Imaging Reveals Delayed HDO Formation at Water Interface

Proton-deuteron exchange in water is often treated as effectively immediate once H₂O and D₂O meet, because the elementary proton-transfer steps occur on picosecond timescales. This paper tests that assumption at a well-defined liquid-liquid interface rather than in a premixed sample. The authors use two impinging cylindrical microjets to form a fast-flowing liquid sheet with an H₂O/D₂O interface, then image wavelength-tunable infrared transmission across the sheet as the interface moves downstream in time.

The result is a direct concentration movie of H₂O, D₂O, and HDO during the first 100 microseconds of contact. The central finding is that the early interface can be diffusively mixed while still chemically out of equilibrium: H₂O and D₂O have overlapped, but the HDO concentration remains below the local equilibrium value.

Core Contribution

The contribution is not a new exchange mechanism. It is a measurement that separates molecular mixing from isotopic equilibration on microsecond length and time scales. The experiment reads out chemically specific vibrational bands: the H-O-H bend near 1640 cm⁻¹, the D-O-D bend near 1210 cm⁻¹, and the H-O-D bend near 1450 cm⁻¹. By fitting each local spectrum as a linear combination of reference spectra, the authors convert infrared images into optical thicknesses for the reactants and product.

Figure 1 shows the experimental idea and why the spectroscopy is sufficient for quantification: the three bending bands are separated enough to follow HDO formation without relying on an indirect proxy.

Technical Approach

The time axis comes from flow, not from repeated pump-probe delay scans. As the liquid sheet travels downstream, positions along the sheet correspond to later reaction times. The paper bins regions at the interface and extracts spectra with 10 µs time resolution. That design matters because it lets the authors compare pure H₂O/D₂O exchange with acid- or base-doped cases under the same imaging geometry.

Figure 2 is the key measurement. In pure H₂O/D₂O, the HDO band grows gradually as the interface evolves. In a 10⁻² M HCl/D₂O liquid sheet, the HDO signal rises much faster. The interpretation is chemical rather than geometric: increasing the available hydronium or hydroxide concentration accelerates the proton-deuteron exchange until it approaches the diffusion-limited case.

The analysis then couples the concentration maps to reaction-diffusion simulations. In the fast model, the local HDO concentration is assumed to reach equilibrium wherever H₂O and D₂O mix. In the slow model, the HDO formation rate is finite and constrained by the available ions that mediate exchange. The fitted diffusion coefficient is taken from the acid/base case, where the exchange is too fast for the apparatus to resolve, and the pure-water case is used to fit the rate constant.

Results and Analysis

The strongest evidence is the contrast between the pure and doped interfaces. The pure H₂O/D₂O system shows a delayed HDO buildup over the first 100 µs, while 10⁻² M acid or base drives the system close to the fast-equilibration limit. That comparison supports the authors' claim that the early pure-water interface is reaction-limited despite the fast elementary proton-transfer step.

Figure 3 makes the physical picture explicit. The fast model reaches local equilibrium concentrations, with 27.5 M corresponding to the 1:1:2 H₂O:D₂O:HDO composition. The slow model stays below that equilibrium line at early times, even though diffusion has already brought the isotopic liquids into contact. This is the paper's most useful distinction: diffusion can erase the spatial separation of the two liquids before the isotope-exchange chemistry catches up.

The evidence is convincing for the specific geometry and chemistry tested. The paper combines spatially resolved spectra, reference-based decomposition, acid/base perturbations, and reaction-diffusion modeling rather than inferring kinetics from a single bulk observable. The extracted rate constant is reported as consistent with picosecond elementary exchange, so the apparent microsecond delay is not a contradiction; it reflects the limited concentration of hydronium and hydroxide ions needed to mediate exchange at the interface.

Caveats in Practice

The measurement depends on converting optical density into concentration through reference spectra and a thickness model for the liquid sheet. The paper addresses this with calibration spectra, thickness measurements, and fits, but those steps still make the quantitative rate extraction model-dependent. The study also focuses on H₂O/D₂O and acid/base variants; its broader claim is best read as a demonstration of a measurement platform for fast liquid-liquid interfacial reactions, not as proof that all such reactions will show the same non-equilibrium delay.