Trapping of ultra cold atoms in a 3He/4He dilution refrigerator
F. Jessen, M. Knufinke, S. C. Bell, P. Vergien, H. Hattermann, P. Weiss, M. Rudolph, M. Reinschmidt, K. Meyer, T. Gaber, D. Cano, A. Guenther, S. Bernon, D. Koelle, R. Kleiner, J. Fortagh
TTrapping of ultra cold atoms in a He/ He dilution refrigerator
F. Jessen, M. Knufinke, S. C. Bell, ∗ P. Vergien, H. Hattermann, P. Weiss, M. Rudolph, M. Reinschmidt,K. Meyer, T. Gaber, D. Cano, A. G¨unther, S. Bernon, † D. Koelle, R. Kleiner, and J. Fort´agh ‡ CQ Center for Collective Quantum Phenomena and their Applications in LISA + ,Physikalisches Institut, Eberhard-Karls-Universit¨at T¨ubingen,Auf der Morgenstelle 14, D-72076 T¨ubingen, Germany We describe the preparation of ultra cold atomic clouds in a dilution refrigerator. The closedcycle He/ He cryostat was custom made to provide optical access for laser cooling, optical manip-ulation and detection of atoms. We show that the cryostat meets the requirements for cold atomexperiments, specifically in terms of operating a magneto-optical trap, magnetic traps and magnetictransport under ultra high vacuum conditions. The presented system is a step towards the creationof a quantum hybrid system combining ultra cold atoms and solid state quantum devices.
PACS numbers: 37.10.Gh 07.20.-N
The development of cold atom/solid state hybrid sys-tems holds the promise of creating a quantum interfacebetween printed electronic circuits, atoms and light [1–10] with applications in quantum electronics and infor-mation processing. Several groups are currently prepar-ing cold atomic clouds in the vicinity of superconduct-ing chips at 77 K, cooled by liquid nitrogen, and at 4 K,cooled by liquid He [11–23]. The vision of quantum statetransfer between superconducting circuits and cold atomsrequires further experimental development, in particularthe preparation of atomic clouds close to millikelvin sur-faces. This low temperature is required to operate super-conducting quantum circuits and also enhances the co-herence time of solid state quantum bits, which must belong enough to realize quantum state transfer to atomicdegrees of freedom.The conditions for cold atom preparation and the op-eration of mK environments are, however, very differ-ent. The first requires several tens of milliwatts of laserpower for laser cooling [24]. The second is highly sen-sitive to heat sources such as laser radiation, since thecooling power of dilution refrigerators is typically lessthan a milliwatt. Here, we describe an experimentalsetup that fulfills the requirements for both the produc-tion of ultra cold atoms and operation of a mK envi-ronment. We trap rubidium atoms in a 6 K environmentinside a He/ He dilution refrigerator capable of mK tem-peratures. We demonstrate the operation of a magneto-optical trap loaded by a beam of slow atoms producedwith a Zeeman slower. The MOT coils and end section ofthe Zeeman slower are constructed with superconductingelectromagnets mounted on an additional 6 K plate of thecryostat. We transfer the atoms into a magnetic trap anddemonstrate the first step in a magnetic transfer schemeto bring the atoms towards the mK environment. ∗ [email protected] † Present adress: Quantronics group, SPEC (CNRS URA 2464),IRAMIS, DSM, CEA-Saclay, 91191 Gif-sur-Yvette, France ‡ [email protected] I. CRYOSTAT AND VACUUM SYSTEM
Cryogenic temperatures are achieved using a closed cy-cle He/ He dilution refrigerator, based on the OxfordInstruments Triton 200 system, shown in figure 1. Thecryostat consists of two pulse tube cooled stages withcooling power of 35 watts at nominal 45 K and 1 watt atnominal 4 K. Due to the heat load on these stages theyoperate at temperatures of 65 K and 6 K. The 6 K stagecools a dilution unit which has three stages operating attemperatures of 1 . . µ m, thus blocking the majority of300 K thermal radiation. The windows are anti-reflectioncoated at 780 nm for high transmission of the cooling andtrapping light. Two optical access ports are also placedbelow the mK stage, for optical diagnosis of the cloud inthe mK environment.The cryostat is mounted inside a large stainless steelvacuum chamber, which allows for the creation of ul-tra high vacuum conditions, as necessary for the coldatom experiments. The vacuum conditions also providethermal insulation between the cryostat cooling stages.The vacuum chamber was initially pumped by a turbo-molecular pump, when the system is cold the cryostatsurfaces serve as a high surface area cryopump and the a r X i v : . [ phy s i c s . a t o m - ph ] S e p additional 6K plate<50 mK100 mK1 K6 K65 K6K optical access 6K connection rodradiationshield (65K)outer vacuumchamber mK optical accessZeemanslower FIG. 1. (Color online) Photograph of the cryostat. Left: System during operation, the cryostat is installed in the stainlesssteel outer vacuum chamber (OVC) and optics for the cold atom experiments surround the chamber. Center: After removal ofthe OVC, the 65 K radiation shield is visible with the optical viewports: at the lower part for cold atom preparation in the 6 Kstage and the upper viewport for cold atom detection in the mK stage. Right: All radiation shields unmounted, showing thedifferent temperature stages of the cryostat, including the additional 6 K plate with the cold atom preparation setup mounted. turbo-molecular pump is isolated from the chamber witha gate valve. A pressure of 10 − mbar is measured with acold cathode gauge in close proximity to the room tem-perature outer vacuum chamber. Due to efficient cryop-umping, the pressure in the cooling and trapping regionis significantly lower, as evident by long magnetic traplifetimes.Thermal anchoring of the superconducting electromag-nets for the cold atom preparation setup was a particu-larly important issue. The superconducting electromag-nets were wound with single filament niobium-titaniumwire (50 µ m diameter, T c =9 . µ m) and insulatedwith a Kapton layer (total diameter 100 µ m). The wireswere thermally anchored to the cryostat by placing theminto channels machined into copper ‘anchor blocks’ andfilled with indium. Despite these thermal anchoring ef-forts the measured critical current of the wires were be-tween 0 . . II. COLD ATOM SETUP
The preparation of cold atoms follows standard cool-ing and trapping techniques [24], with the exception ofthe use of superconducting electromagnets to create therequired magnetic field gradients [11], which have theadvantage of negligible Ohmic heating. An effusive oven,combined with a Zeeman slower [25], creates a beam ofslow atoms which are captured in a magneto-optical trap(MOT) [26] and then transferred into a magnetic trap. Infuture experiments the cold atomic cloud will be trans-ported from this preparation setup in the 6 K environ-ment to the mK environment using a magnetic conveyorbelt, the path of which will be curved to avoid 6 K ther-mal radiation reaching the mK environment. The firststep of magnetic transfer has been realized. A schematicof the cold atom preparation setup is shown in Figure 3.
A. Slow atomic beam
The preparation of a sample of cold atoms begins witha beam of fast atoms effusing out of a hot oven and de-celerated by a Zeeman slower in a zero-crossing configu-ration [27]. The oven is similar to the design presentedin [28]. The rubidium oven is typically operated at areservoir temperature of 80 ◦ C and a collimation tubetemperature of 120 ◦ C, chosen as a compromise between
OVC 300K PT65K 6K 1K DU100mK<50mKadditional 6K plateMCB s l o w a t o m i c bea m laser laserTP FIG. 2. Schematic of the dilution refrigerator indicating thecooling stages and radiation shields. The slow atomic beamand optical access ports through the radiation shields are in-dicated, along with a schematic of the electromagnets for coldatom preparation mounted on the additional 6 K plate. Theconveyor belt to take atoms from the 6 K stage to the mKis also indicated. It follows an intentionally curved path toavoid 6 K thermal radiation reaching the mK environment.OVC: outer vacuum chamber, PT: pulse tube, DU: dilutionunit, TP: turbo-molecular pump and MCB: magnetic con-veyor belt. producing a high atomic flux and preserving the oven life-time. A mechanical shutter at the end of the collimationtube blocks the atomic beam after loading of the MOT.The required magnetic field profile for Zeeman slowingis created by a series of coils, divided into two sections.The coils comprising each section run equal current andthe local magnetic field is given by the number of wind-ings in each coil. One coil section creates the positivemagnetic field (i.e., from the oven until the zero-crossing).This section is at room temperature, constructed with a1 . − . Zeeman slower transfer coilsMOT c on v e y o r be l t t o m illi k e l v i n r eg i on slow atomic beam x z FIG. 3. (Color online) Schematic of the cold atom prepa-ration setup, constructed with superconducting coils on theadditional 6 K plate. The cold section of the Zeeman slower isindicated in yellow, including a compensation coil in orange.The MOT coils are in green and enclosed in the red transfercoils. The first section of the magnetic conveyor belt, whichwill be used to transport atoms to the mK environment isindicated in blue.
Rubidium-87 atoms with an initial velocity below335 m s − are decelerated along the Zeeman slower to afinal velocity of 24 m s − over a distance of 1 .
23 m by aslowing laser beam counter propagating to the atomicbeam. The magnetic field at the beginning of the sloweris 250 G and the exit magnetic field is −
35 G, where pos-itive is defined as the propagation direction of the slow-ing laser. Slowing is achieved using 25 mW of σ + po-larized light (relative to the direction of propagation),tuned 80 MHz below the 5S F =2 → F (cid:48) =3 cy-cling transition. An additional 8 mW of re-pump light,tuned 80 MHz below the 5S F =1 → F (cid:48) =2 tran-sition returns any atoms that are lost from the cyclingtransition. B. Magneto-optical trap
The exit of the Zeeman slower is located at a distanceof 55 mm from the center of the MOT, which was inten-tionally kept short to minimize the diffusion of atoms asthey coast out of the Zeeman slower and into the trap-ping region. A compensation coil is placed between theend of the slower and the MOT to ensure that the mag-netic field of the Zeeman slower is canceled at the centerof the MOT.The magneto-optical trap consists of three orthog-onal pairs of counter-propagating cooling laser beamswith a total power of 65 mW. The Gaussian profile ofthe beams is cut with an aperture at 20 mm diameterto produce a nearly homogeneous beam profile. Thebeams intersect at the center of a pair of superconductingcoils with an inner diameter of 22 mm and separation of24 mm (indicated in green in fig. 3). The coils produce
0 20 40 60 80 100 120 140 a t o m nu m be r loading time (s) FIG. 4. (Color online) A typical curve of the MOT load-ing from the Zeeman slower. Data points were measured byabsorption imaging and the data was fit to an exponentialfunction N ( t ) = N · (1 − exp( − t/τ l )). The fit indicates aloading time of τ l = 19 s. a quadrapole magnetic field with a gradient of approxi-mately 250 G cm − in the z direction, running a currentof 150 mA. The cooling light is detuned 17 MHz belowthe 5S F =2 → F (cid:48) =3 cycling transition in Rb.An additional 8 mW re-pump laser beam, on resonancewith the Rb 5S F =1 → F (cid:48) =2 transition is over-lapped with the cooling beams and re-pumps atoms lostinto the F =1 state back into the cooling cycle.Figure 4 shows the number of atoms in the magneto-optical trap as a function of loading time, measured byabsorption imaging [29]. The red curve in is an exponen-tial loading fit to the data: N ( t ) = N · (1 − exp( − t/τ l )),where N is the steady state atom number, t is time and τ l is the loading constant. The fit indicates a loadingconstant of τ l = 19 s. The MOT loads 5 × atoms in10 seconds and saturates at 1 × atoms after 100 sec-onds. Such a loading rate is considered low for a Zeemanslower system, but is not surprising given that the Zee-man slower is not running at the design specifications dueto current limitation in the superconducting wire.In a typical experimental cycle we load the MOT for10 seconds, trapping 5 × atoms at a temperature of230 µ K. The cooling light is then detuned to 67 MHz andthe magnetic field gradients are simultaneously rampeddown over 12 ms. After a further 5 ms of molasses coolinga 250 µ s optical pumping pulse is applied in a weak ho-mogeneous field to transfer atoms into the magneticallylow field seeking state F =2, m F =2.The magnetic field ramps used during molasses cool-ing and magnetic trapping are intentionally longer than10 ms, exceeding the decay time scale of eddy currents in-duced in the surrounding copper coil supports and cryo-stat components. Future improvements to the systemwill reduce the eddy currents, while maintaining essen-tial high quality thermal contact to the cryostat.
0 20 40 60 80 100 120 a t o m nu m be r holdtime (s)(a) 4.5 5.0 5.5 6.0 6.5 0 20 40 60 80 100 120 F W H M ( mm ) holdtime (s)(b) FIG. 5. (Color online) (a) A curve of the decay of atoms outof the magnetic trap, as measured by absorption imaging.The data was fit to an exponential decay function N ( t ) = N · exp( − t/τ d ), and indicates a magnetic trap lifetime of τ d = 70 s. (b) The FWHM of the cloud in the horizontaldirection over the lifetime of the cloud. The initial increasein the cloud width is consistent with a thermalization time of7 . C. Magnetic trap
We transfer 10 atoms into a magnetic quadrupole trapby increasing the current in the MOT coils to 500 mA,resulting in a field gradient of approximately 30 G cm − in the z direction. The temperature of the ensemble inthe magnetic trap was measured to be 90 µ K by ballis-tic expansion in time-of-flight images [30]. Figure 5 (a)is a logarithmic plot of the atom number as a functionof hold time in the magnetic trap, indicating the traplifetime. The data is well described by an exponentialdecay: N ( t ) = N · exp( − t/τ d ) with a time constant of τ d = 70 s. Plotted in figure 5(b) is the full width halfmaximum (FWHM) of the cloud size in the horizontaldirection over the lifetime of the trap. The initial in-crease in the FWHM of the cloud is consistent with athermalization time scale of 7 .
0 2 4 6 8 10 12 a t o m nu m be r position (mm)atomnumberbackground losses 50 60 70 80 90 100 0 2 4 6 8 10 12 t e m pe r a t u r e ( m K ) position (mm) FIG. 6. (Color online) (a) Atom count during magnetic trans-port of the cloud, as measured by absorption imaging. Thecloud was transported a varied distance from the center of theMOT coils and then back again for imaging. Accounting forbackground losses (red dashed curve), the data indicates thata 70 µ K cloud can be transferred in 1 second with nearly 100%efficiency. (b) The temperature of the cloud during transport,showing no evidence of heating.
D. Magnetic transfer
The atomic cloud is transferred from the magnetic trapcreated by the MOT coils into the first transfer coil pair(indicated in red in fig. 3). This transports the atomiccloud approximately 12 mm in the horizontal direction.Transfer is achieved by increasing the current in thetransfer coils, and simultaneously decreasing the currentin the MOT coils. A transfer time of 1 second was foundto optimize the number of transferred atoms. To assessthe transfer efficiency the cloud was transported a varieddistance from the center of the MOT coils and then backagain for imaging diagnostics (figure 6). The slight atomloss evident in the data can be explained by the magnetictrap lifetime, as indicated in the dashed red curve. Taking these background losses into account, we finda transfer efficiency of nearly 100% for an initial cloudtemperature of 70 µ K. For higher temperatures, the effi-ciency of the transfer is limited by the trap depth whenthe magnetic quadrupoles of the two coil pairs merge.This magnetic transport will play a major role in thenext step in our experiments, in which atoms will betransported from the 6 K cold atom preparation setup tothe mK environment. The full magnetic conveyor beltstructure has been designed and is now being installed.
III. CONCLUSION
We have prepared an ensemble of cold atoms in a closedcycle He/ He cryostat. The magnetic field gradients re-quired for laser cooling and trapping are produced en-tirely with superconducting electromagnets, which carrya current of approximately 0 . IV. ACKNOWLEDGMENTS
This work was supported by the European ResearchCouncil (ERC Advanced Grant SOCATHES) and theDeutsche Forschungsgemeinschaft (SFB TRR21). M.K.and M.R. acknowledge support from the Carl ZeissStiftung. P.V. und H.H. acknowledge support from theEvangelisches Studienwerk e.V. Villigst. [1] M. Wallquist, K. Hammerer, P. Rabl, M. Lukin, andP. Zoller, Physica Scripta , 014001 (2009).[2] Z.-L. Xiang, S. Ashhab, J. Q. You, and F. Nori, Rev.Mod. Phys. , 623 (2013).[3] A. S. Sørensen, C. H. van der Wal, L. I. Childress, andM. D. Lukin, Phys. Rev. Lett. , 063601 (2004).[4] P. Rabl, D. DeMille, J. M. Doyle, M. D. Lukin, R. J.Schoelkopf, and P. Zoller, Phys. Rev. Lett. , 033003(2006).[5] D. Petrosyan and M. Fleischhauer, Phys. Rev. Lett. ,170501 (2008).[6] J. Verd´u, H. Zoubi, C. Koller, J. Majer, H. Ritsch, andJ. Schmiedmayer, Phys. Rev. Lett. , 043603 (2009).[7] D. Petrosyan, G. Bensky, G. Kurizki, I. Mazets, J. Majer,and J. Schmiedmayer, Phys. Rev. A , 040304 (2009). [8] K. R. Patton and U. R. Fischer, Phys. Rev. A , 052303(2013).[9] K. Henschel, J. Majer, J. Schmiedmayer, and H. Ritsch,Phys. Rev. A , 033810 (2010).[10] M. Hafezi, Z. Kim, S. L. Rolston, L. A. Orozco, B. L.Lev, and J. M. Taylor, Phys. Rev. A , 020302 (2012).[11] P. A. Willems and K. G. Libbrecht, Phys. Rev. A ,1403 (1995).[12] T. Nirrengarten, A. Qarry, C. Roux, A. Emmert,G. Nogues, M. Brune, J.-M. Raimond, and S. Haroche,Phys. Rev. Lett , 200405 (2006).[13] T. Mukai, C. Hufnagel, A. Kasper, T. Meno, A. Tsukada,K. Semba, and F. Shimizu, Phys. Rev. Lett , 260407(2007). [14] T. M¨uller, B. Zhang, R. Fermani, K. S. Chan, Z. W.Wang, C. B. Zhang, M. J. Lim, and R. Dumke, New J.Phys. , 043016 (2010).[15] D. Cano, B. Kasch, H. Hattermann, D. Koelle,R. Kleiner, C. Zimmermann, and J. Fort´agh, Phys. Rev.A , 063408 (2008).[16] B. Kasch, H. Hattermann, D. Cano, T. E. Judd,S. Scheel, C. Zimmermann, R. Kleiner, D. Koelle, andJ. Fort´agh, New J. Phys. , 065024 (2010).[17] D. Cano, H. Hattermann, B. Kasch, C. Zimmermann,R. Kleiner, D. Koelle, and J. Fort´agh, Eur. Phys. J. D , 17 (2011).[18] A. Emmert, A. Lupascu, M. Brune, J.-M. Raimond,S. Haroche, and G. Nogues, Phys. Rev. A , 061604(2009).[19] S. Bernon, H. Hattermann, D. Bothner, M. Knufinke,P. Weiss, F. Jessen, D. Cano, M. Kemmler, R. Kleiner,D. Koelle, and J. Fort´agh, Nat. Commun. , 2380 (2013).[20] C. Roux, A. Emmert, A. Lupascu, T. Nirrengarten,G. Nogues, M. Brune, J. M. Raimond, and S. Haroche,Europhys. Lett. , 56004 (2008).[21] F. Shimizu, C. Hufnagel, and T. Mukai, Phys. Rev. Lett. , 253002 (2009).[22] G. Nogues, C. Roux, T. Nirrengarten, A. Lupacu, A. Em-mert, M. Brune, J.-M. Raimond, S. Haroche, B. Plaais, and J.-J. Greffet, EPL (Europhysics Letters) , 13002(2009).[23] B. Zhang, M. Siercke, K. S. Chan, M. Beian, M. J. Lim,and R. Dumke, Phys. Rev. A , 013404 (2012).[24] H. J. Metcalf and P. van der Straten, Laser Cooling andTrapping (Springer, 1999).[25] W. D. Phillips and H. Metcalf, Phys. Rev. Lett. , 596(1982).[26] E. L. Raab, M. Prentiss, A. Cable, S. Chu, and D. E.Pritchard, Phys. Rev. Lett. , 2631 (1987).[27] S. C. Bell, M. Junker, M. Jasperse, L. D. Turner, Y.-J. Lin, I. B. Spielman, and R. E. Scholten, Review ofScientific Instruments , 013105 (2010).[28] Y.-J. Lin, A. R. Perry, R. L. Compton, I. B. Spielman,and J. V. Porto, Phys. Rev. A , 063631 (2009).[29] W. Ketterle, D. S. Durfee, and D. M. Stamper-Kurn,Proc. Int. School of Phys. ”Enrico Fermi” , 67 (1999),cond-mat/9904034.[30] A. Vorozcovs, M. Weel, S. Beattie, S. Cauchi, and A. Ku-marakrishnan, J. Opt. Soc. Am. B , 943 (2005).[31] H.-J. Mundinger, H.-U. H¨afner, M. Mattern-Klosson,H. Klein, and U. Timm, Vacuum , 545 (1992).[32] T. Amthor, J. Denskat, C. Giese, N. N. Bezuglov, A. Ek-ers, L. S. Cederbaum, and M. Weidem¨uller, The Euro-pean Physical Journal D53