Joële Viallon

A study of systematic biases and measurement uncertainties in ozone mole fraction measurements with the NIST Standard Reference Photometer

Sources of bias in the National Institute of Standards and Technology ozone Standard Reference Photometer (SRP) maintained by the Bureau International des Poids et Mesures have been investigated. A relative bias of ?0.4% in the ozone mole fraction measurement caused by a temperature gradient in the gas cells of the instrument was characterized and corrected for in a modified version of the instrument. A second relative bias of +0.5% due to the multiple reflections of light within the gas cells was also corrected. The Guide to the Expression of Uncertainty in Measurement approach was used to develop an uncertainty budget for the modified SRP, including a relative value for the ozone absorption cross-section uncertainty of 2.1% (k = 2). The measurement uncertainty for the bias-corrected SRPs is enlarged compared with earlier studies, but their comparability improved.

Influence of correlation on the assessment of measurement result compatibility over a dynamic range

Comparison of measurement results obtained by different measuring equipment on the same reference object is a key element in equipment qualification. The paper discusses an approach for an integrated assessment of a series of comparison measurements taken on more than one (i.e. many) reference object(s). The approach accounts for the fact that measurements taken by one (and also by the other) instrument on a series of ideally independent objects are nevertheless partially correlated due to common uncertainty contributions arising from the measurement process. The approach is exemplified for international comparisons using a (travelling) transfer instrument for the comparison of national instrument standards with a reference standard over a range of measured ozone mole fractions.

Final report on the on-going key comparison BIPM.QM-K1: Ozone at ambient level, comparison with NMISA, 2008

As part of the on-going key comparison BIPM.QM-K1, a comparison has been performed between the ozone national standard of the National Metrology Institute of South Africa (NMISA) and the common reference standard of the key comparison, maintained by the Bureau International des Poids et Mesures (BIPM). The instruments have been compared over a nominal ozone mole fraction range of 0 nmol/mol to 500 nmol/mol.

Methane Standards Made in Whole and Synthetic Air Compared by Cavity Ring Down Spectroscopy and Gas Chromatography with Flame Ionization Detection for Atmospheric Monitoring Applications

There is evidence that the use of whole air versus synthetic air can bias measurement results when analyzing atmospheric samples for methane (CH4) and carbon dioxide (CO2). Gas chromatography with flame ionization detection (GC-FID) and wavelength scanned-cavity ring down spectroscopy (WS-CRDS) were used to compare CH4 standards produced with whole air or synthetic air as the matrix over the mole fraction range of 1600-2100 nmol mol(-1). GC-FID measurements were performed by including ratios to a stable control cylinder, obtaining a typical relative standard measurement uncertainty of 0.025%. CRDS measurements were performed using the same protocol and also with no interruption for a limited time period without use of a control cylinder, obtaining relative standard uncertainties of 0.031% and 0.015%, respectively. This measurement procedure was subsequently used for an international comparison, in which three pairs of whole air standards were compared with five pairs of synthetic air standards (two each from eight different laboratories). The variation from the reference value for the whole air standards was determined to be 2.07 nmol mol(-1) (average standard deviation) and that of synthetic air standards was 1.37 nmol mol(-1) (average standard deviation). All but one standard agreed with the reference value within the stated uncertainty. No significant difference in performance was observed between standards made from synthetic air or whole air, and the accuracy of both types of standards was limited only by the ability to measure trace CH4 levels in the matrix gases used to produce the standards.

Highly accurate nitrogen dioxide (NO2) in nitrogen standards based on permeation.

The development and operation of a highly accurate primary gas facility for the dynamic production of mixtures of nitrogen dioxide (NO(2)) in nitrogen (N(2)) based on continuous weighing of a permeation tube and accurate impurity quantification and correction of the gas mixtures using Fourier transform infrared spectroscopy (FT-IR) is described. NO(2) gas mixtures in the range of 5 μmol mol(-1) to 15 μmol mol(-1) with a standard relative uncertainty of 0.4% can be produced with this facility. To achieve an uncertainty at this level, significant efforts were made to reduce, identify and quantify potential impurities present in the gas mixtures, such as nitric acid (HNO(3)). A complete uncertainty budget, based on the analysis of the performance of the facility, including the use of a FT-IR spectrometer and a nondispersive UV analyzer as analytical techniques, is presented in this work. The mixtures produced by this facility were validated and then selected to provide reference values for an international comparison of the Consultative Committee for Amount of Substance (CCQM), number CCQM-K74, (1) which was designed to evaluate the consistency of primary NO(2) gas standards from 17 National Metrology Institutes.

Accurate Fourier transform infrared (FT-IR) spectroscopy measurements of nitrogen dioxide (NO2) and nitric acid (HNO3) calibrated with synthetic spectra.

A novel method for determining the accuracy of laboratory-based measurements of nitrogen dioxide (NO2) and nitric acid (HNO3) mole fractions using Fourier transform infrared (FT-IR) spectroscopy 1 cm−1 resolution instruments calibrated with synthetic spectra has been developed. The traceability of these measurement results is to the reference line strength data contained within the high-resolution transmission molecular absorption (HITRAN) database. Incorporating a proper estimate of the uncertainty of this data into the measurement results will ensure that the SI traceable values are encompassed within the uncertainty of the measurement results. The major contributors to the uncertainties of the results are, in descending order of importance, the uncertainty in the line strength values (HITRAN 2004), the uncertainty attributed to the generation of reference spectra (including knowledge of the optical path length of the FT-IR gas cell), and temperature measurements of the gas. The stability of the FT-IR instrument itself is only a minor contributor to the overall uncertainty of the measurements. FT-IR measurements of NO2 mole fractions at nominal values of 10 μmol mol−1 calibrated with synthetic spectra lead to standard uncertainties of 0.34 μmol mol−1 (3.4% relative). In contrast, calibration of the FT-IR instrument with SI traceable gas standards generated by a dynamic weighing system resulted in measurements results with standard uncertainties of 0.04 μmol mol−1 (0.4% relative). When comparing the consistency of measurement results based on the synthetic calibration method against those obtained by calibrations with SI traceable gas standards, the existence of a potential bias of ∼5% was observed, although this was within the stated uncertainties of the results. The FT-IR measurements of HNO3 mole fractions at nominal values of 200 nmol mol−1 calibrated with synthetic spectra resulted in values with standard uncertainties of 23 nmol mol−1 (11% relative) with the dominating uncertainty in this case arising from the stabilization of the mole fraction value within the FT-IR gas cell.

International comparison CCQM-K82: methane in air at ambient level (1800 to 2200) nmol/mol

The CCQM-K82 comparison was designed to evaluate the degrees of equivalence of NMI capabilities for methane in air primary reference mixtures in the range (1800 to 2200) nmol/mol. The balance gas for the standards was either scrubbed dry real air or synthetic air. CH4 in air standards have been produced by a number of laboratories for many years, with more recent developments focused on standards at atmospheric measurement concentrations and aimed at obtaining agreement between independently produced standards. A comparison of the differences in primary gas standards for methane in air was previously performed in 2003 (CCQM-P41 Greenhouse gases. 1 and 2) with a standard deviation of results around the reference value of 30 nmol/mol and 10 nmol/mol for a more limited set of standards. This can be contrasted with the level of agreement required from field laboratories routinely measuring atmospheric methane levels, set by Data Quality Objectives (DQO) established by the World Meteorological Organization (WMO) to reflect the scientifically desirable level of compatibility for CH4 measurements at the global scale, currently set at 2 nmol/mol (1 sigma). The measurements of this key comparison took place from May 2012 to June 2012. Eight laboratories took part in this comparison coordinated by the BIPM and NIST. Key comparison reference values were calculated based on Cavity Ring Down Spectroscopy Measurements performed at the BIPM, combined with participants gravimetric values to identify a consistent set of standards. Regression analysis allowed predicted values for each standard to be calculated which acted as the KCRVs. In this comparison reported standard uncertainties by participants ranged from 0.50 nmol/mol to 2.4 nmol/mol and the uncertainties of individual KCRVs ranged from 0.68 nmol/mol to 0.71 nmol/mol. The standard deviation of the ensemble of standards about the KCRV value was 1.70 nmol/mol. This represents a greater than tenfold improvement in the level of compatibility of methane in air standards compared to that demonstrated in 2003. Further improvements in the compatibility of standards will require improved methods and uncertainties for the measurement of trace level methane in balance gases. Main text. To reach the main text of this paper, click on Final Report. Note that this text is that which appears in Appendix B of the BIPM key comparison database kcdb.bipm.org/. The final report has been peer-reviewed and approved for publication by CCQM, according to the provisions of the CIPM Mutual Recognition Arrangement (CIPM MRA).

Calibration Strategies for FT-IR and Other Isotope Ratio Infrared Spectrometer Instruments for Accurate δ13C and δ18O Measurements of CO2 in Air

This paper describes calibration strategies in laboratory conditions that can be applied to ensure accurate measurements of the isotopic composition of the CO2 in ultradry air, expressed as δ13C and δ18O on the VPDB scale, with either FT-IR (in this case a Vertex 70 V (Bruker)) or an isotope ratio infrared spectrometer (IRIS) (in this case a Delta Ray (Thermo Fisher Scientific)). In the case of FT-IR a novel methodology using only two standards of CO2 in air with different mole fractions but identical isotopic composition was demonstrated to be highly accurate for measurements of δ13C and δ18O with standard uncertainties of 0.09‰ and 1.03‰, respectively, at a nominal CO2 mole fraction of 400 μmol mol-1 in air. In the case of the IRIS system, we demonstrate that the use of two standards of CO2 in air of known but differing δ13C and δ18O isotopic composition allows standard uncertainties of 0.18‰ and 0.48‰ to be achieved for δ13C and δ18O measurements, respectively. The calibration strategies were validated using a set of five traceable primary reference gas mixtures. These standards, produced with whole air or synthetic air covered the mole fraction range of (378-420) μmol mol-1 and were prepared and/or value assigned either by the National Institute of Standards and Technology (NIST) or the National Physical Laboratory (NPL). The standards were prepared using pure CO2 obtained from different sources, namely, combustion; Northern Continental and Southern Oceanic Air and a gas well source, with δ13C values ranging between -35‰ and -1‰. The isotopic composition of all standards was value assigned at the Max Planck Institute for Biogeochemistry Jena (MPI-Jena).

Final report, on-going key comparison BIPM.QM-K1: Ozone at ambient level, comparison with FMI, 2007

As part of the on-going key comparison BIPM.QM-K1, a comparison has been performed between the ozone national standard of the Czech Hydrometeorological Institute (CHMI) and the common reference standard of the key comparison, maintained by the Bureau International des Poids et Mesures (BIPM). The instruments have been compared over a nominal ozone mole fraction range of 0 nmol/mol to 500 nmol/mol.

Temperature measurement and optical path-length bias improvement modifications to National Institute of Standards and Technology ozone reference standards

Ambient ozone measurements in the United States and many other countries are traceable to a National Institute of Standards and Technology Standard Reference Photometer (NIST SRP). The NIST SRP serves as the highest level ozone reference standard in the United States, with NIST SRPs located at NIST and at many U.S. Environmental Protection Agency (EPA) laboratories. The International Bureau of Weights and Measures (BIPM) maintains a NIST SRP as the reference standard for international measurement comparability through the International Committee of Weights and Measures (CIPM). In total, there are currently NIST SRPs located in 20 countries for use as an ozone reference standard. A detailed examination of the NIST SRP by the BIPM and NIST has revealed a temperature gradient and optical path-length bias inherent in all NIST SRPs. A temperature gradient along the absorption cells causes incorrect temperature measurements by as much as 2 °C. Additionally, the temperature probe used for temperature measurements was found to inaccurately measure the temperature of the sample gas due to a self-heating effect. Multiple internal reflections within the absorption cells produce an actual path length longer than the measured fixed length used in the calculations for ozone mole fractions. Reflections from optical filters located at the exit of the absorption cells add to this effect. Because all NIST SRPs are essentially identical, the temperature and path-length biases exist on all units by varying amounts dependent upon instrument settings, laboratory conditions, and absorption cell window alignment. This paper will discuss the cause of, and physical modifications for, reducing these measurement biases in NIST SRPs. Results from actual NIST SRP bias upgrades quantifying the effects of these measurement biases on ozone measurements are summarized. Implications: NIST SRPs are maintained in laboratories around the world underpinning ozone measurement calibration and traceability within and between countries. The work described in this paper quantifies and shows the reduction of instrument biases in NIST SRPs improving their overall agreement. This improved agreement in all NIST SRPs provides a more stable baseline for ozone measurements worldwide.