We report on the generation of a Bose-Einstein condensate in a gas of chromium atoms, which will make studies of the effects of anisotropic long-range interactions in degenerate quantum gases possible. The preparation of the chromium condensate requires novel cooling strategies that are adapted to its special electronic and magnetic properties. The final step to reach quantum degeneracy is forced evaporative cooling of 52 Cr atoms within a crossed optical dipole trap. At a critical temperature of Tc ≈700 nK, we observe Bose-Einstein condensation by the appearance of a two-component velocity distribution. Released from an anisotropic trap, the condensate expands with an inversion of the aspect ratio. We observe critical behavior of the condensate fraction as a function of temperature and more than 50,000 condensed 52 Cr atoms. The essential properties of degenerate quantum gases depend on range, strength and symmetry of the present interactions. Since the first observation of Bose-Einstein condensation in weakly interacting atomic gases, eight different elements have been Bose-Einstein condensed [1,2,3,4,5,6,7,8]. All these elements, mainly alkali atoms, interact dominantly via short-range isotropic potentials. Based on this effective contact interaction, many exciting phenomena have been studied [9,10]. Examples are the realization of four-wave mixing with matter waves [11] as well as the observation of vortices [12,13] and solitons [14,15,16] in degenerate quantum gases. Bose-Einstein condensates (BECs) with contact interaction have also been used to investigate solid-state physics problems like the Mott-metal-insulator transition [17,18]. Tuning the contact interaction, the collapse and explosion ("Bosenova") of Bose-Einstein condensates has been studied [19] and new types of quantum matter like a Tonks-Girardeau gas have been realized [20]. In a chromium Bose-Einstein condensate, one can not only tune the short-range contact interaction using one of the recently observed Feshbach resonances [21] but also investigate effects of the longrange and anisotropic dipole-dipole interaction. This becomes possible because, compared to other Bose-condensed elements, the transition metal chromium has a unique electronic structure. The valence shell of its ground state contains six electrons with parallel spin alignment (electronic configuration: [Ar]3d 5 4s 1 ). For the bosonic chromium isotopes, which have no nuclear spin, this gives rise to a total electronic spin quantum number of 3 and a very high magnetic moment of 6 µ B (µ B is the Bohr magneton) in its ground state 7 S 3 . Since the magnetic dipole-dipole interaction (MDDI) scales with the square of the magnetic moment, it is a factor of 36 higher for chromium than for alkali atoms. For this reason, dipole-dipole interactions which have not yet been investigated experimentally in degenerate quantum gases will become observable in chromium BEC. For example, it was shown in [22] that the MDDI in chromium is strong enough to manifest itself in a well pronounced modif...

We have observed Feshbach resonances in collisions between ultracold 52Cr atoms. This is the first observation of collisional Feshbach resonances in an atomic species with more than one valence electron. The zero nuclear spin of 52Cr and thus the absence of a Fermi-contact interaction leads to regularly spaced resonance sequences. By comparing resonance positions with multichannel scattering calculations we determine the s-wave scattering length of the lowest (2S+1)Sigma(+)(g) potentials to be 112(14) a(0), 58(6) a(0), and -7(20) a(0) for S=6, 4, and 2, respectively, where a(0)=0.0529 nm.

We have investigated both theoretically and experimentally dipolar relaxation in a gas of magnetically trapped chromium atoms. We have found that the large magnetic moment of 6 µ B results in an event rate coefficient for dipolar relaxation processes of up to 3.2 · 10 −11 cm 3 s −1 at a magnetic field of 44 G. We present a theoretical model based on pure dipolar coupling, which predicts dipolar relaxation rates in agreement with our experimental observations. This very general approach can be applied to a large variety of dipolar gases.

We have realized a scheme for continuous loading of a magnetic trap (MT).52 Cr atoms are continuously captured and cooled in a magneto-optical trap (MOT). Optical pumping to a metastable state decouples atoms from the cooling light. Due to their high magnetic moment (6 µB), low-field seeking metastable atoms are trapped in the magnetic quadrupole field provided by the MOT. Limited by inelastic collisions between atoms in the MOT and in the MT, we load 10 8 metastable atoms at a rate of 10 8 atoms/s below 100 µK into the MT. Optical repumping after the loading allows us to realize a MT of ground state chromium atoms. Although multiply loading of a MT has been achieved [10] experiments so far suffer from the absence of a method for efficient cw loading of atoms into a BEC. Hence to date a matter wave analogon to the continuous wave optical laser has not been realized. Alternatively, a cw atom laser based on magnetic guiding in combination with atomic collisions was suggested [11]. In addition, cw loading of low-dimensional optical traps with laser cooled atoms was proposed [12] and recently demonstrated [13].In this letter we report on the cw loading of a three dimensional conservative trap with laser cooled atoms that are decoupled from all light fields present. We show that atoms can be optically pumped within a chromium magneto-optical trap (MOT) [14,15] into metastable "dark" states and stored in a MT built up by the quadrupole magnetic field of the MOT. We present results of systematic studies on the loading process and on the lifetime of the MT. Using the cw loading mechanism and a final repumping process we obtain good starting conditions for experiments towards degenerate quantum gases with ground state chromium atoms.Our cw loading scheme consists of an atomic reservoir and a conservative trap overlapped in space and time. The reservoir is prepared by cooling atoms in a MOT on a transition |g → |e (FIG. 1). A weak decay channel |e → |d allows the transfer of reservoir atoms into an additional long lived and trapped state |d in which atoms can be accumulated. In our realization low field seeking Zeeman substates of |d are trapped in the magnetic quadrupole field of the MOT. The loading can be very efficient if |d atoms are decoupled from the MOT light and if their kinetic energy is smaller than the conservative trap depth. A large decay rate branching ratio (Γ eg /Γ ed ≫ 1) assures a steady state MOT in thermal equilibrium and is expected to greatly reduce reabsorption of transfer photons by atoms in the MT [16]. MT FIG. 1. Relevant part of the52 Cr level scheme. The MOT involves all levels and transitions, the continuous loading process of the magnetic trap (MT) relies on the Λ-system depicted in black (levels |g , |e , |d ).Chromium combines the desired Λ-like level scheme (FIG. 1, black levels and transitions) with a high magnetic moment of up to 6 µ B (µ B =Bohr's magneton). Due to its isotopic composition (3 bosons:52 Cr (84%), 50 Cr (4%), 54 Cr (2%), and one fermion: 53 Cr (10%) ) it is a promising e...

We demonstrate demagnetization cooling of a gas of ultracold 52 Cr atoms. Demagnetization is driven by inelastic dipolar collisions which couple the motional degrees of freedom to the spin degree. By that kinetic energy is converted into magnetic work with a consequent temperature reduction of the gas. Optical pumping is used to magnetize the system and drive continuous demagnetization cooling. Applying this technique, we can increase the phase space density of our sample by one order of magnitude, with nearly no atom loss. This method can be in principle extended to every dipolar system and could be used to achieve quantum degeneracy via optical means.

We present the realisation of a magneto-optical trap for chromium. 52Cr atoms are loaded directly from a thermal beam. The trap lifetime is enhanced by using two red lasers to repump population that has decayed, via intercombination lines, to metastable levels back into the cooling cycle. We have measured the wavelengths of these intercombination lines and observed coherent Raman spectra. The observed density of 108 cm−3 is currently limited by collisions with the hot beam.

We have studied a general technique for laser cooling a cloud of polarized trapped atoms down to the Doppler temperature. A one-dimensional optical molasses using polarized light cools the axial motional degree of freedom of the atoms in the trap. Cooling of the radial degrees of freedom can be modelled by reabsorption of scattered photons in the optically dense cloud. We present experimental results for a cloud of chromium atoms in a magnetic trap. A simple model based on rate equations shows quantitative agreement with the experimental results. This scheme allows us to readily prepare a dense cloud of atoms in a magnetic trap with ideal starting conditions for evaporative cooling.

We have measured the deca-triplet s-wave scattering length of the bosonic chromium isotopes 52 Cr and 50 Cr. From the time constants for cross-dimensional thermalization in atomic samples we have determined the magnitudes |a( 52 Cr)| = (170 ± 39) a0 and |a( 50 Cr)| = (40 ± 15) a0, where a0 = 0.053 nm. By measuring the rethermalization rate of 52 Cr over a wide temperature range and comparing the temperature dependence with the effective-range theory and single-channel calculations, we have obtained strong evidence that the sign of a( Cs [5], Feshbach resonances have been used to tune the scattering length into the right range for achieving BEC. Therefore, a careful measurement of the scattering length is a first requirement before a strategy for the condensation of a new atomic species can be devised.However, elastic collisions are not only essential for evaporative cooling of an atomic gas but they also determine the interaction in a quantum degenerate gas. The presence of this interaction is also responsible for many of the fascinating features of BECs like superfluid behavior or the existence of phonon-like excitations [1]. In the case of chromium, the magnitude of the scattering length will also determine how significant the effects of the anisotropic dipole-dipole interaction will be [6].Up to now, only the s-wave scattering lengths for alkali atoms, hydrogen and metastable helium have been measured. The most precise methods are probably photoassociation spectroscopy [7,8] and Raman spectroscopy [9] between vibrational levels of the electronic molecular ground state. Another possibility is Feshbach resonance spectroscopy [10,11,12] where the position and magnetic field dependence of several Feshbach resonances is determined.For chromium, no molecular spectroscopy data are available for the deca-triplet state which -neglecting the spin-spin interaction -corresponds to the collisional channel for elastic collisions in samples prepared in spinstretched (i.e. |m J | = J) states. Due to the complicated electronic structure with six unpaired outer-shell electrons, accurate ab initio calculations of the molecular potentials are also not available. Fortunately nearthreshold collisions are not sensitive to the details of the short-range interaction potential. They are to a large extent determined by the s-wave scattering length a and by the long-range dispersion interaction, characterized by the leading van der Waals coefficient C 6 .In this paper, we report on a measurement of the swave scattering length of bosonic chromium atoms, spin polarized in the weak-field seeking J = m J = 3 state, from cross-dimensional relaxation in a magnetic trap. By recording the relaxation of an anisotropic temperature distribution in an atomic cloud towards equilibrium, we obtain a relaxation time constant which is proportional to the effective scattering cross-section through the atomic density. This method has been first used by Monroe et al. [13] for 133 Cs and has since been employed to determine the ultra-cold scattering prop...