High-Temperature, High-Pressure Viscosities and Densities of n-Hexadecane, 2,2,4,4,6,8,8-Heptamethylnonane, and Squalane Measured Using a Universal Calibration for a Rolling-Ball Viscometer/Densimeter

The development of reference correlations for viscous fluids is predicated on the availability of accurate viscosity data, especially at high pressure, high temperature (HPHT) conditions. The rolling ball viscometer (RBV) is a facile technique for obtaining such HPHT viscosity data. A new, universal RBV calibration methodology is described and applied over a broad T-p region and for a wide range of viscosities. The new calibration equation is used to obtain viscosities for n-hexadecane (HXD), 2,2,4,4,6,8,8-heptamethylnonane (HMN), and 2,6,10,15,19,23-hexamethyltetracosane (squalane) from 298 – 530 K and pressures to 250 MPa. The available literature data base for HMN is expanded to 520 K and 175 MPa and for squalane to 525 K and 250 MPa. The combined expanded uncertainties are 0.6% and 2.5% for the densities and viscosities, respectively, each with a coverage factor, k = 2. The reliability of the viscosity data is validated by comparison of HXD and squalane viscosities to accepted reference correlations and HMN viscosities to available literature data. The necessity of this new calibration approach is confirmed by the large deviations observed between HXD, HMN, and squalane viscosities determined using the new, universal RBV calibration equation and viscosities determined using a quadratic polynomial calibration equation. HXD, HMN, and squalane densities are predicted with the Perturbed Chain Statistical Associating Fluid Theory using pure component parameters calculated with a previously reported group contribution (GC) method. HXD, HMN, and squalane viscosities are compared to Free Volume Theory (FVT) predictions using FVT parameters calculated from a literature correlation for nalkanes. Although the FVT predictions for HXD, a normal alkane, result in an average absolute percent deviation (∆AAD) of 3.8%, predictions for HMN and squalane, two branched alkanes, are four to 13 times larger. The fit of the FVT model for the branched alkanes is dramatically improved if the FVT parameters are allowed to vary with temperature