The feasibility of welding APC-2 thermoplastic composite material

TWI Bulletin, July 1987

By Nicola Taylor

Nicola Taylor, BSc, is a Research Engineer in the Plastics Joining Section.

Thermoplastic composite materials which can be welded will have considerable advantages over the more commonly used composites. Welding trials detailed here show that the well established methods used for welding unreinforced thermoplastics can indeed be applied to composite materials.

High performance thermoplastic, continuous fibre, composites, originally developed for the aerospace industry, are increasingly interesting to the high volume, mass production industries. Such materials are of interest because they have significantly better properties than currently available thermoset composites, particularly with respect to fracture toughness, damage tolerance and resistance to hot, wet environments. Thermoplastic composites are more difficult to bond adhesively than thermoset materials, but they are hot-formable and therefore should be weldable. Weldable thermoplastic composites have an advantage over the non-weldable thermoset materials.

The aim of the investigation reported here was to examine the feasibility of welding thermoplastic carbon fibre composite laminates by well established methods currently used for joining unreinforced thermoplastics. All the experimental work was conducted on one material, APC-2 (carbon fibres in a polyetheretherketone (PEEK) matrix) manufactured by Imperial Chemical Industries (ICI), which is currently the most widely available thermoplastic composite.

This work is being continued in the form of a Group Sponsored Project, which is further developing the welding techniques described here, as well as assessing other novel techniques for joining APC-2.

Composite materials

Composite materials can be classified in many ways, but there are three main aspects which need to be included in a definition of a composite material for use in structural applications: ^[1]

1. It consists of two or more physically distinct and mechanically separable materials;
2. It can be made by mixing separate materials in such a way that dispersion of one material in the other is achieved in a controlled way to give optimum properties;
3. Certain properties of the composite material are superior, and possibly unique in some specific respects, to those of the individual components.

In fibre reinforced plastics, fibres and plastics (resin) with some excellent physical and mechanical properties are combined to give a material with new and superior properties. Fibres have high strength and modulus but these properties are only developed in fine fibres, with diameters in the range 7-15µm, and such fibres are usually brittle. Plastics may be either ductile or brittle and can have considerable resistance to chemical attack. By combining fibres and resin, a bulk material may be produced which has a proportion of the fibre strength, the chemical resistance of the resin and a combination of other physical and mechanical properties of the fibres and resin. Composite materials may also exhibit specific properties as a direct result of combining the resin and fibres, e.g. resistance to crack propagation and an ability to absorb energy during deformation.

The technology of fibre reinforced plastics has made dramatic progress over the last 30-40 years.^[2] In volume terms, random glass reinforced polyester, wet lay-up composites still dominate the low technology areas of applications such as boat building. The demand for better mechanical and environmental properties, however, has led to the development of epoxy resin based composites, reinforced by woven glass fabrics. Control of the fibre alignment in these materials allows higher strength and stiffness to be achieved, and allows more efficient use of materials. However, the properties of the basic materials still limit their application to relatively lightly loaded components. About fifteen years ago carbon and aramid (aromatic polyamide), high modulus reinforcing fibres were introduced. These resulted in aligned fibre composites with strength/weight and stiffness/weight ratios far superior to those of metals, and extended possible composite material applications to highly loaded lightweight components in aerospace, sports goods and general engineering (Table 1). The good mechanical properties of epoxy resins made them the automatic choice of matrix for most applications.

Table 1 Existing applications for fibre reinforced composite materials based on plastics

Large numbers of composite materials are available, giving a wide range of properties. The most widely used fibre in load bearing applications is glass. For special high performance duties, carbon fibres and aramid fibres are also used. Fibre length varies according to the application and processing methods employed. Short fibres, i.e. < 2mm, are used to reinforce injection moulded components. Long fibres are aligned within the matrix, resulting in an anisotropic material. Continuous fibres are used in laminate composites, resulting in high strength materials. The fibre orientation is used to control the degree of anisotropy in the mechanical properties of the final product.

Thermoplastic composites

Until recently, little attention has been paid to the use of thermoplastic resins as matrix materials for high performance laminated composites. Most work with fibre reinforced thermoplastics has been limited to injection moulding compounds (i.e. with short fibres). The major limitations of conventional thermoplastics have been: low elastic modulus, low softening temperature, poor solvent resistance, and poor fibre/resin bond strength. During the past few years, a range of engineering thermoplastics has been developed which largely overcomes these limitations, and offers mechanical properties and temperature/solvent resistance comparable with the best of the thermosets. Table 2 gives properties of typical thermoset and thermoplastic matrix materials. Whilst strength (at yield) and modulus values for the thermoplastics are generally similar to those for the epoxy system, plastic flow of the thermoplastics under load gives rise to elongation at failure of 35-100%, compared with 1-2% for the epoxies. The improved toughness of the thermoplastics leads to greatly enhanced impact properties in composite form.

Other properties of thermoplastic resins which make them more attractive than thermosets for use in composite materials are:

1. Consistency - because polymerisation is complete, thermoplastic resin users are not concerned with chemical changes taking place during the moulding process. Thermoplastic composites therefore have, in principle, a high level ofquality and consistency.
2. Shelf life - the short shelf lives associated with most thermosetting resin systems do not apply to thermoplastics.
3. Freedom from flame and toxic fume - some thermoplastics are non-flammable and do not give off toxic fumes when subjected to external sources of heat.
4. Ability to thermoform - thermoplastic matrix materials can be shaped using plastic forming (thermoforming) techniques.^[3] The extent to which such techniquesmay be applied to composite materials depends on the form of the fibre reinforcement used.
5. Environmental resistance - for example, the crystalline PEEK matrix is resistant to virtually all solvent attack.

Table 2 Properties of unreinforced polymers

Aromatic polymer composites (APC)

PEEK/carbon fibre composite is an aromatic polymer composite i.e. it combines a semi-crystalline aromatic polymer thermoplastic matrix with a continuous carbon fibre reinforcement.

APC-2 is one particular grade of this material produced by ICI. Sheets of the material are fabricated by consolidating multiple layers of unidirectional pre-impregnated tapes. The tapes are first tacked together using a soldering iron. They are then placed between aluminium foil, and mirror finish press plates, put into a hot platen press and subjected to a temperature above the melting point of the polymer, i.e. 343°C.^[4] Controlled cooling allows the polymer to crystallise and form a rigid structure. ^[5]

APC-2 composites have a complex hierarchy of substructures. The largest sub-units are the stacked 125µm plies of unidirectional impregnated tapes (prepreg). These tapes can be stacked in any number of directions until the required stiffness and strength is achieved. The carbon fibres within each prepreg layer are about 7µm diameter and typically occupy 60% of the volume. The gaps occupied by the polymer matrix between the fibres are therefore in the same size range as the fibres, i.e. 1-10µm. The morphology of the resin matrix is semi-crystalline. The percentage crystallinity usually lies between 20-40% in a completed APC-2 sheet. ^[6]

Although APC-2 is currently the most widely available of the continuous fibre composite materials, others are under development. Examples of these are:

1. K-polymer, a thermoplastic polyimide developed by E I Du Pont De Nemours and Company Inc for use with carbon or Kevlar (R) 49 aramid fibres. This is an amorphous polymer and so does not suffer from problems related to thedevelopment and control of crystallinity. ^[7]
   
   
2. Noryl, a polyphenylene oxide resin for use with continuous fibres being developed by General Electric Company.
   
   
3. Torlon, a poly (amide-imide) resin for use with continuous fibres. Torlon polymers result from the condensation reaction between trimetallic anhydride (TMA) and various diamines. They are manufactured by the Amoco ChemicalsCorporation. ^[8]
   
   
4. Ultem, a polyetherimide resin. This amorphous thermoplastic is based upon regular repeating ether and imide linkages. Ultem is manufactured by the General Electric Company. ^[8]

Properties of these materials are given in Table 3.

Table 3 Mechanical properties of unreinforced resins

Experimental welding trials on APC-2

Objectives and welding techniques

The objective of this programme was to assess a range of techniques for joining APC-2 as follows:

1. Hot plate welding - a heated plate is clamped between the surfaces to be joined until they soften. The plate is then withdrawn and the surfaces brought together under controlled pressure, for a specific period(Fig.1).
2. Ultrasonic welding - the parts to be assembled are held together under pressure and subjected to ultrasonic vibration at right angles to the plane of the contact area (Fig.2). The consequent alternating high frequencystresses generate heat in the plastic and, if the components are properly designed, this heat can be selectively generated at the joint interface.
3. Vibration welding - frictional heat is generated by relative movement between the two parts to be welded which are held together under pressure (Fig.3). The movement consists of linear oscillations. Once molten materialhas been generated at the joint interface, vibration is stopped, the parts are aligned and the weld consolidates on cooling.
4. Inductively heated implant (induction) welding - a metal insert is placed between the two parts to be joined. The insert is heated by induction heating so that the plastics material around the implant melts and fuses to form ajoint. A high frequency field (in this case 2.3MHz) is used to induce an electric current, and hence heat, in the metal implant (Fig.4).

A more detailed description of these processes is given in Ref ^[9]. Details of the welding equipment employed for the various trials are given in Table 4.

Table 4 Specifications of welding equipment

Materials and testpiece design

The welding trials were conducted on 2mm thick (16-ply) quasi-isotropic APC-2 sheet with fibre orientation (0/45/90/-45°)2s. The 's' indicates that the lay-up is symmetrical and the '2' that the lay-up is repeated. The material was prepared by ICI such that the sheet surface was free from mould release lubricants, known to cause problems when adhesively bonding APC-2. The inter-laminar shear strength of the material is claimed, by the manufacturers, to be 56 N/mm². However, this result was obtained by a three point bend test and is not considered comparable to the tensile test performed on the welded lap-shear specimens. Therefore, parent material specimens were produced, by ICI, of the same geometry as the welded specimens. These were laid up from prepreg and consolidated in an autoclave. These specimens, when tested in the same manner as the welded specimens, had strength of approximately 45 N/mm². This value is used as a comparison for the welded specimens.

The continuous fibre nature of the material made it impractical to use a butt joint configuration for the welding trials, as it was impossible to produce fibre continuity across the joint. For this reason a lap joint was used (Fig.5). This specimen configuration is recommended in ASTM standard D 1002-72^[10] for the testing of adhesives.

Specimens 25.4 x 101.6mm were cut from a flat sheet using an MK, ultra-thin rim, diamond rotary blade. Cut specimen edges were deburred with fine wet and dry abrasive papers and were thoroughly washed and dried. Throughout the tests the surface fibre orientation was arranged to be parallel to the length of the specimen.

Welding procedures

Hot plate welding

A hot plate temperature of 365°C was chosen for the initial studies (measured using a Ni/Cr, Ni/Al thermocouple probe and digital thermometer). This lies within the processing temperature range quoted for the material.^[4] Heating and forge pressures of 0.5 and 1.75 N/mm² respectively were set, based on published conditions for the production of APC-1.^[5] Heating times of 5-120sec were investigated at these conditions. The cooling time for the weld, i.e. the time for which the welded specimen was held under forge pressure, was equal to the heating time.

Further trials investigated the effect of increasing the hot plate temperature to as high as 379°C. The heating pressure and forge pressure were as before.

Ultrasonic welding

For ultrasonic welding the variable welding parameters are time, pressure and vibration amplitude. The amplitude 0.04mm was used for the welding trials. This value lies in the middle of the range normally considered when welding semicrystalline materials.^[11] Times in the range 1-2.5sec were evaluated, at welding pressures between 1.0-2.0 N/mm². A total of 10 welds were looked at under these conditions, which covers the range normally used when welding semicrystalline materials. A flat rectangular faced welding horn was employed, which matched the joint configuration.

Vibration welding

Trials were conducted using the standard range of welding parameters for the equipment. The effect of weld times between 1-6sec was investigated with weld pressures in the range of 1-2.5 N/mm² and vibration amplitude between 2-3mm. The direction of vibration amplitude was parallel to the length of the specimen.

Induction welding

Initial work involved constructing a solenoid work coil of suitable dimensions to accept APC-2 specimens and of correct resonant frequency to match the power supply (2.3MHz).

An 11 turn coil, made of 6.4mm outside diameter copper tube was used, with dimensions 50mm internal diameter and 110mm length. 1µm silver wire was used as the metallic insert.

Initially, welds were made with silver wire wrapped around the end of one of the pieces of APC-2 using one turn of wire/mm but this led to excessive heat generation in this piece and insufficient heat generation in the other. Welds were then attempted with silver wire wrapped around both pieces; however, this caused a shorting effect and negligible heat was generated.

Testing and examination procedure

Samples suitable for shear testing were produced to the dimensions given in Fig.5. These samples were shear tested at room temperature using an Avery-Denison universal testing machine (Type 7152) at a cross head displacement rate of 10 mm/min. The jaws were 2mm offset so that the tensile forces were parallel to each half of the welded specimen. The shear strength of the weld was calculated by dividing the maximum failure force achieved by the original overlap area.

Untested welds were examined visually, both externally and by sectioning and polishing. Sections were taken both parallel to and at 90° to the length of the specimen (Fig. 6).

Results

Hot plate welding

The results of the shear tests showed that bonding could be achieved over the total overlap area, resulting in strengths up to 38 N/mm², i.e. 84% parent material strength. Figure 7 illustrates the effect of heating time on joint strength, showing that longer heating times, i.e. 120sec, resulted in the strongest bonds.

Figure 8 illustrates the influence of hot plate temperature on joint strength. A typical fracture is shown in Fig.9a. This illustrates some fibre displacement and flash. The fracture surface showed that, on the top layer, the fibres were no longer parallel to the length of the specimen after welding. This displacement was not limited to the surface layer of fibres. Fibre displacement at the weld is also illustrated in the section shown in Fig.9b. Welds with low failure forces, i.e. those made at high hot plate temperatures, suffered surface lamination foldover (Fig.10a). This was a result of the surface laminations sticking to the heating element when it was withdrawn and folding over during the subsequent forging cycle, giving limited surface contact and thus low failure strengths. The consequent void formation within the weld is shown in Fig.10b.

Ultrasonic welding

Limited trials in ultrasonic welding of APC-2 resulted in maximum weld strengths of only 5 N/mm² (11% parent material strength). A typical weld fracture surface showing that uniform bonding was not obtained across the joint is illustrated in Fig.11. Heat was only generated at point contacts on the specimen faces.

The majority of welded specimens failed during handling.

Vibration welding

Vibration welding of APC-2 using low pressure and vibration amplitude resulted in little or no bonding of the surfaces. Typical strengths were of the order of 1 N/mm² (Fig.12a). Increasing the welding pressure or vibration amplitude resulted in higher strength welds. The results are presented in Fig.12b for an increased welding pressure and Fig.12c for increased vibration amplitude. The latter had the more marked effect, giving maximum strengths of around 17 N/mm², 35% of parent material strength.

At all the welding conditions, the weld strength tended to increase with weld time. However, the degree of flash and fibre displacement, illustrated in Fig.13a and b, also increases with weld time.

Induction welding

The results of the preliminary trials showed that bonding could be achieved by this method, however the strengths obtained were low (maximum 5 N/mm², 11% parent material strength). Figure 14 shows the influence of weld time on joint strength. Within the range investigated joint strength increased with heating time. Complete face bonding was not achieved in any of the welds, and in all cases edge heating effects were incurred in the APC-2. Preferential heating at the edges was caused by inadequate design of the induction coil.

Discussion

The trials conducted have proved that it is feasible to weld the thermoplastic composite material, APC-2. The four processes investigated showed varying degrees of success. Hot plate welding resulted in the highest weld strength, although weld times were long, totalling up to 240sec. The results obtained suggest that raising the hot plate temperature could reduce the weld time. Further work is required to eliminate the problem of sticking to the hot plate with the aim of producing more consistent welds.

The ultrasonic welding process, with its short weld times, is ideally suited to mass production, unlike hot plate welding. Although the weld strengths were low, only limited trials were carried out and further assessment of the effect of amplitude is required. Joint design is also considered to be important in the ultrasonic welding of thermoplastics. The basic requirement of joint design for ultrasonic welds is a small, uniform initial contact area. In the most commonly used joint design, this is provided by a triangular section called an energy concentrator (also known as an energy director or projection). This gives a high concentration of vibrating energy at its apex resulting in rapid heat buildup.^[11] Because APC-2 is laid up as flat sheets, an energy concentrator cannot be included in the design of the joint as it would with a moulded component, therefore, an external energy concentrator may be introduced. Those which have been suggested for further investigation include a PEEK interlayer, metal wire or mesh.

Vibration welding is a technique more suited to the welding of large parts than ultrasonic welding, i.e. complex joints < 2m long. A major disadvantage of the technique, when used on APC-2, is the upset and edge distortion. The acceptable amount of flash for any weld will be determined by the application. The flash may also have a harmful effect on joint strength because of the large degree of fibre displacement.

Induction welding is an attractive technique. The equipment may be made portable and hence it has the potential for welding large areas. The results of these welding trials show the process to be feasible, but work is required to gain a greater understanding of the heat distribution with respect to coil design, before welding parameters and implant characteristics can be optimised.

Summary

Continuous fibre reinforced thermoplastic composite materials are being considered for many future applications. They will be used to replace metals and thermoset materials where their corrosion resistance, high strength/weight ratio and processability are beneficial. For many applications, it will be necessary to join these materials and the development of joining techniques will encourage and increase their use. The welding feasibility trials conducted on APC-2, a thermoplastic composite material, have enabled the following conclusions to be drawn:

1. APC-2 can be successfully joined by hot plate welding. High plate temperatures, 375°C, and long welding times, ~ 120sec, result in welds up to 84% parent material strength in a lap-shear specimen.
   
   
2. It was not found possible to obtain a complete face weld between planar surfaces using ultrasonic welding. Welds of only 11% parent material strength were obtained. The use of an energy concentrator may make ultrasonic welding successful.
   
   
3. Vibration welded shear test specimens gave failure strengths of up to 17 N/mm², 38% parent material strength. The best results were obtained when using 2.5 N/mm² welding pressure and 3mm vibration amplitude.
   
   
4. Induction welding produced only weak welds, up to 11% parent material strength. This is believed to be because the induction coil was of incorrect design.

Acknowledgements

The work was funded jointly by Research Members of The Welding Institute and the Minerals and Metals Division of the UK Department of Trade and Industry.

References