Measuring properties of plastics welds
TWI Bulletin, March/April 1990
Geoff Hale is Head of the Advanced Materials Section in the Engineering Department, where he is responsible for a small group working on engineering, performance, design and related activities associated with polymers, composites and adhesively bonded joints.
Geoff joined TWI in 1983 after completing his PhD at Leeds University. For four years he worked in the Materials Department on the welding metallurgy and corrosion behaviour of stainless steels, but since 1987 he has concentrated on developing TWI's expertise in the engineering aspects of non-metallic materials.
Geoff's current technical interests include development of standardised fracture mechanics test procedures for parent polymers and welded thermoplastic joints, provision of reliable standardised materials design data for welded and adhesively bonded joints, and increasing industrial awareness of potential applications for adhesives technology in manufacturing. Geoff is a Chartered Engineer and a Professional Member of the Institute of Metals.
In this article, the emphasis is on the short-term mechanical property tests employed to assess welded joints in thermoplastics. The properties most commonly measured are tensile performance and impact toughness. Other short-term tests, e.g. determination of flexural modulus, are referred to briefly. On their own, strength and toughness data may be adequate for certain applications, but in load-bearing situations where structural integrity is important, additional properties such as fracture toughness, creep and fatigue may be needed to satisfy all the design requirements. A recent study in support of a design data programme - initiated by the Polymer Engineering Group, the British Plastics Federation and the British Rubber Manufacturers' Association - has indicated that, outside the gas and water industries' well-defined requirements relating to welded joints in pipes, reliable materials design data for welded plastics joints are scarce.
Before looking at tensile testing in particular, a brief resume of the more common short-term tests applied to plastics is appropriate. These tests are used to determine:
- tensile properties;
- compression strength;
- flexural strength and modulus;
- shear strength;
- hardness;
- impact toughness (energy).
For plain materials, the majority of the tests used are covered by national and/or international standards, e.g. BS, DIN, ASTM. However, there are no standards specifically for welded joints. The German Welding Society (Deutscher Verband für Schweisstechnik) has produced a series of qualitative recommendations in document DVS 2203.
At present, there is a proposal before the European Standards body (CEN) to develop European Standards for welded joints in thermoplastics. TWI expects to play an important role in developing such standards.
Tensile testing
For parent material, tensile tests are widely standardised and in the UK are covered by BS 2782: Part 3: Methods 320A to 320F. An additional standard (BS 7008: Part 2: 1988) has been issued recently and provides further recommendations on determination of tensile properties with the aim of improving comparison between sets of data generated by different organisations.
At TWI, a dumb-bell testpiece has been adopted for welded joints based on method 320C of BS 2782, with the weld located at the centre of the waisted section ( Figure 1). In many cases, a simpler parallel-sided testpiece can be used for initial assessments of the influence of changes in welding parameters. However, since the aim is to achieve weld strengths equivalent to those of the parent material, a waisted sample is preferred to avoid fracture at the grips. With this type of specimen, it is possible to show up differences in weld quality ( Figure 2). It is recommended that any weld flash is machined off prior to testing so that effects arising from changes in the size of the weld reinforcement are eliminated and the stress concentrating influence of any notch beneath the flash is removed.
Fig. 1. Tensile specimen geometry (from BS 2782: Part 3: Method 320C), adapted for welded samples (dimensions mm)
Fig. 2. Effect of loading rate on the tensile strength of polypropylene hot plate welds with the flash removed
With thermoplastics, especially parent materials, a testing machine with a long crosshead displacement is needed. Under certain loading rates, extremely high elongations (up to 400%) may occur in some of the more ductile plastics, e.g. polypropylene.
The type of flat testpiece referred to above is ideal for assessing welds in sheet and for butt joints in large diameter pipe, but it is not practical for smaller diameter pipe (say 120mm diameter or less), for which a curved specimen is often used ( Figure 3). The flash on hot plate pipe butt welds is often large and, for the reasons given above, it is again preferable to remove it from this type of specimen, As shown in the figure, small circular-ended side notches ensure that deformation starts in the weld. However, as deformation continues, parent material tends to be pulled out from one side of the weld and the weld itself does not deform. It is thus difficult to measure elongation accurately and the performance of parent material relative to the weld is not clarified. While sharper notches might confine deformation and failure to the weld zone, it is also likely that they would cause premature failure because of the stress concentration effect.
Fig. 3. Tensile test specimen adopted for welded plastic pipe: a) Geometry (dimensions in millimetres)
b) Fractured plastics pipe tensile specimen
With most plastics welding processes, other than hot gas welds where a filler is added, the weld zone is narrow (often less than 0.5mm wide) and it is therefore impossible to obtain an all-weld material testpiece in a similar manner to metal weldments. Thus assessment of the actual material behaviour of the weld zone is difficult. Examination of fracture faces is therefore important to determine whether ductile or brittle features are present, since this gives additional information on the qualitative nature of the bonded region.
To summarise, the tensile test can be used to assess differences in weld quality; it can indicate when a joint is particularly weak, but the weld material itself is not easily sampled. Unless the application dictates otherwise, the loading rates of 5 or 50 mm/min as specified in BS 7008: Part 2: 1988 should be used. In some cases, it may be necessary to check that there is no indication of a transition in fracture behaviour with loading rate.
Impact toughness testing
In many situations, material toughness - in particular, resistance to impact loading - is one of the primary properties required by design engineers. Since all plastics welding processes tend to introduce some form of notch, usually at the joint centreline, it is important to assess the effect of such notches on toughness.
The simplest impact test is to hit a section of welded plastic with a hammer and examine the appearance of the fracture which can indicate whether or not the material/joint is brittle.
Amore rigorous approach is to use a standardised pendulum impact testing machine. Such machines measure the energy absorbed by the material during the test and, for specimens of different sizes, the data can be normalised with respect to ligament area, i.e. the area of material fractured during the test.
Several different impact tests have been standardised for plastics, but TWI tends to use the Charpy test in accordance with BS 2782: Method 359: 1984. When applying this test to plastics welds, a number of points need consideration: notch location, notch sharpness, weld quality, and weld flash. They are discussed briefly below.
Notch location
Most of the welding processes produce a weld bead or flash, the size of which varies depending on the process and the welding parameters used. When the weld bead is small and symmetrical, it is possible to use the centre of the bead as an accurate locator of the weld zone and the notch can be positioned centrally within the weld area. However, when the weld flash is larger and possibly unsymmetrical, as seen with some medium density polyethylene (MDPE) welds, the above approach may be unsuccessful (
Figure 4). Alternative methods, such as short-term etching to reveal the extent of the weld zone, are being investigated.
Fig. 4. The difficulties of accurate notch location when the weld bead is unsymmetrical
Notch sharpness
The V notch generally employed for Charpy tests on plastics has a root radius of 0.25mm. With a larger weld zone, as often seen in PE butt fusion welds for instance, the notch tip can be correctly positioned, but where the weld region is much narrower, e.g. in vibration welds, even the tip of the notch may well extend outside the weld area and thus the measured energy will not be representative of the weld properties.
In addition, with tougher thermoplastics such as MDPE and polyvinylidenedifluoride (PVDF), complete fracture may not occur with the standard V notch at room temperature and hence the precise value of the measured impact energy is unclear.
Since those defects which cause premature failure of a welded joint almost invariably occur at the weld centreline, any test to assess the significance of defects must attempt to sample the interface material in this region. Taken together, the various factors outlined above suggest that a sharper notch could be more appropriate for welded joints in plastics, but even so there may be difficulties locating the tip precisely into an area only 0.1 to 0.2mm in width. Razor/scalpel blades have a sufficiently small notch tip root radius (around 20µm); the strain field generated at the tip of such a notch and its effect on toughness are areas of current investigation at TWI.
Weld quality and flash
Even in its present form, the Charpy test is capable of distinguishing differences in weld quality ( Figure 5). Generally, it is better to remove the weld flash as this tends to result in reduced scatter. When the flash is left intact there are two counterbalancing effects:
- Extra energy is required to fracture the flash material itself;
- The notch beneath the flash may act as a stress concentrator.
These two opposing factors do not necessarily cancel and their relative effects may vary with changes in welding conditions.
Fig. 5. Influence of weld quality and weld flash on the Charpy impact energy of polypropylene hot plate welds
Structural examination
As with tensile testpieces, when possible examination of fracture faces in the scanning electron microscope and of microtome sections by transmitted light microscopy should be undertaken to aid interpretation of the effects of welding on joint toughness. Welds exhibiting higher impact toughness generally fail in a more ductile mode, primarily along the 'fusion boundary' ( Figure 6). This is indicative of good bonding at the weld centreline. Conversely, lower toughness values are associated with planar fracture along the weld centreline which often appears to have a different structure from the remainder of the flowed zone ( Figure 7).
Fig. 6. Fracture morphology observed in a Charpy impact specimen extracted from a good quality polypropylene hot plate weld: a) Microtome section cut transverse to the notch illustrating that failure occurred predominantly along the fusion boundary
b) SEM photograph showing a coarser rippled morphology in the fusion boundary region, but a finer structure where failure tool place at the weld centreline
Fig. 7. Charpy impact fracture appearance in a poorly-bonded polypropylene hot plate weld: a) Failure predominantly along the plane of the weld centreline
b) Boundary between the main fracture face and the smaller raised plateau seen in Figure 7a, illustrating the more ductile character of the fusion boundary fracture
Instrumented impact testing
An alternative to the simple pendulum machine described above is to instrument the striker so that load or force versus time and displacement plots can be obtained. With this approach, it is possible to differentiate between various features in the fracture process, e.g. the energy to initiate damage, whereas with a non-instrumented machine only the total energy for the complete impact event is recorded.
Both instrumented pendulum machines and falling or driven weight instruments are available. The Rosand instrument installed recently at TWI is a falling weight machine ( Figure 8).
Fig. 8. Instrumented falling weight impact machine
At present, instrumented impact testing of plastics welds is relatively uncommon. However, with a fully instrumented system, it is possible to discriminate between different parts of the impact event and this provides a method of establishing how impact fracture proceeds in such welds; answering questions such as: 'Are certain joint morphologies more resistant to fracture than others?' 'Which microstructural features are preferable for good toughness?' None of these factors can be readily studied with a conventional non-instrumented machine. More detailed investigation of impact fracture in welded thermoplastics will be carried out as part of TWI's Co-operative Research Programme over the next 12-18 months.
In conclusion
This article summarises current thinking and knowledge on the application of the most common short-term mechanical property tests applied to welded joints in plastics. The most important factors have been discussed and illustrated with reference to appropriate results and fractographic observations. Where possible, areas of future developments have been highlighted.
Finally, we must accept that there is at present no simple short-term test which can be used as a reliable predictor of the long-term behaviour of plastics welds. That goal has still to be achieved.
