Residual
entropy scaling (RES) of thermal conductivity was applied
to pure refrigerants, including natural and halogenated refrigerants,
and their mixtures. The reference equations of state and the mixture
models implemented in the REFPROP software package were adopted to
calculate the residual entropy, and the critical enhancement of thermal
conductivity was taken into account with the RES approach for the
first time. Experimental data of 39 pure fluids with more than 38,000
data points and of 31 mixtures with more than 7600 points were collected
and analyzed. More than 95.4% of the data (within two standard deviations
of the mean) of pure fluids collapse into a global dimensionless residual
thermal conductivity versus scaled dimensionless residual entropy
curve within 11.1% and those of mixtures are within 8.3%. This smooth,
monotonically increasing curve was correlated with a polynomial function
containing only four fitted parameters and one fluid-specific scaling
factor. Each pure fluid has its individual scaling factor, and a simple
mole-fraction-weighted mixing rule was applied for mixtures. The correlation
function provides a reliable thermal conductivity prediction of pure
fluids and, without any additional parameters, of mixtures. The proposed
model yields a similar level of statistical agreement with the experimental
data as the extended corresponding states model, which is the current
state-of-the-art and has as many as four more parameters for each
pair of components.