10 Measurement of Trace Gases, I: Gas Analysis, Chamber Methods, and Related Procedures Keith A. Smith and Franz Conen The University of Edinburgh, Edinburgh, Scotland I. INTRODUCTION The exchange of gases between the biosphere and the atmosphere has had a profound effect on the development of the Earth environment. Globally, vegetation (prin cipally forests) releases and absorbs about 60 billion tonnes of carbon dioxide, CO 2 , annually and it is the perturbation of this exchange by additional anthropogenic emissions of about a tenth of this quantity that is the principal contribution to global warming—the ‘‘greenhouse effect.’’ Emissions of methane from natural wetlands, rice fields, and landfills, and of nitrous oxide from fertilized agricultural soils and the soils of tropical rainforests, add to global warming. A further contribution comes from tropospheric ozone, produced by reactions involving volatile organic compounds from natural vegetation, and NO x , which comes variously from combustion sources and from soils. Soil surfaces can, on the other hand, act as sinks for many pollutant gases in the atmosphere, through both physicochemical sorption and microbial oxidation. Improved methods of measurement of the fluxes of these gases have become of major environmental importance, both to determine flux levels and in order to improve understanding of the processes involved, to aid the prediction of future trends. This chapter covers the principles of current instrumental techniques for the determination of trace gases, and the experimental procedures used TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. to make flux measurements in the field on a small scale: from 0.1 m 2 of land surface to 10–100 m 2 . Examples of recent applications are included. Chapter 11 contains complementary information on micrometeorological methods for the determination of fluxes at larger scales. Throughout, the commonly used concentration units ppm and ppb will be used for convenience, instead of the more rigorous notation in volumetric units (mLL À1 and nL L À1 ), or partial pressure units (1 ppm ffi 0:1 Pa and 1 ppb ffi 10 À4 Pa) unless the context dictates otherwise. II. GAS CHROMATOGRAPHY Gas chromatography, GC, is the principal analytical method used for the measurement of trace gases in soil air and in the atmosphere above the soil. In particular, it ha s provided most of the data on the fluxes of the greenhouse gases methane and nitrous oxide and is also used in some studies for measuring carbon dioxide released by soil respiration. A. Principles A brief outline of the principles involved in gas chromatography (GC) is given below, but for a more comprehensive discussion, the reader is referred to books solely devoted to this subject, e.g., Bruner (1993) and Braithwaite and Smith (1995). The subject (though more related to the analysis of nongaseous substances) is also covered in Chap. 12. Chromatography is essentially the separation of the components of a mixture resulting from differences in their partition between a stationary phase with a large active surface area and a moving phase that flows over the stationary phase. Depending on the state of the moving phase, a distinction is made between liquid and gas chromatography. Separations that would be very difficult to achieve by any other means are possible by chromatographic methods, because even small differences between the components in their partition between the phases are multiplied many times during their passage through the chromatographic system. GC may be conveniently classified into two types: gas–liquid and gas– solid chromatography. In the former, the stationary phase is an involatile liquid coated either on an inert support material in a packed column, or on the internal wall of an open tubular (capillary) column. In the latter type, the column is packed with an adsorbent of small particle size, or the inner surface of a capillary column is coated with a thin adsorbent layer. For separations of gaseous mixtures, solid adsorbents (see below) have been used more widely than liquid stationary phases. 434 Smith and Conen TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. At constant temperature, pressure, and carrier gas velocity, the rate at which a component travels along the column is related to the partition coefficient K: K ¼ C s C g ð1Þ where C s and C g are the concentrations in the stationary and gaseous phases, respectively, The more strongly bound components stay for a longer time in the stationary phase and thus travel along the column more slowly than the more weakly bound components. The zone occupied by a component broadens as it moves along, due to the diffusion of molecules in the gaseous phase. The separation of the components in a sample depends on the relative values of the partition coefficients. The closer the values are to each other, the longer is the column required to give an adequate separation. Reducing the temperature has the effect of increasing the separation. In practice, the variables that are most commonly exploited to achieve satisfactory separations are 1. The material used as stationary phase 2. The length and diameter of the column 3. The column temperature 4. The flow rate of the moving phase (the carrier gas) The essential parts of all gas chromatographs consist of 1. A carrier gas system, generally in the form of a cylinder supply at pressures up to 20 MPa (200 bar), with a two-stage regulator to reduce the pressure and a flow-rate regulating valve 2. A device for introduction of the sample into the carrier gas stream 3. A column to separate the components of the sample 4. A detector to indicate the elution of each component and produce a response proportional to the quantity of the component present 5. A recording system to provide a permanent record of the detector response The column and detector (and sometimes the injection port) are normally enclosed in separate thermostats (‘‘ovens’’) that can be controlled individually. This is because the optimum column temperature for achieving a desired separation is usually different from the optimum detector temperature. Measurement of Trace Gases, I 435 TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. B. Commercial Instruments A wide range of gas chromatographs (GCs) is available from manufac- turers in several countries (Table 1). For many applications the choice of instrument is not critical, because the substances to be measured are easily within the detection capabilities of any current instrument; where this is so, the cost of the instrument and the availability of after- sales service are more important. Modern GCs usually have the capacity for at least two columns and two detectors (of the same or different types) and also have digital control of all important functions: inlet, oven, and detector temperatures, carrier gas flow rates, detector output settings, etc. If it is desired to add, say, another detector to the original configuration, the necessary electronic control/amplifier circuitry is now usually added as a plug-in board, whereas in earlier generations of GCs it would generally have been housed in a separate unit and linked by cable to the main system. The modern systems have many advantages but do make it more difficult for the user to assemble hybrid systems of units from different manufacturers. For information on the latest developments, manufacturers’ literature and websites should be consulted. The web site addresses of some of the major suppliers are given in Table 1. C. Systems for Gas Analysis 1. Columns and Packings Porous Polymer Beads. Solid adsorbents such as porous polymer beads, marketed under such names as Porapak and HayeSep, are commonly used for separations of permanent gases. These materials are available in different grades that vary in their retention characteristics for particular substances. Porapak Q and its equivalents have been used widely for separating CH 4 ,CO 2 , and/or N 2 O from nitrogen and oxygen in air samples. Columns 1–2 m in length, 1/8–1/4 00 (3.2–6.4 mm) in diameter, packed with these materials, give satisfactory separations of these gases at modest temperatures in the range 30–70  C. Nongraphitized Carbon Molecular Sieve. Thes e materials offer an alternative to porous polymer beads for separating permanent gases. Separations of air (oxygen plus nitrogen ), carbon monoxide, methane, CO 2 , and/or C 2 hydrocarbons can be readily achieved with a 2 m Â1/8 00 packed column of carbon molecular sieve, depending on the column temperature (Fig. 1). 436 Smith and Conen TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. Table 1 Some Manufacturers of Gas Chromatographs Manufacturer Address Website Comments Agilent 395 Page Mill Rd, PO Box 10395, Palo Alto, CA 94303, USA www.agilent.com Formerly Hewlett-Packard Buck Scientific 58 Fort Point St, East Norwalk, CT 06855, USA www.bucksci.com Small transportable instruments suitable for field use CE Instruments Strada Rivoltana, 20090 Rodano, Milan, Italy www.ceinstruments.it Formerly Carlo Erba/Fisons. Now part of Thermoquest GOW-MAC Instruments 277 Brodhead Rd, Bethlehem, PA 18017-8600, USA www.gow-mac.com PerkinElmer Instruments 710 Bridgeport Ave., Shelton, CT 06484-4794, USA www.perkinelmer.com Shimadzu Scientific Instruments 1 Noshinokyo Kuwabarocho, Kyoto 604-8511, Japan and 7102 Riverwood Dr. Columbia, MD 21046, USA www.shimadzu.co.jp www.shimadzu.com SRI Instruments 3882 Del Amo Blvd., Ste 601, Torrance, CA 90503, USA www.srigc.com Range includes small transportable instruments suitable for field use Unicam Chromatography Viking Way, Bar Hill Cambridge, CB3 8EL, UK www.unicam.co.uk Formerly Pye-Unicam. Now part of Onix Process Analysis Varian Instruments 2700 Mitchell Dr., Walnut Creek, CA 94598, USA www.varianinc.com Measurement of Trace Gases, I 437 TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. Capillary Columns and Open Tubular Columns. Complex mixtures of volatile organic compounds (VOCs), emitted to the atmosphere from vegetation, are analyzed using long capillary or open tubular columns. Examples are a Megabore Carbowax 20 M column, 60 m long, with a film thickness of 1.2 mm, and a 50 Â0.22 mm BP1 column, with a film thickness of 0.12 mm (Kesselmeier et al., 1996) (see Sec. IV below). 2. Procedures for Key Gases Methane, CH 4 . Methane is determined with a flame ionization detector, which responds to combustible compounds (in effect, organic compounds) and, to some extent, to oxygen, but which is insensitive to the Figure 1 Effect of temperature on separation of permanent gases on Carbosphere carbon molecular sieve. (Reproduced by permission of Alltech Associated Inc., Deerfield, IL, USA.) 438 Smith and Conen TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. presence of inorganic gases and water vapor. Methane must, therefore, be separated satisfactorily from oxygen and other organic compounds, and this can be done readily on Porapak Q and similar materials. The limit of detection for this gas is of the order of 1 ppb, so there is no problem of determining concentrations at ambient level (1.7 ppm) and above. In some studies, in fact, dilution of samples becomes necessary to avoid nonlinear detector response (see below). Nitrous Oxide, N 2 O. This gas is determined normally with a selective electron capture detector (ECD: see below), operated at high temperature (340–390  C). The ECD responds primarily to gases with a high affinity for electrons, e.g., oxygen, water vapor, and halogenated compounds. Thus effective separation of N 2 O from such gases is very important. This is achieved easily on Porapak Q or HayeSep Q. Analysis times can be reduced by using backflushing techniques to remove the more slowly eluted substances such as water vapor (e.g., Arah et al., 1994). ECDs also have some sensitivity for CO 2 , and separation from the latter gas can also be achieved on the same column packings, if the temperature is low enough. Alternatively, the sample can be passed through a precolumn of soda lime to remove the CO 2 before it enters the analytical column. Carbon Dioxide, CO 2 . If this gas is to be determined by GC rather than with an IRGA, the most common procedure is separat ion from oxygen/nitrogen on Porapak Q and detection with a thermal conductivity detector (TCD). Aternatively an ECD can be used, operated at a lower temperature (ca. 270  C) than is commonly employed for N 2 O; however, if the relative concentrations of CO 2 and N 2 O are in a suitable range, these two gases can be determined simultaneously by this procedure (Thomson et al., 1997). Isoprene and Monoterpenes. These are the most important VOCs. They play an important part in atmospheric chemistry by, for example, contributing to tropospheric ozone production (through photochemical reactions involving NO x ). Separation requires the use of a capillary or open tubular column (see above). Temperature programming, in which the column oven increases in temperature at a predetermined rate during the analysis, is also employed, to speed up the elution of the more involatile components of the mixture. Detection of VOCs is normally by FID, but the mass spectrometer has become an increasingly attractive alternative, because of its high sensitivity and its capability for distinguishing between compounds with similar retention times by their fragmentation patterns (Parrish and Fehsenfeld, 2000). Measurement of Trace Gases, I 439 TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. Simultaneous Measurement of Two or More Gases. The separation and quantification of a family of related compounds, such as the VOC separation discussed in the previous paragraph, may require a long column and take many minutes for a single analysis, but it normally only requires one column and one detector. Sometimes, however, there is a need to determine simultaneously in the same sample different gases that require different columns and/or different detectors. Some earlier multicolumn and multidetector systems for this purpose were described in the previous edition (Smith and Arah, 1991), and more recent examples are given in Sec. IV below. 3. Calibration The concentration of the target gas in a sample is conveniently determined by comparisons of peak height (or peak area) with those obtained with gas standards of known concentration, having first established the relationship between peak height (or area) and detector response. Most modern detector/ amplifier systems have a linear response over a concentration range covering several orders of magnitude. If the response becomes nonlinear but is not saturated, the system can still be used, provided several standards of different concentrations are analyzed in order to construct a response curve. The alternative is to make dilutions, e.g., with air or nitrogen, before sample injection. Better reproducibility of results is usually obtained with a sampling valve than by direct injection with a syringe. 4. Automation When large numbers of samples have to be analyzed, a degree of automation can greatly reduce operator time. Commercial headspace samplers are available that consist of a conveyor/sample changer system carrying 10–50 glass vials sealed with serum caps, and an automatica lly operated gas syringe. The syringe needle punctures the cap of each vial in turn, a gas sample is withdrawn, the syringe moves to a position above the GC injection port, and the sample is injected through the septum. The main problem is expense—a headspace sampler commonly costs more than the GC to which it is attached. Automated systems for the injection of gas samples collected in the field in Tedlar bags, syringes, or other containers are not commercially available. Either the samples have to be transferred to evacuated vials, and thence to a headspace sampler, or a ‘‘do-it-yourself’’ automatic injection system has to be constructed. In one such system used by our group, the loop of a motor-driven gas sampling valve is evacuated by a pump and then coupled, by the operation of a rotary switching valve, to each of a series of 440 Smith and Conen TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. sample containers (syringes, bags, or even small incubation vessels) in turn. Each container is connected to the valve via luer-lock fittings. The gas in the sample vessel expands into the loop, the contents of which are then injected (Fig. 2). Data acquisition and processing is nowadays almost always performed electronically. The chromatographic data are recorded either with a digital integrator or via a computer interface, using specialized software. Statistical evaluation of replicate analyses can take place, and results can be stored for subsequent retrieval, without the manu al transfer of data at any stage. This has the benefit of eliminating errors, as well as saving operator time. III. NONDISPERSIVE INFRARED GAS ANALYSIS A. Principles Gaseous compounds such as CO 2 ,N 2 O, CH 4 , and H 2 O vapor absorb electromagnetic radiation in the infrared (IR) part of the spectrum, due to the particular vibrational/rotational energies of their interatomic bonds. This is, of course, the characteristic that makes these compounds greenhouse gases and thus the focus of much environmental research, but it also provides the basis for sensitive methods of analysis. The main Figure 2 Schematic diagram of automatic GC injection system for gaseous samples. (From Arah et al., 1994.) Measurement of Trace Gases, I 441 TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. constituents of the atmosphere, N 2 and O 2 , being symmetrical diatomic molecules, do not absorb in the IR region; thus analyzers based on IR absorption can be used to determine trace concentrations of CO 2 ,N 2 O, CH 4 ,andH 2 O vapor in air without interference from N 2 or O 2 . These instruments contain three main components: an infrared source, a gas cell through which the radiation passes, and an IR detector. Absorption of IR radiation at any particular wavelength takes place according to the Beer–Lambert law: I ¼ I 0 e Àkcd ð2Þ where I and I 0 are the intensities of the transmitted and incident radiation, respectively, k is the molar extinction coefficient for the particular wavelength, c is the concentration of the absorbing gas, and d is the path length through the cell. Infrared gas analyzers (IRGAs) fall into two main categories: dispersive (DIR) and nondispersive (NDIR). In the former, an IR beam from a source filament passes through a monochromator, and radiation of a selected wavelength that is strongly absorbed by the target gas passes through a gas cell and then to a detector. By changing the wavelength passing through the cell, other gases may be targeted, and if a scanning system is used, an absorption spectrum will be obtained. However, for study of, for example, CO 2 fluxes during photosynthesis or respiration, a simpler approach, NDIR analysis, is generally used. Here there is no attempt to select particular wavelengths. Instead, the broad band of radiation from the source is employed, and the total absorption of all the lines in the 4.26 mm major absorption band of CO 2 is measured (Long and Ha ¨ llgren, 1993). B. Commercial Instruments Current IRGA system include The range of instruments made by LI-COR (Lincoln, Nebraska, USA), such as the LI-6252 CO 2 Analyzer and the LI-6262 CO 2 /H 2 O Analyzer The Rosemount Analytical (Orrville, Ohio, USA) BINOS series of instruments The PP Systems Inc. (Haverhill, Mass., USA) CIRAS-2 portable photosynthesis system The ADC Bioscientific Ltd. (Hoddesdon, Herts., UK) Mini-IRGA These analyzers are compact and por table and can be used in the field under battery power, as well as in the laboratory on a mains AC power 442 Smith and Conen TM Copyright n 2004 by Marcel Dekker, Inc. All Rights Reserved. [...]... greenhouse gases and ozone precursors between soils (or vegetation) and the atmosphere A wide range of ecosystems has been investigated, including agricultural land, natural and plantation forests, natural grassland and moorland, tundra, and wetlands Landfills, which constitute a significant source of methane, have also been TM Copyright n 2004 by Marcel Dekker, Inc All Rights Reserved 464 Smith and Conen... the CO2 and 222Rn concentration profiles were determined, and the 222Rn emission rate was also measured, using a closed chamber and an alpha counter Davidson and Trumbore (1995) used a vented dynamic chamber system and the 222Rn tracer method in deep forest and pasture soils in the eastern Amazon, and concluded that 20–30% of the CO2 emission from the soil surface resulted from root respiration and decay... soil surface (or water surface, in wetlands and rice paddies), and the change in concentration with time of a gas emanating from the soil (or being absorbed by the soil from the air in the box) is measured by one of the instrumental techniques discussed in the preceding sections This section describes various versions of the technique and discusses the advantages and disadvantages that they offer A Closed... tower- and aircraft-based [-d-] estimates, the agreement was satisfactory, although earlier work by two of the authors (Moore and Roulet, 1991) had indicated that static chambers gave flux values 20% lower than dynamic ones Frolking and Crill (1994) investigated climatic controls on CH4 emission from a New England fen with closed static chambers They employed much larger chambers (150 L volume, and approx... (IRGA) This version (LI-COR 6262) has two detectors, for measurement of both CO2 and water vapor (Courtesy of LI-COR Corp., Lincoln, NB, USA.) TM Copyright n 2004 by Marcel Dekker, Inc All Rights Reserved 444 Smith and Conen Figure 4 Relationships between (A) signal ratio (IRGA cell/reference cell) and CO2 concentration, and (B) CO2 concentration and IRGA output voltage (Courtesy of LI-COR Corp., Lincoln,... require sampling of the soil atmosphere This can be, for example, in the context of studying soil conditions for root growth (Strojny et al., 1998), trace gas flux between soil and atmosphere (Gut et al., 1999), soil gas diffusivity (Dorr and Munnich, 1990; Lehmann ¨ ¨ et al., 2000), or for identifying the depth and the rate at which a gas species is produced or consumed in the soil (Neftel et al., 2000)... (12 mL in the LI-COR series instruments) and gold-plated to enhance IR reflection The radiation is collimated by CaF2 lenses and focused on the detector Air from the sample source (e.g., a photosynthesis or soil respiration chamber, see Sec III below) is pumped through the sample cell, while a CO2 standard of known concentration flows through the reference cell Detectors In the LI-6252 and 6262 instruments,... concentrations measured by TGA and by ECD-GC (Reproduced from Ambus and Robertson, 1998, by permission of the Soil Science Society of America.) beam provides the spectrum of the input beam and the absorption by any sample inserted in the beam (Griffith and Galle, 2000) For analysis of trace gases, air is drawn continuously from the sampling point through Teflon tubing and then into two 25-L optical cells (White... sum of the NO and NO2 concentrations, and analyzers such as the Thermo Environmental (Franklin, Mass., U.S.A.) Model 42C automatically cycle between the NO and NOx modes and calculate the NO, NO2 , and NOx concentrations The system is extremely sensitive: the limit of detection for NO, using a 60 s averaging time, is 0.4 ppb To measure soil emissions of NO, air is passed at between 50 and 150 L minÀ1... N2O emissions to the atmosphere, and high fluxes are correlated with rapid N cycling (providing NHþ and NOÀ substrates) and high soil 3 4 water content and temperatures A typical study in a temperate agricultural environment is that of Velthof et al (1997), who found that 5–14% of the N applied to a poorly drained sandy soil in a wet spring was emitted as N2O Smith and Dobbie (2001) compared the cumulative . effect.’’ Emissions of methane from natural wetlands, rice fields, and landfills, and of nitrous oxide from fertilized agricultural soils and the soils of tropical rainforests, add to global warming placed over the soil surface (or water surface, in wetlands and rice paddies), and the change in concentra- tion with time of a gas emanating from the soil (or being absorbed by the soil from the. NO and NO 2 concentrations, and analyzers such as the Thermo Environmental (Franklin, Mass., U.S.A.) Model 42C automatically cycle between the NO and NO x modes and calculate the NO, NO 2 ,andNO x concentrations.