Tuning the p Ka of a pH Responsive Fluorophore and the Consequences for Calibration of Optical Sensors Based on a Single Fluorophore but Multiple Receptors.

Since Sørensen and Bjerrum defined the pH scale, we have relied on two methods for determining pH, the colorimetric or the electrochemical. For pH electrodes, calibration is easy as a linear response is observed in the interesting pH range from 1 to ∼12. For colorimetric sensors, the response follows the sigmoidal Bjerrum diagram of an acid-base equilibrium. Thus, calibration of colorimetric sensors is more complex. Here, seven pH responsive fluorescent dyes based on the same diazaoxatriangulenium (DAOTA) fluorophore linked to varying receptor groups were prepared. Photoinduced electron transfer (PeT) quenching from appended aniline or phenol receptors generated the pH response of the DAOTA dyes, and the position of the p Ka value of the dye was tuned using the Hammett relationship as a guideline. The fluorescence intensity of the dyes in a sol-gel matrix environment was measured as a function of pH in universal buffer, and it was found that the dyes behave as perfect pH responsive probes under these conditions. The response of optical pH sensors is nonlinear and was found to be limited to 2-3 pH units for a precision of 0.01 pH unit. As sensors with a broader sensitivity range can be achieved by mixing multiple dyes with different p Ka values, mixtures of dyes in solution were investigated, and a broad range pH sensor with a precision of 0.006 pH units over a range of 3.6 pH units was demonstrated. Further, approximating the sensor response as linear was considered, and a limiting precision for this approach was determined. As the responses of the pH responsive DAOTA dyes were found to be ideally sigmoidal and as the six dyes were shown to have p Ka values scattered over a range from ∼2 to ∼9, this allows for design of a broad range optical pH sensor in the pH range from 1 to 10. This hypothesis was tested using quaternary mixtures of the different DAOTA dyes, and these were found to behave as a direct sum of the individual components. Thus, while linear calibration is limited to a precision of 0.02 in a range of 2-3 pH units, calibration using ideal sigmoidal functions is possible in the range of 1-10 with a precision better than 0.01, and as good as 0.002 pH units.