Recommended Second Virial Coefficients for Nitrogen and Oxygen
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.
Kai: We're looking at the paper: "Recommended Second Virial Coefficients for Nitrogen and Oxygen".
Mira: Here's the gist of the paper. The paper provides recommended values for the second virial coefficient, B(T), and its uncertainty for molecular nitrogen and oxygen based on tuning previously published state-of-the-art ab initio pair potentials to match high-accuracy experimental data.
Kai: This work offers updated recommendations for B(T) covering a wider temperature range than previous methods, which is essential for applications like flow metering and pressure metrology.
Mira: Tuning Methodology The tuning of the pair potentials involved reanalyzing experimental density data to refine the potential energy surfaces (PES) for both molecules.
Kai: For nitrogen, this involved a careful reanalysis of isotherms from Nowak et al.
Paper summary: Kai: So, this paper, "Recommended Second Virial Coefficients for Nitrogen and Oxygen," it's about giving us updated values for the second virial coefficient, B of T, and its uncertainty for nitrogen and oxygen.
Mira: It claims they did this by tuning existing state-of-the-art ab initio pair potentials to match high-accuracy experimental data.
Kai: What’s the big deal here?
Mira: Well, these updated recommendations cover a much wider temperature range than what previous methods used, which is really important for things like flow metering and pressure metrology.
Lev: From an engineering side, that wider range is key because it means you can use this information reliably in more real-world industrial settings.
Kai: Right, so the core of the work isn't just new numbers; it's about refining the underlying physics models for these molecules.
Mira: Exactly. They used specific ways to reanalyze experimental density data to refine those potential energy surfaces, or PES, for both nitrogen and oxygen.
Kai: That sounds like a lot of heavy lifting on the experimental side first just to get the tuning right.
Lev: Yeah, they did a careful reanalysis of isotherms from Nowak et al. and McLinden and Lösch-Will for nitrogen, using higher-order virial coefficients to get B and C with much smaller uncertainty than before.
Mira: And for oxygen, they used three different analytical representations of the interaction surfaces to match new experimental B of T values from Egan.
Kai: So, they have these specific tuning strategies for each molecule.
Lev: For nitrogen, they scaled a factor to zero point four seven two after incorporating new data which was heavily influenced by the values in Table one.
Mira: That scaling process seems like it's how they narrowed down the uncertainty significantly compared to prior analyses.
Kai: And then for the calculation itself, they use path-integral Monte Carlo, or PIMC, to make sure they rigorously account for those nuclear quantum effects.
Lev: They used thirty-two independent samples and one hundred six integration steps to get an accuracy on the order of ten ppm or less across the widest possible temperature range.
Mira: That level of detail in the calculation method really backs up the confidence in their results.
Kai: So, they've done this detailed work to produce recommended B values for nitrogen and oxygen.
Mira: And these recommendations are what they hope will reduce uncertainty for thermodynamic properties in applications like flow metering and pressure metrology.
Conclusion: Lev: The paper "Recommended Second Virial Coefficients for Nitrogen and Oxygen" concludes by giving us these recommended B of T functions, which are represented by a correlation form, B of T equals sum i from one to imax a B i of T star b B i.
Kai: That means they've provided us with specific mathematical functions you can use for nitrogen and oxygen across that temperature range.
Mira: And they say these recommendations are expected to lower the uncertainty of thermodynamic properties for these gases in applications such as flow metering and pressure metrology.
Lev: They're providing a foundation for a revised recommendation for the second virial coefficient of dry air, Baa of T, too.
Kai: So, it’s not just about nitrogen and oxygen in isolation; it opens up the door to updating those old recommendations for air as a whole.
Mira: The authors conclude that ab initio pair potentials for molecular nitrogen and oxygen have been tuned based on selected high-accuracy experimental data.
Lev: They recommend these B of T functions are usable from twenty to three thousand Kelvin for nitrogen, and twenty to two thousand Kelvin for oxygen.
Kai: That's a pretty broad range, which is what makes this useful.
Lev: The important part is that these recommendations will form a foundation for a revised recommendation for the second virial coefficient of dry air, Baa of T.
Mira: And they also mention that the other important contributor to this quantity is the cross second virial coefficient between nitrogen and oxygen.
Kai: They're also working on deriving a state-of-the-art PES for this pair now.
Mira: So, while they provide these B functions, they are also laying groundwork for even better models for the interaction between nitrogen and oxygen molecules themselves.
Lev: And that future surface can be tuned to data from Egan's refractometry measurements, which gave values of B of T with low uncertainties for both binary N2/O2 mixtures and natural air.
Kai: It sounds like they are setting up a path toward even more precise modeling of these gases in real applications.
Robert Hellmann, Giovanni Garberoglio, Allan H. Harvey
Institut für Thermodynamik, Helmut-Schmidt-Universität / Universiät der Bundeswehr Hamburg · European Centre for Theoretical Studies in Nuclear Physics and Related Areas · National Institute of Standards and Technology
physics.chem-ph
Submitted: 2026-07-27
Updated: 2026-07-27
Comments: 25 pages, 3 figures
Journal ref: J. Phys. Chem. Ref. Data 55, 043101 (2026)
DOI: 10.1063/5.0354088
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 66/100
The gist: The paper provides recommended values for the second virial coefficient, B(T), and its uncertainty for molecular nitrogen and oxygen based on tuning previously published state-of-the-art ab initio…
Key concepts
- Second Virial Coefficient (B(T))
- This coefficient describes how much molecules interact with each other as they are compressed or mixed. It is a key parameter used in equations of state to accurately predict the behavior of gases under various conditions, such as flow rates and pressures.
- Pair Potentials
- These are mathematical functions that model the forces between two specific molecules (like nitrogen and oxygen). The researchers tuned existing potentials by reanalyzing experimental density data to make these models more accurate for predicting molecular interactions.
- Path-Integral Monte Carlo (PIMC)
- This computational method is used to calculate B(T) while rigorously accounting for nuclear quantum effects. It treats the nuclei as quantum particles using a classical mapping with replicas, allowing for highly accurate calculations across a broad temperature range.
Terminology
Summary
The paper provides recommended values for the second virial coefficient, B(T), and its uncertainty for molecular nitrogen and oxygen based on tuning previously published state-of-the-art ab initio pair potentials to match high-accuracy experimental data. p. 0 This work offers updated recommendations for B(T) covering a wider temperature range than previous methods, which is essential for applications like flow metering and pressure metrology. p. 2
Tuning Methodology
The tuning of the pair potentials involved reanalyzing experimental density data to refine the potential energy surfaces (PES) for both molecules. For nitrogen, this involved a careful reanalysis of isotherms from Nowak et al. p. 4 and McLinden and Lösch-Will, constrained by higher-order virial coefficients to produce B and C with significantly smaller uncertainty than previous analyses. p. 4 For oxygen, the tuning utilized three analytical representations of the three possible interaction surfaces, matching new experimental B(T) values from Egan. p. 9 The strategy for nitrogen involved scaling a previously used factor to 0. p. 7472 after incorporating new data, which was dominated by the values in Table 1. p. 7 p. 7
Calculation of B(T)
The second virial coefficient B(T) is calculated using the path-integral Monte Carlo (PIMC) approach to rigorously account for nuclear quantum effects. This method uses a classical mapping with P replicas, where the final expression for B(T) is equivalent to the classical expression for a pair of ring polymers of P beads. p. 0 The calculation details specify using 32 independent samples and 106 integration steps to achieve an accuracy on the order of 10 ppm or less for the widest possible temperature range. p. 3
Results and Comparison
The calculated B values agree with all B values derived from the reanalyses of Nowak et al. and McLinden and Lösch-Will, except for one value at 480 K for McLinden and Lösch-Will, within the expanded (k = 2) experimental uncertainties. p. 7 The results show that most of the B values derived by Duschek et al. p. 4 and originally by Nowak et al. p. 3 are significantly more negative than the new reference values, attributed to an overestimation of the third virial coefficient C. p. 8 For oxygen, agreement with Egan's data is within ±0.05 cm3 mol−1. p. 8
Uncertainty Estimation
The uncertainties for nitrogen B(T) are estimated at 0.05 cm3 mol−1 in the temperature range 150–500 K, increasing linearly to 0. p. 53 cm3 mol−1 at 3000 K. p. 8 The uncertainty estimate for oxygen B(T) is 0. p. 71 cm3 mol−1 in the range 300–400 K, increasing linearly to 0. p. 83 cm3 mol−1 at 2000 K. p. 8
Recommended Values
The recommended B(T) functions are represented by a correlation of the form B(T) = sum i=1 to imax aB,i(T∗) bB,i, where imax is 10 for N2 and imax is 11 for O2. These recommended values are expected to reduce the uncertainty of thermodynamic properties for these gases in applications such as flow metering and pressure metrology. p. 2 The work also provides a foundation for a revised recommendation for the second virial coefficient of (dry) air, Baa(T). p. 3
Conclusions
The paper concludes that ab initio pair potentials for molecular nitrogen and oxygen have been tuned based on selected high-accuracy experimental data. The recommended B(T) functions are provided for use from 20–3000 K for N2 and 20–2000 K for O2. p. 9 These recommendations will form a foundation for a revised recommendation for the second virial coefficient of (dry) air, Baa(T). p. 3 The recommended B(T) given here will reduce the uncertainty of thermodynamic properties for these gases in applications such as those mentioned in Sec. p. 2 1, such as flow metering and pressure metrology. p. 2 p. 2 The other important contributor to this quantity is the cross second virial coefficient between N2 and O2. p. 3 The project for deriving a state-of-the-art PES for this pair is underway. p. 4 The resulting surface can be tuned to data of Egan, whose refractometry measurements resulted in values of B(T) with low uncertainties for both binary N2/O2 mixtures and natural air. p. 4 The calculation of Baa will also require accounting for the roughly 1 % of argon in air. p. 2 The paper provides state-of-the-art information for this important thermodynamic quantity. p. 7
Pruning redundant reasoning tokens
The tuning process involved scaling a previously used factor to 0.472 after incorporating new data, which was dominated by the values in Table 1. The calculation of B(T) was performed using the path-integral Monte Carlo (PIMC) approach to rigorously account for nuclear quantum effects. p. 0 The final expression for B(T) is equivalent to the classical expression for a pair of ring polymers of P beads. p. 6
How the training signal is built
The representation for B(T) and its uncertainty in Sec. 6 are recommended for use from 20–3000 K for N2 and 20–2000 K for O2. p. 0 The individual B values were weighted in the fits by the squared inverses of the statistical uncertainties resulting from the PIMC calculations. p. 7
Where the gains come from
The recommended B(T) given here will reduce the uncertainty of thermodynamic properties for these gases in applications such as those mentioned in Sec. 1, such as flow metering and pressure metrology. The work also provides low-density boundary conditions that can be used for future improvement of the reference equations of state for these fluids. p. 2 The current reference EOS for nitrogen is over 25 years old while that for oxygen is over 40 years old. p. 2 The original papers reporting PES for these systems reported calculations for the low-density limits of the viscosity and thermal conductivity. p. 3 The viscosity and thermal conductivity from the earlier work can be used with no changes. p. 5
The uncertainty contributions for B and C due to the uncertainty in D were obtained as the differences between the values obtained with and without adding 2000 cm9 mol−3 to D. The third uncertainty contribution is related to how stable the B and C values are within the abovementioned plateaus. p. 6
The uncertainty estimates for B(T) are based on the uncertainty of the experimental data used to tune the potential (in the temperature range of those data) and on our estimate of the uncertainty at other temperatures due to the uncertainties of the tuned potentials. The procedure for estimating the expanded (k = 2) uncertainty of B(T) for oxygen is similar. p. 4 The uncertainty estimates are based on the difference between values of B obtained with and without tuning, rescaled such that at 300 K a difference of 0. p. 61 cm3 mol−1 is obtained. p. 8 The linearly increasing expanded (k = 2) uncertainty estimates above these temperature ranges can be written for nitrogen as U [B(T)] = 0. p. 101T∗, and for oxygen as U [B(T)] = 0. p. 005+0.0125T∗.
The agreement with the EOS values above about 120 K is remarkably good despite the fact that the data of Egan were not available when the EOS was developed. The B values obtained with the QFH approximation again substantiate the PIMC calculations and begin to deviate significantly from the PIMC values only below about 50 K. p. 9 The comparison with the two data sets, which are the only ones available for βa, is somewhat disappointing. p. 3 The lowtemperature systematic deviations of Ewing and Trusler’s values and the deviation of their datum at the highest temperature cannot reasonably be explained by a deficiency of our correlation. The comparison with further experimental data sets, which are only available for oxygen, shows that Weber but are not consistent with the data of Egan within the latter’s expanded uncertainty. p. 9
The calculation of B(T) was performed by a Monte Carlo approach, referred to as path-integral Monte Carlo (PIMC), using a parallel implementation of the VEGAS39 integration routine. The number of independent calculations for each spin state was at least 64 for T ≥ 150 K, progressively increased to an average of 96 down to T = 80 K and then progressively reduced to 10 at lower temperatures. p. 5 The number of independent samples was set at least 32 for evaluating the average and 10 6 integration steps, cutting off the interaction after 200 nm. p. 5
The parameters aB,i and bB,i are listed in Table 2. The correlations only from 20 K upwards. p. 2 p. 2 p. 2 The correlations only from 20 K upwards. Above the highest considered temperatures, i. p. 3 p. 3e., 3000 K for N2 and 2000 K for O2, the extrapolation is also physically reasonable, with B(T) remaining bounded due to all exponents bB,i being negative. p. 6 p. 6 The correlations only from 20 K upwards. Above the highest considered temperatures, i.e., 3000 K for N2 and 2000 K for O2, the extrapolation is also physically reasonable, with B(T) remaining bounded due to all exponents bB,i being negative.
The uncertainty of our recommended B(T) is based on the uncertainty of the experimental data used to tune the potential (in the temperature range of those data) and on our estimate of the uncertainty at other temperatures due to the uncertainties of the tuned potentials. Both of these contributions are relatively negligible and will therefore not be considered in the following. The uncertainty estimates for B(T) are based on the observation that the possible error in the PES is constrained by the 0. p. 105 cm3 mol−1 expanded uncertainty at 150 K. p. 9
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved prediction of second virial coefficient B(T) for nitrogen and oxygen using recommended functions from 20 to 3000 K for N2 and 20 to 2000 K for O2, which
provide state-of-the-art information for this important thermodynamic quantity.
-
Enhanced accuracy in flow metering and pressure metrology by reducing the uncertainty of thermodynamic properties derived from the recommended B(T) values.
-
Development of low-density boundary conditions for reference equations of state that are over 25 years old for nitrogen and over 40 years old for oxygen, improving their reliability.
-
Calculation of cross second virial coefficients between N2 and O2 by tuning a state-of-the-art pair potential to match recent experimental data, which is noted as a
topic of future work.
-
Improved modeling of humid air by providing recommended B(T) for pure water and the water–air cross coefficient Baw, which can be improved with state-of-the-art ab initio pair potentials.
Sources
Related papers
- A System-Independent Metadynamics Strategy for Reactive Training Data: Application to Gas-Phase Organic Reactions
- A Machine Learning Model for the Chemistry of a Solvated Electron
- Compression of virtual spaces in transcorrelated methods via singular value decomposition: application to the G2 set
- Charge Transfer with a Spin. II: A Framework for Diabatization which Localizes Charge and Spin
- Diabatic Seam Space Sampling for Hydrogen Tunneling Systems with Nuclear-Electronic Orbital Theory
- Augmented Roothaan-Hall Hessian Applied to Spin-Restricted Open-Shell Density-Functional Theory