Determination of CH4, CO2 and N2O in air samples

0 downloads 0 Views 228KB Size Report
A method for determination of the climate gases CH4, CO2 and N2O in air samples and soil atmosphere was developed using ... The presence of ¡10% by volume of C2H2 (often used experimentally to ..... R. A. Bailey, Academic Press,. 1978.
Determination of CH4, CO2 and N2O in air samples and soil atmosphere by gas chromatography mass spectrometry, GC-MS Dag Ekeberg,*a Gunnar Ogner,b Monica Fongen,c Erik J. Jonerc and Torild Wickstrømc{ a

Department of Chemistry, Biotechnology and Food science, P.O. Box 5003, Agricultural University of Norway, N-1432 A˚s, Norway. E-mail: [email protected]; Fax: +(47) 64 94 77 20 b Drottveien 4, N-1432 A˚s, Norway c Norwegian Forest Research Institute, Høgskoleveien 12, N-1432 A˚s, Norway Received 27th January 2004, Accepted 6th May 2004 First published as an Advance Article on the web 26th May 2004

A method for determination of the climate gases CH4, CO2 and N2O in air samples and soil atmosphere was developed using GC-MS. The method uses straightforward gas chromatography (separation of the gases) with a mass spectrometric detector in single ion mode (specific determination). The gases were determined with high sensitivity and high sample throughput (18 samples h21). The LOD (3s) for the gases were 0.10 mL L21 for CH4, 20 mL L21 for CO2 and 0.02 mL L21 for N2O. The linear range (R2 = 0.999) was up to 500 mL L21 for CH4, 4000 mL L21 for CO2 and 80 mL L21 for N2O. The samples were collected in 10 mL vials and a 5 mL aliquot was injected on column. The method was tested against certified gas references, the analytical data gave an accuracy within ¡5% and a precision of ¡3%. The presence of ¡10% by volume of C2H2 (often used experimentally to prevent N2 formation from N2O) did not interfere with detection for the targeted trace gases.

Aim of investigation Due to concern about climate change, there are increased demands for analysis of climate gases. CO2, CH4, and N2O are important for the climate and for processes in plants and soils. These gases, in air samples collected from various matrixes, have earlier been analysed mostly by complex gas chromatographic systems using multiple detectors with flow switching,1–4 with conversion of CO2 to CH4 prior to detection,5,6 or by photoacoustic infrared spectroscopy techniques.7 The purpose of this work was to simplify the analysis of climate gases in air and soil atmosphere by using straightforward gas chromatography with component separation on one single column, with a specific mass spectroscopic detection. The method for determination of the climate gases should be automatic, have high sensitivity and high sample throughput.

10000 mL L21), and of N2O (1 and 100 mL L21) in He or N2 were used to construct x-point calibration curves for each gas individually. These standards were certified to ¡5% accuracy, (Scott Speciality Gases). A certified mixture of the same gases in He, (8.08 mL L21 for CH4, 4023 mL L21 for CO2 and 8.24 mL L21 for N2O), at an accuracy of ¡3, 1 and 3% for the three gases, respectively, (Norsk Hydro A/S) was also used as standard sample. Another certified mixture of climate gases in He, (0.90 mL L21 for CH4, 977 mL L21 for CO2 and 1.0 mL L21 for N2O, at an accuracy of ¡10%), was used as quality control (Norsk Hydro A/S). Synthetic air (5.5, Norsk Hydro A/S) with certified residual concentrations of CH4 v 0.05 mL L21, CO2 v 0.1 mL L21 and N2O v 5 nL L21 was used for dilutions. Dilution was performed using a gas tight syringe. The matrix was always tied to match that of the sample. Chemical analysis and calculations

Description of the experimental procedures

DOI: 10.1039/b401315h

Samples from air, soil atmosphere and references Evacuated vials (9.8 mL, Chromacol 9CV) sealed with a crimp seal (20 mm, no. 151290, Brown Chromatography Supplies) were used for sampling. The vials were first evacuated to v70 Pa, filled with He 4.6 to 0.2 MPa, evacuated and flushed with He twice more and finally evacuated to v50 Pa. The evacuated vials could be stored for up to 2 weeks without being contaminated by surrounding air. The vials were filled with an air sample, reference sample or calibration standard by injection of a 10.0 mL gas sample from a gas tight syringe without undue delay. The pressure in the vials just after sampling was approximately 2% higher than the atmospheric pressure. The samples could be stored for at least 21 days in semidarkness at room temperature prior to analysis without any change in analyte concentration. Certified standards of CH4 (in concentrations 10, 100 and 1000 mL L21), of CO2 (100, 1000 and { Present address: Amersham Health, Analytical R&D, Nycovn. 2, 0401 Oslo, Norway.

The analysis was performed on a GC-MS system (GC 8000 Top gas chromatograph from CE instruments, Voyager Finnigan quadrupole mass detector, Gilson 222 XL auto sampler, all units run by the Xcalibur software). The sample was injected by a 5 mL sample loop, through a 0.5 m 6 0.32 mm deactivated precolumn, into a 25 m 6 0.32 mm CP-PoraPLOT Q-HT column (Chrompack), kept at 23 uC. Helium was used as carrier gas at 2.0 mL min21. The sampling needle and tubes were flushed with He just until penetration of the vial septum to avoid interference from air and from the previous sample. A scheme of the analytical system is described in Fig. 1. The mass spectrometer (MS) was used in selected ion monitoring (SIM) mode (8 scans s21) with electron ionisation (70 eV). The selected ions were m/z 15 for CH4, m/z 22 for CO2, and m/z 44 for N2O. The retention times for CH4, CO2, and N2O were 1.3, 1.8 and 2.2 min, respectively. The total time for one sample was set to 3.3 min to give sufficient time for sample collection, injection, a 2.5 min chromatogram and data collection. Peak height was used for quantification of CH4 and N2O, peak area for CO2. These data gave the best linear range for the analysis. The Xcalibur software performed the calculation.

This journal is ß The Royal Society of Chemistry 2004

J. Environ. Monit., 2004, 6, 621–623

621

Fig. 1 The sampling and injection of air samples for analysis of climate gases. The on/off valve (1) was switched to its ON position and helium was flushed through the valve and the sample needle (2), and at the same time, 400 mL of He was pumped by the peristaltic pump (3) through the sample loop (4). The needle penetrated the septum of the sample flask (5) immediately after the valve was switched to OFF, and 400 mL of sample was pumped through the sample loop. The sampling valve (6) was then switched for injection and returned to the position shown after 6 s. The sequence was repeated for the next sample after 3.3 min.

Statistical analyses were performed using SAS statistical program (SAS institute INC., Cary, NC, USA) or SigmaPlot 2000 (SPSS ASC b.v., Gorinchem, The Netherlands). Real samples Analysis of trace gases evolving from forest soil was used to test the method on environmental samples. Organic topsoil was collected from a plot of a field experiment with 110 year old spruce that had been fertilized with 40 kg N ha21 for 10 years; a non-fertilized plot was used as a reference.8 1 g fresh weight of 3 homogenized samples from each plot was incubated with 4 mL 10 mM NaNO3 in gas tight 120 mL bottles under aerobic and anaerobic conditions (the latter used an He atmosphere and included repeated treatment with 10% acetylene by volume in the bottle headspace serving as a denitrification inhibitor) at 15 uC for 16 h prior to gas sampling (19 mL) and GC-MS analysis.

Results and discussion Analysis in scan mode (from m/z 12 to m/z 50) is not sufficient for determination of CH4 and N2O in the concentrations usually found in the atmosphere. We therefore used selected ion monitoring (SIM) mode to achieve sufficient sensitivity. For the qualitative analyses of CH4, ions at m/z 15 (which corresponds to CH3+) were used. This ion is present in 89% of the base ion (National Institute of Standards and Technology (NIST) library data). The base peak in the mass spectrum of methane that corresponds to CH4+, at m/z 16, could not be used due to interference from O2 in air (formation of O+). The base ion of N2O is m/z 44. This ion gives satisfactory results when the concentration of CO2 was below 4000 mL L21. At higher concentrations of CO2 the tailing of its peak interfered with N2O analysis. CO2 was analysed at m/z 22, (CO22+), an ion present in 1.4% of the base ion (NIST library data). The base ion for CO2 is found at m/z 44, but this had too high intensity to give good determinations at concentrations above approximately 3000 mL L21 for CO2. A SIM plot of the ions m/z 15,

Fig. 2 The SIM chromatograms of outdoor air as recorded by m/z 15, m/z 22 and m/z 44. The peaks at RT = 1.28 min, RT = 1.81 min and RT = 2.21 min represent CH4 (1.75 mL L21), CO2 (400 mL L21) and N2O (0.33 mL L21), respectively. The chromatographic details are described in the text.

m/z 22 and m/z 44 is presented in Fig. 2 using ambient air as an example. The LOD (3s) for the different gases were 0.10 mL L21 for CH4, 20 mL L21 for CO2 and 0.02 mL L21 for N2O. The high LOD for CO2 was due to interference from CO2 in air. The linear range was up to 500 mL L21 for CH4 (R2 = 0.999), 4000 L L21 for CO2 (R2 = 0.999) and 80 mL L21 for N2O (R2 = 0.999). Analysis of samples with CO2 in the range 4000– 10000 mL L21 for CO2 was performed on diluted samples. Data for accuracy and precision are presented in Table 1. All data had a standard deviation, s, within ¡10% of the mean value. If it was necessary to repeat an analysis, one extra 400 mL sample could be drawn from the sample vial without significant decrease in the concentration determined. Ten following injections from the same vial filled with a calibrated air sample gave a continuous decrease in the CO2 concentration analysed (from 335 ¡ 14 to 235 ¡ 10, (n = 10, mean ¡2s), R2 = 0.86, p v 0.0001) due to reduced pressure in the vial. The long-term instrument stability was tested by analysing 180 samples of laboratory air. The air was sampled at 3.3 min intervals just prior to analysis during a period of approximately 10 h. The CO2 concentration decreased from 312 to 288 mL L21 (R2 = 0.87, p v 0.001), the mean concentration of the first 10 samples was 312 ¡ 4 and the last 10 286 ¡ 2 mL L21 for CO2

Table 1 Determination of certified reference gases (n = 10) at three different concentrations (in mL L21). The low concentration was at approximately 56LOD Low concentration Gas CH4 CO2 N2O 622

Diluted reference 0.50 97.7 0.10

Medium concentration Mean 0.52 105 0.09

¡ ¡ ¡ ¡

2s 0.04 4.0 0.02

J. Environ. Monit., 2004, 6, 621–623

Certified value 0.90 977 1.0

High concentration Mean 1.09 1016 1.01

¡ ¡ ¡ ¡

2s 0.06 74 0.10

Certified value 8.08 4023 8.24

Mean 8.11 4034 8.32

¡ ¡ ¡ ¡

2s 0.2 68 0.18

(mean ¡2s). When recalibrations were performed every 1 h the corresponding values for the first and last set of 10 samples were identical (315 ¡ 4 mL L21 for CO2). The results for CH4 and N2O gave a corresponding stability. Additional information on detector stability was obtained from 22Ne+, the first peak at RT = 1.02 min in the m/z 22 graph of Fig. 2. The first and last set of 10 samples had mean concentrations (¡2s) of 17.7 ¡ 0.6 and 17.9 ¡ 0.4 mL L21 for Ne (relative to a literature value of 18.2 mL L21 for Ne).9 GC-MS analyses of gases in the headspace of bottles with incubated soil showed no differences in production of trace gases between soils with different nitrogen input history under aerobic conditions. Under anaerobic conditions, however, N2O production was markedly higher in the soil having received additional nitrogen for 10 years. The addition of C2H2 had no impact on N2O production, whereas it increased CO2 production by approximately 40% and CH4 production about 200 times. Apparently, C2H2 served as a readily available substrate for microbial methanogenesis. These and other10 results show clearly that this method is satisfactory for environmental gas analysis. In conclusion, the presented GC-MS method for analysis of climate gases has high sample throughput, high sensitivity and performs with sufficient stability to give reliable analytical data.

Acknowledgements We thank E. Molnes, Instrument-Teknikk A/S, for technical assistance.

References 1 2

T. Magnusson, Plant Soil, 1989, 120, 39. B. K. Sitaula, J. Luo and L. R. Bakken, J. Environ. Qual., 1992, 21, 493. 3 A. Vermoesen, H. Ramon and O. van Cleemput, Pedologie, 1991, 41-2, 119. 4 M. Yoh, M. Takeuchi and H. Tode, Jpn. J. Limnol., 1998, 59, 147. 5 T. Nakazawa, T. Machida, K. Esumi, M. Tanaka, Y. Fujii, S. Aoki and O. Watanbe, J. Glaciology, 1993, 39, 209. 6 J. M. Ncki, M. Zimnoch, J. Miroslaw and J. Lasa, Chem. Anal. (Warsaw), 1999, 44, 841. 7 S. Yamulki and S. C. Jarvis, J. Geophys. Res., 1999, 104, 5463. 8 O. J. Kjønnaas, A. O. Stuanes and M. Huse, For. Ecol. Manage., 1998, 101, 227. 9 Chemistry of the Environment, ed. R. A. Bailey, Academic Press, 1978. 10 E. Joner, to be published.

J. Environ. Monit., 2004, 6, 621–623

623