Controlled molecule injector for cold, dense, and pure molecular beams at the European x-ray free-electron laser

The vacuum setup, shown schematically in the lower right corner of Fig. 1, consists of two stainless-steel chambers, highlighted by the blue and purple areas, housing the molecular-beam source and the electrostatic b-type deflector,⁵⁴ respectively. The source chamber is pumped with two turbomolecular pumps (Pfeiffer Vacuum HiPace 2300), resulting in a pumping speed on the order of 4000 l/s for helium. The deflector chamber is pumped with two turbomolecular pumps (Pfeiffer Vacuum HiPace 700), resulting in a pumping speed on the order of 1400 l/s. The typical pressure in the source and deflector chamber when operating the molecular-beam valve at 10 Hz is $∼10−7$ and $∼10−8$ mbar, respectively. Both chambers can be separated with a manual DN250 CF gate valve (VAT 10848-CE01-0008) to allow for sample changes in the source chamber without breaking the vacuum of the deflector chamber and the endstation. The source chamber is equipped with a quick access door (Pfeiffer Vacuum 420KTU250) on a DN250 CF flange to allow for fast sample replacement in the valve’s sample container.⁵⁵ The source and deflector chambers are mounted on rails attached to a stainless-steel platform to facilitate easy connection of the entire setup to the endstation. The total mass of the molecular-beam setup including the pumps is $∼500$ kg. The interface between COMO and an endstation requires a DN250 CF port with a cylindrical-shaped space with a length of 13.5 cm and a diameter of 20.4 cm inside the endstation for the inverse skimmer pot.

A schematic of the COMO molecular-beam setup is shown in Fig. 1. A commercially available long-pulse version of an Even-Lavie (EL) valve⁵⁵ with a nozzle diameter of 70 µm is used to deliver the molecular beam by a supersonic expansion of helium with a trace of molecules into the vacuum. This version of the valve produces molecular-beam pulse durations up to 100 µs by utilizing two current pulses in close succession provided by a modified driver unit. The valve is situated on a 3D manipulator (VAB PM12-300-S2EC) with travel ranges of 1.25, 1.25, and 30 cm in the x, y, and z directions, respectively. The valve body can be heated up to 250 °C to allow for vapor-pressure control of the sample under investigation. Helium is usually used as the carrier gas with a stagnation pressure on the order of tens of bars, up to 100 bar, to generate a cold molecular beam. At EuXFEL, the valve is operated at 10 Hz, matching the x-ray burst rate. A fast beam flux monitor (BFM, MBE-Komponenten BFM 40-150) mounted on a DN40 CF linear feedthrough with a linear travel range of 15 cm perpendicular to the molecular-beam-propagation direction is used to monitor the temporal profile of the molecular beam. The distance between the BFM and the first skimmer’s tip is 19.7 cm. The beam flux monitor is controlled by a modified pressure gauge controller (JEVATEC VCU-B0) that allows for measuring the collector current. The collector current is transimpedance amplified with a current amplifier (FEMTO DHPCA-100), which, in turn, is connected to a digitizer to display and record the temporal profile of the molecular beam.

All of the components used to shape the molecular beam are mounted on motorized translation stages, and the individual positions can be remotely controlled in planes perpendicular to the molecular-beam propagation direction. The first skimmer (Beam Dynamics 50.8, ∅ = 3 mm) is used for differential pumping. The distance between the first skimmer and the valve can be adjusted between 16.5 and 46.5 cm by moving the valve along the molecular-beam propagation direction. The first skimmer is mounted on a home-built flange with a built-in 2D translation stage to adjust the skimmer position in the x and y directions using two stepper motors (Thorlabs DRV014) outside vacuum and linear feedthroughs. The distance between the skimmer tip and the surface of the home-built flange is given by 4.05 cm. The second skimmer (Beam Dynamics 40.5, ∅ = 1.5 mm) is placed 9.5 cm behind the first skimmer tip for further collimation of the beam. An electrostatic b-type deflector,⁵⁴ which disperses the molecules in the beam with respect to their quantum states^31,56 is located 4.4 cm downstream of the tip of the second skimmer. Both electrodes of the electrostatic deflector are connected through two high-voltage feedthroughs (Pfeiffer Vacuum 420XST040-30-30-1) allowing for voltage differences up to 60 kV between the electrodes. The unit of the second skimmer and the electrostatic deflector are mounted on a 2D translation stage to adjust their positions simultaneously in the x and y directions using two stepper motors (Thorlabs DRV014) and linear feedthroughs. This translation stage is mounted on the back side of the same flange as the translation stage for skimmer 1. The upper left inset of Fig. 1 depicts the shape and electric fields of the deflector geometry. The numerically calculated electric-field strengths are depicted in color coding. The dashed circle indicates the typical size and position of the undeflected molecular beam. The dispersed molecular beam is then cut by a vertically adjustable knife edge placed 5.87 cm downstream of the exit of the deflector. This allows for improved sample separation and higher column density of the molecular beam.³⁷ The orientation of the knife edge can be controlled by a motorized rotation stage (Smaract SR-1908) to ensure the molecular beam is cut parallel to the x-ray propagation direction. The vertical position of the knife edge can be adjusted by a linear in-vacuum piezo stage (Smaract SLC 1750) with a travel range of 3.1 cm. The typical pressure in the deflector chamber is on the order of 10⁻⁸ mbar in operation. The molecular beam is further skimmed by a third conical skimmer (Beam Dynamics 50.8, ∅ = 1.5 mm) placed 2.52 cm downstream of the knife edge, providing differential pumping against the interaction chamber. This third skimmer is again mounted on a 2D translation stage, with the position controlled by two linear in-vacuum piezo stages (Smaract SLC 1750). Behind the third skimmer, the molecular beam enters an endstation compatible with COMO, e.g., the nano-size quantum systems (NQS), or the reaction microscope (REMI) experimental stations at the SQS instrument at EuXFEL.^9,57 The distance between the tip of the third skimmer and the interaction region with the x rays is 15.76 cm in the case of the NQS setup.

To characterize the performance of the COMO setup in-house before its delivery to EuXFEL, it was combined with a time-of-flight mass spectrometer (TOF-MS, Jordan C-677) and tested with a Ti:sapphire femtosecond laser system (Coherent Astrella) to strong-field-ionize the sample. The laser pulses with a duration of FWHM_I ∼ 30 fs and a wavelength centered at 800 nm were focused to FWHM_I ≈ 50 µm and directed perpendicular to the molecular beam along the negative x direction to ionize molecules in the extraction region of the TOF-MS. The TOF-MS had a distance of 17.6 cm with respect to the last skimmer tip along the z direction, reflecting the geometry at the EuXFEL SQS instrument. The peak intensity of the laser pulse was I₀ ∼ 6 × 10¹³ W/cm². We used pyrrole as a sample to characterize the temporal and spatial profiles of the molecular beam.⁵⁸ Pyrrole (Sigma-Aldrich Chemie GmbH 131709) was placed in the sample reservoir of the valve and heated up to 70 °C, which resulted in a pyrrole vapor pressure on the order of 100 mbar.⁵⁹ Helium was used as the carrier gas at a stagnation pressure of 20100 bar.

Figure 2(a) shows the temporal density profiles of molecular-beam pulses obtained, gating the signal on the parent cation of pyrrole (m/q = 67 u/e), for a long-pulse (violet) and a standard EL valve (green). A pulse duration of τ ∼ 100 µs is observed for the long-pulse valve in contrast to τ ∼ 25 µs in the case of the normal EL valve. In addition, for the long-pulse EL valve, two maxima can be identified during the pulse, resulting from the two current pulses of the driver unit. Furthermore, post pulses were detected at delays between 700 and 850 µs, which are attributed to the bouncing of the plunger in the EL valve driven by the two current pulses. The post pulses are expected to gradually disappear with increasing stagnation pressure.⁶⁰ The absolute density for the long-pulse EL valve was determined by a laser-pulse-energy scan as described below. The density of the normal EL valve is not to scale, but from a signal strengths comparison, we assume similar densities as obtained for the long-pulse EL valve.

To characterize the separation of pyrrole from the direct molecular-beam, measurements were carried out with and without voltages on the deflector. Figure 2(b) shows the measured (dots) and simulated³¹ (lines) profiles for the direct (deflector off, blue) and the 14 kV (±7 kV) deflected (red) molecular beams. Both profiles were obtained by scanning the molecular beam along the y direction making use of the motorized translation stages to move the beam with respect to the laser propagation axis. The parent-ion yields were obtained by integrating over the pyrrole⁺ peak in the TOF-MS. The timing of the laser pulses with respect to the molecular-beam-valve trigger was 610 µs, as indicated by the vertical dashed line in Fig. 2(a). Comparing both profiles, a clear shift toward negative y values is observed when the deflector is switched on. This is expected, as all quantum states of pyrrole are strong-field seeking at the relevant electric-field strengths in the deflector.⁵⁸ The solid lines in Fig. 2(b) are simulated pyrrole beam profiles obtained from Monte Carlo trajectory calculations that take into account the geometrical constraints of the mechanical apertures and the rotational temperature T_rot of the molecular beam.^58,61 The experimental data of pyrrole’s deflection profiles match well with the simulated results, which assume an initial temperature of T_rot = 1 K for the molecular beam entering the deflector. Furthermore, the results are comparable to those obtained in earlier measurements.^58,62

The pyrrole molecular-beam sample densities were estimated based on a strong-field ionization model:^62–64 The asymptotic slope of an integral ionization signal with respect to the natural logarithm of peak intensity can be expressed as S = ϱπR²Dα. Here, ϱ represents the sample density, R represents the I₀/e radius of the transverse laser intensity distribution, D represents the length of the focal volume in the molecular beam, and α represents the detection efficiency. The detected number of pyrrole ions per laser pulse as a function of the logarithm of the laser peak intensity ln(I₀) is shown in Fig. 3. The vertical laser position, molecular-beam timing, and stagnation pressure were y = 0 mm, t = 600 µs, and 90 bar, respectively. A typical increase and saturation behavior of the ion yield with increasing laser intensity is observed. The purple curve depicts a fit to the asymptotic slope. From the fit, a saturation onset of I₀ = (6.8 ± 1.1_syst) × 10¹³ W/cm² was deduced. A sample density of ϱ = (8.4 ± 4.2_syst) × 10⁷ cm⁻³ was obtained, taking into account the measured laser-beam waist of ω₀ = 2σ = 52(2) µm, the measured molecular-beam diameter of D = 2.0(3) mm, and an estimated detection efficiency of 50%.⁶⁵ The measured molecular-beam density is consistent with typical previously reported values^48,62,64 obtained with an EL valve. The inset in Fig. 3 shows the temporal profiles of the pyrrole molecular beam for stagnation pressures of 20, 50, and 90 bar. A decrease in the pulse duration is observed with increasing stagnation pressure accompanied by a reduction in the peak density. This behavior could be due to the complex gas dynamics in the source chamber and is subject to further investigations. Currently, this limits the long-pulse operation to samples where sufficient cooling is achieved at 20 bar stagnation pressure. Improvements in the valve-driving-electronics parameters are foreseen for samples of controlled and species-selected molecular clusters^40,58,66 or applications of strong molecular orientation,^42,67 which require higher stagnation pressures for efficient formation or cooling.

The COMO setup was assembled and further tested with the nano-sized quantum systems (NQS) experimental station at the SQS instrument at EuXFEL. A pure pyrrole-water (PW) heterodimer sample^58,62 was investigated. The molecular beam was produced using a stagnation pressure of 80 bar of helium with traces of water and pyrrole added. The driver settings were similar to the ones used during the characterization measurements with the optical laser described above. These conditions resulted in a similar spatial molecular-beam profile and a purity of pyrrole-water heterodimer as published previously^58,62 and also shown in Fig. 2(b). The EuXFEL and the optical laser were operated in burst mode at a base frequency of 10 Hz with 20 FEL pulses, synchronized to the same number of 1030 nm optical laser pulses per train at an intra-train repetition rate of 250 kHz and a photon energy of 600 eV. Figure 4 depicts the typical ion-yield spectrum as a function of the relative timing Δt between the molecular-beam arrival time and the train of combined x-ray and 1030 nm optical laser pulses. Vertical red dotted lines indicate the timing of each specific x-ray pulse in the train. The signal envelope as depicted by the blue-shaded area indicates the overall temporal profile of the molecular beam in the deflected part. The pulse duration of the molecular beam was on the order of 40 µs during the commissioning time. The temporal molecular-beam profile is fully covered by the 20 FEL/1030 nm pulses and the specific molecular-beam TOF spectra and the background are detected simultaneously in one pulse train. The shorter pulse duration with respect to the duration obtained with the laser shown in Fig. 3 is attributed to the 80 bar stagnation pressure required to produce pyrrole–water clusters at high density.

The inset in Fig. 4 shows a zoomed-in TOF-MS spectrum. The complete, integrated time trace was split into individual traces to obtain a TOF spectrum for each specific x-ray pulse. Afterward, the TOF spectrum corresponding to a specific x-ray pulse was converted to the mass-over-charge ratio in units of u/e. The main ion peaks of the TOF-MS spectrum are labeled and attributed to the pyrrole-water parent ion (PW⁺), pyrrole-water fragment ions, and remaining seed-gas contributions (He⁺). A detailed analysis of the TOF-MS spectrum is beyond the scope of this paper and will be discussed in a future publication. However, all ion signals induced by the x rays can be clearly resolved in the TOF-MS spectrum, demonstrating that the COMO setup is fully compatible with the chosen burst-mode operation at the EuXFEL. This will significantly increase the data-acquisition duty cycle and, thus, the utilization of x-ray pulses, compared to experiments using standard short-pulse valves up to a factor of $∼4$⁠. It should be mentioned here that there are no relevant restrictions on the number of pulses with which the valve is operated, up to several 100 pulses per second. By modifying the electric driver, longer molecular pulses will be possible to cover the 400 µs duration of the EuXFEL pulse trains. Furthermore, our results demonstrate that the obtained molecular sample densities are clearly sufficient to conduct experiments with x rays at free-electron lasers using COMO. In addition, by exploiting the electrostatic deflector, the x-ray-induced helium-ion signal, from the molecular-beam seed gas, is highly suppressed. However, to implement 100 µs or longer molecular-beam pulses under conditions to produce significant densities of molecular clusters, further optimization of the valve performance is required.

In summary, the permanently available molecular-beam injector setup COMO was developed, set up, and commissioned using TOF-MS experiments both in-house at CFEL with optical laser ionization and at EuXFEL with x-ray ionization detection. The performance of the long-pulse-valve molecular-beam source was characterized and proven to be compatible with the burst-mode operation of the EuXFEL. Our results show that the COMO setup can generate cold (T ∼ 1 K) and dense (ρ ∼ 10⁸ cm⁻³) molecular-beam pulses with pulse durations on the order of 100 µs. The high molecular-beam density allows experiments with data recorded in single-shot mode. In general, the COMO molecular-beam source can be combined with various electron and ion spectrometers, along with large-area x-ray detectors. With the included electrostatic deflector, pure state-, size-, and isomer-selected samples of polar molecules and clusters can be provided to the interaction region.

Hence, COMO proves to be an excellent source for various experiments at the EuXFEL, such as the recording of molecular movies using ion, electron, and x-ray imaging. This enables a diverse range of science from atomic, molecular, and cluster physics to materials and energy science, as well as chemistry and biology.

We thank Uzi Even, Nachum A. Lavie, and Ronny Barnea for designing and making commercially available the long-pulse Even-Lavie valve and for sharing early test results from their laboratory.

We acknowledge the financial support by Deutsches Elektronen-Synchtrotron DESY, a member of the Helmholtz Association (HGF), the use of the Maxwell computational resources operated at Deutsches Elektronen- Synchrotron DESY, 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), the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska-Curie Grant Agreement “Molecular Electron Dynamics investigated by Intense Fields and Attosecond Pulses” (MEDEA, 641789), the Helmholtz Foundation through funds from the Helmholtz-Lund International Graduate School (HELIOS, Helmholtz project number HIRS-0018), and the German Federal Ministry of Education and Research (BMBF) and the Swedish Research Council (VR 2021-05992) through the Röntgen-Ångström cluster (project “Ultrafast dynamics in intermolecular energy transfer: elementary processes in aerosols and liquid chemistry” UDIET, 05K22GUA). We also acknowledge the scientific exchange of the Center for Molecular Water Science (CMWS). For the experiments at EuXFEL, we thank EuXFEL in Schenefeld, Germany, for the provision of XFEL beam time at the SQS instrument and we thank the staff for their assistance.

L.H. acknowledges the support of the National Natural Science Foundation of China (Grant Nos. 11704147 and 92261201) and a fellowship within the framework of the Helmholtz-OCPC postdoctoral exchange program. F.T. acknowledges the funding by the Deutsche Forschungsgemeinschaft (DFG) (Project No. 509471550), Emmy Noether Program and acknowledges the support by the MaxWater initiative of the Max-Planck-Gesellschaft. S.B. and L.S. acknowledge the funding by the Helmholtz Initiative and Networking Fund. A. R. acknowledges the support from the U.S. Department of Energy, Office of Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division under Award No. DE-FG02-86ER13491.

The authors have no conflicts to disclose.

Lanhai He: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Software (equal); Writing – original draft (equal); Writing – review & editing (equal). Melby Johny: Conceptualization (equal); Data curation (equal); Writing – review & editing (equal). Thomas Kierspel: Data curation (equal); Writing – review & editing (equal). Karol Długołęcki: Conceptualization (equal). Sadia Bari: Data curation (equal); Writing – review & editing (equal). Rebecca Boll: Data curation (equal); Writing – review & editing (equal). Hubertus Bromberger: Data curation (equal); Writing – review & editing (equal). Marcello Coreno: Data curation (equal); Writing – review & editing (equal). Alberto De Fanis: Data curation (equal). Michele Di Fraia: Data curation (equal); Writing – review & editing (equal). Benjamin Erk: Data curation (equal); Writing – review & editing (equal). Mathieu Gisselbrecht: Data curation (equal); Funding acquisition (equal); Writing – review & editing (equal). Patrik Grychtol: Data curation (equal); Writing – review & editing (equal). Per Eng-Johnsson: Data curation (equal); Funding acquisition (equal); Writing – review & editing (equal). Tommaso Mazza: Data curation (equal); Writing – review & editing (equal). Jolijn Onvlee: Data curation (equal); Software (equal); Writing – review & editing (equal). Yevheniy Ovcharenko: Data curation (equal); Writing – review & editing (equal). Jovana Petrovic: Data curation (equal); Writing – review & editing (equal). Nils Rennhack: Data curation (equal). Daniel E. Rivas: Data curation (equal); Writing – review & editing (equal). Artem Rudenko: Data curation (equal). Eckart Rühl: Data curation (equal); Writing – review & editing (equal). Lucas Schwob: Data curation (equal). Marc Simon: Data curation (equal). Florian Trinter: Data curation (equal); Writing – review & editing (equal). Sergey Usenko: Data curation (equal); Writing – review & editing (equal). Joss Wiese: Data curation (equal); Software (equal); Writing – review & editing (equal). Michael Meyer: Conceptualization (equal); Data curation (equal); Funding acquisition (equal); Investigation (equal); Project administration (equal); Resources (equal); Software (equal); Supervision (equal); Writing – review & editing (equal). Sebastian Trippel: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Project administration (equal); Software (equal); Supervision (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Jochen Küpper: Conceptualization (equal); Data curation (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal); Project administration (equal); Resources (equal); Supervision (equal); Validation (equal); Visualization (equal); Writing – review & editing (equal).

The data recorded for the experiment at the EuXFEL are available at https://doi.org/10.22003/XFEL.EU-DATA-002388-00. The data recorded in the laser lab at DESY are available from the corresponding author upon reasonable request. The data that support the findings of this study are available from the corresponding author upon reasonable request.