PARTITION COEFFICIENTS OF BIOLOGICAL COMPOUNDS BETWEEN IONIC LIQUIDS AND WATER. THEIR PREDICTION THROUGH THE ?SOLVATION PARAMETER MODEL?

PADRÓ, JUAN M.; PELLEGRINO VIDAL, ROCÍO; DÍAZ, GISELA; PONZINIBBIO, AGUSTÍN; RETA, MARIO

Florianópolis

Congreso; COLACRO XIV - Congresso Latino-Americano de Cromatografia e Técnicas Relacionadas; 2012

Associação Internacional de Cromatografia

PARTITION COEFFICIENTS OF BIOLOGICAL COMPOUNDS BETWEEN IONIC LIQUIDS AND WATER. THEIR PREDICTION THROUGH THE ?SOLVATION PARAMETER MODEL? Juan M. Padró, Rocío Pellegrino Vidal, Gisela Diaz, Agustin Ponzinibbio, Mario Reta The room-temperature ionic liquids (RTILs) are organic salts formed by organic cations asymmetrically substituted by alkyl groups (N alkylpyridinium, N,N´-dialkylimidazolium, etc.) and inorganic anions such as chloride, hexafluorophosphate (PF6-), tetrafluoroborate (BF4-), among others. The RTILs are considered as potential ?green solvents? due to the better compatibility with the environment as compared with the typical organic solvents, to its low vapor pressures and very low flammability. Together with its favorable solvation properties towards organic and inorganic analytes, the RTILs are actually very used in analytical chemistry, specifically, in liquid-liquid microextraction. The phosphonium-based cations some advantages as compared with the previously mentioned Nitrogen-based RTILs, such as better solvation properties and higher immiscibility with water. In this work, the partition coefficients between different RTILs and water, PIL/w, for different probe molecules (training set) as well as for compounds of biological interest (test set) were accurately measured. The Cyphos 101, Cyphos 102, Cyphos 105, Cyphos 109, [OMIM][PF6] and [OPYR][BF4] were used. In order to elucidate the chemical interactions involved in the partitioning process, the ?solvation parameter model? was used. This model considers a lineal regression between the log PIL/w and different solute parameters which measure polarity-polarizability, hydrogen-bond donor, hydrogenbond? acceptor, dispersive interactions and cohesivity. By using the training set, the obtained multiparametric regression coefficients were used to predict the PIL/w values for compounds of the test set. Solute descriptors for the test set were calculated through commercial software. An excellent agreement between calculated and experimental log PIL/w values was obtained, which demonstrate that the obtained multiparametric equations are robust and allow predicting partitioning for any organic molecule in the biphasic systems studied. Cavity formation and hydrogen-bond acceptor interactions domains the partition process. Polarity, hydrogenbond acidity and basicity for the different studied RTILs were measured by using different solvatochromic probes in order to interpret the regression coefficients.