Sage Journals: Discover world-class research

Abstract

Thermoplastic tapes have found a prominent place in automated tape placement (ATP), due to their reduced processing time. ATP also offers significant reduction in labour; however, the most attractive aspect is the use of its welding properties. Welding or diffusion bonding is necessary for two thermoplastic materials to bond to each other through the combined effect of heating and consolidation pressure. The work published in this article shows how various thermoplastic tape materials with different material properties are bonded to each other using a direct flame-type ATP process. Contact angle and differential scanning calorimetry measurements help understanding of the processing needs of the considered materials. The samples obtained after ATP are sent for peel testing using a wedge peel test principle, so that the force required to separate the bonding is identified. A T-peel test/pull test is also employed to cross-compare peel results obtained through wedge peel testing. The main aim of the work is to study the quality of connection between the two plies with different material interfaces and also how friction might contribute to peel force when wedge peeling is used. A numerical model is also implemented to show the effects of this friction.

Keywords

Polymer composites adhesion ATPs peel force interface strength friction co-efficient

Introduction

Composites have been important materials due to their high-end properties and their versatility in a wide range of applications. Composite materials these days constitute a significant proportion of the engineered materials market ranging from everyday products to sophisticated niche applications. The ever-growing research sector in composite materials has pushed for newer polymer resin matrix materials and high-performance reinforcement fibres of glass, carbon and aramid. The penetration of these advanced materials has witnessed a steady expansion in uses and volume. The need for bigger volume has also resulted in cost reductions. High-performance fibre-reinforced polymers can now be found in diverse applications, such as composite armouring designed to resist explosive impacts, cylindrical vessels, offshore/onshore windmill blades and industrial drive shafts, and these materials also play a significant role in automotive or aerospace sectors. The benefits of corrosion resistance and light weight have proven attractive in many low-stress applications. There are also applications in which composites are combined with metal in order to save both costs and weight. One such type of composite material is thermoplastic composites, and they are highly attractive to the aerospace industry due to their combination of high fracture toughness, high damage tolerance, long (indefinite) out life and good recyclability compared to thermoset systems.¹ The processing difficulties associated with them do pose a problem in composite industries. Driven by the need for more cost-effective applications, researchers and design engineers of polymer composites are facing increasing demands for the development of faster and more environment-friendly processing methods. Rapid processing techniques require a better understanding and a more efficient control of phenomena such as heating, impregnation, bonding and solidification. Most of the new techniques promote non-isothermal and transient phenomena in order to achieve faster transformation of the materials. This has led to increasing research in the area of in situ consolidation using automated tape placement (ATP). The automation of the manufacturing process for thermoplastics by ATP can achieve an increased production rate, reduction in labour cost and improved geometric repeatability compared to conventional hand lay-up.² Thermoplastic tape placement also offers the possibility to save an additional consolidation treatment.³ Furthermore, the thermoplastic tape material is much easier to handle and has no limited stock life. The high-quality requirement of the components requires the availability of adequate production processes and a deep understanding for the parameters affecting the process.^4
-6

Using an ATP process, it is possible to add several layers with reduced time and minimum supervision to create complex geometry. The mechanism to join the layers used here is the welding process. Welding of two thermoplastic tapes requires specific physical conditions: an intimate contact between the surfaces at sufficient pressure level, temperature at the nip point high enough to allow diffusion of macromolecules and a time span long enough in order to secure welding quality. Due to the low thermal conductivity of thermoplastic composites, a high temperature at the interface can be reached by local heating. Mantell et al.⁷ introduced a simplification for the welding (bonding) of polymer surfaces that are well accepted for simulations of the tape placement process. Previous research work^8
-10 has been carried out in the field of intimate contact and polymer healing. Intimate contact and molecular diffusion occur at several interfaces, such as between plies of a laminate or between different elements assembled to form the final part. However, the high content of carbon fibre (CF) reinforcement in thermoplastics, resulting in high thermal and electrical conductivity, imposes difficulties such as uneven heating, delamination^11,12 and distortion of the laminates. These problems have posed a big threat when bonding large components.^13,14 In addition, as fibre volume fraction increases, the amount of resin available to melt and reconsolidate into a fused joint is reduced and this can also affect the welding quality. One can say that in situ consolidation might not offer the same benchmark as an autoclave but the research carried out recently reveals that there is huge potential for processed thermoplastic materials using the ATP process. Ageorges et al.¹⁵ explain that the fusion bonding processes offer advantages such as reduced surface preparation requirements, reprocessing, recyclability and improved integrity/durability. Schell et al.¹⁶ confirmed through mathematical modelling and experimental validations the effects of temperature at the interface between two composite parts and hence this can cause randomness in bonding quality. Bourban et al.¹⁷ also illustrate how the material microstructure and transformation phenomena during diffusion bonding and crystallisation control the rapid processing of thermoplastic composites.

Heat input required for thermoplastic welding in an ATP process is carried out either using a high-powered laser,^18

-21 direct flame,²² infrared (IR) heating,²³ ultrasonic heating^24
-26 or a combination of these methods.²⁷ Pressure using rollers then consolidates the heat-affected region to acquire the necessary bonding between the parts. Laser heat source is seen as the most accurate and high-end heat source for ATP process in industries. Rosselli et al.²⁸ showed a range of specimens tested in terms of power variation and the trends showed an increase and a decrease of shear strength typically at about a peak laser power of 70 W. They also confirm that slight misalignment of laser beam can result in shear strength reduction of up to 14%. The study by Ferraro et al.²⁹ overcomes the overheating of the tape using laser heat source by taking the local tape temperature or the actual winding speed as the variable controlling the laser output. Recent investigations of Stokes Griffin et al.^30,31 deal with a model-based approach for near-IR laser heating and selective reinforcement of steel with CF/polyamide 6 (PA6) tapes. Funck et al.³² compare IR, direct flame, hot gas and laser with respect to their processing costs and parameters. The studies found IR to be inexpensive compared to laser but the energy density and response time are not nearly as high as with laser radiation. However, to keep purchasing costs low and to have faster incoming tape speed, direct flame would certainly fit the bill. Hot gas torch (HGT) on the other hand keeps the purchase cost reasonable, but the energy efficiency is low. Nonetheless, recent investigations of Qureshi et al.³³ showed the potential and possibilities with optimised HGT nozzle geometry and its influence on mechanical behaviour in interlaminar shear strength. The study of Funck and Neitzel³² helped to establish that ideal heating source with best contact is required to achieve laying conditions suitable for optimal diffusion bonding. Calawa and Nancarrow’s³⁴ investigations have led them to design a quartz lamp IR heater suitable for heat application right before the application of tow. The main advantages of the quartz IR emitter are short response time, durability in a manufacturing environment and longest wavelength with high power output, and this allows for highly controlled localised heating of the area in a small package. It has also been found by Karpachevskyy³⁵ that increasing the number of individual light sources and optimizing IR geometry allow for both higher efficiency and uniformity. The heater geometries generated by Karpachevskyy’s method achieve excellent efficiencies, with up to 60% of emitted light being absorbed by the tape directly without intermediate reflection and irradiance functions that are visually indistinguishable from being constant. Schledjewski³ conducted a cost-efficiency analysis (CEA) and it was concluded that the tape placement head’s efficiency depends mainly on the type of heat source chosen. This would also mean that the final properties and the quality of the tape would be strongly affected by the heat source selected. With this scoring system, Schledjewski evaluates the different heat sources and confirms that laser, direct flame and HGT systems result in the highest CEA score. Recent developments try to combine the pros of laser heating, namely, pulsed high energy emission (Xenon) for precise controllability without the cons, namely, security issues for human life.^36,37

Typically, bond strength of welded interfaces is characterised using a standard test, like the lap shear test (e.g. according to ASTM D1002³⁸) or double cantilever beam (DCB) test (e.g. according to ISO 15024³⁹). Since those tests require bulk specimens, a characterisation of the bond strength of two thin tapes continuously welded is not possible using such a test. Instead, peel testing is used to characterise the bonding.^40,41 Hulcher et al.⁴² used a wedge peel tests to study the cohesive failure of specimens during peel testing under different loading conditions. For the current study, the wedge peel method was selected since it can be used to characterise the welding of two thin tapes.⁴³ The principle of the wedge peel test mechanism is that a bonded tape is passed over a stationary blade and is pulled with the force until the bond starts breaking along a crack line. Consequently, the strength of the bond is measured as a requirement for the application. Kawashita et al.⁴⁴ use different methods of peel testing to critically examine adhesive strength of two different sets of aerospace metal polymer laminates. Grouve et al.⁴⁵ investigate experimentally the validity of the mandrel peel test on thermoplastic laminates by comparing the measured fracture toughness with the values obtained by the DCB and the end-loaded-split beam test. Rao et al.⁴⁶ use a floating roller peel test method to determine the peel force that is required to separate the eight towpregs from the placement machine’s surface. The research of Rao et al.⁴⁶ identified that the lay-up temperature has the greatest influence in determining the resulting peel force. Comer et al.⁴⁷ conducted a series of manufacturing, mechanical and thermal trials to identify an optimum set of processing parameters for the laser-assisted tape placement using CF/polyether-ether-ketone (PEEK). Wedge peel test is employed here to study the quality of interlaminar strength and the study concluded that for any samples produced with unheated non-metallic tool provided the best peel force (4.1 N/mm) and void content (1%); however, low percentage crystallinity (approximately 20%) resulting from rapid cooling rates will likely compromise mechanical strength, stiffness and chemical resistance of CF/PEEK. Although energy intensive, a heated metallic tool (280°C) can serve to increase the percentage crystallinity to a level generally considered optimal (approximately 35%) but this indeed gave a significant drop in peel force values. Recent developments deal with quality assurance systems to detect gaps, overlaps, twists, bridging or foreign bodies. A promising project was presented by Denkena et al.⁴⁸

This introduction gives a brief insight into how previous research has focused on different aspects of ATP processing: from its industrial requirement to utilizing various mathematical models to predict manufactured parts property through microscale modelling. The present study showcases a full process chain starting from material selection until material testing with all its stages in between (Figure 1) and a separate modelling stage. Therefore, tape materials were pre-characterised in terms of their surface energy (Table 3). Parallel, differential scanning calorimetry (DSC) tests were carried out on all tape materials providing transition temperatures (Table 4) to estimate processing temperatures and energies. Subsequently, the selected tape combinations were placed and welded on an ATP machine considering pre-characterisation results (Table 6). Two mechanical test setups were employed to estimate the bonding strength of selected welded tape combinations, namely, wedge peel test and T-peel/pull test. Both the tests have their pros and cons; consequently, they were compared in terms of their results for selected tape combinations. Additionally, a finite element model was employed comparing peel force against energy release rate during peeling. For the wedge peel test, this finite element model also considers the effect of friction on peeling.

Figure 1.

Process chain of the work carried out in this research work.

Experimental work

Materials selection and cutting

Thermoplastic tape materials were purchased from Suprem (Z.I. Le Bey 17, CH-1400 Yverdon-Ies-Bains, Switzerland) and Celanese (Am Unisys-Park 1, 65843 Sulzbach, Deutschland). Tape materials available for testing are listed in Table 1.

Table 1.

List of available tape materials and their thicknesses.

Manufacturer	Fibre material	Matrix material	Thickness (mm)
Suprem	CF	PES	0.17
Suprem	CF	PEEK	0.18
Celanese	CF	PA6	0.13
Celanese	CF	PP	0.16
Celanese	GF	PP	0.29

PEEK: polyether-ether-ketone; PP: polypropylene; PA6: polyamide 6; PES: polyethersulfone; CF: carbon fibre; GF: glass fibre.

Once the materials were chosen, the tapes needed to be confected using a rotating blade cutter. The tape samples were slit to the specific size of 700 × 24 mm². Visual inspection of the cut tapes was carried out to ensure the integrity of the material after cutting. Moreover, each tape was measured at three distinct points over the 700-mm length to ensure a constant width was kept.

Tape characterisation

Tape characterisation was mainly carried out to understand the bonding and polarity against different other tapes. For these purposes, contact angle (CA) measurements were carried out. CA of the samples was measured with a drop shape analyser (DSA 100; Krüss GmbH, Borsteler Chaussee 85, 22453 Hamburg, Deutschland) at room temperature with a droplet volume of 2 µ _l. CA measurements were averaged over five different positions of the surface and using three different liquids. The used liquids for the CA measurements and their respective surface tensions (γ_l ) are given in Table 2.

Table 2.

Test liquids for CA measurements on tape materials.^49,50

Liquids	Surface tension (γ; mN/m)	Polar part (γ^p ; mN/m)	Disperse part (γ^d ; mN/m)
Water	72.8	51.0	21.8
Diiodomethane	50.8	0	50.8
Ethylene glycol	47.7	16.8	30.9

CA: contact angle.

The surface energies of the tape materials were calculated with the Owens, Wendt, Rabel and Kaeble methods.^51

-54 Therefore, CAs of each tape material, with all three measurement liquids, were determined. These angles and the known surface energies of the measurement liquids served as input parameters for equation (1). The results were plotted accordingly and a linear approximation was performed. Polar surface energy parts are calculated from the slope, while the disperse parts are represented through the constant. The calculated surface energies (γ_s ) of the tape materials are given in Table 3.

\frac{(1 + cos (Θ)) \cdot γ_{l}}{2 \cdot \sqrt{γ_{l}^{d}}} = \sqrt{γ_{s}^{p}} \cdot \frac{\sqrt{γ_{l}^{p}}}{\sqrt{γ_{l}^{d}}} + \sqrt{γ_{s}^{d}}

Table 3.

Surface energy of tape materials.

Samples	Surface tension (γ; mN/m)	Polar part (γ^p ; mN/m)	Disperse part (γ^d ; mN/m)
CF-PES	54.9	6.9	47.6
CF-PEEK	44.7	0.01	44.7
CF-PP	29.5	0.3	29.2
CF-PA 6	31.6	2.0	29.6
GF-PP	29.1	3.5	25.6

PEEK: polyether-ether-ketone; PP: polypropylene; PA6: polyamide 6; PES: polyethersulfone; CF: carbon fibre; GF: glass fibre.

DSC tests for all tape samples at a heating rate of 10 K/min were also carried out to study tape parameters such as glass transition temperature and melting temperature. Table 4 lists the DSC data extracted from DSC curves. The information provided from DSC was used to determine the heating plate temperature for the ATP process of tape welding. According to heating plate, temperatures can be found in Table 6.

Table 4.

Sample materials and properties determined through DSC.

Tape material	Sample weight (mg)	DSC range (°C)	T _g (°C)	T _m (°C)
CF-PP	4.1120	20-350	−	164.07
CF-PEEK	5.4850	0-450	−	342.01
CF-PA6	3.7320	0-350	98.04	219.45
GF-PP	8.3520	0-250	−	162.41
CF-PES	5.8440	0-450	221.77	−

DSC: differential scanning calorimetry; PEEK: polyether-ether-ketone; PP: polypropylene; PA6: polyamide 6; PES: polyethersulfone; CF: carbon fibre; GF: glass fibre; T _g: glass transition temperature; T _m: melting temperature.

ATP processing

Figure 2 shows the ATP head with its schematic used for the purposes of this research work. It consists of a thick steel heating plate with dimensions 700 × 350 mm² to lay tapes horizontally. The most important parts of the ATP head are the consolidation system and heating devices. There are three devices which can alter temperatures:

Figure 2.

ATP test rig used for producing bonded tapes. A schematic of the ATP process shows the different zones after the application of heat source. ATP: automated tape placement.

heating plate for tool and substrate,

consolidation roller and

direct flame.

The heating plate for tool and substrate heating serves two purposes. First, it should provide enough temperature in order to make the tape material tacky. This should make the first layer stick to the tool. Second, it should slow down the temperature decrease after the welding process to maximise polymer healing. The heating source used for nip-point heating is the direct flame. It produces a flammable propane–air mixture controlled by a mechanical circuit. Through the adjusted airflow, a membrane is triggered that adds gas to the air. For igniting the flame, a diode is used. While the direct flame aids in the first step of welding process, a consolidation roller helps to apply the compaction force, F, to provide intimate contact between the two surfaces. This roller is also tempered, since high thermal energy is produced by the direct flame that would heat up the roller, which is supposed to keep a constant temperature.

As one can see in Table 5, the materials’ combinations were chosen on the basis of a full factorial design. The only two exceptions are the combinations of CF-PA6/CF-PEEK and glass fibre (GF)-polyethersulfone (PES)/CF-PEEK, since it was not possible to find a suitable processing window in which the combinations would bond against each other. Other than that all combinations were successfully welded.

Table 5.

Tape material pairing and combination.

Tape material samples	CF-PP	CF-PEEK	CF-PA6	CF-PP	GF-PES
CF-PP	•	•	•	•	•
CF-PEEK		•		•	•
CF-PA6			•	•
GF-PP				•	•
CF-PES					•

PEEK: polyether-ether-ketone; PP: polypropylene; PA6: polyamide 6; PES: polyethersulfone; CF: carbon fibre; GF: glass fibre.

During ATP processing, the process parameters such as consolidation force (900 N) and consolidation roll temperature (80°C) are kept constant. One parameter was adapted according to subjective optical inspection after welding trials, namely, the incoming tape speed. Nonetheless, this parameter affects several other ATP-related phenomena. First of all, the reduced incoming speed automatically increases the net time of direct flame acting on the nip point. Consequently, the surface temperature is believed to rise. Second, the consolidation force acts for a longer duration on the interfaces. This could potentially increase the quality of polymer healing, resulting in higher quality bonding.

After ATP processing, there is a clear increase in width of the welded tape specimens. This is due to the consolidation force acting directly on the tapes. A graphical illustration (Figure 3) is given below to show width factor increase from its initial width of 24 mm. The initial width and final width are all noted before passing the samples for peel testing. High increase in tape width also suggests that the surfaces had a good intimate contact by achieving maximum adhesion.

Figure 3.

Mean increase in tape width profile for different tape material combination due to consolidation force.

Peel testing

Research work carried out in this article investigates how the wedge peel test compares against pull/T-peel test, and how these tests can help to characterise the adhesion properties of bonded thermoplastic tapes. For the wedge peel test, there is an effect of friction on tapes against the wedge blade but in the T-peel test, there is no such effect since the tapes are pulled apart. A schematic illustration can be found in the following figure.

When using a wedge peel test, the peel test controls the movement of the spindle drive to create the force F_P . The output is generated in a form of peel force with respect to tape displacement data. The loading mode is different to that of other readily available peeling methods but the process parameters identified are identical to wedge peel resistance tests and other standard tests like short beam bending or DCB test.⁴² Figure 4 shows the test equipment used for peel testing. Compared to the other peel tests explained in the past research work section, the wedge peel test can only be used vertically and at room temperature. This creates an easy qualitative possibility of fast characterisation. The peel sample is loaded vertically and there are steady fixtures that allow for smooth feeding of tape while it is pulled. The tape feeding unit consists of rollers to ensure vertical guiding for the entire peel length. The cleavage section of the tape is held firmly onto the tape alignment unit with the help of screw profiles. The bronze wedge-shaped blade is housed inside the blade housing unit and is rested on two load cells (Figure 4). The load cells are placed above the sensors and during pulling, the signals from the load cells are sent to the sensors for processing.

Figure 4.

Wedge peel test rig used for peeling of bonded tapes to study adhesion behaviour.

T-peel test is done using a pull/pressure universal testing machine, Zwick (Z1.0; Zwick GmbH & Co.KG, August-Nagel-Straße 11, 89079 Ulm, Germany). The ends of a bonded tape are held on clamps and a pull mechanism is employed to peel the tapes. A load cell of 1 kN is used during pulling. Bending of the specimen could be a problem during this test. Figure 5 shows the working mechanism of the T-peel/pull test. For both types of peel test, peeling speeds of 0.5 and 0.75 mm/s are chosen. A simple schematic drawing showing the forces acting on the tapes along with the tape dimensions during both peel tests can be seen in Figure 6.

Figure 5.

T-peel/pull peel test rig used for peeling bonded tapes to study adhesion behaviour.

Figure 6.

Half geometry of the wedge blade used for mathematical modelling.

Finite element model

It is obvious that for the two test setups presented in Figure 7, the measured forces cannot be directly compared. For the wedge peel test, the force is applied in the delamination crack plane, whereas for the pull test, the force is applied perpendicular to the crack plane. The resistance against delamination of the interface can be described by the energy release rate for mode 1 (i.e. normal) loading G_I . In a two-dimensional finite element model, the two test setups shown in Figure 7 are loaded and the G_I values are evaluated with the virtual crack closure technique.⁵⁵ For both test setups, a symmetric model is used because only the CF-polypropylene (PP)/CF-PP and GF-PP/GF-PP combinations were studied. The plies were described by an isotropic elastic material with a Young’s modulus of 98.1 and 14.7 GPa for the CF-PP and GF-PP, respectively. Those stiffness values were extracted from bending tests. A Poisson’s ratio of 0.3 is used for both materials. The calculations are carried out for a ply thickness of 0.15 mm. The model uses quadratic plane-strain elements. The region of the crack tip is discretised with eight elements over the ply thickness and other regions with four elements over the ply thickness.

Figure 7.

Schematic of wedge and T-peel/pull test when loaded and applied with a force, F_p .

For the wedge peel test, the wedge or knife is modelled rigidly. The half geometry of the knife is shown in Figure 6. Normal contact between wedge and ply is modelled using the augmented Lagrange formulation of SIMULIA/Abaqus and the tangential contact using a penalty contact algorithm with a friction coefficient of 0, 0.2, 0.4 or 0.6. Note that this friction coefficient is unknown in the wedge peel test, and the model thus shows its influence on the occurring peel forces by the variation of friction coefficient values.

For both the wedge peel test and the pull test, displacements are applied at the end of the ply. The forces due to those displacements are evaluated. At four nodes at the crack tip, the reaction forces and displacements are used to calculate the G_I for quadratic elements. The length of the ply from the crack tip to the end, where displacements were applied, is 180 mm. The length of the bonded ply region is 30 mm.

Results and discussion

In total, test samples of 13 different material configurations (Table 6) were produced through ATP and provided for wedge peel testing and T-peel/pull testing, respectively. For both types of testing, two different levels of test speed, 0.5 and 0.75 mm/s, were chosen. In total, this enables the study of three different aspects:

comparison of adhesive properties of the different material configurations as determined by the particular test methods,

comparison of adhesive properties of specific material configurations as determined with the different test methods and

comparison of test results obtained at different test speed levels.

Table 6.

Processing parameters utilised for the production of multi-material bonded tapes using ATP process.

	Test material combination
Configuration number	Tape 1	Tape 2	Incoming tape speed (mm/s)	Processing/heating plate temperature (°C)
1	CF-PP	CF-PP	50	140
2	CF-PP	GF-PP	50	140
3	CF-PP	CF-PA6	50	140
4	CF-PP	CF-PEEK	35	140
5	CF-PP	CF-PES	35	140
6	GF-PP	GF-PP	50	140
7	GF-PP	CF-PA6	50	140
8	GF-PP	CF-PEEK	50	140
9	GF-PP	CF-PES	35	140
10	CF-PA6	CF-PA6	50	190
11	CF-PEEK	CF-PEEK	35	250
12	CF-PEEK	CF-PES	35	200
13	CF-PES	CF-PES	35	200

ATP: automated tape placement; PEEK: polyether-ether-ketone; PP: polypropylene; PA6: polyamide 6; PES: polyethersulfone; CF: carbon fibre; GF: glass fibre.

Wedge peel test results

The test results of a typical wedge peel test are shown in Figure 8 by means of an example chosen from configuration no. 7 (Table 6), that is, a sample made of the material combination CF-PA6/GF-PP. Following a region of steady increase in peel force, that is, from the start of the experiment to a displacement of about 100 mm, the characteristics turn into a section of more stable data, which was further analysed in order to determine a peel force value characteristic for the sample under test.

Figure 8.

Wedge peel test result showing the variation in peel force with respect to position for a tape material combination of CF-PA6/GF-PP. The displayed graph also shows the different regions during peel testing. PP: polypropylene; PA6: polyamide 6; CF: carbon fibre; GF: glass fibre.

The data characteristics were processed by means of a specifically developed ‘envelope filter’, which represents the mean data (highlighted as solid black line in Figure 8) of top and bottom envelope curves, respectively. These in turn were computed as interpolating curves of local peak and valley data, respectively. In order to derive a single peel force value characteristic for the sample under test, the average value of the filtered data (indicated as a dashed line in Figure 8) was computed.

A full overview of peel force values is given in Figure 9 for wedge peel tests conducted at two different test speeds. The results are visualised by means of average and standard deviation values as determined from testing three specimens for each of the material sample configurations.

Figure 9.

Wedge peel test results spectrum for different tape material combinations at varying peel speeds.

T-peel/pull test results

For 5 of the 13 material configurations, in particular configurations 1 (GF-PP/GF-PP), 2 (CF-PP/CF-PP), 5 (CF-PES/CF-PES), 7 (CF-PA6/CF-PP) and 11(CF-PP/CF-PEEK), T-peel tests were conducted. All other material configurations showed critical failure by means of intralaminar fibre breakage. This caused a situation where these specimens could not be characterised.

Figure 10 shows the characteristics typically obtained from a T-peel/pull test as exemplarily chosen from configuration no. 1, that is, a sample from the material configuration GF-PP/GF-PP. Similarly to the results obtained from the wedge peel tests, a section of steady increase of peel force can be seen, which is followed by a section of more stable data. This again was further analysed in order to determine a peel force value characteristic for the sample under test by means of the ‘envelope filter’ and subsequent data averaging.

Figure 10.

Typical pull/T-peel test results showing peel pattern as well as maximum and mean for a tape material combination of GF-PP/GF-PP. PP: polypropylene; GF: glass fibre.

The results are finally compared with respect to the test speed and the two different test methods, as shown in Figure 11. While there is no distinct influence from the test speed, the test method clearly shows a significant impact. The results obtained with the wedge peel test method are considerably higher than those gained with the T-peel/pull test method.

Figure 11.

Wedge peel test result comparison with T-peel/pull test for chosen tape samples.

Modelling results

Figure 12 shows the energy release rate G_I plotted over the peel force F_p evaluated in the pull test model and wedge peel test model. All G_I over F_p curves are practically linear. For the pull test (see Figure 12(a)), a peel force of 50 N corresponds with a G_I value of 2000 J/m² and is practically the same for the CF-PP and GF-PP material. If delamination in the pull test would occur at 50 N, the energy release rate G_I would be 2000 J/m². For the wedge peel test (see Figure 12(b)), the G_I over F_p curve significantly changes with the friction coefficient µ between wedge and ply. For a µ value of 0 (frictionless), the G_I over F_p curve for the wedge peel test corresponds well with the curve for the pull test. At higher µ values, the slope of the G_I over F_p curves decreases – for a measured peel force F_p in the wedge peel test of 50 N, the G_I value would be 2000, 1200, 820 and 600 J/m² for µ values of 0, 0.2, 0.4 and 0.6, respectively. Assuming that the critical G_I value is the crucial physical parameter that describes the resistance of the delamination crack against propagation, this shows that in the wedge peel test, the friction coefficient can significantly change the measured forces. The curves of the pull test and the wedge peel test with a friction coefficient µ of 0 are nearly identical.

Figure 12.

Energy release rate G_I evaluated with the VCCT concept over the peel force for (a) the pull test model and (b) the wedge peel test model for different friction coefficients µ. VCCT: virtual crack closure technique.

Discussion

The work in this research article focussed on bonding characterisation of multimaterial interfaces. The results investigated the tape bonding polarity, process temperature analysis of individual tapes, width increase of tapes during ATP trials, two different peel test result comparisons, using varying peel speeds, data selection and filtering to identify suitable plateau region and using a simulation model to understand the effects of release energy in both peel test types.

The polarity of the tapes is of high importance and hence the measured CA values do help in understanding which tapes can actually bond against each other. The peeling of similar tape pairs depends on cohesion of the matrix, while peeling of tapes with different combinations depends on adhesion between the materials. Therefore, similar tape pairs tend to have higher peel force compared to tape combinations. Figure 9 confirms this trend. Tape combinations with CF-PES/CF-PA6 or PES with PP also showed reproducible results with the least errors when tested under different speeds. Lesser errors confirmed that the samples are reproducible and the adhesion between tapes tends to be more stable.

Cross-comparison of results was carried out using wedge and T-peel test results. Wedge peel test gave higher peel forces compared to that of T-peel/pull test. The samples tested with the wedge peel test should be carefully picked because any microcracks can cause the wedge blade to tear the tape samples. For the T-peel/pull test, the biggest concern was the bending stresses during pulling. When the cleavage end of the tape samples is clamped onto the testing machine (Figure 7(b)), the bonded part of the tape is left hanging under the influence of gravity. Those bending stresses might alter the measured peel forces. A typical peel result exhibits high force oscillations under the influence of staged crack development, as shown in Figure 13. In Figure 13, it can be clearly seen that there is a dramatic decrease in peel force.

Figure 13.

A typical peel result obtained from T-peel/pull test for a tape sample GF-PP/GF-PP, showing dynamic behaviour in peel force recordings. PP: polypropylene; GF: glass fibre.

The results in Figure 12 indicate that in the pull test, the peel force F_p shows a linear dependency on the energy release rate G_I . For the wedge peel test, this dependency is also linear but itself depends on the friction coefficient µ between the fractured surfaces and the wedge. Since those µ values are hard to obtain – no friction tests can be made of the fractured surface and the bronze surface of the wedge beforehand – it is hard to use the wedge peel test even for qualitative comparisons. If a certain ply combination shows higher peel forces than another ply combination, this can be either due to higher G_I values or due to higher µ values. Further studies would have to be made to measure friction coefficients and thus help to extend the applicability of the wedge peel test.

Conclusions

The work presented here compared surface adhesion between tape materials using two different peel-type testing methods. The characterisation was made through its polarity, peel region, bonding capabilities and the flaws during ATP processing of tapes, and also a peel model was made to study the effect of friction on peel force. A string of conclusions can be drawn from the above work and they are listed below:

Peel speed affects the peel force by increasing the force with speed. However, few tape combinations of CF-PES/CF-PES, CF-PA6/GF-PP and GF-PP/CF-PEEK showed negative response to peel speeds. This could be due to the polarity that affects the adhesion between tapes of different surfaces.

For CF-PES/CF-PES, the peel force obtained was low and the obtained peel results showed highly oscillating regions that made the mean value evaluation difficult.

Wedge peel test results showed regions with stable data points, but in the case of the T-peel/pull test, the scatter of data is highly oscillating, hence showed lesser stable data points for evaluation.

The plateau region with steady data region was sometimes difficult to identify when evaluating the mean peel force. There are many reasons why the peel force result can sometimes show oscillating characteristics. Micro-cracks can be one reason and this is confirmed on comparison with the smoother data in the results of the wedge peel tests than those of the T-peel/pull test.

The friction on tape surface can also have an influence on bonding during ATP production because there were times when the consolidation roller that applied pressure was dragged along the tape surface rather than having a controlled rolling motion. This could only reduce the peel forces in those regions where welding was affected due to the friction between roller and tape.

The modelling results confirm the dependency of peel force on the friction coefficient value of tape surface with the wedge blade surface when using a wedge peel test. In the wedge peel test, there is, however, an uncertainty about the friction coefficient between wedge and ply. There is also no straightforward calculation of G_I from peel forces. In the case of a pull test, a direct calculation of G_I from peel forces is possible but there are higher force oscillation amplitudes generated during peeling.

References

EL Dessouky

Lawrence

. Ultra-lightweight carbon fibre/thermoplastic composite material using spread tow technology. Compos B Eng 2013; 50: 91–97.

Lukaszewicz

DHJ

Potter

Eales

. A concept for the in situ consolidation of thermoset matrix prepreg during automated lay-up. Compos B Eng 2013; 45: 538–543.

Schledjewski

. Thermoplastic tape placement process – in situ consolidation is reachable. Plast Rubber Compos 2013; 38: 379–386.

Sonmez

Hahn

. Modeling of Heat Transfer and Crystallization in Thermoplastic Composite Tape Placement Process. J Thermoplast Compos Mater 2016; 10: 198–240.

Sonmez

Hahn

Akbulut

. Analysis of process-induced residual stresses in tape placement. J Thermoplast Compos Mater 2002; 15: 525–544.

Khan

Mitschang

Schledjewski

. Parametric study on processing parameters and resulting part quality through thermoplastic tape placement process. J Compos Mater 2013; 47: 485–499.

Mantell

Wang

Springer

. Thermoplastic tape laying. How Concept Becomes Reality 1991; 36: 1763–1772.

Bastien

Gillespie

. A non-isothermal healing model for strength and toughness of fusion bonded joints of amorphous thermoplastics. Polym Eng Sci 1991; 31: 1720–1730.

Dara

Loos

. Thermoplastic matrix composite processing model. Blacksburg, VA, United States: Dept. of Engineering Science and Mechanics, Virginia Polytechnic Inst. and State Univ. 1985. https://ntrs.nasa.gov/search.jsp?R=19860012148.

10.

Kline

Wool

. Polymer welding relations investigated by a lap shear joint method. Polym Eng Sci 1988; 28: 52–57.

11.

Brzeski

Mitschang

. Deconsolidation and its Interdependent Mechanisms of Fibre Reinforced Polypropylene. Polym Polym Compos 2015; 23: 515.

12.

Chen

. De-consolidation and re-consolidation in CF/PPS thermoplastic matrix composites. Compos A Appl Sci Manuf 2005; 36: 915–922.

13.

Davies

Courty

Xanthopoulos

. Surface treatment for adhesive bonding of carbon fibre-poly(etherether ketone) composites. J Mater Sci Lett 1991; 10: 335–338.

14.

Schuler

Keane

. Adhesive bonding of thermoplastic composites. I- The effect of surface treatment on adhesive bonding. In: International SAMPE Technical Conference, 19 th, Crystal City, VA, pp. 277–290.

15.

Ageorges

Hou

. Advances in fusion bonding techniques for joining thermoplastic matrix composites. Compos A Appl Sci Manuf 2001; 32: 839–857.

16.

Schell

Guilleminot

Binetruy

. Computational and experimental analysis of fusion bonding in thermoplastic composites. J Mater Process Technol 2009; 209: 5211–5219.

17.

Bourban

Bernet

Zanetto

. Material phenomena controlling rapid processing of thermoplastic composites. Compos A Appl Sci Manuf 2001; 32: 1045–1057.

18.

Beyeler

Guceri

. Experimental investigation of laser-assisted thermoplastic tape consolidation. J Heat Transfer 1988; 110: 424–430.

19.

Beyeler

Phillips

Güçeri

. Experimental investigation of laser-assisted thermoplastic tape consolidation. J Thermoplast Compos Mater 1988; 1: 107–121.

20.

Mazumdar

Hoa

. Experimental determination of process parameters for laser assisted processing of PEEK/carbon thermoplastic composites. In: International Sampe Symposium And Exhibition, p. 189.

21.

Maurer

Mitschang

. Laser-powered tape placement process-simulation and optimization. Adv Manuf Polym Compos Sci 2015; 1: 129–137.

22.

Werdermann

Friedrich

Cirino

. Design and fabrication of an on-line consolidation facility for thermoplastic composites. J Thermoplast Compos Mater 1989; 2: 293–306.

23.

Haupert

Chen

Friedrich

. Manufacturing of thermoplastic composite parts by combined filament winding and injection molding. In: Proceeding International Conference Comp. Materials ICCM-10, Whistler, Canada, pp. 381–388.

24.

Benatar

Gutowski

. Review of methods for fusion bonding thermoplastic composites. SAMPE J;(United States) 1987; 23: 35–42.

25.

Bullock

Boyce

. On-line consolidation of thermoplastic matrix composite tape using ultrasonic heating. Moving forward with 50 years of leadership in advanced materials. In: 39th International SAMPE Symposium and Exhibition, Anaheim Convention Center, Anaheim, California, 11–14 April 1994.

26.

Senders

van Beurden

Palardy

. Zero-flow: a novel approach to continuous ultrasonic welding of CF/PPS thermoplastic composite plates. Adv Manuf Polym Compos Sci 2016; 2: 83–92.

27.

Felderhoff

Steiner

. A new compact robotic head for thermoplastic fiber placement. Adv Mater Perform Through Technol Insert 1993; 38: 138–149.

28.

Rosselli

Santare

Güçeri

. Effects of processing on laser assisted thermoplastic tape consolidation. Compos A Appl Sci Manuf 1997; 28: 1023–1033.

29.

Ferraro

Di Vita

Marchetti

Cutolo

Zeni

. Laser Processing of Thermoplastic Matrix Filament Wound Composites. In: Füller

Grüninger

Schulte

Bunsell

Massiah

(eds) Developments in the Science and Technology of Composite Materials. Dordrecht: Springer, 1990.

30.

Stokes Griffin

Compston

. A combined optical-thermal model for near-infrared laser heating of thermoplastic composites in an automated tape placement process. Compos A Appl Sci Manuf 2015; 75: 104–115.

31.

Stokes Griffin

Ehard

Kollmannsberger

. A laser tape placement process for selective reinforcement of steel with CF/PA6 composites. Mater Des 2017; 116: 545–553.

32.

Funck

Neitzel

. Improved thermoplastic tape winding using laser or direct-flame heating. Compos Manuf 1995; 6: 189–192.

33.

Qureshi

Swait

Scaife

. In situ consolidation of thermoplastic prepreg tape using automated tape placement technology. Compos B Eng 2014; 66: 255–267.

34.

Calawa

Nancarrow

. Medium wave infrared heater for high-speed fiber placement, SAE Technical Paper, 1 January 2007. https://www.electroimpact.com/WhitePapers/2007-01-3842.pdf.

35.

Karpachevskyy

. Infrared tape preheater optimization for in situ thermoplastic consolidation. In: SAMPE, 2014.

36.

CATAPULT. Flashlamp heating for composites manufacture. (2016, accessed 28 September 2017).

37.

Williams

. A new light on composite heating. JEC Compos Magaz 2017; 111: 46–48.

38.

Standard

. D1002-10. Standard test method for apparent shear strength of single-lap-joint adhesively bonded metal specimens by tension loading (Metal-to-Metal). West Conshohocken: ASTM International, 2010.

39.

ISO B. Fibre-reinforced plastic composites. Determination of mode I interlaminar fracture toughness, GIC, for unidirectional reinforced materials. BSI, Geneva, Switzerland. 2002.

40.

Villegas

Moser

Yousefpour

. Process and performance evaluation of ultrasonic, induction and resistance welding of advanced thermoplastic composites. J Thermoplast Compos Mater 2013; 26: 1007–1024.

41.

Grouve

Warnet

Akkerman

. Critical assessment of the mandrel peel test for fiber reinforced thermoplastic laminates. Eng Fract Mech 2013; 101: 96–108.

42.

Hulcher

Marchello

Hinkley

. Correlation between double cantilever beam and wedge peel tests for automated tow placement. Materials and Process Affordability- Keys to the future (1998) 1955–1965. In: 43rd International SAMPE Symposium and Exhibition (Proceedings). Part 2 (of 2), Anaheim, US, May 31–Jun 4 1998.

43.

Schledjewski

. Mechanical performance of in-situ consolidated thermoplastic fiber reinforced tape materials. In: Proceedings ECCM. 2004.

44.

Kawashita

Moore

Williams

. Comparison of Peel Tests for Metal–Polymer Laminates for Aerospace Applications. J Adhes 2005; 81: 561–586.

45.

Grouve

WJB

Warnet

Akkerman

. Weld Strength Assessment for Tape Placement. Int J Mater Form 2010; 3: 707–710.

46.

Rao

Umer

Thomas

. Investigation of peel resistance during the fibre placement process. J Reinf Plast Compos 2016; 35: 275–286.

47.

Comer

Hammond

Ray

. Wedge peel interlaminar toughness of carbon-fibre/PEEK thermoplastic laminates manufactured by laser-assisted automated-tape-placement (LATP). SETEC 2014; 14: 10–11.

48.

Denkena

Schmidt

Völtzer

. Thermographic online monitoring system for Automated Fiber Placement processes. Compos B Eng 2016; 97: 239–243.

49.

Ström

Fredriksson

Stenius

. Contact angles, work of adhesion, and interfacial tensions at a dissolving Hydrocarbon surface. J Colloid Interface Sci 1987; 119: 352–361.

50.

Rudawska

Jacniacka

. Analysis for determining surface free energy uncertainty by the Owen–Wendt method. Int J Adhes Adhes 2009; 29: 451–457.

51.

Owens

Wendt

. Estimation of the surface free energy of polymers. J Appl Polym Sci 1969; 13: 1741–1747.

52.

Rabel

. Flüssigkeitsgrenzflächen in Theorie und Anwendungstechnik. Phys J 1977; 33: 151–161.

53.

Kaelble

. Dispersion-Polar Surface Tension Properties of Organic Solids. J Adhes 2008; 2: 66–81.

54.

Fowkes

. Attractive Forces at Interfaces. Ind Eng Chem 1964; 56: 40–52.

55.

Krueger

. Virtual crack closure technique. Appl Mech Rev 2004; 57: 109–143.

Peel strength characterisation on ply/ply interface using wedge and T-peel/pull-type tests

Abstract

Keywords

Introduction

Experimental work

Materials selection and cutting

Tape characterisation

ATP processing

Peel testing

Finite element model

Results and discussion

Wedge peel test results

T-peel/pull test results

Modelling results

Discussion

Conclusions

References