Rayleigh and Brillion Scattering for the Measurement of Thermal Diffusivity and Sound Speed in Multicomponent Mixtures

According to the regulation (EC) No. 2037/2000 of the European Parliament, which controls the use of hydrochlorofuorocarbons, the refrigerant R22 is prohibited due to environmental hazards. Appropriate replacements for this widely used industrial refrigerant are thus necessary. The new refrigerant mixtures R22 M and R22 L currently constitute a direct drop-in solution for existing refrigeration systems. To optimize the performance of existing systems, the thermophysical properties for these ternary mixtures must be determined. Up to now, no experimental data have been available. Since the calculation of the transport properties for multicomponent mixtures based on the properties of the pure components still has an uncertainty range of 10%, measurements for the thermal diffusivity and sound speed, performed in this thesis via Dynamic Light Scattering (DLS), are of technical and scientific interest.

DLS is a thoroughly proven technique and known as a very accurate and absolute method to determine thermophysical properties. For example the thermal diffusivity a, the speed of sound cs, sound attenuation Ds, and the mutual diffusion coefficient D12 of fluids can be investigated in macroscopic gradient, which influences and distorts the results. DLS techniques analyze the decay of microscopic gradients present in thermodynamic equilibrium, resulting from Brownian molecular motion.

Currently available algorithms and methods for the DLS data evaluation are quite elaborate and thus not effective for technical and industrial standard. For evaluating the thermal diffusivity based on a recorded auto correlation function (ACF), the measure data have to be fitted to an ideal mathematical function, a fit model. Current methods can only handle ACFs free of any disturbing signals. Consequently, developing a new data analysis routine is necessary and the main goal of this thesis. The new method must be capable of handling erroneous ACFs in order to achieve the highest possible precision in the final result for the corresponding property.

 

At first in this thesis, an introduction to the basic principles of DLS will be presented, followed by a description of the experimental setup measuring in the time domain. The latter section also includes a detailed discussion about disturbing signals which have to be considered during the experiment and regarding the data evaluation procedure. Then the development of a user-firendly software, which also considers disturbing signals, with the derivation of an algorithm dealing with non-ideal ACFs, is desribed in detail. Several evaluation technique, models, and procedures are tested and discussed with helpf of experimental data as well as synthetic dta sets. This discussion leads to one evaluation procedure, which provides the bet reuslts regarding the resulting error. In the next section, the results obtained by the DLS experiment are used to finally discuss and compare the different evaluation procedures. For comparison, the transport properties of the samples are also predicted based on the properties of the pure mixture components, using of the software "RefProp" provided by the national Institute of Standard and Technology (NiST). Furthermore, simple prediction methods are compared with the DLS results.