"Laser Cooling and Trapping of Atoms" Steven Chu Physics Department Stanford University, Stanford CA 94305

Transcription

1 Proceedings or the 6th International Symposium on Advanced Nuclear Energy Research -INNOVATIVE LASER TECHNOLOGIES IN NUCLEAR ENERGY- "Laser Cooling and Trapping of Atoms" Steven Chu Physics Department Stanford University, Stanford CA The basic ideas of laser cooling and atom trapping will be discussed. These techniques have applications in spectroscopy, metrology, nuclear physics, biophysics, geophysics, and polymer science. Keywords : Laser Cooling, Atom Trapping 1. Laser Cooling The field of laser cooling and trapping of atoms and other particles has blossomed during the last ten years and a short article such as this can not hope to adequately explain the major developments. This article will give a snapshot in time of the current state of the art in laser cooling, paying particular attention to those developments that this author believes will have general applicability. For a more thorough review of these advances, the interested reader is referred to a number of review articles. 1,2 The electromagnetic forces used to manipulate atoms are feeble: for example, the average kinetic energy of an atom at room temperature is a thousand times greater than depth of our deepest optical traps. The rapid progress in atom manipulation techniques is based on the fact that atoms can be readily cooled in "optical molasses" to microkelvin temperatures. As originally conceived by Hansch and Schawlow in , laser cooling is based on the Doppler effect for a two-level atom. If an atom is irradiated by counterpropagating laser beams tuned to the low frequency side of the atomic resonance, a moving atom will Doppler shift the beam opposing its motion to increase the scattering probability while the beam copropagating with the atom will be frequency shifted away from the resonance. Thus, there will be a net scattering of photons opposing the motion of the atom with each scattering event transferring an average momentum of #k to the atom. By surrounding the atom with three sets of counterpropagating beams along the x,y,and z axes, a drag force opposing the velocity of the atom can be generated. This configuration of laser beams was predicted* to cool 35-

2 atoms to a minimum temperature of k n T=/2l72 for a detuning Av of the laser equal to 172 in the limit of low laser intensity. The first three dimensional laser cooling of atoms was demonstrated by Chu, ct al., in The first measured temperatures and confinement times were consistent with the Doppler cooling, but a number of subsequent experimental findings were in strong disagreement with theoretical predictions including insensitivity to intensity imbalance between the counterpropagating beams, 6 and anomalously long lived molasses in which the storage times of the atoms increased by a factor of 50 if the laser beams were misaligned. 7 The biggest failing of two-level theory of optical molasses was uncovered by the NIST group. They found that sodium atoms could be cooled to temperatures at least six times colder than the predicted minimum temperature. 8 This discovery prompted Dalibard and Cohen-Tannoudji 9 and Chu and co-workers 10 to develop a new theory of laser cooling. The basic idea of the new theory is based on a combination of optical pumping effects, induced light shifts, and the motion of atoms in light fields with polarization gradients. Polarization gradients can not be avoided in any three dimensional configuration of light beams. As a simple case, consider two counterpropagating, linearly polarized laser beams with orthogonal polarizations. The two light fields will generate regions in space where the local polarization has <x + helicity followed by linearly polarization and then a helicity. An atom at rest in the <r + light field will optically pump into the lowest energy ground state, ni F = +F if the light is driving a F->F+1 transition, where F is the angular momentum state. If the atom then moves into a region where the local polarization of the light has changed to a polarization during a time faster than the optical pumping time, it will find itself in a superposition of magnetic substates with higher internal energy. The increase in the internal energy comes at the expense of the kinetic energy of the atom. If the atom remains in the same atomic state, it can recover the kinetic energy by returning to a region of space which has the original polarization. The irreversible step in the cooling occurs when the internal energy is dissipated by spontaneous emission in the optical pumping process. Experimentally, one finds that 3-dimensional polarization gradient cooling can cool atoms to temperatures on the order of k T~ (4fik)72M. Atoms can also be cooled below the photon recoil limit, defined as the temperature where the rms velocity is equal to /zk/m, the velocity change due to the recoil of a single photon. One technique, referred to as "velocity selective population trapping", depends on quantum mechanical interference to generate a superposition of atomic ground states that is not connected to the excited state. 11 Consider an atom with two ground states gi,pi> and jg 2,p 2 > with external momenta p, and p 2 connected to an excited state e> via light fields with quantum mechanical amplitudes A ( and A 2. In the case where A, = A 2 and the energies - 36-

3 of the two ground states are equal, the state (la / 2)(Jg,,p, > - jg 2,p 2 >) is "dark" in the sense that it is not be connected to the excited state by the radiation field. (This state is one of the new, re-diagonalized eigenstates of the system of jgi>, g 2 >, and e> after one turns on the light fields.) The debroglie frequencies of the two states w, 2=E, 2/fi are not equal since E, 2 is the total energy (internal plus external) of the atomic states unless the magnitude of the momenta are the same. Thus, the "dark" state will oscillate into the orthogonal state (la/2)( g,,p,> + g 2,p 2 >) which is strongly coupled to the radiation field. As the atoms spontaneously scatter many photons, they may randomly scatter into a state where p, -ftk and p 2 ~ +#k so that the dark state remains dark for an appreciable amount of time. The width of the momenta distribution decreases the longer an atom remains in the radiation field since only atoms with momenta approaching the exact non-coupled dark state will not oscillate into the coupled state. If one considers the width of each of the momenta peaks about +hk as a measure of the "temperature" of the atoms, it is possible to cool the atoms to temperatures less than the recoil energy. The second technique, which uses velocity selective stimulated Raman transitions to push atoms towards the v=0 state, 12 is similar to the frequency chirp method of cooling that was first employed to slow and stop atoms in an atomic beam. 13 Stimulated Raman transitions can be extremely Doppler sensitive since the width of the transition is determined by the time of the transition (alternately the Rabi frequency of the driving light fields) and not the lifetime of the excited state. 14 Thus, atoms cooled to optical molasses temperatures can be further cooled with Raman transitions by pushing the atoms towards the v=0 state. Consider an atom with ground states 11 > and 12 > and an excited state 13 >. Atoms are initially optically pumped into state l,p>. A stimulated Raman transition moves a subset of the atoms into state 12,p+2/»k> where the direction of k is chosen so as the reduce the velocity of the atom. All atoms with velocities vy^o can be pushed towards the v=0 state. An atom put into the j2> state is optically pumped back into the state j l,p'>. In the spontaneous emission process, the new momenta p' may remain nea. v=0 or could be kicked out to a higher momentum state. If the atom remains near v=0, subsequent Raman pulses are designed not the excite the atom but if it receives a spontaneous kick to away from the v=0 state, a subsequent Raman pulse will have another chance to push it back to the v=0 state. This cooling process was applied to sodium pre-cooled with optical molasses to 35 /xk. Roughly 50% of the atoms were cooled to 100 nanokelvin, corresponding to an increase in velocity phase space of an order of magnitude. 12 Both of these cooling techniques can be extended to two and three dimensions. -37-

4 2. Atom Trapping The first two techniques used to trap atoms exploit either the permanent magnetic dipole or induced dipole moment of atoms. Since div B = 0 and div E = 4irp = 0 in a region of space free of charge, one can not construct a static E or B field that has a maximum value in a region of space void of charges or currents. Magnetic trapping of atoms 15 is based on the idea that a local minimum in the a static magnetic can be made. If the magnetic moment is aligned anti-parallel to the magnetic field, the particle will be driven to the region of space where its energy W=-ji«B is minimized. As long as the particle moves slowly in the magnetic field, its quantum mechanical alignment with respect to the magnetic field is preserved. Magnetic traps have been used to confine spin-aligned hydrogen in order to avoid recombination of the hydrogen atoms into hydrogen molecules that occurs at the walls of a conventional cryostat. 16 Atoms in their ground state do not have permanent electric dipole moments, and trapping must be done through an induced dipole moment. A focused laser beam produces a time varying electric field maximum at the focal point, and as long as the driving electric field is below the resonances of the atom, the induced dipole moment will be in phase with the E field. For this type of trap, 17 the particle's energy is minimized by seeking regions of space where the electric field is strongest. If the laser field is tuned above the atomic resonance, the induced dipole moment will be aligned anti-parallel with the driving field and the atom will seek an E field minimum. Atomic traps based on a repulsive trampoline bowl of light and gravity have been proposed. 18 If the dipole trapping light is tuned far from resonance, the probability of scattering photons is greatly diminished and the trap begins to behave like a purely conservative potential well. 19 Sodium atoms with a resonant frequency at 589 nm have been recently trapped with light at 1.06 microns in our laboratory. We have also combined cooling via stimulated Raman transitions with this far off-resonant dipole trap and have cooled atoms along the axial direction of the trap to a velocity spread of 0.6fik/M. This method can be generalized to three dimensions in order to produce very cold, moderate density quantum gases. 2 One of the most widely used atom traps is a hybrid magneto-optical trap. 21 A weak magnetic field is used break the degeneracy of the Zeeman sub-levels of the atom. The slight shift in the energy levels cause counterpropagating a + and a laser beams to scatter more photons in the direction towards B=0. This type of trap can be thought of as mostly optical molasses with a magnetic field to tell the atoms where to collect. A great advantage of the magneto-optic trap is that the laser intensity in each of the 38-

5 3. Applications The first demonstration of laser cooling and trapping of atoms occurred less than a decade ago, and we have only begun to recognize the myriad of potential applications. With cold atoms, one can create an "atomic fountain" of atoms in which microkelvin atoms are tossed upwards in a ballistic trajectory. 26 The energy levels of these atoms, free of any perturbation except gravity, can be measured with great precision because of the long observation time available. Our second generation atomic fountain measured the hyperfine splitting of the cesium ground state with a short term stability that exceeds the atomic clocks maintained by standards laboratories by an order of magnitude. 27 An analysis of the systematic errors anticipated with an atomic fountain clock estimates that the absolute accuracy will be in the range of Ai>fa~ 10~ 15 to 10" Much more dramatic improvement is expected for fountain clocks that measure higher frequency transitions such as the 600 Ghz transition in Mg or near infrared or optical transitions in atoms such as magnesium, calcium and silver. 29 A number of different atom manipulation techniques have been devised which are analogous to optical components such as lens, mirrors, gratings. 30 Despite the proliferation of these methods, it is important to realize that "atom optics" can, in principle, be more powerful than photon optics. Photon optics is limited by a "brightness" conservation theorem beams is on the order of a few milliwatts/cm 2 so large volume traps can be made with low intensity lasers. Wieman and collaborators observed that the trap could collect atoms directly from a sealed, low-vapor pressure cell (~ 10" 8 torr). 22 The magneto-optic trap currently serves as a convenient starting point: as many as 4xl0 10 atoms at a density of ~ 10" atoms /cm 3 can be collected within 0.5 second. 23 Following a few millisecond exposure of polarization gradient molasses (accomplished by turning off the magnetic field) a dense sample of atoms at a temperature of k T=~(3-4flk) 2 /2M can then be used in a variety of experiments. Recently, 21 Na with a half-life of 22.5 sec was trapped in a magneto-optic trap Na decays by positron emission to its "mirror" nuclei 21 Ne and can be used in the precision test of the vector-axial structure of the weak interactions. The 21 Na was produced with the reaction 24 Mg (p,a) 21 Na using a 25 MeV proton beam. The sodium atoms effusing from an atomic beam oven were collimated by transverse optical molasses, slowed by an opposing laser beam in an inliomogeneous magnetic field 25 and brought to rest in a magneto-optic trap approximately 2 meters away from the production target. -39-

6 which states that optics can never be used to image a light source onto a image plane with an increase in the intensity of the light per unit bandwidth divided by the average divergence of the light. In contrast to photon optics, we can cool atoms and thereby increase the brightness of an atomic source. The first illustration of this idea was the "atomic funnel", where atoms from an atomic beam are captured by a magneto-optic trap, cooled and compressed in the trap, and then allowed to escape from the trap. 31 The phase space density of the initial atom beam was increased by four orders of magnitude in the first version of this "phase space compressor", and another factor of 10 4 is possible if one incorporates all the tricks in laser cooling that have been realized after this work was completed. Atom optics have been used to construct atom interferometers. 32 These devices are especially sensitive measuring devices when slow atoms are used in the atomic fountain geometry. For example, the acceleration due to gravity has been measured with a precision of 3 parts in 10" 8, 33 and we are working to improve the relative precision by three to four orders of magnitude and hope to measure g with an absolute precision of one part in Comparably sensitive gyroscopes and gravity gradiometers are being designed. A portable gravity meter or gravity gradiometer based on a compact Cs cell and diode lasers will find practical applications in oil and mineral exploration and the measurement of small land or ocean level changes that can not be resolved by the global positioning satellite system. Continued improvements in both cooling and trapping techniques may finally lead to the condensation of a dilute Bose gas or the creation of a degenerate Fermi gas. These novel quantum systems will provide unique opportunities to study the collective properties of fundamental quantum systems. Furthermore, the prospect of condensing a majority of the atoms into a single quantum state opens up the possibility of creating an atomic source with unprecedented properties analogous to a laser operating in the TEMQQ mode. Bose condensation can also be viewed as the theoretical limit to brightness enhancement of an incoherent atomic source. Optical traps based on a single focused laser beam can also trap particles between 0.02 to 10 jxm. 34 Individual living cells and even organelles inside a cell can be manipulated and viewed simultaneously. 35 Individual molecules can not be directly held by a laser beam at room temperature, but the attachment of micron-sized polystyrene "handles" to biological marcomolecules such as DNA 36 or actin or microtubule filaments 37 have enabled researchers to quantitatively measure the mechanical behavior of individual molecules. Fundamental questions in polymer physics he also been addressed with these single molecule manipulation techniques. 38 The ability to hold macroscopic particles with light opens up the possibility of 40-

7 suspending "target" materials in a manner such that the supporting mechanism (the light) will not interfere with the experiment. An example where a non-interacting support structure may be useful is the suspension of micro-spheres of tritium for hydrogen implosion experiments. The field of laser cooling and trapping is young. Much of the progress has been made within the last decade and there are no signs that progress in the field is waning. New trapping and cooling ideas will no doubt lead to exciting new applications. References: 1. For an excellent review of many aspects of laser cooling, see Laser Manipulation of Atoms and Ions, Proceedings of the International School of Physics "Enrico Fermi'', Course CXV1I1, eds. E. Arimondo, W.D. Phillips, and F. Strumia, (North Holland, Amsterdam, 1992). 2. For other reviews, see S. Chu, Science, 253, 861, (1991); Fundamental Systems in Quantum Optics, Les Ilouches, Session LII, 1990, eds. J. Dalibard, J.M. Raimond and J. Zinn-Justin (Elsevier Science Publishers, Amsterdam, 1992); C. Cohen-Tannoudji and W.D. Phillips, Phys. Today, 43, 33 (Oct 1990); the special issue of the J. Opt. Soc. Am. B6, eds. S. Chu and C. Wieman, (1989). 3. T.W. Hansch and A.L. Schawlow, Opt. Comm. 13, 68 (1975). 4. See, for example D. Wineland and W. Itano, Phys. Rev. A 20, 1521 (1979), and J. Gordon and A. Ashkin, Phys. Rev. A 21, 1606 (1980). 5. S. Chu, L. Hollberg, J.E. Bjorkholm, A. Cable, and A. Ashkin, Phys. Rev. Lett. 55, 48 (1985). 6. P.L. Gould, P.D. Lett, and W.D. Phillips, in Laser Spectroscopy VII, S. Svanberg and W. Persson, eds. (Springer-Verlag, Berlin, 1987). 7. S. Chu, M.G. Prentiss, A. Cable, and J.E. Bjorkholm, in Laser Spectroscopy VII, W. Persson and S. Svanberg, eds., (Springer-Verlag, Berlin, 1988) pp 64-67; Y. Shevy, D.S. Weiss, and S. Chu, in Spin Polarized Systems, S. Stringari, ed. (World Scientific, Singapore, 1989), pp P.D. Lett, R.N. Watts, C.I. Westbrook, W.D. Phillips, P.L. Gould, and H.J. Metcalf, Phys. Rev. Lett. 62, 1118 (1988). 9. J. Dalibard and C. Cohen-Tannoudji, J. Opt. Soc. Am. B 6, 2023, (1989). 10. P.J. Ungar, D.S. Weiss, E. Riis, S. Chu, J. Opt. Soc. Am. B 6, 2058, (1989). -41-

8 11. M. Kasevich and S. Chu, Phys. Rev. Lett. 69, 1741 (1992). 12. N. Davidson, H.J. Lee, M. Kasevich, and S. Chu, Phys. Rev. Lett. 72, 3158 (1994). 13. W. Ertmer, R. Blatt, J.L. Hall and M. Zhu, Phys. Rev. Lett. 54, 996 (1985). 14. M. Kasevich, D.S. Weiss, E. Riis, K. Moler, S. Kasapi, and S. Chu, Phys. Rev. Lett. 66,2297 (1991). 15. A. Migdall, J.V. Prodan, W.D. Phillips, T.H. Bergman, and H. Metcalf, Phys. Rev. Lett. 54, 2596 (1985). 16. H.F. Hess, G.P. Kochanski, J.M. Doyle, N. Masuhara, D. Kleppner, and T.J. Geytak, Phys. Rev. Lett. 59, 672 (1987). 17. S. Chu, J.E. Bjorkholm, A. Ashkin, and A. Cable, Phys. Rev. Lett. 57, 314 (1986). 18. M. Kasevich, D. Weiss, and S. Chu, Optics Lett. 15, 667 (1990); A.M. Steane, et al. Europhys. Lett. 14,231 (1991). 19. J.D. Miller, R.A. Cline, and D.J. Heinzen, Phys. Rev. A47, R4567 (1993). 20. H.J. Lee, C.S. Adams, N. Davidson, M. Kasevich, and S. Chu, to be published. 21. E.L. Raab, M. Prentiss, A.E. Cable, S. Chu, and D.E. Pritchard, Phys. Rev. Lett. 59, 2631 (1987). 22. C. Monroe, W. Swann, H. Robinson, and C.E. Wieman, Phys. Rev. Lett. 65, 1571 (1990). 23. K.E. Gibble, S. Kasapi, and S. Chu, Optics Letters 17, 526 (1992). 24. Z-T. Lu, era!., submitted to Phys. Rev. Letters, J. Prodan, et al., Phys. Rev. Lett. 54, 992 (1985). 26. M. Kasevich, E. Riis, S. Chu and R. Devoe, Phys. Rev. Lett. 63, 612 (1989); A. Clairon, et al., Europhys. Lett. 16, 165 (1991). 27. K. Gibble and S. Chu, Phys. Rev. Lett. 70, 1771 (1993). 28. K. Gibble and S. Chu, Metrolgia 29, 201 (1992). 29. For discussions of potential atom candidates for clocks, see F. Strumia, in Laser Science and Technology, eds. A.N. Chester, V.S. Letokov, and S. Martellucci, (Plenum Press, N.Y., 1988) pp ; J.L. Hall, M. Zhu, and P. Buch, J. Opt. Soc. Am. B6, 2194 (1989). 30. For a review, see C.S. Adams, M. Segel, and J. Mlynek, Phys. Rep. 240, 143 (1994). 31. E. Riis, D.S. Weiss, K. Moler and S. Chu, Phys. Rev. Lett. 64, 1658 (1990). -42-

9 Carnal and J. Mlynek, Phys. Rev. Lett. 66, 2689 (1991); D. Keith, C. Ekstrom, O. Turchette, and D. Pritchard, Phys. Rev. Lett. 66, 2693 (1991); F. Riehle, Th. Kisters, A. Witte, S. Helmeke, and Ch. Borde, Phys. Rev. Lett. 67, 177 (1991); M. Kasevich and S. Chu, Phys. Rev. Lett, 67, 181 (1991). 33. M. Kasevich and S. Chu, Appl. Phys. B 54, 321 (1992). 34. A. Ashkin. J.M. Dziedzic, J.E. Bjorkholm, and S. Chu, Opt. Lett. 11, 288 (1986). 35. A. Ashkin and J.M. Dziedzic, Science 253, 1517 (1987). 36. S. Chu, Science 253, 861 (1991). 37. K. Svoboda, C.F. Schmidt, B.J. Schnapp, and S.M. Block, Nature 365, 721 (1993); J. Finer, R.M. Simmons, and J. A. Spudich, Nature 368, 113 (1994). 38. T.T. Perkins, D.E. Smith and S. Chu, Science 264, 819 (1994); T.T. Perkins, S.R. Quake, D.E. Smith, and S. Chu, Science 264, 822 (1994). -43-