| 28 Sep 2026
High-precision gravimetric primary standards for atmospheric Ar/N2 measurements and redetermination of atmospheric Ar abundance
Abstract. Accurate determination of atmospheric Ar/N2 ratios requires primary standard gases with well-characterized uncertainties. However, an initial preparation approach based on the transfer of premixed gases introduced systematic deviations caused by transfer-induced composition changes, which limited the achievable accuracy. In this study, we developed an improved gravimetric preparation method for highly precise Ar standards (HPAs), in which CO2, Ar, O2, and N2 are introduced independently from source gases. This approach eliminates transfer-induced compositional changes and allows the Ar amount to be determined directly with high precision. The resulting uncertainty in Ar/N2 was reduced to approximately 5 per meg, corresponding to an Ar molar fraction of 0.05 µmol mol−1. Validation confirmed that the gravimetric and measured values for CO2, O2/N2, and Ar/N2 agreed within their uncertainties. Using the validated HPAs, the absolute Ar/N2 ratio of the AIST reference air was redetermined as 0.01195567 ± 0.000000011, a 8-fold reduction in uncertainty relative to the previous value. Application to atmospheric observations yielded an absolute Ar/N2 ratio of 0.01195497 ± 0.00000013 and an Ar mole fraction of 9335.71 ± 0.10 µmol mol−1 in 2021. The numbers following the symbol ± represent the standard uncertainty. These results establish a metrologically traceable framework for high-precision atmospheric Ar measurements.
General comments
The manuscript describes the development of SI-traceable Ar standard gas mixtures to support accurate long-term monitoring of the atmospheric Ar/N2 ratio. The authors developed a new weighing system and evaluated the contributions of potential sources of uncertainty in a modifed gravimetric preparation procedure for Ar standard gas mixtures. The results show that the uncertainties of the newly developed SI-traceable Ar standard gas mixtures are smaller than the measurement uncertainty. The Ar standard gas mixtures are therefore suitable for use as calibration gas mixtures for monitoring small variations in the atmospheric Ar/N2 ratio.
Overall, the manuscript is well organized and presents a sound analysis of the measurement data. The manuscript can be considered for publication after the authors adequately address the comments below.
Specific comments
Throughout the manuscript: “molar fraction” and “mole fraction” should be replaced with “amount fraction” since “molar” is an adjective related to the mole as a unit rather than the name of the quantity.
Page 2 line 42: What is meant by metrologically traceable calibration scale? Please clarify whether this refers to traceablility to a specific calibration scale or to the SI.
Page 4 line 101: the unit for purity should be expressed as cmol/mol rather than %
Page 8 line 167-171: The statement that “the isotopic compositions of the source gases were assumed to correspond to those of the components in the HPAs and the isotopic composition of the source gases can therefore be inferred from the measured isotope ratios” is valid only if no isotope fractionation occurred during the transfer of Ar or CO2 from the pure source cylinders into the 0.9 L cylinders. The same consideration applies to O2 and N2.
Table 1: the values should be checked again. For example, the O2/N2 ratio for CPB12763 is 0.2679049 in Table 1, but it is calculated as 0.2679044 from the listed values.
Page 14 line 320-325. For the measurement of Ar impurity in pure nitrogen by GC-TCD, where gases are introduced using an MFC, possible bias arising from the matrix difference should be considered or experimentally negated (or demonstrated to be negligible). There is a substantial matrix difference between the sample (N2)and the calibration gas mixtures (He), which could affect accurate quantification of Ar.
Page 17 line 398-399: “the sample cylinder was cooled to 1-3 K below the reference cylinder to reproduce the temperature difference…. “ this sentence is a bit awkward. this sentence should be revised to state that the temperature of the sample cylinder was 1-3 K lower than that of the reference cylinder.
Page 20, CO2 validation: the Picarro CRDS (G2301) measures 12C16O16O rather than total CO2. Therefore, for a proper comparison, the gravimetrically assigned values should correspond to the 12C16O16O amount fractions. And How was the CRDS calibrated for these measurements?
Page 23, Effect of compositional changes during CO2/Ar transfer: How was the possible adsorption loss of CO2 onto the internal surface of the cylinder considered? CO2 can adsorb onto cylinder surfaces, unlike Ar, O2, or N2, and this could affect its amount fraction.
Page 23, line 555-556: In the statement “…provides a basis for improving consistency among atmospheric observation networks through traceable Ar/N2 calibration.” I assume that the authors mean SI traceable Ar/N2 calibration. Please clarify.
Page 25, line 612-613: I assume that by “traceable absolute” the authors mean “SI traceable”. This terminology should be clarified.
As I understand it, measurements made against a scale reference (or standard) are traceable to that scale, whereas measurements made against an SI reference (or standard) are traceable the SI. The expression “metrologically scale traceable” should therefore be avoided. The aim of this study appears to be the development of SI traceable primary reference (standard) gases to support accurate long-term atmospheric Ar/N2 monitoring. It would therefore be helpful to distinguish clearly between scale traceability and SI traceability throughout the manuscript, rather than using expression such as “traceable” or “traceable absolute”.