Polyethylene and polypropylene hot plate welds - a structural evaluation

TWI Bulletin, January/February 1994

Sheila Stevens is a Principal Research Chemist in TWI's Chemical Laboratory. Her work at TWI has involved analysis of steel weld metal, fluxes, wire and wrought products, specialising in modern instrumental methods. This has extended to the study of theoretical aspects of optical emission spectroscopy.

Sheila has also worked on projects involving sampling and analysis of welding fume and gases, including local extraction ventilation, determination of free nitrogen by hydrogen hot extraction, and extraction and analysis of gases from pores in weld metal. The last five years have been a change of emphasis in her work which is now directed towards use of chemical and physical techniques for characterisation of structural properties in weld areas in thermoplastics.

By a combination of optical and scanning electron microscopy Sheila Stevens shows that it is possible to relate features of hot plate welds in polyethylene and polypropylene to the welding procedure adopted.

Polyethylene (PE) and polypropylene (PP) are two of the most widely used semi-crystalline thermoplastics, finding uses in gas and water pipes, and in a variety of injection moulded industrial components. It is, therefore, important to be able to join these materials in a reliable fashion, and welding is often used.

An inherent part of any investigation into joining these materials is a correlation of the microstructure of the welded joint with the welding parameters and mechanical properties. Preparation and examination of welded joints in metals is a well established technique, generally involving reflected light microscopy on polished and etched samples.

Reflected light microscopy can be used to examine polished thermoplastics, but in general this approach is only used for very hard materials, or those which are filled with glass fibres or carbon fibres. Soft thermoplastics, such as PE and PP, are usually examined by transmitted light microscopy on thin sections, and this method is routinely applied to welded joints at TWI. In addition, scanning electron microscopy (SEM) has been used on etched parent materials, and, although acid etching has been applied to welds, there have been no reports of the use of solvent etching.

The purpose of this article is to show that both transmitted light microscopy and scanning electron microscopy are important for examination of welded joints in thermoplastics and also to explain the nature of the etching mechanism: in order to do this, references are made to crystallinity and helical content results which are detailed in a TWI report. ^[1]

Sample details

The samples examined in this article were hot plate pipe welds made from Dupont Aldyl A (PEA) medium density PE (125mm diameter by 11mm wall thickness) and George Fischer Dekaprop Type 2 isotactic PP randomly copolymerised with ethylene (180mm diameter by l0mm wall thickness). The welding conditions are given in Table 1; the material manufacturer's recommended conditions were used for A1 and D1, but A10 and D9 were made using increased plate removal times. The aim was to give welds with different structures, resulting in the tensile properties which are also given in Table 1.

Table 1 Welding details and tensile results

Transmitted light microscopy

Since PE and PP are soft and flexible at room temperature, it was necessary to use a cryomicrotome and temperatures of -30°C(PP) and 45°C (PP) to produce sections around 10µm thick. These were mounted on glass sides using castor oil and a cover slip, and examined using polarised light and a first order red tinting plate. Photomicrographs ( Fig.1a, d and Fig.2a, d) show that the morphology of the welds and the parent pipe was spherulitic. The weld centreline was also revealed, and experience with PP welds has shown that the centreline of D1 is composed of smaller spherulites than the rest of the weld region, and that the centreline of D9 comprises a transcrystalline row structure.

Fig. 1. Polyethylene weld A1, x75 (approx for SEM):

a) Transmitted light micrograph (i - HAZ, ii - weld region);

b) Scanning electron micrograph, acid etched;

c) Scanning electron micrograph, solvent etched. Polyethylene weld A10, x30 (approx for SEM):

d) Transmitted light micrograph;

e) Scanning electron micrograph, acid etched;

f) Scanning electron micrograph, solvent etched

It was apparent that the width of the welds increased as the hot plate removal time increased. In the standard welds, A1 and D1, the weld was separated from the parent pipe by a heat affected zone (HAZ) comprising a region of deformed spherulites where partially molten material had been sheared by the displacement process. This region was narrow and ill defined in the non-standard welds A10 and D9.

SEM examination

To study the etching mechanism, Fourier transform infrared (FTIR) - microspectrometry and differential scanning calorimetry (DSC) were carried out on sections before and after acid etching. Analysis of toluene etched sections was not possible as these disintegrated on contact with the hot toluene, so instead, the material extracted from the polymers by the toluene was analysed by FTIR.

Acid etching

In both materials, acid attack took the form of pitting, and this was more severe in PP. The relative sizes of the weld regions revealed by acid etching, Fig.1b, e and Fig.2b, e correlated well with those revealed using transmitted light microscopy, for PE and PP, as did the HAZ in D1. However, a comparison of the acid etched and light micrographs, Fig.2d and e showed that in D9 the full extent of the HAZ was only revealed by etching.

Fig. 2. Polypropylene weld D1, x30 (approx for SEM):

a) Transmitted light micrograph (i - HAZ, ii - weld region);

b) Scanning electron micrograph, acid etched;

c) Scanning electron micrograph, solvent etched. Polypropylene weld D9, x30 (approx for SEM):

d) Transmitted light micrograph;

e) Scanning electron micrograph, acid etched;

f) Scanning electron micrograph, solvent etched.

In all the samples, preferential attack occurred at the weld/pipe interface, and in the PE welds this was greatest in A1 where the deformed region was also visible. There was also preferential attack at the weld centreline, and this was more apparent as the plate removal time increased.

The DSC results on the parent materials ( Table 2) showed that the measured crystallinity increased after etching, and this was also the case for the FTIR results ( Table 2) at the weld centreline in the PE weldment sections. However, in the PP weldment sections the helical content (which can be correlated with crystallinity) decreased slightly at the weld centreline, presumably as a result of removal of helices within the amorphous phase (since PP is a three component system with a non-helical amorphous component, a helical amorphous component, and a helical crystalline component). The crystalline orientation functions had increased in both the PE and PP weld. Thus acid etching was preferentially removing material from the amorphous phase, and also molecules orientated perpendicular to the centreline.

Table 2 FTIR and DSC results on samples before and after etching

Polyethylene is fairly resistant to oxidative attack because of the stability of the backbone chain, and therefore high temperatures and long times during etching are necessary. Acid etching of PE works by accelerated oxidative degradation, which causes chain cleavage in amorphous regions. In PP, chromic acid/sulphuric acid destroys amorphous material among and within spherulites. ^[4] PP is more readily attacked than PE because the C-C bond involving the methyl group is relatively easily oxidised, while these bonds are more easily accessible in amorphous regions. ^[5] It has been reported that, in PE, chromic acid/nitric acid attacks frozen in stresses, ^[6] while chromic acid attack at the weld centreline may be a result of incomplete adhesion ^[5] enabling the acid to come into greater 'contact' with the polymer.

Thus, the preferential attack at the weld/pipe interface in A1 was probably on deformed spherulites, and in A10 could have been caused by higher residual stresses and greater removal of amorphous regions, where the crystallinity was least. If the weld centreline attack were only due to crystallinity variations, A1 would have been affected most, as this had the lowest crystallinity, but A10 was the most severely attacked, presumably because of higher residual stresses and/or incomplete adhesion.

In the PP welds, it is possible that the HAZ was revealed because of residual stresses, and that the less severe attack in this region was a result of the higher helical content.

Preferential attack at the weld/HAZ interface probably stemmed from residual stresses rather than crystallinity changes, as the helical content did not drop to a minimum here. At the centreline, residual stresses and incomplete adhesion were probably causes of attack, especially as the orientation changed sharply. Attack on amorphous regions as a result of decreased cyrstallinity could be ruled out because the helical contents were at a maximum at the centreline.

Solvent etching

In the solvent etched samples, the visible form of attack was different from that of acid etching, as no pitting was observed. As with acid etching, the relative sizes of the weld regions ( Fig.1c, f and Fig.2c, f) correlated well with those revealed by light microscopy but both PP welds apparently had much wider HAZs. In the PE samples, the weld was removed to the greater depth than the pipe; deformation at the weld/pipe interface was visible in A1, and preferential attack occurred at the weld centreline, being severest in A10 ( Fig.1f). Higher SEM magnification of the parent material revealed the spherulitic structure. The PE spherulite shown in Fig.3a was about l0µm diameter, and the fibrils radiating from the centre were the lamellae: the concentric ring structure was associated with periodic twisting of the lamellae. Following welding with a short hot plate removal time, the spherulites at the weld centreline were smaller, about 5µm diameter, and the lamellae were more disordered ( Fig.3b). The spherulites were of similar size to the pits in the acid etched samples, and this suggests that the pits were caused by preferential attack at the interspherulitic boundaries, causing the spherulites to drop out. Work at the Paton Institute using a plasma etching technique has also revealed the presence of ringed spherulites in a plasma etched PE hot plate weld. ^[7]

Fig. 3. Scanning electron micrographs of spherulites in solvent etched polyethylene samples, x5000 (approx):

a) Parent pipe;

b) Weld centreline

In the PP welds, the morphology of the weld regions ( Fig.4a) was similar to that of the parent material ( Fig.4b), consisting of a smooth surface with occasional holes. The weld centreline was only just visible in D1, but became a ridge in D9 ( Fig.2c and f). The HAZs were also much wider than for the acid etched samples, and appeared to be raised above the surface of the weld and parent material. The surface of the HAZ was smooth, and lacked the holes which had been evident in the parent material. There also seemed to be a ridge in the HAZ in D9. In both samples, there was a region between the parent material and the raised outer edge of the HAZ, characterised by circular bumps on a flat surface ( Fig.4c) which must presumably be part of the HAZ.

Fig. 4. Scanning electron micrographs of solvent etched polypropylene samples, x300 (approx):

a) Weld centreline in D9;

b) Parent pipe;

c) Pipe just outside edge of HAZ in D1

The spherulitic structure in PE revealed by toluene etching was similar to that produced by ion and nitric acid etching, ^[8] the latter being reported to attack fold surfaces, tie molecules and other disordered material preferentially, leaving the crystalline material relatively unattacked. Molecular weight (MW) segregation can occur during segregation from the melt and low MW material is concentrated at the spherulitic boundaries and between the main lamellae within the spherulites. This low MW material may crystallise at a lower temperature than the main lamellae and can be removed with xylene at 100°C. ^[9] FTIR analysis ( Fig.5a) showed that PE and additives had been extracted by the toluene, and it is therefore probable that the toluene was removing low MW disordered material, and some high MW disordered material, including tie molecules. The recessed appearance of the welds was probably a result of greater removal of material in the amorphous interlamellar regions, since the crystallinity was lower in the welds than in the parent material.

Fig. 5. Comparisons of FTIR spectra from materials extracted by toluene, with pipe spectra:

a) PE;

b) PP

Dekaprop PP contains predominantly the α-crystalline form, with about 6% of the β -crystalline form. Spherulites containing the β-form are more readily attacked by hot solvents than α-form spherulites, and indeed hot toluene can remove β-form spherulites while leaving the α-spherulites virtually unaffected. The holes in the parent material and weld regions may be caused by removal of β -spherulites. A possible explanation for the lack of holes in the HAZ is that the β-form had been converted to the more stable α-form during relatively slow heating while welding.

FTIR analysis ( Fig.5b) showed that isotactic PP and atactic PP had been extracted by the toluene, which is in agreement with the observation that PP spherulites contain fibrils of lamellae comprising mainly highly ordered high MW isotactic material, and the interlamellar regions contain atactic low MW material, and some high MW isotactic material rejected here during crystallisation. ^[10] The spherulitic boundaries would probably be rich in disordered high MW isotactic material. Assuming that α-form spherulites are less affected by the toluene, then the FTIR results indicate that the toluene dissolved the interlamellar and disordered boundary material in the β-spherulites, causing these to be completely removed. The fact that the parent pipe and weld regions were recessed with respect to the HAZ and centreline suggested that the higher crystallinity and/or lack of β-spherulites in the HAZ and centreline restricted the access of toluene to the amorphous regions and thus relatively less material was removed.

In conclusion

Welds made in the two materials using standard welding conditions (2sec hot plate removal time) clearly contained similar morphological features, and differed from those made using non-standard conditions (20-70sec hot plate removal time). The latter lacked a well defined region of deformed spherulites between the weld and the parent pipe, and were much wider.

Examination of acid and solvent etched samples by SEM revealed information additional to that obtained by transmitted light microscopy, such as the apparent extent of the HAZ in the PP welds, and the different centreline morphologies of the PP welds made using standard and non-standard conditions. The mode of etching, for both materials, and for acid and solvent etching, was consistent with the removal of low and high MW disordered interlamellar material, and high MW disordered material at the spherulitic boundaries. In the PP material, it is probable that the main attack was on the β-form spherulites.

It is apparent from the information presented in this article, that welding changes the structure of a polymer, and hence can be expected to affect the mechanical properties. Optical and scanning electron microscopy are two important tools which should be used as part of any investigation involving structure/property relationships in polymer welds.

References