Effects of storage time and temperature on greenhouse gas samples in Exetainer vials with chlorobutyl septa caps

Graphical abstract


Method details
Measurement of greenhouse gas (GHG) flux occurs in a variety of ecosystems and landscapes, often with the goal of assessing carbon (C) and nitrogen (N) dynamics and associated ecosystem contributions to global warming potential. The static chamber method for collection of GHG samples is one of the most commonly used methods for determining GHG flux from agricultural soils over the past 30 years and requires samples be stored in a vessel before analysis [1]. In many cases, analysis of samples begins as soon as possible after collection [2,3]. However, situations may arise whereby analysis is delayed and samples must be stored (e.g., time lag due to transport of samples from field to laboratory; gas chromatograph immediately unavailable due to sample analysis backlog; gas chromatograph or autosampler is inoperable for maintenance, servicing, or troubleshooting). Studies and associated guidance can be found that established Exetainer1 vials with chlorobutyl septa (Labco Limited, Lampeter, UK) as the recommended vessel for GHG storage, along with other associated vial and sampling recommendations [2][3][4][5][6][7]. Unfortunately, guidance is lacking regarding appropriate storage temperature and time for GHG samples retained in septa-capped vials.
Segschneider et al. [4] observed only butyl rubber septa provided an effective seal for up to seven days among butyl, ethylene propylene diene monomer rubber, silicone, and Teflon faced butyl stoppers and septa. Duration of sample storage for less than 15 days in glass vials was ranked as "very good" by Rochette and Eriksen-Hamel [7] based on contamination and leakage potential. Some studies evaluated storage of gas samples using the amount of time vials maintained evacuation volume. Rochette and Bertrand [5] observed vials with double-wadded septa could be evacuated up to 63 days prior to use and evacuated Exetainer vials maintained evacuation volume well for 135 days. Use of both butyl rubber and silicone septa decreased the loss of vacuum with time (from 11% with only the rubber septa to 2% with both) and provided a more reliable seal [5]. Parkin and Venterea [2] found screw cap vials with butyl rubber septa held pressure better after 3 and 13 days compared to crimp top serum vials.
Most studies examining impacts of storage time on GHG samples focused on nitrous oxide (N 2 O), leaving information gaps regarding storage of carbon dioxide (CO 2 ) and methane (CH 4 ). No significant leakage for C isotope value was detected in Exetainer vials for up to 14 days [8]. Laughlin and Stevens [9] tested 12-mL Exetainer vials monthly for one year of storage of 15 N-labeled N 2 O and N 2 . Concentration of N 2 O did not decrease significantly for up to 8 weeks, but had decreased by 34% after a year of storage at room temperature in Exetainer vials [9]. Samples with N 2 O concentration of 10 mL L À1 and storage periods of 14 and 126 days had between 92 and 98% retention from vials using butyl rubber and double-wadded septa, respectively [5]. A butyl rubber septa-sealed vial with N 2 O concentration of 1 mL L À1 retained 90% of the gas for storage period of 365 days [3]. Another suggested alternative for accurate N 2 O concentration determination was storage and analysis of standards along with samples [9]. Only one study, evaluated the effects of storage temperature, observing storage at 5 and 20 C had no effect on N 2 O concentration in Exetainer vials [9].
Due to the lack of studies documenting effects of storage time and temperature on GHG samples available in peer-reviewed literature or guidance documents, the objective of this study was to determine effects of storage time and temperature on carbon dioxide (CO 2 ), methane (CH 4 ), and nitrous oxide (N 2 O) in septa-capped vials. This will be the first study to systematically evaluate a broad range of both storage temperatures and times, will improve reliability of GHG values reported in the literature, and provide guidance for appropriate storage temperature and time for GHG samples. We hypothesized colder storage temperatures (walk-in cooler and freezer) would provide greater storage stability of GHG samples than at room temperature or 25 C (incubator) and storage of GHG samples longer than 28 days would result in unacceptable amounts of leakage of measured gases.

Sample preparation and storage
Four storage locations differing in ambient temperature were used in the study: laboratory workbench at ambient room temperature, incubator set to 25 C, walk-in cooler set to 4 C, and freezer at À10 C. Storage times were 3, 7, 14, 28, and 84 days. To reduce effects of light, vials for each storage temperature were placed in corrugated cardboard boxes. An Omega SC-GG-K-30-36-PP thermocouple (Omega Engineering, Inc., Stamford, CT) was positioned in each box such that storage temperature measurements could be monitored periodically using a VWR traceable dual laser infrared thermometer (VWR International LLC, Radnor, PA). Mean AE standard deviation temperatures over the 84 day study were 20.4 AE 0.8 C for room temperature, 23.5 AE 0.4 C for incubator, 5.2 AE 0.5 C for walk-in cooler, and À11.3 AE 1.2 C for freezer ( Fig. 1).
Before placement in their respective storage locations, Exetainer1 vials sealed with chlorobutyl rubber septa (Labco Limited, Lampeter, UK) were prepared by purging with ultrapure helium (Praxair Distribution, Inc., Danbury, CT) for 10 s and evacuated for 15 s. After preparation, one set of vials were injected with 15 mL of "low" concentration CO 2 , CH 4 , and N 2 O blended standard and another set of vials were similarly injected with a "high" concentration blended standard. Standards were certified by the National Oceanic and Atmospheric Administration Global Monitoring Division (Boulder, CO) on 22 June 2015 with analysis certificate numbers CB11245-A and CB11061-A for "low" and "high" standards, respectively. Concentrations in the low standard were 380 mL L À1 CO 2 , 1.790 mL L À1 CH 4 , and 313 nL L À1 N 2 O, while concentrations in the high standard were 1000 mL L À1 CO 2 , 2.500 mL L À1 CH 4 , and 360 nL L À1 N 2 O.
For each storage temperature and time treatment combination, four replicate vials of low and high standards were prepared. After the appropriate storage time had passed, vials were placed on a laboratory workbench and allowed to set at room temperature for 30 min before analysis. At this time, vial temperatures had equilibrated to approximately 21.0 C as measured by a VWR traceable dual laser infrared thermometer (VWR International LLC, Radnor, PA).

Sample analysis
Concentrations of CO 2 , CH 4 , and N 2 O were determined using gas chromatography on a Varian CP-3800 Gas Chromatograph with CombiPAL autosampler and Varian Star Workstation Version 6.41 software (Agilent Technologies, Santa Clara, CA). Samples were injected and split into two sample loops. One sample loop used ultra-high purity helium (Praxair Distribution, Inc., Danbury, CT) to carry 1 mL of sample through a column and to a thermal conductivity detector for CO 2 and flame ionization detector for CH 4 . Hydrocarbon-free air and ultra-high purity hydrogen (Praxair Distribution, Inc) were used for combustion in the flame ionization detector. The second sample loop used ultra-high purity helium and a 5% methane, 95% argon blend (Praxair Distribution, Inc., Danbury, CT) to carry 0.5 mL of sample through a column to a 63 Ni electron capture detector for N 2 O. The gas chromatograph was calibrated with three standard blends of CO 2 (350, 424, 1998 mL L À1 ), CH 4   For quality assurance and quality control purposes, each run of samples included hydrocarbon-free air checks every ten samples (n = 5 for each run). Mean (AEstandard deviation) percent recoveries of expected concentrations for these quality control air checks were 94.4 AE 4.8% for CO 2 ,103 AE 7% for CH 4 , and 95.3 AE 8.1% for N 2 O. Furthermore, six vials each of the low and high standards used in the study were filled the day of each run. Mean (AEstandard deviation) percent recoveries of expected concentrations for the low standard were 99.3 AE 4.8% for CO 2 , 98.7 AE 6.0% for CH 4 , and 109.9 AE 3.4% for N 2 O. Mean (AEstandard deviation) percent recoveries of expected concentrations for the high standard were 90.7 AE 6.0% for CO 2 , 97.9 AE 2.1% for CH 4 , and 99.6 AE 2.5% for N 2 O.

Statistical analyses
All statistical analyses were conducted in R 3.3.1 [10] using an alpha level of 0.05 to indicate statistical significance. Analysis of variance (ANOVA) was used to test for statistical differences between treatments. Fixed effects were storage temperature, time, and storage temperature x time interaction. Tukey honest significant difference (HSD) post-hoc tests were used if ANOVA tests indicated treatment effects. Normality and variance homogeneity assumptions were met as indicated by Shapiro-Wilk and Bartlett's tests, respectively. ANOVAs and Tukey HSD tests were conducted individually for CO 2 , CH 4 , and N 2 O concentrations for the low and high standards.

Results
Storage temperature significantly affected CH 4 concentrations of the low standard (F 3, 57 = 8.58; P < 0.001) (Fig. 2) and CO 2 concentrations of the high standard (F 3, 60 = 6.11; P = 0.001) (Fig. 3). Similarly, a significant storage temperature x storage time interaction was observed for CH 4 low standard concentrations (F 12, 57 = 3.03; P = 0.002) and CO 2 high standard concentrations (F 12, 60 = 5.04; P < 0.001). Conversely, storage time significantly influenced CO 2 , CH 4 , and N 2 O concentrations for both low and high standards (All P < 0.001) (Figs. 2 and 3). This time effect primarily resulted from significant differences between storage for 84 days and all other storage times. Concentrations of GHGs measured were generally within 5% of certified concentrations regardless of storage temperature and time with one exception-N 2 O concentrations of the low standard were consistently 30-40 nL L À1 greater than the certified concentration of 313 nL L À1 (Fig. 2).

Brief discussion
Due to the importance of obtaining accurate and precise concentration data and extensive use of static chamber methodology for calculation off GHG fluxes [1], this study examined effects of storage temperature and time on CO 2 , CH 4 , and N 2 O concentrations in septa-capped vials. Significant effects of both storage time and temperature were observed. Although significant differences by storage temperature were observed for CH 4 low standard and CO 2 high standard, differences were small and likely the result of statistical power to detect differences between treatments and interaction effects between storage time and temperature. The interactive effect of time was apparent in both circumstances such that 84 days of storage resulted in consistently higher than expected CH 4 low concentrations and lower than expected CO 2 high concentrations (Figs. 2 and 3). Other possible mechanisms could include leakage or contamination of GHG samples, both of which increase with storage duration [5].
In support of the hypothesis, storage time affected concentrations of CO 2 , CH 4 , and N 2 O in both low and high standards, with the most apparent effect after 84 days of storage (Figs. 2 and 3). On average, CO 2 concentrations for both low and high standards after 84 days were less than concentrations for other storage times and 5% and 10% lower than certified concentrations of the standards, respectively. Concentrations of CH 4 in both low and high standards were generally greater than concentrations for other storage times and at least 10% greater than the certified concentrations. Concentrations of CH 4 greater than certified concentrations in low and high standards are difficult to explain, but could be due to sample contamination over the course of 84 days of storage.
However, CO 2 concentrations less than certified in both standards after 84 days can be explained with literature regarding losses of N 2 O during storage. Possible explanations offered by Laughlin and Stevens [9] for greater N 2 O percent loss than for N 2 included: a larger concentration gradient between Exetainers and the atmosphere; diffusion through septa; and adsorption to septum or glass. Since losses were observed for CO 2 , but not CH 4 or N 2 O, the most likely explanation is leakage through or around septa because leakage rate from vials is proportional to the concentration gradient between the vials and ambient according to diffusion theory [5,9]. The certified CO 2 concentration in the low standard (380 mL L À1 ) and high standard (1000 mL L À1 ) were at least one and two orders of magnitude greater, respectively, than the CH 4 high (2.5 mL L À1 ) and N 2 O high (0.360 mL L À1 ) standard concentrations. Furthermore, loss of CO 2 after 84 days of storage is analogous to increased probability of leaks at the collar/soil or collar/chamber interface with increasing deployment duration and soil gas diffusivity [1]. It is possible that leakage occurred around the septa of vials as a dent was observed in septa after 84 days, but not after any other storage time (Fig. 4). Although the severity of denting in septa was variable and not associated with any particular treatment, dents were observed in all septa after 84 days of storage, regardless of storage temperature.

Conclusions
Based on results of this study, it is recommended that GHG samples in Exetainer vials with chlorobutyl septa be stored no longer than 28 days before analysis for CO 2 , CH 4 , and N 2 O. Though it may be possible to store samples up to or longer than 84 days, decreases in CO 2 concentrations would be expected under similar conditions evaluated in this study. All storage temperatures evaluated in this study were found to be appropriate for storage of GHG samples. Accordingly, while a temperaturecontrolled environment with minimal light is ideal, a laboratory room with minimal temperature fluctuations is sufficient for GHG sample storage. These results provide scientifically supported guidance for storage of GHG samples. If the need arises for sample storage due to sample transport, analysis backlog, or inoperable instrumentation, then this study provides the scientific community confidence that samples can be properly stored for a period of 28 days without impacting data integrity. Furthermore, following guidance provided in this study will improve reliability of GHG values reported in the literature, ensure consistency, and help editors and reviewers assess data quality if samples were stored. Future studies should be conducted to further refine guidance on storage temperature and time for GHG samples.