Quinolinium-Based Fluorescent Probes for Dynamic pH Monitoring in Aqueous Media at High pH Using Fluorescence Lifetime Imaging

Spatiotemporal pH imaging using fluorescence lifetime imaging microscopy (FLIM) is an excellent technique for investigating dynamic (electro)chemical processes. However, probes that are responsive at high pH values are not available. Here, we describe the development and application of dedicated pH probes based on the 1-methyl-7-amino-quinolinium fluorophore. The high fluorescence lifetime and quantum yield, the high (photo)stability, and the inherent water solubility make the quinolinium fluorophore well suited for the development of FLIM probes. Due to the flexible fluorophore-spacer–receptor architecture, probe lifetimes are tunable in the pH range between 5.5 and 11. An additional fluorescence lifetime response, at tunable pH values between 11 and 13, is achieved by deprotonation of the aromatic amine at the quinolinium core. Probe lifetimes are hardly affected by temperature and the presence of most inorganic ions, thus making FLIM imaging highly reliable and convenient. At 0.1 mM probe concentrations, imaging at rates of 3 images per second, at a resolution of 4 μm, while measuring pH values up to 12 is achieved. This enables the pH imaging of dynamic electrochemical processes involving chemical reactions and mass transport.


SI-2. Photophysics of HQ and Q at high pH values.
The photophysics of alkylated quinolinium probes is depicted in Figure S10 N Figure S10. Photo physics of 1-methyl-7-aminoquinolinium probes at high pH values.
In Figure S10 k F and k IC are rate constants for fluorescence, non-radiative decay by internal conversion, respectively. r ESPT and r GSPT are the rates of excited state proton transfer and ground state proton transfer. K A2 and are K A2 * are the acid dissociation constants in the ground and the excited state, respectively.
In the ground state HQ is the only species that is detected by absorption spectroscopy. This is the case because pK A has a value well above 14. In the excited state, the equilibrium composition is determined by equation S2 in which pK A2 * has values between 11.5 and 12.5. At pH values well below pK A2 * only HQ + * is formed, because there is no driving force for deprotonation in the excited state according to Eq. S2. At pH values well above pK A2 *, nonemissive Q* is formed exclusively upon excitation of HQ + . This is the case because the [HQ + *]/[Q*] ratio described by Eq. S2 approaches 0 and because at high pH values the rate of deprotonation r ESPT (= k ESPT [OH -]) is high enough to fully deprotonate HQ + * during its lifetime. At pH values close to pK A2 *, mixtures compose of HQ + * and Q* will be formed. The ratio [HQ + *]/[Q*] will be determined by the equilibrium ratio described by Eq. S2, the rate of deprotonation r ESPT and the excited state lifetime that is available to reach equilibrium. If the ESPT process is fast compared to the lifetime, i.e. r ESPT >> k IC ,the equilibrium constant K A * measured from intensity versus pH plots will equal K A2 *. If the ESPT process proceeds slower, deprotonation lags behind and the apparent dissociation pK A * will have a higher value. However, in this equation k ESPT is a rate "constant" that is pH dependent, is the unknown. As is the case in the intensity measurements, a slower ESPT process will result in a higher experimental pK A * value. In a follow-up manuscript we will address the kinetics of the ESPT process.

SI-4. Fitting the fluorescence lifetime vs pH curves
Fluorescence lifetime curves were fitted using Equation S5 that is similar to Equation 5, that we used for fitting fluorescence intensity curves. Equation S5 lacks a solid physical foundation, because lifetimes are proportional to the concentrations of the constituents in a mixture. Still Eq. S5 provides excellent fitting curves and the obtained pK A * values and lifetime enhancements are reported in Table 1.
The symbols used in Eq. S6 are explained in sections SI-3 and SI-1.

SI-6. Toggel FLIM camera settings and setup
All lifetime measurements were performed with a Toggel FLIM camera. The solutions were placed in a cuvette and an image was taken at experimental settings shown in Figure S17. The LIFA software calculates the lifetime from the phase shift of the laser at every pixel. The lifetimes shown in the figures (i.e. Figure 3) were calculated by taking an average over all the pixels ( Figure S18). This method is also known as the "Frequency domain" Note on the error bars: The error bars in all graphs are the standard deviations of the FLIM measurements. We acknowledge that the error bars of our measurements (in figures 4, 5, 6, S1-S7) are relatively high. There are two main reasons for this:  At high lifetimes (>10 ns) -The FLIM camera calculates the lifetime at >200000 pixels and takes an average over all these pixels to calculate the average lifetime ( Figure S18). However, exponential decay is a random process and at every pixel the lifetime can therefore be slightly different. At larger lifetimes the excited state molecule is more stable and therefore results in a larger spread of times (which in turn results in a larger standard deviation). Also, in all measurements the modulation frequency was 20 MHz, which is optimal for detecting lifetimes of around 8 ns. The higher lifetimes (12 ns+) would be more accurate if a modulation frequency of ~15 MHz was used.  At low light intensities -At low light intensities we also observed large standard deviation, this is mainly due to the poor signal to noise ratio at these conditions. The noise in the measurements was usually around 50-100 counts, which was similar to the signal at the lower (1%) fluorescence intensities.
The setup is shown in Figure S19.
A more standard method of determining the lifetime of a fluorescent probe is by looking at the exponential decay of the light intensity, which is also known as the "Time domain".
In order to translate lifetimes to pH values reliably, we have compared lifetimes measured in the time domain with those measured in the frequency domain in our FLIM microscope setup. The data collected for probe 2b, presented in Figure S20, clearly demonstrate that lifetimes determined by both methods are identical.