Knife edge skimming for improved separation of molecular species by the deflector
Sebastian Trippel, Melby Johny, Thomas Kierspel, Jolijn Onvlee, Helen Bieker, Hong Ye, Terry Mullins, Lars Gumprecht, Karol Długołęcki, Jochen Küpper
NNOTE: Knife edge skimming for improved separation of molecular speciesby the deflector
Sebastian Trippel,
1, 2, a) Melby Johny,
1, 2, b) Thomas Kierspel,
1, 2, 3, b) Jolijn Onvlee, Helen Bieker,
1, 2
HongYe,
1, 2, 3
Terry Mullins, Lars Gumprecht, Karol Długołęcki, and Jochen Küpper
1, 2, 3 Center for Free-Electron Laser Science, Deutsches Elektronen-Synchrotron DESY, Notkestraße 85, 22607 Hamburg,Germany Center for Ultrafast Imaging, Universität of Hamburg, Luruper Chaussee 149, 22761 Hamburg, Germany Department of Physics, Universität Hamburg, Luruper Chaussee 149, 22761 Hamburg, Germany (Dated: 5 September 2018)
A knife edge for shaping a molecular beam is described to improve the spatial separationof the species in a molecular beam by the electrostatic deflector. The spatial separation ofdifferent molecular species from each other as well as from atomic seed gas is improved. Thecolumn density of the selected molecular-beam part in the interaction zone, which correspondsto higher signal rates, was enhanced by a factor of 1.5, limited by the virtual source size ofthe molecular beam.Molecular-beam methods are important in physicalchemistry and molecular physics, as they provide uniqueopportunities to obtain fundamental insight into mecha-nisms and dynamics of elementary molecular and chemi-cal processes. Furthermore, industrial applications usingmolecular beams range from the fabrication of thin filmsto the production of artificial structures such as quantumwires and dots.Supersonic expansion of a gas into vacuum providesextreme cooling, in the case of atomic or seeded molecularbeams typically from ambient or elevated temperaturesdown to ∼ K . This approach is used for a large varietyof experiments. In many applications the molecular beamis shaped by skimmers, knife edges, razor blades, slits,slit-skimmers, or gratings to select only the most intensepart of the beam . Furthermore, molecular beams canbe manipulated by electric and magnetic fields whichallow, e. g., the separation of the molecules from a seedgas .Spatial separation of different species is achieved by theelectrostatic deflector . Experiments are in this casetypically performed at the edge of the deflected molecularbeam to maximize the separation or to reduce the amountof signal originating from the seed gas. However, atthis position of the beam profile the column density ofmolecules is rather low. In this note, we present thecombination of a knife edge with the electrostatic deflector,which allows for a better separation of the different speciesof a molecular beam as well as an increase in columndensity in the interaction region.A schematic of the experimental setup is shown inFIG. 1. A pulsed molecular beam was provided by expand-ing a few millibar of indole and a trace of water in 80 bar ofhelium through a position-adjustable Even-Lavie valve .The valve was operated at a temperature of ◦ C andat a repetition rate of 250 Hz. Two transversely, in X – Y ,adjustable conical skimmers (Beam Dynamics, model 50.8with ∅ = 3 . , model 40.5 with ∅ = 1 . mm) were placed a) b) These authors contributed equally.
DeflectedUndeflectedWithoutknife edge Withknife edge Knife edgeLaserbeama bEL-valve Skimmers DeflectorKnife edge TOF-spectrometerImagingdetectorMotorLaser beamMolecular beam
X Y Z Y X
MovableKnife edgeMolecular beam Skimmed part p pr BA c b Motor MovableKnife edgeMolecular beam
FIG. 1. (Color online) Schematic of the experimental setupand the definition of the coordinate system. a) Sketch of thecross sections of the molecular beam and the laser beam toillustrate the working principle of the knife edge. b) Zoominto the knife edge region, showing the mechanical setup andmotorization. c) Definition of the volumes A and B , the beamradius r , and the width p used for the theoretical limit; seetext for details. , the molecularbeam was dispersed according to the specific quantumstates of the molecular species . The vertically, Y ,adjustable knife edge was placed 1.3 cm behind the endof the deflector.For the measurements with knife edge its vertical po-sition was chosen such that the undeflected molecularbeam was cut roughly in its center. For the measure-ments without knife edge it was moved vertically out ofthe molecular beam. A third, transversely adjustableskimmer (Beam Dynamics, model 50.8 with ∅ = 1 . mm)was placed 2.5 cm downstream of the front of the knifeedge. The molecular beam entered a time of flight massspectrometer (TOF-MS) centered 17.6 cm downstream ofthe last skimmer, where the molecules and clusters werestrong-field ionized by a laser pulse with a pulse durationof 30 fs, centered at a wavelength of 800 nm, and focusedto ∅ ≈ µm. FIG. 1a shows a cross section, in the X – Y a r X i v : . [ phy s i c s . a t m - c l u s ] S e p plane, of the molecular beam to schematically illustratethe working principle of the knife edge. On the left, amolecular beam profile defined by the shape of a roundskimmer is depicted. Its deflected part is shown by avertical shift. On the right, the corresponding profilesare depicted for the case with the knife edge. The laserprobes the molecules in the deflected part of the beam,resulting in a higher column density compared to the casewithout knife edge. FIG. 1b highlights the region of thesetup where the knife edge was located. It depicts theknife edge with its holder which was mounted on a motor(SmarAct SLC-1750-S-UHV) which allows to position theknife edge vertically. The molecular beam is indicated bythe green cylinder which is cut into halves by the knifeedge.We used the separation of indole and indole-water clus-ters to demonstrate the advantage of using the knife edgein combination with the electrostatic deflector. FIG. 2ashows the measured vertical density profiles of the un-deflected and deflected molecular beam when the knifeedge was used. The TOF mass spectrum was gated onspecific masses, which corresponded to either parent ionsor specific fragments, to obtain each individual profile.The undeflected (0 kV) profile of the signal correspondingto the indole mass of m = 117 u is shown in dark blue. Allmolecules and clusters were deflected downwards whenvoltages of ± kV were applied to the deflector electrodes,as all quantum states were high-field seeking at the elec-tric field strengths experienced inside the deflector . Thedeflection profiles for the gates set to the masses of indole,indole ( H O ) , indole ( H O ) and ( indole ) are shown inred, black, green, and orange, respectively. The profilesfor indole ( H O ) , indole ( H O ) , and ( indole ) were multi-plied by a factor of five. The indole ( H O ) cluster wasnot observed in the mass spectrum. Furthermore, theprofile of ( indole ) ( H O ) had the same shape as the onefor ( indole ) and is not shown in the figure. Several edgeswere observed in the profiles which correspond to variousmolecules and fragments. Going from left to right, theoutermost edge at -1.25 mm is attributed to indole ( H O ) because this cluster showed the largest Stark effect of allmolecules and clusters to be considered and was, therefore,deflected the most . The shape of this edge matches thecorresponding edge in the indole-ion profile, which con-firms that the indole ( H O ) ion was fragmenting to indoleion with a probability of ∼ %. The edge at -0.9 mmin the indole-cation signal was attributed to the indolemonomer, since indole had the second largest Stark effect.The edge on top of the indole ( H O ) profile at -0.6 mm wasproduced by indole ( H O ) clusters which fragmented intoindole ( H O ) with a probability of ∼ %. A better sepa-ration of indole ( H O ) from indole and higher clusters wasobserved in comparison to our previous experiments onthis system without the knife edge . Furthermore, theedge for the indole ( H O ) cluster has now been observedfor the first time.FIG. 2b shows the measured deflection profiles for in-dole corrected by the known fragmentation probabilitiesto account for the fragmentation for the case with andwithout knife edge (Knife) and the deflector switched on(red) and off (blue). The profiles for the case withoutknife edge were shifted by 0.975 mm to the left to match Y (mm)00.51 I o n s ( a r b . un i t ) a Ind (0 kV)IndInd-(H O) x5Ind-(H O) x5(Ind) x5 -2 -1 0 1 2 Y (mm)00.51 I o n s ( a r b . un i t ) b Ind (0 kV) K-EInd K-EInd (0 kV)Ind
FIG. 2. (Color online) a) Column density profile withknife edge of indole (dark blue), and deflection curves ofindole (red open circles), indole ( H O ) (black open triangles),indole ( H O ) (green open triangles), and ( indole ) (orangeopen squares). b) Column density profiles with deflectorswitched off without knife edge (blue triangles), deflectorswitched off with knife edge (dark blue circles), deflectorswitched on without knife edge (light red open triangles),and deflector switched on with knife edge (red open circles). the edges on the left side for a better direct comparison.In both cases – deflector on and off – the left edge wassteeper for the measurements with knife edge. This isattributed to the higher column density as a result of theknife edge. Placing the probe laser at − . mm in the de-flected profile results in an enhancement factor of R = 1 . at this position. The measured molecular beam diam-eter of r = 2 mm matches exactly the expected radiusfrom geometry arguments assuming a point source for themolecular beam. The distances between the valve and thethird skimmer and the interaction region are 53.4 cm and71 cm, respectively. This results in a magnification factorof . / . . , in excellent agreement with the ratiobetween the measured molecular-beam diameter and theskimmer diameter given by . / . . . The deflectedpart of the molecular beam is, therefore, also expectedto be far out of the geometric helium profile. Additionalbroadening mechanisms for the molecular beam, such asthe finite temperature or deviations from a point sourceare not taken into account. The influence of these contri-butions to the purity of the molecular beam are beyondthe scope of this manuscript.The maximum enhancement factor R for the increasein column density can be estimated, assuming an uniformmolecular beam emitted from a point source and a uni-form deflection force, from the molecular beam radius r inthe interaction region and the width of the, for the interac-tion with the molecules relevant, volume p . For p (cid:28) r theenhancement factor is given by R = A/B = 3 / (cid:112) r/p ,see FIG. 1c. Taking the radius of our measured molec-ular beam profile of r = 1 . mm and the diameter ofthe ionization laser p = 50 µm resulted in an expectedenhancement factor of R ≈ . . We attribute the reducedexperimentally observed enhancement factor of R = 1 . tothe following reasons: The experimental molecular beamprofile was not completely collimated and, therefore, theedges of the profiles are not infinitely steep. This is as-cribed to the finite size of the virtual source, which weestimate to be in the order of 0.6 mm. Due to the finitesource size and the geometrical constraints given by theskimmers we furthermore expect it be be advantageousto place the knife edge behind the deflector, comparedto using, e. g., a slit-skimmer before the deflector since itdecreases the effective virtual source. Secondly, the impor-tant volume for the interaction of the molecules with theionization laser was unknown and might be broader thanthe measured diameter in intensity. A third contributionto the reduced enhancement is attributed to the fact thatthe deflector acts as a thick lens for the dispersion of themolecular beam which leads to a softening of the edges.A further contribution could be a misalignment of theknife edge with respect to the propagation direction ofthe probe laser.The combination of the knife edge with the electrostaticdeflector is of general use for all molecular beam experi-ments that benefit from a strong separation of molecularspecies or a strong separation from the seed gas. The pre-sented approach is also especially useful for applicationswith low count rates or restricted measurement times,e. g., beamtimes at large facilities such as free electronlasers (FELs), synchrotrons, or high-power-laser facilities,where typically only a few days of beamtime are availablefor the measurements. Furthermore, for probing a colli-mated molecular beam with r = 2 mm by the generallysmall x-ray beams, e. g., p = 5 µm, a theoretical enhance-ment factor, according to the model described above, of R > is obtained. Taking into account the finite virtualsource size and the resulting measured reduction of theenhancement factor by about / . leads to an expectedenhancement factor of 6 in line with preliminary resultsfrom a recent beamtime at the LCLS. ACKNOWLEDGMENTS
We acknowledge Benjamin Erk and the CAMP teamfor a significant equipment loan.Besides DESY, this work has been supported by theexcellence cluster “The Hamburg Center for UltrafastImaging – Structure, Dynamics and Control of Matter atthe Atomic Scale” (CUI, DFG-EXC1074), the EuropeanResearch Council under the European Union’s SeventhFramework Programme (FP7/2007-2013) through theConsolidator Grant COMOTION (ERC-Küpper-614507),by the European Union’s Horizon 2020 research and in-novation program under the Marie Skłodowska-CurieGrant Agreement 641789 “Molecular Electron Dynam-ics investigated by Intense Fields and Attosecond Pulses”(MEDEA), and by the Helmholtz Association through theVirtual Institute 419 “Dynamic Pathways in Multidimen-sional Landscapes” and the “Initiative and NetworkingFund”. J.O. gratefully acknowledges a fellowship by theAlexander von Humboldt Foundation. G. Scoles, ed.,
Atomic and molecular beam methods , Vol. 1 & 2(Oxford University Press, New York, NY, USA, 1988 & 1992). W. R. Gentry and C. F. Giese, “High-precision skimmers forsupersonic molecular beams,” Rev. Sci. Instrum. , 104 (1975). W. Demtröder,
Laser Spectroscopy 1: Basic Principles (SpringerVerlag, Berlin, 2008). R. Subramanian and M. Sulkes, “Production of a slit skimmer foruse in cold supersonic molecular beams,” Rev. Sci. Instrum. ,016101 (2008). K. Hornberger, S. Gerlich, P. Haslinger, S. Nimmrichter, andM. Arndt, “Colloquium: Quantum interference of clusters andmolecules,” Rev. Mod. Phys. , 157 (2012). S. Y. T. van de Meerakker, H. L. Bethlem, N. Vanhaecke, andG. Meijer, “Manipulation and control of molecular beams,” Chem.Rev. , 4828 (2012). Y.-P. Chang, D. A. Horke, S. Trippel, and J. Küpper, “Spatially-controlled complex molecules and their applications,” Int. Rev.Phys. Chem. , 557 (2015), arXiv:1505.05632 [physics]. N. F. Ramsey,
Molecular Beams , The International Series ofMonographs on Physics (Oxford University Press, London, GB,1956) reprinted in
Oxford Classic Texts in the Physical Sciences (2005). F. Filsinger, J. Küpper, G. Meijer, L. Holmegaard, J. H. Nielsen,I. Nevo, J. L. Hansen, and H. Stapelfeldt, “Quantum-state selec-tion, alignment, and orientation of large molecules using staticelectric and laser fields,” J. Chem. Phys. , 064309 (2009),arXiv:0903.5413 [physics]. U. Even, J. Jortner, D. Noy, N. Lavie, and N. Cossart-Magos,“Cooling of large molecules below 1 K and He clusters formation,”J. Chem. Phys. , 8068 (2000). J. S. Kienitz, K. Długołęcki, S. Trippel, and J. Küpper, “Improvedspatial separation of neutral molecules,” J. Chem. Phys. ,024304 (2017), https://doi.org/10.1063/1.4991479. S. Trippel, Y.-P. Chang, S. Stern, T. Mullins, L. Holmegaard,and J. Küpper, “Spatial separation of state- and size-selectedneutral clusters,” Phys. Rev. A86