On continuous loading of a U-magneto-optical trap (U-MOT) on atom-chip in ultra high vacuum
aa r X i v : . [ phy s i c s . a t o m - ph ] D ec On continuous loading of a U-magneto-optical trap(U-MOT) on atom-chip in ultra high vacuum
Vivek singh
1, 2 , V. B. Tiwari
1, 2 and S. R. Mishra
1, 2 Laser Physics Applications Section, Raja Ramanna Centre for AdvancedTechnology, Indore-452013, India. Homi Bhabha National Institute, Anushaktinagar, Mumbai-400094, IndiaE-mail: [email protected]
December 2019
Abstract.
Here, we report our studies on the continuous loading of a U-magneto-optical trap (U-MOT) on atom-chip from background rubidium (Rb) vapor generatedusing a dispenser source in ultra high vacuum (UHV) environment. Using the U-MOTloading curves, the partial pressure due to Rb vapor and pressure due to backgroundgas have been estimated near the MOT cloud position. The estimated pressure due toRb vapor increased from ∼ . × − Torr to ∼ . × − Torr as Rb-dispensercurrent was increased from 2.8 to 3.4 A. The increase in dispenser current also resultedin decrease in loading as well as lifetime of the MOT cloud. This study is useful formagnetic trapping experiments where accurate information of pressure in chamber isimportant for the lifetime of the magnetic trap.
1. Introduction:
An Atom-chip setup [1, 2, 3, 4, 5] provides a platform to manipulate cold atoms onminiaturized scale to achieve atom trapping [4], guiding [6], beam splitting [7] andBose-Einstein condensation (BEC) [8], etc for practical applications. In addition tothis, an atom-chip offers advantage of tighter magnetic traps with possibility of highthermalization rates which can significantly reduce the time needed for evaporativecooling from minutes to seconds [9]. This makes possible the achievement of BEC atmoderate level of ultra high vacuum ( ∼ × − Torr). Therefore, the atom-chip setupsworld wide are becoming popular with the single MOT loaded in an UHV chamber(pressure ∼ × − Torr) where MOT loading, magnetic trapping and evaporativecooling can be done at the same place. The starting point for these experiments isusually the loading of a MOT for initial cooling of atoms. A popular method to loadMOT on atom-chip is loading an U-magneto-optical trap (U-MOT), which is formed byreflecting MOT laser beams on atom-chip surface and applying quadrupole magneticfield generated by bias coils and a current carrying U-shaped copper wire placed behindthe atom-chip element [10]. Atoms in the MOT are captured either from the background oading studies... ∼ − Torr) and second MOT is loaded in UHV chamber( < − Torr) by transferring cold atoms from the first MOT. An alternative approachis to load MOT in UHV chamber by using thermal atomic beam slowed down by Zee-man slower. However, both these schemes involve some complications. For example, thedouble-MOT setup involves difficulties of differential maintenance of the vacuum as wellas transfer of cold atoms from first MOT to second MOT. The Zeeman slower loadedMOT involves the complications of design and implementation of Zeeman slower device.Thus, preparation of MOT directly in the UHV chamber by temporally controlling theatomic gas vapor pressure makes the UHV MOT loading much simplified [17, 18, 19].A fast rise in partial pressure of atomic vapor can be easily achieved by passing highcurrent pulse in the metal dispensers [19]. However, the resistive heat generated bypassing high current through dispenser needs to be managed properly to achieve fast re-covery of vacuum after switching-off dispenser current [19]. In addition to this, repeatedhigh current pulses also reduce the lifetime of the dispenser. Therefore, the continuousloading of MOT in UHV chamber seems a better option for this purpose, in-spite ofthe continuous loading of MOT in UHV conditions being difficult due to low numberdensity of atomic vapor. Though, a continuous MOT loading in UHV condition hasbeen reported earlier [20] in brief. However, the dependence of pressure, MOT-loadingtime and lifetime on dispenser current in a continuously loaded UHV-MOT needs to beinvestigated.In this letter, we report detailed study on the continuous loading of a U-magneto-optical trap (U-MOT) of Rb atoms from the Rb-atom vapor in the UHV chamber.We have investigated the effect of variation in Rb-dispenser current on number of atomstrapped in the U-MOT in UHV chamber, the rise in pressure in the UHV chamber,MOT loading time and lifetime of the MOT cloud. In addition to this, the Rb vaporpressure at the MOT position has been estimated at a given dispenser current usingthe MOT loading data. Our results show that the partial Rb pressure near MOTposition is different from the pressure measured at sputter ion pump (SIP) position. Thisestimation of Rb-vapor partial pressure is important for magnetic trapping experimentsto be performed in this setup, where the actual value of pressure determines the lifetimeof atom cloud in the magnetic trap. oading studies... Figure 1.
The schematic diagram of the experimental setup of U-magneto-optical trap(U-MOT). Two MOT-beams in reflection geometry in y-z plane are shown whereasanother two MOT-beams perpendicular to this plane ( ± x-direction) are not shown inthis diagram. PD represents photodiode for detection of fluorescence.
2. Experimental setup
The experimental setup used for this work consists of an octagonal vacuum chamber. Aturbo molecular pump (TMP) (77 l/s), sputter ion pump (SIP) (300 l/s) and titaniumsublimation pump (TSP) are used to achieve the UHV in chamber with base pressure ∼ . × − Torr as read by SIP controller. Four independent MOT-beams, eachof which is made by superposing a cooling beam and repumping beam, are used toform U-MOT as shown in figure 1. The physical size of MOT-beams are ∼
18 mm indiameter. During the U-MOT operation, a current of ∼
54 A is supplied to the copperU-wire in presence of bias fields ( B y ∼
11 G and B z ∼ ∼
17 cm from the centre of the octagonal chamber.The Rb vapor is injected into the vacuum chamber by passing a current through thisfeed-through. The pressure inside the vacuum chamber was dependent on the currentsupplied to the dispenser. The SIP controller displayed the pressure of ∼ . × − Torr at the dispenser of 2.8 A and ∼ . × − Torr at a higher dispenser currentvalue of 3.4 A.The emission of Rb vapor depends on temperature of dispenser which can be con-trolled by the value of current passing through dispenser. At a given dispenser current,the emission rate from the dispenser gets saturated as the thermal equilibrium is achievedat the dispenser [24]. The growth of Rb atoms concentration in the experimental cham-ber for different values of Rb dispenser current was monitored by collecting Rb-vaporfluorescence generated by MOT-beams using a high sensitivity photodiode (PD). Thegrowth of fluorescence signal (i.e. number density of Rb atoms in the chamber) is shownin figure 2(a) for the Rb-dispenser current of 3.4 A. The photodiode signal gets satu-rated in ∼
150 s. It was found that saturation time is nearly same ( ∼
150 s) for all oading studies... Rb dispensr on P ho t od i od e s i gn a l ( m V ) Time (s) (a) P ho t od i od e s i gn a l ( m V ) Time (s)
Rb dispenser off (b)
Figure 2. (a) The photodiode signal measuring the fluorescence from background Rb-atoms in chamber for dispenser current of 3.4 A. (b) The variation in Rb fluorescencesignal with time after the dispenser current of 3.4 A was switched-off. In plot (b), theexperimental data are shown by continuous curve whereas fitted data are shown bydashed curve. values of dispenser current used during the experiment. But, the saturated photodiodesignal value increased with increase in the dispenser current. It was observed that thephotodiode signal of Rb fluorescence at dispenser current of 3.4 A was ∼
30 times moreas compared to that at dispenser current of 2.8 A. The decay of fluorescence signal startsafter switching-off the Rb-dispenser current. The fluorescence decays exponentially withtime constant (1/e) of ∼
3. Results and discussions
The fluorescence from the U-MOT cloud is detected using a high sensitivity photodiode(PD) as well as using a charge-coupled device (CCD) camrea. From the optical power(P) measured by this calibrated PD, we can estimate the total number of atoms N inthe MOT cloud using the formula [21], P = hcλ ( I/I sat )(Γ / I/I sat ) + (2∆ L / Γ) ) N Ω4 π , (1)where I is the total intensity of all the cooling beams, I sat (1 . mW/cm ) is satura-tion intensity for the transition for circularly polarized beam, ∆ L is detuning of coolinglaser beam and Ω is the solid angle subtended by fluorescence collecting imaging lenson atom cloud.The loading of atoms in a MOT is determined by the rate equations, dNdt = R − Nτ L (2) oading studies... N is the number of atoms at any time (t) in the MOT, R is the rate at whichatoms are being trapped in the MOT from the background Rb vapor and τ L is loadingtime constant. The loading time ( τ L ) of U-MOT depends upon partial Rb pressure( P Rb ) in the chamber and is given by [25], τ L = 1 βP Rb + γ b (3)where the term βP Rb represents the loss rate due to collisions with untrapped Rbatoms in the background and γ b represents collisional loss rate due to other (non-Rb)background atoms/molecules.The solution of equation (2) is an exponential growth of number of cold atoms in theMOT which can be written as N = N s [1 − exp ( − t/τ L )] , (4)where N s = Rτ L = αP Rb τ L . (5)The N s represents the steady state number of cold atoms in the MOT and α is a constantrepresenting MOT trapping cross section. After using equations (3) and (5), the relationbetween N s and τ L is given by N s = αβ (1 − γ b τ L ) . (6)Since, N s and τ L can be measured from MOT loading graph. It is obvious fromequation (6) that the N s and τ L are related, but values of α / β and γ b govern the vari-ation of N s with τ L . Hence, by measuring the variation of N s with τ L for different Rbpressure values (by varying dispenser current), it is possible to estimate the Rb pressure(from α / β ) as well as non-Rb background pressure (from γ b ).By measuring the fluorescence signal, the loading of U-MOT has been investigatedfor different values of dispenser current. The dispenser current is switched on ∼ ∼ − Torr. For MOT loadingstudies, the fluorescence from the U-MOT atom cloud was detected using the same pho-todiode (PD) as a function of time after switching on the MOT-beams. For estimationof number of atoms in the MOT, the photodiode signal was corrected after subtractingthe signal corresponding to the background Rb-vapor. Figure 3(a) shows the loadingcurve of U-MOT i.e. rise in number of atoms in U-MOT with time, for different valuesof dispenser current. The observed fluorescence signal is increased as dispenser cur-rent was varied form 2.8 A to 3.4 A. The loading curves are fitted with exponentialgrowth function to determine the loading time constant ( τ L ) for different values of Rbdispenser current. It can be seen from the graph that τ L decreases with increase indispenser current. This is due to availability of more Rb number density in the cham-ber at higher dispenser current. Figure 3(b) shows the fluorescence decay curve of theU-MOT atom cloud for different values of dispenser current. This measurement was oading studies... F l uo r e s ce n ce s i gn a l fr o m U - M O T ( m V ) Time (s) (a) F l uo r e s ce n ce s i gn a l fr o m U - M O T ( m V ) Time (s) (b)
Figure 3. (a) Loading curve of U-MOT for different values of dispenser current. Theexperimental data are shown by continuous curve whereas fitted data is shown bydashed curve. The trapping beams are turned on at t = 0 s and observed data is fit toequation (4). (b) The fluorescence decay curve of the U-MOT after switching-off theRb dispenser current for different values of current. From the exponential fit of themeasured data in plot (b), the MOT lifetime ( τ D ) is ∼ ∼ mW/cm and detuning ∼ is - 12MHz throughout the experiment. Table 1.
The table of measured MOT loading time and lifetime of the MOT cloudfor different values of Rb-dispenser current. All other parameters were fixed duringthe measurements.Dispenser current Loading time Lifetime(A) ( τ L )(s) ( τ D )(s)2.8 A 27.72 ± ± ± ± ± ± ± ± done by switching-off the dispenser current and measuring the fluorescence signal fromMOT cloud as function of time, while MOT laser beams and quadrupole magnetic fieldwere kept on. The fluorescence decay curve is fitted with exponential decay function fordifferent values of Rb-dispenser current to determine the MOT lifetime ( τ D ). The vari-ation in the values of loading time constant ( τ L ) and MOT lifetime ( τ D ) with dispensercurrent is presented in Table-1. It is evident from the table that the loading time ( τ L )aswell MOT lifetime ( τ D ) decreased as dispenser current was increased from 2.8 A to 3.4 A.The variation in MOT loading rate (R) and pressure inside vacuum chamber (asmeasured by SIP controller) with increase in Rb-dispenser current are shown in figure4(a). As dispenser current was increased, the loading rate and pressure both increased.The loading rate was increased to ∼ . × atoms/s and pressure in chamber was oading studies... Dispenser current (A) L o a d i ng R a t e ( R ) ( x a t o m s / s ) R (atoms/s)
P (Torr) P r e ss u r e ( P ) ( x - T o rr) (a) N u m b e r o f c o l d a t o m s ( x ) Dispenser current (A) (b)
Figure 4. (a) The variation of loading rate (R) and pressure read by SIP controllerinside the chamber with dispenser current. The filled circles show loading rate and thestars symbol show pressure inside the chamber. (b) The variation in number of coldatoms in U-MOT with dispenser current. N s ( x ) Loading time ( L ) (s) (a) P a r ti a l R b p r e ss u r e ( x - T o rr ) Dispenser current (A) (b)
Figure 5. (a) The variation of steady state cold atom number ( N s ) with MOT loadingtime ( τ L ). The experimental data is fit to the equation (6). The fitting provides thevalues of αβ = (6 . ± . × and γ b = (2 . ± . × − s − . (b) The variationin estimated Rb-pressure with dispenser current. increased to ∼ . × − Torr, as dispenser current was increased to 3.4 A. Figure 4 (b)shows the variation in cold atom number with the Rb-dispenser current. As dispensercurrent was increased from 2.8 A to 3.4 A, the number of Rb atoms in MOT cloudincreased from ∼ . × to ∼ . × . The enhancement in number of cold atomsis due to availability of high number density of Rb atoms inside the vacuum chamberfor higher values of dispenser current.Using the experimental data on MOT loading, we can find N s and τ L for differentdispenser currents. Figure 5(a) shows this variation in N s in U-MOT with loading timeconstant ( τ L ). Using this N s vs τ L plot and fitting the measured data with equation (6), oading studies... αβ and γ b . From the fit, we obtain γ b = (2 . ± . × − s − and αβ = (6 . ± . × . The value of γ b is related to background (non-Rb) pressure ( P )and it is converted to pressure by using relation γ b /P = 4 . × T orr − s − [25, 26].The partial Rb pressure can be estimated using the value β = 4 . × T orr − s − asgiven in literature [25, 26]. This gives value of α = (2 . ± . × T orr − s − . Usingthese values of α and γ b , the Rb atoms pressure as well as background gas (non-Rb atomsand molecules) pressure in the chamber can be estimated respectively. The estimatedRb pressure is ∼ . × − Torr at dispenser current of 2.8 A and ∼ . × − Torrat dispenser current of 3.4 A. The estimated value of partial pressure of Rb-vapor inchamber is plotted with Rb-dispenser current in figure 5(b).Using the value of γ b , the estimated background pressure in our chamber is ∼ . × − Torr. It is around three times higher than the base pressure mea-sured by SIP controller ( ∼ × − Torr) without any current in dispenser. Thisdifference is due to several following possible reasons. One reason is that the methodadopted above to estimate the background pressure from the γ b value is approximatewith factor of 2 which depends upon the variation in trap parameters, background gascomposition or trapped alkali-metal species [26]. Another reason for the discrepancybetween estimated pressure and SIP controller measured pressure could be the locationof the SIP in the low conductance region.This approach of estimating the Rb and non-Rb pressure in chamber can be helpfulin optimizing dispenser current for magnetic trapping experiments, where lower pressureand sufficient number of cold atoms in the MOT are desirable parameters. For exam-ple in our U-MOT operation at the dispenser current of 3.0 A, the Rb vapor pressure ∼ . × − Torr with cold atom number is ∼ . × . Such a parameter regimecan be suitable for magnetic trapping experiments.
4. Conclusion
In conclusion, the loading of a U-magneto-optical trap (U-MOT) in continuous modefrom background vapor in UHV environment is studied. It is found that the loadingrate and pressure inside chamber increases with increase in the dispenser current.There is monotonic decrease in MOT lifetime and MOT loading time with increasein Rb-dispenser current. Using the MOT loading studies, the partial Rb pressure andbackground pressure in the chamber have been estimated. The estimated Rb pressurewas increased from ∼ . × − Torr to 4 . × − Torr with increase in dispensercurrent from 2.8 A to 3.4 A. These results can be helpful in optimizing dispenser currentfor magnetic trapping experiments where lower partial pressure of Rb atoms havingsufficient cold atom number in the MOT is desirable. oading studies...
5. Acknowledgement
We acknowledge the help provided by Shri Amit Chaudhary during the experiments.We are thankful to S. P. Ram and S. Singh for helpful discussions during this work. Weare also thankful to A. Kak for fabrication of vacuum feed-throughs and R. Shukla andC. Mukherjee for fabrication of atom-chip.
6. ORCID iDs
Vivek Singh https://orcid.org/0000-0002-8132-8504
References [1] Folman R, Kruger P, Schmiedmayer J, Denschlag J H and Henkel C 2002 Microscopic atom optics:from wires to an atom chip
Adv. At. Mol. Opt. Phys. , 263[2] Fortagh J and Zimmermann C 2007 Magnetic microtraps for ultracold atoms Rev. Mod. Phys. ,235[3] Dendchlag J, Cassettari D, Chenet A, Schneider S and Schmiedmayer J 1999 A neutral atom anda wire: towards mesoscopic atom optics. Appl. Phys. B , 291[4] Reichel J, Hansel W and Hansch T W 1999 Atomic Micromanipulation with Magnetic SurfaceTraps, Phys. Rev. Lett. , 3398[5] Eckel S, Barker D S, Fedchak J A, Klimov N N, Norrgard E, Scherschligt J, Makrides C andTiesinga E 2018 Challenges to miniaturizing cold atom technology for deployable vacuummetrology, Metrologia , S182[6] Muller D, Anderson D Z, Grow R J, Schwindt P D D and Cornell E A 1999 Guiding Neutral AtomsAround Curves with Lithographically Patterned Current-Carrying Wires Phys. Rev. Lett. ,5194[7] Cassettari D, Hessmo B, Folman R, Maier T and Schmiedmayer J 2000 Beam Splitter for GuidedAtoms Phys. Rev. Lett. , 5483[8] Hansel W, Hommelhoff P, Hansch T W and Reichel J 2001 Bose-Einstein condensation on amicroelectronic chip Nature , 498[9] Aubin S, Myrskog S, Extavour M H, Leblanc L J, Mckay D, Stummer A and Thywissen J H 2006Rapid sympathetic cooling to Fermi degeneracy on a chip
Nature Physics , 384[10] Wildermuth S, Kruger P, Becker C, Brajdic M, Haupt S, Kasper A, Folman R and SchmiedmayerJ 2004 Optimized magneto-optical trap for experiments with ultracold atoms near surfaces PhysRev A , 030901(R)[11] Phillips W D and Metcalf H 1982 Laser Deceleration of an Atomic Beam Phys. Rev. Lett. , 596[12] Pedrozo-Penafiel E, Vivanco F, Castilho P, Paiva R R, Farias K M and Bagnato V S 2016 Directcomparison between a two-dimensional magneto-optical trap and a Zeeman slower as sources ofcold sodium atoms Laser Phys. Lett.
13 065501[13] Castilho P C M, Pedrozo-Penafiel E, Gutierrez E M, Mazo P L, Roati G, Farias K M and BagnatoV S 2019 A compact experimental machine for studying tunable Bose-Bose superfluid mixtures
Laser Phys. Lett.
16 035501[14] Ram S P, Tiwari S K, Mishra S R and Rawat H S 2013 Push beam spot-size dependence of atomtransfer in a double magneto-optical trap setup
Rev. Sci. Instrum , 073102[15] Myatt C J, Newbury N R, Ghrist R W, Loutzenhiser S and Wieman C E 1996 Multiply loadedmagneto-optical trap Opt. Lett. , 290[16] Gibble K, Chang S and Legere R 1995 Direct observation of s-wave atomic collisions Phys. Rev.Lett. , 2666 oading studies... [17] McDowall P D, Grunzweig T, Hilliard A and Andersen M F 2012 An atomic beam source for fastloading of a magneto-optical trap under high vacuum Rev. Sci. Instrum. , 055102[18] Moore K L, Purdy T P, Murch K W, Leslie S, Gupta S and Stamper-Kurn D M 2005 Collimated,single-pass atom source from a pulsed alkali metal dispenser for laser cooling experiments Rev.Sci. Instrum. , 023106[19] Dugrain V, Rosenbusch P and Reichel J 2014 Alkali vapor pressure modulation on the 100 msscale in a single-cell vacuum system for cold atom experiments Rev. Sci. Instrum. , 083112[20] Treutlein P 2008 Coherent manipulation of ultra cold atoms on atom chips
PhD Thesis Ludwig-Maximilians-Universitat Munchen[21] Pritchard M J 2006
Manipulation of ultracold atoms using magnetic and optical fields
PhD ThesisUniversity Durham[22] Singh V, Tiwari V B, Singh K A P and Mishra S R 2018 On loading of a magneto-optical trap onan atom-chip with U-wire quadrupole field
J. Mod. Opt. , 2332[23] Singh V, Shukla R, Mukherjee C, Tiwari V B and Mishra S R 20-23 December 2017Characterization of in-house fabricated atom chip for magnetic trapping of Rb atoms Proceedings of National Laser Symposium, Mumbai, India, .[24] Fortagh J, Grossmann A, Hansch T W and Zimmermann C 1998 Fast loading of a magneto-opticaltrap from a pulsed thermal source
J. Appl. Phys. , 6499[25] Moore R W G, Lee L A, Findlay E A, Campo L T, Bruce G D and Cassettari D 2015 Measurementof vacuum pressure with a magneto-optical trap: A pressure-rise method Rev. Sci. Instrum. ,093108[26] Arpornthip T and Sackett C A 2012 Vacuum-pressure measurement using a magneto-optical trap Phys. Rev. A85