INTRODUCTION
The global atmospheric methane (CH4) concentration has more than doubled since the 19th century and is increasing even more rapidly since 2007 (Saunois et al. Reference Saunois, Stavert, Poulter, Bousquet, Canadell and Jackson2020). After carbon dioxide (CO2), CH4 is considered to be the second most important greenhouse gas with a 28–36 times larger greenhouse gas warming potential compared to CO2. When looking at its impact over 100 years, 1 tonne of CH4 is equivalent to 28–36 tonnes of CO2 (IPCC AR5; Myhre et al. Reference Myhre, Shindell, Bréon, Collins, Fuglestvedt, Huang, Koch, Lamarque, Lee and Mendoza2013.). Therefore, CH4 is more powerful at trapping heat in the atmosphere than CO2 on a per molecule basis and thus has an important influence on the rate of climate change (Saunois et al. Reference Saunois, Jackson, Bousquet, Poulter and Canadell2016).
CH4 in the atmosphere originates from many different sources, which are difficult to identify and quantify. Presently, the dominant global sources of anthropogenic CH4 emissions are agriculture, fossil fuel production and combustion (Kirschke et al. Reference Kirschke, Bousquet, Ciais, Saunois, Canadell and Dlugokencky2013; Turner et al. Reference Turner, Frankenberg, Wennberg and Jacob2017; Maasakkers et al. Reference Maasakkers, Jacob, Sulprizio, Scarpelli, Nesser and Sheng2019). While agricultural CH4 emissions derive mainly from ruminant animals, organic matter degradation through methanogens is responsible for wetland CH4 emissions. Whereas the extraction, storage, and transportation of oil, natural gas, and coal release CH4 generated by thermogenic (geological) processes. These CH4 sources have characteristic 13C and 14C isotopic signatures. CH4 produced by methanogenesis of fresh organic matter is depleted in its 13C content and enriched in 14C, whereas thermogenic degradation of organic matter in sedimentary rocks generates CH4 that contains more 13C compared to biogenic sources and no 14C.
To distinguish between C sources, numerous studies have applied carbon isotopic analysis of CH4 emissions over the last years. While many of these investigations use 13C analysis (Lowry et al. Reference Lowry, Holmes, Rata, O’Brien and Nisbet2001; Fisher et al. Reference Fisher, Sriskantharajah, Lowry, Lanoisellé, Fowler and James2011; Townsend-Small et al. Reference Townsend-Small, Botner, Jimenez, Schroeder, Blake and Meinardi2016; Lopez et al. Reference Lopez, Sherwood, Dlugokencky, Kessler, Giroux and Worthy2017; Maazallahi et al. Reference Maazallahi, Fernandez, Menoud, Zavala-Araiza, Weller and Schwietzke2020), other studies applied 14C analysis to distinguish CH4 sources in the atmosphere (Graven et al. Reference Graven, Hocking and Zazzeri2019; Zazzeri et al. Reference Zazzeri, Xu and Graven2021), in peatland (Garnett et al. Reference Garnett, Gulliver and Billett2016; Cooper et al. Reference Cooper, Estop-Aragonés, Fisher, Thierry, Garnett and Charman2017) or in aquatic systems (Pohlman et al. Reference Pohlman, Kaneko, Heuer, Coffin and Whiticar2009; Joung et al. Reference Joung, Leonte and Kessler2019) or the combination of both isotopes has been applied (Gonzalez Moguel et al. Reference Gonzalez Moguel, Vogel, Ars, Schaefer, Turnbull and Douglas2022).
Unfortunately, in contrast to 13C, the direct 14C analysis of CH4 is not possible and requires an elaborate pre-treatment routine during which the CH4 is purified and subsequently oxidized to CO2 for further analysis. Gas samples taken from the environment or from laboratory incubations contain other C carrying gases that need to be separated from CH4 by utilizing carrier gases (noble gas or synthetic air, Pohlman et al. Reference Pohlman, Kaneko, Heuer, Coffin and Whiticar2009) or pressure differences (Garnett et al. Reference Garnett, Murray, Gulliver and Ascough2019) by which the sample is moved through chemical adsorbents (Garnett et al. Reference Garnett, Murray, Gulliver and Ascough2019; Zazzeri et al. Reference Zazzeri, Xu and Graven2021; Gonzalez Moguel et al. Reference Gonzalez Moguel, Vogel, Ars, Schaefer, Turnbull and Douglas2022) and cryogenic traps (Petrenko et al. Reference Petrenko, Smith, Brailsford, Riedel, Hua and Lowe2008; Pack et al. Reference Pack, Xu, Lupascu, Kessler and Czimczik2015) or a combination of those (Garnett et al. Reference Garnett, Murray, Gulliver and Ascough2019; Gonzalez Moguel et al. Reference Gonzalez Moguel, Vogel, Ars, Schaefer, Turnbull and Douglas2022). The approach presented in this study utilizes a combination of methods taken from recent studies including a set of cryogenic traps, an oxidation furnace and synthetic air as carrier gas operating below ambient pressure.
Environmental gas samples most prominently contain CO2 and water (H2O) but also carbon monoxide (CO) and other hydrocarbon gases in smaller quantities. As these gases are potentially emitted from different organic or inorganic sources having different isotopic compositions, it is necessary to eliminate them (Petrenko et al. Reference Petrenko, Smith, Brailsford, Riedel, Hua and Lowe2008). H2O is removed using dry ice slurries (DI) or adsorbents (Garnett et al. Reference Garnett, Murray, Gulliver and Ascough2019; Zazzeri et al. Reference Zazzeri, Xu and Graven2021), while CO2 is trapped in liquid nitrogen (LN; Petrenko et al. Reference Petrenko, Smith, Brailsford, Riedel, Hua and Lowe2008; Pack et al. Reference Pack, Xu, Lupascu, Kessler and Czimczik2015) or adsorbed through molecular sieves containing zeolite (Garnett et al. Reference Garnett, Murray, Gulliver and Ascough2019; Zazzeri et al. Reference Zazzeri, Xu and Graven2021). Depending on the origin and composition of the gas sample, the removal of CO may be essential because it is oxidized in the furnace along with the CH4, which may bias the result of the CH4 (Pack et al. Reference Pack, Xu, Lupascu, Kessler and Czimczik2015). Ultimately, the purified CH4 is oxidized in a furnace with an oxygen donor and converted to CO2 and H2O. The choice of the catalyst varies from laboratory to laboratory. Most prominently, cupric oxide (CuO) filled columns (Kessler and Reeburgh Reference Kessler and Reeburgh2005; Pack et al. Reference Pack, Xu, Lupascu, Kessler and Czimczik2015), platinized alumina beads contained in quartz glass tubes (Garnett et al. Reference Garnett, Murray, Gulliver and Ascough2019; Gonzalez Moguel et al. Reference Gonzalez Moguel, Vogel, Ars, Schaefer, Turnbull and Douglas2022) or platinized quartz wool are used as catalysts (Petrenko et al. Reference Petrenko, Smith, Brailsford, Riedel, Hua and Lowe2008; Sparrow and Kessler Reference Sparrow and Kessler2017; Zazzeri et al. Reference Zazzeri, Xu and Graven2021). After the reaction, the CH4-derived CO2 can be quantified and prepared for 14C analysis via gas injection or graphitization, i.e., conversion to elemental C.
Here we present a flow-through vacuum system coupled with cryogenic traps for the purification and conversion of CH4 samples to CO2 for the purpose of radiocarbon analysis at the CologneAMS facility. The system operates at low pressures (30 mbar) and is by design applicable for the processing of gaseous samples of different origin via syringes or capillaries that can be attached to the system. The system can handle large and small samples, but it was mainly designed for 15–50 µg C, that is the typical sample size for CO2 analysis using the gas ion source of our AMS at Cologne University.
METHODS
System Overview
The CH4 oxidation system consists of (1) the mixing unit, where samples or standard gas mixtures are prepared and injected via the gas supplies, (2) the CH4 purification unit consisting of furnaces and cryotraps, and (3) the sealing unit where the sample can be quantified and CO2 aliquots corresponding to 10–50 µg C are sealed off for subsequent AMS measurement (Figure 1).
Figure 1 Schematic overview of the oxidation rig consisting of the mixing (green), purification (red) and sealing (blue) units. (Please see online version for color figures.)
In the mixing unit, each connection is individually maintainable by a respective valve allowing the injection and mixing of defined volumes of gas mixtures. CH4 and CO2 are supplied either in pressurized 15 L bottles that are directly connected to the mixing unit with a pressure regulator and regular ferrule fittings (Swagelok®, USA), from a self-assembled stainless-steel cylinder (304L-HDF4-1GAL, Swagelok®, USA) that is equipped with a pressure gauge (-1 – 3 bar, PGI-63B-BC3-LAQX, Swagelok®, USA) and a septum port (SS-4-TA-1-4STKZ, Swagelok®, USA) or the gases are taken from sealed serum bottles. Similar to the sample cylinder, a septum port (S, Figure 1) is used for syringe injection of gases with a 10 mL gas tight syringe (Gastight 1010 LTN, Hamilton®, USA). In addition, a 2.5 mL stainless-steel vacuum syringe (SV, Figure 1, KDScientific Inc., USA) with a screw thread for standard vacuum fittings was installed in order to mix bottled gas with gases injected with a syringe. The valves are permanently regulating the gas flow so that a gas stream of 60 mL min–1 cannot be exceeded. The flow is monitored by a flow meter (F, Figure 1, FMA-1606, Omega™ Engineering Inc., USA) that is installed prior to the purification unit of the oxidation rig.
Gases are flushed from the mixing unit through a U-tube submerged in a Dewar flask filled with dry ice (DI) trapping moisture, while a second U-tube in a Dewar flask filled with liquid nitrogen (LN) traps CO2 from the gas stream. After removal of moisture and CO2, the gas flows through a furnace via a CuO-filled quartz glass tube that is heated to 290°C. Thereby CO is removed from the gas phase and oxidized to CO2, which is then trapped in another LN trap installed behind the furnace. Subsequently, the gas stream enters a second furnace (Carbolite Gero GmbH & Co. KG) with a CuO-filled quartz glass tube that is heated at 1000°C in order to oxidize CH4 to CO2 (red frame, Figure 1). In a final set of cryogenic traps, H2O and CO2 are collected within U-tubes submerged in DI and LN, respectively. Any non-reactive and non-condensable gases left in the gas stream are now evacuated through the vacuum pump and the isolated CO2 is transferred to the sealing unit.
The sealing unit (blue frame, Figure 1) is constructed like a standard vacuum rig with pressure transducers, calibrated volumes, and ultra-torr fittings equipped with glass ampoules (4 or 6 mm OD), that are used to seal CO2 samples (Wotte et al. Reference Wotte, Wordell-Dietrich, Wacker, Don and Rethemeyer2017). The sample CO2 is first trapped in a LN-filled Dewar flask and then transferred into a calibrated volume for quantification. Defined amounts of CO2 can be transferred via a second calibrated volume into glass ampoules with the help of LN. Non-condensable gases are removed by briefly opening the valves (V) prior to sealing the ampoules with a hand-torch.
Sample Preparation
Prior to sample processing, the tube furnace is heated up to 1000°C for 30 min to ensure temperature equilibration. The mixing and purification units of the system are flushed three times with argon (Ar) for 5 min (40 mL min–1) and evacuated afterwards below 10–3 mbar. This is also done between samples to maintain a low line blank and prevent memory effects (Pack et al. Reference Pack, Xu, Lupascu, Kessler and Czimczik2015). The synthetic air is set to a flow rate of 10 mL min–1 during the entire sample preparation procedure. The pressure in the system is < 20 mbar throughout the sample processing, which still ensures the removal of other hydrocarbon gases from the gas phase into the LN trap and, more importantly, prevents the condensation of oxygen, which might pose a safety hazard (Sparrow and Kessler Reference Sparrow and Kessler2017). The sealing part of the system is not flushed and kept at a constant vacuum below 10–7 mbar to maintain cleanliness.
Depending on the type of sample or standard, different injection and mixing procedures are possible at the mixing unit of the oxidation rig that will be addressed individually in more detail below. Injected gas mixtures are transported via the synthetic air stream through the furnaces and cryogenic traps in the purification unit into the calibrated volumes and glass ampoules at the sealing unit.
Preparation of Standards
Carbon Free Standards
Different types of standards were prepared to test the overall cleanliness of the setup as well as of the trapping efficiency and the CH4 to CO2 conversion rate of the oxidation rig. The overall C blank of the line was tested by C-free N2 gas with the system (99.999% purity, Linde GmbH, Germany). The N2 was filled via a Teflon tube attached to sterile injection needles (Sterican® size 18, B. Braun SE, Germany) into 100 mL serum bottles sealed with crimped butyl rubber stoppers. Ambient air was released from the bottle via a second injection needle. The serum bottles were previously washed with Milli-Q water (MQ; Millipore, USA), combusted at 450°C for 3 hr and flushed at ambient pressure with N2 with at least three bottle volumes to ensure that no ambient air remained inside. A 10 mL gas tight syringe, which was pre-cleaned with dichloromethane (DCM; SupraSolv®, MERCK KGaA, Germany) to remove residual lubricants, was used to extract 10 mL N2 from the sealed serum bottle and inject the gas into the oxidation rig via the septum port (S, Figure 1).
Efficiency of CO Removal
The efficiency of CO removal was tested using a N2/CO mixture (CO concentration 100 ppm, 12 L ALLCAN, All-in-Gas E.K., Munich, Germany). Due to safety measures, we were not able to handle pure bottled CO. Several pre-washed 100 mL serum bottles sealed with butyl rubber stoppers were flushed at about ambient pressure with this gas mixture via injection needles. A Teflon tube with injection needles on both ends was then used to transfer the N2/CO mixture via the septum port into the oxidation rig. The flowrate was increased from 60 mL/min to 120 mL/min during each injection, while the furnace was heated at 250°C and, in another series, at 290°C, comparable to Pack et al. (Reference Pack, Xu, Lupascu, Kessler and Czimczik2015). The gas mixture was flushed from the serum bottles into the vacuum system.
Modern Standards
14C-enriched CO2 standards (Ox-II; NIST SRM 4990C; nominal value 1.3407 F14C) were prepared with sealed tube combustion of oxalic acid crystals according to Melchert et al. (Reference Melchert, Stolz, Dewald, Gierga, Wischhöfer and Rethemeyer2019). In summary, oxalic acid powder equal to 3 mg of C was weighed into DCM-cleaned Sn boats (4 × 4 × 11 mm, Elementar, Germany) and transferred into quartz tubes (MQ-washed and pre-combusted at 900°C) along with CuO as a combustion catalyst. The quartz tubes were evacuated at a vacuum rig below 10–3 mbar and sealed using a blow torch and combusted at 900°C for 4 hr. After combustion, the cool tubes were wiped with acetone (PESTINORM® SUPRA TRACE, grade ≥99.9%, VWR® chemicals, Germany), to remove potential dust and cloth fibres from the surface and put in a tube cracker at the vacuum rig. The CO2 was flushed with a He stream through a dry-ice ethanol slurry to remove any excess moisture and then through a LN-filled cryotrap to fixate the CO2. Subsequently, the amount of CO2 was quantified and sealed off in a glass ampoule that was carefully inserted inside a 100 mL serum bottle sealed with a butyl stopper and flushed with He for 10 min (40 mL min–1) via injection needles at ambient pressure. After flushing, the serum bottle was thoroughly shaken to break the glass ampoule inside in order to release the CO2. Lastly, a 10 mL aliquot of the CO2:He mixture was extracted using the gas tight syringe and injected into the oxidation rig via the syringe port.
Fossil Standards
Two mL of bottled 14C-free CH4 (99.995% purity, Westfalen AG, Germany) were injected at ambient pressure as a 14C-free standard and, in a second series, mixed with 14C-enriched CO2 from oxalic acid. With this second series, we tested the efficiency to trap CO2 that does not originate from CH4 and separate it from CH4-derived CO2. A ratio of 2 mL CH4 to 10 mL of He:CO2 gas mixture was injected to provide a sensitive measure for the separation of 14C-free CO2 derived from CH4 and 14C-enriched CO2 originated from Ox-II. Therefore, the synthetic air stream was closed and the mixing unit was evacuated. Then, the CH4 bottle was opened and set up at 1 bar to fill the 2 mL vacuum syringe. The valves next to the vacuum syringe and the CH4 supply were then closed and the system left to evacuate. Subsequently, the synthetic air stream was opened again and set up at 10 mL min–1. Up to this point, no cryogenic traps were yet attached and the residual CH4 was pumped out of the system. 10 mL aliquots of the Ox-II derived CO2:He mixture were extracted from the 100 mL serum bottles using the gas tight syringe and inserted into the septum port. The DI ethanol slurry and LN traps were attached to the U-tubes, the valves to the line were opened and the gases were flushed through the setup via the stream of synthetic air. After the injection, the valves to the septum port and vacuum syringe were closed and the system was flushed for 10 min with synthetic air to ensure that the gas mixture was moved completely through the setup.
Biogas Mixture
Biogas collected from a nearby biogas facility (RheinEnergie Biogasanlage Randkanal-Nord, Dormagen, Germany) in a pre-evacuated stainless-steel cylinder was used as a near-modern CH4 laboratory standard. At the biogas facility, the cylinder was directly connected to one of the maintenance gas outlets via a Teflon coated tube. The valves were opened after connecting the tube to flush the cylinder for about 5 min. The biogas mainly consists of CH4 and CO2 among other trace gases (56% CH4, 44% CO2, quantified by the supplier). 10 mL aliquots were extracted from the cylinder at ambient pressure by connecting a second valve with septum port to the cylinder and using the gas tight syringe and immediately injected into the oxidation rig, similar to the procedure for the previously described Ox-II CO2 standards.
The 14C results for the biogas were background corrected using the average 14C concentration from the injection of pure 14C-free CH4 (0.003 ± 0.003 F14C, Table 1).
Table 1 Comparison of different standard series handled on the CH4 oxidation rig, expected and measured F14C (average value with standard deviation), as well as recovery efficiency.
RESULTS AND DISCUSSION
In order to test our oxidation system for recovery efficiency and contamination, a total of 40 aliquots from multiple 2 mL and 10 mL injections containing between 10 and 50 µg C were processed for AMS analysis. The results of the different series are summarized in Figure 2 and Table 1, and the raw AMS results are summarized in the supplementary material (Table S1).
Figure 2 Results of AMS 14C measurements for the biogas mixture, of modern and 14C-free standards as well as of the CH4:CO2 gas mixture (14C-free:14C-enriched).
14C Analysis of Standards
CO2 produced from pure 14C-free CH4 had an average F14C of 0.003 ± 0.003 (n = 11). The same CH4 that was mixed with Ox-II derived CO2 contained slightly more 14C (0.006 ± 0.006 F14C; n = 13). The blank values obtained from these series are comparable to blank standards of a similar size prepared on our CO2 vacuum line that is used for small samples (10–50 µg C) including compound-specific radiocarbon analysis (F14C 0.005 ± 0.0004; n = 4; Melchert et al. Reference Melchert, Stolz, Dewald, Gierga, Wischhöfer and Rethemeyer2019). However, the blank values shown here are up to one order of magnitude higher compared to other studies (Table 2), which we think is most likely related to the much smaller size of our samples. In contrast to other studies, the CO2 is processed for direct injection into the AMS, instead of graphitization and AMS analysis as solid target, making the samples more sensitive to contributions by extraneous C. The 14C content of pure CO2 produced from Ox-II (n = 7) is close to that of the consensus value with a F14C of 1.346 ± 0.015, while the injection of carbon-free N2 gas gave no quantifiable amount of C via AMS analysis. This indicates that the system operates without quantifiable leakages. The results of the analysis of 14C-enriched CO2, 14C-free CH4 and C-free N2 further indicate that the sample handling, which includes the sealing of gas in serum bottles and transfer of aliquots via syringes, does not introduce substantial amounts of contamination.
Table 2 Overview over blank values of this and other studies as well as measured (S) and contamination corrected (R) 14C data for biogas (own data are mean values with propagated errors) as well as amounts of contaminants (mmodern; mdead).
*1 Background corrected for 0.004 ± 0.001 F14C
*2 background corrected for 0.003 ± 0.003 F14C
*3 background corrected for 0.0017 F14C
a AMS measurement of CO2
b AMS measurement of graphite.
Estimating Extraneous Carbon
The amount of extraneous C introduced during sample processing was determined from the 14C data of standard CH4, which was on average 0.20 ± 0.23 µg modern C. This amount of modern contamination is similar to values reported by Pack et al. (Reference Pack, Xu, Lupascu, Kessler and Czimczik2015) and Sparrow and Kessler (Reference Sparrow and Kessler2017). The 14C data of the 14C-enriched standards were expected to show a size dependency, i.e., an increase in F14C with decreasing sample size and vice versa, following the concept of the introduction of a constant amount of contamination (Ruff et al. Reference Ruff, Szidat, Gäggeler, Suter, Synal and Wacker2010; Rethemeyer et al. Reference Rethemeyer, Fülöp, Höfle, Wacker, Heinze, Hajdas, Patt, König, Stapper and Dewald2013; Melchert et al. Reference Melchert, Stolz, Dewald, Gierga, Wischhöfer and Rethemeyer2019). Only weak to moderate correlation coefficients after Pearson could be determined for the distribution of F14C versus sample size for the injected series of samples (14C-free CH4 r = 0.5; 14C-free CH4:Ox-II CO2 r = 0.1; Ox-II CO2 r = -0.64; biogas r = 0.46). The results of the Ox-II standards summarized in Figure 2 and the negative correlation coefficient, do not show an assessable pool of fossil contamination, although, the standards were sealed in serum bottles closed with butyl rubber stoppers, which have been shown to introduce fossil CO2 by outgassing (Gao et al. Reference Gao, Xu, Zhou, Pack, Griffin and Santos2014; Pack et al. Reference Pack, Xu, Lupascu, Kessler and Czimczik2015). This could either indicate, that during the extra steps of oxalic acid handling (sealed tube combustion, sealing in ampoules, storing in serum bottles) no more additional contaminant has been added or that the size of the standards (10 mL Ox-II:He mixture) is too large to identify additional contamination. Because of this, the thorough assessment of extraneous C via a model of constant contamination is not supported by the data distribution. Furthermore, pure N2 gas from sealed serum bottles contained no C, which either indicates that the outgassing of butyl only happens after longer storage or is below detection limit of the pressure transducers at the oxidation system.
14C Analysis of Biogas
Near modern biogas was collected at a nearby biogas facility and used as a nearly modern in-house CH4 standard. Buyable CH4 gas is mostly produced from natural gas and oil residues and is therefore 14C-depleted. The sizes of the biogas samples ranged from 20 to 50 µg C (n = 9) and gave 14C contents in a narrow range of 0.967 to 1.000 F14C (Table 2). We observed a small decrease in F14C with decreasing sample size. The 14C data of processed blank standards were used for the correction of the biogas data, which did not significantly increase 14C contents of the biogas. Because the presented data are the first tests with biogas of this facility, we have no distinct reference values like other authors (Palonen et al. Reference Palonen, Uusitalo, Seppälä and Oinonen2017; Garnett et al. Reference Garnett, Murray, Gulliver and Ascough2019).
CH4 Oxidation Efficiency and Recovery
For several injection series, an average of about 94.4% of 14C-free CH4 from the CH4:CO2 gas mixture was oxidized to CO2. An even higher amount of 97.3% CH4-derived CO2 was recovered after the oxidation of the CH4:CO2 gas mixture. The 14C content of the CH4:CO2 gas mixture was similar to the 14C content of the pure CH4 standards underlining the efficient separation of CO2 from CH4.
A temporary installation of a second LN trap behind the oxidation furnace (not included in Figure 1), as suggested by various authors (Petrenko et al. Reference Petrenko, Smith, Brailsford, Riedel, Hua and Lowe2008; Pack et al. Reference Pack, Xu, Lupascu, Kessler and Czimczik2015; Sparrow and Kessler Reference Sparrow and Kessler2017), did not increase the recovery efficiency, from which we deduced that our traps work sufficiently. The oxidation efficiency and calculated recoveries may be slightly biased by small losses of sample material related to the manual sample extraction and injection with the syringe that was not specifically accounted for. Furthermore, the oxidation efficiency and recovery are likely linked to the type of catalyst and to the packing and compaction of the catalyst inside the oxidation tube (Sparrow and Kessler Reference Sparrow and Kessler2017). We decided to use CuO rods, because they are easy to handle, cost efficient and are already routinely used in our laboratory. Pack et al. (Reference Pack, Xu, Lupascu, Kessler and Czimczik2015) reported a very high oxidation efficiency of about 100% with CuO, while Sparrow and Kessler (Reference Sparrow and Kessler2017) used platinized quartz wool to achieve consistently high efficiencies. The CuO is rather loosely packed in the quartz tube over almost the entire length of about 30 cm and fixated on each side with quartz wool, which should allow the gas to heat up properly and react with the catalyst without restricting gas flow.
The injection of pure CO2 at different flow rates of synthetic air up to 80 mL min–1 gave no changes in the amount of CO2 trapped with LN as observed by Pack et al. (Reference Pack, Xu, Lupascu, Kessler and Czimczik2015), who installed an additional CO2 trap to improve recoveries. Achieved flow rates and the pressure inside the system are highly dependent on the scale of the setup, i.e., length and diameter of the capillaries and components. Therefore, advice given on optimal setup conditions are, unfortunately, not universally applicable.
CO Oxidation Efficiency
Multiple injections of N2/CO mixtures (100 ppm CO) from sealed 100 mL serum bottles indicated an oxidation efficiency of CO to CO2 of 98–99% at 290°C (n = 5). Variations of the flow rate from 60 mL/min up to 120 mL/min neither influenced oxidation nor trapping efficiency. However, decreasing the temperature of the CO furnace to 250°C significantly decreased the oxidation efficiency to 12.5 % (n = 3). From this finding we conclude that the efficiency of the setup is dependent on the furnace temperature and not on the flow rate, which was similarly concluded by Pack et al. (Reference Pack, Xu, Lupascu, Kessler and Czimczik2015), although a stable oxidation rate was reported there for 250°C using the same catalyst (CuO).
Limitation of Sample Sizes
In this manuscript we only evaluated small volume samples containing higher CH4 concentrations than natural samples. From these high concentration samples, we deduced that neither the combustion furnaces, nor our cryogenic traps were overloaded. Thus, samples that are less concentrated in CH4, for example atmospheric samples that on average contain 2 ppm CH4 and are sampled in bags exceeding 100 L (Townsend-Small et al. Reference Townsend-Small, Tyler, Pataki, Xu and Christensen2012; Espic et al. Reference Espic, Liechti, Battaglia, Paul, Röckmann and Szidat2019), should not overload the traps either. However, their processing will take much longer using the flow rates established in this study. This practically excludes the processing of very low concentration samples (i.e., atmospheric) because it would take multiple hours to fully inject such a sample into the system. Thus, the presented system is suitable for the processing of gases from laboratory incubation experiments as well as natural wetland emissions provided from canisters or gas bags.
CONCLUSION
The need for radiocarbon dating of CH4 required the development of methods and systems for processing such samples at the CologneAMS dating center. The constructed vacuum system, which converts CH4 to CO2 for direct injection of CO2 into the AMS, was tested with 14C-free and modern standards and a near modern biogas. Sample sizes measured with the gas ion source were all in the range of 10 and 50 µg C. The results of these tests reveal the quantifiable contribution of about 0.20 ± 0.23 µg modern extraneous carbon that is introduced during sample handling. The collection and analysis of a CH4:CO2 biogas mixture representing a “natural” gas gave reproducible results (0.967–1.000 F14C). The conversion rate of CH4 to CO2 that was calculated from the analyses of standards and biogas was about 94–97% and may be further improved in the future by changing the types and packing of the oxidation catalyst inside the furnace as shown in previous studies. In summary, the preparation and handling of CH4 derived CO2 for AMS analysis is operational at CologneAMS. The current pre-treatment methods and handling of suitable samples within 1 h is time efficient and the usage of CuO catalysts and cryogenic traps instead of chemicals is cost efficient.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2023.109
ACKNOWLEDGMENTS
We thank Marcus Brand from Rheinenergie for the possibility of visiting the biogas facility Randkanal Nord and the opportunity to sample gas directly from the facility.
PREVIOUS PUBLICATION
An early version of this manuscript was used as part of a doctoral dissertation and had to be published in the online repository at University of Cologne (http://kups.ub.uni-koeln.de/id/eprint/55058). The publication of this early manuscript iteration was done for academic purposes only.
Open access
