A Fluorogenic Reaction for Monitoring Cross-Coupling with Turn-On Ratio Greater than a Hundred Thousand

Abstract

Fluorogenic reactions convert nonfluorescent reactants into fluorescent products and are ubiquitous in molecular sciences, allowing discovery, quantification, and imaging of a range of chemical and biological processes. However, the “non-fluorescent” reactants typically have substantial residual fluorescence, reducing the turn-on ratio of the reaction. Limits on the turn-on ratio then impose constraints on the fluorogenic reaction’s application, such as lowering the maximum allowable concentration of reactants or increasing the minimum detectable amount of product. Here, we report a design scheme producing exceptionally high turn-on ratios of up to 207,000. The scheme relies on a condensation reaction uniting electron-rich and electron-deficient reactants, resulting in a product molecule with “push-pull” character and significantly altered color and photophysical behavior. This approach is demonstrated using the Suzuki cross-coupling to produce a highly fluorescent dicyanomethylenedihydrofuran (DCDHF) product. The product has a peak absorption nearly 1 eV redder than the nearest reactant, while also exhibiting a large Stokes shift of nearly 0.5 eV. As a result, the reactants are negligibly excited, leading to the exceptionally high turn-on ratio, and enabling reactions to be performed at high concentrations of starting material without drowning out the product signal. The reaction is demonstrated in bulk conditions and in microdroplets, where differences in reaction rate are observed.

Introduction

Fluorogenic reactions, where the product molecules are strongly fluorescent, but the starting substrates are not, find many uses in molecular sciences. In the context of chemical reaction development, their turn-on behavior makes them useful to optically and noninvasively monitor reaction processes, allowing the efficient screening of varying conditions in combinatorial studies − and rapid, sensitive kinetic experiments. , In chemical biology and biophysics, fluorogenic systems have found utility in a variety of in vitro assays, as well as intracellular sensing and imaging. − More broadly, fluorogenic reactions have enabled quantification of pollutant metal ions, enabled tracking of lithium ions in model systems for batteries, and found a role in teaching laboratories. A particularly dynamic application area for fluorogenic reactions is their use in single-molecule microscopy, where the vast majority of studies rely on fluorescence as a readout, as the separation of Stokes-shifted fluorescence photons from their excitation source leads to an ideally “zero-background” measurement. , Among single-molecule applications, fluorogenic systems are applied in the monitoring of single enzymes ,− and nanoparticle and inorganic catalysts, − super-resolution microscopy of biological systems, − and in enabling high-concentration single-molecule microscopy. , Especially important in ensemble and single-molecule studies related to catalysis is the notion that the concentration of fresh substrate molecules in these experiments can be effectively kept constant and new fluorescent products can be continuously created, minimizing the effects of photobleaching on measurements.

The ideal fluorogenic system should possess several key attributes. Several of these properties are common to all fluorescence imaging. For example, the newly formed fluorophore should have a high fluorescence quantum yield, to minimize the energy lost to nonradiative processes, and a high extinction coefficient, to minimize the required excitation intensity, which serves to reduce background and rate of photobleaching. The fluorophore should also possess a large Stokes shift, the gap between peak excitation and emission wavelengths, to reduce self-absorption, where fluorescence emitted from a sample is reabsorbed by the sample itself. A large Stokes shift also facilities optical experiments, as the excitation light source can be more easily filtered out with minimal loss of desirable fluorescence photons. However, ideal fluorogenic systems require an additional feature: the fluorogenic reaction should exhibit a large turn-on ratio, where a massive increase in fluorescence signal is seen at a particular pump wavelength when a given number of reactant molecules are replaced by an equivalent amount of product molecules. A high turn-on ratio enables the detection of trace amount of product formation, enabling detection of even weak “hits” in screening for new reactivity, or potentially allowing observation of single-molecule turnover, when in the presence of a finite amount of background. Importantly, as the turn-on ratio rises, higher concentrations of reactants can be accessed, leading to the capability of following reactivity in more varied and realistic chemical environments.

In this work, we report a strategy for producing extremely high turn-on ratios in fluorogenic condensation reactions based on tailoring the inductive properties of the reactants and products. In practice, most “non-fluorescent” reactants used in fluorogenic systems are merely weakly fluorescent. For example, in the well-known fluorogenic transformation of resazurin to resorufin deployed in many catalysis experiments including at the single turnover level, − , even as the resorufin product possesses a quantum yield of 0.97, the resazurin reactant still possesses a quantum yield of 0.09, , for an overall turn-on ratio of ∼ 15 after a small spectral shift is included (see SI). To access the large turn-on ratios that would be required to enable trace product molecule detection even at high concentrations of reactant, fluorescence from the reactant molecules (at a specific excitation wavelength and emission band optimized for the product molecule) must truly be infinitesimal. To achieve this goal, we designed a fluorogenic system such that in addition to being very weakly fluorescent at even the wavelength of maximum absorption (λ_max) of the reactant, the λ_max of the reactant is also significantly blue-shifted as compared to the λ_max of the product. By designing one reactant to have strong electron-donating character and the other to have strong electron-withdrawing character, a large red-shift of the λ_max of the product relative to the reactants is created. Now, having reactants that are both weakly emitting and absorbing, as opposed to just being weakly emitting, enables access to a turn-on ratio exceeding 100,000. The strategy is deployed for monitoring a carbon bond-forming cross-coupling reaction.

Results and Discussion

Molecular Design

We demonstrate our strategy using a fluorophore from the dicyanomethylenedihydrofuran (DCDHF) family, originally designed for its nonlinear optics properties. , Specifically, our approach is based on the Suzuki coupling between the aryl halide 2-dicyanomethylen-3-cyano-5,5-dimethyl-4-(4-bromophenyl)-2,5-dihydrofuran (compound 1) with a boronic acid, 4-(N,N-dimethylamino)­phenylboronic acid (DMAPBA). The product of this reaction is a 2-dicyanomethylene-3-cyano-2,5-dihydrofuran (DCDHF)-based fluorophore (compound 2, IUPAC name 3-cyano-2-dicyanomethylen-5,5-dimethyl-4-[(4′-dimethylamino-biphenyl)-4-yl]-2,5-dihydrofuran). This reaction is shown below in Figure , with the general experimental procedure detailed in the Experimental. DMAPBA has a strong electron donating character, while compound 1 has the DCDHF group, endowing it with strong electron accepting character. Upon formation of cross-coupling product molecule 2, the donor and acceptor groups are joined over what becomes an extended π-conjugated system, yielding a fluorescent product with highly red-shifted λ_max of absorption that can be selectively excited without pumping the reactants. The progress of this reaction can be quantitatively monitored in situ by an increase in fluorescence corresponding to an increasing concentration of coupled product.

1.

Production of a DCDHF fluorophore via Suzuki coupling.

The dramatic photophysical properties exhibited by DCDHF dyes can be traced back to the push–pull nature of their design. , The eponymous DCDHF group serves as an electron-withdrawing “pull” group and in product 2, is separated from a dialkylamino donor “push” group by a conjugated π system. This asymmetric distribution of charge, spread over a long distance, leads to a large ground state dipole moment. The HOMO of 2 is raised due to the electron-donating influence of the amine, while the LUMO is lowered due to the presence of the electron-withdrawing DCHDF functional group, resulting in a narrowed HOMO–LUMO gap and consequent red-shifted λ_max of absorption. Alternatively, one can also view the electron donating and withdrawing groups at either end of 2 as favoring the quinoidal form, leading to increased conjugation length and red-shifted λ_max. While conjugation length naturally increases in a Suzuki cross-coupling reaction due to the formed aryl–aryl linkage, this reaction effectively gains an additional “boost” in conjugation because of the product’s push–pull character. The strong charge-transfer properties of these dyes leads to red-shifting of their fluorescent emission, and correspondingly large Stokes shifts.

DCDHF dyes are effective reporters on their immediate chemical and physical surroundings with viscometric and polarity sensitivity due to dynamics in their excited state S₁. − The S₁ state can relax to one of two local energy minima, a planar structure resulting primarily from a twist around the phenyl-furan single bond, or a twisted intramolecular charge transfer (TICT) structure of slightly lower energy arising primarily from a twist around the dicyano-furan double bond. Relaxation from the dicyano-furan twisted state proceeds via conical intersection and is extremely rapid and nonradiative. Partitioning between these dual pathways of relaxation make the dye sensitive to its local environment; if the rotation of the dicyano group is impeded, the radiatively relaxing planar state will be preferred and the quantum yield will be higher. Rigid systems that prohibit intramolecular twists, such as viscous solutions or physically constrained geometries, limit the population of excited molecules relaxing through the nonradiative pathway, increasing the observed emission. ,− The viscometric character of DCDHF dyes has been characterized by comparing DCDHF quantum yields in poly­(methyl methacrylate) film with those in toluene, increasing concentration of glycerol in methanol as DCDHF solvent, and freezing an aqueous DCDHF solution. , In addition, the asymmetric structure of a push–pull chromophore leads to solvatochromic effects. Both excited states exhibit a high degree of charge transfer, and are stabilized by more polar solvents. Polar solvents result in a more dominant nonradiative relaxation and decreasing quantum yield, although there is not consensus as to the mechanism through which this process proceeds. , The particularly high quantum yields of DCDHFs in aromatic solvents such as toluene or benzene has been hypothesized to be due to solvent-fluorophore interaction such as π-π stacking that favors the more planar conformation of the excited state and inhibits intramolecular rotations.

The photophysics of the DCDHF family of molecules has made it into a valuable class of fluorophores for single-molecule microscopy, , including in live cells. DCDHF fluorophores are also known to be robust to photobleaching under bulk and single-molecule conditions, and variants have been prepared for lipid or aqueous solubility, or incorporation into oligonucleotides as a molecular beacon. DCDHF fluorophores have also been used in a photoactivatable context for super-resolution imaging. ,,

Photophysical Characterization

Compounds 1, 2 and DMAPBA were characterized by UV–vis and fluorescence spectroscopy in toluene (Figure ). The photophysical properties reported in this section are summarized in Table . Both reactants possess absorption peaks in the ultraviolet region. 1 displays very weak fluorescence when pumped at its λ_max, whereas DMAPBA emission was undetectable in the presence of residual scatter of excitation light. Product 2, in contrast, shows a λ_max of absorption at 488 nm, a shift of 137 nm (8000 cm^–1, 0.99 eV) from the spectrally nearest starting material, and a λ_max of emission at 605 nm, for a Stokes shift of 117 nm (3960 cm^–1, 0.491 eV). In comparison, fluorescein and Rhodamine 6G, two dyes commonly used as fluorescent labels, each exhibit Stokes shifts under 30 nm, with exact values dependent on solvent conditions. The fluorescence quantum yield of 2 was measured as 0.51 ± 0.01, using Rhodamine 6G as a standard.

2.

Optical characterization of key compounds. (a) UV–vis absorption spectra of DMAPBA. (b) UV–vis absorbance and fluorescence emission spectra (excitation at 351 nm) of reactant 1. The small peak at 700 nm is from residual scatter of the excitation light at the second order of the diffraction grating. (c) UV–vis absorbance and fluorescence emission spectra (excitation at 488 nm) of fluorophore 2. (d) Picture of reactants (left) and fluorophore 2 (right) upon UV excitation. All spectra are in toluene.

1. Photophysical Properties of Reactants and Products, as Experimentally Determined by UV–vis and Fluorimetry .

molecule	DMAPBA	1	2
absorptive λmax	306 nm	351 nm	488 nm
emissive λmax	NA	469 nm	605 nm
Stokes Shift	NA	118 nm	117 nm
molar absorptivity (ε)	35,000 M–1 cm–1	17,500 M–1 cm–1	19,100 M–1 cm–1
fluorescent quantum yield (ϕfluor)	NA	NA	0.51 ± 0.01

^aAll spectra were recorded in Optima toluene. Quantum yields were determined relative to Rhodamine 6G (ϕ_fluor= 0.95). No fluorescence was detected from DMAPBA, while emission from 1 was difficult to resolve from elastic scattering, suggesting negligible quantum yields of emission for both reactants.

As discussed above, turn-on ratio is one of the most important figures of merit of a fluorogenic reaction, allowing detection of minimal product in the presence of a high concentration of starting material. Here, turn-on ratio was determined by integrating bulk fluorimetry spectra of the mixed starting materials and of pure product samples and calculating the ratio between them, normalized for concentration. Mixtures of DMAPBA and 1 were made at 0.5 mM, while product 2 was made at 0.0067 mM. This disparity in concentration was necessary since the fluorescence from the starting materials was so weak with visible excitation. Excitation was set to be at 488 or 532 nm, and the emission spectra were integrated from 538 to 800 nm for 488 nm excitation, and from 542 to 800 nm, for 532 nm excitation, with different slit widths. For 488 nm excitation, the turn-on ratio was calculated to be 127,000 ± 3000 for 5 nm emission slit width, and 69,000 ± 300 for 10 nm emission slit width, where the detailed uncertainty is the standard deviation. At 532 nm excitation, the turn-on ratio was 207,000 ± 8000 for 5 nm emission slit width, and 171,000 ± 3000 for 10 nm emission slit width. With 532 nm excitation, all measurements of turn-on ratio values comfortably exceeded the extraordinarily high value of 100,000.

In Situ Reaction Monitoring

The identification of new carbon–carbon bond-forming reactions continues to be a major frontier in reaction chemistry, and major motivator for the development of fluorogenic reactions given their role in combinatorial approaches. The key value of a fluorogenic reaction is the ease with which desirable reactivity can be identified and progress over time monitored. As a demonstration, fluorescence intensity was monitored while running the Suzuki cross-coupling reaction with DMAPBA and 1 as starting materials and Pd­(PPh₃)₄, a common cross-coupling catalyst (2.5 mM starting materials, 10 uM or 0.4 mol % catalyst) in a cuvette capped to minimize solvent evaporation. Scans were taken every ten minutes over the span of two hours with excitation at 532 nm (Figure ). As shown, fluorescence increases monotonically, and the fluorogenic nature of this system makes it easy to quantify changes in product concentration over time. An experimental control of the system without catalyst was also performed, and negligible fluorescence was observed. After the two-hour reaction time had elapsed, the reaction was quenched with brine and the product separated from remaining starting materials via column chromatography. Isolated yield was 35%. A survey of other attempted conditions is provided in the Supporting Information.

3.

Monitoring production of compound 1 from a Suzuki reaction over time by fluorimetry. Spectra were recorded before catalyst addition (t = −1 min), immediately after catalyst addition (t = 0 min), and every 10 min thereafter for 120 min. (a) Fluorescence emission spectra of the reaction at various time points when excited at 532 nm. (b) Peak fluorescence intensity over time. (c) Qualitative color change of the fluorogenic Suzuki reaction over time.

The extremely high fluorescence turn-on ratio of this reaction makes it ideal for the detection of trace amounts of product. To demonstrate this capability, the reaction was performed at extremely low catalyst loadings, down to 0.023 mol percent (0.65 μM). Reactions were set up in vials at room temperature without stirring for two hours, then measured via fluorimetry exciting at compound 2’s λ_max of absorption (Figure , left). Their emission was compared to a linear calibration of fluorescence from known concentrations of 2 in toluene, allowing the quantification of the concentration of 2 in each vial postreaction. Turnover number (TON), the number of times each palladium complex catalyzes the reaction, was then determined by comparing the number of product molecules formed to the number of catalyst molecules in solution (Figure , right). The sharp drop in reactivity below catalyst loadings of approximately 0.3 mol percent (10 μM) suggests that Pd­(PPh₃)₄ is likely most active in nanoparticle form or otherwise requires the presence of multiple catalyst molecules, as has been proposed in literature.

4.

Fluorescence from reactions with varying catalyst concentration (left) indicative of reaction yield can be used to measure turnover number (right) of the catalyst at varying concentrations, down to 0.023 mol percent.

Reaction Monitoring in Emulsions

Widefield fluorescence microscopy can also be used to measure the rise in fluorescence indicative of catalyst turnover in the droplets of an emulsion (Figure ). Reactions in dispersed media are of interest in a variety of reaction contexts, where the inherently small scale of reaction both reduces quantities of reagent needed and waste generated while allowing increased assay portability and close control of reaction. , Microdroplets are used in high-throughput screening of reaction conditions, ,, including drug candidate screenings , and the isolation and study of single cells. , Reactions performed in mixed phases and emulsions are also key processes in the industrial toolkit to produce specialty polymeric and hybrid materials on large scales. − Microdroplets have also been proposed as functional units for chemical computing, where an optical readout is necessary. , Additionally, a rapidly evolving field of study has emerged on altered chemical kinetics sometimes observed when reactions to are confined to small droplets, with the mechanistic understanding of this phenomenon still under debate. − Despite the utility of nanoconfined volumes across many fields, the transient nature and complexity of emulsion droplets makes their study challenging. Fluorescence microscopy-based methods are uniquely powerful tools for the study of emulsions, due to the high spatial resolution they offer and the potential benefit of a nonfluorescent continuous phase on signal-to-background. An outstanding need is chemically specific optical readouts, such as a recently reported means of monitoring polymerization in droplets. A fluorogenic reaction provides another chemically specific optical readout.

5.

Fluorogenic Suzuki coupling preformed in microscopic emulsion droplets. The brightfield of the droplets (a) and their fluorescence emission (b) can be overlaid (c) with fluorescence in false color (slightly magnified), showing the reaction is localized to the organic droplet phase and does not extend into the aqueous continuous phase. In emulsions with catalyst present, the fluorescence of the system increases on the time scale of hours in individual droplets (d, thin gray) and for the mean of all drops in the field of view (d, color). In control experiments (d, black), emulsions were made with both starting materials but without catalyst, and no increase in fluorescence was observed.

The fluorogenic reaction shown in Figure was performed in an emulsion with toluene droplets (Figure a), dispersed in an aqueous continuous phase, which was observed to be stable on the time scale of hours without any added surfactant. The toluene droplets displayed a range of diameters, centered at 0.7 μm but extending from 0.3 to 1.8 μm (see SI, Figure S3). Fluorescence from droplets is easily identified (Figure b) and correlated with droplet location (Figure c). Emulsions made with all reactants (2.5 mM each DMAPBA and 1, 10 μM Pd­(PPh₃)₄ catalyst), show an increase in brightness over time (Figure d). Control emulsions, made with all reactants except catalyst, remain dark, as the reaction does not proceed and the starting materials are nonfluorescent (Figure d, black trace). The reaction’s high turn-on ratio makes the fluorescent readout extremely sensitive, enabling facile detection of early reactivity and quantification of turnovers from low catalyst loadings. Widefield measurements allow a broad field of droplets to be monitored in a single field of view, enabling the simultaneous acquisition of dozens to hundreds of data points. Additional trajectories from individual droplets are shown in Figure S4.

Interestingly, the behavior of the reaction was different when measured in emulsion than when measured under bulk conditions as above. The experiments in Figure and Figure were run with the same concentrations of starting materials, including catalyst. While the increasing fluorescence in the bulk-scale reaction of Figure leveled out after about an hour, the brightness of the droplets in Figure was still trending upward linearly after two hours. Additionally, average droplet brightness increased by over a factor of 20 in the emulsion, between the first image taken 20 min after catalyst addition and the final image taken at 130 min. In comparison, the fluorescence intensity in the experiment run in a cuvette increases by less than a factor of 2 between t = 20 min and t = 120 min (both with intensity as measured at the λ_max of emission and as measured by the integration of the fluorescence spectrum). Differences in observed reaction kinetics could derive from previously identified droplet enhancement mechanisms, − but other differences in geometry are also present. Here, the emulsion experiment measures dynamics in droplets surrounded by an aqueous continuous phase, which could affect the environmentally sensitive DCDHF product. The exposure time for the emulsion was considerably shorter (100 ms vs 10 s) than the bulk reaction since spectral resolution was not required, though excitation intensity was greater. Increasing the exposure time or frame rate for the emulsion experiment resulted in a decrease of fluorescence intensity, likely due to product photobleaching. The conspicuous discrepancies in activity indicate the utility of fluorogenic systems applied to the study of nanoconfined chemistry.

Finally, the prospect of reaching the single-catalyst limit can be discussed, as reached in the classic microdroplet confined beta-galactosidase experiments of Rotman. The droplets shown in Figure are approximately 0.7 μm in diameter, so assuming a hemispherical geometry, have a volume of 0.1 femtoliters. To track the increasing fluorescence in these droplets due to the activity of a single molecule of catalyst, catalyst concentration would need to be single-digit nanomolar, 4 orders of magnitude lower than the activity drop-off shown in Figure . As volume is proportional to the cube of droplet radius, droplets of 0.1 μm diameter would only need catalyst concentrations of around 1 μM for single-molecule studies. The TON for Pd­(PPh₃)₄ at that concentration under these conditions is around 1 molecule of product formed for each molecule of catalyst.

For the particular reaction shown in Scheme 1, pushing to higher TON at such low concentrations of catalyst is likely not possible. Increasing solvent polarity tends to increase the rate of similar cross-coupling reactions, and may even preserve the molecular character of the catalyst, but any solvent more polar than toluene results in substantial fluorescence quenching of DCDHF dyes, a major limitation of the reported scheme. , While our reported paradigm of using fluorogenically produced fluorophores with high push–pull character and high turnover ratio allows access to the needed high reactant concentrations that may ultimately enable the single molecular catalyst limit to be reached, a reaction system will need to be identified that maintains high fluorescent quantum yield in somewhat more polar solvents. One way to achieve this goal is to modify the reduction and oxidation potentials of the electron rich and electron poor reaction partners, respectfully, through careful functionalization to slow the rate of photoinduced charge transfer and allow radiative decay to prevail, even in more polar solvents. Such a reaction system, by having access to more polar solvents, would also likely result in higher yields and molecular catalytic behavior. However, if the redox potentials are lowered too much, the shift in absorption spectrum of the product might decrease and the high turn-on ratio might suffer. The search for new candidates is in progress.

Conclusions

In this work, we report a fluorogenic reaction with an exceptionally high turn-on ratio greater than 10⁵. This high turn-on ratio derives from the electronic structure of the fluorophore relative to the component reactants, whereby a massive shift in absorption in the product relative to the reactants and a large Stokes shift are observed. Both of these features originate in the large charge transfer character of the product fluorophore. These features allow the fluorophore to be efficiently excited while limiting emission from the reactants to a nearly undetectable amount. This chemistry is deployed in a cross-coupling reaction to allow the reaction to be monitored quantitatively in real time as it proceeds, reporting on the activity of the catalytic species. In bulk fluorimetry experiments, trace reactivity was detectable at catalyst loadings as low as 0.023 mol percent. This fluorogenic C–C cross coupling reaction adds an additional tool for high throughput reaction screenings, as demonstrated by its use in emulsions. One limit of the reaction is that the high turn-on ratio is only accessible in highly nonpolar solvent. This paradigm can be extended to other condensation reactions and other fluorophores of similar “push-pull” character.

Experimental Section

Materials and General Suzuki Coupling Procedure

Synthesis and NMR characterization (Figures S1, S2) of reactants and intermediates are included in the Supporting Information.

Compound 1 and DMAPBA were combined 1:1 in Optima grade toluene at concentrations of 2.5 mM. In a separate vial, Pd­(PPh₃)₄ was dissolved in toluene, and at the designated t = 0 min of reaction, an appropriate volume of catalyst was added to reach desired concentration. Volume was filled to 20 mL with toluene to keep concentrations consistent across experiments. Reactions were carried out for two hours in the dark without stirring under ambient atmosphere, temperature, and pressure. Thin layer chromatography was used in conjunction with colorimetric and fluorescence methods to check for the presence of product as well as residual starting material. On glass-backed silica plates in 45:55 hexane:ethyl acetate, R_f values were measured to be 0.71 for 1, 0.61 for 2, and 0.38 for DMAPBA, visualized by UV light at 254 nm. Reactions were quenched with brine and extracted into dichloromethane, then purified by column chromatography (45:55 hexane:ethyl acetate). More standard literature conditions for a Suzuki coupling (catalyst loading of 2 mol % and excess Na₂CO₃, air-free techniques, stir 20 h) yielded product 2 at a 35.5% yield after purification.

UV–vis and Fluorimetry Acquisition

Absorption spectra were acquired on a Varian Cary 50 UV–vis Spectrophotometer (version 3.00) with a scan rate of 600 nm/min in Dual Beam mode. Scan Software Version 3.00 (182) was used for data collection. Samples were measured in a Hellma QS quartz glass cuvette with a 1 cm path length.

The fluorimetry data in Figure , and for quantum yield and turn-on ratio measurements was acquired on a F-7100 Hitachi fluorescence spectrophotometer with a Xe lamp in Fluorescence mode with a scan speed of 1200 nm/min and recorded with FL Solutions v 4.2. Figure c was background subtracted using a toluene blank, while Figure b is presented without background subtraction due to data processing artifacts from the extremely weak signal, which is weaker than elastically scattered light. For timelapse measurements and quantum yield, the excitation slit width was 1.0 nm and emission slit width 2.5 nm, but for turn-on ratio measurements excitation slit width was 2.5 nm and emission slit width was 5.0 nm, widened to collect the trace fluorescence from the starting materials. The fluorimetry data in Figure was acquired on an F-4500 FL Spectrophotometer with Xe lamp. The scan speed was 1200 nm/min and excitation and emission slit width were each 5.0 nm. Samples were measured in a Hellma QS quartz glass cuvette with a 1 cm path length. Between fluorescence scans, the instrument’s shutter was closed to minimize the sample’s exposure to light and resultant bleaching behavior of the dye product. Time points were taken once every 10 min in bulk fluorimetry with a fluorescence scan of 200 at 1200 nm/min taking 10 s per time point.

Microscopy Sample Preparation and Data Collection

Triethoxyoctylsilane (TEOS) was used to create a hydrophobic coverslip surface where droplets of toluene could be immobilized and imaged on hours-long time scales in a procedure adapted from previously published work by the Goldsmith lab. Samples were prepared for microscopy in “sandwiches” with a bottom TEOS-functionalized coverslip and a top, plasma-cleaned, unfunctionalized glass coverslip. After emulsifying a solution of 95% Milli-Q water (continuous phase) and 5% organic reaction phase (Optima toluene, DMAPBA, compound 1, Pd­(PPh₃)₄ catalyst) via vortexing, 5 μL of the emulsified mixture was pipetted onto a 1% TEOS functionalized coverslip. This was diluted with an additional 10 μL of Milli-Q water, to ensure that droplets visualized were organic in water, rather than water in organic. A nonfunctionalized coverslip was rinsed with HPLC grade methanol and Milli-Q water, dried under filtered nitrogen, and cleaned for 5 min with ambient oxygen plasma (Harrick Plasma PDC-001-HP, 45W), then placed on top of the drop of emulsion. The sides of this sandwich were sealed with clear nail polish to prevent solvent evaporation, and the sample was placed on the microscope where droplets could be found in brightfield. This sample preparation process produced polydisperse droplets ranging from 1 to 100 μm in diameter, and field of view was selected to center a monodisperse and well-separated droplet selection. Emulsions prepared in this way were observed to be stable on the time scale of hours to days.

Microscopy experiments were performed on a modified inverted microscope (Nikon Eclipse Ti-E) with Perfect Focus System and excitation from a 200 mW 532 nm laser (Lasertack). An electronic shutter at the start of the beam path was used to prevent excess photobleaching by keeping the sample from exposure to excitation between images. The laser was attenuated with ND filters and spatially filtered by a pinhole, then expanded via telescoping lenses to a diameter slightly larger than the back aperture of the objective (model). A multiedge dichroic mirror (Semrock Di01-R405/488/532/635–25 × 36) directed the excitation light to the sample. Emission light from the sample collected by the objective passed back through the dichroic and was sent through two stacked 532 nm long pass filters (Semrock LP03–532RU-25) to filter out background excitation light. Images were collected on an Andor sCMOS camera (SONA-4BV11) air-cooled to −25 °C.

Custom LabVIEW code was used to synchronize the shutter and camera exposure times to automatically collect images at each time point. The shutter was closed between exposures to minimize photobleaching. Time points were taken once every five minutes for a duration of two hours, with an exposure time of 100 ms. A custom MATLAB code was used to analyze the stack of images collected. The first frame of the video was used to identify the droplets as regions of interest (ROIs) via the TrackMate plugin to FIJI’s ImageJ. The MATLAB script performed a local background subtraction from the brightness intensity of each droplet. The fluorescence in each droplet was then recorded for each time point, reporting on the amount of product 2 formed over time.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/cbmi.5c00103.

• Synthesis and characterization of reactants, characterization of microdroplets and microdroplet dynamics, other reactions and cross-coupling reactions attempted, and discussion of the Resazurin/Resorufin fluorogenic system (PDF)

The authors declare no competing financial interest.