No abstract

Techniques to determine and compensate for the force transmission error (FTE), including the magnetic effects of the fluid being measured, in magnetic suspension densimeters are presented. For a two-sinker densimeter, the forces on the balance are expressed for each of the weighings comprising a density determination (i.e., the two sinkers plus balance calibration and tare weights). This yields a system of four equations, which are solved for the fluid density, a balance calibration factor, a coupling factor (related to the FTE), and a quantity related to the balance tare. For a single-sinker densimeter, an in situ weighing of the sinker in vacuum compensates for the FTE of the apparatus itself. A determination of the fluid-specific effect requires measurements with two different sinkers-analogous to the two-sinker analysis, but with the measurements spread out over time. The apparatus part of the FTE is generally less than ±20 ppm. Measurements on propane, helium, neon, nitrogen, argon, toluene, and air are analyzed for the fluid-specific effect; this effect is correlated with the magnetic susceptibility of the fluid together with an apparatus constant. With this analysis, the force transmission "error" becomes an effect that can be accounted for rather than a significant source of uncertainty in density measurements carried out with magnetic suspension densimeters.

This paper comprehensively describes those types of densimeters that allow very accurate density measurements of fluid substances over large ranges of temperature and for high pressures and that also cover the entire density range from low gas densities up to high densities of compressed liquids. Densimeters meeting these requirements are based on special applications of hydrostatic balances in combination with magnetic suspension couplings. These densimeters can be subdivided into two-sinker densimeters, and single-sinker densimeters which allow pρT measurements with total uncertainties in density of about 0.01% to 0.02% (level of confidence 95%) depending on their individual design. Besides a discussion of the design of such densimeters, it is also shown which substances in which ranges of the fluid state have been investigated with these instruments. Using nitrogen as an example, the experimental results obtained with these densimeters are discussed and compared with the pρT data obtained with other types of densimeters. A description of the features of other types of densimeters is outside the scope of this paper, but they are briefly summarized at the beginning of this paper.

No abstract

No abstract