University of Utrecht Debye Institute Cold Atom Nanophotonics Group An Optical Conveyor for Light-Atom Interaction Supervisors: Author: B. O. Mussmann M.Sc. S. Greveling B.Sc. A. J. van Lange M.Sc. Dr. D. van Oosten December 22, 2013 Abstract In the Cold Atom Nanophotonics group, the interaction between 87Rb and sur- faceplasmonswillbeinvestigated. Forthatreason,atomsneedtobepositioned in the near field of a gold film. This is the most challenging part as the near field extend to a distance of ∼ λ/2 above a sample surface, where λ denotes the wavelength of light. The first step is to trap and cool the atoms using a magneto-optical trap. From this trap the atoms are loaded into a far detuned optical dipole trap. By displacing the focus of this trap towards the sample, the atoms are displaced. When the atoms are directly above the sample they are loaded into a red-detuned optical conveyor, which displaces the atoms towards the sample surface to a height of λ/2. Using two synchronized pairs of rotating mirrors, the optical conveyor is translated in two-dimensions over the sample surface to position the atoms above the correct structure. Inthisthesiswedescribetheopticaldipoletrapthatweusedtotrap2.1×105atoms. We have measured the atomic lifetime in the optical dipole trap for different trap depths. Using a laser power of 10.4W, we obtained a lifetime of up to 6.4±2.9s. Our magneto-optical trap contains (4.0±0.2)×1010atoms. We achieved a transfer efficiency from the magneto-optical trap into the optical dipole trap of ∼ 4×10−4%. In future experiments this transfer efficiency can be increased by readily implementing changes to the loading procedure of the dipole trap. For the movement of the atoms towards the sample surface, we will use a red- detuned optical conveyor. In this optical conveyor atoms are trapped at the anti-nodes of a standing wave. To displace these anti-nodes, we give the beam producing the standing wave a slight frequency offset. We tested this method interferometrically. We conclude that we can displace the anti-nodes over a distanceof>2mmwithanaccuracyof150nm. Theaccuracyismostlikelydue to mechanical vibrations. We developed a feedback mechanism to increase the accuracy of the displacement of the anti-nodes. Finally, we designed a setup for the optical conveyor. This setup includes the translation of the optical conveyor in two-dimensions. It is to be implemented into the experiment. Contents Contents 1 Introduction 1 2 Magneto-Optical Trap: Theory and Setup 3 2.1 Theory of magneto-optical trapping . . . . . . . . . . . . . . . . . . . 3 2.1.1 Doppler cooling . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.1.2 Magnetic trapping . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.1.3 Multi-level atoms in a magneto-optical trap . . . . . . . . . . . 7 2.2 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2.1 Rubidium source . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.2.2 Vacuum chambers . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2.3 Magnetic fields . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2.4 Diode lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3 Optical Dipole Trap: Theory and Setup 12 3.1 Theory of optical dipole trapping . . . . . . . . . . . . . . . . . . . . 12 3.1.1 Classical Lorentz model . . . . . . . . . . . . . . . . . . . . . . 12 3.1.2 Multi-level atoms in an optical dipole trap . . . . . . . . . . . 16 3.1.3 Trap parameters of the optical dipole trap . . . . . . . . . . . 17 3.2 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.2.1 Bottom layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.2.2 Top layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.3 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.3.1 Imaging techniques . . . . . . . . . . . . . . . . . . . . . . . . 25 3.3.2 CCD camera . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.3.3 Imaging setup . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.3.4 Experimental control . . . . . . . . . . . . . . . . . . . . . . . 28 i Contents 4 Optical Dipole Trap: Results 31 4.1 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 4.1.1 Number of atoms in the optical dipole trap . . . . . . . . . . . 31 4.1.2 Number of atoms in the magneto-optical trap . . . . . . . . . 34 4.2 Atomic lifetime in the optical dipole trap . . . . . . . . . . . . . . . . 35 4.2.1 Experimental conditions . . . . . . . . . . . . . . . . . . . . . 35 4.2.2 Experimental results . . . . . . . . . . . . . . . . . . . . . . . 36 5 Optical Conveyor: Theory and Test Setups 41 5.1 Theory of atoms in moving optical lattices . . . . . . . . . . . . . . . 41 5.1.1 Trap parameters of a 1D optical lattice . . . . . . . . . . . . . 43 5.1.2 Hopping time between lattice sites . . . . . . . . . . . . . . . . 44 5.2 Frequency sweep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 5.2.1 Heating due to frequency sweep . . . . . . . . . . . . . . . . . 45 5.2.2 Block signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 5.2.3 Cosine signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 5.3 Experimental setups . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 5.3.1 Electronic setup . . . . . . . . . . . . . . . . . . . . . . . . . . 56 5.3.2 Optical conveyor prototype . . . . . . . . . . . . . . . . . . . . 56 6 Optical Conveyor: Test Results 58 6.1 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 6.2 Electronic setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 6.3 Optical conveyor prototype . . . . . . . . . . . . . . . . . . . . . . . . 61 7 Optical Conveyor: Final Design 65 7.1 Lower arm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 7.1.1 Displacement deviation . . . . . . . . . . . . . . . . . . . . . . 68 ii Contents 7.2 Probe level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 7.3 Upper arm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 7.4 Side view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 8 Conclusion 78 9 Acknowledgements 80 Appendix A 87Rb D hyperfine structure 82 2 Appendix B Optical conveyor design 83 iii Introduction 1 Introduction The interesting fields of nano-plasmonics and ultracold atoms have gained great pop- ularityoverthepastdecades. Thepopularityfornano-plasmonicscanbefoundinits possible applications which have the potential to revolutionize the telecommunica- tions industry [1,2]. The ultracold atoms have been a “hot” topic ever since the first Bose-Einstein condensate was achieved, which in 2001 was awarded with the Nobel Prize in Physics [3]. In 1997 the field was also awarded the Nobel Prize in Physics for the development of methods to cool and trap atoms using laser light [3,4]. In the field of nano-plasmonics, the interaction of light with metal structures on the nanoscale is studied [1,2]. The coupling between electromagnetic waves and collectivechargeoscillationsgiverisetosurfaceplasmonpolaritons,whichcanextend both inside the metal and on the interface of the metal [1,2,5]. What makes these plasmons so interesting is their ability to confine light to nanoscale regions, much smaller than the wavelength of light. By controlling light on the nanoscale, possible applications for nano-plasmonics range from biomedical sensors to plasmon-assisted solar energy conversion [5,6]. In the field of ultracold atoms, optical and magnetic fields are used to trap and cool atoms [3,7–9]. Temperatures associated with this cooling are typically in the range of 10(cid:181)K, allowing for great control and manipulation of the atoms [3]. Because of these low temperatures, atoms are described quantum mechanically, making ultra- cold atoms a beautiful field for the study of for example Bose-Einstein condensates, degenerate Fermi gasses and quantum information processing [3,10,11]. In our experiment, we will for the first time combine the interesting fields of nano- plasmonics and ultracold atoms by controlling and studying the coupling between atoms and light close to a plasmonic nanostructure sample. The atoms are placed close to the sample surface where they are trapped optical in the evanescent field of nanoplasmonic structures [12]. With this experimental setup we could for instance investigate what influence an atom has on the optical properties of a nanohole and vice versa. For the cooled atoms, isotope 87 of Rubidium (Rb) is used. This isotope has a positive scattering length which means that it is mutually repulsive at low tempera- tures [13]. 87Rb also has strong absorption lines at wavelengths that can be reached with commercially available diode lasers. We need a way to accurately and reproducibly position the atoms above the sample surface. The first step is to obtain a source of cold 87Rb. To this end, an ampul of natural Rb is heated to obtain a vapor of Rb. The atoms in this vapor are cooled and collected in a two-dimensionally shaped cigar using a 2D-magneto-optical trap [3,7–9,14–16]. From this trap, the atoms are pushed into a 3D-magneto-optical trap(3D-MOT).Inthistraptheatomsareconfinedinthreedimensionsatthecenter 1 Introduction of the vacuum chamber. The theory and experimental setup of the magneto-optical trap are discussed in Section 2. Thesamplecannotbeplacedunderthecloudofcooledatomsasthiswillblockoneof the lasers used to produce the 3D-MOT. The atoms are therefore displaced towards the sample. From the 3D-MOT, the atoms are therefore loaded into a far detuned optical dipole trap. Since the optical dipole trap is red-detuned, atoms are attracted to high intensities i.e. the focus of the optical dipole trap [17]. By displacing the focus of the optical dipole trap, the atoms will also be displaced. With the optical dipole trap the atoms are positioned above the sample. Using a vertically aligned optical conveyor, the atoms are moved down towards the sample surface. The optical conveyor consists of two counter-propagating beams with iden- tical wavelengths [18–20]. Due to their interference a standing wave is formed. Using red-detunedlasers, theatomsareattractedtointensitymaximai.e. theanti-nodesof the standing wave. By changing the frequency of one of the two counterpropagating beams slightly, the standing wave and thus the anti-nodes will move [18–20]. With theopticalconveyortheatomsarepositionedatλ/2abovethesamplesurface, where λ denotes the wavelength of the lasers. Withtheopticalconveyorwepositiontheatomsverticallywithrespecttothesample surface. However, we also want to be able to move the atoms in two-dimensions over the sample surface. This means we need to translate the optical conveyor in the plane of the sample while keeping the counterpropagating beams overlapped. This is the most unique and challenging part of the setup. The translation of a single laser beam in two dimensions is achieved using two mirror galvanometers. Each mirror galvanometer consists of a mirror mounted directly on the shaft of an electromotor. It thus rotates around its axis when applying a current to the motor. Appropriately placed lenses convert the change in angle to a change in position. By actively synchronizing the translation of the counterpropagating beams, they remain overlapped [21]. In this thesis we trap atoms in an optical dipole trap. We measure the lifetime of the atoms in this trap. In Section 3, the theory and experimental setup of the optical dipoletraparediscussed. Alsoincludedisthedescriptionoftheexperimentalcontrol. The results of the atom lifetime are discussed in Section 4. In addition, we have developed a method to displace the anti-nodes of the standing wave. We investigated this method of displacement both electronically and optically. InSection5,thetheoryandexperimentalsetupsoftheopticalconveyorarediscussed. A method to test the accuracy of the displacement of the anti-nodes and the results of these tests are discussed in Section 6. Furthermore, a detailed design of the final setup of the optical conveyor is presented in Section 7. This design includes the translation of the optical conveyor. The conclusions on both the optical dipole trap and optical conveyor are given in Section 8. 2