Single-Micelle and Single-Zinc-Particle Imaging Provides Insights into Physical Processes Underpinning Organozinc Reactions in Water

Abstract

Micelles on the surfaces of individual metallic zinc particles are imaged by fluorescence microscopy with sensitivity up to single micelles. These micelles are made fluorescent to enable imaging, through incorporation of boron dipyrromethene (BODIPY) fluorophores as representative organic molecular “cargo”. Highlighting an advantage of this in situ and sensitive fluorescence technique, the same micelles are not visible by ex situ SEM/EDS analysis. Examination of micellar solutions with zinc reveals an aging process: micelles do not immediately adhere to the zinc surfaces upon mixing but rather build up over time. Furthermore, at longer times, smaller zinc particles become fully encased in micelle “shells”. Once adhered, micelles remain in local regions of the zinc surface for the duration of the imaging experiments (>2 h). Single micelles are imaged in solution, and their molecular contents are characterized. Two-color fluorescence crossover experiments show that micelles adhered to the surface of the zinc exchange molecular contents with micelles in solution, achieving molecular exchange equilibrium in ~2.5 h. Unique (non-ensemble averaged) exchange kinetics are displayed by micelles at different locations on the zinc surface, consistent with exchange kinetics of single micelles or small local clusters of micelles. The aging of the micellar solutions and the rate of exchange while on the surface of the zinc suggest that micelle mass transport processes may contribute to overall reaction barriers in sustainable organozinc cross-coupling reactions in micellar water. The observed aging of the system suggests routes for improvement of preparative, bench-scale synthetic reactions involving micellar preparations of organozinc compounds.

Graphical Abstract

Introduction

Water holds particular promise as a sustainable solvent for organic and organometallic chemistry,^1,2 if not for its seemingly fundamental conflicts of reaction incompatibility. Nonpolar organic substrates are poorly soluble in water, and organometallic compounds tend to suffer from decomposition via protodemetalation.^3,4 Micelle-forming surfactants provide a plausible way to adapt organic substrates to aqueous conditions.^5–8 However, micellar approaches are so far not widely adopted in organic and organometallic chemistry. Even in cases where micelles are successfully used, their role in the reaction mechanism remains poorly understood.⁹ This poor understanding arises from existing analytical limitations that prevent characterizing the behavior of micelles in the complicated, often multiphasic reaction mixtures germane to organic and organometallic synthesis. We posit that this poor understanding is a primary factor limiting the generalization of micellar reaction methods to organic synthesis.

Here we exploit the spatial resolution power and high sensitivity of fluorescence microscopy to develop in situ imaging of micelles and to study micelle physical behavior as they transport organic “cargo” in the presence of metallic zinc powder. Images acquired through this technique provide early insights into real-time micelle dynamics under preparative organic chemistry reaction conditions: similar surfactant concentrations, ambient temperature, aqueous media, and commercial zinc powder composed of single particles ~2–20 μm with irregular physical and chemical surfaces.

Catching our attention, in 2009, Lipshutz demonstrated that palladium-catalyzed cross-coupling reactions with organozinc intermediates (generated in situ from commercial zinc powder) proceed under aqueous conditions with the use of self-assembled micelles (Figure 1).¹⁰ This demonstration is remarkable because organozinc compounds are water reactive and control experiments in the absence of micelles showed that the organozinc intermediates partially protodemetalated.^11,12 Thus, the micelles provided not only solubility enhancement but also protection against decomposition of the organometallic species. Palladium-catalyzed cross-coupling reactions are robust and have a broad range of academic and industrial uses;^13–15 heavy interest in this reaction class has generated the need for sustainable processes like this example.^1,14

Figure 1.

Reaction schematic of previous work, showing unknown micelle mechanistic behavior.

Despite the impressive synthetic outcome, the reaction lacks a granular mechanistic characterization that would facilitate generalization. Lipshutz proposed that the micelles adhere to the metallic zinc surface and protect the water-sensitive organozinc reagent prior to the cross-coupling step.¹⁶ The absence of mechanistic characterization evokes fundamental questions regarding the role of micelles: 1) Do micelles adhere to the zinc surface? 2) How is micelle adhesion to zinc influenced by time, stirring, and cosolvent? 3) Is micelle adhesion increased by the availability of direct insertion/oxidative addition chemistry of their organic substrate contents? 4) Does the sensitivity of in situ fluorescence microscopy enable detection of micelles on zinc surfaces that are not detectable with prior micelle characterization techniques? 5) Do micelles dissociate from zinc rapidly? Are micelles mobile or stationary once adhered to the zinc surfaces? 6) Do micelles exchange organic contents (e.g., substrates, intermediates, products) while on the surface of zinc? 7) Are the mass transport physical processes involving the micelles slow enough to limit the rate of the overall cross-coupling reaction kinetics? We here provide answers to these questions by developing and applying single-zinc-particle imaging techniques with up to single-micelle sensitivity.

To our knowledge, these are the first fluorescence microscopy imaging studies of micelles interacting with irregular metal surfaces. Micelles have previously been visualized through SEM¹⁷ and TEM,^17,18 but these ex situ methods are not suitable for studies on irregular surfaces, nor for in situ characterization of mechanisms, as they typically require smooth surfaces, ex situ high vacuum, and/or cryogenic temperatures. For these reasons, previous studies involving micelle interactions with metal surfaces either involved atomically smooth surfaces or were limited to computational methods.^19–22 In contrast, the metallic zinc powder employed in this preparative organic chemistry reaction has highly irregular physical and chemical surfaces and is employed in solvent at or above ambient temperature. Bulk methods such as NMR spectroscopy, while able to determine the kinetics of micelle exchange in solution,^23,24 lack the spatial resolution that would inform on the nature and variations of their interactions with zinc.²⁵

Results and Discussion

Question 1: Do micelles adhere to the zinc surface?

To examine this question under conditions similar to those reported by Lipshutz, a solution of 2%/wt polyoxyethanyl-±-tocopherol sebacate (PTS) surfactant in water was prepared.¹⁰ This concentration is well above the critical micelle concentration for self-assembly (0.028%)²⁶. The micelle solution was then loaded with fluorescent imaging agent 1 in a vial (Figure 2). The uptake of nonpolar 1 by the micelles was anticipated to make them brightly fluorescent and enable imaging. Boron dipyrromethene (BODIPY) imaging agent 1 was chosen as the spectator imaging agent due to its high quantum yield, chemical inertness, and small size relative to the micelle.²⁷ Zinc powder of identical supplier and mesh to that reported by Lipshutz¹⁰ was then treated with the micellar solution of 1, and stirred for 2 h. The resulting suspension was transferred from the vial to a microscopy imaging chamber without further workup and imaged on a widefield epifluorescence microscope. The image shown in Figure 2b is representative from a spatial survey of the sample (see SI Movie 3). Bright green fluorescent “hot spots” consistent with micelles were visible on the surface of otherwise dark, irregularly shaped, individual zinc particles (Figure 2b).²⁸

Figure 2.

(a) Schematic overview of fluorescence microscopy to image micelle behavior in heterogeneous mixtures at single-micelle and single-zinc-particle level. (b) Zinc powder treated in a fluorescent micellar solution with imaging agent 1. Created with BioRender.com.

Closer examination of the image in Figure 2b show that individual zinc particles, ~2–20 μm, were visible against a bright diffuse green background. The bright background was caused by BODIPY-loaded micelles that remained in solution and by out-of-focus zinc particles. Some in-focus zinc particles were relatively dark, consistent with being sparsely coated with micelles, and others were relatively bright, consistent with being densely coated with micelles. A control experiment showed the zinc/micelle mixture in the absence of 1 to be nonfluorescent (see SI, Figure S14a); thus, the fluorescent signal in the presence of 1 was attributed to 1 and not to autofluorescence of or light scattering from the zinc or micelles.

In order to determine if fluorophore imaging agent 1 was currently inside micelles in these fluorescent “hot spots,” we turned to fluorescence lifetime imaging microscopy (FLIM). Fluorescence lifetime is a highly sensitive gauge for fluorophore microenvironment.²⁹ An alternative origin hypothesis for these “hot spots,” for example, could be that the micelles had deposited their organic contents on the zinc surface and had resolubilized, leaving behind fluorophore physisorbed to the zinc surface yet outside of a micelle.

To perform FLIM imaging, zinc powder first was stirred with 1 for 2 h in water in the absence of PTS. This process caused 1 to physisorb to the zinc surface. This sample served as a control that enabled measurement of the average fluorescence lifetime of 1 on zinc in the absence of micelles. This sample was imaged by confocal microscopy with FLIM capabilities: Individual particles of zinc coated in physisorbed 1 appear false-colored blue, corresponding to τ_{Avg_Amp} = 2.4 ± 0.2 ns (representative image, Figure 3a).

Figure 3.

FLIM images showing the lifetime differences between: (a) BODIPY physisorbed to zinc surface without micelles, (b) BODIPY on zinc surface after addition of micellar solution to sample a, (c) BODIPY inside micelles, stirred with zinc. (d) Fluorescent lifetimes in samples a–c (from a minimum of different 3 particles measurements per sample).

For comparison, zinc powder was stirred with 1 for 2 h in 2% PTS micellar water. These conditions were identical to those which produced the previous widefield epifluorescence data in Figure 2b. Individual particles of zinc coated in 1 appear false colored green, corresponding to τ_{Avg_Amp} = 4.5 ± 0.1 ns (representative image, Figure 3c). These highly different fluorescence lifetimes between samples a and c clearly indicate that 1 is in different microenvironments in the absence and presence of micelles, even when 1 is on the zinc surface in both cases. This shorter lifetime of physisorbed 1 in the absence of micelles is consistent with partial quenching by the zinc caused by 1 being closer to the surface.³⁰ The longer lifetime of 1 in the presence of PTS is consistent with 1 being further away from the zinc surface. In turn, these values are consistent with 1 being inside of micelles when PTS is present rather than directly physisorbed to zinc.

When the water was decanted from sample a and the medium was replaced with 2% PTS micellar solution, 1 on zinc exhibited a fluorescence lifetime of τ_{Avg_Amp} = 3.6 ± 0.4 ns (Figure 3b, t ~ 3 min after addition of PTS). This fluorescence lifetime was intermediate between that which occurred in samples a and c. Sample b also exhibited a rapid time-dependent fading of fluorescence intensity (see SI, Figure S13). Both the intermediate fluorescence lifetime and the rapid fading of fluorescence are consistent with previously physisorbed 1 becoming taken and solubilized by micelles during the imaging experiment (~7 min).

Although the same single zinc particle cannot be found before and after the solvent change in a and b, the increase in lifetime upon addition of 2% PTS was representative of fluorescence lifetime changes across the sample. Summarized FLIM data for samples a–c are tabulated in Figure 3d. Error is reported as standard deviation from triplicate or quintuplet different locations in the same sample (i.e., from different particles), and the narrow deviation range shows the sample-wide representative nature of the measurements (see SI, Figures S11–S13). Control experiments in the absence of imaging agent 1 confirmed that the micelles and zinc surface alone are nonfluorescent, consistent with assignment of the fluorescence signal for fluorescence lifetime analysis as arising from 1 (see SI, Figure S14b).

Together, these FLIM experiments led to the conclusion that the fluorescent signal arising from 1 on the surface of zinc in the presence of 2% PTS, including the previously described “hot spots” in epifluorescence widefield images, arose from 1 inside micelles. The rapid rate of uptake of previously physisorbed 1 by micelles (~7 min) indicates that 1 will be present inside micelles when 2% PTS is the imaging medium from the start of the experiment. The “hot spots” on the surfaces of zinc in Figure 2b are therefore assigned as micelles, made fluorescent by the incorporation of imaging agent. Thus, micelles adhere to the zinc surface. Spatial distributions of the micelles on the zinc surfaces were heterogeneous, with selective adhesion in certain regions and on certain particles. The spatial distributions are consistent with a physical and chemical heterogeneity of the commercial zinc surface a (see below for additional discussion).

Question 2: How is micelle adhesion to zinc influenced by time, stirring, and cosolvent?

Lipshutz used dynamic light scattering (DLS) to measure the size of the PTS micelles in water and concluded that their sizes in solution ranged between 10 to 50 nm, with an average size of 22 nm.³¹ Micelles are known to increase in size and/or agglomerate, most notably with the addition of a cosolvent.^18,32–34 Addition of 10% THF, for example, has been shown to accelerate the rate of organic reactions in micellar water with PTS; the larger size of micelles in the presence of THF has been proposed as the cause of this reaction acceleration.³³ Thus, time, stirring, and cosolvent plausibly impact the interaction of micelles with zinc particles although the nature of these interactions have not been directly observed or characterized.

To observe and characterize these interactions, we now use fluorescence microscopy. As shown in Figure 2b, after stirring for 2 h, the observable diameters of the micelle “hot spots” on the zinc surfaces ranged from ~0.25–2.0 μm (see SI, Figure S2 and Table S1). The upper limit corresponds either to larger micelles or to groups of micelles regionally clustered in specific locations on the zinc surface. This lower limit was diffraction-limited and corresponds to single micelles or small local agglomerates of micelles. Due to the diffraction limit, any fluorescent object smaller than ~250 nm will have an apparent size of ~250 nm in the image, although its actual size may be smaller. Diffraction-limited locations were determined by a 2D Gaussian fit of the point spread functions (PSF) of the potential candidates for single-micelles and small micelle agglomerates.³⁵ This initial data served as a benchmark for determining how time, stirring, and cosolvent changed micelle adhesion.

To investigate the influence of time on micelle adhesion to zinc, a micellar solution was added to zinc powder, followed by 1. The resulting sample was briefly agitated, transferred to a microscopy imaging chamber, and imaged immediately. No micelles were visible on the surface of the zinc, as evidenced by dark zinc particles that lacked green “hot spots” (t = 0; Figure 4a). This data showed that micelle adhesion to the zinc surface is not immediate upon mixing.

Figure 4.

(a) Image at t = 0 and (b) t = 2 h without stirring show limited-to-no micelle adhesion to zinc surfaces. (c) Images at t = 4 h after soaking show large micelles, including ones that form “shells” encapsulating smaller zinc particles. (d) 10% THF cosolvent, t = 2 h, showing micelle adhesion to zinc particles. (e) t = 4 d, soaking, large micelles have settled to bottom of imaging chamber and are diffusing in solution (not attached to zinc surfaces); single-micelle solution diffusion tracks are overlaid.

To investigate the influence of stirring on micelle adhesion to zinc, a micellar solution was added to zinc powder, followed by imaging agent 1. The mixture was soaked (rather than stirred) for 2 h. Upon imaging, zinc particles showed no visible micelles on their surface (Figure 4b; compare with previous Figure 2b at the same time point). Thus, stirring was critical for micelle adhesion at 2 h. This result indicates that the stirring in the reported cross-coupling reaction likely assists the reaction by increasing the rate at which micelles adhere to zinc.

Samples were next soaked for 24 h. Large micelles or micelle agglomerates were visible on the surface of the glass coverslip and attached to zinc particles (bright green shapes, Figure 4c, 2 different samples showing variation). Larger micelles formed thick coatings on a subset of zinc particles, resulting in “shells” fully encapsulating some of the smaller zinc particles. The size of these larger micelles or micelle clusters ranged between 0.9 and 3.5 μm, significantly larger than those observed with fluorescence microscopy at shorter times and with dynamic light scattering observed by Lipshutz³⁶ at shorter times³⁷.

These shells suggest that smaller zinc particles could become fully encased in micelles during the organozinc formation reactions and subsequent palladium-catalyzed cross-coupling reactions; however, this process takes time. Degradation by chemical reaction and mechanical stirring of the zinc powder into smaller particles³⁸, or employment of nanozinc as the starting reagent,^22,39 likely increases the rapidity with which the zinc particles can be fully encased. Full encapsulation plausibly provides enhanced protection the resulting organozinc intermediates, so long as the rate of the zinc insertion reaction into the alkyl iodide is slow relative to the rate of micelle adhesion/encapsulation.⁴⁰

The impact of 10% THF cosolvent on micelle adhesion was next examined. After 2 h stirring with micellar 1, stationary micelles were visible on the surface of the zinc particles (Figure 4d), visually similar to those in the absence of cosolvent (compare Figure 2b). After 4 d soaking (rather than stirring), large discrete micelles or spherically symmetrical micelle clusters of diameter ~1–2 μm were observed diffusing in the solvent, partially trapped between the zinc particles and the glass, not adhered to zinc particles. These micelles are moving slowly enough to both identify and track (see SI Movie 2). Figure 4e shows these micelles overlaid with a subset of their single-particle diffusion tracks. The significance of this observation is two-fold: first, it establishes the ability to see slowly diffusing micelles in solution through this widefield epifluorescence microscopy method and, second, the diffusion tracks show how far these micelles move, which is less than the width of most zinc particles in a few seconds. These data establish an upper bound for how large micelles can become under synthetic conditions (synthetic reaction times up to ~24 h with stirring).¹⁰

Taken together, the experimental data in this section lead to time-dependent aging models for micelle interaction, adhesion, and morphology with zinc surfaces. Specifically, and potentially most impactful for synthetic optimization, these data show that micelle physical adhesion to zinc is not immediate upon mixing. Drawing from these observations: 1) Micelle adhesion is dependent on aging of the micellar solution, providing potential batch-to-batch variability that impacts synthetic reproducibility during cross-coupling reactions and in situ organozinc formation, and 2) Micelle adhesion to zinc may be slow relative to chemical steps in the reaction (e.g., relative to undesired competitive oxidative addition of the substrate to zinc outside of the micelle that results in protodemetalation).

Question 3: Is micelle adhesion increased by the availability of direct insertion/oxidative addition chemistry of organic substrate contents?

In the zinc-mediated cross-coupling reactions reported by Lipshutz, alkyl iodides carried by the micelles undergo direct insertion with the zinc metal in situ.¹⁰ We therefore investigated whether micelles that contain chemically active alkyl iodides as contents adhere to the zinc surface stronger than micelles containing chemically inert contents. Such stronger adherence could occur if, for example, the micelles are partially disintegrated at the metal surface to expose the contents, as has previously been proposed.^19,21 As a result, the subsequent formation of a carbon–zinc bond between material within the micelle and the zinc surface could lead to increased persistence of the residual micellar components on the zinc surface.⁴¹

In order to examine this hypothesis, the degree of adhesion of micelles containing chemically inert imaging agent 1 (X = H), as previously employed, was compared to that of micelles containing direct-insertion reactive imaging agent 2 (X = I)⁴². Experiments were performed under otherwise similar conditions. Micelles containing chemically reactive alkyl iodide imaging 2 agent showed similar size, quantity, and brightness of micelle adhesion to those containing chemically inert 1 (compare Figure 5a with previous Figure 2b). The image shown in Figure 5a is representative from a spatial survey of the sample (see SI, Movie 4). Further, the average fluorescence intensity of zinc particles in samples treated with 1 or 2 were quantitatively compared and there was no statistically significant difference in fluorescence intensity (see SI, Table S3). Thus, the degree of adhesion of the micelles to the zinc surface was independent of the chemical reactivity of the contents of the micelles.⁴³ Further, there was no evidence of oxidative addition of 2 to the surface of the zinc during imaging, as assessed by the similar appearance of images arising from 1 and 2.⁴⁴

Figure 5.

(a) Comparison with direct-insertion reactive alkyl iodide. SEM images of lyophilized zinc sample treated (b) with micelles and (c) in the absence of micelles. (d) Summary of EDS data by %/wt in samples.

A limitation of this conclusion of the chemical independence of the degree of micelle adhesion arises from the fact that small differences in fluorescence intensity are not distinguishable, due to the inherent particle-to-particle variations of the samples (i.e., a range of intensities of particles even within a single sample; Table S3); thus, small differences in micelle adhesion would not be detectable. An additional limitation of this conclusion arises from the low concentrations of 1 and 2 during imaging (220 nM overall, with concentration inside the micelles presumably higher). This concentration was chosen because it produced images with the requisite high spatial resolution and single-particle sensitivity; higher concentrations of imaging agents resulted in “washing out” of the spatial details in the images due to bright signal in the background. The concentration of organic substrates within micelles under synthetic preparative reaction conditions, however, is higher and may have a more pronounced adhesion-impacting effect.

Question 4: Does the sensitivity of in situ fluorescence microscopy enable detection of micelles on zinc surfaces that are not detectable with ex situ SEM/EDS?

Because there had not been previous reports of imaging micelles on irregular metal surfaces by SEM/EDS that we were aware of, but only on smooth metal surfaces, we anticipated that SEM/EDS would be ineffective at detecting micelles on these highly irregular commercial zinc metal powder surfaces used in preparative synthetic reactions. The expected fundamental analytical limitations arise because the natural height/physical variations of the commercial zinc powder used preparatively are anticipated to be much greater than the size of the micelles. Additionally, micelles become somewhat flattened once collapsed under the high vacuum conditions needed for SEM, further reducing the capability to detect them on physically irregular surfaces. Similarly, the commercial zinc powder used preparatively is composed of chemically variable surfaces, with spatially heterogeneous locations of zinc metal, carbonates, and oxides exposed on the surface. We anticipated that this chemical variation would prevent detection of micelles on the zinc surfaces by EDS.

Nevertheless, for evaluation of available analytical techniques, we subjected two samples to analysis by SEM/EDS under conditions that were similar to those of successful micellar imaging by fluorescence microscopy: A micellar (2% PTS) sample with zinc powder, and a control water-only treated sample with zinc powder. Both were separately stirred for 2 h prior to SEM/EDS analysis, identical to the stirring and time that led to imageable micelles on the surface by fluorescence microscopy. Both samples were then lyophilized, to reduce bursting of micelles under vacuum,⁴⁵ and mounted with copper to limit carbon contamination during SEM/EDS.

Figure 5b–d show SEM images and data from surface elemental composition information from EDS (average data from 8 surface measurement points in each sample; error is reported as standard deviation of these measurements (see SI Table S10). As expected, the commercial zinc surfaces were highly irregular physically and showed substantial chemical variation even in the control sample that was not treated with micelles. No significant differences in either physical features or surface chemical compositions at specific point locations occurred between the micelle-treated and water-only control samples of zinc powder; notably, no features were attributable to micelles. Therefore, these data confirm that the higher sensitivity and in situ imaging capabilities of fluorescence microscopy are necessary for observing micelle adhesion to these synthetically relevant highly irregular zinc metal powder surfaces.

Question 5. Do micelles dissociate from zinc rapidly? Are micelles mobile or stationary once adhered to the zinc surfaces?

To examine if micelles rapidly dissociate from the surface of zinc, we designed fluorescence recovery after photobleaching (FRAP) experiments (Figure 6a, b). These experiments were performed on a widefield epifluorescence microscope to enable photobleaching a wide region of the sample for investigation of the rate of recovery. If micelles were stationary, bright zinc particles with micelles at t = 0 would recover fluorescence intensity after photobleaching but dark zinc particles without micelles at t = 0 would not (Figure 6b). In contrast, non-selective recovery of fluorescence on both types of zinc particles would suggest that micelles could exchange with those in solution or with each other on the surface, leading to spatial redistribution during the imaging experiment. Figure 6b shows a schematic of these two scenarios and how the resulting data would be interpreted.

Figure 6.

(a) Experimental schematic for FRAP imaging. (b) Schematic of two hypotheses showing anticipated imaging outcomes of FRAP. (c) Fluorescence images before, immediately after, and 60 min after photobleaching in presence of micelles. Created with BioRender.com.

This side-by-side comparison established that fluorescence recovery was selective (Figure 6c). Further, the recovered particles have similar spatial patterns before and after photobleaching. These observations indicate that the micelles are “sticky”: They remain on the surface of the zinc, at or near their starting location for the duration of imaging (60 min).

Question 6: Do micelles exchange organic contents (e.g., substrates, intermediates, products) while on the surface of zinc?

We designed a series of crossover experiments to observe the rate of exchange of organic “cargo” between micelles on the zinc surface. The goal of these experiments was to determine if mass transport of material between different micelles and from micelles to the zinc surface could be a limiting factor in the overall rate of organozinc formation and subsequent cross-coupling under preparative conditions.

Although molecular exchange kinetics can be measured by alternative analytical techniques when the micelles are in solution,⁴⁶ two relative rate hypotheses were considered here that are unique to micelle interactions with the zinc surface and therefore without prior measurement. Hypothesis 1: Surface micelles exchange material with micelles in solution at a slower rate than micelles dissociate from zinc particles. Hypothesis 2: Surface micelles exchange material with micelles in solution at a faster rate than micelles dissociate from zinc particles.

In order to determine which hypothesis is correct, we examined the rate of a two-color two-fluorophore crossover of the molecular contents of micelles on the zinc surface. For these experiments, we introduced spectrally complementary orange imaging agent 3. Zinc particles were treated with solutions containing green imaging agent 1 or separately orange imaging agent 3. Each separate sample was stirred for 2 h under identical conditions except for the identity of the imaging agent. This process formed two separate samples with green- or orange- micelle coated zinc particles. After 2 h, the supernatant was removed from both samples to reduce fluorescence background. (Attempted imaging without this supernatant removal step resulted in images where the details on the zinc surfaces were “washed out” by the too-bright background.) A clean 2% micellar PTS/water solution without fluorophore was added to both samples, to restore the presence of micelles in the sample. The green and orange samples were then combined and swirled to mix. If mass exchange between micelles or particles occurred slowly upon mixing, then separate green-micelle- and orange-micelle-coated zinc particles would be visible (Figure 7a). If mass exchange between micelles or particles occurred rapidly upon mixing, then only yellow-micelle-coated zinc particles would be visible (a combination of green and orange appears yellow).

Figure 7.

(a) Proposed mechanism for micelle content exchange on zinc. (b) Experimental schematic and epifluorescence images, showing discrete green- and orange-micelle zinc particles (c) Confocal microscopy images; diffraction-limited single micelles in red tringles diffusing in solution, and micelles on zinc surfaces shown in red circles. (d) Intensity ratio of green-to-orange with time. Some individual locations (0.37 × 0.37 μm²) corresponding to single micelles or small clusters of micelles show unique exchange kinetics, distinct from whole-particle averages. Smoothed data contours shown for ease of visualization of trends. Created with BioRender.com.

Widefield epifluorescence microscopy images taken shortly after combining (t = 4 min) showed distinct green and orange particles (Figure 7b). The persistent color integrity shows that micelles stay on particles even with agitation, and do not fully exchange their contents before imaging. Thus, mass transport of the contents of the micelles on the surface of the zinc is not immediate upon mixing.

While the widefield epifluorescence images show zinc particles with distinctly green or distinctly orange micelles on the surface, the fluorescence background was too bright to glean quantitative data on the rate of exchange from the images (e.g., see bright background in Figure 7b). For this reason, we switched to confocal fluorescence imaging. Two-color confocal microscopy imaging was performed with a 485 nm excitation laser (well-matched to excite green imaging agent 1) and a 532 nm excitation laser (to excite orange imaging agent 3). Emission from each fluorophore with each laser was detected using time gating and spectral filters in “green channel” and “orange channel” detectors.

This technique enabled visualization of the micelles on the surface with substantially reduced background fluorescence compared to epifluorescence microscopy. The background was reduced to the degree that removal of the supernatant and replacement with clean micellar solution was no longer required prior to imaging. Due to this low background, excellent spatial resolution and spectral assignment of micelles on the surface of the zinc was achieved.

As with imaging from the prior widefield epifluorescence instrument, images obtained shortly after combining separate green- and orange-micelle samples showed zinc particles that predominantly retained their source colors (Figure 7c). Rates of molecular content exchange were quantitatively measurable due to the absence of confounding background signal. Quantitative kinetics measurements were performed by measuring intensity values in the green channel and intensity in the orange channel. For ease of consideration, the ratio of these two values was obtained by dividing the intensity in the green channel by that in the orange channel. Change in this ratio with time indicated mass exchange (i.e., an increase in the ratio indicated the region was becoming enriched in green 1, and a decrease indicated enrichment in orange 3).

These samples showed all possible micelle dynamics occurring in real time as molecular exchange equilibria were reached: Whole particles that started green became more orange. Particles that started orange became greener. (The combination of green and orange appears yellow in display images.) Surprisingly, some individual particles showed adjacent regions that diverged independently into two different colors. The top image series in Figure 7c shows multiple zinc particles that started mostly orange. Yet, two of these particles (top left and top right) developed pronounced “stripes” of both orange and green, with clearly distinct local regions and behaviors. A different particle in the same field of view (bottom right) more uniformly shifted more orange without developing stripes or other distinct color regions. The lower image series in Figure 7c shows a zinc particle that started mostly green and then gradually became more orange overall, without developing distinct color regions. Thus, all possible crossover combinations were observed. These exchange processes appeared to reached equilibrium in ~2.5 h, after which little additional change was observed.

Mass/color equilibria were reached while individual “hot-spots” of single micelles or small clusters of local micelles remained present through the course of imaging. This observation is most consistent with Hypothesis 2: Surface micelles exchange material with micelles in solution at a faster rate than whole micelles migrate off zinc particles. These experiments examine behavior of the micelles under conditions where the surfaces of the zinc particles are not concurrently degraded chemically (as would happen during oxidative addition/cross-coupling), but nevertheless provide relevant insight into the possible spatiotemporal distribution of exchange behavior of micelles.

To interpret these crossover exchange data quantitatively, control experiments were performed under otherwise identical conditions but with samples of only orange or only green micelles (Figure 8a, b). These control samples showed that the green-micelle samples were intrinsically brighter by ~6 times. Thus, a mixed two-color sample with a surface region displaying a ratio of ~6:1 green-intensity-to-orange-intensity had approximately a 1:1 molar ratio of 1:3. Single color controls also showed a substantial variation in brightness per zinc particle, consistent with the heterogeneous nature of the samples and with the prior epifluorescence images. Finally, critical to the ultimate interpretation of the data, these single-color controls showed stable, unchanging color compositions on the micellar surfaces of the zinc particles (Figure 8b). (Figure 8b shows intensity values rather than ratios for the green-only control on the y-axis, because some orange intensity values were zero, and therefore the ratio of green:orange intensity was undefined; see SI Tables S7 and S8 for raw data).

Figure 8.

Single-color control experiments. (a) Fluorescence microscopy images of green 1 only and of orange 3 only. (b) Intensity values of green 1 from a and ratio of green:orange from b over time, showing intensity stability.

Next, individual locations on the surfaces 0.37 × 0.37 μm² were analyzed, corresponding to individual hot spots and thus to plausible single micelles or small local clusters of micelles. Remarkably, these hot spots exchanged contents at individualistic rates that were both different from their neighbors and from the whole-particle average exchange rates (Figure 7d). For example, the left graph in Figure 7d shows that neighboring locations on the sample particle in the range from no change (location 2) to net increase in green (location 3), to highly variable (location 4); all individual locations differ from the gradual increase in green in the whole-particle average. The apparent increases and decreases in location 4 are consistent with quantized single micelle or small numbers of changing micelles in the same location. Each micelle contains many fluorophores, such that the quantized kinetics behavior arises from small numbers of discreet micelles, not from small numbers of single fluorophores/single molecules. In contrast, the right graph in Figure 7d shows locations 6–10, each reasonably well-described by a shift to orange. Thus, the average exchange rate can be a poor descriptor of the true variation of behavior of individual micelles or local clusters of micelles. The reasons for these individualistic behaviors are not yet fully understood, but they may be caused by different degrees of carbonate and oxide coatings on the surfaces of these commercial zinc particles as previously characterized by SEM and EDS,⁴² or by the variable sizes and shapes of the micelles, which prior reports suggest range in size from 10–50 nm and can form either spheres or nanorods³⁶. Graphs in Figure 7d show different initial measurement time points due to particle settling, leading to different x axis time ranges in Figure 7d.

Importantly, many single micelles were observed as diffraction-limited spots (~250 nm), freely diffusing in solution in the spaces between zinc particles (see SI, Movie 2 and Table S2). These single micelles are visible as solution-phase “hot spots” in individual time frames in the top image series in Figure 7c; two of many examples are noted by red triangles (see also Supplementary Movie). The green-intensity-to-orange-intensity data of single micelles is captured for one time frame before the micelle diffuses out of the focal plane. For example, the single micelles denoted in red triangles in Figure 7c had intensity ratios of 4.0 and 6.3, corresponding to mostly orange contents and similar amounts of green and orange contents, respectively. This measurement is the characterization of the contents of single micelles^47–49 under conditions relevant to preparative organic chemistry.

Diffraction-limited spots, many of them candidates for single micelles, were also clearly visible on the surface of the zinc particles (Figure 7c). These surface spots, however, may alternatively arise from local clusters of micelles with sizes below the diffraction limit. Assuming that fluorescence intensity is proportional to number of micelles allows its use to estimate the true number of micelles in a given diffraction-limited area of the zinc surface. Thus, to estimate the number of micelles in locations 1–10, the fluorescence intensities of locations 1–10 were compared to the fluorescence intensities of the single solution-phase micelles noted in red triangles in Figure 7c.

The intensity values for locations 1–5 ranged from similar to the single micelles in solution to about 3 times as bright, corresponding to between 1 and 3 micelles per location at t = 55 min. In contrast, the intensity values for locations 6–10 ranged from 5–7 times the intensity values of single micelles in solution, corresponding to between 5 and 7 micelles per location at t = 55 min. The estimation of the number of micelles using intensity proportionality is primarily limited by two factors: 1) The inherent variation in intensity of individual micelles, and 2) The unknown degree to which attachment near zinc may partially quench fluorescence, even when fluorophores are within micelles, leading to dimmer than expected micelles on zinc. Lending validity to this proportionality assumption, however, is that the equilibration kinetics for locations 1–5 exhibit abrupt ratio jumps consistent with single-micelle or small collections of micelles, as plausibly expected for about 1–3 micelles (Figure 7d). In contrast, locations 6–10 exhibit more uniform kinetics without abrupt ratio jumps. This behavior is more representative of the particle average and is consistent with a larger number of micelles per location, plausibly 5–7 micelles. Thus, the exchange kinetics and the intensity estimates for numbers of micelles are in general agreement.

A control experiment was performed in which imaging agent 1 and imaging agent 3 were physisorbed to zinc surfaces by stirring in the absence of micelles (see SI Figure S25). These samples were then combined and imaged under otherwise identical experimental conditions. These control experiments showed no time-dependent change in the ratio of green-to-orange fluorescence intensities on the surfaces of zinc (see SI Figure S26). Thus, micelles were required for the change in color ratios on the surfaces of the zinc as previously observed. The mechanism of color change is therefore attributed to the mass transport activity of the micelles. A plausible mechanism for this exchange, consistent with this suite of data, is thus proposed (Figure 7a): 1) Micelles from the original color solution remain attached for the duration of imaging. 2) Single micelles from solution land on top of locations with prior micelles and exchange or add organic contents. 3) Equilibrium is reached (in ~2.5 h under imaging conditions). 4) Individual micelles or small local clusters of micelles exchange molecular contents by unique kinetics (i.e., by kinetics that are poorly described by whole-particle averages).

Question 7. Are the mass transport physical processes involving the micelles slow enough to limit the rate of the overall cross-coupling reaction kinetics?

Plausibly, yes. Imaging data from time and from stir-vs-soak experiments (Figure 4) show that micelle adhesion to zinc is not immediate upon addition of zinc to micellar PTS solution. Imaging data from FRAP experiments (Figure 6) show micelles present in similar locations on the same particles for 60 minutes, suggesting that once adhered, micelles are “sticky” and slow to dissociate. Imaging data from two-color crossover experiments (Figure 7) show that full exchange of contents of organic material inside of micelles on zinc surfaces with material from micelles in solution is not rapid upon mixing, meaning that the original micelles are still present with their molecular contents for extended times. While these mass transport processes are likely to be faster in stirred systems, the data herein show that several micelle mass transport processes are reasonable contributing factors to the overall reaction barriers of the reported preparative cross-coupling processes.

Conclusion

Mechanistic information learned.

The exquisite sensitivity and in situ imaging capabilities enabled mechanistic lessons extendable to the preparative reaction system (Figure 9): 1) The observed adhesion of micelles to the surface of the zinc particles is consistent with the hypothesis made by Lipshutz wherein the micelles protect the insertion of the zinc into the sp³ carbon–halogen bonds;¹⁶ such presence of the micelles at the zinc surface had never before been directly observed (Figure 2b). 2) Micellar solutions undergo aging that impacts their interactions with zinc powders (Figure 4b–e). Similarly, micelle adhesion to zinc is not immediate upon mixing but requires time (Figure 4a). 3) Once adhered, the adhesion of the micelles to the surface of the zinc particles is sufficiently strong to allow decanting the supernatant, replacement of the solution with fresh solvent, and imaging >2 h (Figure 7b). Thus, the micelles act as “sticky” mass transport agents, raising the possibility that several steps in the synthetic cross-coupling reaction may occur within one persistent micelle without ejection from the zinc surface. 4) Microscopy reveals heterogeneous behaviors: Some zinc particles are covered in fluorescent micelles, while others are not (Figure 2b). Similarly, the kinetics of molecular exchange display individualistic behaviors that are poorly described by whole-particle averages (Figure 7c,d). These mechanistic conclusions are uniquely available from the analytical technique herein and are not accessible through alternative approaches to studying micelles (e.g., NMR spectroscopy, dynamic light scattering, SEM, EDS, or TEM). These conclusions hold under the current imaging conditions of low substrate concentration and in the absence of TMEDA.

Figure 9.

Overall conclusions from microscopy. Created with BioRender.com.

These observations suggest new granular additions to the mechanistic picture originally shown Figure 1. The micelles buildup gradually with time (Figure 10), in spatially heterogeneous fashion and with particle-to-particle variations (including possible full encapsulation of smaller zinc particles). Early reaction times may be more prone to oxidative addition of alkyl halide to the surface of the zinc in the absence of micellar protection, resulting in increased protodemetalation; later reaction times or otherwise increased coatings of the zinc particles may result in increased protection of water-sensitive organozinc reagents from protodemetalation of intermediates on or near the zinc surface (Figure 10).

Figure 10.

Schematic of micelles accumulating over time and plausibly protecting water-sensitive organozinc surface species from protodemetalation. Created with BioRender.com.

Technical achievements.

Single-micelle imaging is achieved. Single micelles are imaged in solution and with concentrations of surfactant and with commercial zinc powder similar to preparative conditions. The sensitivity and spatiotemporal resolution allowed characterization of the molecular contents of these single micelles as they diffused in solution (Figure 7c). Resolution of single micelles and/or small clusters of local micelles on the surface of zinc is also achieved, as is measurement of local, non-ensemble averaged molecular exchange kinetics (Figure 7c,d). Many micelles overlap on the zinc surfaces such that each fluorescent signal on the surface does not arise from a single micelle. What is interesting, in our view, is that a subset of these signals is diffraction-limited. Combined with kinetics data, these signals are candidates for single-micelles and/or small clusters of micelles. This spatial resolution and sensitivity far exceed anything previously for studies of organometallic chemistry in micellar solutions.

Suggestions for improvements of bench-scale preparative reactions.

These mechanistic observations underpin specific suggestions for optimization of synthetic organozinc and cross-coupling reactions under aqueous conditions. For example, it may be helpful to premix micellar solutions with zinc powders prior to addition of organic substrate to enable pre-encapsulation of the zinc. Alternatively, it may be helpful to “age” micellar solutions for days before use to allow larger micelles to form, which are better able to encapsulate zinc. Both options may reduce undesired protodemetalation pathways by limiting the exposure of sensitive organozinc intermediates to water. Such options are particularly promising for improving the performance of micellar solutions with highly sensitive substrates such as tertiary alkyl halides¹⁰ or with preparations employing highly reactive nanozinc that currently may lead to undesired competitive protodemetalation⁴⁰. Ongoing work includes examination of the palladium-catalyzed step in this preparative reaction system, higher concentrations of substrates, and the effect of different surfactants on micelle adhesion.