Pure Molecular Beam of Water Dimer

TL;DR

This study demonstrates the spatial separation of water dimer molecules with high purity and low temperature using an electric deflector, and investigates their ionization behavior under strong laser fields, revealing proton transfer and dissociation pathways.

Contribution

First to produce a pure, cold water dimer molecular beam via electric deflection and analyze its strong-field ionization dynamics.

Findings

Water dimer beam with 93% purity and 1.5 K temperature achieved.

Proton transfer observed in ionized water dimer.

46% of ionized dimers dissociate into H3O+ and neutral OH.

Abstract

Spatial separation of water dimer from water monomer and larger water-clusters through the electric deflector is presented. A beam of water dimer with $93~\%$ purity and a rotational temperature of $1.5~$K was obtained. Following strong-field ionization using a $35~$fs laser pulse with a wavelength centered around $800~$nm and a peak intensity of $10^{14}~\text{W}/\text{cm}^2$ we observed proton transfer and $46~\%$ of the ionized water dimer broke apart into a hydronium ion $\text{H}_3\text{O}^+$ and neutral OH.

Full text

I Introduction

Hydrogen bonding between water molecules plays an important role in aqueous systems, e. g., for biomolecules that are surrounded by solvents. It is responsible for the unique properties of water, such as its high boiling point Jeffrey (1997). While hydrogen bonds have been studied extensively in many different molecular systems Liu et al. (1996); Nauta and Miller (2000); Dunning et al. (2015); Berden et al. (1996); Korter et al. (1998); Sobolewski and Domcke (2001); Ren et al. (2018), one of the most important models remains the water dimer, somehow the smallest drop of water. Numerous studies have been conducted on this benchmark system and its structure with a single hydrogen bond is well known Odutola and Dyke (1980); Yu and van Gunsteren (2004); Dyke et al. (1977); Dyke and Muenter (1972).

Water molecules and water-clusters have been studied using various techniques to describe dynamics such as proton motion Marchenko et al. (2018) or chemical processes, e. g., reactive collisions Kilaj et al. (2018). For investigations of ultrafast molecular dynamics, such as energy and charge transfer across hydrogen bonds in molecular systems, photoion-photoion coincidence measurements at free-electron lasers are developing as a powerful tool Boll et al. (2016); Kierspel (2016); Ren et al. (2018) and this approach was also used to study hydrogen bonding in the water dimer at a synchrotron Jahnke et al. (2010). Other spectroscopic techniques utilizing synchrotron facilities Winter et al. (2007); Smith et al. (2004) or table-top laser-systems Keutsch and Saykally (2001); Berden et al. (1996); Korter et al. (1998); Zwier (1996) further improved the knowledge about hydrogen bonding in water and water-clusters.

Most of these experiments investigating the dynamics of hydrogen-bonded systems would benefit from samples of identical molecules in a well-defined initial state. The widely used supersonic expansion technique provides cold molecular beams down to rotational temperatures of $<1$ K Scoles (1988); Even et al. (2000); Johny et al. (2019). However, cluster expansions do not produce single-species beams, but a mixture of various cluster stoichiometries. Hence, only low concentrations of specific species can be achieved. In the case of water molecules, supersonic expansion produces a cold beam of various water clusters Liu et al. (1996) with a water dimer concentration of only a few percent Jahnke et al. (2010); Paul et al. (1997). This leads to small experimental event rates and requires long measurement times, e. g., in coincidence detection schemes Jahnke et al. (2010); Kierspel (2016). These experiments with a mixture of molecules in a molecular beam are only feasible if it can be disentangled which molecule was actually measured. Therefore, these mixtures severely limit the applicable techniques. A pure beam of water dimers would significantly speed up the measurements, when unwanted backgrounds from carrier gas and larger water-clusters are avoided, or simply enable such experiments.

The electrostatic deflector is an established method to spatially separate the molecules of interest from the carrier gas and to separate different species within a cold molecular beam Chang et al. (2015). This includes the separation of molecular conformers Filsinger et al. (2008, 2009a); Kierspel et al. (2014); Teschmit et al. (2018), individual quantum states of small molecules Nielsen et al. (2011); Horke et al. (2014), as well as specific molecular clusters Trippel et al. (2012); You et al. (2018); Johny et al. (2019). The deflector was previously utilized in investigations of water, e. g., to determine the rotational temperatures of “warm” molecular beams of water Moro et al. (2007), to separate its para and ortho species Horke et al. (2014), and to measure the dipole moment of small water-clusters Moro et al. (2006). Alternatively, separation by the cluster species’ distinct collision cross sections, i. e., by the transverse momentum changes due to scattering with a perpendicular rare-gas beam, was demonstrated Buck and Huisken (2000); this method is especially amenable to larger cluster sizes Pradzynski et al. (2012). Such spatially separated single-species samples enable, for instance, advanced imaging applications of water-clusters using non-species-specific techniques, as well as the study of size-specific effects and the transition from single-molecule to bulk behavior.

II Experimental Methods

Here, the electrostatic deflector was used to spatially separate water dimers from water monomers as well as larger water-clusters in a molecular beam formed by supersonic expansion. The experimental setup was described previously Chang et al. (2015); Trippel et al. (2018). Briefly, liquid water was placed in the reservoir of an Even-Lavie valve Even et al. (2000), heated to $55\,^{\circ}{}\text{C}$, seeded in 100 bar of helium, and expanded into vacuum with a nominal driving-pulse duration of 19.5 µs and at a repetition rate of 250 Hz. The produced molecular beam was doubly skimmed, 6.5 cm ($\varnothing=3$ mm) and 30.2 cm ($\varnothing=1.5$ mm) downstream from the nozzle, directed through the electrostatic deflector Kienitz et al. (2017) of 154 mm length and with a nominal field strength of 50 $\text{kV}/\text{cm}$ with an applied voltage of 8 kV across the deflector, before passing through a third skimmer ($\varnothing=1.5$ mm). The deflector was placed 4.4 cm behind the tip of the second skimmer. In the center of a time-of-flight (TOF) mass spectrometer, 134.5 cm downstream from the nozzle, molecules were strong-field ionized by a 35 fs short laser pulse with a central wavelength around 800 nm and a pulse energy of 170 µJ. Focusing to 65 µm yielded a peak intensity of $\mathord{\sim}10^{14}~{}\text{W}/\text{cm}^{2}$. The generated ions were accelerated toward a microchannel-plate detector combined with a phosphor screen and the generated signal was recorded with a digitizer. The valve, skimmers, and deflector were placed on motorized translation stages, which allowed movement of the molecular beam through the ionization laser focus and the recording of vertical molecular-beam-density profiles without moving the laser focus, resulting in fixed imaging conditions Küpper et al. (2014); Stern et al. (2014); Filsinger et al. (2009b).

While the employed strong-field ionization is a general, non-species specific ionization technique, it can also lead to fragmentation of molecules, such that recorded mass spectra (MS) do not directly reflect the composition of the molecular beam. In combination with the species-specific deflection process, however, this can be disentangled and, thus, even allows for the investigation of strong-field-induced fragmentation processes of a single species.

III Results and Discussion

TOF-MS of the direct and the deflected beams are shown in Fig. 1 . The spectrum of the undeflected beam shows water-cluster ions $(\text{H}_{2}\text{O})_{n}^{+}$ up to $n=2$ and protonated water-cluster ions $(\text{H}_{2}\text{O})_{n}\text{H}^{+}$ up to $n=10$. Even larger clusters were likely formed in the supersonic expansion, but were not observed in the recorded TOF interval. We point out that all clusters that reach the interaction region are neutral clusters of the type $(\text{H}_{2}\text{O})_{n}$, and protonated clusters must result from the interactions with the femtosecond laser, i. e., due to fragmentation during or after the strong-field-ionization process.

Vertical molecular-beam-density profiles for water ions $(\text{H}_{2}\text{O})^{+}$, water dimer ions $(\text{H}_{2}\text{O})_{2}^{+}$, and protonated water-cluster ions $(\text{H}_{2}\text{O})_{n}$H+ up to $n=4$, with a potential difference of 8 kV applied across the deflector, are shown in Fig. 2 . For comparison, a field-free vertical profile for the water ion with 0 kV across the deflector is also shown. The vertical molecular-beam-density profiles have been normalized to the area of the field-free spatial profile of the water ion. For visibility the water dimer profile has been scaled by a factor of 100 after normalization. While the field-free molecular beam profile is centered around 0 mm, application of a voltage of 8 kV to the deflector shifted the peak of water ions, water dimer ions, and protonated water-cluster ions by +0.5 mm, as indicated by the red arrow in Fig. 2 . In addition, water dimer ions showed a broadening and an increase of signal at around +3 mm, indicated by a black arrow in Fig. 2 .

In the inset of Fig. 2 the region around $+3$ mm is shown enlarged with a magnification factor of 5 applied to $(\text{H}_{2}\text{O})_{2}^{+}$ and $(\text{H}_{2}\text{O})_{n}$H+ with $n=1\ldots 4$. The corresponding TOF-MS in the deflected part of the beam at a position of +3 mm is highlighted in Fig. 1 by the blue line. Not just water dimer ions, but also hydronium ions, $\text{H}_{3}\text{O}^{+}$, and water ions, $\text{H}_{2}\text{O}^{+}$, showed an increased signal in the deflected beam. The shape of the vertical beam profiles for these ions matched the water dimer profile in the region of 2.8–3.5 mm, indicating that they originated from the same parent molecule.

The water dimer ion was the largest non-protonated cluster measured in this setup. To verify that the water dimer ion was originating from the water dimer, the deflection behaviour of water-clusters inside the electrostatic deflector was simulated. Therefore, the Stark energies and effective dipole moments of water monomers and water-clusters were calculated with the freely available CMIstark software package Chang et al. (2014) using rotational constants, dipole moments, and centrifugal distortion constants from the literature DeLucia et al. (1974); Coudert and Hougen (1990); Shostak et al. (1991); Malomuzh et al. (2014); Dyke et al. (1977), see Suppl. Inf. Table I; contributions of the polarizability to the Stark effect could safely be ignored Maroulis (2000); Kienitz et al. (2017). The rotational constants of the water dimer are significantly smaller than for the water monomer, leading to a larger effective dipole moment for the water dimer than for water and a larger acceleration in the electric field in the deflector, see Fig. 3 and Table I of the Suppl. Inf. for further information.

The simulated vertical molecular-beam-density profiles of the water monomer and the water dimer are shown in Fig. 2 . The deviations between the measured and simulated undeflected vertical beam profiles are ascribed to imperfect alignment of experimental setup, which was not taken into account in the simulations. Due to the rotational-state dependence of the Stark effect, the deflection of a molecular beam in an electrostatic field depends on the rotational temperature of the molecular ensemble Chang et al. (2015) and the best fit for the profiles of the water monomer and the water dimer at a deflector voltage of 8 kV was obtained assuming a Boltzmann population distribution of rotational states corresponding to $1.5(5)$ K.

Not only deflection of water-clusters measured as a mass of 36 amu, but also of water-clusters detected as protonated-clusters have been measured, for instance, for $(\text{H}_{2}\text{O})_{2}^{+}$, as indicated by the red arrow and symbols in Fig. 2 . Trajectory simulations for water-clusters $(\text{H}_{2}\text{O})_{n}^{+}$ with $n=3\ldots 7$ using a rotational temperature of $T_{\text{rot}}=1.5(5)$ K were performed to understand the origin of this deflection behavior. For the water hexamer three and for the water heptamer two conformers have been simulated, assuming an equal population of the conformers. These showed that, based on the different effective dipole moments, a different deflection is expected for different water-clusters, see Fig. 5 of the Suppl. Inf.. Since the detected protonated water-clusters arose from the strong-field fragmentation of larger neutral clusters in the interaction region, the measured vertical protonated-cluster density profiles are a superposition of several neutral water-cluster density profiles. Thus, it is not possible to compare the individual simulated molecular-beam-density profiles of neutral clusters directly with the measured protonated water-cluster density profiles. Instead, at each position of the deflection profile the signal from all water-clusters has been summed up, both for the simulated and the measured molecular-beam-density profiles. The result yields a comparable amount of deflection for simulated and measured molecular-beam-density profiles, see Suppl. Inf. Fig. 6. The shift of 0.5 mm can, therefore, originate from the superimposed molecular-beam-density profiles from different larger clusters due to fragmentation into smaller water-clusters. The same shift is visible for $\text{H}_{2}\text{O}^{+}$ and $(\text{H}_{2}\text{O})_{2}^{+}$, which indicates that water-clusters are also fragmenting into $\text{H}_{2}\text{O}^{+}$ and $(\text{H}_{2}\text{O})_{2}^{+}$. Nevertheless, the simulation for water-clusters $n=1\ldots 7$ shows that the water dimer deflected the most, reaching a position of +3 mm and above, see Suppl. Inf. Fig. 4 and Fig. 5. Of all the other clusters considered, only the water hexamer in its prism and book forms reaches to a position up to 3.2 mm with the falling edge of the profile. In our experiments the water hexamer and higher order clusters have only been measured as fragments, such that the concentration and size distribution of neutral clusters in the molecular beam is unknown. However, the measured fragment distributions strongly suggest that significantly larger clusters are not present, since the ion signals decay exponentially and it is known that clusters primarily fragment through loss of single water molecules Angel and Stace (2001); Belau et al. (2007); Liu et al. (2011).

The TOF-MS in the deflected part of the beam, shown in Fig. 1 , contains peaks corresponding to H*+, O+, OH+*, $\text{H}_{2}\text{O}^{+}$, and $\text{H}_{3}\text{O}^{+}$, in addition to the water dimer ion. As mentioned before the short-pulse ionization can lead to fragmentation. For the water dimer, two fragmentation channels were reported for electron-impact ionization with 70 eV electrons Angel and Stace (2001): either an $\text{H}_{3}\text{O}^{+}$ ion and a neutral OH are formed or a $\text{H}_{2}\text{O}^{+}$ ion and a neutral water monomer $\text{H}_{2}\text{O}$. Using a size-selection method and infrared spectroscopy, $\text{H}_{2}\text{O}^{+}$ has been reported as a fragment of the water dimer Buck and Huisken (2000). Comparison of the vertical molecular-beam-density profiles of the deflected molecules allowed further investigation of the fragmentation channels of the water dimer. The measured vertical molecular-beam-density profiles of these molecules showed a similar deflection behavior in the region of 2.8 to 3.5 mm as the water dimer, see Fig. 2 and Suppl. Inf. Fig. 1. The observed constant ratio of those fragments over this spatial region indicates that all these fragments originated from the water dimer.

Comparing the intensity of the fragments of the water dimer, $\text{H}_{2}\text{O}^{+}$ and $\text{H}_{3}\text{O}^{+}$ and ($\text{H}_{2}\text{O})_{2}^{+}$, in the deflected beam, the fragmentation ratios of the intact water dimer were estimated. These showed that $46(7)~{}\%$ of the water dimer fragmented into one ionized water molecule, while $46(4)~{}\%$ of the water dimer underwent most likely proton transfer and formed a hydronium ion. Only $8(2)~{}\%$ of the water dimer present in the molecular beam stayed intact after ionization.

The actual number of water dimer molecules per shot in the deflected molecular beam was estimated to $\mathord{\sim}0.8$ within the laser focus using the known fragmentation ratios of $\text{H}_{2}\text{O}^{+}$ and $\text{H}_{3}\text{O}^{+}$, while the fragmentation channels of H*+, O+, OH+* have not been included. Taking the known fragmentation channels into account, the fraction of the water dimer within the molecular beam was evaluated. Comparing the ratios between the water dimer and all other species visible in the TOF, a water dimer fraction of $3.9(6)~{}\%$ in the center of the undeflected beam and of $93(15)~{}\%$ in the deflected beam, at a position of +3 mm, was achieved. Thus, using the electrostatic deflector the fraction of the water dimer within the interaction region could be increased by nearly a factor of 24.

IV Conclusions

In summary, a high-purity beam of water dimers was created using the electrostatic deflector, which spatially separated water dimers from other species present in the molecular beam. The resulting water dimer sample had a purity of $93(15)~{}\%$. The fragmentation products and ratios of the water dimer following strong-field ionization using a 35 fs laser pulse with a wavelength centered around 800 nm and peak intensity of $\mathord{\sim}10^{14}~{}\text{W}/\text{cm}^{2}$ were studied, with $46(4)~{}\%$ of the water dimer found to form a hydronium ion and $46(7)~{}\%$ fragmenting into one water cation and one neutral water monomer, while $8(2)~{}\%$ of the water dimer stayed intact. The deflection profiles could be simulated using a rigid-rotor model and an initial rotational temperature of 1.5(5) K.

The produced clean samples of water dimers are well suited for non-species-specific experiments, e. g., reactive-collisions, diffractive imaging, or ultrafast spectroscopies Kilaj et al. (2018); Hensley et al. (2012); Küpper et al. (2014). Even for experiments that can distinguish different species, for example photoion-photoion coincidence measurements Ren et al. (2018); Kierspel et al. (2018), the produced clean beams will enable significantly faster measurements of this important hydrogen-bonded model system, e. g., because unwanted backgrounds are avoided. Furthermore, the electrostatic separation technique can be used to separate different conformers Chang et al. (2015), which could be highly interesting in the purification and studies of larger water-clusters that exhibit multiple conformers Gregory et al. (1997).

Acknowledgments

This work has been supported by the Clusters of Excellence “Center for Ultrafast Imaging” (CUI, EXC 1074, ID 194651731) and “Advanced Imaging of Matter” (AIM, EXC 2056, ID 390715994) of the Deutsche Forschungsgemeinschaft (DFG), by the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie Grant Agreement 641789 “Molecular Electron Dynamics investigated by Intense Fields and Attosecond Pulses” (MEDEA), by the European Research Council under the European Union’s Seventh Framework Program (FP7/2007-2013) through the Consolidator Grant COMOTION (ERC-Küpper-614507), and by the Helmholtz Gemeinschaft through the “Impuls- und Vernetzungsfond”. L.H. acknowledges a fellowship within the framework of the Helmholtz-OCPC postdoctoral exchange program and J.O. gratefully acknowledges a fellowship by the Alexander von Humboldt Foundation.

Supporting Information Description

Supporting Information Available: Description of the fragmentation correction method and the trajectory simulations