Declining sulphur hexafluoride permeability of polytetrafluoroethylene membranes causes overestimation of calculated ruminant methane emissions using the tracer technique

Predictable release of sulphur hexafluoride (SF6) from permeation tubes is critical for determination of enteric methane emissions from ruminants using the calibrated tracer technique. Experiments comparing respiration chambers and the SF6 technique indicate that prolonged deployment of SF6 tubes causes overestimation of methane emissions. We report three studies investigating release rate of SF6 from permeation tubes. Experiment 1 investigated if SF6 was released by routes other than the polytetrafluoroethylene (PTFE) membrane. Replacement of the membrane with an impervious disk prevented SF6 release, confirming that SF6 release occurs only through the PTFE membrane. Experiment 2 investigated the effect of frit exposure to the rumen environment and reticulo-rumen residence duration upon SF6 release rate. Three treatments were applied: (i) tubes with exposed frits (Control), (ii) tubes with exposed frits previously incubated for 152 d in the reticulo-rumen (Exposed-frit) and (iii) tubes incorporating an external membrane to prevent frit contamination during 152 d incubation in the reticulo-rumen (Protected-frit). These tubes were then used to determine methane yield. Tubes of each treatment were also retained in an incubator to determine SF6 release rate concurrent to the in vivo experiment. Decline of SF6 release was not related to frit exposure to the reticulo-rumen environment. Methane yield determined using Exposed-frit and Protected-frit tubes was 21% greater than that estimated with Control tubes (P<0.01). Concurrent SF6 release, from Exposed-frit and Protected-frit tubes retained within an incubator was similarly reduced by 22% relative to Control tubes, indicating that decline of SF6 release rate is a function of time elapsed post-calibration, unrelated to tube reticulo-rumen residence duration per se. Experiment 3 was conducted to determine the cause of the in vivo decline in SF6 release. Three pre-treatments were applied to PTFE membranes; untreated (Control-PTFE), 165 d rumen exposure (Rumen-PTFE) and 165 d SF6 exposure (SF6-PTFE). Pre-treated membranes were used to construct new permeation tubes. Release rate from SF6-PTFE tubes was 36% less than from control tubes and 35% less than from Rumen-PTFE tubes (P<0.001). Release rate from Rumen-PTFE tubes did not differ from control tubes (P>0.05). Post-calibration decline in SF6 release rate from permeation tubes results from exposure of PTFE membranes to SF6 rather than reticulo-rumen exposure. We conclude that ‘dry’ incubation of permeation tubes in a laboratory to determine the rate of SF6 release is valid. However, post-calibration decline in SF6 release must be accounted for to prevent over-estimation of methane emissions.