Surfactant Micellar and Vesicle Microenvironments and Structures Under Synthetic Organic Conditions

Abstract

Fluorescence lifetime imaging microscopy (FLIM) reveals vesicle sizes, structures, microenvironments, reagent partitioning, and system evolution with two chemical reactions for widely used surfactant–water systems under conditions relevant to organic synthesis, including during steps of Negishi cross-coupling reactions. In contrast to previous investigations, the present experiments characterize surfactant systems with a representative organohalide substrates at the high concentrations (0.5 M) that are reflective of the preparative-scale organic reactions performed and reported in water. In the presence of representative organic substrates, 2-iodoethylbenzene and 2-bromo-6-methoxypyridine, micelles swell into emulsion droplets that are up to 20 μm in diameter, which is 3–4 orders of magnitude larger than previously measured in the absence of organic substrate (5–200 nm). The partitioning of reagents in these systems is imaged through FLIM—demonstrated here with nonpolar, amphiphilic, organic, basic, and oxidative-addition reactive compounds, a reactive zinc metal powder, and a palladium catalyst. FLIM characterizes the chemical species and/or provides microenvironment information inside micelles and vesicles. These data show that surfactants cause surfactant-dictated microenvironments inside smaller micelles (<200 nm), but that addition of a representative organic substrate produces internal microenvironments dictated primarily by the substrate rather than by the surfactant, concurrent with swelling. Addition of a palladium catalyst causes the internal environments to differ between vesicles—information that is not available through nor predicted from prior analytical techniques. Together, these data provide immediately actionable information for revising reaction models of surfactant–water systems that underpin the development of sustainable organic chemistry in water.

Graphical Abstract

INTRODUCTION

Aqueous surfactant mixtures provide a promising sustainable approach for performing organic reactions in water instead of in organic solvents. Recently, applications of this approach have been developed in both academic^1–9 and industrial settings,^8,10–15 including on scale, encompassing a range of modern synthetic reactions (e.g., oxidative addition to metal powders, palladium-catalyzed cross-coupling,^8,11,16–19 amination,^20,21 and sulfonations¹²). Yet, optimization of these surfactant-based reaction systems remains largely screening-based, including selection of which surfactants and concentrations in each component. Considerable effort has thus been applied to improving the fundamental understanding of these systems so as to advance beyond this purely empirical approach;^22,23 however, fundamental understanding remains sparce, primarily due to analytical limitations that have so far restricted studies to idealized systems or calculations.

For example, the size of micelles made from commonly applied surfactants in these organic-synthesis systems have been measured by cryo-transmission electron microscopy (cryo-TEM)^3,12,24 and dynamic light scattering (DLS)^12,24–26 to be ~5–169 nm (Table 1). These size measurements, however, were necessitated to be performed on idealized samples that contained only the surfactant and water (and sometimes organic cosolvent) but without additional reaction components. Most notably, the organic substrate, a key component that typically comprises a high concentration of the reaction mixture, was missing. These prior measurements also required characterization at significantly lower surfactant concentrations than those employed under synthetic reaction conditions. DLS experiments are fundamentally limited to dilute, transparent surfactant solutions in order to permit sufficient light scattering, and were previously performed at 0.02–0.3 wt % surfactant,^27,28 whereas typical synthetic conditions are 10–100 times that (2 wt %)^4,19. A publication with photographs of multiple systems under synthetic conditions clearly shows optically opaque reaction mixtures that are not amenable to DLS characterization,²⁷ similar to such reaction mixtures in our own hands. Further, because DLS is an ensemble technique, it misses spatially resolved location information such as reagent partitioning. Cryo-TEM is an ex situ technique that requires cold temperatures and high vacuum, making it incompatible with representative substrate concentrations (typically 0.01–1 M^3,11,19,29) and incapable of characterizing chemical and physical changes to the systems in real time. Because of these marked differences in conditions, the idealized systems suitable to prior analytical techniques and the synthetically relevant systems may be particularly distinct.

Table 1.

Surfactant Structures and Size Comparison (Previously by DLS/Cryo-TEM without Substrate and/or Low Surfactant %, and Here by FLIM Under Synthetic Conditions)

Surfactant	Chemical structure	Micelle/Vesicle Diameter
		Previously in absence of substrate (nm)	Here with 0.5 M 2-iodoethyl benzene (nm)
PS-750-M11		169	≤ 20000
PTS3		24	≤ 20000
TPGS23		35	≤ 15000
Brij© 3023		110	≤ 25000
Brij© 3524		15	≤ 20000
CTAC25		5	≤ 10000
none	-	-	≤ 15000

Given the importance of size, structure, and composition on determining the partitioning and microenvironments of reactants, and thus reasonably on reaction kinetics and selectivity, these distinctions have the potential to govern the true chemical reactivity of these systems. Further, such distinctions may dictate system stability with time and thus reusability/recyclability. To date, several conceptual and calculation-derived models have been put forth.^30–32 Crucially, these models have largely extrapolated from ideas that the micelles retain similar nano size ranges in the presence of reaction components and at reaction concentrations (substrate, reagents, catalyst, metal powders, higher surfactant concentrations, etc.), which we here show is not necessarily the case. In 2018, for example, a calculated model suggested that the surfactant TPGS-750-M produces clustered micelle aggregates with ca. 33% water in the interior and with a total diameter of 50–60 nm, and this model was used to rationalize reactivity in related synthetic systems.²⁴ The authors note, “The internal micellar shape…is unknown and cannot be predicted.”²⁴

Thus, the prior studies did not provide relevant size, nor partitioning, nor spatially resolved environmental information, nor reveal real-time synthetic system evolution information with reagent addition—knowledge gaps filled by this study. This fundamental information is critical for the development and improvement of aqueous–surfactant systems for synthetic organic chemistry. Fluorescence lifetimes of small-molecule organic fluorophores are exquisitely sensitive to their individual microenvironments.^33,34 Fluorescence lifetime imaging microscopy (FLIM) offers high spatial resolution. The spatial resolution of the imaging technique is key for unmasking potential system heterogeneities that would otherwise be lost to ensemble averaging in other forms of analytical measurements. FLIM has not been applied previously to the study of aqueous–surfactant systems at relevant concentrations of organic substrates. We hypothesized that these characteristics may combine to reveal previously hidden size, structural, environmental, and reagent partitioning information in these aqueous–surfactant systems under conditions relevant to organic synthesis (e.g., high concentrations of representative organic substrate, 2% surfactant, ambient temperatures and pressure). This FLIM approach is here developed, and the resulting system information thus revealed is described.

RESULTS AND DISCUSSION

Imaging with amphiphilic imaging agent 1.

Studies were initiated using commercially available imaging agent 1 (at 1 μM). Imaging agent 1 contains a boron dipyrromethene (BODIPY) fluorophore core and a carboxylic acid tail. The BODIPY core is a well-established hydrophobic fluorophore.^35–43 This fluorophore class was chosen for present studies due to its hydrophobicity (which makes it a good model for representative organic additives), high quantum yield, small size relative to the micelles/vesicles, and established chemical inertness.^{30,38,42–48} Specific amphiphilic imaging agent 1 was of particular interest due to the hydrophilic carboxylic acid group combined with its hydrophobic BODIPY core—possessing an amphiphilicity that could inform on interactions between organic substrate, surfactant, water, and reagents. 2-Iodoethylbenzene was chosen as a model organic substrate at 0.5 M, representative of the organohalide oxidative-addition partners and high substrate concentrations used preparatively in carbon–carbon cross-coupling reactions in water in the presence of zinc, palladium catalyst, and surfactants.¹⁹ Six surfactants, PS-750-M, PTS, TPGS-750-M, Brij-30, Brij-35, and CTAC, were chosen for initial study due to their application in organic synthesis reactions in water and their representative range of structures.^{2,13,14,17–19,49,50}

Surfactant solutions (1.8 wt % in water) with amphiphilic imaging agent 1 (1.1 μM) and representative organic substrate 2-iodoethylbenzene (0.5 M) were stirred for 2 h to mimic synthetic reaction conditions¹⁹ and then transferred to a microscopy vial and imaged with time-resolved confocal fluorescence microscopy (Figure 1). A comparison sample was prepared similarly but by using water in the absence of any surfactant. Images were obtained at a depth of ~3 μm above the imaging vial bottom. The FLIM images shown in Figure 1 are representative of spatial surveys of each sample (additional survey images are available in the SI).

Figure 1.

a. Structure of amphiphilic imaging agent 1. b. FLIM images of 1 with 0.5 M 2-iodoethylbenzene, surfactants and no surfactant; vesicle sizes, vesicle shapes, reagent partitioning, and microenvironments are characterized. Histograms of fluorescence lifetime pixel distributions for each image.

The intensity component within each image shows the location and thus partitioning of amphiphilic imaging agent 1 within that surfactant–water–substrate mixture. In these images, vesicles appear as roughly spherical shapes against a background in each sample. We here use the phrase “vesicle” for the larger emulsion droplets that form under these conditions due to their substrate-carrying properties.^51–54

Immediately noteworthy in the images in Figure 1 is that the sizes of the vesicles are ca. 3–4 orders of magnitude larger than previously measured in the absence of substrate, for all examined surfactants (Table 1).^27,28 Thus, in contrast to previously proposed models,^24,27,28,51 addition of representative organic substrate at relevant concentrations of substrate and surfactant is sufficient to cross from a nano regime to a micro regime.⁵¹ Vesicles are also visible in the no-surfactant sample. Further, all seven systems show a range of vesicle diameters within each sample, demonstrating a previously uncharacterized size heterogeneity within single surfactant samples. For PS-750-M, a range of sizes had been previously inferred for different mPEG chain lengths by NMR spectroscopy,¹³ and characterized by ex situ TEM,^13,14 but this range remained in the nanoscale regime and both were characterized in the absence of organic substrate.^13,14,28 A limitation of the fluorescence microscopy imaging approach for size measurement is that larger vesicles are more likely to settle to the bottom and be imaged; thus, this imaging method provided an upper limit for vesicle size in the full sample. This partitioning is stable (unchanged) to 45 min after transfer to the microscopy vial without further stirring (see SI Figure S9).

Patterns of inter-vesicle behavior begin to become apparent in the images in Figure 1: for example, the cationic surfactant CTAC reproducibly produced smaller vesicles that adhered to larger vesicles. The reasons for this “stickiness” are not yet understood, but this behavior may impact chemical processes, e.g., by accelerating mass-transport kinetics between vesicles.

Strikingly, for surfactants PTS and TPGS-750-M, amphiphilic imaging agent 1 strongly favors partitioning inside the vesicles, as shown by the bright fluorescence intensity inside the circular vesicles; however, for the samples with Brij-30 and for no surfactant, the opposite partitioning occurs, with 1 favoring the exterior aqueous environment, as evidenced by the dark spheres against bright backgrounds. The similar partitioning induced by Brij-30 and no surfactant may plausibly be due to Brij-30 being a shorter-chain surfactant compared to the others—that is, it may create an interior-vesicle environment most similar to that when “no surfactant” is present. Thus, the interior chemical/physical environment in both may be similarly described as that dictated by 2-iodoethylbenzene. The reason for the “blurrier” boundaries of the vesicles in the image of the no-surfactant sample compared to the Brij-30 sample is currently not fully understood but is plausibly assigned at least in part to faster motion/motion blur during imaging of the vesicles diffusing in the absence of surfactant, due to the lower overall viscosity of the medium. This substantially increased motion is visible in movies and in the apparent distortion of the spherical shapes (see especially Figure 3, vide supra).

Figure 3. Chemical reactions change reagent partitioning and microenvironments.

(a) Schematic of chemical process. FLIM images of 1 (1 μM), 2-iodoethylbenzene (0.5 M), and (b) PS-750-M (1.8%) or (c) TPGS-750-M (1.8%). Addition of TMEDA (0.2 M) deprotonates the carboxylic acid of 1, changing partitioning of the carboxylate of 1 into the aqueous layer, followed by addition of metallic zinc powder that shows surface-coordination consistent with the carboxylate of 1. All images in a given reaction sequence are displayed at identical brightness–contrast settings.

PS-750-M and Brij-35 result in more equitable partitioning of 1 between the exteriors and interiors of their respective vesicles, as observed by both the bright background and bright vesicle interiors. The distinct partitioning in each system can be uniquely characterized by the sensitivity and spatial resolution of FLIM, with information and characterization resolved for single vesicles. The mechanistic cause of the difference in partitioning in each system is the differential polarity/hydrogen bonding capability of the external medium and internal vesicle environment as created by each surfactant.

Together, these data show that reagent partitioning is surfactant-dependent. These differences in reagent partitioning caused by surfactant choice may be particularly impactful on reaction rates and selectivities when, e.g., promotors, ligands, or catalysts are added to reaction mixtures and their potential colocalization with other reaction components changes.

Microenvironment information from FLIM.

The images in Figure 1 are rainbow-false-colored to display the fluorescence lifetime of amphiphilic 1, which is indicative of its local microenvironment. In turn, this information enables environmental sensing. Longer fluorescence lifetimes are displayed as red, and shorter fluorescence lifetimes are displayed as blue.

Upon acquisition of the data in Figure 1, it was readily deduced that 1 displayed markedly distinct lifetimes inside and outside of the vesicles, enabling facile assignment of location/partitioning of multiple reaction components. Specifically, 1 displayed longer fluorescence lifetimes in the exterior of the vesicles (the aqueous phase), appearing red (τ₂ ~5–6 ns) in the aqueous backgrounds of the PS-750-M, Brij-30, Brij-35, CTAC, and no-surfactant samples, and shorter lifetimes, appearing blue (τ₁ ~2–2.8 ns) in the organic phase in the centers of spheres in PS-750-M, PTS, TPGS-750-M, Brij-35, and CTAC.

The relative contributions of these two environmental-sensing lifetimes to the pixel distribution of images for each system are displayed in histograms alongside the FLIM images in Figure 1. Each histogram shows intensity average fluorescence lifetime (τ_ave, black), τ₁ (orange) and τ₂ (blue), derived from biexponential fitting. No data in Figure 1 was well-fit by a monoexponential, even those that did not show two distinct populations that were separated by a visual gap in their respective histograms. Surfactants that caused the most even partitioning of amphiphilic 1 into both interior and exterior phases showed the most visually pronounced bimodal distributions. Triexponential fitting did not improve the fit (see SI Section 2.5), leading to the conclusion that the apparent green halos observed around vesicles were caused by the spatial overlap of the original two microenvironments and not by a third, spatially distinct microenvironment. The impact of this conclusion is to indicate that the interiors of the vesicles are spatially uniform in structure, a piece of information that was previously missing from reaction models.

The assignment of vesicle interiors as predominantly composed of 2-iodoethylbenzene in all systems was made as follows: 1) In separate bench-scale experiments, 2-iodoethylbenzene was found to partially quench the fluorescence of the BODIPY core, providing a physical mechanism by which environmental proximity to 2-iodoethylbenzene would lower the lifetime of 1 in a concentration-dependent fashion (see SI for Stern-Volmer plot); and, 2) Dissolution of BODIPY fluorophore in neat 2-iodoethylbenzene in the absence of any additional components gave rise to τ_ave = 2.7 ns, strongly supporting its assignment as the lower lifetime environment sensed by 1 (Table 2). A similar environmental–lifetime consideration led to the assignment of the exteriors of the vesicles as containing very little 2-iodoethylbenzene in all studied systems. Thus, the partitioning of this representative organic substrate could be clearly determined without the need to attach an exogenous fluorophore to the representative organic substrate itself. The availability of partitioning information for multiple components without the requirement to chemically label each component accelerates the analysis of aqueous–organic surfactant systems by FLIM.

Table 2.

Fluorescence lifetime (τ_{ave_int}) of 2 characterizes environment in presence and absence of 2-iodoethylbenzene

Surfactant	Fluorescence lifetime with substrate (ns)	Fluorescence lifetime without substrate (ns)
PS-750-M	2.70 ± 0.07	5.94 ± 0.01
PTS	2.78 ± 0.03	5.69 ± 0.01
TPGS-750-M	2.65 ± 0.02	5.4 ± 0.2
Brij-30	2.96 ± 0.06	5.89 ± 0.10
Brij-35	2.70 ± 0.01	5.83 ± 0.12
CTAC	2.68 ± 0.03	5.88 ± 0.10
No surfactant	2.74 ± 0.02	5.22 ± 0.03
Neat (no water)	2.67 ± 0.01
Lifetime range	0.31	0.72

The summary of FLIM environmental data presented in Figure 1 is that different surfactants produced different partitioning of the same reagent (1). Thus, the nature of the surfactant dictated the environment of 1, as inside the vesicle (hydrophobic environment), outside the vesicle (hydrophilic environment), or both. At the same time, the representative organic substrate, 2-iodoethylbenzene, remained nearly exclusively inside the vesicles. This differential partitioning of reagents dictated by different surfactants may affect chemical reactions on the bulk scale, for example, by altering reaction chemo- or stereoselectivity by changing the (co)locations of reagents, additives, ligands, or molecular catalysts.

Imaging with hydrophobic imaging agent 2. The partitioning and environmental assignments were further supported by investigation of the same water–surfactant–2-iodoethylbenzene systems, but next with nonpolar, non-hydrogen-bonding imaging agent 2 (at 220 nM; Figure 2). BODIPY 2 has no polar functional groups and therefore cannot participate in hydrogen bonding, providing a complementary partitioning and environmental probe of the vesicle system. In all cases, 2 partitioned inside the vesicles—no significant partitioning of 2 into the aqueous exterior was observed regardless of surfactant nature or absence. This inside-the-vesicle exclusivity is observable both in the images themselves and in their respective histograms (e.g., absence of “double hump” or “shoulder” distributions). As previously with 1, the spatial distribution of lifetimes of 2 was uniformly distributed within the interior of the vesicle, suggesting uniform interior structure (see SI Section 2.5).

FIGURE 2.

a. Structure of hydrophobic 2. b. FLIM images showing near-exclusive partitioning of 2 inside vesicles regardless of surfactant; vesicle sizes, shapes, and chemical composition, and microenvironments are also characterized. Histograms of fluorescence lifetime distributions with accompanying biexponential best fit.

The fluorescence lifetimes of 2 in the interiors of the vesicles were within a narrow 0.3 ns range (Table 2; (τ_ave = 2.5–2.8 ns) regardless of nature or presence of surfactant. The standard deviations shown in Table 2 are derived from multiple measurements of different regions of the same sample to account for variations due to the heterogeneous nature of the sample. The uniformity of data in Figure 2 indicates a general surfactant independence on the partitioning and accessed microenvironments of nonpolar reagents, and thus indicates that reactions with hydrophobic reagents/additives/ligands/catalysts may be less effected by surfactant choice than are amphiphilic ones, at least when paired with a nonpolar substrate.

For comparison, fluorescence lifetimes were also measured in the absence of 2-iodoethylbenzene (Table 2). In the absence of 2-iodoethybenzene, the average lifetimes were strikingly longer (τ_ave = 5.2–5.9 ns) than in its presence and the range was somewhat larger between different surfactants (Δτ_ave ~0.7 ns). Two noteworthy points arise from this data: 1) The physical/chemical environments inside micelles/vesicles with surfactant alone is clearly and measurably different than when representative organic substrate is also present; and, b) The somewhat broader range of lifetimes reveals that the inside environment may be slightly more surfactant-dependent in the absence of organic substrate. We interpret these data as suggesting that high concentrations of substrate appear to partially “overwhelm” environments that would otherwise arise from the different surfactants.

Prior work in our laboratory established that in the absence of relevant concentrations of organic substrate, smaller micelles are formed, which diffuse rapidly and present challenges to imaging.^41,55 For this reason, the comparison data in Table 2 in the absence of substrate was obtained via a 30 s point-time-trace measurement, a technique that characterized species diffusing through a stationary confocal volume. Thus, these values contain a significant contribution of aqueous-phase measurement. Differences in microenvironment caused by different surfactants are nevertheless reflected by the range of the data (Δτ ~ 0.7 ns). Values in Table 2 are reported as the average and standard deviation of three or more different regions/images or point-time-trace measurements to account for the heterogeneity of the samples.

In all studies, the BODIPY imaging agents were present in low concentration (1.1 μM in 1 and 200 nM in 2). In contrast, the organic substrates were present at 0.5 M. This means that the ratio of substrate–imaging agent is ~500,000:1 (for 1) or ~2,500,000:1 (for 2). Similarly, the surfactants were present at 20–60 mM (1.8 wt %). This means that the ratio of surfactant–imaging agent was ~50,000:1 (for 1) or ~150,000:1 (for 2). At these ratios, the majority of supramolecular effects are anticipated to be dictated by the substrate and surfactant and not the imaging agents.

Evolution of Chemical Systems: Reaction Monitoring

Traditional methods of monitoring reactions with time in surfactant systems provide product yield and ex situ information about catalyst speciation,⁵⁶ but do not provide critical information about real-time changes in environment or reagent partitioning with reaction progress. This dynamic environment and partitioning information is a foundation for mechanistic insight into the evolution, stability, reactivity, and potential medium reuse of these systems. We envisioned that FLIM can fill these gaps. In this manuscript, two demonstrations of reaction monitoring are demonstrated, showing the complementarity of the FLIM method to fill knowledge gaps that exist with traditional analytical techniques.

Reaction-Monitoring Demonstration 1:

Reaction of alkyl iodides with zinc metal, a sequence in Negishi-like cross-coupling reactions. Negishi-like cross-coupling reactions conceptually contain two parallel processes: 1) oxidative addition of an organohalide (typically iodide) to metallic zinc to form an organozinc reagent, and 2) oxidative addition of an organohalide (typically bromide) to a palladium catalyst. The first sequence (reaction of alkyl iodide with zinc) was initially examined. Specifically, the chemical and physical evolution of the PS-750-M and TPGS-750-M systems were investigated in the presence of 1 (1 μM) and 2-iodoethylbenzene (0.5 M) (Figure 3). To these systems was added tetramethylethylenediamine (TMEDA), a chelating ligand previously demonstrated to enhance yields of organozinc reagents in aqueous surfactant systems,^5,17,19,57 and zinc metal powder, previously demonstrated to produce oxidative addition/direct insertion with organohalides (here represented by 2-iodoethylbenzene) in the presence of TMEDA in aqueous surfactant systems.^17,19,58

Figure 3b shows the PS-750-M surfactant system. Upon addition of TMEDA (0.2 M), the carboxylic acid in 1 was deprotonated, with repartitioning of the resulting carboxylate nearly exclusively into the aqueous phase (Figure 3). Both the deprotonation and reagent repartitioning were rapid, having occurred within ~20 s of shaking and ~3 min of refocusing the sample to acquire an “after” image. The chemical composition of the aqueous phase had changed, reflective of an environment containing TMEDA, as reported by the lowered lifetime of 1 in the solution phase after addition of TMEDA (τ = 3.6 ns) than before addition of TMEDA (τ = 4.7 ns). TMEDA is a known quenching agent of this BODIPY core class,⁴¹ which is likely the cause of this lowered lifetime and decreased overall brightness, with these features resulting in an environmental-change-readout from the probe as observed by FLIM. Previous size measurements for the objects resulting from this surfactant remain ≤1 μm even with aggregation (measured in absence of organic substrate),¹³ in contrast to the 5–20 μm sizes occurring at all reaction stages in Figure 3b.

Figure 3c shows the TPGS-750-M system. This surfactant system was previously demonstrated to be particularly effective for Negishi-like cross-coupling reactions in synthesis.²⁰ In this case, deprotonation of 1 upon addition of TMEDA was again accompanied by a rapid repartitioning of the carboxylate into the aqueous layer. In this case, however, the presence of only a small change in the lifetime (τ = 2.2 ns to τ = 2.4 ns) indicated that the microenvironments in the interior and exterior of the vesicles were somewhat similar. This similarity indicates that the more organoiodide partitions into the aqueous layer with TPGS-750-M than with PS-750-M. Comparison of the two data sets in Figure 3b and 3c thus provides previously unobtainable information about surfactant-dependent differences in partitioning and reagent environments during this step of the reaction. Such partitioning and environmental differences may underpin reaction-yield differences with surfactant choice as observed on the bulk scale.²⁰

In both surfactant systems, upon addition of zinc metal powder (1 equiv relative to 2-iodoethylbenzene), an additional repartitioning of 1 occurred. FLIM reported on this repartitioning at the individual-zinc-particle level. Partial scavenging of 1 from the solution phase by the zinc surface occurred (presumably as the zinc carboxylate), as evidenced by the accumulation of 1 on the surface of the three individual and irregular zinc particles (~5–20 μm in length) in Figure 3b, and on one zinc particle in Figure 3c. The images in Figure 3 are representative of sample-wide surveys (additional examples available in SI). Previous control experiments established that the zinc surface is dark in the absence of exogenous imaging agent.⁴¹

Reaction-Monitoring Demonstration 2:

alkyl iodide and bromo pyridines with palladium catalyst, a sequence in Negishi-like cross-coupling reactions. In the prior reaction sequence with zinc in Figure 3, zinc was a heterogeneous reagent, and FLIM enabled imaging the zinc surface. In order to determine if the physical and chemical environments inside different vesicles differ during an ongoing reaction with a homogeneous palladium catalyst, the palladium-mediated portion of the reaction was examined next (Figure 4).

Figure 4.

Reaction system with palladium catalyst revealed different chemical/physical microenvironments in different vesicles (with 2 h stirring). z-Scan FLIM images of 2 (1 μM), 2-iodoethylbenzene, 2-bromo-6-methoxypyridine, Pd catalyst, and PTS (1.6 %/v in water). The red border outlines the enlarged slice.

Specifically, a PTS surfactant reaction system was examined with 2-iodoethylbenzene, 2-bromo-6-methoxypyridine, TMEDA, Pd(amphos)₂Cl₂, and 2 (1 μM), using similar substrates, catalysts, and concentrations to those previously reported for successful catalytic carbon–carbon bond formation in aqueous–surfactant systems.⁵⁹ This reaction system proceeds at room temperature without external heating or cooling, and with alkyl iodides and bromopyridines that are insoluble in water but are enabled for aqueous reaction due to the PTS surfactant, making it a prime example of the power of such surfactant systems.⁵⁹

After stirring at room temperature for 2 h, this reaction system was imaged (Figure 4). The organic substrate-based vesicles again formed, but unlike previous experiments that showed similar lifetimes between all vesicles, now different vesicles displayed different lifetimes (~2.5–3.1 ns; Figure 4). A z-scan was performed in 0.5 μm slices from the sample bottom to 10 μm into the sample. The fluorescence lifetimes of the individual vesicles did not depend on height in z, establishing that an individual vesical was spatially uniform in lifetime in x, y, and z. (Note: The lifetimes are false-colored with a narrower range in Figure 4 to clearly show this distribution; see SI Section 2.7 for comparison with other data at a similarly narrow range, which do not show a distribution in the absence of palladium).

In order to identify the chemical source of the variation in environments, each component was imaged separately, and then in combinations (see SI, Section 2.7). These experiments showed that, at a minimum, the palladium catalyst and 2-bromo-6-methoxypyridine needed to be present for the system to demonstrate different lifetimes between different vesicles (Figures S46–S79), pinpointing this effect to a specific chemical reaction pair. The difference in lifetimes characterize that each vesicle has its own microenvironment and therefore likely has different reactivity—but the order of cause and effect are intriguingly unknown (that is, whether different reactivity caused the different environments, or if the different environments caused/cause different reactivity). For example, because palladium and bromopyridine are both required for the environmental distribution, it is conceivable that oxidative addition to palladium by the bromopyridine may have progressed at different rates in different vesicles, leading to the different fluorescence lifetimes. Regardless of the root cause, these findings are in contrast to current models that present surfactant systems as vesicle-homogeneous.^13,53 These findings highlight the ability of FLIM to spatially reveal such physical and chemical variances in the system and to pinpoint their causes to specific reagents—data that was not previously reported and could not have been obtained through the prior analytical techniques applied to these synthetic organic systems.

CONCLUSION

Development and application of a FLIM technique bridged previous knowledge gaps of micelle/vesicle sizes, reagent partitioning, and microenvironments with aqueous–surfactant conditions in organic synthesis. This technique provided environmental and partitioning information about multiple system components without the requirement to label each component with a fluorophore. These imaging data included characterization of a surfactant-free emulsion system—providing information about a system unexplorable by prior cryo-TEM or DLS techniques due to the absence of stable micelles.

Under synthetically relevant surfactant and organic substrate (0.5 M) concentrations, the sizes of the surfactant vesicles were 3–4 orders of magnitude larger than previously reported in the literature for micelles in the absence of substrate and/or under dilute surfactant conditions (Figures 1 and 2, Table 1). Reagent partitioning was surfactant-dependent for amphiphilic compounds (Figure 1) and less so for hydrophobic compounds (Figure 2), showing that (co)localization changes of additives, ligands, or molecular catalysts may underpin surfactant–reactivity and concentration–reactivity relationships in preparative systems using surfactant.

Addition of an organic base and a reactive metal powder into a vesicle-rich solution caused rapid changes in the partitioning of a carboxylic acid, with environments that depended on the surfactant choice (Figure 3). A system with a palladium catalyst and oxidative-addition reactive bromopyridine revealed different microenvironments between different vesicles (Figure 4). Together, these experiments provide size, partitioning, microenvironment, and system-evolution-upon-reaction information that was not previously reported and could not have been learned through previously ex situ, low-concentration-required, or non-spatially resolved/ensemble averaged analytical techniques. These findings enable immediate revision of models of size (micro for all surfactants rather than nano), internal environments (substantially dictated by organic substrate, variations pinpointed to specific reagents), and partitioning (dependent on each surfactant) in these systems, suitable for improving the development of sustainable aqueous–organic reactions.^{8,10–15,53,60–62}