A new type of optical cw atom laser design is proposed that should operate at high intensity and high coherence and possibly record low temperatures. It is based on an "optical-shepherd" technique, in which far-off-resonance blue-detuned swept sheet laser beams are used to make new types of high-density traps, atom waveguides, and other components for achieving very efficient Bose-Einstein condensation and cw atom laser operation. A shepherd-enhanced trap is proposed that should be superior to conventional magneto-optic traps for the initial collection of molasses-cooled atoms. A type of dark-spot optical trap is devised that can cool large numbers of atoms to polarization-gradient temperatures at densities limited only by three-body collisional loss. A scheme is designed to use shepherd beams to capture and recycle essentially all of the escaped atoms in evaporative cooling, thereby increasing the condensate output by several orders of magnitude. Condensate atoms are stored in a shepherd trap, protected from absorbing light, under effectively zero-gravity conditions, and coupled out directly into an optical waveguide. Many experiments and devices may be possible with this cw atom laser.
A new type of optical cw atom laser design is proposed that should operate at high intensity and high coherence and possibly record low temperatures. It is based on an "optical-shepherd" technique, in which far-off-resonance blue-detuned swept sheet laser beams are used to make new types of high-density traps, atom waveguides, and other components for achieving very efficient Bose-Einstein condensation and cw atom laser operation. A shepherd-enhanced trap is proposed that should be superior to conventional magneto-optic traps for the initial collection of molasses-cooled atoms. A type of dark-spot optical trap is devised that can cool large numbers of atoms to polarization-gradient temperatures at densities limited only by three-body collisional loss. A scheme is designed to use shepherd beams to capture and recycle essentially all of the escaped atoms in evaporative cooling, thereby increasing the condensate output by several orders of magnitude. Condensate atoms are stored in a shepherd trap, protected from absorbing light, under effectively zero-gravity conditions, and coupled out directly into an optical waveguide. Many experiments and devices may be possible with this cw atom laser.
We discuss our recent experimental demonstrations of three novel atom traps which are all based on the optical dipole force in far-detuned laser light. In the gravito-optical surface trap (GOST), Cs atoms are trapped and cooled by means of an evanescent-wave mirror very closely above a dielectric surface. This trap allows for optical cooling at very high densities and provides good starting conditions for evaporative cooling to quantum degeneracy and the realization of a 2D quantum gas. The conical atom trap (CAT) uses a hollow blue-detuned laser beam in combination with gravity to trap atoms in a dark spatial region. This trap is experimentally simple and provides high loading efficiency, tight confinement, efficient sub-Doppler cooling, and strong suppression of collisional losses. The standing-wave red-detuned (STAR) trap allows one to confine atoms in a shallow optical potential at a very low photon scattering rate, and is thus an interesting tool for studies of the atomic ground-state dynamics in external electro-magnetic fields. We have studied atomic spin polarization in this trap.
In 1925, Einstein predicted the condensation of bosons into the ground state of the system for low (but finite) temperatures. Several phenomena, including superfluidity and superconductivity have been associated with Bose-Einstein condensation, but these systems interact strongly with their environment and pure Bose-Einstein condensation could not be established. It took 70 years, in which time the laser was discovered, and laser cooling techniques to manipulate atoms in a dilute atomic gas, before Bose-Einstein condensation in dilute atomic gases could be demonstrated in 1995. In the first condensation experiments, BECs were created in a magnetic trap. Since in a magnetic trap not all mF states of the atom can be trapped simultaneously, thereby limiting the number of experiments that can be done, other ways of trapping and generating BECs were sought and found. In 2001, the first all-optical BEC was made, where the dipole force was used to trap atoms in the crossing of two far red detuned laser beams. In an optical dipole trap not only atoms in different internal states can be trapped, but also different atomic species simultaneously. In this thesis, the formation of an all-optical Bose-Einstein condensate with rubidium atoms is presented. Conventional all-optical BECs are usually created in high power CO2 laser dipole traps, or have complicated laser cooling schemes and complex dipole trap setups. In our simple and straightforward setup, we load rubidium atoms from a magneto-optical trap into a crossed optical dipole trap created by a single frequency Yb:YAG laser with a wavelength at 1030 nm. The small wavelength allows for a small diffraction limit, and permits us to use standard optical materials, thus making the experimental setup cost effective. Other attempts to achieve Bose-Einstein condensation in a multi-mode (frequency) fiber laser at 1064 nm failed, because the atom loss was quite high. It is assumed that the multi-mode character of the fiber laser induces Raman transitions in rubidium atoms, thereby heating them. We can trap about ∼ 5E7 atoms in a single beam dipole trap out of ∼ 5E9 atoms trapped in theMOT, and ∼ 350,000 atoms can be trapped in a crossed beam dipole trap due to the smaller trap volume. 70% of the atoms in the dipole trap is optically pumped into one mF state. Quantum degeneracy is reached by evaporatively cooling the atoms in the crossed dipole trap by ramping down the laser power with three linear ramps. We can independently change the power of each beam by an AOM. This allows us to use one beam as an atom waveguide for future experiments. After evaporation, we typically have about 10,000 atoms at a temperature below the critical temperature. We have proved Bose-Einstein condensation by using the anisotropic expansion of a quantum degenerate gas trapped in an anisotropic potential. The aspect ratio of our atom cloud changed during a time of flight from 0.7 to 1.2 in 10 ms, thus proving that we have reached quantum degeneracy. We have about 5,000 condensed atoms in our optical dipole trap at a temperature less than 100 nK. The remaining atoms are thermal. Bose-Einstein condensation is obtained within 8 s, and we can repeat the experiment every 30 s. It should be mentioned that the Bose-Einstein experiment was moved from the ”Leibniz Universität Hannover” to the ”Technische Universität Darmstadt”, and had to be completely rebuilt. All-optical Bose-Einstein condensation was reached within one year after the move. Our Bose-Einstein condensation setup presents an ideal starting point for using our condensates in combination with miniaturized atom optical setups based on our novel microfabricated optical elements. With our microlenses, we can create a number of possible dipole trap configurations, such as the dipole trap array or the cylindrical microlens array. Using microlenses in miniaturized atom optical setups opens a completely new field of coherent atom optics. Also because the tight confinement of the microtraps allows us to load a 3D BEC, a 1D BEC, or a Tonks-Girardeau gas in the micropotentials depending on the density.
This thesis reports on the design, setup, and operation of a new atom trap apparatus, which is developed to confine few rubidium atoms in ultrahigh vacuum by means of taylored and dynamically variable light potentials (dipole traps), thus making the atoms available for spectroscopy and controlled manipulations. To maintain low background pressure, atoms of a vapour cell are transferred into a cold atomic beam by laser cooling techniques, and accumulated by a magneto-optic trap (MOT) in a separate part of the vacuum system. The laser cooled atoms are then transferred into dipole traps made of focused far-off-resonant laser fields in single or crossed-beam geometry, which are superimposed with the center of the MOT. Gaussian as well as hollow Laguerre-Gaussian (LG01) beam profiles are used with red-detuned and blue-detuned light. Microfabricated dielectric phase objects allow the efficient and robust mode conversion of Gaussian into Laguerre-Gaussian laser beams. Trap geometries can easily be changed due to the highly flexible experimental setup. The dipole trap laser beams have powers up to several hundred milliwatts and are focused to around 10 m. At a detuning of several tens of nanometers from the atomic resonance, the resulting traps are a few millikelvin deep, and the trapped atoms exhibit motional frequencies near 30 kHz, trap light scattering rates of a few hundred photons per second per atom, and storage times of several seconds. The number of dipole-trapped atoms ranges from more than ten thousand to below ten. The dipole-trapped atoms are detected either by a photon counting system with efficient straylight discrimination, or by recapture into the MOT, which is imaged onto a sensitive photodiode and a CCD-camera. Due to the strong AC-Stark shift imposed by the high-intensity trapping light, energy-selective resonant excitation and detection of the stored atoms is possible. The measured energy distribution is consistent with a harmonic potential shape and allows for the determination of temperatures and heating rates. From initial measurements, the thermal energy is found to be about 10 % of the trap depth. Using a crossed-beam geometry with red-detuned laser light, efficient transfer of atoms between the beams is observed. Optimum transfer occurs when the two beams cross at a radial offset, which can qualitatively be understood when the particle energy and geometrical properties of the two-beam trapping potential are considered. Numerical simulations reproduce the general features of the measured transfer efficiency vs. radial beam offset. This experiment is a first step towards dynamically variable potentials. Atoms have been radially confined in a blue-detuned hollow beam, and their decay time has been determined. This configuration is currently extended to a three-dimensionally confining blue-detuned dipole trap. For advanced laser cooling, state manipulation and spectroscopy of trapped atoms, a system of two diode lasers has been set up which are phase-coherently locked with a difference frequency near 6.834 GHz, to drive Raman transitions between the hyperfine-split ground states of 87Rb atoms. Resonances of such a transition have been recorded in a buffer gas loaded rubidium vapour cell with linewidths below 100 Hz.
Summary form only given. Bose-Einstein condensates (BEC) are generally produced by evaporative cooling of laser cooled atomic samples in magnetic traps. Though magnetic traps are indispensable for the production of a BEC, they pose restrictions on the possible studies with a BEC. Previous experiments of our group have demonstrated that it is possible to trap a BEC in the circular focus of a far-detuned infrared laser beam. This first realization of an optical dipole trap (ODT) for BEC was used to prepare and investigate spinor condensates of Na atoms in the F=1 hyperfine manifold. However, the initial peak density in this ODT after loading from the magnetic trap reached 10/sup 15/ atoms/cm/sup 3/, leading to rapid losses due to three-body collisions. To realize an ODT capable of storing large condensates for a long time, we have used a laser beam with an elliptical beam profile. The optical trap is loaded by superimposing the light field onto the magnetic trap and then slowly turning off the magnetic fields. The transfer efficiency is close to 100%, and thus more than 10/sup 7/ Bose-condensed atoms can be loaded into the ODT.
A detailed experimental and theoretical investigation of a magneto-optical trap for caesium atoms is presented. Particular emphasis has been placed on achieving high spatial number densities and low temperatures. Optimizing both of these together enables efficient evaporative cooling from a conservative trap, a procedure which has recently led to the first observations of Bose-Einstein condensation in a dilute atomic vapour. The behaviour of a magneto-optical trap is nominally determined by four independent parameters: the detuning and intensity of the light field, the magnetic field gradient and the number of trapped atoms. A model is presented which incorporates previous treatments into a single description of the trap that encompasses a wide range of its behaviour. This model was tested quantitatively by measuring the temperature of the cloud and its spatial distribution as a function of the four parameters. The maximum density was found to be limited both by the reabsorption of photons scattered within the cloud and by a reduction of the confining force at small light shifts. The nonlinear variation with position of the restoring force was found to be significant in limiting the number of atoms confined to a high density. A maximum density in phase space (defined as the number of atoms in a box with sides of dimension one thermal de Broglie wavelength) of (1.5 ± 0.5) x 10-5 was observed, with a spatial density of 1.5 x 1011 atoms per cm3. Cold collision losses from a caesium magneto-optical trap have been studied with the purpose of assessing their influence on spatial densities. In contrast to previous measurements of similar quantities, these measurements did not require the use of an ultra-low (< 10-10 Torr) background vapour pressure. The dependence of the cold collision loss coefficient β on the trapping intensity was measured to permit identification of the different cold collision processes. The largest loss rates observed were those due to hyperfine structure-changing collisions, with a coefficient β = (2±1) x 10-10cm3s-1. A study is presented of a modified magneto-optical trap in which a fraction of the population is shelved into a hyperfine level that does not interact with the trapping light. In this so-called "dark" magneto-optical trap, improved densities of nearly 1012cm-3 have been previously reported for sodium. The application of the technique to caesium is not straightforward due to the larger excited state hyperfine splittings. A simple theory for caesium is presented and its main predictions verified by measurements of density, number and temperature. A density of nearly 1012cm,-3 was indeed obtained but at a temperature substantially higher than in the conventional magneto-optical trap.
A detailed experimental and theoretical investigation of a magneto-optical trap for caesium atoms is presented. Particular emphasis has been placed on achieving high spatial number densities and low temperatures. Optimizing both of these together enables efficient evaporative cooling from a conservative trap, a procedure which has recently led to the first observations of Bose-Einstein condensation in a dilute atomic vapour. The behaviour of a magneto-optical trap is nominally determined by four independent parameters: the detuning and intensity of the light field, the magnetic field gradient and the number of trapped atoms. A model is presented which incorporates previous treatments into a single description of the trap that encompasses a wide range of its behaviour. This model was tested quantitatively by measuring the temperature of the cloud and its spatial distribution as a function of the four parameters. The maximum density was found to be limited both by the reabsorption of photons scattered within the cloud and by a reduction of the confining force at small light shifts. The nonlinear variation with position of the restoring force was found to be significant in limiting the number of atoms confined to a high density. A maximum density in phase space (defined as the number of atoms in a box with sides of dimension one thermal de Broglie wavelength) of (1.5 ± 0.5) x 10-5 was observed, with a spatial density of 1.5 x 1011 atoms per cm3. Cold collision losses from a caesium magneto-optical trap have been studied with the purpose of assessing their influence on spatial densities. In contrast to previous measurements of similar quantities, these measurements did not require the use of an ultra-low ( A study is presented of a modified magneto-optical trap in which a fraction of the population is shelved into a hyperfine level that does not interact with the trapping light. In this so-called dark magneto-optical trap, improved densities of nearly 1012cm-3 have been previously reported for sodium. The application of the technique to caesium is not straightforward due to the larger excited state hyperfine splittings. A simple theory for caesium is presented and its main predictions verified by measurements of density, number and temperature. A density of nearly 1012cm,-3 was indeed obtained but at a temperature substantially higher than in the conventional magneto-optical trap.