Dangers ... what dangers?

TWI Bulletin, March/April 2008

Just what products evolve during hot gas welding of fluoropolymers

The health and safety ramifications of joining polymers can be more demanding than we think

Mike Troughton is Technology Manager for Plastics at TWI. He joined TWI in 1993 with a PhD in polymer physics, after spending seven years at BP Chemicals. His main areas of expertise are the welding, mechanical testing and inspection of welds in thermoplastics as well as the certification of plastics welding personnel. He is also the chairman of a number of standards committees, both in the UK and world-wide.

The fabrication of structures made from fluoropolymers such as polytetrafluoroethylene (PTFE), polyvinylidenefluoride (PVDF), ethylene chlorotrifluoroethylene (ECTFE), perfluoroalkoxy (PFA) and fluorinated ethylene propylene (FEP) is becoming more widespread in the chemical, semi-conductor, paper and pharmaceutical industries. This is mainly due to the excellent chemical resistance of these materials, combined with good mechanical and physical properties. However, instances of polymer fume fever have been recorded when welding these materials and, for this reason, the Health & Safety Executive asked TWI to participate in a study to investigate the products evolved when welding fluoropolymers as Scott Andrews and Mike Troughton explain.

Hot gas welding is a technique used for fabricating and repairing structures such as tanks and pipelines. Of all the plastics welding techniques, hot gas welding gives the most cause for concern from the point of view of health and safety. Firstly, this is because it is a manual technique, which means that the operator is in close proximity to the working area, so that typically the face is within 500mm of the weld region. This is often made worse by the frequently encountered situation of working in enclosed spaces with the attendant risk of build up of fume and pollutants. Secondly, this technique is widely used (there are over 40 000 units in use in the UK), and often by totally untrained personnel. In fact, it is only within the past decade that training and certification courses have become available in the UK.

From information provided by the fluoropolymer manufacturers, the products generated on decomposition of these materials include hydrogen fluoride, hydrofluoric acid, fluoro-olefins, chlorinated hydrocarbons and carbonyl fluoride. However, it is not known in what quantities these products are generated during welding, or indeed, whether they are generated at all. It is, however, well known in the fabrication industry that inhaling the fumes during hot gas welding of fluoropolymers can cause influenza-like symptoms, known as Teflon fever or polymer fume fever.

This article summarises the findings of a research project that was performed by the Health and Safety Laboratory (HSL) and TWI, on behalf of the Health and Safety Executive (HSE), to identify and measure the amounts of products evolved and the exposure to the operator, during hot gas welding of four fluoropolymers: PTFE, ECTFE, FEP and PFA. PVDF was not included in this study because an initial study had shown that the welding of this material causes a minimal health risk compared to the other fluoropolymers.

Experimental

Materials and Equipment

The materials that were used for the trials are given in Table 1.

Table 1 Details of fluoropolymer materials

A Triple Plus+ gas monitor, manufactured by Crowcon Detection Instruments Ltd and supplied by ESS Ltd, was used to determine whether hydrogen fluoride was present in the laboratory. Local exhaust ventilation (LEV) was provided by a Nederman portable extractor which had an output hose inserted into a laboratory fume cupboard. In order to provide emergency extraction in the event of a power cut, a portable extractor, supplied by Enviro-Vac Sales & Service Ltd, was connected to a portable generator.

The hot gas welding trials were carried out using a Leister PID hot gas gun connected to either a Leister blower or, in the case of ECTFE, also to a nitrogen cylinder. A speed welding nozzle was used in all cases.

Preparation of laboratory

To prevent any fume escaping from the laboratory into the main building, the ceiling was sealed using polyethylene film.

Hot gas welding trials

The experimental set-up for the hot gas welding trials is shown in Fig.1 and Fig.2.

Fume sampling stations were positioned at three locations in the laboratory ( Fig.1); one in front of the welding operation (Station A), one in the breathing zone of the welder (Station B) and one behind the welder (Station C). In addition, a number of ultrafine particle detectors were also positioned in the laboratory ( Fig.2): one in front of the welding operation (Detector A), one attached to a probe positioned in the breathing zone of the welder (Detector B) and one behind the welder (Detector C, not shown). Close-ups of the welding operation are shown in Figures 4 and 5.

The welder wore Tyvek pro-tech coveralls and positive demand full-face mask airline breathing apparatus, which had an assigned protection factor of 2000. A summary of the welding trials performed is given in Table 2.

Table 2 Results of hot gas welding trials

Measurement of airborne material

During the hot gas welding trials, air sampling was performed to establish the airborne concentrations of various substances that could potentially be evolved from the hot fluoropolymer.

Volatile organic compounds (VOCs)

VOCs were sampled using pumped sampling onto sorbent tubes. Since there was the potential for a variety of VOCs to be evolved from hot gas welding of fluoropolymers, three different sorbents were used. All VOC samples were analysed by thermal desorption - gas chromatography with mass spectrometric detection.

Total inhalable particulate (TIP)

TIP was measured using pumped sampling onto PTFE membrane filters. The filters were mounted in stainless steel cassettes, contained in Institute of Occupational Medicine (IOM) sampling heads. TIP was determined using microgravimetry. The filters were then solvent desorbed (into acetone) and the resulting solutions were analysed by gas chromatography with mass spectrometric detection. This was done to check for the presence of higher boiling point organic material (semi volatile organic compounds) which may not have been detected from the VOC samples.

Aldehydes

Aldehydes were measured using pumped sampling onto glass fibre filters impregnated with a derivatising agent (dinitrophenylhydrazine, DNPH). After sampling, these filters were desorbed into acetonitrile and analysed for aldehydes using high pressure liquid chromatography (HPLC) with ultraviolet spectrophotometric detection.

Carbonyl halides

The major carbonyl halide of interest in this work was carbonyl fluoride (COF ₂). However, since the ECTFE molecule contains chlorine as well as fluorine, there was the potential for this material also to generate carbonyl chloride (phosgene) upon decomposition. Samples for these two compounds were taken using pumped sampling onto XAD-2 (a porous polymer sorbent) impregnated with a derivatising agent (2-(hydroxymethyl) piperidine, HMP). After sampling, the sorbent material was desorbed into toluene. Initially these samples were analysed using gas chromatography with a nitrogen-phosphorous detector. As the project progressed the method was developed to use gas chromatography with mass spectrometric detection.

Hydrogen fluoride

This was measured by sampling onto water washed silica gel. The samples were desorbed into aqueous carbonate/bicarbonate buffer solution, which was analysed for anions using ion chromatography with suppressed conductimetric detection.

Ultrafine particles

Ultrafines were measured using a scanning mobility particle sizer configured to monitor for particles in the size range 10 to 535 nm. This instrument gives information regarding the size distribution of the particles it measures. However, it soon became apparent that the response time of this instrument (two minutes to scan over the size range) was too slow to be of use in a dynamic welding situation where airborne concentrations of particles fluctuate rapidly over time. Hence, ultrafine measurements were also made using a TSI condensation particle counter (measuring particles from 0.01 µm to 1 µm) and 2 TSI Ptrak model 8525 particle detectors (measuring particles from 0.02 µm to 1 µm). These instruments measure airborne particle concentrations in real time and log the data so that it can be downloaded later.

Results and discussion

Fume analysis

PTFE

Neither carbonyl fluoride nor hydrogen fluoride was detected at any stage during the welding trials with PTFE/PFA. Large amounts of ultrafine particles were detected during the welding trials, with a significant amount of very small particles (10 to 20nm diameter) being generated. These were generated in bursts, although it is not clear what caused these bursts.

ECTFE

ECTFE was welded using both air and nitrogen as the hot gas. Hydrogen fluoride was not detected under any welding conditions. When welded in air at the maximum temperature, carbonyl fluoride was detected in the static sample positioned close to the welding operation, but not in any other sample, including samples taken on the welding operator. Neither carbonyl fluoride nor hydrogen fluoride was detected during any of the welding trials using nitrogen gas. Large amounts of ultrafine particles were detected during the welding trials. The highest concentrations were measured around the breathing zone of the operator, with maximum levels being detected when the LEV system was not switched on. The ultrafine results obtained when welding under nitrogen were very similar to those obtained when welding in air.

FEP

Hydrogen fluoride was not detected from FEP under any welding conditions. Carbonyl fluoride was detected at concentrations between 0.3 and 0.6ppm in static samples taken close to the welding operation during welding at the maximum temperature. However, concentrations in the operator's breathing zone were below the limit of detection (0.3 ppm). Carbonyl fluoride was not detected in any samples taken whilst welding at the recommended temperature. Large amounts of ultrafine particles were detected from FEP during the welding trials, with a significant amount of very small particles (10 to 20nm diameter) being generated. Again, these were generated in bursts.

PFA

During the welding trials, neither hydrogen fluoride nor carbonyl fluoride was detected from PFA under any conditions. However the detection of carbonyl fluoride during the heating trials indicates the potential for generation of this substance when PFA is overheated. Large amounts of ultrafine particles were detected from PFA during previous welding trials, with a significant amount of very small particles (10 to 20nm diameter) being generated.

Conclusions and recommendations

The results of this study showed that:

• Significant levels of carbonyl fluoride can be generated during hot gas welding of PFA and FEP. Carbonyl fluoride can also be generated from ECTFE, but only when welding in air. If this material is welded in nitrogen, as recommended, then carbonyl fluoride is not generated.
• Significant levels of ultrafine particles are generated by the hot gas gun itself during the welding process.
• Ultrafine particles also appear to be generated from PFA and ECTFE during hot gas welding.
• Only trace levels of relatively non toxic compounds were detected during the hot gas welding of PTFE.

Air monitoring for carbonyl fluoride and/or hydrogen fluoride may be useful when investigating cases of polymer fume fever. However, it should be remembered that, where ultrafine particles are also present, these chemical agents may exhibit toxic effects at much lower levels than where they are present alone. Urinary fluoride monitoring may be useful when investigating polymer fume fever. However, this would require further investigation.

Fluoropolymers should be hot gas welded at the lowest possible temperature to reduce the potential for causing polymer fume fever to operators.

If temperature control is not sufficient to prevent episodes of polymer fume fever, a good standard of local exhaust ventilation (LEV) should be used. LEV systems should be designed to enclose the welding process as much as possible. If portable, 'flexible arm' type LEV systems are used; the operators should receive adequate training to allow them to be used effectively.

Correctly used, P3 rated respiratory protective equipment can offer good protection against the ultrafine particulate material which is implicated in the cause of polymer fume fever. As always, however, RPE should only be used when engineering controls alone do not offer adequate protection.

Any industrial process which involves gross overheating of fluoropolymers, such as laser cutting, should be the subject of a rigorous risk assessment.