aa r X i v : . [ phy s i c s . a t o m - ph ] A ug Bright, continuous beams of cold free radicals
J. C. Shaw and D. J. McCarron ∗ Department of Physics, University of Connecticut,196 Auditorium Road, Unit 3046, Storrs, Connecticut 06269, USA.
We demonstrate a cryogenic buffer gas-cooled molecular beam source capable of producing bright,continuous beams of cold and slow free radicals via laser ablation over durations of up to 60 seconds.The source design uses a closed liquid helium reservoir as a large thermal mass to minimize heatingand ensure reproducible beam properties during operation. Under typical conditions, the sourceproduces beams of our test species SrF, containing 5 × molecules per steradian per second inthe X Σ( v = 0 , N = 1) state with a rotational temperature of 1.0(2) K and a forward velocity of140 m/s. The beam properties are robust and unchanged for multiple cell geometries but dependcritically on the helium buffer gas flow rate, which must be ≥
10 standard cubic centimeters perminute to produce bright, continuous beams of molecules for an ablation repetition rate of 55 Hz.
Beams of cold and slow molecules from cryogenicbuffer gas sources have played a central role in recentimproved precision measurements [1, 2], high-resolutionspectroscopy [3, 4] and the direct laser cooling and trap-ping of molecules at ultracold temperatures [5, 6]. Di-rect cooling methods for molecules have the potentialto produce a chemically diverse range of diatomic andpolyatomic species at ultracold temperatures which arewell-suited for proposed applications including tests offundamental physics [7], and controlled chemistry [8].Cryogenic buffer gas beam sources rely on a flow ofcold inert gas, usually helium or neon, to sympatheti-cally cool the molecular species of interest and collapsethe occupied rovibrational state distribution [9–12]. Thisthermalization takes place inside an enclosed cell as thespecies of interest becomes entrained within the inert gasflow and exits the cell through a small hole to form abeam. These molecular beams have forward velocitiesbetween ∼
50 and 200 m/s and advances in slowingtechniques using radiation pressure [13, 14] have enabledmolecules below ∼
10 m/s to be captured and cooled bymagneto-optical traps (MOTs) [15]. Today’s molecularMOTs provide confining and damping forces comparableto those in atomic MOTs but can only capture 10 − molecules at densities up to ∼ cm − due to the lowtrappable flux and short loading times ( ∼
20 ms) attain-able when loading single pulses of molecules. While thefirst interactions between laser-cooled molecules were re-cently observed [16], many proposed applications requirelarger trapped samples at higher density and increasingthe trappable flux remains a key challenge. More efficientslowing techniques are currently being pursued by multi-ple groups [17–21], and the production of brighter, slowermolecular beams remains an active and complementaryarea of research [22–26].This Rapid Communication presents a cryogenic ∗ [email protected] source capable of producing bright, continuous beamsof cold free radicals via laser ablation, thereby realiz-ing the first step towards longer MOT loading times andthe continuous accumulation of conservatively trappeddark-state molecules using an intermediate MOT stage.Our source uses helium buffer gas at 2 . ∼ ∼
16 J/K) toboth dampen temperature oscillations from the refriger-ator and allow the source to absorb high thermal loadsfrom the ablation laser with limited heating. A constantsource temperature is desirable to ensure reproduciblemolecular beam properties including flux, forward veloc-ity and rotational state distribution. We note that anequivalent thermal mass using copper alone at 2.6 K isimpractical and would require cooling ∼
400 kg of ma-terial. However, rare-earth alloy plates have been suc-cessfully used to dampen thermal oscillations in a sim-ilar manner [28]. While beneficial once cold, the closedhelium reservoir leads to longer source cool-down andwarm-up times. To counter this increase, our designlimits the additional thermal mass to 6 kg of machinedaluminum and copper parts while largely replicating thesource geometry of ref. [10] (fig. 1b). Our design coolsfrom 295 K to 2.5 K in ∼ ∼ . FIG. 1. Overview of the cryogenic source. a Photo of thesource from behind with the rear radiation shields removed toshow (i) the refrigerator second-stage, (ii) liquid helium reser-voir and (iii) cell. b Schematic cross-section of the source fromabove showing the ablation and absorption beam paths. c Typical source cool-down and warm-up curves measured overseveral hours for the refrigerator first-stage (black), second-stage (red) and cell (blue). d Short-term temperature stabil-ity for the same three regions as c . Temperature oscillationsat the 1.4 Hz period of the pulse-tube refrigerator are visibleat all three regions. and refrigerator second-stage are stable to ± ±
60 mK respectively (fig. 1d). For reference, without thehelium reservoir the second-stage temperature stability istypically ±
200 mK as the refrigerator pulses [29]. Theselarger oscillations have been reported to correlate with a ∼
25 % peak to peak variation in molecular beam flux[10], forcing several experiments to synchronize their rep-etition rates to the period of the pulse-tube refrigeratorto recover reproducible pulses of molecules [3, 13].This work uses SrF molecules to characterize thesource performance by ablating a SrF target, mountedinside the cell at 30 ◦ relative to the molecular beam axis,using 15 mJ pulses of 532 nm light with 6 ns durationfrom a Nd:YAG laser. This light is tightly focused ontothe surface of the target using a 200 mm focal length lensoutside the vacuum chamber and the pulse energy is sta-ble to within 1 %. Cold helium buffer gas enters the cellthrough a fill line at the rear and exits through a conicalface with a 40 ◦ half-angle and a 3 mm diameter aperture(fig. 1b). The typical helium buffer gas flow rate is 15standard cubic centimeters per minute (sccm), equivalentto an in-cell steady-state helium density of 10 cm − anda Reynolds number of ≈
60. At this flow rate the vac-uum inside the cryogenic source chamber is 10 − Torr,maintained by ∼
700 cm of cold charcoal cryopump. Properties of the molecular beam are typically probedon the X Σ( v = 0 , N = 1) to A Π / ( v ′ = 0 , J ′ = 1 / ≈
140 m/s with a FWHM of ≈
50 m/s. This was measured through the Doppler shiftbetween two fluorescence profiles recorded using probelasers transverse and counter-propagating to the molec-ular beam. The measured FWHM transverse velocityspread is 80 m/s, corresponding to a FWHM angularspread of 30 ◦ . The rotational temperature of the molec-ular beam is 1 . Σ( v = 0 , N = 0 −
4) using flu-orescence signals from the X Σ( v = 0 , N = 0 −
4) toA Π / ( v ′ = 0 , J ′ = 1 / − /
2) transitions and calculatedbranching ratios to account for the varying line strengths[30]. Here molecules cool rotationally to below the celltemperature due to isentropic cooling near the cell aper-ture [31] and at our rotational temperature ≈
50 % ofthe molecules populate the X Σ( v = 0 , N = 1) state.All of these parameters are in good agreement with mea-surements performed on a source with similar geometry[10].To extract the number of molecules exiting the sourcein the X Σ( v = 0 , N = 1) state we use the time-integralof the resonant absorption signal, Doppler broadened ab-sorption cross section [32], and assume a uniform densityover the cross sectional area of the molecular beam [10].At ablation repetition rates of 1-2 Hz, where other heliumbuffer gas sources typically operate [10, 33, 34], the sourceproduces 10 molecules per steradian per pulse with neg-ligible heating. We note that all reported numbers canvary by ≈ ±
50 % depending on the spot ablated on the
FIG. 2. Absorption and cell temperature traces measuredover 1.5 seconds of source operation for ablation rates of (a)10 Hz (blue), (b) 20 Hz (red) (c) and 55 Hz (black). Duringthese measurements the source temperature (d) increased by40, 80 and 200 mK respectively. The time needed to coolback to 2.64 K was 1 s, 30 s and 70 s for 10, 20 and 55 Hzoperation respectively. target. At ablation repetition rates of 10 and 20 Hz thepulses of molecules are unchanged and the source pro-duces 10 and 2 × time-averaged molecules/sr/secrespectively (figs. 2a and 2b). At 55 Hz we consistentlyobserve a ∼
10 % decrease in brightness ( ∝ the time-integrated absorption signal) over the first 2-5 pulses,with negligible change in rotational temperature, andtypically produce 5 × molecules/sr/sec (fig. 2c). In-cell absorption measurements show that this initial de-crease in brightness is correlated with decreasing in-cellmolecular density and the extraction efficiency from thecell remains unchanged at ∼
50 %. This decrease in yieldis temporary and a 100 ms pause in ablation pulses is suf-ficient to recover the original yield from the next pulseusing the same ablation spot. This behavior is presum-ably due to heating within the cell, which is measured toincrease by 40, 80 and 200 mK over 1.5 seconds of opera-tion at 10, 20 and 55 Hz respectively (fig. 2d). Once thecell reaches thermal equilibrium, we typically measure atemperature increase of ≈ FIG. 3. Absorption versus helium buffer gas flow at a 55 Hzablation repetition rate. From top to bottom, the helium flowrates were 2, 5, 10, 15 and 20 sccm. At 2 and 5 sccm some ab-lation pulses produce ∼ µs bursts of molecules, these arenot detected at higher flow rates. Molecules are continuouslydetected leaving the source for flow rates ≥
10 sccm. Thesedata were recorded in a random order using the same ablationspot on the target and highlight the temporary nature of thedecrease in brightness measured over the first 2-5 pulses. are detected in the beam after 250 ms of operation whilea flow of 5 sccm is sufficient to consistently produce pulsesof molecules and realize 10 molecules/sr/sec. Note thatthe sporadic spikes in fig. 3 visible at flows of 2 and5 sccm are ∼ µs pulses of 10 − molecules and arenot optical pickup due to the ablation laser. At flows of10 sccm we begin to continuously detect molecules exit-ing the cell with a brightness of 3 × molecules/sr/sec,increasing to 7 × molecules/sr/sec at 20 sccm. At15 sccm the beam brightness is typically modulated at55 Hz by ∼
80 % 20 mm downstream of the cell. Weproject that this modulation will decrease to ∼
60 % afurther 1.5 m downstream by convolving the measuredmolecular pulse temporal and velocity distributions.The source performance is robust at 55 Hz for he-lium buffer gas flow rates ≥
10 sccm while a flow rateof ≥ > . Σ( v = 0 , N = 1) state by ≈
30 %. This factor of 2 increase in rotational temper-ature highlights the importance of limiting heating dur-ing source operation to ensure reproducible beam prop-erties. This source design has also proven to be straight-forward to replicate and a second unit is now operationalin our group and produces similar continuous beams ofmolecules.As a demonstration of the stable and continuous na-ture of these molecular beams we produce uninterruptedpulses of SrF molecules at 55 Hz over a 60 second dura-tion using a buffer gas flow rate of 15 sccm (fig. 4). Dur-ing this time the ablation spot was moved every ∼
10 sto restore the decaying ablation yield and we measure amean brightness of ∼ × molecules/sr/sec along-side a cell temperature increase of 0.6 K. Currently themain limit on beam brightness during continuous opera-tion of our source is the durability of the ablated target.Free radical production methods that ablate metals inthe presence a reactant gas (e.g. SF ) have been shownto produce brighter, more reproducible beams [34], andcould potentially work well with our source design at high FIG. 4. A continuous beam of SrF molecules produced over60 s showing (a) absorption and (b) the cell temperature.The mean brightness measured during this minute was ∼ × molecules/sr/sec and afterwards the cell required ∼ ablation repetition rates. Assuming a durable target isused, the absolute limit on operation time is set by satu-ration of the charcoal cryopump. In our current design, acontinuous flow of 15 sccm of helium can be maintainedfor 10 hours before saturation and this is readily extendedby increasing the cryopump surface area.In general, beams of molecules from helium buffergas sources are preferable over neon-based beams formolecular laser cooling and trapping experiments dueto their lower forward velocities, reduced divergence andcolder rotational temperatures [11, 37]. Our source de-sign combines these advantages alongside the ability toabsorb high thermal loads with limited heating, similarto neon-based sources. We note that while the largethermal mass in our design restricts heating, the max-imum input power is determined by the cooling power ofthe pulse-tube refrigerator. In our setup the maximumload is 2 W, permitting ablation repetition rates beyond100 Hz and possible access to beams with more than10 molecules/sr/sec, similar to the brightest beams offree radicals from neon-based sources [12]. At present,liquid helium fills ∼
25 % of the closed reservoir andincreasing the helium mass would further improve thesource temperature stability at the expense of longercool-down and warm-up times. In principle, one couldalso pump on this reservoir to cool the source towards1 K and produce colder, slower molecular beams, butwith substantially reduced cooling power [37, 38].In summary, we have realized a robust cryogenic buffergas source capable of producing continuous beams ofcold free radicals via laser ablation with a time-averagedbrightness of up to 7 × molecules/sr/sec in a sin-gle rovibrational state. Crucially, the performance ofour source does not depend critically on the cell geom-etry or temperature. This suggests that other groupscurrently using helium-based sources could immediatelyadopt our method to realize brighter molecular beams,provided that there is sufficient buffer gas flow and oper-ation times are short to limit heating. These molecularbeams represent the first step towards longer MOT load-ing times to trap larger samples at higher density andare well-suited for continuous beam slowing and cooling techniques such as centrifugal deceleration [39, 40], Zee-man slowing [21], and ZeemanSisyphus deceleration [18].For reference, today’s molecular MOTs load single pulsesof molecules over ∼
20 ms. If similar loading times wereused for atomic MOTs only ∼ − atoms wouldbe trapped [41, 42]. Given that typical molecular MOTlifetimes are short, ∼
100 ms [5], the continuous accumu-lation of conservatively trapped molecules via an inter-mediate MOT stage is a particularly promising approach,a method which has been successfully demonstrated withchromium atoms [43]. These advances have the potentialto increase the numbers and densities realized in laser-cooled samples of molecules by orders-of-magnitude andprovide routine access to the molecule-molecule interac-tions required for many proposed applications.Drawings of the machined parts needed to replicatethis cryogenic source design and assembly instructionscan be provided by contacting the corresponding author.We thank D. DeMille and M. Steinecker for helpfuldiscussions and J. Schnaubelt for carefully reading themanuscript. We acknowledge financial support from theNSF (CAREER Award [1] J. Baron, W. C. Campbell, D. DeMille, J. M. Doyle,G. Gabrielse, Y. V. Gurevich, P. W. Hess, N. R. Hut-zler, E. Kirilov, I. Kozyryev, B. R. O’Leary, C. D.Panda, M. F. Parsons, E. S. Petrik, B. Spaun, A. C.Vutha, and A. D. West (The ACME Collaboration),Science , 269 (2014).[2] V. Andreev, D. G. Ang, D. DeMille, J. M. Doyle,G. Gabrielse, J. Haefner, N. R. Hutzler, Z. Lasner,C. Meisenhelder, B. R. OLeary, C. D. Panda, A. D. West,E. P. West, and X. Wu (The ACME Collaboration), Na-ture , 355 (2018).[3] E. B. Norrgard, E. R. Edwards, D. J. McCarron, M. H.Steinecker, D. DeMille, S. S. Alam, S. K. Peck, N. S. Wa-dia, and L. R. Hunter, Phys. Rev. A , 062506 (2017).[4] S. Truppe, S. Marx, S. Kray, M. Doppelbauer, S. Hofs¨ass,H. C. Schewe, N. Walter, J. P´erez-R´ıos, B. G. Sartakov,and G. Meijer, Phys. Rev. A , 052513 (2019).[5] D. J. McCarron, J. Phys. B: At. Mol. Opt. Phys. ,212001 (2018).[6] M. R. Tarbutt, Contemporary Physics , 356 (2018).[7] I. Kozyryev and N. R. Hutzler,Phys. Rev. Lett. , 133002 (2017).[8] N. Wells and I. C. Lane,Physical Chemistry Chemical Physics , 19036 (2011).[9] S. E. Maxwell, N. Brahms, R. deCarvalho, D. R.Glenn, J. S. Helton, S. V. Nguyen, D. Patter-son, J. Petricka, D. DeMille, and J. M. Doyle,Phys. Rev. Lett. , 173201 (2005).[10] J. F. Barry, E. S. Shuman, and D. DeMille,Phys. Chem. Chem. Phys. , 18936 (2011).[11] N. R. Hutzler, M. F. Parsons, Y. V. Gurevich, P. W. Hess, E. Petrik, B. Spaun, A. C. Vutha,D. DeMille, G. Gabrielse, and J. M. Doyle,Phys. Chem. Chem. Phys. , 18976 (2011).[12] N. R. Hutzler, H.-I. Lu, and J. M. Doyle,Chemical Reviews , 4803 (2012).[13] J. F. Barry, E. S. Shuman, E. B. Norrgard, and D. De-Mille, Phys. Rev. Lett. (2012).[14] S. Truppe, H. J. Williams, N. J. Fitch, M. Hambach,T. E. Wall, E. A. Hinds, B. E. Sauer, and M. R. Tarbutt,New Journal of Physics , 022001 (2017).[15] J. F. Barry, D. J. McCarron, E. N. Norrgard, M. H. Stei-necker, and D. DeMille, Nature , 286 (2014).[16] L. Anderegg, L. W. Cheuk, Y. Bao, S. Burch-esky, W. Ketterle, K.-K. Ni, and J. M. Doyle,Science , 1156 (2019).[17] D. DeMille, J. F. Barry, E. R. Edwards, E. B. Norrgard,and M. H. Steinecker, Mol. Phys. , 1805 (2013).[18] N. J. Fitch and M. R. Tarbutt,ChemPhysChem , 3609 (2016).[19] L. Aldridge, S. E. Galica, and E. E. Eyler,Phys. Rev. A , 013419 (2016).[20] X. Yang, C. Li, Y. Yin, S. Xu, X. Li, Y. Xia, and J. Yin,Journal of Physics B: Atomic, Molecular and OpticalPhysics , 015001 (2017).[21] M. Petzold, P. Kaebert, P. Gersema,M. Siercke, and S. Ospelkaus,New Journal of Physics , 042001 (2018).[22] E. P. West, A Thermochemical Cryogenic Buffer GasBeam Source of ThO for Measuring the Electric DipoleMoment of the Electron , Ph.D. thesis, Harvard University(2017).[23] V. Singh, A. K. Samanta, N. Roth, D. Gusa, T. Os-senbr¨uggen, I. Rubinsky, D. A. Horke, and J. K¨upper,Phys. Rev. A , 032704 (2018).[24] D. Xiao, D. M. Lancaster, C. H. Allen, M. J. Taylor, T. A.Lancaster, G. Shaw, N. R. Hutzler, and J. D. Weinstein,Phys. Rev. A , 013603 (2019).[25] A. Jadbabaie, N. H. Pilgram, J. Kos, S. Kotochigova,and N. R. Hutzler, New J. Phys. , 022002 (2019).[26] T. Gantner, M. Koller, X. Wu, G. Rempe, and M. Zeppenfeld,J. Phys. B: At. Mol. Opt. Phys. , 145302 (2020).[27] Temperature oscillation damping pot, Cryomech Inc..[28] K. Allweins, L. M. Qiu, and G. Thummes,AIP Conference Proceedings , 109 (2008).[29] Cryomech Inc., Private Communication.[30] B. E. Sauer, J. Wang, and E. A. Hinds,J. Chem. Phys. , 7412 (1996).[31] H. Pauly, Atom, Molecule, and Cluster Beams (Springer,Berlin, 2000).[32] D. Budker, D. F. Kimball, and D. P. DeMille,
AtomicPhysics: An Exploration Through Problems and Solu-tions, 2nd. Ed. (Oxford Univ. Press, 2008).[33] G. Z. Iwata, R. L. McNally, and T. Zelevinsky,Phys. Rev. A , 022509 (2017).[34] S. Truppe, M. Hambach, S. M. Skoff, N. E. Bulleid, J. S.Bumby, R. J. Hendricks, E. A. Hinds, B. E. Sauer, andM. R. Tarbutt, J. Mod. Opt. , 648 (2017).[35] L. Anderegg, B. L. Augenbraun, E. Chae,B. Hemmerling, N. R. Hutzler, A. Ravi, A. Col-lopy, J. Ye, W. Ketterle, and J. M. Doyle,Phys. Rev. Lett. , 103201 (2017).[36] A. L. Collopy, S. Ding, Y. Wu, I. A. Finneran, L. An-deregg, B. L. Augenbraun, J. M. Doyle, and J. Ye,Phys. Rev. Lett. , 213201 (2018).[37] D. Patterson, J. Rasmussen, and J. M. Doyle, New Jour-nal of Physics , 055018 (2009).[38] C. Wang and B. Lichtenwalter,AIP Conference Proceedings , 1387 (2014).[39] S. Chervenkov, X. Wu, J. Bayerl, A. Rohlfes,T. Gantner, M. Zeppenfeld, and G. Rempe,Phys. Rev. Lett. , 013001 (2014).[40] X. Wu, T. Gantner, M. Koller, M. Zeppenfeld, S. Cher-venkov, and G. Rempe, Science , 645 (2017).[41] X. Xu, T. H. Loftus, J. L. Hall, A. Gallagher, and J. Ye,J. Opt. Soc. Am. B , 968 (2003).[42] M. L. Harris, P. Tierney, and S. L. Cornish,J. Phys. B: At. Mol. Opt. Phys. , 035303 (2008).[43] P. O. Schmidt, S. Hensler, J. Werner,T. Binhammer, A. G¨orlitz, and T. Pfau,J Opt. B: Quantum Semiclass. Opt.5