Preparation and Validation of Fully Synthetic Standard Gas Mixtures

Jan 10, 2014 - Paul J. Brewer,*. ,† ... National Physical Laboratory, Analytical Science Division, Hampton Road, Teddington, Middlesex TW11 0LW, U.K...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Preparation and Validation of Fully Synthetic Standard Gas Mixtures with Atmospheric Isotopic Composition for Global CO2 and CH4 Monitoring Paul J. Brewer,*,† Richard J. C. Brown,† Michael N. Miller,† Marta Doval Miñarro,† Arul Murugan,† Martin J. T. Milton,†,§ and George C. Rhoderick‡ †

National Physical Laboratory, Analytical Science Division, Hampton Road, Teddington, Middlesex TW11 0LW, U.K. National Institute of Standards and Technology, 100 Bureau Drive, MS-8393 Gaithersburg, Maryland 20899-8393, U.S.A.

‡

ABSTRACT: We report the preparation and validation of the first fully synthetic gaseous reference standards of CO2 and CH4 in a whole air matrix with an isotopic distribution matching that is in the ambient atmosphere. The mixtures are accurately representative of the ambient atmosphere and were prepared gravimetrically. The isotopic distribution of the CO2 was matched to the abundance in the ambient atmosphere by blending 12C-enriched CO2 with 13C-enriched CO2 in order to avoid measurement biases introduced by measurement instrumentation detecting only certain isotopologues. The reference standards developed here have been compared with standards developed by the National Institute of Standards and Technology and standards from the WMO scale. They demonstrate excellent comparability.

T

for comparable reference standards. An infrastructure to disseminate reference standards prepared gravimetrically that are traceable to the International System of Units (SI) offers a means of broadening availability. These could overcome the cost and complexity of sampling air under global background conditions which can only be carried out at remote locations. It promises the possibility for the provision of standards from more than one source and would be enabled and supported by the global agreement under the International Committee for Weights and Measures (CIPM) Mutual Recognition Arrangement (MRA)5 to which the WMO became a signatory is 2010. However, for CO2, the challenge goes beyond the already difficult task of preparing high accuracy gravimetric mixtures that are SI traceable and comparable to the WMO scale within the stringent Data Quality Objectives (DQOs). The introduction of cavity ring-down spectrometers (CRDS) to the commercial market has significantly improved the precision and reliability of routine measurement of these components, resulting in a substantial increase in the number of field measurements. Most commercial CRDS instruments respond only to 12C16O16O and cannot detect the presence of the other major isotopologues (12C18O16O, 13C18O16O, and 13C16O16O). Due to fractionation during the processes employed by industry to purify CO2, samples of the pure gas have a larger 12C/13C

he Global Atmosphere Watch (GAW) program of the World Meteorological Organization (WMO) is a major source of data that contributes toward our understanding of the atmosphere’s natural and anthropogenic change, and this program helps to improve the understanding of interactions between the atmosphere, the oceans, and the biosphere. An essential part of the quality assurance system implemented by the WMO is the designation of a Central Calibration Laboratory (CCL) that maintains the Primary Standard (PS) for each measurand. The CCL for CO2 and CH4 is the Earth System Research Laboratory of the National Oceanic and Atmospheric Administration (NOAA) in Boulder, CO.1,2 Its work is based on the use of a manometric system to assign absolute CO2 values to the primary standards3 and then the use of these to transfer the WMO scale to secondary standards using nondispersive infrared (NDIR) instruments.4 There are currently 15 primary CO2/air standards ranging from 250−520 μmol/mol. These are calibrated at regular intervals (between 1 and 2 years) with the manometric system. Reference standards of CH4 in air are prepared by gravimetry and cover the nominal range of 300−2600 nmol/mol, which is representative of air extracted from glacial ice to contemporary background atmospheric conditions. Standards for both components are samples of ambient air sampled in a global background location and are disseminated in high-pressure gas cylinders. They have been the subject of intensive research to determine their accuracy and lifetime.2 As the requirement for data that is comparable to the WMO scale increases, there is a corresponding increase in the demand © 2014 American Chemical Society

Received: December 9, 2013 Accepted: January 10, 2014 Published: January 10, 2014 1887

dx.doi.org/10.1021/ac403982m | Anal. Chem. 2014, 86, 1887−1893

Analytical Chemistry

Article

Figure 1. Hierarchy of reference gas standards for CO2 and CH4. Standards E, F, G, H, and I form set 1, and standards J, K, L, and M form set 2. All standards were prepared in a dry scrubbed whole air matrix (Scott Marrin, Inc.) except standards F and K (boxes shown with dotted lines), which were prepared in a synthetic air matrix (BOC, metrology grade).

scale within the DQOs set by the WMO of 0.10 μmol/mol for the northern hemisphere and half this for the southern hemisphere. The methods used here to produce synthetic standards with SI traceability can be replicated; hence, this work offers the prospect of the wider availability of standards for the global monitoring of two high impact greenhouse gases.

ratio than the isotopic distribution in the ambient atmosphere. The difference is significant and biases the response of the CRDS by as much as 0.17 μmol/mol.6 Hence, reference mixtures must be isotopically matched to ambient composition to ensure equivalence to the WMO scale, when measured by techniques employed in the field. Moreover, CRDS instruments are also influenced by the matrix gas composition.7 This results from absorption line broadening and narrowing for CO2 molecules due to random thermal motion and collisions. Random thermal motion and intermolecular collisions produce line-broadening effects (referred to as Doppler and Lorentzian broadening effects, respectively), whereas the velocity changing collisions produce line-narrowing effects. Commercial CRDS instrumentation models the line shape of the target gas which describes the line broadening and narrowing effects. However, the magnitudes of these effects depend on the matrix gas composition. Nara et al.7 observed that for a 1% change in the amount fraction of O2 in an air matrix, the change in analyzer response to CO2 was approximately 1 μmol/mol. The same experiment was performed for N2 (2.5% shift in amount fraction) and Ar (20% shift in amount fraction). The analyzer response changed by approximately 1 and 0.1 μmol/mol, respectively. A similar bias was observed for CH4 when the amount fraction of these components are altered in the matrix gas. These findings highlight the importance of matching the matrix to ambient composition and the benefits of using whole air with the CO2, CH4 and water removed (dry scrubbed whole aira) as the matrix for synthetic reference standards because it already has the correct composition. We show that reference mixtures of CO2 and CH4 can be prepared synthetically with an isotopic distribution matching atmospheric abundance, which are equivalent to the WMO

■

GRAVIMETRIC REFERENCE STANDARDS OF CARBON DIOXIDE AND METHANE IN AIR Figure 1 shows the hierarchy of the reference mixtures prepared at the National Physical Laboratory (NPL). All mixtures were prepared by gravimetry, in accordance with ISO 6142,8−10 in 10 L aluminum cylinders (BOC) with a DIN1 outlet diaphragm valve (Ceodeux). The cylinders were treated internally with a proprietary passivation process (BOC Spectraseal) to inhibit adsorption of target components. The cylinders were evacuated using an oil-free pump (Scrollvac SC15D, Leybold Vacuum) and turbo molecular pump with magnetic bearing (Turbo vac 340M, Leybold Vacuum) to a pressure of approximately 1 × 10−7 mbar. A binary reference standard of CH4 in synthetic air was prepared at a nominal amount fraction of 1000 μmol/mol (standard A). A 1/16″ tube (Swagelok, electro-polished stainless steel) was purged several times with pure CH4 (grade 6.0, CK Gas Products) and then used to transfer a sample to an evacuated cylinder. The cylinder was weighed before and after the addition using a single pan balance (Mettler Toledo ID7). Nitrogen (Air Products, BIP+) and oxygen (N6.0, BOC Specialty Gases) were then added and weighed in turn following the same procedure used for CH4. Nitrogen was added first in order to avoid the mixture passing the explosive limit of 5%. Three binary reference mixtures of CO2 in the dry scrubbed whole air matrix were prepared at a 1888

dx.doi.org/10.1021/ac403982m | Anal. Chem. 2014, 86, 1887−1893

Analytical Chemistry

Article

nominal amount fraction of 4000 μmol/mol (standards B, C, and D) following the same method used for CH4. Standards B and C were prepared from pure “natural” CO2 (a mixture of “industrial” CO2 (BOC, 99.999%) spiked with 99% 13CO2 (CLM-185-5, Cambridge Isotope Laboratories) in a ratio of 3398:1, calculated in order to achieve an isotopic distribution closely matching ambient).11 Standard D was prepared directly from pure “industrial” CO2 (BOC, 99.999%). Standards E, G, H, I, L, and M were all prepared by diluting standard A with one of standards the B, C, or D and dry scrubbed whole air as detailed in Figure 1. Standard J was prepared directly from pure “industrial” CO2 by diluting with standard A and dry scrubbed whole air. In order to interrogate the influence of matrix on the analytical technique employed here, standards F and K were prepared by dilution of standard B and the pure “industrial” CO2, respectively, with synthetic air (BOC, metrology grade), which contained no nitrous oxide and only a negligible amount of argon compared to the dry scrubbed whole air. The dry scrubbed whole air matrix used here (Scott Marrin, Inc.) is air sampled from the ambient atmosphere and with the CO2, CH4 and water removed catalytically. The amount fraction of CH4 and CO2 in the dry scrubbed whole air was accurately determined by comparison to gravimetric reference standards at 7 nmol/mol for CH4 and at 250 and 400 nmol/ mol for CO2, using a gas chromatograph (GC) fitted with a methaniser and a flame ionization detector (FID). Similar to the manometric measurement, GC analysis measures total CO2 irrespective of the isotopologue distribution. Purity analysis of the pure CO2 and pure CH4 was also conducted. The results are shown in Table 1. Analysis of CH4 and CO2 in the high

broadening effects in the CRDS used here.7 A GC with a helium discharge ionization detector (HDID) was used to compare the dry scrubbed whole air to primary reference gas standards of argon, nitrogen, and oxygen. The sequence of preparations resulted in two sets of standards: Set 1 consisted of standards E, F, G, H, and I, which all contain CO2 with an isotopic distribution that matches ambient air. Set 2 consisted of standards J, K, L, and M, which all contain CO2 with a isotopic distribution enriched in 12CO2 due to fractionation processes during the purification of the CO2. Sets 1 and 2 are distinguished in Figure 1 by gray and white boxes, respectively. The carbon isotope distribution of the CO2 was confirmed by isotope ratio mass spectrometry (IRMS).9 Figure 2 shows how the atomic weight of carbon increases as the proportion of 13C present increases. The graph presents the

Table 1. Purity Analysis of the Major Components in Pure Industrial CO2, Pure CH4, and Dry Scrubbed Whole Air

Figure 2. Atomic weight of carbon plotted as a function of the proportion of 13C present, denoted by the delta scale where Peedee Belemnite (PDB) has a δ13C value of zero.

source gas industrial CO2

methane

dry scrubbed whole air

component

amount fraction (μmol/mol)

expanded uncertainty (μmol/mol)

CO2 O2 Ar CO N2 CH4 O2 Ar CO2 CO N2 N2 O2 Ar CO2 CO CH4

999 999 0.1 0.05 0.2 0.2 999 999 0.1 0.04 0.035 0.2 0.2 781 280 209 350 9365 0.05 0.004 0.0032

1 0.1 0.04 0.2 0.2 1 0.1 0.04 0.035 0.2 0.2 500 100 30 0.05 0.002 0.001

difference in atomic weight for naturally occurring CO212 compared to the CO2 purified by industrial processes measured for the purposes of this study and their difference on the δ scale. The isotopic distribution of naturally occurring CO2 is shown to the right of the graph. The filled circle shows the measurement of a standard referenced to the WMO scale using IRMS. The open circles show IRMS measurements of the CO2 purified by industrial processes (with the most negative δ value) and the same gas spiked with pure 13CO2, both diluted to ambient amount fractions with synthetic air (standards K and F, respectively). The relative shift in the δ scale is 28.7 δ13C ‰, which confirms that the spiked mixture has a similar isotopic distribution to atmospheric abundance and the reference standards used to define the WMO scale. The remaining discrepancy of 2.83 ‰ is estimated to produce less than a 0.015 μmol/mol bias on the CRDS measurements. The graph illustrates that knowledge of the CO2 composition is crucial not only for negating any bias introduced from analytical instrumentation that is only sensitive to a subset of the CO2 isotopologues13−16 but also for assigning the correct atomic weight for the calculation of gravimetrically prepared mixtures, which can change the amount fraction by as much as 0.0044 μmol/mol.

purity nitrogen and oxygen used in the preparation of the nominal 1000 μmol/mol CH4/air standard was not carried out as their relative contribution to the uncertainty budget is negligible. Measurements of argon, nitrogen, and oxygen in the dry scrubbed whole air were conducted to ensure that their amount fractions matched natural abundance within target limits (0.778−0.783 mol/mol for nitrogen, 0.207−0.211 mol/ mol for oxygen, and 0.00887−0.00980 mol/mol for argon) in order to negate any biases introduced from the pressure 1889

dx.doi.org/10.1021/ac403982m | Anal. Chem. 2014, 86, 1887−1893

Analytical Chemistry

■

Article

ANALYSIS USING CAVITY RING-DOWN SPECTROSCOPY A CRDS (Picarro G2301) was used in the analysis. The instrument employs two lasers for the simultaneous measurement of CO2 and CH4, a high-precision wavelength monitor, a high finesse optical cavity with three high-reflectivity mirrors (>99.995%), a photodetector, and a computer. During the measurements, light at a specific wavelength from a laser is injected into a cavity through a partially reflecting mirror. The light intensity then builds up over time and is monitored through a second partially reflecting mirror using a photodetector located outside the cavity. The “ring-down” measurement is made by rapidly turning off the laser and measuring the time constant of the light intensity as it exponentially decays. This measurement is used to calculate the absorbance. Comparison of the decay time of the cavity, with and without target gas, enables the concentration of the target gas to be calculated. The lasers are tuned to scan over the individual spectral lines of 12C16O2 at the wavelength of 1603 nm, 12C1H4, and H216O at the wavelength of 1651 nm, producing a highresolution spectrum of each. The 1σ precision of 12C16O2 and 12 1 C H4 measurements over an optimum measurement period are reported by the manufacturer as