Comments on temperature calibration and uncertainty estimate of the vibrating tube densimeter operated at atmospheric pressure

• A calibration procedure of the atmospheric vibrating tube densimeter is introduced for temperatures between 273 and 363 K. • An uncertainty analysis is introduced based on the accurate measurements with ultrapure water and dry air. • The calibration was verified on a series of reference liquids with viscosity ranging from 0.3 mPa s to 600 mPa s. • Standard uncertainty of liquid density was determined between 0.030 kg m −3 and 0.14 kg m −3 for typical and highly viscous liquids, respectively. Vibrating tube densimeter (VTD) is a popular instrument providing a quick and accurate measurement of density for large variety of liquids and gases. Sensitive commercial instruments have resolution of down to 0.001 kg · m - 3 and their accuracy is declared to reach 0.007 kg · m - 3 under ideal conditions at temperatures close to ambient. However, the uncertainty of the obtained data can be significantly higher when it comes to measurements over wider temperature range. We present calibration procedure for a commercial borosilicate glass VTD (namely Anton Paar DMA 5000 M) over the temperature range from 275 to 363 K together with an uncertainty analysis. The approach is similar to that of Fritz et al. [J. Phys. Chem. B 104 (2000) 3463] when the density can be obtained from the relative oscillation period PQ , damping difference Δ D 0 and calibration parameters A , B , V 1 , and V 2 as ρ = A ( 1 + V 1 Δ D 0 + V 2 Δ D 0 2 ) PQ 2 - B . Temperature dependencies of the calibration parameters A and B were determined from a series of precise measurements with ultrapure water and dry air correlated to the densities calculated from the IAPWS-95 equation of state for water and the IAPWS G8-10 guideline for humid air. The calibration procedure was verified on measurements with toluene, ethanol, ethylene glycol, and glycerol. A detailed analysis of the uncertainty budget resulted in the standard uncertainty around 0.030 kg · m - 3 for typical low-viscosity samples. In case of highly viscous liquids such as low-temperature ethylene glycol or glycerol, the uncertainty can reach 0.060 kg · m - 3 or even 0.14 kg · m - 3 , respectively, at the dynamic viscosity exceeding 50 mPa · s . Other influences are also discussed, such as the relation between fluid viscosity and damping, the isotopic composition of the calibration water, the measurement procedure covering VTD cleaning and filling, and the effect of water contained in samples.