engleski

Multivariate regression methods for fitting the dissociation constants from the UV/VIS spectra of self-assembled dyes

Multivariate analysis and regression present powerful tools for the analysis and interpretation of large data matrices in analytical chemistry1. In the past decades, methods such as the principal component regression (PCR), were also successfully applied for fitting thermodynamic models to sets of absorption spectra measured for varied thermodynamic equilibrium conditions. A very interesting example is the application of PCR for fitting the dissociation constants from absorption spectra measured at varied pH, when the protonation species present relatively similar spectral responses, and when these spectra are contaminated with variable interferences2, 3. The poster demonstrates the mathematical and computational methodology for fitting the dissociation constants of indicator dyes from their UV/VIS absorption spectra, without using the pure species spectra (calibration). The studied dyes include the in-house synthesized amphiphilic cationic pyridylporphyrins, as an example of photosensitizers that in aqueous medium self-assemble into liposomes4. Due to the strong cationic head, the dissociation constant of the porphyrin nitrogens lie in the range 1<pK<2.5, making the fully protonated state inaccessible for studying at moderate ionic strengths. The aggregation state influences the absorption spectrum by shifting the molecular electronic states as well as the scattering intensity. Consequently the measured spectra at varied pH exhibit shifts that make standard interpretation rather difficult. Therefore, we employed a variety of methods including PCR and partial least squares regression (PLSR) for fitting the protonation isotherms from the absorption spectra measured at varied pH. These methods enable analyzing only the spectral features which accurately represent the protonation equilibrium. In turn, this allowed us to determine the dissociation constants with accuracy that is in all cases better than 0.05 units, plot the protonation speciation diagrams, and generate predicted spectra of the pure protonation species. (1) Geladi, P. ; Kowalski, B. R. Anal. Chim. Acta 1986, 185, 1–17. (2) Kubista, M. ; Sjöback, R. ; Nygren, J. Anal. Chim. Acta 1995, 302 (1), 121–125. (3) Ritschel, T. ; Totsche, K. U. Chemosphere 2017, 172, 175–184. (4) Lovell, J. F. ; Jin, C. S. ; Huynh, E. ; Jin, H. ; Kim, C. ; Rubinstein, J. L. ; Chan, W. C. W. ; Cao, W. ; Wang, L. V ; Zheng, G. Nat. Mater. 2011, 10 (4), 324–332.

multivariate regression ; pK, model ; spectrophotometry ; UV/VIS ; porphyrin ; dye ; amphiphilic

nije evidentirano