betaFeedback

WorksPhysics and Astronomy

Single-molecule cold chemistry: Isotopic branching ratios of the H2O++HD\mathrm{H_2O^+ + HD} reaction at cryogenic temperatures

Abstract

Isotopic branching in ion-molecule reactions influences interstellar deuterium fractionation, yet its dynamics remain poorly understood for many important systems at astrophysically relevant temperatures. Here, we investigate H- and D-atom transfer in the H2O++HD\mathrm{H_2O^+ + HD} reaction using single sympathetically cooled water ions in a cryogenic radiofrequency trap. Nondestructive identification of the products from 441 individual reaction events at a mean collision energy of 7 K reveals an H3O+/H2DO+\mathrm{H_3O^+/H_2DO^+} product branching ratio of 1.50±0.151.50 \pm 0.15 demonstrating a preferential H-atom transfer. The experiments are rationalized using quasi-classical-trajectory calculations which indicate a transition from direct dynamics at higher energies to a complex-mediated reaction under the present experimental conditions and point to an important role of intramolecular vibrational redistribution affecting the isotopic branching in this regime. The present results extend measurements into an astrochemically relevant energy regime and establish cryogenic single-ion chemistry as a new platform for resolving competing reaction pathways.

Ion–neutral reactions are often barrierless and thus dominate the chemical evolution in low-temperature environments where temperatures can be as low as a few Kelvin such as the interstellar medium (ISM) [1, 2]. In this context, the water cation H2O+\mathrm{H_2O^+} is a key intermediate in the chemistry leading to gas-phase water formation in the ISM [3, 4]. Its reactions with hydrogen isotopologues serve as a sensitive probe of H/D transfer and the resulting deuterium fractionation under the cold conditions of interstellar clouds [5, 6, 7]. In particular, the reaction of H2O+\mathrm{H_2O^+} with deuterium hydride HD is of key interest as it can proceed through two energetically allowed product channels,

H2O++HD⟶{H3O++D,H2DO++H.(1)\mathrm{H_2O^+ + HD \longrightarrow\begin{cases} H_3O^+ + D, \\ H_2DO^+ + H. \end{cases}} \tag*{(1)}

Previous laboratory work on this reaction employed rovibrationally state-selected H2O+\mathrm{H_2O^+} ions to measure integral cross sections and product branching ratios for both the H-transfer and D-transfer channels over a collision-energy range Ecoll=0.03E_{\mathrm{coll}} = 0.03–10 eV [5, 7, 8]. These experiments revealed a strong dependence of the branching ratios on the collision energy suggesting that the degree of deuterium fractionation reflects not only spatial variations in the HD abundance of the interstellar medium [9], but also the local temperature, which controls the balance between competing reaction pathways [6, 7]. Accompanying quantum wave-packet and full-dimensional quasi-classical trajectory (QCT) calculations modeled the energy-dependence of the H/D-transfer rates and rationalized them in terms of the differences in the centrifugally corrected potentials for both reaction channels [6]. However, the lowest collision energy explored in these studies (0.030.03 eV ≈350\approx350 K) remained orders of magnitude above the regime of a few Kelvins prevailing in cold molecular clouds. Extrapolations to low temperatures where subtle quantum effects such as quantum tunnelling, zero-point energy and resonances can become decisive [10] may fail considering the pronounced energy dependence of this reaction. Thus far, to our knowledge the only low-temperature data was reported by Hugo et al. [11] who used it as an in situ calibration method for determining the contamination of HD in an H2\mathrm{H_2} sample. Branching fractions of 57% and 43% were reported for the H3O++D\mathrm{H_3O^+ + D} and H2DO++H\mathrm{H_2DO^+ + H} product channels, respectively, determined in a temperature-variable 22-pole ion trap. Although their apparatus operated down to 13 K, the exact reaction temperature for this specific measurement was not reported [11].

Here, we introduce an experimental platform for studying reactions of single Coulomb-crystallized molecular ions trapped under cryogenic conditions in ultrahigh vacuum. Our approach is related to cryogenic Coulomb-crystal experiments reported recently [12], but relies on trapping single ions instead of large ensembles in order to achieve a high degree of control over the reaction partners and to faithfully emulate conditions prevalent in the ISM. In the present work, we used this experimental method for the investigation of the low-temperature reaction of single H2O+\mathrm{H_2O^+} ions with HD molecules. The neutral HD molecules were cooled by repeated surface capture and desorption enabling collision energies as low as Ecoll/kB=7E_{\mathrm{coll}}/k_{\mathrm{B} = 7} K (0.6 meV). Our data confirm a pronounced isotope effect that persists down to the low temperatures of the present study. The experimental results were rationalized by QCT calculations which indicate a switch of reaction mechanism from direct to indirect at the experimentally relevant collision energies. The current results establish cryogenic single-ion chemistry as an attractive platform for exploring ion–molecule reactions under astrochemically relevant conditions and for benchmarking theoretical models for low-temperature chemical reactivity.

The core of the present experimental setup is constituted by a linear-quadrupole radiofrequency (RF) ion trap [13] maintained at 5.4 K and surrounded by two cryogenic shields (Methods). Each experimental cycle was initiated with the preparation of a Coulomb crystal of several tens of laser-cooled Ca+\mathrm{Ca^+} ions which served as a coolant for the water ion (Figure 1(b)(Figure 1i) and Methods). Single H2O+\mathrm{H_2O^+} ions were generated from a molecular beam of water by [2+1] resonance-enhanced multi-photon ionization (REMPI) inside the trap and their translational temperature was sympathetically cooled to a few millikelvin by interaction with the laser-cooled Ca+\mathrm{Ca^+} ions [13]. Molecular ions embedded in the crystal appeared as non-fluorescent (dark) spots within the Coulomb crystal (white dashed circle in Figure 1(b)(ii)). Subsequently, calcium ions were selectively ejected from the trap by transiently adjusting the RF trapping potentials [14] to prepare a two-ion Ca+\mathrm{Ca^+}–H2O+\mathrm{H_2O^+} string (Figure 1(b)(iii) and (iv)).

Figure 1. Generation, sympathetic cooling, and reactions of trapped single H2O+\mathrm{H_2O^+} ions. (a) Schematic illustrating the main components of the experimental setup. See text for details. (b) False-color fluorescence images showing the preparation of a two-ion Ca+−H2O+\mathrm{Ca^+-H_2O^+} crystal: (i) a Coulomb crystal of laser-cooled Ca+\mathrm{Ca^+} ions; (ii) the crystal after loading a single sympathetically cooled, nonfluorescing H2O+\mathrm{H_2O^+} ion (dashed circle); (iii) progressive size reduction of the crystal; and (iv) the resulting two-ion Ca+−H2O+\mathrm{Ca^+-H_2O^+} string prior to reaction. (c) Representative experimental signature of the reaction obtained from monitoring the position of the fluorescence of the Ca+\mathrm{Ca^+} ion along the longitudinal trap axis as a function of time. The blurring of the position of the Ca+\mathrm{Ca^+} ion results from the repeated swapping of the positions of the two ions during resonant motional excitation. A discrete change in the fluorescence response at the point of reaction (vertical dashed line) signals a reaction of H2O+\mathrm{H_2O^+} with HD heralded by the breakdown of the resonant excitation due to a change of the mass of the molecular ion. The horizontal dashed lines indicate the two characteristic positions of Ca+\mathrm{Ca^+} during motional excitation. Inset: images of the Ca+\mathrm{Ca^+} fluorescence (i) before reaction during resonant excitation, and (ii) after reaction with motional excitation suppressed.

To verify the chemical identity of the sympathetically cooled ions and track changes resulting from chemical reactions, the mass of the dark ion in each experiment was measured by a resonant excitation of the motion of the two ions in the trap using a time-varying electric field applied to one of the trap electrodes (Methods). After verification of the initial presence of a water ion, the ions were continuously motionally excited at the frequency of the axial in-phase mode of the Ca+−H2O+\mathrm{Ca^+-H_2O^+} two-ion crystal while being exposed to the co-reactant HD molecules (inset (i) in Figure 1(c)). A reaction event of the H2O+\mathrm{H_2O^+} ion was heralded by a disruption of the motional excitation at the original drive frequency due to a change of the mass of the sympathetically cooled ion (inset (ii) in Figure 1(c)). The identity of the product ion was subsequently established by resonant-excitation attempts at frequencies corresponding to the masses of the expected products. Under the present experimental conditions, the motional frequencies of the possible products which differ by a single mass unit were sufficiently different so as to enable their unambiguous identification (see Table S1 in the Supplementary Information (SI)).

Ion pairModeFrequency (kHz)
Ca+\mathrm{Ca}^{+}–H2O+\mathrm{H}_{2}\mathrm{O}^{+}Axial IP531.9 kHz
Ca+\mathrm{Ca}^{+}–H3O+\mathrm{H}_{3}\mathrm{O}^{+}Axial IP529.2 kHz
Ca+\mathrm{Ca}^{+}–H2DO+\mathrm{H}_{2}\mathrm{DO}^{+}Axial IP526.4 kHz

Table S1.

For studying the target reaction, neutral HD gas (Cambridge Isotope Laboratories Inc., isotopic purity ∼\sim97%) was dosed into the cryogenic environment using a leak valve. Before reaching the trap to react with the ions, the HD molecules had to diffuse through small apertures into the volume enclosed by the cryogenic shields of the vacuum chamber (Figure 1(a)). Within the space enclosed by the shields, the molecules undergo multiple collisions, adsorption and desorption events with the cryogenic housings leading to thermalization and equilibration at the temperature of the walls. The partial pressure of HD in the reaction zone was determined to be PHD=5.7(2.5)×10−12P_{\mathrm{HD}} = 5.7(2.5) \times10^{-12} mbar (see Sec. S2 of the SI for details).

In the absence of the HD co-reactant, the trap lifetime of H2O+\mathrm{H_2O^+} was primarily constrained by its reactivity with H2\mathrm{H_2} which typically constitutes the dominant residual gas in ultrahigh vacuum (UHV). Loss of H2O+\mathrm{H_2O^+} due to reactions H2O++H2⟶H3O++H\mathrm{H_2O^+ + H_2 \longrightarrow H_3O^+ + H} therefore limited the achievable storage time τbg\tau_{\mathrm{bg}} of the ions and constituted the major background process affecting the study of the title reaction. The mean storage time τbg\tau_{\mathrm{bg}} of H2_2O+^+ limited by background reactions was subsequently determined from 20 independent reaction events under identical conditions. A result of τbg=4.5(1.0)\tau_{\mathrm{bg}} = 4.5(1.0) h was obtained, from which a residual H2_2 partial pressure PH2=3.0(0.7)×10−14P_{\mathrm{H}_2} = 3.0(0.7) \times10^{-14} mbar was determined (see § 6 of the SI for details). This value is about two orders of magnitude lower than the partial pressure of the neutral reactant gas used in the the present experiments.

The present [2+1] REMPI scheme has been shown to prepare H2_2O+^+ in its vibronic ground state [15]. It can further be assumed that the population is initially spread over several low-lying rotational states dictated by rotational photoionization dynamics [16]. However, as the time-to-reaction is typically of order 100 s under the present conditions (see Figure S1 in the SI), the ions can at least partially rotationally thermalize by spontaneous emission and optical pumping by blackbody-radiation (BBR) in the cryogenic environment during that time. Since the trapping region is enclosed predominantly by the 5.4 K cryogenic shield, the BBR field experienced by the ions is dominated by this temperature, with only minor contributions from warmer surfaces visible to the ions through small apertures. The rotational temperatures can therefore be assumed to be close to thermal equilibrium: the JKaKc=000J_{K_aK_c} = 000 rotational state accounts for approximately 96% of the population, whereas the 111111, 202202, and 211211 states contribute approximately 3.5%, 0.04%, and 0.01%. Here, JJ denotes the rotational-angular-momentum quantum number and KaK_a and KcK_c the asymmetric-top pseudo quantum numbers. In summary, it can thus be assumed that the ions were translationally cooled to mK temperatures, prepared in their ground vibrational state and rotationally relaxed close to thermal populations.

Figure S1. Lifetimes of trapped H2_2O+^+ limited by reactive collisions. (a) Survival of trapped H2_2O+^+ ions limited by reactions with background H2_2 gas. The empirical survival probability as a function of time tt (blue points/line), defined as the fraction of ions with lifetime ≥t\ge t, was fit with a single-exponential model using maximum-likelihood estimation (MLE), yielding a lifetime τbg=4.5±1.0\tau_{\mathrm{bg}} = 4.5 \pm1.0 h (red dashed line; shaded band indicates the fit uncertainty). (b) Lifetime of trapped H2_2O+^+ ions limited by reactive loss in the presence of HD. The lifetimes determined on each measurement day (blue points with error bars) were obtained from a maximum-likelihood estimation (MLE) analysis performed independently for each day. The mean reactive-loss lifetime, τHD=98±44\tau_{\mathrm{HD}} = 98 \pm44 s, was calculated as the average of the daily values.

Product branching ratios for the title reaction were determined by counting the number of H3_3O+^+ and H2_2DO+^+ products over 441 independent reaction events measured over 16 measurement days. The product mass and the time-to-reaction (defined as the interval between loading the dark H2_2O+^+ ion and the first observed reaction signature) were recorded for each event. Because the HD sample employed was only ∼97%\sim97\% isotopically enriched, this implied a ∼3%\sim3\% contribution from reactions with H2_2 present as major impurity in our HD gas sample. The measured product branching ratios were consequently corrected for the H2_2 impurity (see § 7 in the SI for details).

Figure 2 (a) and (Figure 2b) shows the statistics of the individual reaction outcomes as well as the H3_3O+^+/H2_2DO+^+ product branching ratios determined on each measurement day, respectively. The resulting impurity-corrected branching ratio, defined as the ratio of H-transfer to D-transfer product yields H3_3O+^+/H2_2DO+^+, was found to be 1.50(±\pm0.15) averaged over all data. The quoted uncertainty represents the 1σ\sigma statistical uncertainty assuming Poissonian statistics.

Figure 2. Product branching in the H2_2O+^+ + HD reaction. (a) Counts of product ions of the title reaction obtained from single-ion reaction experiments on 16 different measurement days. (b) Measured branching fractions for formation of H3_3O+^+ (H transfer) and H2_2DO+^+ (D transfer) computed from the data in panel (a) for each measurement day. Error bars denote 1σ\sigma statistical uncertainties, assuming independent Poisson counting statistics. (c) H3_3O+^+/H2_2DO+^+ product branching ratio as a function of collision energy comparing the present result with quasi-classical trajectory (QCT) calculations and experimental data from the literature at higher energies. See text for details.

To evaluate the effective collision energy associated with this result, it needs to be considered that the single H2_2O+^+ ions were sympathetically cooled to translational temperatures of a few millikelvin [13], whereas the neutral HD molecules were assumed to thermalize with the cryogenic housing and therefore exhibit a Maxwell–Boltzmann velocity distribution at a temperature of 5.4 K. Due to continuous RF excitation of the ions, a minor energy contribution from driven ion motion has to be considered as well. Overall, a mean collision energy Ecol/kB≃7E_{\mathrm{col}}/k_{\mathrm{B}} \simeq7 K (0.6 meV) was determined (see Sec. S5 of the SI for details).

To rationalize the experimental observations, QCT calculations were performed to compute the branching ratio of the H2_2O+^+ + HD reaction. The results within the collision-energy range from 1 meV to 1 eV are displayed together with the present and previous [7] experimental data in Figure 2 (c), as well as with previous theoretical results from Ref. [6]. The present experimental value for the branching ratio of 1.50 ±\pm 0.15 at 0.6 meV is in qualitative agreement with the theoretical value of 1.15 at 0.9 meV. The QCT calculations predict that the branching ratio increases with collision energy, reaches a maximum around 40–80 meV, and then declines again. This turnover is understood as the result of the superposition of two competing trends: a direct mechanism that proceeds without the reaction partners spending significant contact time, and a complex-forming (indirect) mechanism in which the reactants are temporarily captured as an ion–molecule complex in the pre-reaction well of the potential energy surface (see Figure 1(a) of Ref. [6] for a schematic) in which they can undergo extensive intramolecular vibrational energy redistribution (IVR). The competition of the two mechanisms is evidenced in Figure 3 which shows their contributions to the total branching ratio (panel (a)) as well as the fraction of trapped vs. direct trajectories (panel (b)) as a function of the collision energy. The indirect mechanism prevails at low energies, but gives way to direct dynamics in the energy region around the maximum in the branching ratio.

Figure 3. Theoretical results from QCT calculations for the H2_2O+^+ + HD reaction (a) Branching ratio plotted as a function of collision energy for the different trajectories contributing to the reaction. (b) Fraction of trajectories ff corresponding to the direct and indirect ("trapped") mechanism as a function of the collision energy. See text for details.

The decline of the branching ratio with increasing collision energy beyond 100 meV has already been rationalized in Ref. [6]. There, it was concluded that the reaction proceeds via a direct mechanism in which the centrifugal barrier for the D-transfer channel is substantially higher than that for the H-transfer channel, thus favoring the production of H3O+\mathrm{H_3O^+}. However, as the collision energy increases and thus more partial waves contribute, the difference between the two effective barriers diminishes, leading to a monotonic decline in the branching ratio.

At the low collision energies at which the indirect mechanism prevails, the branching ratio obtained from the trapped trajectories decreases monotonically with decreasing EcolE_{\mathrm{col}}. The prolonged residence of the reaction partners in the pre-reaction well allows IVR to efficiently redistribute their initial translational energy among internal modes of the complex which seems to attenuate an isotopic preference of the reaction. The proposed role of IVR for the reaction is further substantiated by the correlation between the residence time in the pre-reaction well and the product branching ratio, as presented in Figure S3 of the SI. For trajectories with short lifetimes (i.e., those that escape the well quickly), the branching ratio is high. However, as the lifetime of the complex increases, the branching ratio decreases indicating that longer IVR periods lead to an increasing loss of the initial mass preference.

Figure S3. Branching ratio as a function of the continuous residence time in the pre-reaction well for trapped trajectories at a collision energy of 0.02 kcal/mol (10 K).

The quantitative discrepancy between the present experimental and computed branching ratios at low energies may be due to the neglect of quantum effects in the theory, e.g., hydrogen tunneling, or inaccuracies in the potential energy surface. In addition, experimental factors such as isotope-dependent diffusion kinetics of the neutral reactant in the experimental chamber and fractionation induced by different condensation kinetics of HD and H2_2 may play a role as well. The possible ramifications of such effects are discussed in § 7 of the SI.

In summary, the astrochemically relevant reaction between H2O+\mathrm{H_2O^+} and HD was investigated at collision energies Ecoll/kB=7E_{\mathrm{coll}}/k_{\mathrm{B}} = 7 K using a novel cryogenic single-ion experimental approach. The present work extends laboratory studies for this reaction towards the low-temperature regime of interest for interstellar chemistry. By directly counting reaction outcomes for isolated events, branching fractions for the competing H-atom and D-atom transfer channels were determined under isolated conditions. The observed preference for H-transfer observed at higher collision energies was found to persist into the cold regime revealing a significant intramolecular isotope effect which is, however, less pronounced than observed at specific higher energies. The experimental results are interpreted with the help of dynamics calculations which reveal a change of reaction mechanism from direct to indirect dynamics at the low energies of relevance here. The switch of mechanism directly impacts the product branching ratios and points to an increasing role of IVR at low energies.

We have demonstrated an experimental platform for studying cold ion-molecule reactions at the single-particle level by combining long-lived trapping of individual molecular ions, internal cooling, and non-destructive product identification. More broadly, this work establishes cold single-ion chemistry as an attractive tool for disentangling ion–molecule reaction dynamics and for testing theoretical reaction models under well-defined conditions. In the future, this platform will enable studies of state-dependent collisions of molecular ions, including measurements of elastic-to-inelastic scattering ratios relevant for buffer-gas cooling. Integration with Stark- or Zeeman-decelerated neutral beams [17] and of minimally perturbative state-readout techniques [18] offers pathways toward simultaneous control of collision energy and internal state, and ultimately toward full state-to-state resolution.

Methods

Experimental methods

Experimental apparatus: The experimental apparatus used in the present study consisted of a molecular-beam source for generating a supersonic molecular beam of water molecules and a linear radiofrequency (RF) quadrupole ion trap [13, 14] for reaction studies housed within a cryogenic environment, see Figure 1(a) for a schematic. The system is divided into differentially pumped regions, with pressures decreasing from 5×10−75 \times10^{-7} mbar in the source chamber to 4×10−144 \times10^{-14} mbar in the science chamber. The regions were separated by gate valves. The trap is thermally anchored to a cryogenic platform connected to a custom-made low-vibration cryostat with base temperatures down to 4 K. Two concentric cryogenic shields surrounding the trap are held at 5.4 K and 23 K under operating conditions providing cryopumping of the chamber and suppressing blackbody radiation at the location of the trap. To further improve cryopumping and reduce the residual gas load, 1.5 g of activated charcoal sorbent were introduced into the cryogenic assembly.

Preparation of Ca+^{+} Coulomb crystals: Each experimental cycle was initiated with the loading of Ca+^{+} ions into the center of the trap. For that purpose, an effusive beam of neutral Ca atoms was resonantly ionized using diode-laser radiation at 423 nm and 375 nm to generate Ca+^{+} ions [14]. The Ca+^{+} ions were Doppler laser cooled by driving the closed cycling transitions (4s) 2S1/2↔(4p) 2P1/2↔(3d) 2D3/2(4s)\,^{2}S_{1/2} \leftrightarrow(4p)\,^{2}P_{1/2} \leftrightarrow(3d)\,^{2}D_{3/2} using diode lasers at 397 nm and 866 nm [13] to form Coulomb crystals in the trap as shown in Figure 1(b)(Figure 1i). The resonance fluorescence from the laser-cooled ions was collected by an objective and imaged onto an electron-multiplied charge-coupled device (EMCCD) camera.

Generation and trap loading of single water ions: A pulsed supersonic beam of H2_2O molecules was generated by expanding water vapor seeded into 1.5 bar of high-purity argon (99.996%) as a carrier gas through a piezoelectric valve. The resulting beam was collimated by two skimmers with apertures of 1.0 mm and 0.5 mm before entering the trap. In the trap center, single water ions were generated by [2+1] resonance-enhanced multi-photon ionization (REMPI) via the C~\tilde{C} electronic state of H2_2O [19]. The required radiation at 248 nm was produced by a dye laser (LIOPTEC) operating with Coumarin 307 dye, pumped by the third harmonic of an Nd:YAG laser (Innolas Spotlight 1500) at 10 Hz. The output of the dye laser was frequency-doubled to obtain 1.5 mJ of the target wavelength, which was focused into the trap center by a lens with a focal length of 50 cm. Once the molecular ion was loaded, the gate valve isolating the trap chamber was closed to prevent streaming of background gas into the reaction region.

Resonant motional excitation: For a two-ion crystal composed of a 40Ca+^{40}\mathrm{Ca}^{+} ion of mass mCam_{\mathrm{Ca}} and a dark ion of mass mdm_{\mathrm{d}}, the motion along the longitudinal symmetry axis of the linear trap is described by two normal modes [20]. Their frequencies f±f_{\pm} can be expressed in terms of the mass ratio μ=md/mCa\mu= m_{\mathrm{d}}/m_{\mathrm{Ca}} and the independently measured motional frequency of a single Ca+^{+} ion, fz,Caf_{z,\mathrm{Ca}}, as [21, 22],

f±=fz,Ca1μ[(1+μ)±1−μ+μ2].(2)f_{\pm} = f_{z,\mathrm{Ca}} \sqrt{\frac{1}{\mu}\left[(1+\mu) \pm\sqrt{1-\mu+\mu^{2}}\right]}. \tag*{(2)}

The lower-frequency mode f−f_{-} corresponds to an in-phase (IP) motion, while the higher-frequency mode f+f_{+} corresponds to an out-of-phase (OP) mode. In the present experiments, a weak sinusoidal voltage Vtsin⁡(ωt)V_{\mathrm{t}}\sin(\omega t) was applied to a trap electrode to excite the IP mode of the two-ion crystal [23, 24, 25, 26]. When the drive frequency ff matched a resonance according to Eq. (2), the amplitude of the motion of the ions in the trap increased, reducing the efficiency of laser cooling because of the Doppler effect [20] thus causing a change in the fluorescence of the laser-cooled Ca+^{+} ion. This manifested itself as a blurring and swapping of the Ca+^{+} fluorescence in the images due to periodic modulation [23, 25] (inset (i) in Figure 1(b)). By measuring the fluorescence response as a function of ff and comparing the observed mode frequency with the theoretical values for a two-ion crystal according to Eq. (2), the mass of the trapped molecular ion was determined.

Theoretical methods

Quasi-classical trajectory calculations The quasi-classical trajectory (QCT) calculations were performed using the standard technique implemented in VENUS [27] on our previous potential energy surface [28]. The trajectories were initiated with a 13 Å separation between reactants (H2_2O$⁺ and HD), and terminated when products reached a separation of 6 Å, or when the reactants were separated by 12 Å for non-reactive trajectories. The collision energy was systematically varied from 0.02 to 20 kcal/mol (0.9 meV to 0.9 eV). At the specified collision energy, the initial vibrational and rotational states for both reactants were kept in their ground states. The maximal impact parameter (bmaxb_{\mathrm{max}}) was determined at each collision energy using small batches of trajectories with trial values and ranged from 12 Å at 0.02 kcal/mol to 6.0 Å at 10.0 kcal/mol. Energy conservation was carefully maintained for all trajectories, with a variation of less than 0.04 kcal/mol, by setting the propagation time step to 0.02 fs. If a trajectory took longer than 4 ps to propagate, it was discarded. Statistical accuracy was ensured by running batches of 20,000 trajectories at each collision energy. The reaction probabilities for the H-atom transfer and D-atom transfer reaction, PrHP_r^H and PrDP_r^D, were given by the ratio between the number of reactive trajectories (NrHN_r^H and NrDN_r^D) and the total number of trajectories (NtotalN_{\mathrm{total}}), respectively, and their branching ratio is PrH/PrD=NrH/NrDP_r^H/P_r^D = N_r^H/N_r^D.

Criteria for classifying direct and trapped trajectories A trajectory was considered to be residing in the pre-reaction well if the following three conditions are met simultaneously: (1) the distance between the oxygen atom and the center of mass of the incident HD lies between 1.6 and 4.0 Å, indicating that the system is within the short range interaction region but not yet in the tight product complex; (2) the internuclear distance of the incident HD is less than 1.1 Å, ensuring that the HD bond remains largely intact and has not undergone significant elongation or rupture; and (3) the distance from the oxygen atom to the third nearest H/D atom is greater than 1.35 Å, which excludes configurations where a new O–H/D bond has already been formed. A trajectory was classified as “trapped” if these three conditions were continuously satisfied for more than 200 fs, corresponding to a sufficiently long residence time in the well to allow for efficient IVR (see § 9 for details). Conversely, if the continuous well-residence time is 200 fs or less, the trajectory is considered “direct”, implying that the system passes through the interaction region and forms products without substantial trapping.

Acknowledgements

We thank Georg Holderied, Dominik Baumgartner, Anatoly Johnson, Grischa Martin and Philipp Knöpfel for technical support. This work was supported financially by the Swiss National Science Foundation (grant nr. TMAG-2_209193) and the University of Basel. P.P. thanks the G. H. Endress foundation for financial support. A.L. acknowledges the support of the National Natural Science Foundation of China (Grant No.22273072). H.G. is supported by Department of Energy (Grant No. DE-SC0015997).

Supporting Information

Resonance frequencies of two-ion strings

Table S1: Resonance frequencies at the present trapping conditions for the excitation of the axial in-phase (IP) mode of a two-ion string consisting of single Ca+\mathrm{Ca}^{+} and molecular ions of relevance in the present study.

Partial pressures of HD and H2_2

Partial pressures were estimated from the measured reactive-loss lifetimes at T=5.4T = 5.4 K using

pi=fikBTτ∑jfjkL,j,(3)p_i = \frac{f_i k_\mathrm{B}T}{\tau\sum_j f_j k_{L,j}}, \tag*{(3)}

where fif_i is the gas fraction (fi=1f_i = 1 for the background H2_2 measurement) and kL,jk_{L,j} is the Langevin rate coefficient for the gas species jj in the mixture.

Table S2: Estimated HD and H2_2 partial pressures assuming a fraction of fHD=0.97f_\mathrm{HD} = 0.97 and fH2=0.03f_\mathrm{H_2} = 0.03 in the HD gas sample used.

ConditionkL,jk_{L,j} (cm3^{3} s−1^{-1})LifetimePressure (mbar)
HD after dosing1.29×10−91.29 \times10^{-9}98±4498 \pm44 s(5.7±2.5)×10−12(5.7 \pm2.5) \times10^{-12}
H2_2 after dosing1.54×10−91.54 \times10^{-9}98±4498 \pm44 s(1.8±0.8)×10−13(1.8 \pm0.8) \times10^{-13}
Background H2_21.54×10−91.54 \times10^{-9}4.5±1.04.5 \pm1.0 h(3.0±0.7)×10−14(3.0 \pm0.7) \times10^{-14}

Table S2.

Characterization of background reactions

The mean lifetime τbg\tau_\mathrm{bg} of the trapped water ions limited by reactions with background gas in the vacuum chamber was determined from 20 independent reaction events under identical conditions. Ion loss was modeled as a Poissonian process with a time-independent loss rate. Accordingly, the survival probability, defined as the fraction of ions remaining after a storage time tt was described by,

S(t)=exp⁡(−tτbg),(4)S(t) = \exp\left(-\frac{t}{\tau_\mathrm{bg}}\right), \tag*{(4)}

where τbg\tau_\mathrm{bg} is the background-gas-limited trap lifetime. τbg\tau_\mathrm{bg} was determined using a maximum-likelihood estimation (MLE) applied directly to the set of measured lifetimes, yielding τbg=4.5(1.0)\tau_\mathrm{bg} = 4.5(1.0) h (Figure S1), with the quoted uncertainty given by the statistical error of the MLE fit. Assuming Langevin kinetics and the ideal gas law, the background H2_2 partial pressure was estimated from the measured reaction time τbg\tau_\mathrm{bg} as

PH2=kBTkL,H2τbg,(5)P_{\mathrm{H_2}} = \frac{k_\mathrm{B}T}{k_{L,\mathrm{H_2}}\tau_\mathrm{bg}}, \tag*{(5)}

where kLk_L the Langevin rate coefficient (Table S2). The resulting value was found to be PH2=3.0(0.7)×10−14P_{\mathrm{H_2}} = 3.0(0.7) \times10^{-14} mbar.

Correction of observed branching ratio for sample impurities

Correction for H2 contamination for product branching ratios

For gas fractions fHD+fH2=1f_{\mathrm{HD}} + f_{\mathrm{H_2}} = 1 and bimolecular rate constants kHDk_{\mathrm{HD}} and kH2k_{\mathrm{H_2}}, the fraction of reactive events due to H2_2 is

α=fH2kH2fHDkHD+fH2kH2.(6)\alpha= \frac{f_{\mathrm{H_2}} k_{\mathrm{H_2}}}{f_{\mathrm{HD}} k_{\mathrm{HD}} + f_{\mathrm{H_2}} k_{\mathrm{H_2}}}. \tag*{(6)}

Assuming H2_2 produces only H3_3O+^+ and HD produces both channels, the observed product ratio robs=N(H3O+)/N(H2DO+)r_{\mathrm{obs}} = N(\mathrm{H_3O}^+)/N(\mathrm{H_2DO}^+) must consequently be corrected. We define Atot=N(H3O+)A_{\mathrm{tot}} = N(\mathrm{H_3O}^+) and Btot=N(H2DO+)B_{\mathrm{tot}} = N(\mathrm{H_2DO}^+) as the totally measured counts of H3_3O+^+ and H2_2DO+^+, respectively, and robs=Atot/Btotr_{\mathrm{obs}} = A_{\mathrm{tot}}/B_{\mathrm{tot}}. The total number of detected events is Ntot=Atot+BtotN_{\mathrm{tot}} = A_{\mathrm{tot}} + B_{\mathrm{tot}}. Thus, the corrected product branching ratio is obtained as,

rcorr=Atot−αNtotBtot(7)r_{\mathrm{corr}} = \frac{A_{\mathrm{tot}} - \alpha N_{\mathrm{tot}}}{B_{\mathrm{tot}}} \tag*{(7)}
=Atot−α(Atot+Btot)Btot(8)= \frac{A_{\mathrm{tot}} - \alpha(A_{\mathrm{tot}} + B_{\mathrm{tot}})}{B_{\mathrm{tot}}} \tag*{(8)}
=robs(1−α)−α.(9)= r_{\mathrm{obs}}(1-\alpha) - \alpha. \tag*{(9)}

The correction to the product ratio according to eq. (9) is shown in Figure S2. The correction assumes capture-limited Langevin reaction rates with the contaminant contribution weighted by kL,H2/kL,HD=1.19k_{\mathrm{L,H_2}}/k_{\mathrm{L,HD}} = 1.19. Using the measured yields of 271 H3_3O+^+ and 170 H2_2DO+^+ events, the expected H3_3O+^+ contribution from contaminant H2_2 is subtracted to obtain the intrinsic HD branching ratio. For an assumed H2_2 contamination of 3%, the corrected ratio is H3_3O+^+/H2_2DO+=1.50^+ = 1.50.

Figure S2. Corrected H3O+/H2DO+\mathrm{H_3O^+}/\mathrm{H_2DO^+} product ratio as a function of the assumed H2\mathrm{H_2} contamination in the HD sample. For the present sample contamination of 3% H2\mathrm{H_2}, a correct product ratio of 1.50 is obtained (dashed lines).

Discussion of possible processes affecting the gas composition in the trap volume

The gas composition in the trap volume can potentially vary from the nominal composition of the sample and be modified by several processes in the cryogenic environment that are not independently constrained. Under conditions in which the reactant gas effuses into the reaction volume through small apertures, the partial pressure of the gas near the ion trap may not directly reflect the original mixture at room temperature but depends on the conductance through the apertures approximately scaling as 1/m1/\sqrt{m} where mm is the molecular mass [29]. However, on the long time scales of the present experiment, we expect mass-dependent effusion effects to be negligible. In addition, at temperatures below ∼10\sim10 K, isotope-dependent condensation on the trap walls may introduce further fractionation [30]. At the operating temperature of the trap, cryosorption on the charcoal provides an additional selective and time-dependent loss channel for H2\mathrm{H_2} and HD [31]. Because the sticking coefficients and residence times of these species on charcoal are not known at the relevant temperatures, and may also depend on surface coverage of absorbent, the ramifications of these effects cannot be quantified at present. Thus, the actual H2\mathrm{H_2} partial pressure near the trap may be higher than the nominally expected value which may imply an overestimation of the present product branching ratio.

Collision energy calculation

Assuming isotropic collisions between H2O+\mathrm{H_2O^+} and HD, the mean collision energy is given by

⟨Ecoll⟩=12μ(⟨vH2O+2⟩+⟨vHD2⟩),(10)\langle E_{\mathrm{coll}}\rangle= \frac{1}{2}\mu\left(\langle v_{\mathrm{H_2O^+}}^2\rangle+\langle v_{\mathrm{HD}}^2\rangle\right), \tag*{(10)}

where μ=mH2O+mHD/(mH2O++mHD)\mu= m_{\mathrm{H_2O^+}}m_{\mathrm{HD}}/(m_{\mathrm{H_2O^+}}+m_{\mathrm{HD}}) is the reduced mass.

HD is assumed to thermalize to the copper housing at THD=5.4T_{\mathrm{HD}} = 5.4 K. For a Maxwell–Boltzmann distribution, this corresponds to ⟨vHD2⟩=3kBTHD/mHD=4.45×104 m2 s−2\langle v_{\mathrm{HD}}^2\rangle= 3k_{\mathrm{B}}T_{\mathrm{HD}}/m_{\mathrm{HD}} = 4.45 \times10^4\ \mathrm{m^2\,s^{-2}}.

The Ca+\mathrm{Ca}^{+}–H2O+\mathrm{H_2O}^{+} crystal is driven at the axial in-phase mode frequency f=531.9 kHzf = 531.9\ \mathrm{kHz}. Molecular dynamics simulations [32] give the H2O+\mathrm{H_2O}^{+} kinetic energy of approximately 0.04 meV0.04\ \mathrm{meV} corresponding to ⟨vH2O+2⟩=4.25×102 m2 s−2\langle v_{\mathrm{H_2O}^{+}}^2\rangle= 4.25 \times10^2\ \mathrm{m^2\,s^{-2}}. Additionally, residual RF fields and coupling between motional modes may contribute to micromotion, which we estimate to introduce an additional collision energy contribution of ≈0.01 meV\approx0.01\ \mathrm{meV}.

Using μ=2.57 amu\mu= 2.57\ \mathrm{amu} in Eq. (10), we obtain

⟨Ecoll⟩=0.6 meV=7 K.(11)\langle E_{\mathrm{coll}}\rangle= 0.6\ \mathrm{meV} = 7\ \mathrm{K}. \tag*{(11)}

The collision energy is therefore governed by the thermal motion of HD.

Product branching ratio as a function complex lifetime

References

  1. [1]Snow, T. P.; Bierbaum, V. M. Ion chemistry in the interstellar medium. Annu. Rev. Anal. Chem. 2008, 1, 229–259.
  2. [2]Larsson, M.; Geppert, W. D.; Nyman, G. Ion chemistry in space. Rep. Prog. Phys. 2012, 75, 066901.DOI
  3. [3]Kumar, S. S.; Grussie, F.; Suleimanov, Y. V.; Guo, H.; Kreckel, H. Low temperature rates for key steps of interstellar gas-phase water formation. Sci. Adv. 2018, 4, eaar3417.DOI
  4. [4]Tran, T. D.; Rednyk, S.; Kovalenko, A.; Roučka, Š.; Dohnal, P.; Plašil, R.; Gerlich, D.; Glosík, J. Formation of H₂O⁺ and H₃O⁺ Cations in Reactions of OH⁺ and H₂O⁺ with H₂: Experimental Studies of the Reaction Rate Coefficients from T = 15 to 300 K. Astrophys. J. 2018, 854, 25.DOI
  5. [5]Ard, S. G.; Li, A.; Martinez Jr, O.; Shuman, N. S.; Viggiano, A. A.; Guo, H. Experimental and Theoretical Kinetics for the H₂O⁺ + H₂/D₂ → H₃O⁺/H₂DO⁺ + H/D Reactions: Observation of the Rotational Effect in the Temperature Dependence. J. Phys. Chem. A 2014, 118, 11485–11489.DOI
  6. [6]Song, H.; Li, A.; Yang, M.; Guo, H. Competition between the H- and D-Atom Transfer Channels in the H₂O⁺ + HD Reaction: Reduced-Dimensional Quantum and Quasi-Classical Studies. Phys. Chem. Chem. Phys. 2017, 19, 17396–17403.
  7. [7]Xu, Y.; Xiong, B.; Chang, Y. C.; Ng, C. Y. Isotopic and quantum-ro\vibrational-state effects for the ion–molecule reaction in the collision energy range of 0.03–10.00 eV. Phys. Chem. Chem. Phys. 2017, 19, 8694–8705.
  8. [8]Xu, Y.; Xiong, B.; Chang, Y. C.; Ng, C. Y. Communication: Rovibrationally selected absolute total cross sections for the reaction H₂O⁺(X²B₁; v₁⁺v₂⁺v₃⁺ = 000; N⁺Kₐ⁺K_c⁺) + D₂: Observation of the rotational enhancement effect. J. Chem. Phys. 2012, 137, 241101.
  9. [9]Balashev, S. A.; Kosenko, D. N. HD/H₂ ratio in the diffuse interstellar medium. Mon. Not. R. Astron. Soc.: Lett. 2020, 492, L45–L49.DOI
  10. [10]Wild, R.; Nötzold, M.; Simpson, M.; Tran, T. D.; Wester, R. Tunnelling measured in a very slow ion–molecule reaction. Nature 2023, 615, 425–429.DOI
  11. [11]Hugo, E.; Asvany, O.; Schlemmer, S. H₃⁺ + H₂ isotopic system at low temperatures: microcanonical model and experimental study. J. Chem. Phys. 2009, 130, 164302.DOI
  12. [12]Wang, C.; Gan, W.-T.; Peng, W.-C.; Tong, X. Cold Chemistry Probes Isotope Effects in the Be⁺ + HD Reaction. J. Phys. Chem. Lett. 2026, 17, 8173–8178.
  13. [13]Willitsch, S. Coulomb-crystallised molecular ions in traps: methods, applications, prospects. Int. Rev. Phys. Chem. 2012, 31, 175–199.DOI
  14. [14]Meir, Z.; Hegi, G.; Najafian, K.; Sinhal, M.; Willitsch, S. State-selective coherent motional excitation as a new approach for the manipulation, spectroscopy and state-to-state chemistry of single molecular ions. Faraday Discuss. 2019, 217, 561–583.
  15. [15]Uselman, B. W.; Boyle, J. M.; Anderson, S. L. Multiphoton ionization vibrational state selection of H₂O⁺, D₂O⁺ and HOD⁺. Chem. Phys. Lett. 2007, 440, 171–175.
  16. [16]Willitsch, S.; Merkt, F. Rovibronic photoionization dynamics of asymmetric-top molecules. Int. J. Mass Spectrom. 2005, 245, 14–25.DOI
  17. [17]Haas, D.; Scherb, S.; Zhang, D.; Willitsch, S. Optimizing the Density of Stark Decelerated Radicals at Low Final Velocities: A Tutorial Review. EPJ Tech. Instrum. 2017, 4, 6.
  18. [18]Sinhal, M.; Meir, Z.; Najafian, K.; Hegi, G.; Willitsch, S. Quantum-non demolition state detection and spectroscopy of single trapped molecules. Science 2020, 367, 1213–1218.
  19. [19]Yang, C. H.; Sarma, G.; ter Meulen, J. J.; Parker, D. H.; Western, C. M. REMPI spectroscopy and predissociation of the C̃¹B₁(v = 0) rotational levels of H₂O, HOD and D₂O. Phys. Chem. Chem. Phys. 2010, 12, 13983–13991.
  20. [20]Drewsen, M.; Mortensen, A.; Martinussen, R.; Staanum, P.; Sørensen, J. L. Nondestructive identification of cold and extremely localized single molecular ions. Phys. Rev. Lett. 2004, 93, 243201.
  21. [21]James, D. F. V. Quantum dynamics of cold trapped ions with application to quantum computation. Appl. Phys. B 1998, 66, 181–190.
  22. [22]Leibfried, D.; Blatt, R.; Monroe, C.; Wineland, D. Quantum dynamics of single trapped ions. Rev. Mod. Phys. 2003, 75, 281–324.DOI
  23. [23]Kim, K.; Chang, M.-S.; Korenblit, S.; Islam, R.; Edwards, E. E.; Freericks, J. K.; Lin, G.-D.; Duan, L.-M.; Monroe, C. Quantum simulation of frustrated Ising spins with trapped ions. Nature 2010, 465, 590–593.
  24. [24]Schmidt, J.; Louvradoux, T.; Heinrich, J.; Sillitoe, N.; Simpson, M.; Karr, J.-P.; Hilico, L. Trapping, Cooling, and Photodissociation Analysis of State-Selected H₂⁺ Ions Produced by (3 + 1) Multiphoton Ionization. Phys. Rev. Appl. 2020, 14, 024053.
  25. [25]Home, J. P.; Hanneke, D.; Jost, J. D.; Leibfried, D.; Wineland, D. J. Normal modes of trapped ions in the presence of anharmonic trap potentials. New J. Phys. 2011, 13, 073026.DOI
  26. [26]Fan, M.; Holliman, C. A.; Shi, X.; Zhang, H.; Straus, M. W.; Li, X.; Buechele, S. W.; Jayich, A. M. Optical Mass Spectrometry of Cold RaOH⁺ and RaOCH₃⁺. Phys. Rev. Lett. 2021, 126, 023002.DOI
  27. [27]Hu, X.; Hase, W. L.; Pirraglia, T. Vectorization of the general Monte Carlo classical trajectory program VENUS. J. Comput. Chem. 1991, 12, 1014–1024.DOI
  28. [28]Li, A.; Guo, H. A nine-dimensional ab initio global potential energy surface for the H₂O⁺ + H₂ → H₃⁺ + H reaction. J. Chem. Phys. 2014, 140, 224313.DOI
  29. [29]Chiggiato, P. Vacuum Technology for Ion Sources; CERN Yellow Report CERN-2013-007; Geneva, Switzerland: CERN, 2013, pp 463–502.DOI
  30. [30]Roder, H. M.; Childs, G. E.; McCarty, R. D.; Angerhofer, P. E. Survey of the Properties of the Hydrogen Isotopes Below Their Critical Temperatures; NBS Technical Note 641; Washington, DC: National Bureau of Standards, 1973.
  31. [31]Scannapieco, M.; Day, C. Experimental investigation on charcoal adsorption for cryogenic pump application. IOP Conf. Ser.: Mater. Sci. Eng. 2017, 278, 012160.
  32. [32]Rouse, I.; Willitsch, S. Superstatistical velocity distributions of cold trapped ions in molecular-dynamics simulations. Phys. Rev. A 2015, 92, 053420.DOI

Paper details

vlnc:W_ryufq3ax
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 14

downloads · 4

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents