adhesive. In the case of adhesive A where significant irreversible degradation was observed, all these parameters varied approximately linearly with G_cs. In the case of adhesive B, aging did not result in permanent degradation (G_cs was unchanged) and so all these fracture surface parameters also remained unchanged after aging. The results indicate that quantifying fracture surface parameters as a post-failure analysis can be of use in the estimation of the fracture toughness at which a practical joint fails.     

Cleavage and Static Toughness

The cleavage of adhesive joints allows the experimental study of the process of fracture in the low speed range. The value of the fracture energy deduced from the fracture length is the static toughness of the adhesive. This value, which determines the endurance limit of the joint, is much larger than can be explained by the current theories. It depends on the surface treatment of the substrate and results from the damage of the adhesive bonds. To take into account these results, the equation describing the fracture of adhesive joints as it was proposed by A. N. Gent and J. Schultz has to be extended. When that is done, it applies to viscoelastic adhesives, whether pressure sensitive or hot melts, and probably also to cross-linked adhesives.

If G is the fracture energy of the joint, the equation G = G ₀ + α K ₂ ·v^a accounts for most experimental results and even for the fatigue of adhesive joints.     

The Effect of WS2 Nanotubes on the Properties of Epoxy-Based Nanocomposites

In this paper we evaluated the effect of embedding inorganic nanotubes (INT) of tungsten disulfide (WS₂) in an epoxy matrix, on the mechanical, thermal and adhesion properties of the resulting nanocomposites. The nanotube content spanned a range of values (0, 0.1, 0.3, 0.5 and 1.0 wt%), and the nanotube incorporation process consisted of a combination of both distributive (magnetic stirring) and dispersive (ultrasonic mixing) methods. The adhesion of the nanocomposites to an aluminum substrate was characterized by both a single lap shear and a T-peel test. The fracture toughness (K _IC) of the nanocomposites was characterized by a standard compact tension (CT) plane-strain fracture test. The thermal properties of the nanocomposites were determined by dynamic mechanical thermal analysis (DMTA). Overall, the addition of INT-WS₂ was found to improve the shear strength and peel properties of the nanocomposite, and to significantly improve its fracture toughness and glass transition temperature. The extent and character of the nanotube–epoxy interaction were examined by electron microscopy, as was the energy dissipation mechanisms during fracture.     

Temperature and Loading Rate Effects on the Fracture Behaviour of Adhesively Bonded GFRP Nylon-6,6

The combined effect of varying test temperature and loading rate on the Mode II fracture toughness of plasma-treated GFRP Nylon-6,6 composites bonded using a silica-reinforced epoxy adhesive has been studied. End notch flexure tests have shown that the adhesive system used in this study offers a wide range of fracture energies that are extremely sensitive to changes in temperature and loading rate. Increasing the test temperature resulted in a substantial reduction in the Mode II fracture toughness of the adhesive, with the value of G_IIc at 60°C being approximately one-half of the room temperature value. In contrast, increasing the crosshead displacement rate at a given temperature has been shown to increase the value of G_IIc by up to 250%. Compression tests performed on bulk adhesive specimens revealed similar trends in the value of [sgrave]_y with temperature and loading rate. In addition, it was found that the plasma treatment employed in this study resulted in stable crack propagation through the adhesive layer under all testing conditions.

A more detailed understanding of the effect of varying temperature and loading rate on the failure mechanisms occurring at the crack tip was achieved using the double end notch flexure (DENF) geometry, which was considered in tandem with the fracture surface morphologies. Here, changes in the degree of matrix shear yielding and particle-matrix debonding were used to explain the trends in [sgrave]_y and G_IIc.     

Fracture toughness of adhesive joints. II. Temperature and rate dependencies of mode I fracture toughness and adhesive tensile strength

It has been known that adhesive strength shows temperature and rate dependencies reflecting visoelastic properties of an adhesive. Similarly, a critical strain energy release rate is expected to show temperature and time dependencies deformation and fracture of the adhesive occurs at the time of measurement of the critical strain energy release rate, which is a kind of fracture mechanical parameter for adhesive joints. The term “critical strain energy release rate” has usually been called “fracture toughness.” In this study, the critical strain energy release rate (G_IC) of the opening mode was called mode I fracture toughness. G_IC was measured over a wide range of temperatures and rates, and then a master curve was obtained by applying the temperature–rate superposition principle to the obtained data. Also, on the relation between G_IC and adhesive tensile strength is discussed. © 1995 John Wiley & Sons, Inc.     

Adhesive Joint Durability Assessed Using Open-faced Peel Specimens

This paper presents an investigation of the durability of two aluminum-epoxy adhesive systems by means of open-faced peel specimens. A peel analysis model was used to determine the fracture energy from the peel data. Both wet and dry peel tests were conducted in order to distinguish between the reversible and the permanent effects of water. The effects of water on the cohesive properties of the adhesives were also assessed by tension tests. It was found that, for the two-part epoxy adhesive, which plasticized to a large extent, the peel testing should be carried out in a dry state to assess the interfacial weakening. It was also observed that the two-part adhesive was much stiffer in the dry, degraded state, and it was important to take account of such permanent changes in the cohesive properties associated with water uptake when determining the fracture energy from the peel data. In contrast, the one-part epoxy system did not suffer from appreciable cohesive changes, either reversible or permanent. In this case, both wet and dry failure loci were interfacial, and some of the interfacial damage was found to be reversible. Finally, surface analyses of the peel failure surfaces were carried out, and the formation of micro-debonds was identified as a possible mechanism of degradation for the two-part system.     

Prediction of mechanical behaviour of adhesively bonded CFRP scarf jointed specimen under tensile loading using localised DIC and CZM

The present study focuses on the mechanical behaviour of both single and double tapered scarf adhesively bonded joint of Carbon fibre reinforced polymer (CFRP) laminate as adherend subjected to tensile loading. The layup sequence of the CFRP adherend having unidirectional (UD) [0⁰]₁₆ and quasi [+45/−45/0/90]_2S are studied. The adhesive used here is Araldite 2015 supplied by Huntsman which is a two part epoxy system of intermediate toughness grade. Here, 2D digital image correlation (DIC) technique is used for capturing the whole field longitudinal, peel and shear strain distribution over the adhesive bond line of the CFRP specimen. Further, a localised DIC measurement is also carried out using microscopic tube lens for precisely capturing strain field over concentrated zones where damage initiation occurs. The evolution of whole field strain distribution with increasing load is captured to predict the mechanical behaviour and failure mechanism of a tapered scarf joint specimen. In addition, 2-D finite element analysis (FEA) of scarf joint model is carried out for validating the DIC results. In the finite element model cohesive zone elements are used for the modelling of both adhesive layer and inter/intra laminar interface of the composite laminate. Initially, to verify the proposed numerical model, joint's initial stiffness, failure load and corresponding displacement obtained from FEA are compared against the experimental load – displacement results. Later, qualitative and quantitative comparison of longitudinal, peel and shear strain values obtained over the adhesive layer by DIC and FEA is carried out to confirm the accuracy of the DIC results. A decent correlation is found to exist between the DIC predictions and numerical results thereby confirming the accuracy of the DIC technique. Analytical solutions are also derived for the same problem based on mechanics of material and further it is compared with both FEA and DIC predictions for completeness.     

Kevlar fiber–epoxy adhesion and its effect on composite mechanical and fracture properties by plasma and chemical treatment

Kevlar 49 fibers were surface-modified by NH₃-, O₂-, and H₂O-plasma etching and chlo-rosulfonation and subsequent reaction with some reagents (glycine, deionized water, eth-ylendeiamine, and 1-butanol) to improve the adhesion to epoxy resin. After these treatments, the changes in fiber topography, chemical compositions of the fiber surfaces, and the surface functional groups introduced to the surface of fibers were identified by SEM, XPS, and static SIMS. Interlaminar shear strength (ILSS) and T-peel strenght between the fiber and opoxy resin, as measured by the short-beam test and T-peel test, were remarkedly improved by gas plasma and chlorosulfonation (0.1% and 0.25% CISO₃H at 30 s). However, from the results of similar G_IC values of the treated and untreated fiber composites, it is clear that the fiber/matrix interfacial bond strength is only a minor contributor to G_IC. SEM was also used to study the surface topography of the fracture surfaces of composites in T-peel test. It could be seen from SEM observations that the improvement of fiber/matrix interfacial bond strength often accompanied a change in fracture mode from the interface of fiber/epoxy resins to the fiber fibrillation and the resins. © 1996 John Wiley & Sons, Inc.     

Evaluation of Mode-II Fracture Energy of Adhesive Joints with Different Bond Thickness

A study on the mode-II edge-sliding fracture behaviour of aluminium-adhesive joints was carried out. Compact pure shear (CPS) adhesive joints of different bond thickness were produced using a rubber-modified epoxy resin as the adhesive. An analytical model was developed to calculate the stress distribution along the bond line of the joint. A crack-closure technique was used to evaluate the mode-II strain energy release rate. G _II, as a function of the adhesive bond thickness. The results indicated that for a given applied load, G _II increased gradually with the bond thickness. A finite element model (FEM) was also developed to evaluate the stress state along the bond line and the strain energy release rate of the CPS specimens. Consistent results were obtained between the theoretical model and finite element analysis. Scanning electron micrographs of the fracture surface illustrated a mainly interfacial fracture path between the adherends and the adhesive for all adhesive joint specimens. The critical fracture load increased very rapidly with bond thickness in the range 0.02 mm to 0.1 mm but remained constant thereafter. However, the mode-II critical fracture energy rose more gradually as the bond thickness was increased.