Reply to the Correspondence on “Can Charge Transfer Across C–H···O Hydrogen Bonds Stabilize Oil Droplets in Water?”

Abstract

Water‐to‐oil charge transfer (CT) across weak C–H···O hydrogen bonds is localized for hexane and hexadecane, and is too small to explain oil‐droplet stability and electrokinetic experiments. Collective intraphase density rearrangements should not be confused with interfacial CT. Spectroscopic evidence cannot simultaneously imply weak CT for the C‐H blue shift and large CT for the O‐H/D red shift. A weakness of the C–H···O CT mechanism is that it can't explain other oil emulsion chemistry without this specific chemical motif.

1. Introduction

We thank Singh, Rani, Pullanchery, and Roke for their Correspondence [1] regarding our article, “Can Charge Transfer Across C–H···O Hydrogen Bonds Stabilize Oil Droplets in Water?” [2] Our article examined a specific contention of Roke and co‐workers that charge transfer (CT) across C–H···O hydrogen bonds is quantitatively sufficient to account for the interfacial charge density, surface potential, and electrokinetic response associated with stable oil droplets in water. Their Correspondence discusses several issues related to their CT mechanism as the molecular origin of oil charging in water, including (1) hydroxide adsorption versus charge transfer (CT) across C–H···O hydrogen bonds; (2) limitations of cluster calculations; (3) possible role of collective interfacial effects; (4) conductivity‐based interpretations of electrophoretic mobility; (5) surface‐specific spectroscopy, and (6) the suitability of hexane as a model alkane for CT compared to hexadecane [1]. We welcome this discussion because the origin of oil charging remains an important and unresolved problem [3, 4, 5, 6].

As we clarify below, the points raised in the Correspondence do not change the main conclusion of our analysis: the net water‐to‐oil CT across saturated C–H···O hydrogen bonds is localized, and the resulting net CT by area is insufficient to explain oil‐droplet stabilization and electrokinetic experiments. We further emphasize that collective polarization or intraphase density rearrangement should not be confused with net interfacial CT, and that spectroscopic evidence for weak C–H···O interactions does not establish a persistent and quantitatively significant charge separation across the oil–water interface, it only supports Pauli repulsion confinement at the given density. Changing hexane to dodecane or hexadecane makes no qualitative nor quantitative difference in our conclusions. Finally, (7) a particular weakness of the CT through weak C–H···O interactions model is that it can't explain any other oil chemistry without this specific chemical motif.

2. Results and Discussion

(1) Hydroxide Adsorption Versus CT Across C–H···O Hydrogen Bonds.

The Correspondence contrasts hydroxide adsorption with CT through improper C–H···O hydrogen bonds as two explanations for the negative electrophoretic mobility of oil droplets. While we agree that the hydroxide hypothesis remains controversial [1, 7], uncertainty about hydroxide adsorption does not establish CT as the preferred and only alternative mechanism. Any CT‐based explanation must independently demonstrate a persistent and quantitatively sufficient transfer of charge from water to oil. Our calculations showed that forward and backward CT largely cancel, leaving only a negligibly small net CT density that furthermore is in the direction from oil to water [2]. Hence, the resulting net CT is insufficient to explain oil‐droplet stabilization as a viable explanation on its own merit.

(2) Cluster Calculations Capture CT, Which is Short‐Ranged. The Correspondence argues that cluster‐based calculations are intrinsically unable to capture collective effects present at extended interfaces. However, our analysis was not limited to isolated water–alkane dimers [2]. We evaluated large complexes derived from interfacial clusters extracted from molecular dynamics simulations of an extended water–alkane interface under periodic boundary conditions. We note that Amante and co‐workers failed to acknowledge our extended interfacial systems [8]. These extended interfacial hydrophobic‐water clusters thereby represent the cooperative effects of the surrounding hydrogen‐bond network [2].

We showed that including this extended surface hydrogen‐bond environment did not change the net negligible CT. Instead, the CT balance remained primarily controlled by the immediate water–oil contact geometry, with forward and backward CT nearly compensating. To further assess whether longer‐range environmental effects could enhance the net CT, we recalculated the first ten interfacial clusters used in Figure 2e of our original 2025 study with QM/MM point‐charge embedding of the remaining water and oil molecules. The embedded CT distributions closely overlap with the original extended cluster results (Figure 1a), indicating that the surrounding electrostatic environment only weakly perturbs the directional CT components. Thus, the short‐ranged CT response remains governed by the local water–hexane interaction geometry, with forward and backward CT nearly compensating. This behavior is consistent with the short‐ranged nature of intermolecular CT across weak hydrogen bonds [9, 10]. Longer‐range polarization can certainly modulate the electrostatic environment, but it cannot generate a large net transfer of electronic charge across the oil–water boundary, since the definition of polarization involves no net charge flow.

FIGURE 1.

Electronic‐structure analysis and density rearrangement in representative water–oil model systems. (a) Effect of QM/MM point‐charge embedding on interfacial charge‐transfer distributions. 10 representative interfacial clusters were used. Solid lines show the original cluster calculations, and dashed lines show QM/MM embedding calculations for the same 10 representative water–hexane interfacial clusters. Forward CT corresponds to water‐to‐hexane electron transfer, backward CT corresponds to hexane‐to‐water electron transfer, and net CT is the sum of the two directional CT components. The close overlap between the solid and dashed distributions shows that including the surrounding point‐charge environment does not substantially increase the net CT. (b) Dominant COVPs associated with the polarization process for a water–hexane complex, using the same optimized structure as in (c) of our original 2025 study [2] (c) Electron‐density‐difference isosurfaces for a representative water/oil interfacial cluster selected from the water–hexane clusters analyzed in Figure 2 of our original 2025 study [2] The density differences are shown relative to the HF density for the same cluster, using the BLYP and $\omega$B97X‐V functionals. Positive density is shown in red and negative density in blue, with wireframe isosurfaces plotted at $\pm$0.007. The BLYP result shows a pronounced apparent electron‐density depletion on water and accumulation on oil, whereas the $\omega$B97X‐V result is weaker, consistent with the smaller delocalization error expected for $\omega$B97X‐V relative to BLYP. (d) Representative optimized structure of a perfluoromethyldecalin–water complex used to examine CT at fluorinated oil–water contacts. The net CT direction is from perfluoromethyldecalin to water, opposite to the water‐to‐oil CT direction required by the proposed hydrocarbon C–H···O CT mechanism.

(3) Collective Interfacial Enhancement: EDA and Limitations of DFT. While interface formation inevitably induces electron‐density redistribution, it does not demonstrate persistent interphase electron transfer. A proposed CT mechanism must further establish that a net charge is transferred from water to oil with sufficient magnitude to account for oil‐droplet stabilization or electrophoretic mobility [2]. In a previous study [11], small changes in atomic charges across interfacial molecules were interpreted as a collective polarization/CT effect and were argued to accumulate into a sizable interfacial charge profile. However,“collective polarization” is not a defined concept, and in fact the underlying charge‐partitioning changes contain mixed contributions from overlap‐driven density rearrangements with frozen orbitals, dominated in this case by permanent electrostatics, in addition to polarization and CT. A quantitative decomposition of frozen, polarization, and CT terms is essential before assigning oil charging to collective water‐to‐oil electron transfer.

Energy‐decomposition analysis (EDA) provides a route to separate these effects [9, 12, 13, 14] EDA shows that the net water‐to‐oil CT density is very small, whereas the larger intraphase charge shifts are better understood as polarization density rearrangement rather than persistent interphase electron transfer. As shown in Figure 1b, the polarization component produces substantial charge redistribution within each phase. Moreover, because electron density is sensitive to the electronic‐structure method, particular care is required to avoid overestimating polarization and CT through the well‐known delocalization error introduced by density functional theory (DFT) [15] As shown in Figure 1c, by contrast with the better range‐separated hybrid density functional $\omega$B97X‐V [16] used in our work [2], the semi‐local functional BLYP [17, 18] produces a much more pronounced apparent electron‐density depletion on water and accumulation on oil. This behavior is consistent with the greater delocalization error expected for semi‐local functionals, which can artificially amplify both polarization‐like density rearrangement and apparent CT. Quantitative occupied‐virtual orbitals for chemical valence (OVOCV) analysis [19, 20] for the same interfacial cluster, containing a hydrophobic phase of 8 associated hexanes and a water phase of 35 water molecules over an interfacial area of approximately 4 ${\mathrm{nm}}^2$, gives an apparent charge difference of 0.63 ${\mathrm{e}}^{-}$ from the delocalization free HF density to BLYP density, while it decreases to 0.23 ${\mathrm{e}}^{-}$ with $\omega$B97X‐V. Hence, the resulting accumulated interfacial charge profile cannot be attributed to a “collective‐polarization” mechanism or taken as direct evidence for persistent water‐to‐oil CT when using DFT functionals ill‐suited to this analysis, as done in previous studies used to interpret the oil‐water emulsion experiments [11].

(4) Bulk Conductivity Does Not Explain Electrophoretic Mobility. The Correspondence also suggests that the pH dependence of droplet mobility may reflect changes in bulk aqueous conductivity rather than changes in interfacial charge [1, 21]. However, a conductivity‐only interpretation is not self‐consistent as a microscopic explanation for oil charging.

The molar conductivity of ${\mathrm{OH}}^{-}$ is lower than that of ${\mathrm{H}}_3{\mathrm{O}}^{+}$ [22, 23]. If electrophoretic mobility simply scaled with bulk conductivity, acidic solutions should produce larger mobilities than basic solutions. This is not the general experimental trend associated with oil droplet charging, where droplets approach an isoelectric point or become unstable under acidic conditions, and larger negative electrophoretic mobilities occur primarily under basic conditions [24, 25, 26, 27]. The observation that droplets cannot be straightforwardly studied under acidic conditions because they become unstable is thus not naturally explained by a conductivity‐only picture. If surface charge were independent of pH and mobility merely reflected bulk conductivity, then the higher conductivity of acidic solutions should enhance, rather than suppress, stability and electrophoretic response.

The acid‐induced loss of droplet stability instead points to changes in interfacial charge stabilization and the electrokinetic double layer [26, 28, 29]. This is also difficult to reconcile with the claim that C–H···O CT provides the surface charge responsible for droplet stabilization, because the underlying hydrocarbon C–H motifs should remain present under acidic conditions. More generally, electrophoretic mobility reflects the motion of a charged object and its surrounding electrokinetic double layer under an applied field [28, 29, 30]. A local, short‐ranged electronic density rearrangement at the oil–water interface is not equivalent to mobile ionic charge in a diffuse layer, nor does it automatically provide a sustained charge‐transport mechanism coupled to electrophoretic motion.

This is why our paper distinguished between electrostatic models appropriate for mobile ionic charge distributions and those appropriate for localized CT. A Poisson–Boltzmann description is appropriate for mobile ionic charge in an electrochemical double layer, whereas localized CT advocated by Roke and co‐workers is more naturally treated as a fixed interfacial charge distribution in Poisson electrostatics [2, 28, 31, 32, 33]. Still, under either theoretical model, the CT densities obtained from our analysis are too small to produce the surface potentials required for droplet metastability.

This distinction also suggests a possible experimental test of the conductivity‐based interpretation. Nonionic surfactant‐stabilized droplets, such as Triton X‐100‐stabilized oil droplets, can remain stable under acidic conditions where bare droplets may destabilize, while the surfactant layer deliberately modifies the interfacial hydrophilicity, slipping plane, and local ion distribution. If electrophoretic mobility were controlled mainly by bulk conductivity, the mobility should continue to follow the conductivity order of the aqueous phase even after surfactant stabilization. In contrast, if mobility is governed by interfacial charge and the electrokinetic double layer, surfactant adsorption should substantially modify or suppress the $\zeta$‐potential and its pH dependence. Reported reductions in the $\zeta$‐potential of Triton X‐100‐modified hexadecane droplets are consistent with this latter expectation [34].

(5) Interpretation of C–H and O–D Frequency Shifts are Inconsistent. The spectroscopic argument based on C–H and O–D frequency shifts also does not establish a large interfacial CT mechanism. In the original sum‐frequency scattering interpretation, the broad distribution of high‐frequency O–D modes at the oil droplet interface was assigned to non‐hydrogen‐bonded O–D groups (i.e., dangling O–D groups) interacting with the oil phase, while the accompanying blue shift of interfacial C–H modes was taken as evidence for water‐to‐oil CT [35]. However, these two spectral observations place different demands on the strength of the proposed C–H···O interaction.

The Correspondence notes that the C–H blue shift arises because Pauli repulsion dominates over the relatively weak CT contribution in C–H···O hydrogen bonds [1]. This point is consistent with our work and that of many others [2, 10, 36, 37]. However, it also means that the C–H blue shift indicates that CT is too small to overcome the blue‐shifting contribution from the dominant Pauli repulsion, and therefore cannot be used as evidence for a large net charge transfer [10, 36, 37].

This also creates an inconsistency in assigning the broad O–D band that is strongly red shifted to the same weak C–H···O CT interaction. If the C–H···O contact is weak enough that Pauli repulsion dominates the C–H response, it is not obvious why the same interaction should also strongly perturb the dangling O–D stretching mode and produce a large red shift. Such a large red shift should reflect a large perturbation such as strong local electric fields [34], interfacial ionic species [38], charged impurities [6, 39], and hence electrostatic in origin [6, 34, 39]. Weak saturated C–H···O hydrogen bonds alone have not been shown to produce an O–D red shift of the required magnitude. Hence, an O–D band that is strongly red shifted cannot be uniquely and solely assigned to C–H···O CT, which in itself is so weak that a C–H blue shift is observed via the same C–H···O contact.

Thus, spectroscopy supports the presence of interfacial water–oil interactions, including weak C–H···O hydrogen bonds that are blue‐shifted. In fact the spectroscopy establishes that these hydrogen bonds do not generate a macroscopic, persistent, and quantitatively sufficient net water‐to‐oil CT.

(6) Hexane as a Model for Saturated C–H···O Hydrogen Bonds. The Correspondence questions our use of hexane as a model system, emphasizing that hexane and hexadecane differ in solubility, wetting behavior, interfacial structure, and macroscopic droplet stability [1]. Furthermore, experiments must also demonstrate that saturated C–H···O hydrogen‐bonds in longer alkanes generate a much larger intrinsic CT interaction or a substantially larger net CT density.

However, the question addressed in our article was whether saturated hydrocarbon C–H groups can accept enough charge through C–H···O hydrogen bonds to generate the surface charge required for droplet stabilization. For this purpose, hexane is a chemically appropriate minimal model because it contains the same saturated C–H motifs central to the proposed CT mechanism [2]. However, to further test this point computationally, we extended our analysis to water–dodecane and water–hexadecane interfacial models using the same analysis protocol. The surface densities of interfacial C–H···O hydrogen bonds are very similar across the alkane series, increasing only slightly from 12.1 H‐bonds ${\mathrm{nm}}^{-2}$ for hexane to 12.5 and 12.6 H‐bonds ${\mathrm{nm}}^{-2}$ for dodecane and hexadecane, respectively. These results indicate that increasing alkane chain length does not substantially increase the number of C–H···O hydrogen‐bond geometries available for CT per unit interfacial area. The corresponding net CT densities are 0.0007 ${\mathrm{e}}^{-}$ ${\mathrm{nm}}^{-2}$ for hexane, 0.0005 ${\mathrm{e}}^{-}$ ${\mathrm{nm}}^{-2}$ for dodecane, and 0.0003 ${\mathrm{e}}^{-}$ ${\mathrm{nm}}^{-2}$ for hexadecane, all of which are similar to each other and are far below the charge density required to explain oil‐droplet stabilization or electrophoretic mobility [2, 26]. Thus, while longer alkanes provide a more realistic representation of experimentally stable oil droplets, the central conclusion is unchanged: increasing alkane chain length does not convert weak, nearly compensating C–H···O CT into a sufficiently large net interfacial charge.

There is also a quantitative point concerning the solubility comparison of hexane versus hexadecane in the Correspondence. The Correspondence states that n‐hexane has a water solubility of 110 mM, whereas commonly tabulated values at room temperature are closer to 110 $\mu \mathrm{M}$. This correction to the Correspondence does not eliminate that there is a distinction between hexane and hexadecane in regards stability, but it reduces the magnitude of the contrast by three orders of magnitude, and weakens the argument that the use of hexane alone invalidates a local CT analysis.

(7) C–H···O CT is not a Universal Explanation for Oil Charging. The proposed CT mechanism relies specifically on weak or improper C–H···O hydrogen bonds, in which small amounts of CT from water to oil are proposed to accumulate collectively over the droplet surface [1, 11, 35, 40]. This mechanism faces a basic limitation: oils lacking C–H hydrogen‐donor motifs, including perfluorinated oils, have also been reported to exhibit negative zeta potentials and pH‐dependent electrokinetic behavior comparable to hydrocarbon oils. Such observations are difficult to reconcile with a general charging mechanism that depends specifically on C–H···O hydrogen bonds [25, 41].

We have further examined representative perfluorinated oil–water interactions using energy‐decomposition analysis. In contrast to the proposed water‐to‐oil CT across C–H···O hydrogen bonds, the relevant weak interaction motif for perfluorinated oils is better described as an F···H–O contact, for which the CT direction is from the fluorinated oil molecule toward water rather than from water toward oil Figure 1d. Thus, even when a weak donor–acceptor interaction exists at a perfluorinated oil–water interface, its CT direction is opposite to that required to produce a negatively charged oil phase through water‐to‐oil electron transfer. This provides a direct molecular counterexample to the idea that weak interfacial CT generically produces negative oil charging. Discussed points (1)–(6) do not change with this new chemical motif.

3. Conclusions

We agree with the Correspondence that the molecular origin of oil charging in water remains unresolved and that hydroxide adsorption is not established by direct interfacial spectroscopies such as VSFS under many experimental conditions [1]. However, the absence of definitive evidence for hydroxide adsorption does not automatically default to CT as the origin of oil charging. The proposed CT mechanism must independently satisfy a quantitative criterion: it must generate enough persistent water‐to‐oil charge separation to account for droplet stabilization and electrophoretic mobility.

Our analysis shows that it does not. Saturated C–H···O hydrogen bonds exhibit weak CT, but forward and backward contributions nearly compensate, leaving a small net CT density [2]. Extended hydrogen‐bond environments around extended hydrophobic surfaces and longer alkane chains do not convert this weak local interaction into a large macroscopic charge. Intraphase collective polarization and density rearrangement are real interfacial effects, but they are not equivalent to net interphase CT. The conductivity‐based interpretation of mobility does not explain the contrasting behavior of acidic and basic solutions. C–H blue shifts are consistent with Pauli‐dominated weak hydrogen bonds, not with large water‐to‐oil charge transfer [10, 36, 37]. Finally, the net CT associated with C–H···O hydrogen bonds is too small and too chemically specific to serve as a general mechanism for oil‐droplet charging.

We therefore maintain the central conclusion of our article [2]: CT across C–H···O hydrogen bonds is quantitatively too small to stabilize oil droplets in water or to explain their stability and electrophoretic mobility.

Author Contributions

R.Z. and T.H.‐G. conceived the scientific direction of the reply. R.Z. prepared the draft, and all authors provided comments on the analysis and manuscript.

Conflicts of Interest

The authors have no conflicts of interest.

Data Availability Statement

All data will be made available upon reasonable request from the authors.