Introduction

AMT

Atmospheric Measurement Techniques

AMT

Atmos. Meas. Tech.

1867-8548

Copernicus Publications

Göttingen, Germany

10.5194/amt-10-4613-2017

Validation of spectroscopic gas analyzer accuracy using gravimetric standard gas mixtures: impact of background gas composition on CO2 quantitation by cavity ring-down spectroscopy

Lim

Jeong Sik

https://orcid.org/0000-0002-7326-1725

Park

Miyeon

Lee

Jinbok

Lee

Jeongsoon

leejs@kriss.re.kr Center for Gas Analysis, Chemical and Medical Metrology Division, Korea Research Institute of Standards and Science (KRISS), Gajeong-ro 267, Yuseong-gu, Daejeon 34113, Republic of Korea

Jeongsoon Lee (leejs@kriss.re.kr)

1December2017

10 12 46134621 23February2017 24May2017 11August2017 16September2017

This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/3.0/

This article is available from https://amt.copernicus.org/articles/.html

The full text article is available as a PDF file from https://amt.copernicus.org/articles/.pdf

The effect of background gas composition on the measurement of CO2 levels was investigated by wavelength-scanned cavity ring-down spectrometry (WS-CRDS) employing a spectral line centered at the R(1) of the (3 00 1)III ← (0 0 0) band. For this purpose, eight cylinders with various gas compositions were gravimetrically and volumetrically prepared within 2σ=0.1 %, and these gas mixtures were introduced into the WS-CRDS analyzer calibrated against standards of ambient air composition. Depending on the gas composition, deviations between CRDS-determined and gravimetrically (or volumetrically) assigned CO2 concentrations ranged from -9.77 to 5.36 µmol mol-1, e.g., excess N2 exhibited a negative deviation, whereas excess Ar showed a positive one. The total pressure broadening coefficients (TPBCs) obtained from the composition of N2, O2, and Ar thoroughly corrected the deviations up to -0.5 to 0.6 µmol mol-1, while these values were -0.43 to 1.43 µmol mol-1 considering PBCs induced by only N2. The use of TPBC enhanced deviations to be corrected to ∼ 0.15 %.

Furthermore, the above correction linearly shifted CRDS responses for a large extent of TPBCs ranging from 0.065 to 0.081 cm-1 atm-1. Thus, accurate measurements using optical intensity-based techniques such as WS-CRDS require TPBC-based instrument calibration or use standards prepared in the same background composition of ambient air.

Introduction

Emission of carbon dioxide (CO2), the most important greenhouse gas, has been reported to increase, resulting in global climate change (Messerschmidt et al., 2011; Solomon et al., 2007). According to the IPCC Fourth Assessment Report (Solomon et al., 2007), CO2 is the major contributor to global warming, with a 62.9 % share of the total radiative force caused by long-lived greenhouse gases. Although it is not yet feasible to quantify all its sources and sinks within small uncertainties (Conway et al., 1988; Schulze et al., 2009), all countries have agreed to consistently control CO2 emissions, necessitating accurate measurements of atmospheric CO2 mole fractions. Gas chromatography coupled with flame ionization detection (GC-FID) (van der Laan et al., 2009), nondispersive infrared spectroscopy (NDIR) at 4.26 µm (Lee et al., 2006; Min et al., 2009; Crawley, 2008; Tohjima et al., 2009), Fourier transform infrared (FTIR) spectroscopy (Griffith et al., 2012), tunable diode laser absorption spectroscopy (TDLAS) (Durry et al., 2010), wavelength-scanned cavity ring-down spectroscopy (WS-CRDS) (Crosson, 2008), and other cavity-enhanced absorption spectroscopies (O'Shea et al., 2013) are well-known techniques for quantifying atmospheric CO2. Despite exhibiting the advantage of high measurement precision, GC-FID suffers from long acquisition time due to delayed CO2 retention in the separation column (typically several tens of minutes). NDIR shows better performance than GC-FID in real-time measurements due to using filtered spectral fingerprints of CO2 instead of relying on analyte separation. However, frequent calibrations are required to correct NDIR response drifts. Recently, WS-CRDS has attracted attention because of its high precision and low drift. In contrast to intensity-based techniques such as NDIR and TDLAS, CRDS is immune to laser shot noise and detector electric noise due to employing the ring-down count method. Furthermore, the increased path length offered by the resonant optical cavity provides excellent sensitivity, i.e., signal-to-noise ratio, and high precision. Since a CO2 inter-laboratory compatibility of ±0.1 µmol mol-1 in the Northern Hemisphere was set as a goal by the World Meteorological Organization (WMO), WS-CRDS is viewed as a competitive technique for measuring atmospheric greenhouse gas levels (Rella et al., 2013).

Accurate measurements of atmospheric CO2 levels by WS-CRDS require the removal of water vapor, which causes spectral interference, and an empirical cubic polynomial model for correcting the water background has been developed (Rella et al., 2013). Nevertheless, CO2 mole fraction measurements can be adversely affected by spectral line broadening if calibration gas mixtures whose background composition is different from the natural N2 : O2 : Ar ratio in the atmosphere are used (Nara et al., 2012). In this study, standard gas mixtures containing ambient levels of CO2 in synthetic air (N2 + O2 + Ar) were gravimetrically prepared for utilization as calibration standards and measuring targets for investigating the impact of background gas composition on WS-CRDS responses, owing to the excellent uncertainty of gravimetric gas mixtures. Furthermore, an empirical equation for correcting the “matrix effect” was derived in terms of total pressure broadening. The good agreement achieved between CO2 mole fractions of the calibration standards and synthetic samples of arbitrary composition validated the measurement accuracy of matrix-effect-corrected WS-CRDS.

Purities of raw carbon dioxide and background gases (N2, O2, and Ar).

Impurity Mole fraction Detectorsa component (µmol mol-1) CO2 N2 O2 Ar H2 < 0.1 < 0.1 < 0.1 < 0.1 PDDb O2 < 0.1 0.003 ± 0.003 – 0.003 ± 0.002 PDD Ar < 0.1 21.6 ± 4.32 < 1.0 – TCDc N2 12.8 ± 2.56 – 3.1 ± 0.62 2.4 ± 0.48 PDD CO 0.3 ± 0.06 < 0.005 0.08 ± 0.016 < 0.005 PDD and FIDd CH4 2.6 ± 0.52 < 0.005 < 0.005 < 0.005 PDD and FID CO2 – 0.002 ± 0.001 0.195 ± 0.039 < 0.002 PDD and FID H2O 4.5 ± 2.25 1.6 ± 0.8 1.1 ± 0.55 0.9 ± 0.45 Dew point meter C2 2.8 ± 0.56 – – – AEDe C3–C5 0.7 ± 0.35 – – – AED Purity (%) (k=2) 99.9976 ± 0.0007 99.9976 ± 0.0009 99.9995 ± 0.0002 99.9996 ± 0.0001

a Tabulated detectors were coupled to the main body of the gas chromatograph (Agilent 6890A). b Pulsed discharge detector. c Thermal conductivity detector. d Flame ionization detector. e Atomic emission detector.

Materials and methods Preparation of standard gas mixtures

Gas mixtures were prepared using gravimetric and volumetric methods, based on ISO 6142-1 (2015) and ISO 6144 (2003), respectively. The gravimetric method featured filling pure CO2 (MG industries, USA) and N2 (Deokyang Energen, South Korea) gases into a clean aluminum cylinder. Subsequently, pure O2 (Praxair Co., South Korea) and Ar (Deokyang Energen, South Korea) gases were added to the obtained CO2 / N2 mixture to obtain an ambient level of CO2 in a matrix of synthetic air. The amounts of filled gases were determined based on their weight, which was obtained by weighing the aluminum cylinder before and after filling. The weights used for calibrating the weighing balance (Mettler Toledo, XP 26003L, USA) were calibrated against the national kilogram standard to ensure measurement traceability. For high weighing precision, an automatic weighing machine patented by KRISS was used to control the loading position on the weighing pan of the top loading balance, resulting in a typical weighing uncertainty of less than 0.005 %. A circular turntable was used to support tare and sample cylinders. During weighing, the drift of the weighing balance and the buoyancy effect exerted by the cylinders were effectively corrected or canceled out by using the following bracketing sequence: tare – cylinder A – tare – cylinder B – tare – cylinder C. The preparation of standard gas mixtures based on this technique has been reported in detail elsewhere (Wessel, 2008). The CO2 mole fraction in the resulting mixture can be computed as follows: yj=∑A=1Pxj,A⋅mA∑i=1nxi,A⋅Mi∑A=1PmA∑i=1nxi,A⋅Mi. Here, yj is the mole fraction of component j in the gas mixture, P is the total number of parent gases, n is the total number of components in the final mixture, mA is the measured mass of parent gas A, Mi is the molar mass of component i, and xi,A or xj,A is the mole fraction of component i or j in parent gas A. Therefore, quantification of impurities present in pure parent gases is needed to determine the composition of each parent gas. Hence, impurities in N2, O2, Ar, and CO2 were analyzed by gas chromatography employing various detection methods, e.g., thermal conductivity detection (TCD), pulsed discharge detection (PDD), FID, and atomic emission detection (AED), with detector assignments for all impurities given in Table 1. Purity, namely the mole fraction of the dominant component in “pure” parent gas (xpure), was determined as follows: xpure=1-∑i=1Nxi, where N is the number of impurities likely to be present in the final mixture. For selecting target impurities, the source and its purification process were considered. If the expected impurity was not detected, its mole fraction was set to half of the limit of detection (LOD/2), and the associated standard uncertainty was defined as the assigned mole fraction divided by √3, e.g., LOD/(2 ⋅ √3), as expected for a uniform probability density function ranging from 0 to LOD (ISO 6142-1, 2015). In particular, it was very important to accurately analyze the mole fractions of target components (N2, O2, Ar, and CO2) in the respective raw gases, since the weighed target component amount in the obtained mixture could be biased by the presence of the same component in other raw gases as an impurity. For instance, the mole fractions of CO2 in pure N2, O2, and Ar gases were determined as 0.002, 0.195, and < 0.002 µmol mol-1, respectively. Thus, the amounts of CO2 in pure N2 and Ar gases were negligible and did not impact final mixtures with CO2 fractions above 300 µmol mol-1. However, the large amount of CO2 in pure O2 led to a bias of 0.04 µmol mol-1, which was comparable to the uncertainty level of the final mixture. Table 1 summarizes the reference values and associated uncertainties of major impurities in raw gases.

For CO2, a verification test was representatively performed to determine the potential systematic error of the gravimetric procedure described above, relying on comparing the detection sensitivity of CO2 in different gas mixtures using GC-FID coupled with an MS-5A (molecular sieve 5A, 4 m) separation column. The column oven was kept at 30 ∘C, and high-purity N2 (99.999 %, Deokyang Energen) was used as a carrier gas. Sample gas flows were carefully controlled to ensure that the same amount of gas was introduced into the sample loop regardless of its composition; for this purpose, mass flow controllers (MFCs) were calibrated using a flow meter (digital flow calibrator, cat no. 20123, Restek Inc., USA). Therefore, the CO2 mole fraction uncertainty of prepared mixtures included uncertainties associated with the weighing process, raw gas purities, and verification tests, resulting in a gravimetric preparation uncertainty of less than 0.1 µmol mol-1 (1σ).

The standard gas mixture denoted as EBXXXXXXX (Table 2) was prepared by the static volumetric method (ISO 6144, 2003; Waldén, 2009). Ambient air was collected with a pressurizing pump through a chemical moisture trap containing Mg(ClO4)2 in order to yield the complementary gas, namely dry air. The amount of N2 was then varied by diluting the dry air with high-purity N2 (> 99.999 %), which eventually led to a variation in the mole fractions of the major components, N2, O2, Ar, and CO2. In this way, the mole fractions of the background gas composition can be easily predicted by using the measured pressure ratio of the filled gas. In the case of the CO2 mole fraction, three volumetric cylinders (EBXXXXXXX) were calibrated against the gravimetric standards (Table 2) because the mixing ratio of atmospheric CO2 varies each day. Eventually, the compositions of EB0006391 and ME0434 closely reflected the atmospheric ratio of the major components. Notably, all prepared gas mixtures were maintained under very dry conditions, with the mole fraction of H2O being less than 5 µmol mol-1.

Mole fractions of gas mixtures.

Cylinder no. Gas composition Preparation (cmol mol-1) method CO2

N2 O2 Ar DF4560 400.61 (0.05 %) 99.96 – – gravimetry EB0011591 351.78 (0.10 %) 83.45 16.48 0.04 volumetry EB0011528 353.08 (0.10 %) 80.97 18.19 0.81 volumetry ME5590 386.94 (0.05 %) 78.33 21.63 – gravimetry EB0006391 406.40 (0.10 %) 78.16 20.87 0.93 volumetry ME0434 402.25 (0.05 %) 78.07 21.03 0.87 gravimetry ME5502 384.35 (0.05 %) 77.57 20.53 1.86 gravimetry ME5537 385.35 (0.05 %) 70.98 18.85 10.12 gravimetry

* Numbers denote the mole fraction (µmol mol-1) of CO2 and its relative preparation uncertainty.

Schematic diagram depicting the gas supply to the WS-CRDS analyzer. The acronym SUS represents the stainless steel.

The 12C / 13C ratio of CO2 raw gas for the gravimetric standards was similar to the atmospheric level of approximately -11 ‰, which suggests similar isotope ratios would occur across the prepared cylinders as determined by gravimetry and volumetry. Nevertheless, isotope effects biasing the CRDS response seemed to be hardly discernable in this study because verification (calibration) of the CO2 mole fractions in the prepared gravimetric (volumetric) standards was carried out by GC-FID, which measured the total carbon isotopes.

Summary of CRDS calibration results.

Cylinder no. CO2 mole fraction (µmol mol-1) Difference Gravimetrically Before CRDS After CRDS (B - A) (B - A)/A × 100 assigned value (A) calibration calibration (B) (µmol mol-1) (%) ME0424 371.22 371.18 371.29 0.07 0.0193 ME0485 380.31 380.23 380.28 -0.03 -0.0088 ME5552 384.76 384.66 384.67 -0.09 -0.0222 ME0434 402.25 402.41 402.30 0.05 0.0117

Cavity ring-down spectroscopy

CRDS is an ultrasensitive technique introduced by O'Keefe and Deacon in 1988 (Chen et al., 2010; Rothman et al., 2005). In principle, the leakage rate of the trapped laser source in the optical cavity can be fitted by monoexponential decay, and absorbance at wavelength λ can then be calculated from the difference of ring-down signal decay rates in the presence and absence of the target gas. Alternatively, the absorbance at λ can be determined from the ring-down time at the non-absorbing wavelength λ0 in the presence of the target gas. In this study, a commercial WS-CRDS (G-1301, Picarro, USA) was employed. Since the WS-CRDS system has been described elsewhere (Chen et al., 2010; Nara et al., 2012), only a brief description is provided here. The WS-CRDS analyzer, operating at a wavelength of 1.603 µm that corresponds to R(1) of the (3 00 1)III ← (0 0 0) band, is comprised of diode lasers, a high-precision wavelength monitor, a high-finesse cavity defined by three high-reflectivity mirrors (< 99.995 %), a photodiode detector, and a data acquisition computer. Laser light confined in the cavity traveled along the triangular optical axis, exhibiting an effective path length of 15–20 km. Ambient air or gas from a pressure-regulated tank was supplied to the optical cavity backed by a built-in diaphragm pump, which was conditioned to a highly controlled pressure and temperature of 140 ± 0.05 Torr and 40 ± 0.01 ∘C, respectively.

Result of WS-CRDS calibration using gravimetric standards (ambient air background composition, see main text for details). Good agreement between gravimetric and CRDS-determined CO2 concentrations was observed.

CO2 concentrations determined by gravimetry and measured by well-calibrated CRDS, together with the correction due to N2-induced pressure broadening (PBC(N2)) and total pressure broadening coefficient (TPBC). Differences between the measured (corrected) and assigned concentrations are also listed.

Cylinder no. CO2 mole fraction (µmol mol-1) Difference Gravimetrically CRDS measured PBC (N2) TPBC

DSTD-CRDS

(B - A)/A × 100 (C - A)/A × 100 (D - A)/A × 100 assigned value value corrected corrected (B - A) (%) (%) (%) (A) (B) (C) (D) (µmol mol-1) DF4560 400.61 390.84 401.09 400.82 -9.77 -2.44 0.12 0.05 EB0011591 351.78 349.62 351.79 351.97 -2.16 -0.61 0.00 0.05 EB0011528 353.08 352.05 353.00 353.15 -1.03 -0.29 -0.02 0.02 ME5590 386.94 386.51 386.17 386.47 -0.43 -0.11 -0.20 -0.12 EB0006391 406.40 406.39 405.97 406.15 -0.01 0.00 -0.11 -0.06 ME0434 402.25 402.34 401.87 402.09 0.09 0.02 -0.09 -0.04 ME5502 384.35 384.80 384.09 384.17 0.45 0.12 -0.07 -0.05 ME5537 385.35 390.71 386.78 385.95 5.36 1.39 0.37 0.16

For this study, a gas flow rate of 400 mL min-1 and a pig-tailed bypass-out were combined to achieve a steady gas flow undisturbed by laboratory pressure fluctuation, yielding a constant pressure in the CRDS cavity (Fig. 1). The outer diameters of stainless steel tubes connecting highly pressurized cylinders to the MFC (5850E, Brooks Inc., USA) inlet and the MFC to the spectrometer equaled 0.125 and 0.0625 in., respectively. High-purity nitrogen was used for flushing the gas lines and CRDS analyzer between switching cylinders.

The measured spectral line consisting of ∼ 10 points was fitted by the Galatry profile to obtain quantitative information, based on the assumption that the CRDS read-out was influenced only by variations in the CO2 concentration of tested samples and not by variations of background gas composition (Chen et al., 2010). This assumption implies that the peak height of the fitted profile was regarded as a CRDS read-out instead of the corresponding integrated area (Nara et al., 2012). As described in Table 3, CRDS responses were calibrated against gravimetric standards, in which N2, O2, and Ar ratios are close to that in the atmosphere ratio, with CO2 concentrations very similar to those of ambient air (between 360 and 410 µmol mol-1). Absorbance was found to be linearly proportional to the concentration of light-absorbing gas, as indicated by the straight-line fit of CRDS responses with R2 ∼ 0.9999 (Fig. 2 and Table 3), supporting the validity of the attempted calibration and the hypothesis proposed in this study. In other words, deviations from expected sensitivity (i.e., CRDS response divided by the gravimetric concentration of CO2) were due to deviations in the composition of background gas from that of ambient air, namely the extent of alien gas line broadening or narrowing.

Results and discussion

To investigate the effect of background gas composition on CRDS responses, gas mixtures were analyzed against ambient-air-like standards using a well-calibrated CRDS (Table 4).

Deviations of CO2 concentrations determined by CRDS from those assigned by gravimetry (or volumetry) ranged from -2.44 to 1.39 %. CRDS responses of EB0006391 and ME0434 were in good agreement with the assigned CO2 concentrations, showing deviations of less than 0.1 µmol mol-1, whereas extreme deviations of greater than 1 % were observed for cylinders DF4560 and ME5537. In particular, the CO2 concentration of DF4560 (CO2 in pure N2) showed a deviation of -9.77 µmol mol-1. Therefore, it can be conjectured that N2-induced broadening is more important than that induced by other background gases, O2 and Ar. Since the optical cavity was kept at constant pressure and temperature, Doppler broadening was not considered. Instead, collision-induced broadening (or narrowing) was invoked in the case of variable composition. The collisional half-width, i.e., the total pressure broadening coefficient (γTPB), can be expressed as follows:

N2-induced line broadening (x axis) vs. difference between CRDS-measured and assigned CO2 levels of standard gas mixtures (y axis).

γTPB=∑i=1nγi⋅pi, where γi is the pressure broadening coefficient (PBC) of component i, and pi is the partial pressure of component i, e.g., its molar fraction multiplied by the cavity pressure of 18 kPa. The maximum peak height of the Galatry profile at a given background gas composition, G(γ), can be assumed to be linearly proportional to the PBC for a sufficiently narrow interval of pi, Δpi (Varghese and Hanson, 1984). In view of the dominance of N2-induced pressure broadening, the difference between CRDS-determined and gravimetrically (volumetrically) assigned CO2 concentrations of the measured sample, DSTD-CRDS, can be determined as follows: DSTD-CRDS∝Gγ∝γN2⋅pN2. As shown in Fig. 3, a linear relationship between DSTD-CRDS and N2-induced line broadening was found at given partial pressures (i.e., mole fractions multiplied by cavity pressure) in the optical cavity.

The PBC of N2 was set to 0.08064 cm-1 atm-1, as reported by Nakamich et al. (2006). Since N2 showed the largest PBC among those of other background components, positive (or negative) deviations between CRDS-determined and assigned CO2 concentrations of tested cylinders, i.e., the lower (or higher) extent of pressure broadening, were observed at N2 concentrations below (or above) the ambient value of 78 cmol mol-1 corresponding to ME5590 (Table 4). Thus, the CO2 concentration could be corrected based on the following linear fit: ycorrected=yCRDS--606.63⋅γN2⋅pN2+38.656, where γN2⋅pN2 is the N2-induced pressure broadening, yCRDS is the value obtained by WS-CRDS, and ycorrected is the CO2 concentration corrected for N2-induced pressure broadening. Corrected CO2 concentrations exhibited good agreement (within 0.4 %) with the regression fit (R2 ∼ 0.9736). This correction error significantly exceeded the instrumental precision (reported as 0.01 % (1σ); Nara et al., 2012), strongly suggesting the presence of other error sources.

The pressure broadening correction of ME5537 showed the highest deviation of 0.4 %. The background gas composition of ME5537 (70.98 % N2, 18.85 % O2, and 10.13 % Ar) implied that the Ar content should be taken into account for the correction. Since CO2 self-broadening is negligible due to the low concentration of CO2 compared to that of other components (N2, O2, and Ar) in the investigated gas mixtures, the total pressure broadening coefficient (TPBC) could be expressed as a function of alien gas PBCs and the partial pressures of the corresponding components: γTPBC=γN2pN2+γO2pO2+γArpAr. Table 5 shows the reported PBCs for N2, O2, and Ar, and Table 6 shows TPBCs of all cylinders, with (a), (b), and (c) denoting results obtained independently by Pouchet et al. (2004), Nakamichi et al. (2006), and HITRAN2004 (Rothman et al., 2005), respectively.

Summary of N2-, O2-, and Ar-related pressure broadening coefficients in cm-1 atm-1. All parameters were taken by using the Voigt function.

Pouchet et Nakamichi et HITRAN 2004 al. (2004) al. (2006)

γN2

0.0721 0.08064 0.0778

γO2

0.0660 0.06695 0.0702

γAr

– 0.06312 –

γair

– – 0.0758

Since the coefficients of Ar have not been reported by Pouchet et al. (2004) and HITRAN2004, the corresponding TPBCs include only N2- and O2-related pressure broadening (Table 6). Therefore, the TPBCs in (a) and (c) were underestimated in comparison to those in (b). For instance, TPBCs of 0.0636 and 0.0685 were obtained for cylinder ME5537 in the cases of (a) and (c), respectively, with the value for (b) equaling 0.07625. As shown in Table 6, the TPBC of ME5537 exhibited the largest deviation of 20 %, originating mainly from the Ar mole fraction. Figure 4 shows DSTD-CRDS values (taken from fourth column in Table 4) as a function of calculated TPBCs (taken from Table 6).

Pressure broadening for investigated gas mixtures based on pressure broadening coefficients from different sources.

Cylinder no. Pouchet et Nakamichi et HITRAN 2004* al. (2004)* al. (2006) DF4560 0.0721 0.08061 0.0778 EB0011591 0.0710 0.07835 0.0765 EB0011528 0.0704 0.07798 0.0758 ME5590 0.0708 0.07765 0.0761 EB0006391 0.0701 0.07759 0.0755 ME0434 0.0702 0.07758 0.0755 ME5502 0.0695 0.07747 0.0748 ME5537 0.0636 0.07625 0.0685

* Pressure broadenings were estimated without Ar due to the absence of a broadening coefficient in the corresponding studies.

Total pressure broadening coefficient vs. difference between CRDS-measured and assigned CO2 levels of standard gas mixtures. Due to the lack of γAr, correlations (a) and (c) exhibit poor fits.

TPBC values reported by Nakamichi et al. (2006) exhibited a linear correlation with CRDS responses within the investigated background composition interval. In practice, Huang and Yung (2004) reported that the Lorentzian width is inversely proportional to the peak value of the Voigt function for a fixed Gaussian width. The results shown in Fig. 4 reveal that DSTD-CRDS values decreased with increasing TPBCs, in agreement with previous reports (Huang and Yung, 2004). Only the result of (b) exhibited a fairly linear behavior; however, nonlinearity was observed when the broadening coefficients of O2 or Ar were not taken into account. The following equation was derived for correcting CRDS-determined concentrations: ycorr.TPB=yCRDS--3382.1⋅γTPBC+262.65. Here, yCRDS is the CRDS-measured value of the standard gas mixture, and ycorr.TPB is the corresponding corrected CRDS response computed using the relation in (b) (Fig. 4). Table 4 summarizes the results obtained after correction using Eq. (7), showing that the correction was improved from 0.68 (N2 PBC) to 0.33 µmol mol-1 (TPBC) in terms of standard deviations (1σ) of differences (corrected minus gravimetry-assigned). Furthermore, R2 was improved to 0.99 when pressure broadening related to three main components of air (N2, O2, and Ar) was taken into account. For every cylinder, excellent agreement was observed after implementing the TPBC corresponding to the assigned values. In particular, even cylinders DF4560, ME5590, and ME5537, whose background gas compositions were significantly different from that of ambient air, exhibited good correlation of CO2 concentrations determined by CRDS with those assigned by gravimetry or volumetry. It is worth noting that the quality of the TPBC correction can be improved further by using quality standards with lower background composition uncertainties, including 13CO2 isotopologues and precisely measured broadening coefficients that are deduced from advanced line-shape functions such as Galatry and Rautian profiles.

Conclusions

In this study, we investigated the impact of background gas composition on spectroscopic quantitation of CO2 at ambient concentration. Standard gas mixtures with various background compositions were prepared by gravimetry or volumetry for use as calibration standards and test samples. Purity analysis and gravimetric weighing showed high accuracy and precision. For purity analysis, analytical techniques such as GC-PDD, TCD, FID, AED, and dew point metering were used. Raw gas (N2, O2, Ar, and CO2) purities were obtained within uncertainties of less than 0.001 % (1σ). Moreover, biasing impurities in N2, O2, and CO2 were accurately crosschecked. With a weighing precision of 0.007 %, the preparation uncertainties of gravimetric and volumetric mixing were demonstrated to be lower than 0.05 and 0.1 % (2σ), respectively, after performing verification tests. The preparation uncertainty of volumetry was slightly higher than that of gravimetry, still being sufficiently satisfactory to distinguish error sources for “matrix effect” correction. Based on the composition accuracy of the prepared gas mixtures, CO2 levels were determined by WS-CRDS for eight standard gas mixtures with different background compositions. An injection unit with a bypass-out was used to ensure a precise and moderate gas inflow from a highly pressurized cylinder to the WS-CRDS, which was calibrated against well-certified standard gas mixtures of air composition with CO2 levels of 360–410 µmol mol-1. Among the eight cylinders, the CRDS responses of EB0006391 and ME0434 were well matched to the corresponding preparative values, whereas the values obtained for other cylinders exhibited large deviations between +5.36 and -9.77 µmol mol-1. For a N2-enriched mixture (DF4560), the CRDS-determined CO2 concentration was 2.44 % lower than the preparative value. Since CRDS calibration was performed using standards with ambient air composition, the fact that CRDS responses tended to be negative for N2-enriched and positive for Ar-enriched mixtures was in good agreement with the results obtained in earlier experimental (Nara et al., 2012; Zhao et al., 1997) and theoretical studies (Huang and Yung, 2004), reflecting the dependence of line broadening on alien gas composition.

Therefore, a linear shift of CRDS responses was observed for TPBCs above 0.05 cm-1 atm-1, which covers 20 % N2-enriched and 10 % Ar-enriched gas mixtures. TPBC-corrected CRDS responses were in good agreement with the gravimetric (or volumetric) concentration of the investigated gas mixtures within 0.15 % (±0.6 µmol mol-1). Considering the instrumental uncertainty of 0.01 % (1σ), the improved PBC uncertainties should lead to lower discrepancies of corrected CRDS responses. The correction presented in Eq. (7) works only for the designated vibrational transition, i.e., R(1) of the (3 00 1)III ← (0 0 0) band at 1.603 µm, and referred PBCs, but a similar calibration strategy can be used for determining gas mixing ratios by other intensity-based optical measurement techniques.

No data sets were used in this article.

JiL prepared the certified reference materials, JSL and MP performed measurements and analysis. JeL designed experiments, and JSL prepared the manuscript with contributions from other co-authors.

The authors declare that they have no conflict of interest.

Acknowledgements

This work was funded by the Korea Meteorological Administration Research and Development Program under grant no. KMIPA 2015-2032. Edited by: Christian Brümmer Reviewed by: Zoe Loh and Jooil Kim

References 1

Chen, H., Winderlich, J., Gerbig, C., Hoefer, A., Rella, C. W., Crosson, E. R., Van Pelt, A. D., Steinbach, J., Kolle, O., Beck, V., Daube, B. C., Gottlieb, E. W., Chow, V. Y., Santoni, G. W., and Wofsy, S. C.: High-accuracy continuous airborne measurements of greenhouse gases (CO2 and CH4) using the cavity ring-down spectroscopy (CRDS) technique, Atmos. Meas. Tech., 3, 375–386, 10.5194/amt-3-375-2010, 2010.

Conway, T. J., Tans, P., Waterman, L. S., Thoning, K. W., Masarie, K. A., and Gammon, R. H.: Atmospheric carbon dioxide measurements in the remote global troposphere 1981–1984, Tellus B, 40, 81–115, 10.1111/j.1600-0889.1988.tb00214.x, 1998.

Crawley, L. H.: Application of non-dispersive infrared (NDIR) spectroscopy to the measurement of atmospheric trace gases, MSc Thesis, University of Canterbury, Christchurch, New Zealand, 2008.

Crosson E. R.: A cavity ring-down analyzer for measuring atmospheric levels of methane, carbon dioxide, and water vapor, Appl. Phys. B, 92, 403–408, 10.1007/s00340-008-3135-y, 2008.

Durry, G., Li, J. S., Vinogradov, I., Titov, A., Joly, L., Cousin, J., Decarpenterie, T., Amarouche, N., Liu, X., Parvitte, B., Korablev, B., Gerasimov, M., and Zéninari, V.: Near infrared diode laser spectroscopy of C2H2, H2O, CO2 and their isotopologues and the application to TDLAS, a tunable diode laser spectrometer for the Martian PHOBOS-GRUNT space mission, Appl. Phys. B, 99, 339–351, 10.1007/s00340-010-3924-y, 2010.

Griffith, D. W. T., Deutscher, N. M., Caldow, C., Kettlewell, G., Riggenbach, M., and Hammer, S.: A Fourier transform infrared trace gas and isotope analyser for atmospheric applications, Atmos. Meas. Tech., 5, 2481–2498, 10.5194/amt-5-2481-2012, 2012.

Huang, X. and Yung, Y. L.: A common misunderstanding about the Voigt line profile, J. Atmos. Sci., 61, 1630–1632, 10.1175/1520-0469(2004)061<1630:ACMATV>2.0.CO;2, 2004.

ISO 6142-1: Gas analysis- Preparation of calibration gas mixture-Part 1: Gravimetric method for class 1 mixtures, available at: https://www.iso.org/standard/59631.html, last access: August 2015.

ISO 6144: Gas analysis- Preparation of calibration gas mixture-Static volumetric method, available at: https://www.iso.org/standard/24666.html, last access: February 2003.

Lee, J.-Y., Hee-Soo, Y., Marti, K., Moon, D. M., Lee, J. B., and Kim, J. S.: Effect of carbon isotopic variations on measured CO2 abundances in reference gas mixtures, J. Geophys. Res., 111, D053302, 10.1029/2005JD006551, 2006.

Messerschmidt, J., Geibel, M. C., Blumenstock, T., Chen, H., Deutscher, N. M., Engel, A., Feist, D. G., Gerbig, C., Gisi, M., Hase, F., Katrynski, K., Kolle, O., Lavric, J. V., Notholt, J., Palm, M., Ramonet, M., Rettinger, M., Schmidt, M., Sussmann, R., Toon, G. C., Truong, F., Warneke, T., Wennberg, P. O., Wunch, D., and Xueref-Remy, I.: Calibration of TCCON column-averaged CO2: the first aircraft campaign over European TCCON sites, Atmos. Chem. Phys., 11, 10765–10777, 10.5194/acp-11-10765-2011, 2011.

Min, D., Kang, N., Moon, D. M., Lee, J. B., Lee, D. S., and Kim, J. S.: Effect of variation in argon content of calibration gases on determination of atmospheric carbon dioxide, Talanta, 80, 422–427, 10.1016/j.talanta.2009.03.019, 2009.

Nakamichi, S., Kawaguchi, Y., Fukuda, H., Enami, S., Hashimoto, S., Kawasaki, M., Umekawa, T., Morino, I., Suto, H., and Inoue, G.: Buffer-gas pressure broadening for the (3 00 1)III ← (0 0 0) band of CO2 measured with continuous-wave cavity ring-down spectroscopy, Phys. Chem. Chem. Phys., 8, 364–368, 10.1039/b511772k, 2006.

Nara, H., Tanimoto, H., Tohjima, Y., Mukai, H., Nojiri, Y., Katsumata, K., and Rella, C. W.: Effect of air composition (N2, O2, Ar, and H2O) on CO2 and CH4 measurement by wavelength-scanned cavity ring-down spectroscopy: calibration and measurement strategy, Atmos. Meas. Tech., 5, 2689–2701, 10.5194/amt-5-2689-2012, 2012.

O'Shea, S. J., Bauguitte, S. J.-B., Gallagher, M. W., Lowry, D., and Percival, C. J.: Development of a cavity-enhanced absorption spectrometer for airborne measurements of CH4 and CO2, Atmos. Meas. Tech., 6, 1095–1109, 10.5194/amt-6-1095-2013, 2013.

Pouchet, I., Zéninari, V., Parvitte, B., and Durry, G.: Diode laser spectroscopy of CO2 in the 1.6 µm region for the in situ sensing of the middle atmosphere, J. Quant. Spectrosc. Ra., 83, 619–628, 10.1016/S0022-4073(03)00108-0, 2004.

Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K., Tignor, M. M. B., and Miller, H. L. (Eds.): Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, New York, N. Y., 2007.

Rella, C. W., Chen, H., Andrews, A. E., Filges, A., Gerbig, C., Hatakka, J., Karion, A., Miles, N. L., Richardson, S. J., Steinbacher, M., Sweeney, C., Wastine, B., and Zellweger, C.: High accuracy measurements of dry mole fractions of carbon dioxide and methane in humid air, Atmos. Meas. Tech., 6, 837–860, 10.5194/amt-6-837-2013, 2013.

Rothman, L. S., Jacquemart, D., Barbe, A., Benner, D. C., Birk, M., Brown, L. R., Carleer, M. R., Chackerian Jr., C., Chance, K., Coudert, L. H., Dana, V., Devi, V. M., Flaud, J.-M., Gamache, R. R., Goldman, A., Hartmann, J.-M., Jucks, K. W., Maki, A. G., Mandin, J.-Y., Massie, S. T., Orphal, J., Perrin, A., Rinsland, C. P., Smith, M. A. H., Tennyson, J., Tolchenov, R. N., Toth, R. A., Vander Auwera, J., Varanasi, P., and Wagner, G.: The HITRAN 2004 molecular spectroscopic database, J. Quant. Spectrosc. Ra., 96, 139–204, 10.1016/j.jqsrt.2004.10.008, 2005.

Schulze, E. D., Luyssaert, S., Ciais, P., Freibauer, A., Janssens, I. A., Soussana, J. F., Smith, P., Grace, J., Levin, I., Thiruchittampalam, B., Heimann, M., Dolman, A. J., Valentini, R., Bousquet, P., Peylin, P., Peters, W., Rödenbeck, C., Etiope, G., Vuichard, N., Wattenbach, M., Nabuurs, G. J., Poussi, Z., Nieschulze, J., Gash, J. H., and the CarboEurope Team: Importance of methane and nitrous oxide for Europe's terrestrial greenhouse-gas balance, Nat. Geosci., 2, 842–850, 10.1038/ngeo686, 2009.

Tohjima, Y., Katsumata, K., Morino, I., Mukai, H., Machida, T., Akama, I., Amari, Y., and Tsunogai, U. Theoretical and experimental evaluation of the isotope effect of NDIR analyzer on atmospheric CO2 measurement, J. Geophys. Res., 114, D13302, 10.1029/2009JD011734, 2009.

van der Laan, S., Neubert, R. E. M., and Meijer, H. A. J.: A single gas chromatograph for accurate atmospheric mixing ratio measurements of CO2, CH4, N2O, SF6 and CO, Atmos. Meas. Tech., 2, 549–559, 10.5194/amt-2-549-2009, 2009.

Varghese, P. L. and Hanson, R. K.: Collisional narrowing effects on spectral line shapes measured at high resolution, Appl. Optics, 23, 2376, 10.1364/AO.23.002376, 1984.

Waldén, J. (Ed.): Metrology of Gaseous Air Pollutants, No. 79 (Finnish Meteorological Institute Contributions), Finnish Meteorological Institute, Finland, 2009.

Wessel, R. M., van der Veen, A. M. H., Ziel, P. R., Steele, P., Langenfelds, R., van der Schoot, M., Smeulders, D., Besley, L., da Cunha, V. S., Zhou, Z., Qiao, H., Heine, H. J., Martin, B., Macé, T. P., Gupta, K., di Meane, E. A., Sega, M., Rolle, F., Maruyama, M., Kato, K., Matsumoto, N., Kim, J. S., Moon, D. M., Lee, J. B., Murillo, F. R., Nambo, C. R., Caballero, V. M. S., de Jesús Avila Salas, M., Castorena, A. P., Konopelko, L. A., Kustikov, Y. A., Kolobova, A. V., Pankratov, V. V., Efremova, O. V., Musil, S., Chromek, F., Valkova, M., Milton, M. J. T., Vargha, G., Guenther, F., Miller, W. R., Botha, A., Tshilongo, J., Mokgoro, I. S., and Leshabane, N.: International comparison CCQM-K52: Carbon dioxide in synthetic air, Metrologia, 45, 08011, 10.1088/0026-1394/45/1A/08011, 2008.

Zhao, C. L., Tans, P. P., and Thoning, K. W.: A high precision manometric system for absolute calibrations of CO2 in dry air, J. Geophys. Res., 102, 5885–5894, 10.1029/96JD03764, 1997.

</app></app-group></back> </article>