The thematics treated in this manuscript illustrate two specificities of ultracold metastable Helium gases: the ability to compare experimental measurement with precise theoretical evaluations (simplicity of the atomic structure) and to easily detect products of Penning ionisations. We present the building and the characterisation of a novel magnetic trap with large optical access, allowing to combine the production of a Bose-Einstein condensate with its in situ loading into a 3D optical lattice. The theoretical fundaments of the experiments planned in those optical potentials are then detailed. In a crossed dipole trap, the influence of the magnetic field, now a free parameter, on the Penning collision rates can be measured and compared to a new theoretical evaluation. Concerning Helium in optical lattices, two subjects are developped: the effect of the confinement on the inelastic Penning collisions (1D optical lattice) and the modelisation of the Penning losses via a dissipative Bose-Hubbard model (3D lattice). Finally we present the first direct measurement of the dipolar magnetic transition 23S1 to 23P2, linking the singlet and triplet families in Helium 4. This spectroscopy experiment, realised in collaboration with the group of W. Vassen (LaserLab - Amsterdam), combines cold atoms knowledge with the frequency comb technique, in order to reach an accuracy of 5 kHz.
The thematics treated in this manuscript illustrate two specificities of ultracold metastable Helium gases: the ability to compare experimental measurement with precise theoretical evaluations (simplicity of the atomic structure) and to easily detect products of Penning ionisations. We present the building and the characterisation of a novel magnetic trap with large optical access, allowing to combine the production of a Bose-Einstein condensate with its in situ loading into a 3D optical lattice. The theoretical fundaments of the experiments planned in those optical potentials are then detailed. In a crossed dipole trap, the influence of the magnetic field, now a free parameter, on the Penning collision rates can be measured and compared to a new theoretical evaluation. Concerning Helium in optical lattices, two subjects are developped: the effect of the confinement on the inelastic Penning collisions (1D optical lattice) and the modelisation of the Penning losses via a dissipative Bose-Hubbard model (3D lattice). Finally we present the first direct measurement of the dipolar magnetic transition 23S1 to 23P2, linking the singlet and triplet families in Helium 4. This spectroscopy experiment, realised in collaboration with the group of W. Vassen (LaserLab - Amsterdam), combines cold atoms knowledge with the frequency comb technique, in order to reach an accuracy of 5 kHz.
About 10 4 triplet metastable helium atoms have been trapped in a magneto-optical trap, based on the 2 3 S 1 –2 3 P 2 transition at 1.083 μm (volume 0.1 mm 3 , temperature 1 mK). The losses of the trap are analysed by monitoring the ion current due to Penning collisions either with background gases, or between trapped atoms. The rate coefficient for these collisions between trapped metastable helium atoms has been found to be unusually large, and the crucial role of the quasi-resonant laser light has been experimentally demonstrated. The rate coefficient of the dominant process (He*(2 3 S 1 ) + He*(2 3 P 2 ) Penning collisions) has been measured by two different methods, and found to be K sp ≃ 10 -7 cm 3 s -1 within an estimated factor of 4. The corresponding cross-section is 6 · 10 6 Å 2 .
The main subject of this thesis is to demonstrate that helium is a promising candidate for matter wave interferometry, with the goal of providing an improved value for the fine-structure constant, α. In particular, fast and efficient momentum transfer via Bloch oscillations in an optical lattice is demonstrated, together with the enhanced sensitivity offered by an ion-detector based detection scheme that is unique to metastable helium (He*). The experiments are performed in a setup that was built to produce, slow down and trap ultracold (~0.2 μK) He-4 atoms. The approach to realising an ultracold sample of He* atoms consists of a sequence of trapping and cooling stages. The experiment starts at the plasma discharge source, where helium atoms are excited to the so-called metastable 2-3S1 state, with a lifetime of ~7800 seconds, 20 eV above the atomic ground state. The atomic beam exiting from the source is 2-D collimated by 1083 nm laser light, resonant to the closed 2-3S1 - 2-3P2 cooling transition. Next, the atoms are slowed down from about 1000- to 50 m/s in a two meter long Zeeman slower through continuous absorption of resonant 1083 nm light. The slowed atoms are captured in a Magneto-Optical Trap (MOT), that holds up to 5x10^8 atoms at a temperature of 1 mK. The MOT cloud is compressed, spin-polarised and transferred to a cloverleaf Magnetic Trap (MT), which provides both axial and radial trapping through a set of coils outside the vacuum. In the MT, 1D-Doppler cooling is applied to the atomic sample to reduce the temperature to about 150 μK. Following this step, the temperature is reduced further by forced evaporative radio-frequency cooling, resulting in the onset of a Bose-Einstein Condensate (BEC). The atoms are transferred from the MT to a crossed horizontal optical dipole trap (ODT), in which the sample is further evaporatively cooled to produce a BEC of 5x10^6 atoms. The BEC is then transferred from the ODT to either a horizontal or vertical optical lattice, consisting of two counter-propagating laser beams at 1557 nm. By detuning the frequency of one of the lattice beams, the standing wave can be accelerated. The trapped helium atoms can follow the accelerating lattice through a process called Bloch oscillations (BO's), which describes the behaviour of particles in a periodic potential that are subjected to a force. Specifically, in this scenario, particles' velocities increase in increments of two recoil velocities, instead of continuously. Bloch oscillations of cold atoms in periodic potentials formed by optical lattices were first studied in 1996. Importantly, BO’s provide a way to efficiently accelerate atoms by coherently transferring a well controlled, large number of photon momenta, which has been used to increase the resolution of recoil measurements with caesium and rubidium atoms. In this thesis, Bloch oscillations with a metastable helium BEC are demonstrated. Due to its light mass, helium can be efficiently accelerated to high velocities which makes it a candidate for experiments that benefit from splitting the atomic wave-function over large distances. Additionally, Rabi-oscillations in an optical lattice are shown, which serve to probe the depth of the lattice, and a measurement of gravity through Bloch oscillations in a static vertical optical lattice is demonstrated, with a statistical uncertainty of 4x10^-5. Finally, the first proof of principle demonstration of a pulsed light He* atom interferometer is shown. The measurements illustrate the high signal-to-noise that can be obtained using an MCP detector. Next to that, we have demonstrated a method for determining the polarizability of atoms at specific wavelengths, by observing the induced change of the characteristic Bloch oscillation frequency by introducing that wavelength to the system.
We report on interference studies in the internal and external degrees of freedom of metastable triplet helium atoms trapped near quantum degeneracy in a $$1.5 \,\upmu {\mathrm{m}}$$ optical dipole trap. Applying a single π/2 rf pulse we demonstrate that 50% of the atoms initially in the m=+1 state can be transferred to the magnetic field insensitive m=0 state. Two π/2 pulses with varying time delay allow a Ramsey-type measurement of the Zeeman shift for a high precision measurement of the 23S1 – 21S0 transition frequency. We show that this method also allows strong suppression of mean-field effects on the measurement of the Zeeman shift, which is necessary to reach the accuracy goal of 0.1 kHz on the absolute transition frequencies. Theoretically the feasibility of using metastable triplet helium atoms in the m=0 state for atom interferometry is studied demonstrating favorable conditions, compared to the alkali atoms that are used traditionally, for a non-QED determination of the fine structure constant.
Trapped gases of ultracold metastable helium atoms are a unique experimental sys- tem. The trapped cloud’s low temperatures, the ability to selectively trap and ma- nipulate individual atomic states and the isolation provided in a vacuum chamber are all conducive to the measurement of precision atomic parameters, such as atomic transition rates. Due to it’s atomic structure helium - the simplest multi-electron atom - is a favourite test-bed for precision QED theoretical calculations compared to experimentally determined parameters. This thesis reports on a number of tran- sition rates measured for the first time with high accuracy, including the first excited metastable state lifetime, found to be 7870(510)s, the longest of any neutral atomic species yet measured. Transition rates from the 23P level to the ground state were also measured for the first time and found to be in excellent agreement with theory. Further cooling of helium produces a Bose-Einstein condensate (BEC), where all atoms are in an identical quantum state, the analogue of an optical laser. One par- ticularly interesting property a BEC posesses is long-range coherence to all orders, meaning all component particles are described by a single wavefunction. While first order coherence is measured in an interference experiment like Young’s double slit, higher order coherence is the ability for correlations to exist between groups of two, three or more particles, quantified by the measurement of correlation functions. Prior to the work in this thesis correlation functions for atoms had only been measured for second order. Here we develop a new technique which we use to show that a BEC is coherent to second and third order to 0.1% accuracy, compared to a thermal cloud which we show is incoherent. At small inter-particle spacings, there is an enhanced probability of measuring double and triple events in the thermal cloud (termed bunching). A drawback of all previous atomic correlation measurements is that the small de-Broglie wavelengths compared to the detector resolution leads to a significant reduction in the measured bunching amplitudes compared to theory. By altering our trapping configuration and temperature we are able to measure a two-particle bunching amplitude which is 88% of the theoretical maximum. This system also allowed the measurement of correlations up to 5th order in thermal gases. The near ideal bunching allowed us to demonstrate perturbations to the correlation functions due to the 1D nature of the system. These advances should lead to correlation measurements being used as a probe to characterise more exotic system properties such as low dimensionality, engineered quantum states and paticle pairing. Finally, a waveguide was created using a dipole laser beam and used to guide atoms in analogy to an optical laser in a fibre. Guided ‘atom lasers’ are likely to be important components for practical devices such as precision sensors. We demonstrate that a guided atom laser has a gaussian profile matching the lowest transverse mode of the waveguide and is first and second order coherent. In contrast guided thermal atoms show matter-wave speckle and are incoherent.
Abstract We report on interference studies in the internal and external degrees of freedom of metastable triplet helium atoms trapped near quantum degeneracy in a $$ 1.5\kern0.166667em \upmu \mathrm{m} $$ 1.5 μ m optical dipole trap. Applying a single π/2 π / 2 rf pulse we demonstrate that 50% of the atoms initially in the m=+1 m = + 1 state can be transferred to the magnetic field insensitive m=0 m = 0 state. Two π/2 π / 2 pulses with varying time delay allow a Ramsey-type measurement of the Zeeman shift for a high precision measurement of the 23S1 2 3 S 1 – 21S0 2 1 S 0 transition frequency. We show that this method also allows strong suppression of mean-field effects on the measurement of the Zeeman shift, which is necessary to reach the accuracy goal of 0.1 kHz on the absolute transition frequencies. Theoretically the feasibility of using metastable triplet helium atoms in the m=0 m = 0 state for atom interferometry is studied demonstrating favorable conditions, compared to the alkali atoms that are used traditionally, for a non-QED determination of the fine structure constant.
Two groups, in Orsay and in Paris, have recently obtained experimental evidence for Bose-Einstein condensation on He4 atoms in the metastable state 23S1. The condensates which have been obtained are the first ones where atoms are condensed, not in their electronic ground state, but in an excited state, in this case a metastable state about 20 eV above the ground state. In principle, Penning collisions should very rapidly destroy the cloud of cold metastable atoms and it seems hopeless to observe Bose-Einstein condensation in such a system. But, as suggested by Shlyapnikov et al, the conservation of the total spin during the collision reduces the Penning collision cross-sections by several orders of magnitude and this explains why large enough densities of metastable atoms can be obtained to observe the condensation. We briefly describe the first experimental results which have been obtained in Paris. Atoms are optically detected by optical imaging at 1083 nm (wavelength of the transition connecting the metastable state 23S1 to the triplet state 23P2). The evolution of the anisotropy of the images during a ballistic expansion gives clear evidence for the condensation. Orders of magnitude are also obtained for the scattering length describing the collision between 2 metastable atoms. The large value of this scattering length (of the order of 10 nm) allows one to reach the so called hydrodynamic regime where the mean free path between 2 collisions in the thermal cloud becomes smaller than the size of the cloud. Preliminary results showing the transition between the collisionless regime and the hydrodynamic regime are presented and compared to theoretical predictions.