
==== Front
Proc Natl Acad Sci U S A
Proc Natl Acad Sci U S A
PNAS
Proceedings of the National Academy of Sciences of the United States of America
0027-8424
1091-6490
National Academy of Sciences

38466847
202321064
10.1073/pnas.2321064121
datasetDatasetresearch-articleResearch ArticlechemChemistry410
Physical Sciences
Chemistry
Direct electrochemical evidence suggests that aqueous microdroplets spontaneously produce hydrogen peroxide
Krushinski Lynn E. a https://orcid.org/0000-0003-2896-4605

Dick Jeffrey E. jdick@purdue.edu
a b 1 https://orcid.org/0000-0002-4538-9705

aDepartment of Chemistry, Purdue University, West Lafayette, IN 47907
bElmore Family School of Electrical and Computer Engineering, Purdue University, West Lafayette, IN 47907
1To whom correspondence may be addressed. Email: jdick@purdue.edu.
Edited by Chad Mirkin, Northwestern University, Evanston, IL; received December 7, 2023; accepted January 30, 2024

11 3 2024
19 3 2024
11 9 2024
121 12 e232106412107 12 2023
30 1 2024
Copyright © 2024 the Author(s). Published by PNAS.
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).

Significance

For centuries, scientists have assumed that chemistry is independent of the volume in which it is being studied. For very small volumes, such as microdroplets, one cannot neglect important interfacial effects, which have been implicated in recent observations of microdroplet curious chemistry. One observation is the spontaneous generation of hydrogen peroxide in microdroplets. Here, we provide direct electrochemical evidence confirming the formation of hydrogen peroxide in aqueous microdroplets suspended in a nonaqueous phase. Our results provide a direct measurement platform for curious chemistry in small volumes.

Recent reports have detailed the striking observation that electroactive molecules, such as hydrogen peroxide (H2O2) and radical water species (H2O.+/H2O.−), are spontaneously produced in aqueous microdroplets. Stochastic electrochemistry allows one to study reactions in real-time occurring inside subfemtoliter droplets, one droplet at a time, when a microdroplet irreversibly adsorbs to an ultramicroelectrode surface (radius ~ 5 µm). Here, we use stochastic electrochemistry to probe the formation of hydrogen peroxide (H2O2) in single aqueous microdroplets suspended in 1,2-dichloroethane. The oxidation of H2O2 at alkaline pH (11.5) differs from near-neutral conditions (6.4), allowing us to create a digital, turn-off sensing modality for the presence of H2O2. Further, we show that the stochastic electrochemical signal is highest at the mass transfer limitation of the H2O2 couple and is dampened when the potential nears the formal potential. We validate these results by showing that the addition of a H2O2 selective probe, luminol, decreases the stochastic electrochemical response at alkaline pH (11.5). Our results support the observation that H2O2 is generated in water microdroplets at concentrations of ~100 s of µM.

microdroplet
hydrogen peroxide
electrochemistry
HHS | NIH | National Institute of General Medical Sciences (NIGMS) 100000057 R35GM138133 Jeffrey E Dick
==== Body
pmcPrevious groups have reported on enhanced reactivity in small droplets (r < ~20 μm) and the ability of these droplets to drive thermodynamically unfavorable reactions (1–3). Zare and coworkers have reported on the spontaneous generation of hydrogen peroxide (H2O2) at the water|air interface (4). George and coworkers provide experimental evidence for the spontaneous generation of hydroxyl radicals (OH•) at the air|water interface of aerosolized droplets (5). Cooks and coworkers have implicated the water radical cation and anion pair in driving redox reactions in microdroplets (6). Studies that comment on the spontaneous generation of H2O2 and other reactive species in microdroplets mostly consist of electrospray or ultrasonic atomization for droplet generation with mass spectrometry and indirect fluorometric assays as methods to quantify (7–9). Such studies typically suggest the high electric field strength at the air|water interface as the driving force behind the generation of reactive species in water droplets (10).

Here, we probe whether aqueous microdroplets suspended in 1,2-dichloroethane spontaneously produce H2O2 using stochastic electrochemistry. Stochastic electrochemistry is a highly sensitive measurement technique capable of analyzing species in single subfemtoliter droplets. Using this technique, we provide direct evidence for the spontaneous generation of H2O2 in water microdroplets suspended in oil. The oil|water interface also has an associated electric field, which may play a role in our observations (11). Our results confirm the presence of a stable and reactive species at a concentration of 100 s of µM that has heterogeneous pH reactivity at platinum ultramicroelectrodes at a rather mild applied bias (E0 ~ 0 to 0.4 V vs. Ag/AgCl). We verify the identity of this electroactive species by performing extensive background voltammetry in bulk and on adsorbed droplets, changing the applied bias and observing changes in signal as we approach the E1/2 of H2O2 (–0.05 V vs. Ag/AgCl), tuning the pH, and dampening the electrochemical signal via addition of a H2O2 selective probe (luminol).

Results and Discussion

Ultramicroelectrodes have previously been used for the study of microdroplets and other micro- and nanoscale entities (i.e., nanoparticles, enzymes, vesicles) because of their high sensitivity and ability to observe steady-state (time-independent) electrochemical responses (12). We used such electrodes to probe the electrochemical behavior of H2O2 in near-neutral (pH 6.4) and alkaline (pH 11.5) pH at high salt concentrations (1 M KCl). Fig. 1 shows steady-state voltammetry of 10 mM H2O2 in both 10 mM NaOH 1 M KCl (pH 11.5) (red curve) and 1 M KCl (pH 6.4) (black curve). Voltammetry is plotted in polarographic convention (negative potentials are plotted to the right, and anodic current is plotted negative).

Fig. 1. Cyclic voltammetry of 10 mM H2O2 in 1 M KCl (black) (pH 6.4) and in 10 mM NaOH and 1 M KCl (red) (pH 11.5). Voltammetry was collected after 10 min of bubbling with N2 with a Pt UME working electrode (r = 5 μm) from 0.2 to –0.25 V vs. Ag/AgCl at 50 mV/s. A glassy carbon rod was used as a counter electrode. Polarographic convention used.

From the voltammograms in Fig. 1, H2O2 oxidation does not occur readily in our system at near-neutral pH, as minimal anodic current is observed at potentials <0.2 V vs. Ag/AgCl. However, at alkaline pH, H2O2 oxidation is observed at potentials greater than –0.05 V vs. Ag/AgCl. The difference in oxidation potential due to pH allows us to create an assay for H2O2 formation, where anodic current should be observed at positive potentials in alkaline pH, but not neutral. Oxidation of H2O2 occurs even in the presence of oxygen (O2) at basic pH (SI Appendix, Fig. S1) even though O2 reduction occurs at similar potentials to H2O2 reduction in the tested solutions (SI Appendix, Fig. S2) (~ –0.15 V vs. Ag/AgCl for O2 and H2O2). To avoid overlapping oxygen reduction and take advantage of the heterogenous pH reactivity, we chose to solely probe the oxidation of H2O2 by analyzing potentials ≥0 V vs. Ag/AgCl. It is also apparent that in all background voltammetry in both aqueous solutions and the continuous oil phase, no significant anodic current is observed at potentials >0.4 V vs. Ag/AgCl without the addition of H2O2 (SI Appendix, Fig. S2). From voltammetry in solutions of varying pH (SI Appendix, Fig. S3), cathodic current from H2O2 reduction is observed at all pH values tested (6.4 to 12), but anodic current is only observed at pH greater than 10 at potentials greater than –0.05 V vs. Ag/AgCl.

The mechanism for H2O2 electro-oxidation is complicated and has been previously studied by Hart and coworkers (13–15). This highly surface-dependent process has been studied most extensively on platinum electrodes. What is important in our study is that there exists a range of potentials where H2O2 can be detected using stochastic electrochemistry without competing reactions (e.g., water oxidation, oxygen reduction, hydroxide oxidation). According to Hart’s literature, the reason for this likely has to do with our unbuffered solutions, high concentration of chloride at neutral pH, and the inability to oxidize the platinum electrode surface to create a favorable binding site for H2O2 oxidation (13–15). From the voltammetry in Fig. 1, it is obvious that the heterogeneous oxidation of H2O2 is more favorable under alkaline conditions. This observation allows us to create a type of assay to probe for H2O2: Anodic signal should be present under basic conditions and absent under more neutral/acidic conditions. In our experiments, we chose to probe for peroxide under pH 11.5 and 6.4 conditions.

Stochastic electrochemistry is a subset of electrochemistry that allows access to reactivity of single atoms, molecules, and nanoparticles (16–18). The electrochemical exploration of spontaneously generated electroactive species like H2O2 is difficult to achieve with more conventionally static electrochemical techniques (e.g., cyclic voltammetry) because such reactive molecules may catalytically decompose on reactive surfaces like platinum. In stochastic electrochemistry, reactivity of single droplets can be rigorously quantified one at a time (16, 17) immediately when a droplet irreversibly adsorbs to a microelectrode surface. This ability to probe dynamic processes with high temporal resolution is a strength of the technique. If one were to probe H2O2 production in microdroplets statically using cyclic voltammetry (or some other method that does not involve the constant monitoring of droplets as they collide with the electrode), the catalytic nature of most working electrode materials will prohibit positive identification. SI Appendix, Fig. S4 shows voltammetry after the adsorption of water droplets to a Pt working electrode. This not only demonstrates that H2O2 cannot be observed through voltammetry in this manner (even at a 10 mM concentration) but also that there is no faradaic current generated by adsorbed 1 M KCl or 10 mM NaOH 1 M KCl droplets at oxidative potentials up to 0.4 V (vs. Ag/AgCl). For these reasons, stochastic electrochemistry is one of the only direct measurement techniques capable of probing H2O2 in droplets.

Fig. 2A shows a schematic representation of the stochastic electrochemistry experiment. When droplets loaded with a redox-active species irreversibly adsorb to the microelectrode, a change in current can be observed in the amperometric i-t response that reports on the rate of consumption of species within the droplet. Fig. 2B shows the collision of single microdroplets with the microelectrode loaded with 10 mM H2O2 in 1 M NaOH. The characteristic transient that results from the collision of a single droplet with the microelectrode is shaped like a blip (rapid rise followed by an exponential decay), and the current versus time character can be explained by bulk electrolysis theory (12, 17). From Fig. 2B, stochastic electrochemistry demonstrates the capability to detect single droplets containing added H2O2.

Fig. 2. (A) Schematic depicting collisions of emulsion droplets on Pt UME and subsequent oxidation of H2O2 with a potential mechanism. (B) Collisions of aqueous microdroplets loaded with 10 mM H2O2 in 1 M NaOH. Amperometry was held at 0.3 V vs. Ag/AgCl for 100 s with a sample rate of 0.0167 s.

A natural question follows: Can stochastic electrochemistry observe H2O2 that is spontaneously produced in aqueous microdroplets at rather low concentrations? Previously, groups have used secondary methods (i.e., fluorescence) to quantify H2O2 (4, 7–9). The amount of H2O2 that is spontaneously produced in such droplets is generally on the order of several 10 s to 100 s of μM (4, 8). One can use bulk electrolysis theory to predict the lowest concentration of hydrogen peroxide measurable with stochastic electrochemistry (10). For these experiments, a detectable signal should be more than thrice the noise (or a signal-to-noise ratio greater than 3). The background noise within our experimental setup is ~200 fA, which can be seen in the background chronoamperometry in SI Appendix, Fig. S5. The background noise was minimized by running the experiment within a well-grounded Faraday cage on a floating table. If the contact radius between the droplet and the electrode is 100 nm, the minimum detectable concentration is ~10 μM, well below previously reported concentrations of spontaneously generated H2O2. Therefore, a positive identification in stochastic electrochemical experiments favors larger droplets (15).

Because the oxidation of H2O2 occurs at relatively mild potentials in basic pH and not in neutral pH, as seen in Fig. 1, we propose that at a mild applied potential (0.4 V vs. Ag/AgCl), oxidative current spikes could be seen upon the collision of unloaded pH 11.5 droplets but not pH 6.4 droplets. In fact, the amperograms taken in an emulsion solution containing pH 6.4 droplets, as seen in Fig. 3, are generally void of anodic current transients. When the same experiment is performed with basic droplets (pH 11.5), oxidative current transients become apparent, as shown in Fig. 3.

Fig. 3. Chronoamperograms of 1 M KCl (black) and 10 mM NaOH 1 M KCl (red) droplets. Chronoamperometry was collected with a Pt UME working electrode (r = 5 μm) at 0.4 V vs. Ag/AgCl for 100 s after a 3-s quiet time at 0 V vs. Ag/AgCl at a sample interval of 0.0167 s. A glassy carbon rod was used as a counter electrode. Polarographic convention used. An example of an integrated collision event from 10 mM NaOH 1 M KCl droplets (red) with a representative baseline is in-laid.

Replicates of these experiments show the described trend remains consistent between measurements (SI Appendix, Fig. S6). Dynamic light scattering (DLS) analysis of 1 M KCl and 10 mM NaOH 1 M KCl containing emulsions can be seen in SI Appendix, Figs. S7 and S8. The DLS analysis shows that the droplets within our system are ~2.5 μm in diameter for both 1 M KCl and 10 mM NaOH 1 M KCl, which is smaller than the proposed critical diameter of 20 μm for H2O2 generation (4). If we integrate the current versus time blips from the anodic transients (inlaid in Fig. 3), we can determine the charge passed during the collision events (Q). Using Faraday’s Law and the average spherical droplet radius from DLS, we can arrive at an expression to calculate concentration: (18).[1] Credox=3Q4πnFr3,

where r is the droplet radius [1.25 μm from DLS (SI Appendix, Fig. S8)], Q is the integrated charge (C, coulombs), n is the number of electrons transferred per molecule (2 for H2O2 oxidation (Fig. 2A)], F is Faraday’s constant (96,485 C mol−1), and Credox is the concentration of H2O2. From the average integrated charge (~0.5 pC), we can calculate the average H2O2 concentration to be ~300 μM. The average calculated concentrations from each replicate experiment shown in SI Appendix, Fig. S6 are 150, 190, 250, 620, and 690 µM, respectively. While this is on the larger end of the proposed H2O2 concentration, we are likely overestimating this average. As mentioned above, we are sampling the larger distribution of droplets with this method, meaning that we will be overestimating the concentration when we assume the sample average droplet size retrieved from Dynamic Light Scattering. We also have a polydisperse distribution of droplet sizes, limiting our ability to robustly correlate H2O2 concentration directly to a specific droplet size (or perhaps the amount of time the droplet has existed in the oil phase). Even so, we are still able to estimate the concentration of H2O2 within the droplets from our experiments, and the values align well with previous reports.

So far, we assumed that H2O2 is the electroactive species for the sake of the calculation, but from the results thus far, we have not confirmed that H2O2 is the electroactive species being probed. However, we can rule out several potential redox species. Recently, Cooks and coworkers have discussed the reactivity of the water dimer radical cation (H2O-OH2)+ and its participation in redox chemistry in microdroplets (6, 19). Zhang and coworkers suggest that the reported water dimer radical cation is actually a hydroxyl radical combined with a hydronium cation through hydrogen bonding (OH•-H3O+) (20). In any case, reactive water species are a possible source of reactivity in microdroplets (21). The stochastic electrochemical method and its high sensitivity allows us to validate the presence of electroactive species that are generated in water microdroplets suspended in oil. The species heterogeneously reacts on a platinum ultramicroelectrode depending on solution pH and is stable long enough to make the electrochemical measurement (with transients observed at 10 s of seconds with lifetimes of several milliseconds) at a rather mild applied potential (Eapplied ~ +0.4 V vs. Ag/AgCl). Reactive water species or other radical species would not have such a lifetime, and as such, we do not expect such species to be detectable via our method (especially at the chosen sample rate of 16.7 ms) Furthermore, given the previous discussion about the limit of detection in single droplets, electroactive species observed herein are likely at concentrations greater than 10 μM. Therefore, the method we are using is not sufficiently sensitive to detect all potential kinetic products of the interfacial reaction occurring.

The electroactive species could potentially be hydroxide. However, the oxidation of hydroxide will not occur at the potentials applied in our experiments, as seen by control experiments (in SI Appendix, Figs. S2 A and S4 A). In the first of these experiments, voltammetry in bulk 10 mM NaOH 1 M KCl reveals no significant anodic current (>0.5 nA) at even the highest applied potential (0.4 V vs. Ag/AgCl). Ciampi and coworkers have postulated that the oxidation of hydroxide anions can be shifted as much as 0.7 V at the gas|liquid interface (22). To ensure that this is not the case in our system, droplets were adsorbed to the electrode surface (by allowing the electrode to sit in the emulsion solution for ~30 s) before performing cyclic voltammetry (SI Appendix, Fig. S4A). From this experiment, no potential shift is observed from the microdroplets, where the observed anodic current is similarly low to bulk measurements (<0.5 nA) at all relevant potentials. Hart’s literature cited above demonstrates the stability of H2O2 oxidation at neutral pH under buffered conditions. We performed similar experiments in phosphate-buffered saline (PBS) (pH 6.4), where H2O2 oxidation is observed at 0.4 V vs. Ag/AgCl on a Pt UME (SI Appendix, Fig. S9A). Amperometry of PBS droplets both loaded (SI Appendix, Fig. S9C) and unloaded (SI Appendix, Fig. S9B) droplets reveals a similar trend, as observed before, providing evidence that further rules out hydroxide oxidation and supports our conclusion of H2O2 oxidation.

All other background voltammetry (SI Appendix, Figs. S2 and S4) eliminates the added salts and solvents as the source of anodic current. Other contaminants, such as trace metals that may be present in the used salts, are also not expected to be the cause of anodic transients as such contaminants would likely precipitate out at alkaline pH. The local formation of Pt oxides upon contact of the electrode with a highly alkaline droplet could also cause anodic signal in these experiments. However, one could calculate the theoretical Pt oxide film thickness (example calculation can be found in Supporting Information) from such an event from the average integrated charge from the transients (0.5 pC), which in this case would give an unrealistic oxide layer thickness.

Previously, Tschulik and Compton reported capacitive collisions with stochastic electrochemistry (23). If capacitance played a role in the response, we would expect to see transients at neutral pH, but transients are not observed. To further investigate the species responsible for the transients seen in pH 11.5, as seen in Fig. 3, the collision experiments were performed at various potentials, as shown in Fig. 4. Additionally, one would expect a time constant for double-layer relaxation on the order of 10 s of µs or lower for microdroplet collisions, much smaller than our sampling rate (16.7 ms).

Fig. 4. (A) Stacked chronoamperograms of unloaded 10 mM NaOH 1 M KCl (pH 11.5) droplets at various applied potentials. Chronoamperometry was collected with a Pt UME working electrode (r = 5 μm) at 0.4 (black traces), 0.2 (red traces), or 0 (purple traces) V vs. Ag/AgCl for 100 s after a 3-s quiet time at a potential 0.05 V vs. Ag/AgCl more negative than the applied potential at a sample interval of 0.0167 s. A glassy carbon rod was used as a counter electrode. Chronoamperograms are plotted in polarographic convention. (B) Observed collision frequency as a function of applied potential for loaded (black) and unloaded (red) pH 11.5 droplets.

To further investigate the species responsible for the transients seen in pH 11.5 microdroplets (Fig. 3), collision experiments were performed at various potentials, as seen in Fig. 4. When the potential approaches the formal potential of H2O2, the current approaches 0. Therefore, the oxidation of H2O2 in aqueous microdroplets should be dampened as the potential approaches the formal potential of the redox couple. In SI Appendix, Fig. S9A, collisions of native 1 M KCl droplets are observed via chronoamperometry at 0.4, 0.2, and 0 V vs. Ag/AgCl. Low frequency and magnitude collision events at 0.2 and 0.4 V vs. Ag/AgCl can be seen in these measurements. At 0 V vs. Ag/AgCl, cathodic-type transients begin to be distinguishable in unloaded droplets. These responses are very similar to those of pH 6.4 droplets intentionally loaded with 10 mM H2O2 (SI Appendix, Fig. S9A). In the case of pH 11.5 loaded and unloaded droplets, anodic transients can be seen at both 0.2 and 0.4 V vs. Ag/AgCl (Fig. 4A). It should be noted that H2O2 reduction is complicated by overlapping oxygen reduction on platinum, as shown in SI Appendix, Fig. S1, and makes it a poor choice for observation by stochastic electrochemistry. With this, we focused on potentials more positive of the formal potential of oxygen reduction in our system (0 V vs. Ag/AgCl). We chose 0 V vs. Ag/AgCl as our most mild applied potential. Again, the transients are similarly behaved to those resulting from experiments performed at pH 11.5 with 10 mM H2O2 added (SI Appendix, Fig. S9B). The experimental frequency of collision events at these various applied potentials can be seen in Fig. 4B. From this analysis, both the loaded and unloaded droplets demonstrate a decrease in frequency as the potential nears the E1/2 of H2O2 (~ −0.15 V). Because the electrochemical signal for both the loaded and unloaded droplets shows a dependence on potentials near to and far from the formal potential of H2O2 and there is no dependence of the calculated concentration on the frequency of collision and applied potential, we suspect that the species that is most likely responsible for the observed transients in the negative control is H2O2. Additionally, the calculated average concentration from each applied potential was ~320, 230, and 590 µM for 0.4, 0.2, and 0 V vs. Ag/AgCl, respectively, giving an overall average of ~400 µM. If we perform the same calculation for droplets purposefully loaded with 10 mM H2O2 (SI Appendix, Fig. S10B), we obtain a value of ~20 mM, sufficiently similar considering the polydispersity of droplet size.

We further confirmed the presence of H2O2 by adding a reagent, luminol, that will consume H2O2. SI Appendix, Fig. S10 shows the addition of luminol decreases the steady-state limiting current of H2O2. Stochastic electrochemical experiments were then performed with droplets containing 1 mM luminol in 10 mM NaOH 1 M KCl. As seen in SI Appendix, Fig. S11, collision transients are diminished in the presence of luminol. The observed collision frequencies in the experiments with added luminol are nearly half that of native pH 11.5 droplets (where the average observed collision frequencies are 0.04 ± 0.02 Hz and 0.07 ± 0.04 Hz, respectively), implicating H2O2 as the species responsible for the observed transients. Luminol solubility in the nonaqueous phase prohibits robust correlated experiments (despite our best efforts) and limits the amount of H2O2 consumption within the aqueous microdroplets. Even so, we demonstrate with this experiment that the addition of a molecule that consumes H2O2 decreases the stochastic electrochemical signal.

Conclusion

Over the past 10 y, many groups have reported on the curious chemistry of small volumes using a variety of powerful measurement tools (24–26). In this article, we used stochastic electrochemistry to probe reactivity of water microdroplets suspended in a 1,2-dichloroethane continuous phase. When water microdroplets irreversibly adsorb to an ultramicroelectrode, heterogeneous reactivity can be probed on a droplet-by-droplet basis. For basic (pH ~ 11.5) water microdroplets, we observed anodic transients in the amperometric i-t curve in the negative control. Upon further investigation, we found that the heterogeneous reaction that is occurring in the droplets electrochemically behaves similarly to H2O2. This observation is accentuated by a careful study of the amperometric response as a function of potential, which showed diminished transients as the applied potential approached that of the formal potential of H2O2. This would be expected as the transients transition from potentials where mass transfer controls the rate of electron transfer to potentials where electron transfer kinetics dominate. Our studies indicate that the electroactive species generated in the aqueous microdroplets are active at mild potentials in basic pH and tend to be diminished as the formal potential of H2O2 is approached, implicating H2O2 as the species being generated perhaps spontaneously in aqueous microdroplets. We further confirm H2O2 as the electroactive species by adding a species which is known to consume peroxide (luminol) and observing a decrease in the collision frequency, where the observed frequency with luminol is nearly half that of the frequency without. We also calculate the H2O2 concentration within the droplets to be ~400 μM on average, assuming a mean droplet diameter measured by DLS. We do not attempt to determine the mechanism by which H2O2 is generated, so the effect of species like O2 and O3, which have been proposed previously to play a role in H2O2 formation (7, 8), was not investigated. Our results add to the growing discussion regarding the curious chemistry of water microdroplets, yielding potential insights into key environmental and biological processes.

Materials and Methods

Reagents and Materials.

Hydrogen peroxide (H2O2, certified ACS 30%), potassium chloride (KCl), nitric acid (HNO3, certified ACS), phosphate buffered saline (1X solution), and sodium hydroxide (NaOH, Certified ACS) were purchased from Thermo Fisher Scientific. Tetrabutylammonium perchlorate [(TBA)(ClO4), for electrochemical analysis, >99.0%], Luminol (≥97%, HPLC), and 1,2-dichloroethane (DCE, 98%) were purchased from Sigma-Aldrich. All chemicals were used as received. Aqueous solutions were prepared with ultrapure water (Millipore Milli-Q, 18.2 MΩ cm). Electrochemical experiments were performed using a CHI model 6284E potentiostat (CH Instruments, Austin, Texas) with a three-electrode cell placed in a Faraday cage. A glassy carbon rod (7 mm diameter) (Alfa Aesar, Stoughton, MA) was used as the counter electrode, and an Ag/AgCl reference electrode (CH Instruments, Austin, Texas) connected by a salt bridge (1 M KCl suspended in agarose) was used for every measurement. A platinum UME (r = 5 µm) (CH Instruments, Austin, TX) working electrode was used. The working electrode was rinsed thoroughly with ultrapure water, dipped into a vigorously stirring solution of 1 M HNO3 for 30 s, and then polished on a wet microcloth pad (Buehler, Lake Bluff, IL) for 1 min before each measurement. DLS measurements were performed using a Zetasizer Ultra (Malvern, Westborough, MA) within 30 s of emulsion sonication.

Cyclic Voltammetry of H2O2.

A 10 mM H2O2 solution was made in 10 mM NaOH 1 M KCl and in 1 M KCl. Voltammetry was collected with a Pt UME working electrode (r = 5 μm) from 0.2 to −0.25 V vs. Ag/AgCl at 50 mV/s within a minute of sample preparation. A glassy carbon rod was used as a counter electrode.

Collisions of Aqueous Microdroplets.

An aliquot of 20 µL of aqueous solution was added to 2 mL of 0.1 M [TBA][ClO4] in DCE in a 4-mL glass vial (Thermo Fisher Scientific). The two-phase solution was then ultrasonicated (500 W, 40% amplitude) with a pulsing method (5 s on 5 s off, 10 cycles). For collisions of droplets loaded with 10 mM H2O2 in 1 M NaOH, an emulsion was prepared as described above. Amperometry was held at 0.3 V vs. Ag/AgCl for 100 s with a sample rate of 0.0167 s within 30 s of emulsion preparation. For nonloaded droplet collisions, either 10 mM NaOH 1 M KCl or 1 M KCl was used as the aqueous solution, and an emulsion was prepared as described above. For PBS and luminol experiments, the solutions were prepared in the same manner as above with PBS or with 1 mM luminol in NaOH as the aqueous phase respectively. Chronoamperometry was run for 100 s at 0.4 V vs. Ag/AgCl with a sample rate of 0.0167 s after a 3-s quiet time held at 0 V vs. Ag/AgCl. All estimations of droplet concentration assume a droplet diameter of 2.5 µm, which was independently measured using dynamic light scattering (SI Appendix, Fig. S8).

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

This work was completed with financial support from the National Institute of General Medical Sciences under Grant R35138133. We thank Prof. Christophe Renault (Loyola University Chicago) and Drs. Kathryn Vannoy and Silvia Voci for helpful discussions.

Author contributions

L.E.K. and J.E.D. designed research; L.E.K. performed research; L.E.K. and J.E.D. analyzed data; and L.E.K. and J.E.D. wrote the paper.

Competing interests

The authors declare no competing interest.

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.

Supporting Information

This article is a PNAS Direct Submission.
==== Refs
1 L. Qui, M. D. Psimos, R. G. Cooks, Spontaneous oxidation of aromatic sulfones to sulfonic acids in microdroplets. J. Am. Soc. Mass Spectrom. 33 , 1362–1367 (2022).35312307
2 D. T. Holden, M. N. Morato, R. G. Cooks, Aqueous microdroplets enable abiotic synthesis and chain extension of unique peptide isomers from free ammino acids. Proc. Natl. Acad. Sci. U.S.A. 119 , e2212642119 (2022).36191178
3 W. Wang , Water microdroplets allow spontaneously abiotic production of peptides. J. Phys. Chem. Lett. 12 , 5774–5780 (2021).34134488
4 J. K. Lee, K. L. Walker, H. S. Han, R. N. Zare, Spontaneous generation of hydrogen peroxide from aqueous microdroplets. Proc. Natl. Acad. Sci. 116 , 19294–19298 (2019).31451646
5 K. Li , Spontaneous dark formation of OH radicals at the interface of aqueous atmospheric droplets. Proc. Natl. Acad. Sci. U.S.A. 120 , e2220228120 (2023).37011187
6 L. Qui, R. G. Cooks, Simultaneous and spontaneous oxidation and reduction in microdroplets by the radical cation/anion pair. Angew. Chem. Int. Ed. Engl. 134 , e202210765 (2022).
7 A. Gallo , On the formation of hydrogen peroxide in water microdroplets. Chem. Sci. 13 , 2574 (2022).35340850
8 M. A. Mehrgardi, M. Mofidfar, R. N. Zare, Sprayed water microdroplets are able to generate hydrogen peroxide spontaneously. J. Am. Chem. Soc. 144 , 7606–7609 (2022).35451822
9 F. Rezende, R. P. Brandes, K. Schröder, Detection of hydrogen peroxide with fluorescent dyes. Antioxid. Redox Signal. 29 , 585–602 (2018).29054131
10 S. M. Kathmann, I.-F.W. Kuo, C. J. Mundy, Electronic effects on the surface potential at the vapor−liquid interface of water. J. Am. Chem. Soc. 130 , 16556–16561 (2008).19554692
11 H. Xiong, J. K. Lee, R. N. Zare, W. Min, Strong electric field observed at the interface of aqueous microdroplets. J. Phys. Chem. Lett. 11 , 7423–7428 (2020).32804510
12 L. R. Faulkner, A. J. Bard, Electrochemical Methods: Fundamentals and Applications (John Wiley & Sons, New York, NY, ed. 2, 2001).
13 S. B. Hall, E. A. Khudaish, A. L. Hart, Electrochemical oxidation of hydrogen peroxide at platinum electrodes. Part IV. Phosphate buffer dependence. Electrochim. Acta 44 , 4573–4582 (1999).
14 S. B. Hall, E. A. Khudaish, A. L. Hart, Electrochemical oxidation of hydrogen peroxide at platinum electrodes. Part II: Effect of potential. Electrochim. Acta 43 , 2015–2024 (1998).
15 S. B. Hall, E. A. Khudaish, A. L. Hart, Electrochemical oxidation of hydrogen peroxide at platinum electrodes. Part I: An adsorption-controlled mechanism. Electrochim. Acta 43 , 579–588 (1998).
16 K. J. Vannoy, I. Lee, K. Sode, J. E. Dick, Electrochemical quantification of accelerated FADGH rates in aqueous nanodroplets. Proc. Natl. Acad. Sci. U.S.A. 118 , e2025726118 (2021).34161273
17 A. J. Bard, High speed controlled potential coulometry. Anal. Chem. 35 , 1125–1128 (1963).
18 B.-K. Kim, A. Boika, J. Kim, J. E. Dick, A. J. Bard, Characterizing emulsions by observation of single droplet collisions—Attoliter electrochemical reactors. J. Am. Chem. Soc. 136 , 4849–4852 (2014).24641496
19 M. Wang , Abundant production of reactive water radical cations under ambient conditions. CCS Chem. 4 , 1224–1231 (2022).
20 D. Xing , Capture of hydroxyl radicals by hydronium cations in water microdroplets. Angew. Chem. Int. Ed. Engl. 61 , e202207587 (2022).35700155
21 L. Qiu, N. M. Morato, K.-H. Huang, R. G. Cooks, Spontaneous water radical cation oxidation at double bonds in microdroplets. Front. Chem. 10 , 903774 (2022).35559217
22 Y. B. Vogel , The corona of a surface bubble promotes electrochemical reactions. Nature Comm. 11 , 6323 (2020).
23 J. Poon, C. Batchelor-McAuley, K. Tschulik, R. G. Compton, Single graphene nanoplatelets: Capacitance, potential of zero charge and diffusion coefficient. Chem. Sci. 6 , 2869–2876 (2015).28706674
24 M. Senske, A. E. Smith, G. J. Pielak, Protein stability in reverse micelles. Angew. Chem. Int. Ed. Engl. 55 , 3586–3589 (2016).26854977
25 A. Fallah-Araghi , Enhanced chemical synthesis at soft interfaces: A universal reaction-adsorption mechanism in microcompartments. Phys. Rev. Lett. 112 , 028301 (2014).24484045
26 C. J. C. Jordan, M. P. Coons, J. M. Herbert, J. R. R. Verlet, Spectroscopy and dynamics of the hydrated electron at the water/air interface. Nat. Commun. 15 , 182 (2024).38167300
