Milliliter-Scale Setup for the Efficient Characterization of Multicomponent Vapor-Liquid Equilibria Using Raman Spectroscopy

Vapor-liquid equilibrium (VLE) data are of major importance for the chemical industry. Despite significant progress in predictive methods, experimental VLE data are still indispensable. In this work, we address the need for experimental VLE data. Commonly, the characterization of VLE requires significant experimental effort. To limit the experimental effort, VLE measurements are frequently conducted by synthetic methods which employ samples of known composition and avoid complex analytics and sampling issues. In contrast, analytical methods provide independent information on phase compositions, commonly based on sampling and large amounts of substance.
In the first part of this work, we employ a synthetic method, the well-established Cailletet setup, to characterize the high pressure VLE of two promising binary biofuel blends. The Cailletet method serves as a state of the art reference method that enables collecting data of remarkable accuracy. However, extensive infrastructure is needed.
In the second part, to avoid extensive infrastructure and overcome limitations of previous methods, we develop a novel analytical milliliter-scale setup for the noninvasive and efficient characterization of VLE: RAMSPEQU (Raman Spectroscopic Phase Equilibrium Characterization). The novel setup saves substance and rapidly characterizes VLE. Sampling and its associated errors are avoided by analyzing phase compositions using Raman spectroscopy. Thereby, volumes of less than 3 ml are sufficient for reliable phase equilibrium measurements. To enable rapid data generation and save substance, we design an integrated workow combining Raman signal calibration and VLE measurement. As a result, RAMSPEQU gives access to up to 15 pT xy-data sets per workday. RAMSPEQU is successfully validated against pure component and binary VLE data from literature.
However, mixtures with only two components rarely depict real industrial applications. As the number of experiments increases strongly with a rising number of components, the efficient RAMSPEQU setup seems particularly suited for multicomponent systems. In the third part of this work, we employ the RAMSPEQU setup for the characterization of a quaternary system and its binary subsystems. 22 ml and 105 ml of the binary and quaternary mixtures are sufficient for an extensive VLE characterization.
The RAMSPEQU setup and its integrated workow enable the characterization of multicomponent VLE while saving significant amounts of substance and laboratory time.