1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2: 

3: %% Trim Size: 9.75in x 6.5in

4: 

5: %% Text Area: 8in (include Runningheads) x 5in

6: 

7: %% ws-mplb.tex   :   10 September 2002

8: 

9: %% TeX file to use with ws-mplb.cls written in Latex2E.

10: 

11: %% The content, structure, format and layout of this style file is the

12: 

13: %% property of World Scientific Publishing Co. Pte. Ltd.

14: 

15: %% Copyright 1995, 2002 by World Scientific Publishing Co.

16: 

17: %% All rights are reserved.

18: 

19: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

20: 

21: 

22: 

23: \documentclass{ws-mplb}

24: 

25: 

26: 

27: \begin{document}

28: 

29: 

30: 

31: \markboth{F. Sorrentino, G. Ferrari, N. Poli, R. Drullinger and G.

32: M. Tino}{Laser Cooling of Atomic Strontium}

33: 

34: 

35: 

36: %%%%%%%%%%%%%%%%%%%%% Publisher's Area please ignore %%%%%%%%%%%%%%%

37: 

38: %

39: 

40: \catchline{}{}{}{}{}

41: 

42: %

43: 

44: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

45: 

46: 

47: 

48: \title{LASER COOLING AND TRAPPING OF ATOMIC STRONTIUM FOR ULTRACOLD ATOMS

49: PHYSICS, HIGH-PRECISION SPECTROSCOPY AND QUANTUM SENSORS}

50: 

51: 

52: 

53: \author{\footnotesize F. Sorrentino, G. Ferrari, N. Poli, R. Drullinger\footnote{on leave from NIST, 325 Broadway, Boulder, Colorado 80305} and G. M. Tino}

54: 

55: 

56: 

57: \address{Dipartimento di Fisica and LENS - Universit\`a di Firenze, INFN,

58: INFM,\\ I-50019 Sesto Fiorentino (FI), Italy\\

59: guglielmo.tino@fi.infn.it}

60: 

61: 

62: 

63: \maketitle

64: 

65: 

66: 

67: \begin{history}

68: 

69: \received{(received date)}

70: 

71: \revised{(revised date)}

72: 

73: %\accepted{(Day Month Year)}

74: 

75: %\comby{(xxxxxxxxxx)}

76: 

77: \end{history}

78: 

79: 

80: 

81: \begin{abstract}

82: 

83: This review describes the production of atomic strontium samples

84: at ultra-low temperature and at high phase-space density, and

85: their possible use for physical studies and applications. We

86: describe the process of loading a magneto-optical trap from an

87: atomic beam and preparing the sample for high precision

88: measurements. Particular emphasis is given to the applications of

89: ultracold Sr samples, spanning from optical frequency metrology to

90: force sensing at micrometer scale.

91: \end{abstract}

92: 

93: 

94: 

95: 

96: 

97: 

98: 

99: \section{Introduction}

100: 

101: \label{Introduction}

102: 

103: 

104: 

105: Recently, laser-cooled atomic strontium has been the subject of

106: active research in several fields spanning from all-optical

107: cooling towards quantum degeneracy for

108: bosonic,\cite{KatoriMOT1999,Ferrari2006} and

109: fermionic\cite{Mukaiyama2003} isotopes, cooling

110: physics,\cite{Xu2003,Loftus2004} continuous atom

111: laser,\cite{Katori2001} detection of ultra-narrow

112: transitions,\cite{Takamoto2003,Courtillotclock,Ferrari2003,Ido2005}

113: multiple scattering,\cite{Bidel2002} and collisional

114: theory\cite{Derevianko2003}.

115: 

116: 

117: 

118: Much of that interest relies on the features of the electronic

119: level structure of alkali-earth atoms, that make them ideal

120: systems for laser manipulation and for the realization of quantum

121: devices. Among the alkali-earth metals, strontium summarizes most

122: of the useful properties both for the preparation of ultracold

123: samples and for applications.

124: 

125: 

126: 

127: The ground-state Sr atom presents a strong and quasi-closed

128: optical transition well suited for efficient trapping in

129: magneto-optical traps (MOT's) from thermal

130: samples,\cite{Raab1987,Kurosu1990} and a multiplet of weak

131: intercombination transitions with large interest in laser cooling

132: and optical metrology. All of these transitions are easily

133: accessible with compact solid-state laser sources, as described in

134: \ref{ExperimentalSetup}. Second-stage cooling on the narrow

135: $^1$S$_0$-$^3$P$_1$ intercombination line was proven to be an

136: efficient method to reduce the sample temperature down to the

137: photon recoil limit.\cite{KatoriMOT1999}

138: 

139: 

140: 

141: The absence of a nuclear spin in all of the bosonic isotopes

142: greatly simplifies the electronic energy spectrum with respect to

143: the already simple structure of alkali atoms, and allows a direct

144: verification of theories in the field of light

145: scattering\cite{Chanelière2004} and laser

146: cooling.\cite{Loftus2004}  Moreover, the lack of nuclear spin

147: results in minimum sensitivity to stray magnetic fields: this has

148: important consequences on high-precision measurements, as

149: discussed in sections  \ref{Microsensor} and \ref{1S3PFreqMeasure}.

150: 

151: 

152: 

153: Interatomic collisions may represent an important source of

154: perturbations in atomic quantum devices. In this respect, the

155: $^{88}$Sr atom exhibits excellent features due to its remarkably

156: small collisional cross-section, resulting in the longest

157: coherence time for Bloch oscillations

158: observed so far.\cite{FerrariBloch}

159: 

160: 

161: 

162: As a consequence of their zero magnetic moment, ground-state

163: bosonic Sr isotopes cannot be magnetically trapped. However,

164: optical dipole trapping with far-off resonant laser fields

165: provides an effective method for trapping in conservative

166: potentials. By properly choosing the wavelength of the trapping

167: laser field, it is possible to design optical potentials that

168: shift equally levels belonging to the singlet and triplet

169: manifolds, hence avoiding perturbations to a given

170: intercombination line.\cite{KatoriFORT1999} This feature is

171: the basis of laser cooling in dipole traps and allows one to reach

172: a temperature close to the recoil limit with a phase-space density

173: close to quantum degeneracy both with bosonic and fermionic

174: isotopes.\cite{Ido2000,Mukaiyama2003,Ferrari2006}

175: 

176: 

177: 

178: Several groups are presently working on laser cooled strontium.

179: The Tokyo group first demonstrated the double-stage optical

180: cooling down to sub-$\mu$K temperatures;\cite{KatoriMOT1999} they

181: also proposed and realized a Sr ``optical lattice

182: clock'',\cite{Takamoto2005} and made pioneering studies on

183: electric microtraps for strontium.\cite{Katori2006}  The JILA

184: group has made interesting studies on cooling

185: physics,\cite{Xu2003,Loftus2004} and on metrological

186: applications.\cite{Ido2005} Our group is concerned both with

187: ultracold physics studies,\cite{Ferrari2006} and with applications

188: to optical metrology\cite{Ferrari2003} and quantum

189: devices.\cite{FerrariBloch} The BNM-SYRTE group is mainly

190: concerned with metrological

191: applications,\cite{Courtillotclock,Lemonde2005,Targat2006} the Houston

192: group with ultracold Sr physics,\cite{Nagel2005,Mickelson2005} while the

193: Nice group has studied multiple scattering of light from cold Sr

194: samples,\cite{Chanelière2004} and analyzed several details of the

195: Sr cooling mechanisms.\cite{Chanelière2005}

196: 

197: 

198: 

199: 

200: 

201: From the experimental point of view, some of the techniques and

202: solutions involved in atomic strontium cooling are specific to

203: this atom and to some extent uncommon in the field of atomic

204: physics and laser cooling. In this paper we give a detailed

205: description of how to prepare an ultra-cold strontium sample which

206: is well suited for high-precision spectroscopy, the study of

207: quantum degenerate gases and quantum sensors. The presentation has

208: the following structure:

209: 

210: 

211: 

212: \begin{itemize}

213: \item section \ref{Strontiumatom} summarizes the properties of the

214: strontium atom, \item section \ref{MOT} describes the process of

215: laser cooling of strontium starting from the slowing of a thermal

216: beam, to the cooling down to the photon recoil limit and trapping

217: in a conservative potential, \item section

218: \ref{CollisionalMeasure} deals with the study of ultracold

219: collisions on the ground-state bosonic Sr isotopes, \item in

220: section \ref{SrBEC} we show some recent advances towards the

221: realization of a Bose-Einstein condensate (BEC) of strontium,

222: \item in section \ref{Microsensor} we present the use of ultracold

223: Sr atoms as force sensors at micrometric scale, \item in section

224: \ref{1S3PFreqMeasure} we review the recent frequency measurements

225: on the Sr intercombination transitions with application to optical

226: metrology, \item in \ref{ExperimentalSetup} we give a detailed

227: description of the experimental setup employed in laser cooling of

228: strontium, namely the vacuum apparatus and the laser sources.

229: \end{itemize}

230: 

231: 

232: 

233: 

234: 

235: 

236: 

237: \section{Properties of the strontium atom}

238: \label{Strontiumatom}

239: 

240: 

241: The main properties of atomic strontium are common to almost all

242: alkaline-earth metals. At ambient temperature it appears as a

243: metal. Its vapor pressure is quite low, and it reaches 1 mTorr at

244: 1000\,Kelvin. The Sr atom is rather reactive: it forms compounds

245: with oxygen, nitrogen, water and silicates, while it is inert

246: against sapphire. Thus in working with Sr vapors it is common to

247: employ sapphire windows.\cite{Neuman1995}

248: 

249: 

250: 

251: Strontium has four natural isotopes, whose properties are listed

252: in tab. 1. The bosonic (even) isotopes have zero nuclear spin,

253: thus they are perfect scalar particles in the $J=0$ states. This

254: has important consequences for applications to optical metrology

255: and quantum sensors (see sections \ref{Microsensor} and

256: \ref{1S3PFreqMeasure}).

257: 

258: 

259: 

260: 

261: 

262: \begin{table}[pt]

263: \tbl{Natural Sr isotopes (NIST data)} {\begin{tabular}{@{}cccc@{}}

264: \toprule Isotope & Relative atomic mass & Relative abundance &

265: Nuclear spin  \\ \colrule

266: $^{88}$Sr &   87.905 6143(24) & \hphantom{0}82.58(1)\% & 0 \\

267: $^{86}$Sr &  85.909 2624(24) & \hphantom{00}9.86(1)\% & 0 \\

268: $^{87}$Sr &  86.908 8793(24) & \hphantom{00}7.00(1)\% & 9/2 \\

269: $^{84}$Sr &  83.913 425(4) & \hphantom{000}0.56(1)\% & 0\\

270: \botrule

271: \end{tabular}}

272: \end{table}

273: 

274: 

275: 

276: 

277: 

278: Concerning the electronic level structure, due to the presence of

279: two electrons in the outer shell the atomic states can be grouped

280: into two separate classes: singlets and triplets. Since the

281: spin-orbit interaction breaks the spin symmetry, intercombination

282: transitions between singlet and triplet states are weakly

283: allowed.\cite{LibroFisicaAtomica} A simplified scheme of relevant

284: Sr levels and transitions is shown in figure \ref{LevelScheme}.

285: 

286: 

287: 

288: \begin{figure}[t]

289: \centerline{\psfig{file=fig1,width=8cm}} \vspace*{8pt}

290: \caption{Electronic level structure of atomic strontium. The

291: transitions relevant for laser cooling and optical manipulation

292: are indicated as well as their linewidths.}\label{LevelScheme}

293: \end{figure}

294: 

295: 

296: 

297: 

298: 

299: The strong $^1$S$_0$-$^1$P$_1$ transition at 461 nm has a natural

300: width of 32 MHz, and it is used for laser cooling and

301: trapping.\cite{Raab1987,Kurosu1990} Such transition is not

302: perfectly closed, due to a small leakage towards the 4d\,$^1$D$_2$

303: state (branching ratio $\sim 10^{-5}$). The direct

304: $^1$D$_2$-$^1$S$_0$ decay channel is forbidden in dipole

305: approximation, and atoms from the $^1$D$_2$ basically decay

306: towards the 5p\,$^3$P$_2$ (branching ratio 33\,\%) and

307: 5p\,$^3$P$_1$ (branching ratio 67\,\%) states.

308: 

309: 

310: 

311: The line strength of the three 5s$^2$\,$^1$S-5s5p\,$^3$P

312: intercombination transitions range from the relatively high value

313: (7.6\,kHz) of the 689\,nm $^1$S$_0$-$^3$P$_1$ line to the

314: virtually zero value of the $^1$S$_0$-$^1$P$_0$ line for the

315: even isotopes in the absence of external fields. In $^{87}$Sr the

316:  0-0 line has a natural

317: linewidth of

318: $\sim 1$\,mHz due to hyperfine mixing.

319: 

320: 

321: 

322: 

323: 

324: 

325: \section{Laser cooling and trapping}

326: \label{MOT}

327: 

328: 

329: A magneto-optical trap operated on the

330: $^1$S$_0$-$^1$P$_1$ line requires the use of a blue laser source,

331: that can be realized through second-harmonic generation from a

332: semiconductor laser, as described in section \ref{BlueLaser}.

333: 

334: 

335: 

336: The final temperature in such a ``blue MOT'' is limited to few mK

337: by the linewidth of the $^1$S$_0$-$^1$P$_1$ transition: the ground

338: state of alkaline-earth even isotopes has no hyperfine or Zeeman structure, and this

339: prevents the application of sub-Doppler cooling techniques.

340: However, the presence of narrow intercombination transitions

341: allows efficient second-stage Doppler cooling down to sub-recoil

342: temperatures, as described in \ref{RedMOT}. In $^{87}$Sr, the

343: ground-state hyperfine structure offers the chance for sub-Doppler

344: cooling on the

345: $^1$S$_0$-$^1$P$_1$ transition, as demonstrated by the JILA

346: group.\cite{Xu2003} Alternatively, it is possible to apply sub-Doppler

347: cooling techniques to atoms trapped in the metastable $^3$P$_2$ level, as

348: already demonstrated on Ca.\cite{Grunert}

349: 

350: 

351: 

352: In this section we illustrate the Sr cooling and trapping in

353: detail, as performed in our laboratory. We start by preparing a mK

354: sample in the blue MOT from a Zeeman-slowed atomic beam, using optical pumping to recycle atoms

355: from the metastable $^3$P$_2$ level. Then we transfer the atoms to

356: a ``red MOT'' operated on the $^1$S$_0$-$^3$P$_1$ intercombination

357: line, where we cool them down to $\mu$K temperatures. Our

358: apparatus allows us to trap different Sr isotopes simultaneously,

359: as described in \ref{mixture}. The final step consists in loading

360: an optical dipole trap.

361: 

362: 

363: 

364: 

365: 

366: 

367: \subsection{Zeeman slowing and atomic beam collimation} \label{BeamCollimation}

368: 

369: 

370: The sequence for cooling and trapping on the 461\,nm resonance

371: line follows standard techniques, as already reported by other

372: groups.\cite{KatoriMOT1999,Xu2003,Kurosu1990,Oates1999} We slow

373: the thermal atomic beam to few tens of m/s in a 30-cm long

374: Zeeman-slower\cite{Prodan1982} based on a two-stages tapered coil

375: with a zero crossing magnetic field, and a counter-propagating

376: laser beam frequency shifted by 480\,MHz to the red of the

377: $^1$S$_0$-$^1$P$_1$ transition. The beam, with typical power of

378: 40\,mW, has a 7\,mm $1/e^2$ radius at the MOT and it is focused on

379: the capillaries (see appendix \ref{VacuumSystem}). The distance

380: between the capillaries and the MOT region is 100\,cm.

381: 

382: 

383: The atomic beam brightness can be increased with a 2-D transverse

384: cooling stage before the Zeeman slower.\cite{Shimizu,Rasel1999}

385: This improves the MOT loading both by increasing the atomic flux

386: coupled into the differential pumping tube (see

387: \ref{VacuumSystem}), and by effectively reducing the final

388: diameter of the atomic beam after the Zeeman slower.

389: 

390: 

391: 

392: We implement a 2D optical molasses with two 461\,nm beams sent

393: perpendicularly to the atomic beam, in multipass geometry on two

394: orthogonal planes. The beams are red detuned by 20\,MHz with

395: respect to the $^1$S$_0$-$^1$P$_1$ resonance. At given optical power, the

396: multi-pass geometry improves the transverse cooling efficiency  by

397: increasing the length of the interaction region. In the present setup we

398: use beams with

399: $1/e^2$ diameter of 2.5\,mm$^2$ bouncing about 12 times to cover an

400: interaction length of 4\,cm. The interaction starts 15\,cm after the

401: capillaries (see

402: \ref{VacuumSystem}). In that region the beam diameter has already

403: reached about 11\,mm.

404: 

405: 

406: 

407: Without beam collimation we can couple into the differential

408: pumping tube (placed 24\,cm far from collimation region) about

409: 10\,\% of the total atomic flux. With the 2D molasses we increase

410: the coupling by a factor of 4 in optimal conditions (laser detuning

411: $\sim 20$\,MHz, optical power $\sim 20$\,mW). This simple scheme can also

412: be used to deflect the atomic beam.

413: 

414: 

415: 

416: 

417: 

418: 

419: 

420: 

421: \subsection{Blue MOT} \label{BlueMOT}

422: 

423: 

424: 

425: The slowed atoms are then cooled and trapped in a MOT operated

426:  on the $^1$S$_0$-$^1$P$_1$ transition 40\,MHz to the red

427: with respect to resonance. Three retro-reflected beams with a

428: 5\,mm $1/e^2$ radius cross in the MOT region with almost mutually

429: orthogonal directions. The vertical beam is collinear with the

430: magnetic quadrupole axis of an anti-Helmholz pair of coils. The

431: field gradient at the quadrupole center is 50\,Gauss/cm. Taking

432: into account the coupling efficiency of the blue laser source into

433: the fiber (70\,\%) and the efficiency of the AOMs ($75 \div

434: 80$\,\%), the remaining 120\,mW are split into the three channels:

435: typical power values are respectively 60\,mW for MOT beams, 40\,mW

436: for Zeeman slower and 20\,mW for transverse cooling beams. The

437: total 461\,nm light incident on the MOT region amounts to $\sim

438: 90$\,mW/cm$^2$.

439: 

440: 

441: 

442: 

443: 

444: We have a rough estimate of the number of trapped atoms by

445: collecting the 461\,nm fluorescence on a calibrated photodiode. By

446: changing the blue laser detuning we are able to separately trap

447: the four natural Sr isotopes, as shown in figure

448: \ref{IsotopeFluorescence}. For a more accurate measure of the atom

449: number we perform absorption imaging on a CCD with a resonant

450: 461\,nm horizontally propagating probe beam. We measure the sample

451: temperature with standard time-of-flight technique.

452: 

453: 

454: 

455: 

456: 

457: \begin{figure}[t]

458: \centerline{\psfig{file=MOTfluor,width=7cm}} %\vspace*{8pt}

459: \caption{Fluorescence of blue MOT as a function of laser detuning.

460: Resonances corresponding to the different isotopes are visible.

461: }\label{IsotopeFluorescence}

462: \end{figure}

463: 

464: 

465: 

466: 

467: 

468: The lowest temperatures measured in the blue MOT are typically

469: higher than the Doppler limit by at least a factor two. The Nice

470: group has shown that the extra-heating mechanisms causing such

471: discrepancy can be explained in terms of intensity fluctuations in the MOT

472: laser beams.\cite{Chanelière2005}

473: 

474: 

475: 

476: 

477: 

478: 

479: 

480: \subsection{Optical repumping from metastable states} \label{OpticalPumping}

481: 

482: 

483: 

484: 

485: 

486: The $^1$S$_0$-$^1$P$_1$ transition used in the first stages of

487: cooling and trapping is not perfectly closed due to the decay

488: channel of the 5s5p\,$^1$P$_1$ towards the 5s4d\,$^1$D$_2$ state,

489: that has a branching ratio of $2 \times 10^{-5}$. Atoms in the

490: latter state may decay back to the ground state trough the

491: 5s5p\,$^3$P$_1$ within less than 1 \,ms, or may decay to the

492: metastable 5s5p\,$^3$P$_2$ and be lost (see figure

493: \ref{LevelScheme}). Under typical MOT conditions this process

494: limits the MOT lifetime to few tens of milliseconds.

495: 

496: 

497: 

498: Some groups already circumvented this limitation by optical

499: pumping atoms from the metastable 5s5p\,$^3$P$_2$ to the ground

500: state via the 5s6s\,$^3$S$_1$ state with 707\,nm

501: light.\cite{Vogel1999} Since the 5s6s\,$^3$S$_1$ state is also

502: coupled to the metastable 5s5p\,$^3$P$_0$ an additional laser at

503: 679 nm is necessary in this scheme. To reduce the number of

504: repumping lasers we use a different approach which involves the

505: 5s5d\,$^3$D$_2$ state and requires one single laser at 497 nm.

506: Concerning the optical pumping, the $^3$D$_2$ state is coupled

507: basically to the $^3$P$_2$ and $^3$P$_1$ states, then efficient

508: pumping is insured within few absorption cycles. To this purpose

509: during the loading we illuminate the blue MOT with light at

510: 497\,nm, produced by the source described in section

511: \ref{BlueLaser}, which is kept on resonance with the repumping

512: transition. Figure \ref{BlueMOTFluorescence} shows the effect of

513: the repumping field on the $^{88}$Sr MOT loading in different

514: conditions. Without the repumper the MOT loading time ($1/e$) is

515: 10.2 ms regardless to the presence of the atomic beam transverse

516: cooling, and correspondingly to the flux of atoms captured. In

517: presence of the repumper we observe a 0.24 s lifetime when few

518: atoms are present in the MOT, and a decreasing lifetime down to

519: 0.11 s at full MOT charging. This reduction in lifetime is

520: explained in terms of light assisted collision.\cite{Vogel1999}

521: 

522: 

523: 

524: 

525: 

526: \begin{figure}[t]

527: \centerline{\psfig{file=MOTChargingBlue,width=9cm}} %\vspace*{8pt}

528: \caption{Fluorescence of the charging MOT. Solid line: with

529: transverse cooling and repumper. Dash: with repumper. Dot: with

530: transverse cooling. Dash-dot: without repumper or transverse

531: cooling.}\label{BlueMOTFluorescence}

532: \end{figure}

533: 

534: 

535: 

536: 

537: 

538: Due to the difference in the natural isotope abundance (see

539: section \ref{Strontiumatom}), the loading flux into the MOT varies

540: correspondingly. In typical conditions, i.e. when operating with

541: the transverse cooling and optical repumping from the metastable

542: state, and with the Sr oven kept at $\sim 700$\,K, the

543: steady-state blue MOT population amounts to $\sim 10^8$ atoms for

544: $^{88}$Sr and $\sim 10^7$ atoms for $^{86}$Sr.

545: 

546: 

547: 

548: 

549: 

550: 

551: 

552: \subsection{Red MOT} \label{RedMOT}

553: 

554: 

555: 

556: The Tokyo group first realized a strontium MOT operated on the

557: $^1$S$_0$-$^3$P$_1$ line.\cite{KatoriMOT1999} Such system has been

558: exhaustively studied by the JILA group, both theoretically and

559: experimentally.\cite{Loftus2004} The dynamics of laser cooling on

560: narrow transitions presents several interesting features. Unlike

561: in the case of ordinary magneto-optical traps, for the

562: $^1$S$_0$-$^3$P$_1$ transition in alkaline-earth atoms the natural

563: linewidth $\Gamma$ is of the order of the recoil frequency shift

564: $\Gamma_R$. In such conditions both mechanical and thermodynamical

565: MOT properties cannot be explained by the ordinary semiclassical

566: Doppler theory of laser cooling. In particular, the role of

567: gravity becomes non negligible, and the atomic temperature can be

568: lower than the recoil limit $T_R = \frac{\hbar \Gamma_R}{k_B}$.

569: 

570: 

571: 

572: When the laser detuning $\delta$ is negative and larger than the

573: power-broadened linewidth $\Gamma\sqrt{1+\frac{I}{I_{sat}}}$, the

574: atoms interact with the MOT laser beams only on a thin ellipsoidal

575: shell, corresponding to the surface where the laser frequency

576: offset compensates for the Zeeman shift. The maximum radiation

577: force is only one order of magnitude larger than gravity. As a

578: result, the atoms sag on the bottom of the ellipsoid, as

579: shown in the inset of figure \ref{RedMOTtemperature}. In such

580: conditions the atomic temperature is mainly determined by the

581: interaction with the upward propagating vertical MOT beam. With

582: typical laser intensities of $I \simeq 10^1 \div 10^4 I_{sat}$,

583: measured temperatures are in good agreement with a semiclassical

584: model, and can be expressed as

585: 

586: 

587: 

588: \begin{equation}

589: T=\frac{\hbar \Gamma \sqrt{1+\frac{I}{I_{sat}}}}{2 k_B},

590: \end{equation}

591: which is lower than the Doppler temperature at the same detuning.

592: At low laser intensity ($I \approx I_{sat}$), the cooling

593: mechanism becomes fully quantum-mechanical, and the minimal

594: attainable temperature is $T_R \over 2$.\cite{Loftus2004} Figure

595: \ref{RedMOTtemperature} shows the dependence of the MOT

596: temperature on the 689\,nm laser intensity, as measured in our

597: laboratory.

598: 

599: 

600: 

601: The JILA group also studied the case of positive laser detuning,

602: revealing the appearance of novel striking phenomena such as

603: momentum-space crystals. They showed that driving the atomic

604: system with 689\,nm light blue detuned from the

605: $^1$S$_0$-$^3$P$_1$ resonance may result in a periodic pattern in

606: the atomic velocity distribution.\cite{Loftus2004}

607: 

608: 

609: 

610: Second-stage cooling of the odd Sr isotope is more complex,

611: due to the hyperfine structure of both ground and excited states.

612: The Tokyo group has shown that using two lasers at 689\,nm

613: it is possible to cool and trap $^{87}$Sr atoms at phase-space

614: densities close to the Fermi degeneracy.\cite{KatoriFermion}

615: 

616: 

617: 

618: We here illustrate the red MOT for bosonic Sr isotopes as realized

619: in our laboratory. The 689\,nm light for the MOT beam (14\,mW) is

620: provided by a slave laser injection-locked with light coming from

621: the stable master laser described in \ref{RedLaser}. The beam spot

622: is then enlarged to the same radius as the 461\,nm MOT beams and

623: overlapped to the blue beams with a dichroic mirror. After that

624: mirror, the two wavelengths share the same broad-band optics:

625: polarizing beam splitters, mirrors and waveplates.

626: 

627: 

628: 

629: The linewidth of the intercombination transition is not sufficient

630: to cover the Doppler broadening corresponding to the velocity

631: distribution of the sample trapped in the blue MOT. An efficient

632: solution consists in broadening the spectrum of the 689\,nm laser

633: field.\cite{KatoriMOT1999} On the contrary, groups working on

634: other alkaline-earth atoms (i.e. Ca and Mg) employ a resonant

635: coupling to some higher level to quench the $^3$P$_1$

636: lifetime,\cite{Binnewies2001} thus increasing the effective

637: strength of the intercombination transition, as the maximum

638: radiation force would be otherwise lower than gravity.

639: 

640: We add a frequency modulation at 50\,kHz in the first 200\,ms. The

641: total frequency span is 2\,MHz, corresponding to 40 sidebands,

642: with an intensity of 120\,$\mu$W/cm$^2$ per sideband (the

643: saturation intensity is $I_{sat} = 3$\,$\mu$W/cm$^2$). The central

644: frequency is red detuned by 1 MHz with respect to resonance. At

645: the end of the FM recapture, we normally obtain a cloud at a

646: temperature of about 20\,$\mu$K. After that we remove the

647: frequency modulation, set a fixed red detuning from

648: $^1$S$_0$-$^3$P$_1$ transition and reduce the beam intensity in

649: the last 60\,ms of cooling. Working at 350\,kHz below resonance

650: and reducing the total light intensity on the MOT to

651: 70\,$\mu$W/cm$^2$, we can then transfer about 10\,\% of the atoms

652: from the blue MOT to the red MOT at temperatures below 1\,$\mu$K.

653: 

654: 

655: 

656: \begin{figure}[t]

657: \centerline{\psfig{file=redMOTtemp,width=9cm}} %\vspace*{8pt}

658: \caption{Measured atomic temperature and population of the red MOT

659: as a function of the laser intensity with our apparatus. The inset

660: shows in-situ and free-fall images of the trapped

661: atoms.}\label{RedMOTtemperature}

662: \end{figure}

663: 

664: 

665: 

666: 

667: 

668: 

669: 

670: \subsection{Cooling and trapping isotopic mixtures} \label{mixture}

671: 

672: 

673: Among the experiments on ultracold atoms, much work is being

674: concentrated on the study of mixtures of different atomic

675: species\cite{Santos1995,Modugno2001,Murdrich2002} or different

676: isotopes of the same

677: species.\cite{Mewesl999,Loftus2001,Honda2002,Stas2004} Mixtures

678: offer a way to exploit collisional physics not applicable in

679: single species samples.\cite{Schreck2001} They also offer

680: additional degrees of freedom, such as sympathetic cooling, in

681: order to achieve the degenerate quantum regime with atoms for

682: which evaporative cooling is not efficient.\cite{Khaykovich2001}

683: 

684: 

685: 

686: For simultaneous trapping of different isotopes previous

687: experiments employed laser sources delivering the necessary

688: frequency components for each isotope

689: involved.\cite{Suptitz1994,Mewesl999} In the case of the strontium

690: blue MOT, this approach may be difficult to apply because of the

691: complexity of the laser sources and the limited laser power. An

692: alterative solution is presented by the use of the magnetically

693: trapped $^3$P$_2$  state as a dark atom

694: reservoir.\cite{Stuhler2001,Nagel2003,Poli2005} During the blue

695: MOT phase without the repumper, the small loss channel of the

696: excited $^1$P$_1$ state towards the metastable $^3$P$_2$ state

697: provides a continuous loading of atoms into the potential given by

698: the MOT's magnetic quadrupole. Figure \ref{BlueIsotopeLifetime}

699: reports our lifetime measurements on the magnetically trapped

700: metastable isotopes. By using the same blue laser source, one can

701: sequentially load different isotopes into the magnetic potential

702: by just stepping the laser frequency to the different resonances.

703: 

704: 

705: 

706: We typically start by collecting one isotope (say $^{86}$Sr) for a

707: few seconds, then we tune the blue laser on resonance to the other

708: isotope (say $^{88}$Sr). Once the isotopic mixture is prepared in

709: the $^3$P$_2$ state, the blue light is switched off, and the FM

710: red MOT is switched on as well as the repumping beam. The isotopic

711: shift on the repumping transition is smaller than the resonance

712: width of the $^3$P$_2$-$^3$D$_2$ transition observed on the blue

713: MOT fluorescence, which results in efficient, simultaneous optical

714: pumping of $^{88}$Sr and $^{86}$Sr on a time scale short with

715: respect to the red MOT capture time. The loading of a single

716: isotope into the magnetic potential was already described by Nagel

717: et al.,\cite{Nagel2003}  and we did not observe significant

718: differences in the behavior when loading two isotopes. Figure

719: \ref{BlueIsotopeLifetime} shows the measurement of the lifetime

720: for each isotope, both when individually trapped and in the

721: presence of the blue MOT working on the other isotope. The

722: measured lifetime values are all of the order of 5\,s, close to

723: the background pressure limited lifetime of 7\,s.

724: 

725: 

726: 

727: 

728: 

729: \begin{figure}[t]

730: \centerline{\psfig{file=lifetime3P2,width=7cm}} %\vspace*{8pt}

731: \caption{Decay of the $^3$P$_2$ state when trapped in a 56\,G/cm magnetic quadrupole. Circles

732: ($^{86}$Sr) and squares $^{88}$Sr refer  to data taken for the individual isotopes with the blue

733: MOT switched off; diamonds ($^{86}$Sr) and triangles ($^{88}$Sr) with the blue MOT operating on

734: the undetected isotope. The measurements are taken after red MOT recapture. Reprinted figure with

735: permission from N. Poli et al., Phys. Rev. A 71, 061403(R) (2005). Copyright (2005) by the

736: American Physical Society.}\label{BlueIsotopeLifetime}

737: \end{figure}

738: 

739: 

740: 

741: The laser source for the operation of the two-isotope red MOT is

742: composed of two slave lasers injected from the same frequency

743: stabilized master with a frequency offset corresponding to the

744: isotopic shift of 163 817.3\,kHz.\cite{Ferrari2003} Subsequently,

745: the frequency and intensity of the two beams are controlled by

746: double pass AOMs driven by the same RF, the beams are superimposed

747: on a polarizing beam splitter, and then they are overlapped to the

748: blue MOT beams as described previously. Comparing the two-isotope

749: red MOT with the single isotope one, with the same atom numbers we

750: do not observe any effect in the transfer efficiency and final

751: temperature due to the presence of the second isotope. In this

752: way, we obtain samples with up to 10$^7$ (10$^6$) atoms of

753: $^{88}$Sr ($^{86}$Sr) at a temperature 2\,$\mu$K (1\,$\mu$K). We

754: attribute the difference in the loading to the natural abundances

755: and to minor differences in the red MOT parameters. By varying the

756: order of loading and the loading times of the two isotopes we can

757: vary almost arbitrarily the final ratio of populations.

758: 

759: 

760: 

761: The atom number and temperature are measured independently on the

762: two isotopes by absorption imaging with the resonant 461\,nm probe

763: beam, and the contribution of the non-resonant isotope is taken

764: into account.

765: 

766: 

767: 

768: 

769: 

770: 

771: 

772: \subsection{Optical dipole trap} \label{FORT}

773: 

774: 

775: 

776: In magneto-optical traps the atomic temperature and lifetime are

777: fundamentally limited by resonant photon scattering and

778: light-assisted collisions. In all experiments requiring long

779: storage times or ultra-low temperatures, it is convenient to

780: transfer the atoms into a conservative trap. The ground-state even

781: isotopes of alkaline-earth atoms cannot be magnetically trapped,

782: due to the absence of Zeeman structure. Though the Tokyo group has

783: recently demostrated a clever scheme for Sr trapping with AC elecric

784: fields,\cite{Katori2006}  in most cases the best choice consists in

785: optical dipole traps. Moreover on two-electrons atoms it is possible to

786: apply light-shift cancellation techniques to the intercombination

787: transitions,\cite{KatoriFORT1999} opening the way to optical spectroscopy

788: with ultimate accuracy. However, the optical dipole trap is widely

789: employed with magnetic atoms as well,\cite{opticaltrapping} since it

790: generally produces a stronger confinement than magnetic traps, up to the

791: Lamb-Dicke regime in optical lattices,\cite{Dicke} and permits trapping

792: in all the magnetic sub-levels.

793: 

794: 

795: 

796: In an optical dipole trap the confining force originates from the

797: energy level gradient produced by the intensity-dependent light

798: shift. The energy level shift $\Delta E$ of an atom in an optical

799: field is proportional to the light intensity $I$ and, in the rotating-wave approximation, inversely

800: proportional to the frequency detuning $\delta$ from resonance:

801: 

802: 

803: 

804: \begin{equation}

805: \Delta E = \frac{\hbar \Gamma^2 I}{8 \delta I_{sat}},

806: \end{equation}

807: where $\Gamma$ and $I_{sat}$ are the linewidth and the saturation

808: intensity of the resonance transition, respectively, while the

809: photon scattering rate $R_S$ scales as the light intensity and the

810: inverse square of the laser detuning:

811: 

812: 

813: 

814: \begin{equation}

815: R_S = \frac{\Gamma^3 I}{4 \delta^2 I_{sat}}.

816: \end{equation}

817: Thus, at a given trap potential depth, it is possible to reduce

818: the heating due to photon scattering by increasing $\delta$ and

819: $I$ proportionally. In far-off resonant optical dipole traps

820: (FORT) the effect of photon scattering is negligible for most

821: practical purposes.

822: 

823: 

824: 

825: The trapping radiation couples singlet and

826: triplet states independently. This results in a different

827: dependence on the trapping wavelength for the light shift of the

828: $^1$S$_0$ and $^3$P$_1$ levels, so that the differential shift

829: vanishes at a ``magic wavelength''. This enables optical cooling

830: into the dipole trap, since the detuning of the cooling radiation

831: is not position-dependent.

832: 

833: 

834: 

835: For our optical dipole trap we employ the apparatus described in appendix

836: \ref{IRlaser}. In most cases we use the horizontally propagating

837: beam alone. This single-beam FORT has a trap depth of $90\,\mu$K.

838: The computed radial and axial trap frequencies are $\omega_r =

839: 2\pi \times 2$ kHz and $\omega_a = 2\pi \times 26$ Hz

840: respectively. The vertical beam produces a maximum potential depth

841: of about $12\,\mu$K. We find good agreement between computed and

842: measured trap frequencies. The resonant 461\,nm probe beam for

843: time-of-flight absorption imaging propagates horizontally at an

844: angle of about $30^{\circ}$ with the horizontal FORT beam.

845: 

846: 

847: 

848: The transfer efficiency from the red MOT to the FORT results from

849: a balance between loading flux and density dependent losses due to

850: light assisted collisions.\cite{KatoriFORT1999} Figure

851: \ref{DipoleTrap} shows typical loading curves, together with an

852: in-situ image of the trapped atoms.

853: 

854: 

855: 

856: 

857: 

858: \begin{figure}[t]

859: \centerline{\psfig{file=dipolo,width=12cm}} %\vspace*{8pt}

860: \caption{Crossed-beams FORT loading. The inset shows an in-situ

861: image of trapped atoms. The graph shows the measured FORT

862: population as a function of the time overlap between FORT and red

863: MOT. Reprinted figure with permission from N. Poli et al., Phys.

864: Rev. A 71, 061403(R) (2005). Copyright (2005) by the American

865: Physical Society.}\label{DipoleTrap}

866: \end{figure}

867: 

868: 

869: 

870: 

871: 

872: The AC Stark shift of the intercombination line depends on the

873: direction of the FORT field with respect to the bias magnetic field

874: $\vec B$. We use the MOT quadrupole field to resolve the Zeeman structure

875: of the $^1$S$_0$-$^3$P$_1$ line, and we keep the polarization of the FORT

876: beam linear and orthogonal to

877: $\vec B$. In such conditions the resulting light shift is not critical

878: for laser cooling in the FORT. The wavelength used for the dipole trap is

879: only 7\,nm away from the ``magic wavelength'' for the intercombination

880: $^1$S$_0$-$^3$P$_1$ transition.\cite{Ido2003} When the polarization of

881: the dipole trapping light is orthogonal to the magnetic field the Stark

882: shift for the $^1$S$_0$-$^3$P$_1$ transition is lower than 10\,kHz.

883: Thus, it is possible to cool the atoms while loading the

884: dipole trap.\cite{Ido2000}

885: 

886: 

887: 

888: 

889: 

890: 

891: 

892: \section{Collisional studies on ground-state even Sr isotopes}

893: \label{CollisionalMeasure}

894: 

895: 

896: 

897: The study of atomic collisions at low temperature has undergone a

898: rapid development in recent years following the advent of a number

899: of cooling techniques for diluite atomic gases. The measurement of

900: collisional parameters gives insight into an interesting and

901: promising physics, allowing tests of theoretical models for

902: interatomic potentials and molecular

903: wavefunctions.\cite{Vogel1999,Fedichev1996} Moreover, a precise

904: knowledge of collisional properties is essential for experiments

905: aiming to explore regimes of quantum degeneracy in atomic

906: gases.\cite{Delannoy2001,Yamashita2003}

907: 

908: 

909: 

910: Here we describe a systematic analysis on the ground-state collisional

911: properties of the two most abundant bosonic Sr isotopes ($^{88}$Sr

912: and $^{86}$Sr) performed with our apparatus. More specifically, we

913: evaluated the elastic cross-sections $\sigma_{i-j}$ ($i,j =86,88$)

914: for both intra and inter-species collisions, and the three-body

915: recombination coefficients $K_{i}$ ($i =86,88$) in the FORT. The

916: elastic cross-sections were deduced by driving the system out of

917: thermal equilibrium and measuring the thermalization rate together

918: with the sample density. For the inelastic collisions, we measured

919: the density dependence of the trap loss rate. The key point of our

920: experimental procedure is a precise knowledge of the atomic

921: density. This in turn requires a proper control of the trap

922: frequencies. We assumed the equilibrium phase-space distribution

923: in an ideal harmonic trap to infer the atomic density from the

924: measured number of atoms $N$ and temperature $T$. Our assumption

925: is supported by the fact that the ratio $\eta=\frac{U}{k_B T}$ of

926: trap depth and sample temperature was larger than 5 in all of our

927: measurements. The equilibrium peak atomic density in a harmonic trap is

928: 

929: 

930: 

931: \begin{equation}

932: n_0 = N {\bar \nu}^3 \left({2 \pi m \over k_{B}T}\right)^{3 \over

933: 2}

934: \end{equation}

935: where $\bar \nu$ is the average trap frequency, $m$ is the atomic

936: mass and $k_B$ is the Boltzmann constant. Our results show

937: significant differences in the collisional properties of the two

938: isotopes. Both the elastic cross-section and the three-body

939: collision coefficient were found to be several orders of magnitude

940: larger in $^{86}$Sr than in $^{88}$Sr, and the inter-species

941: cross-section $\sigma_{86-88}$ is much larger than the

942: intra-species cross-section $\sigma_{88-88}$.

943: 

944: 

945: We adopted the standard technique of observing the

946: cross-thermalization between orthogonal degrees of

947: freedom.\cite{Monroe1993,Arndt1997,Hopkins2000,Schmidt2003} The

948: single-beam FORT geometry is not well suited for this experiment,

949: as the temperature in the horizontal modes cannot be univocally

950: determined with our apparatus, and because with $^{86}$Sr in our

951: experimental conditions the system turned out to be in

952: hydrodinamic regime along the axial

953: direction.\cite{Odelin1999,Gensemer2001} Given these constraints,

954: we chose the crossed-beams geometry for the thermalization

955: measurements.

956: 

957: 

958: 

959: In order to drive the system out of thermal equilibrium we

960: performed a selective optical cooling along the vertical direction

961: with resonant 689 nm light. For this purpose we took advantage of

962: the fact that the FORT is located at the center of the red MOT,

963: where atoms are not resonant with the horizontal MOT

964: beams.\cite{Loftus2004} We chose the optical intensity and cooling

965: time in such way to create a detectable thermal anisotropy without

966: introducing dramatic losses.

967: 

968: 

969: 

970: We measured the intra-species thermalization rate $\tau_{th}$ by

971: separately loading one isotope (either  $^{88}$Sr or $^{86}$Sr) and

972: observing the temporal evolution towards thermal equilibrium (see

973: figure \ref{86Thermalization}). In both cases the total trap loss

974: rate was much lower than the thermalization rate, so the atom

975: number was constant within shot-to-shot fluctuations.

976: 

977: 

978: 

979: 

980: 

981: \begin{figure}[t]

982: \centerline{\psfig{file=86therm,width=7cm}} %\vspace*{8pt}

983: \caption{Measurement of the thermalization rate for $^{86}$Sr. The

984: graph shows the temporal evolution of the horizontal and vertical

985: temperatures after cooling on the vertical direction. The solid

986: curve is an exponential fit to the vertical temperature

987: data.}\label{86Thermalization}

988: \end{figure}

989: 

990: 

991: 

992: 

993: 

994: We deduced the elastic cross-section from the measured

995: thermalization rate, by assuming a value of 2.7 for the average

996: number of collisions to reach thermalization, that is,

997: $\tau_{th}=2.7

998: \tau_{coll}$.\cite{Monroe1993,Arndt1997,Kavoulakis2000} The

999: collision rate is given by

1000: 

1001: 

1002: 

1003: \begin{equation}

1004: {1\over\tau_{coll}} = \bar n \sigma \bar v

1005: \end{equation}

1006: where $\bar n$ is the average atom density, $\bar v$ is the

1007: average relative velocity of colliding atoms, and $\sigma$ is the

1008: cross-section. We computed $\bar n$ and $\bar v$ from the measured

1009: values for atom number $N$, average sample temperature $T$ and

1010: average trap frequency $\bar \nu$. With our assumptions the

1011: average density is given by $\bar n = {n_0 / 2 \sqrt 2}$, while

1012: the average relative velocity is given by $\bar v = 4 \sqrt {k_B

1013: T/\pi m}$.

1014: 

1015: 

1016: 

1017: We repeated our measurements for different values of the atom

1018: density, in order to check whether the observed

1019: cross-thermalization was due to elastic collisions or ergodic

1020: mixing between different degrees of freedom. The resulting values

1021: are $\sigma_{88-88} = 3(1) \times 10^{-13}$ cm$^2$  and

1022: $\sigma_{86-86} = 1.3(0.5) \times 10^{-10}$ cm$^2$. The

1023: uncertainty is mainly due to shot-to-shot fluctuations in the FORT

1024: population, that reflect in both density and temperature

1025: instabilities. Such results might be compared with the

1026: zero-temperature cross sections deduced from scattering length

1027: values through the relation $\sigma=8\pi\lambda^2$, where

1028: $\lambda$ is the $s$-wave scattering length. Scattering length

1029: values were obtained from photoassociation spectra by the Tokyo

1030: group for $^{88}$Sr,\cite{yasuda2004} and by the Houston group for

1031: both isotopes.\cite{Mickelson2005} As concerning $\sigma_{88-88}$,

1032: our value is consistent with the Tokyo work, while the Houston

1033: group predicts an even smaller cross-section. On the contrary,

1034: there is a fair agreement between our $\sigma_{86-86}$ value and

1035: the Houston work. All of these results are summarized in tab. 2

1036: 

1037: 

1038: 

1039: \begin{table}[pt]

1040: \tbl{Intra-species elastic cross-section for $^{88}$Sr and

1041: $^{86}$Sr, in cm$^2$} {\begin{tabular}{@{}cccc@{}} \toprule

1042: Reference &

1043: $\sigma_{88-88}$  & $\sigma_{86-86}$  & Method  \\

1044: \colrule

1045: Tokyo group\cite{yasuda2004} &  \, $3(1) \times 10^{-13}$ &  & $\lambda$ from PA spectra \\

1046: Houston group\cite{Mickelson2005} &  $<1.2\times 10^{-13}$ & $2.6 \times 10^{-10} < \sigma < 3.7 \times 10^{-9}$ & $\lambda$ from PA spectra \\

1047: Our group\cite{Ferrari2006} & \, $3(1) \times 10^{-13}$ & $1.3(0.5) \times 10^{-10}$ & cross-thermalization \\

1048: \botrule

1049: \end{tabular}}

1050: \end{table}

1051: 

1052: 

1053: 

1054: 

1055: 

1056: We adopted a similar approach for the measurement of the

1057: inter-species cross section.\cite{Goldwin,K-Rb,Mosk,Delannoy} The

1058: single-beam FORT was loaded with a mixture of $^{88}$Sr ($10^5$

1059: atoms) and $^{86}$Sr ($4 \times 10^4$ atoms). We then applied a

1060: selective optical cooling stage (duration 5\,ms) to $^{88}$Sr and

1061: observed the temperature evolution of the two samples (see figure

1062: \ref{MixThermalization}). The $^{88}$Sr is heated by $^{86}$Sr to

1063: the equilibrium temperature in few tens of ms. For comparison, in

1064: absence of $^{86}$Sr the temperature of $^{88}$Sr grows by less

1065: than 5\,\% in 100\,ms. The resulting inter-species cross-section

1066: was found as $\sigma_{88-86} = 4(1) \times 10^{-12}$ cm$^2$. This

1067: value is significantly larger than the intra-species cross-section

1068: $\sigma_{88-88}$, suggesting the way to a novel and efficient

1069: sympathetic cooling mechanism.\cite{Ferrari2006}

1070: 

1071: 

1072: 

1073: \begin{figure}[t]

1074: \centerline{\psfig{file=mixtherm,width=7cm}} %\vspace*{8pt}

1075: \caption{Measurement of the interspecies thermalization rate. The

1076: graph shows the temperature evolution of $^{88}$Sr and $^{86}$Sr

1077: after selective cooling on $^{88}$Sr. The solid curve is an

1078: exponential fit to the $^{88}$Sr data. Trap populations are $10^5$

1079: atoms for $^{88}$Sr and $4.5 \times 10^4$ atoms for

1080: $^{86}$Sr.}\label{MixThermalization}

1081: \end{figure}

1082: 

1083: 

1084: 

1085: We studied the inelastic collisions by loading a single isotope in

1086: the single-beam FORT  (either  $^{88}$Sr or $^{86}$Sr) and looking

1087: at the evolution of the number and temperature of trapped atoms in

1088: the FORT after the MOT operation was finished. We found no

1089: evidence for non-exponential decay with $^{88}$Sr. The measured

1090: lifetime of 7\,s is consistent with the residual background gas

1091: pressure of $10^{-8}$\,torr. With the initial atom density at trap

1092: center of $3 \times 10^{13}$ cm$^{-3}$ this gives an upper limit

1093: of $10^{-27}$ cm$^6$s$^{-1}$ for the $K_{88}$ coefficient.

1094: 

1095: 

1096: 

1097: 

1098: 

1099: \begin{figure}[t]

1100: \centerline{\psfig{file=recomb86,width=9cm}} %\vspace*{8pt}

1101: \caption{Decay of trap population when loading $^{86}$Sr atoms

1102: alone in the single-beam FORT. The sample temperature is

1103: 12\,$\mu$K, trap depth is 90\,$\mu$K.}\label{Recombination86}

1104: \end{figure}

1105: 

1106: 

1107: 

1108: 

1109: 

1110: When loading the single-beam FORT with $^{86}$Sr only, we observed

1111: a clear non-exponential decay (figure \ref{Recombination86}). We

1112: deduced the three-body recombination coefficient from the density

1113: dependence of the loss rate.\cite{Burt} Integrating the loss rate

1114: equation

1115: 

1116: 

1117: 

1118: \begin{equation}

1119: \dot N = -\Gamma_b N - K_{86}{\int {n^3 (\vec r, t) d^3 \vec r}}

1120: \end{equation}

1121: where $\Gamma_b$ is the linear loss rate for background gas

1122: collisions, with our assumptions we can write

1123: 

1124: 

1125: 

1126: \begin{equation}

1127: \ln {N \over N_0} = -\Gamma_b t - {K_{86} (2\pi)^3 m^3 \bar \nu ^6

1128: \over  3^{3\over 2}}{\int {{N^2 (t) \over [k_B T(t)]^3} dt}}

1129: \end{equation}

1130: This formula is valid as long as additional losses due to

1131: evaporation can be neglected. In order to limit such effect, we

1132: select the coldest atoms by means of forced evaporation before the

1133: measurement.

1134: 

1135: 

1136: 

1137: We deduced the linear loss rate from the wing of figure

1138: \ref{Recombination86}. This value is consistent with the loss rate

1139: measured with $^{88}$Sr. Then we performed a linear fit of $\ln {N

1140: \over N_0} + \Gamma_b t$ versus $\int {{N^2 (t) \over [k_B

1141: T(t)]^3} dt}$ to derive the recombination constant. We repeated

1142: the measurement several times to average out density fluctuations

1143: reflecting in large uncertainty on $K_{86}$. The final result was

1144: $K_{86} = 1.0(0.5) \times 10^{-24}$ cm$^6$s$^{-1}$.

1145: 

1146: 

1147: 

1148: 

1149: 

1150: 

1151: 

1152: 

1153: 

1154: \section{Towards a BEC of strontium} \label{SrBEC}

1155: 

1156: 

1157: 

1158: 

1159: Laser cooling is a very effective technique to reach phase-space

1160: densities within few orders of magnitude from quantum degeneracy.

1161: The limits in cooling at high density are set by the optical depth

1162: of the sample, i.e. reabsorption of scattered light, and

1163: light-assisted atom-atom collisions. Forced evaporative cooling

1164: represents the common way to circumvent these

1165: limits.\cite{Hess1986} However, this procedure is not effective

1166: with all atoms. In particular, among the atoms cooled with optical

1167: methods, none of the alkali-earth atoms reached quantum degeneracy

1168: so far, except ytterbium which has an alkali-earth-like electronic

1169: structure.\cite{Takasu2003} A phase-space density of $\simeq

1170: 10^{-1}$ was reported for Sr but BEC could not be

1171: reached.\cite{Ido2000}

1172: 

1173: 

1174: 

1175: On this respect, the results of the collisional measurements

1176: reported in \ref{CollisionalMeasure} suggest that evaporative

1177: cooling on pure samples of either $^{86}$Sr or $^{88}$Sr cannot be

1178: very efficient in our experimental conditions. $^{86}$Sr presents

1179: an extremely large elastic cross-section, a good point to

1180: establish a fast thermalization, but the 3-body recombination rate

1181: introduces a loss channel that is fatal with the typical

1182: geometries accessible through optical dipole trapping. An optical

1183: trap with a much larger trapping volume would partially suppress

1184: this loss channel.\cite{Weber2002} $^{88}$Sr instead turns out to

1185: be stable against 3-body decay, but the small elastic

1186: cross-section results in a long thermalization time compared to

1187: typical trap lifetime.

1188: 

1189: 

1190: 

1191: On the other hand, our results suggest a novel all-optical sympathetic

1192: cooling scheme.\cite{Ferrari2006} In the isotope mixture the relatively

1193: large inter-species cross-section results in thermalization times

1194: typically of the order of few milliseconds. This thermalization is

1195: fast even on the time scale of laser cooling on the

1196: intercombination $^1$S$_0$-$^3$P$_1$ transition. Moreover, due to the

1197: 164\, MHz

1198: $^{86}$Sr-$^{88}$Sr isotopic shift and the natural linewidth of 7.6\,kHz

1199: for the $^1$S$_0$-$^3$P$_1$ transition, laser cooling on one

1200: isotope has negligible effect on the other one. It

1201: is then possible to cool sympathetically a dense and optically thick cloud

1202: of one isotope (for instance

1203: $^{88}$Sr) via optical cooling of a small sample of the other

1204: isotope ($^{86}$Sr). Continuous laser cooling of $^{86}$Sr

1205: provides heat dissipation in the sample, while the small optical

1206: depth on $^{86}$Sr does not limit the achievable minimum

1207: temperature. Sympathetic cooling with neutral atoms normally

1208: requires a thermal bath with heat capacity large with respect to

1209: that of the sample to be cooled. This is due to the fact that when

1210: the thermal bath is cooled by evaporative cooling, each lost atom

1211: carries an energy of the order of few times the temperature of the

1212: sample. In the case of optical-sympathetic cooling, each

1213: laser-cooled atom can subtract an energy of the order of the

1214: optical recoil in a time corresponding to a few lifetimes of the

1215: excited state, without being lost.

1216: 

1217: 

1218: 

1219: \begin{figure}[t]

1220: \centerline{\psfig{file=sympatheticdynamics,width=7cm}} %\vspace*{8pt}

1221: \caption{Dynamics of an optically trapped $^{88}$Sr cloud

1222: sympathetically cooled with laser cooled $^{86}$Sr. Filled

1223: circles: $^{88}$Sr atom number. Open circles: $^{86}$Sr atom

1224: number. The number of $^{88}$Sr atoms remains constant during the

1225: process, while $^{86}$Sr decays exponentially with a 80\,ms time

1226: constant. Under optimized conditions, the temperature (triangles)

1227: decreases with a 150\,ms time constant and the mixture is always

1228: at thermal equilibrium. Reprinted figure with permission from G.

1229: Ferrari et al., Phys. Rev. A 73, 023408 (2006). Copyright (2006)

1230: by the American Physical Society.}\label{SimpatheticCooling}

1231: \end{figure}

1232: 

1233: 

1234: 

1235: 

1236: 

1237: We implemented the optical-sympathetic cooling scheme by extending

1238: the temporal overlap between the FORT and the $^{86}$Sr red MOT

1239: after switching off the $^{88}$Sr MOT. Figure

1240: \ref{SimpatheticCooling} reports the dynamics of

1241: optical-sympathetic cooling, starting from an initial temperature

1242: of 15-20\,$\mu$K, limited by density dependent

1243: heating.\cite{Poli2005,Weber2002} It can be observed that the

1244: cooling process does not induce losses on $^{88}$Sr while the

1245: number of $^{86}$Sr atoms exponentially decays with a 80\,ms

1246: lifetime. About 100\,ms after switching off the $^{88}$Sr red MOT,

1247: we observe that the mixture attains thermal equilibrium.

1248: 

1249: 

1250: 

1251: 

1252: 

1253: Under optimized conditions (overall optical intensity $100

1254: I_{sat}$) the temperature decays from the initial value with a

1255: 150\,ms time constant. The minimum attainable temperature depends

1256: both on the intensity of the $^{86}$Sr cooling beam, and the

1257: $^{88}$Sr density. By keeping the cooling parameters on $^{86}$Sr

1258: fixed at the optimum value and by varying the number of trapped

1259: $^{88}$Sr, we determined the dependence of the final temperature

1260: on the $^{88}$Sr density. For $6 \times 10^5$

1261: $^{88}$Sr atoms trapped in the FORT, the final temperature is

1262: 6.7\,$\mu$K at a peak density of $1.3 \times 10^{14}$ cm$^3$; the

1263: corresponding phase-space density is $5 \times 10^{-2}$. This

1264: value is only a factor of two lower than what was

1265: previously obtained without forced evaporation,\cite{Ido2000} but

1266: it exhibits more favorable conditions for starting evaporative

1267: cooling, considering the higher spatial density (more than one

1268: order of magnitude higher) and the number of trapped atoms (gain

1269: $2 \div 10$).

1270: 

1271: 

1272: Indeed we applied an evaporative

1273: cooling stage on $^{88}$Sr by reducing the FORT intensity after the optical sympathetic cooling.

1274: This produced an

1275: increase in phase-space density by roughly a factor of 4, giving a

1276: maximum of $2 \times 10^{-1}$. Such result is consistent with a numerical simulation of

1277: forced evaporation in our experimental conditions. The gain in phase-space density during evaporative

1278: cooling is basically limited by the  $^{88}$Sr elastic cross-section. A promising way towards

1279:  Bose-Einstein condensation seems to be the use of a dipole trap with variable

1280: geometry, to compensate for the reduction in thermalization rate during

1281: evaporation.\cite{Weiss2003}

1282: 

1283: 

1284: 

1285: %\begin{figure}[t]

1286: %\centerline{\psfig{file=asymptemperature,width=9cm}} %\vspace*{8pt}

1287: %\caption{Asymptotic temperature of the $^{88}$Sr cloud for

1288: %different intensities of the $^{86}$Sr laser cooling beam on the

1289: %$^1$S$_0$-$^3$P$_1$ transition ($I_{sat} = 3 \mu$W/cm$^2$). The

1290: %$^{88}$Sr atom number is $6 \times

1291: %10^5$.}\label{AsymptoticTemperature}

1292: %\end{figure}

1293: 

1294: 

1295: 

1296: 

1297: 

1298: \begin{figure}[t]

1299: \centerline{\psfig{file=phasespacedensity,width=7cm}} %\vspace*{8pt}

1300: \caption{Temperature and phase-space density of the $^{88}$Sr

1301: cloud sympathetically cooled with laser cooled $^{86}$Sr, as a

1302: function of the $^{88}$Sr atom number. Reprinted figure with

1303: permission from G. Ferrari et al., Phys. Rev. A 73, 023408 (2006).

1304: Copyright (2006) by the American Physical

1305: Society.}\label{PhaseSpaceDensity}

1306: \end{figure}

1307: 

1308: 

1309: 

1310: 

1311: 

1312: Figure \ref{PhaseSpaceDensity} shows the dependence of $^{88}$Sr

1313: temperature after the optical sympathetic cooling (without

1314: evaporative cooling)

1315:  on the number of atoms

1316: in the trap. We determine the density-dependent heating $dT/dn \simeq 2

1317: \mu$K/($10^{14}$\,cm$^{-3}$), which is 20 times lower than the

1318: equivalent value for pure laser cooled

1319: $^{88}$Sr.\cite{KatoriMOT1999} This large reduction is a direct

1320: consequence of the strong selectivity of the $^1$S$_0$-$^3$P$_1$

1321: transition with respect to the two isotopes. The limit on the

1322: $^{86}$Sr temperature of $4 \mu$K for zero $^{88}$Sr density can

1323: be attributed to the laser cooling dynamics in the tightly

1324: confining potential of the FORT.

1325: 

1326: 

1327: 

1328: 

1329: 

1330: 

1331: 

1332: \section{Ultracold Sr atoms as quantum sensors} \label{Microsensor}

1333: 

1334: 

1335: 

1336: Ultracold atomic strontium is particularly suited for applications

1337: in the field of quantum sensors. Atom interferometry has already

1338: been used with alkali-metals for precision inertial

1339: sensing,\cite{Peters1999,Gustavson2000} for measuring fundamental

1340: constants,\cite{Wicht2002,Clade2006,Stuhler2003} and testing

1341: relativity.\cite{Fray2004} The extremely small size of ultracold

1342: atomic samples enables precision measurements of forces at

1343: micrometer scale. This is a challenge in physics for studies of

1344: surfaces, Casimir effect,\cite{Antezza2005} and searches for

1345: deviations from Newtonian gravity predicted by theories beyond the

1346: standard model.\cite{Long2003,Dimopoulos2003,Samullin2005}

1347: 

1348: 

1349: 

1350: An interesting class of quantum devices is represented by

1351: ultracold atoms confined in an optical lattice, that is a dipole

1352: trap created by a laser standing wave.\cite{Bloch2005} In

1353: particular, Bloch oscillations were predicted for electrons in a

1354: periodic crystal potential in presence of a static electric

1355: field\cite{Bloch1929} but could not be observed in natural

1356: crystals because of the scattering of electrons by the lattice

1357: defects. They were directly observed using atoms in an optical

1358: lattice.\cite{Raizen1997}

1359: 

1360: 

1361: 

1362: The combination of the periodic optical potential and a linear

1363: potential produced by a constant force $F$ along the lattice

1364: wave-vector gives rise to Bloch oscillations at frequency $\nu_B$

1365: given by $\nu_B = \frac{F \lambda_L}{2h}$, where $\lambda_L$ is

1366: the wavelength of the light producing the lattice, and $h$ is

1367: Planck’s constant. Since $\lambda_L$ is well known, the force

1368: along the lattice axis can be accurately determined by measuring

1369: the Bloch frequency $\nu_B$. In order to perform a sensitive force

1370: measurement, a long coherence time with respect to the measurement

1371: duration is required. The most common effects limiting the

1372: coherence time for ultracold atoms are perturbations due to

1373: electromagnetic fields and atom-atom interactions. $^{88}$Sr is in

1374: this respect an ideal choice because in the ground state it has

1375: zero orbital, spin and nuclear angular momentum. This makes it

1376: virtually insensitive to stray magnetic fields. In addition, as

1377: shown in section \ref{CollisionalMeasure} $^{88}$Sr has remarkably

1378: small atom-atom interactions. Such properties make Sr in optical

1379: lattices a unique sensor for small-scale forces with better

1380: performances and reduced complexity compared to proposed schemes

1381: using degenerate Bose or Fermi gases.\cite{Anderson1998,Roati2004}

1382: This enables new experiments on gravity in unexplored regions.

1383: 

1384: 

1385: 

1386: We tested such idea by measuring the gravitational acceleration in

1387: our laboratory with a $^{88}$Sr sample in a vertical optical

1388: lattice. To this end, we cool $\sim 5 \times 10^5$ atoms in the

1389: red MOT down to the recoil temperature (see section \ref{RedMOT}),

1390: so that the vertical momentum distribution is narrower than the

1391: width of the first Brillouin zone.\cite{Bloch1929}. Then we

1392: release the atoms from the MOT and we switch on adiabatically a

1393: one-dimensional optical lattice.

1394: 

1395: 

1396: The lattice potential is originated by a single-mode

1397: frequency-doubled Nd:YVO$_4$ laser ($\lambda_L = 532$\,nm)

1398: delivering up to 350\,mW on the atoms with a beam waist of $200

1399: \mu$m. The beam is vertically aligned and retro-reflected by a

1400: mirror producing a standing wave with a period

1401: $\frac{\lambda_L}{2} = 266$\,nm. The corresponding photon recoil

1402: energy is $E_R = \frac{h^2}{2m\lambda^2} = k_B \times 381$\,nK. We

1403: populate about 100 lattice sites with $2 \times 10^5$ atoms at an

1404: average spatial density of $\sim 10^{11}$\,cm$^{-3}$. We leave the

1405: atoms in the optical lattice for a variable time $t$, then we

1406: switch off the lattice adiabatically and we measure the momentum

1407: distribution of the sample by time-of-flight imaging, after a free

1408: fall of 12\,ms.

1409: 

1410: 

1411: 

1412: We integrate along the

1413: horizontal direction the optical thickness obtained by absorption imaging.

1414: The resulting curve gives

1415: the vertical momentum distribution of the atomic sample: in figure

1416: \ref{TwoPeaks} we show a typical plot after the integration.

1417: 

1418: 

1419: 

1420: 

1421: 

1422: \begin{figure}[t]

1423: \centerline{\psfig{file=TwoPeakDistrib,width=7cm}} %\vspace*{8pt}

1424: \caption{Vertical momentum distribution of the atoms at the Bragg

1425: reflection.}\label{TwoPeaks}

1426: \end{figure}

1427: 

1428: 

1429: 

1430: 

1431: 

1432: We fit the measured momentum distribution  with the sum of two

1433: Gaussian functions. From each fit we extract the vertical momentum

1434: center of the lower peak and the width of the atomic momentum

1435: distribution. We find that the latter is less sensitive against

1436: noise-induced perturbations to the vertical momentum. We can

1437: observe $\sim 4000$ Bloch oscillations in a time $t = 7$\,s (see

1438: figure \ref{BlochOsc}), with a damping time $\tau\sim 12$\,s. To

1439: our knowledge, the present results for number of Bloch

1440: oscillations, duration, and the corresponding number of coherently

1441: transferred photon momenta, are by far the highest ever achieved

1442: experimentally in any physical system.

1443: 

1444: 

1445: 

1446: 

1447: 

1448: \begin{figure}[t]

1449: \centerline{\psfig{file=fig13.EPS,width=9cm}} %\vspace*{8pt}

1450: \caption{Time evolution of the width of the atomic momentum

1451: distribution, showing Bloch oscillation of $^{88}$Sr atoms in the

1452: vertical 1-dimensional optical lattice under the effect of

1453: gravity. From the the data fit, a Bloch frequency $\nu_B =

1454: 574.568(3)$\,Hz is obtained with a damping time $\tau\sim 12$\,s

1455: for the oscillations.}\label{BlochOsc}

1456: \end{figure}

1457: 

1458: 

1459: 

1460: From the measured Bloch frequency $\nu = 574.568(3)$\,Hz we

1461: determine the gravity acceleration along the optical lattice

1462: $g_{meas} = 9.80012(5)$\,ms$^{-2}$.

1463: %The expected value at the

1464: %laboratory site is $g_{ref} =

1465: %9.805046(9)$\,ms$^{-2}$.\cite{Petelski} The disagreement by $5

1466: %\times 10^{-4} g$ is due to a vertical misalignment of $\sim

1467: %20$\,mrad of the optical lattice beam.

1468: The estimated sensitivity is $5 \times 10^{-6} g$. We expect that

1469: such precision may be increased by one order of magnitude by using

1470: a larger number of atoms, and reducing the initial temperature of

1471: the sample.

1472: 

1473: 

1474: 

1475: 

1476: 

1477: 

1478: 

1479: 

1480: \section{Frequency measurements on the Sr intercombination lines}

1481: 

1482: \label{1S3PFreqMeasure}

1483: 

1484: 

1485: 

1486: The advent of laser cooling techniques had important consequences

1487: on time-frequency metrology, by reducing the uncertainty on the

1488: frequency of atomic clocks caused by atom motion.

1489: 

1490: 

1491: 

1492: The recent progresses in two related fields, namely

1493: high-resolution laser spectroscopy and direct optical-frequency

1494: comb generation, opened the way to a new generation of frequency

1495: standards based on transitions in the optical domain. The use of

1496: frequency combs based on self-mode-locked femtosecond lasers has

1497: made possible, for the first time, relatively simple

1498: optical-frequency measurements.\cite{Udem1999,Diddams2000} On the

1499: other hand, the realization of lasers with ultra-narrow emission

1500: band now enables spectroscopy on forbidden optical transitions

1501: with quality factor $Q = \frac{\nu}{\Delta\nu}$ in excess of

1502: $10^{15}$.

1503: 

1504: 

1505: 

1506: Because of their higher frequency, optical transitions have the

1507: potential for greatly improved accuracy and stability compared to

1508: conventional atomic clocks based on microwave frequency

1509: transitions.\cite{Udem2002} Different transitions are now

1510: considered as optical-frequency standards, involving single ions

1511: and neutral atoms.\cite{Udem2001,Gill2005} While single ions offer

1512: an excellent control on systematic effects, clouds of laser cooled

1513: atoms have the potential for extremely high precision, as the

1514: large number of atoms reduces the quantum projection noise.

1515: Perhaps the use of optical lattices at the magic wavelength to

1516: confine neutral atoms in the Lamb-Dicke regime, as proposed for

1517: the first time on strontium,\cite{KatoriFORT1999,Katori2003}

1518: summarizes the best of both worlds; that is, a large number of

1519: quantum absorbers with negligible shift of the optical clock

1520: transition due to external fields, Doppler effect and collisions.

1521: 

1522: 

1523: 

1524: Among the neutral atoms, Sr has long been considered as one of the

1525: most interesting candidates.\cite{Hall1989} Several features, some

1526: of which are specific to this atom, allow different possibilities

1527: for the realization of a high precision optical clock. The visible

1528: intercombination 5$^1$S-5$^3$P lines from the ground state are

1529: easily accessible with semiconductor lasers. Depending on the

1530: specific fine-structure component and on the isotope,  a wide

1531: choice of transitions with different natural linewidths is

1532: possible (see section \ref{Strontiumatom}).

1533: 

1534: 

1535: 

1536: The first phase-coherent absolute frequency measurement of a Sr

1537: intercombination line was performed by our group on the

1538: 5$^1$S$_0$-5$^3$P$_{1}$ transition, using a thermal atomic

1539: beam.\cite{Ferrari2003} This represented an improvement by several

1540: orders of magnitude with respect to previous data.\cite{Tino1992}

1541: Our result is in agreement with subsequent phase-coherent

1542: measurements performed by the BNM-SYRTE group on a thermal

1543: sample,\cite{Courtillotclock} and by the JILA group on a

1544: free-falling ultracold sample using a red MOT.\cite{Ido2005} The

1545: JILA measurement provided a further improvement in the accuracy by

1546: more than two orders of magnitude.

1547: 

1548: 

1549: 

1550: However, the $^1$S$_0$-$^3$P$_{1}$ transition is not best suited

1551: as a final frequency reference, due to its natural linewidth of

1552: 7.6\,kHz. Some groups recently began working on the 698\,nm

1553: $^1$S$_0$-$^3$P$_{0}$ line, that is strictly forbidden in the even

1554: isotopes since it is a $J=0\to J=0$ transition. In $^{87}$Sr the

1555: hyperfine mixing enables direct $^1$S$_0$-$^3$P$_{0}$ excitation

1556: with a transition probability of $\sim 1 $\,mHz. The BNM-SYRTE

1557: group first measured such transition in a blue MOT with an

1558: uncertainty of 15\,kHz.\cite{Courtillotclock} The Tokyo group

1559: performed the first absolute frequency measurement in an optical

1560: lattice on this transition,\cite{Takamoto2005} followed by the

1561: JILA and BNM-SYRTE groups.\cite{Ludlow2005,Targat2006} All groups

1562: estimated an uncertainty $\leq 20$\,Hz for the absolute transition

1563: frequency. The corresponding 578\,nm $^1$S$_0$-$^3$P$_{0}$

1564: transition in Ytterbium was observed at NIST in a $\sim 70 \mu$K

1565: sample on the two odd isotopes, $^{171}$Yb and $^{173}$Yb, with an

1566: uncertainty of $\sim 4$\,kHz.\cite{Hoyt2005}

1567: 

1568: 

1569: 

1570: In spite of its large quality factor, the $^1$S$_0$-$^3$P$_{0}$

1571: transition in odd Sr and Yb isotopes suffers from residual

1572: sensitivity to stray magnetic fields and optical lattice

1573: polarization, besides a complex line structure due to the presence

1574: of many magnetic sublevels. Several groups are now looking at the

1575: even Sr or Yb isotopes as the best candidates to represent optical

1576: frequency standards based on neutral atoms. Some groups have

1577: proposed different methods to directly excite the clock transition

1578: on the even Sr or Yb isotopes, by properly engineering the atomic

1579: level structure to create a finite and controllable

1580: $^1$S$_0$-$^3$P$_{0}$ transition probability. These methods

1581: basically consist in coupling the metastable $^3$P$_{0}$ level to

1582: other electronic states by using either multiple near-resonant

1583: laser fields,\cite{Santra2005,Hong2005} or simply a small static

1584: magnetic field.\cite{Taichenachev2005} The latter scheme has been

1585: experimentally demonstrated on $^{174}$Yb at

1586: NIST.\cite{Barber2005} The possible instability due to

1587: site-to-site tunneling in optical lattice clocks has been

1588: addressed by the BNM-SYRTE group.\cite{Lemonde2005} For accuracy

1589: goals at the $10^{-18}$ level they propose the use of vertical

1590: optical lattices, in order to lift the degeneracy between adjacent

1591: potential wells.

1592: 

1593: 

1594: 

1595: 

1596: 

1597: 

1598: 

1599: \subsection{Frequency measurement on the $^1$S$_0$-$^3$P$_1$

1600: transition with a thermal beam} \label{OpticalFreqMeasure}

1601: 

1602: 

1603: 

1604: 

1605: 

1606: In this section we discuss our precision frequency

1607: measurements on the intercombination 5$^1$S$_0$-5$^3$P$_{1}$

1608: transition.\cite{Ferrari2003} Using a femtosecond laser comb, we

1609: determined the absolute frequency of the transition for $^{88}$Sr

1610: and $^{86}$Sr and an accurate value for the isotope shift.

1611: 

1612: 

1613: 

1614: 

1615: 

1616: The frequency measurements have been performed through saturation

1617: spectroscopy on a thermal beam with the apparatus described in

1618: appendix \ref{RedLaser}. As frequency-comb generator we employed a

1619: commercial system based on a Kerr-lens mode-locked Ti:Sa laser

1620: with a repetition rate of 1\,GHz.\cite{Cundiff2001} Its repetition

1621: rate and carrier envelope offset frequency were locked to a GPS

1622: stabilized quartz oscillator. The strontium atomic beam is

1623: obtained from the metal heated to 830\,K. Residual atomic beam

1624: divergency is 25\,mrad and the typical atomic density in the

1625: detection region is 10$^8$ cm$^{-3}$. We derive a Doppler-free

1626: signal using a retro-reflected laser beam perpendicular to the

1627: atomic beam. The fluorescence light from the laser excited atoms

1628: is collected on a photomultiplier tube.

1629: 

1630: 

1631: 

1632:  We estimate the indetermination on the reflection angle

1633: of the laser beam to be less than 10\,$\mu$rad. The peak beam

1634: intensity of 60\,$\mu$W/cm$^2$ (to be compared to the saturation

1635: intensity of 3\,$\mu$W/cm$^2$) was chosen to obtain sufficient

1636: signal-to-noise for the RC lock onto the atomic resonance. A

1637: uniform magnetic field of 10\,G (see figure \ref{689Setup})

1638: defines the quantization axis in the interrogation region such

1639: that the light is $\pi$ polarized. The double pass AOM next to the

1640: atomic detection (AOM3) is frequency modulated at 10\,kHz to

1641: derive the locking signal of the cavity onto the atomic line.

1642: 

1643: 

1644: 

1645: \begin{figure}[t]

1646: \centerline{\psfig{file=RedFluorescence,width=9cm}} %\vspace*{8pt}

1647: \caption{Fluorescence spectrum of the strontium

1648: $^1$S$_0$-$^3$P$_1$ line at 689\,nm. The lines of the two bosonic

1649: isotopes $^{86}$Sr and $^{88}$Sr, together with the hyperfine

1650: structure of the fermionic $^{87}$Sr, can be resolved. The

1651: linewidth corresponds to the residual first order Doppler

1652: broadening in the thermal beam. Inset: sub-Doppler resonance of

1653: $^{88}$Sr recorded by saturation spectroscopy using two

1654: counterpropagating laser beams. The amplitude of the dip is 10\,\%

1655: of the Doppler signal. Reprinted figure with permission from G.

1656: Ferrari et al., Phys. Rev. Lett. 91, 243002 (2003). Copyright

1657: (2006) by the American Physical Society.}\label{RedFluor}

1658: \end{figure}

1659: 

1660: 

1661: 

1662: Figure \ref{RedFluor} shows the Doppler broadened resonances of

1663: $^{88}$Sr, $^{86}$Sr, and the hyperfine structure of $^{87}$Sr.

1664: The residual atomic beam divergency produces a residual Doppler

1665: broadening of 60\,MHz FWHM. In the inset, the sub-Doppler signal

1666: for $^{88}$Sr is shown. Two independent measurements of the

1667: sub-Doppler resonance show a FWHM of about 50\,kHz, which is in

1668: agreement with the expected value considering the saturation and

1669: transit time broadening, and the recoil splitting.

1670: 

1671: 

1672: 

1673: 

1674: 

1675: \begin{figure}[t]

1676: \centerline{\psfig{file=FreqPlot,width=8cm}} %\vspace*{8pt}

1677: \caption{Chronological plot of the optical frequency measurements.

1678: The error bars correspond to the standard deviation for each data

1679: set. Reprinted figure with permission from G. Ferrari et al.,

1680: Phys. Rev. Lett. 91, 243002 (2003). Copyright (2006) by the

1681: American Physical Society.}\label{FreqencyMeasure}

1682: \end{figure}

1683: 

1684: 

1685: 

1686: 

1687: 

1688: Figure \ref{FreqencyMeasure} shows the result of the measurement

1689: of the $^{88}$Sr transition frequency taken over a period of

1690: several days. Each data point corresponds to the averaging of the

1691: values resulting from consecutive measurements taken with a 1\,s

1692: integration time over $100 \div 200$\,s. The error bars correspond

1693: to the standard deviation for each data set. We evaluated first-

1694: and second-order Doppler and Zeeman effects, ac Stark shift,

1695: collisional shifts, and mechanical effects of light. The resulting

1696: value for the $^{88}$Sr transition frequency, including the

1697: corrections discussed previously, is 434 829 121 311 (10)\,kHz,

1698: corresponding to a 1$\sigma$ relative uncertainty of $2.3 \times

1699: 10^{-11}$.

1700: 

1701: 

1702: 

1703: 

1704: 

1705: 

1706: \subsection{$^{86}$Sr - $^{88}$Sr isotopic shift measurement} \label{IsotopeShift}

1707: 

1708: 

1709: 

1710: With a minor change in the apparatus, we locked simultaneously the

1711: frequency of two laser beams to the sub-Doppler signals of

1712: $^{86}$Sr and $^{88}$Sr. This system allowed us to measure the

1713: isotopic shift of the $^1$S$_0$-$^3$P$_1$ transition by counting

1714: the beat note between the two interrogating beams. For this

1715: purpose, the reference cavity is locked to the $^{88}$Sr resonance

1716: as described previously and the light for $^{86}$Sr is derived

1717: from the same laser beam and brought to resonance through AOMs.

1718: The two beams are overlapped in a single mode optical fiber and

1719: sent to the interrogation region. By frequency modulating the

1720: beams at different rates and using phase sensitive detection we

1721: get the lock signal for both the isotopes from the same

1722: photomultiplier. In this isotope-shift measurement most of the

1723: noise sources are basically common mode and rejected. The measured

1724: $^{88}$Sr-$^{86}$Sr isotope shift for the $^1$S$_0$-$^3$P$_1$

1725: transition is 163 817.4 (0.2)\,kHz. This value represented an

1726: improvement in accuracy of more than three orders of magnitude

1727: with respect to previously available data.\cite{Buchinger1985} The

1728: $^{86}$Sr optical frequency then amounts to 434 828 957 494

1729: (10)\,kHz.

1730: 

1731: 

1732: 

1733: 

1734: 

1735: \section{Conclusions}

1736: \label{Conclusion}

1737: 

1738: The strontium atom is an attractive candidate both for physical

1739: studies and for applications. We have shown how the properties of

1740: atomic strontium  are suited for laser cooling and trapping, for

1741: the study of ultracold atomic physics, and for applications to

1742: optical frequency metrology or to micrometric force sensors.

1743: 

1744: Future work on ultracold strontium offers intriguing perspectives,

1745: including the possible realization of a nearly non-interacting

1746: Bose-Einstein condensate, the realization of an optical clock with

1747: ultimate stability, or direct experimental tests of theories

1748: beyond the standard model.

1749: 

1750: 

1751: 

1752: 

1753: 

1754: \appendix

1755: 

1756: 

1757: 

1758: 

1759: 

1760: \section{Experimental setup}

1761: \label{ExperimentalSetup}

1762: 

1763: 

1764: 

1765: The typical experimental setup for Sr laser cooling and trapping

1766: basically includes a vacuum system, a blue laser source for the

1767: atom collection and first cooling stage, a red laser for second

1768: stage cooling and precision spectroscopy, and an infrared laser

1769: source for the optical dipole trap. In the following we illustrate

1770: these main parts as they are realized in our laboratory.

1771: 

1772: 

1773: 

1774: 

1775: 

1776: 

1777: 

1778: \subsection{Vacuum system} \label{VacuumSystem}

1779: 

1780: 

1781: 

1782: The apparatus consists in three major parts: the oven, the Zeeman

1783: slower, and the MOT chamber. The oven generates an atomic

1784: strontium vapor by sublimation from metallic strontium kept at

1785: $\sim 800$ Kelvin. The vapor is collimated into an atomic beam

1786: passing through a nozzle filled with about 200 stainless steel

1787: capillaries 8\,mm long which insure a ballistic divergence of the

1788: atomic beam of 20\,mRad.\cite{Courtillot2003} Keeping the

1789: capillaries at a higher temperature prevents internal vapor

1790: condensation and clogging.

1791: 

1792: 

1793: 

1794: 

1795: 

1796: \begin{figure}[t]

1797: \centerline{\psfig{file=vacuum,width=12cm}} \vspace*{8pt}

1798: \caption{Vacuum aparatus. O: oven for St sublimation; IP: ion

1799: pumps; TC: cell for transverse cooling; ZS: differential pumping

1800: tube for the Zeeman slower; C: MOT cell; TS: titanium sublimation

1801: pump; W: sapphire window for the Zeeman slower laser

1802: beam.}\label{VacuumPicture}

1803: \end{figure}

1804: 

1805: 

1806: 

1807: 

1808: 

1809: Following the atomic trajectory (see the sketch in figure

1810: \ref{VacuumPicture}), the atoms pass through a region with radial

1811: optical access for 2D transverse cooling,\cite{Shimizu,Rasel1999}

1812: they are decelerated through the Zeeman slower,\cite{Prodan1982}

1813: and finally stopped in the cell that hosts the MOT. The optical

1814: access for for 2D transverse cooling is provided by a CF35 cube

1815: aligned on the atomic beam. On the two free direction two pairs of

1816: windows, AR coated for 461 and 689\,nm, are sealed with modified

1817: copper gaskets.\cite{WindowSeal} A differential pumping stage

1818: between the transverse cooling region and the Zeeman slower

1819: insures decoupled background pressure between the oven and MOT

1820: region. The oven is pumped by a 20\,l/s ion pump while the MOT is

1821: pumped by a 20\,l/s ion pump and a titanium sublimation pump. With

1822: this setup, in operation condition, we achieve a pressure of

1823: $10^{-7}$ Torr in the transverse cooling region, and a pressure of

1824: $10^{-9}$ Torr on the MOT cell. A BK7 window on the atomic beam

1825: axis provides access for the Zeeman slower decelerating beam. To

1826: prevent chemical reaction of strontium and darkening, the

1827: anti-reflection is only present on the outer side of the window ,

1828: and the window is heated to $\sim 450$\,Kelvin. Paying attention

1829: to block the Sr beam when unnecessary, the window presents a dim

1830: shadow after two years of operation. In the future we plan to

1831: exchange the window with a sapphire one, again anti-reflection

1832: coated on the outer side, from which deposited Sr can be easily

1833: removed.

1834: 

1835: 

1836: 

1837: 

1838: 

1839: \begin{figure}[t]

1840: \centerline{\psfig{file=BlueSetup.ps,width=8cm}} \vspace*{8pt} \caption{Scheme of the laser at 461

1841: nm. A distributed feedback laser (DFB) is amplified in a semiconductor tapered amplifier (TA), and

1842: frequency doubled on a periodically poled KTP crystal. The non-linear conversion is improved by

1843: placing the KTP crystal into an optical resonator. Optical isolators (OI) between the DFB and TA,

1844: and between the TA and the frequency doubling stage. Solid lines represent the optical path,

1845: dashed lines represent electrical connections. BSO: beam shaping optics.}\label{BlueSetup}

1846: \end{figure}

1847: 

1848: 

1849: 

1850: 

1851: 

1852: \subsection{Blue laser sources} \label{BlueLaser}

1853: 

1854: 

1855: 

1856: The light at 461\,nm (see figure \ref{LevelScheme}) used for the

1857: atomic beam deceleration and capture in the MOT is produced by

1858: second-harmonic generation (SHG) of a 922\,nm semiconductor laser

1859: (see figure \ref{BlueSetup}). We generate the 922 nm radiation

1860: with a master oscillator-parametric amplifier system (MOPA). An

1861: anti-reflection coated laser diode mounted in an extended cavity

1862: in Littrow configuration (ECDL)\cite{Littrow1991} delivers 50\,mW

1863: at 922 nm, that is amplified to 1.2 W in a tapered amplifier (TA).

1864: Optical isolators are placed between the ECDL and the TA, and

1865: between the TA and the frequency doubler cavity, in order to

1866: prevent optical feedback into the master laser, and optical

1867: damages on the amplifier. The frequency doubler is composed of a

1868: 20\,mm long periodically-poled KTP crystal, placed in an optical

1869: build up cavity. The crystal facets are anti-reflection coated

1870: both at 922 and 461\,nm (R\,$<$\,0.2\,\%) and the poling period is

1871: chosen to fulfill quasi-phase matching of our wavelength at room

1872: temperature. The resonator has an input coupling mirror with

1873: 11\,\% transmission and it is held in resonance with the input

1874: light feeding the error signal from a H\"{a}nsch-Couillaud

1875: detector\cite{Couillaud81} to a PZT controlled folding mirror.

1876: Under optimal conditions we obtain 300\,mW in the blue and

1877: routinely we work with 200\,mW. We frequency stabilize this blue

1878: source to the $^1$S$_0$-$^1$P$_1$ line of $^{88}$Sr by means of

1879: conventional saturated spectroscopy in a strontium heatpipe. The

1880: servo loop acts on the piezo of the ECDL.

1881: 

1882: 

1883: The light used for atomic manipulation and detection is brought to

1884: the vacuum system through single-mode polarization-maintaining

1885: optical fibers to improve the long-term beam pointing stability.

1886: 

1887: 

1888: 

1889: 

1890: 

1891: 

1892: As we discussed in section \ref{OpticalPumping}, we employ a

1893: turquoise laser at 497\,nm to increase the MOT lifetime and number

1894: of atoms loaded. As in the case of 461\,nm source, there are no

1895: laser diodes available at 497\,nm and the simplest method to

1896: produce this light is frequency doubling an IR laser at 994\,nm.

1897: In this application 1\,mW of light is sufficient to saturate the

1898: process, then we do not need any amplification of the IR light

1899: before frequency doubling. The source at 497\,nm differs from that

1900: at 461\,nm in few parts. The master laser is an anti-reflection

1901: coated diode in ECDL Littrow configuration. After the beam shaping

1902: optics, the IR light is coupled into a bow-tie cavity that

1903: contains a 17\,mm long, b-cut potassium niobate crystal which is

1904: kept at 328\,K in order to satisfy non-critical phase matching for

1905: SHG. The crystal facets are AR coated both for the fundamental and

1906: the blue light and, like the 461\,nm source, the cavity is kept

1907: resonant to the 994\,nm laser with a H\"{a}nsch-Couillaud locking.

1908: Since this laser operates among Sr excited states it is not

1909: possible to lock the laser to the atomic line on a simple

1910: heatpipe. Possible frequency stabilization methods include locking

1911: to a reference cavity or spectroscopy on an atomic sample with a

1912: suitable fraction of excited atoms, such as a hollow cathode lamp,

1913: or an heatpipe with gas discharge, or simply the MOT. We chose the

1914: latter system, in spite of the fact that the blue MOT fluorescence

1915: signal is not continuously available during our measurement

1916: cycles. In fact the short-term laser frequency stability is

1917: sufficient to keep it in resonance with the $^3$P$_2$-$^3$D$_2$

1918: transition for some tens of seconds. Thus we leave the laser free

1919: running, and we manually adjust its frequency at the beginning of

1920: the measurement cycle by acting on the piezo of the ECDL to

1921: maximize the MOT fluorescence.

1922: 

1923: 

1924: 

1925: 

1926: 

1927: 

1928: \subsection{Red laser source} \label{RedLaser}

1929: 

1930: 

1931: 

1932: 

1933: 

1934: The 689\,nm source is composed of a laser diode frequency-locked

1935: to an optical cavity whose modes are stabilized to keep the laser

1936: on resonance with the atomic line. A scheme of the experimental

1937: setup is given in figure \ref{689Setup}. An extended cavity

1938: laser-diode mounted in the Littrow configuration delivers 15\,mW

1939: at 689 nm. Optical feedback into the ECDL is prevented by a 40\,dB

1940: optical isolator and a single pass acusto-optic modulator in

1941: cascade. The laser linewidth is reduced by locking the laser to an

1942: optical reference cavity (RC) with standard Pound-Drever-Hall

1943: technique;\cite{Pound-Drever-Hall} the phase modulation is

1944: produced by an electro-optic modulator (EOM) driven at 11\,MHz and

1945: leaves 85\,\% of the power in the carrier. The reference cavity

1946: has a free spectral range (FSR) of 1.5\,GHz and a finesse of $\sim

1947: 7000$. On one side of the quartz spacer we glued a concave mirror

1948: (R$=50$\,cm) while on the other side a PZT is glued between the

1949: spacer and a flat mirror in order to steer the modes of the cavity

1950: by more than one FSR.

1951: 

1952: 

1953: 

1954: 

1955: 

1956: \begin{figure}[t]

1957: \centerline{\psfig{file=689Setup.eps,width=10cm}} \vspace*{8pt}

1958: \caption{Scheme of the 689\,nm laser. Experimental setup used for

1959: the frequency measurement on the Sr intercombination line. Optical

1960: isolators (OI) and acusto-optic modulators (AOM) eliminate

1961: feedback among the master laser (ECDL), the slave laser, the

1962: electro-optic modulator (EOM) and the reference cavity (RC).  QWP:

1963: quarter wave-plate. CO: collimation optics. PMF: polarization

1964: maintaining fiber. PMT: photo multiplier tube. The same apparatus

1965: is used for Sr second stage cooling and trapping, but the thermal

1966: beam is replaced by an heat-pipe.}\label{689Setup}

1967: \end{figure}

1968: 

1969: 

1970: 

1971: 

1972: 

1973: The locking loop includes a low frequency channel acting on the

1974: PZT of the ECDL (1\,kHz bandwidth), and a high frequency channel

1975: acting on the laser-diode current supply ($\sim 3$\,MHz

1976: bandwidth). Under lock condition more than 55\,\% of the incident

1977: light is transmitted through the cavity. From the frequency noise

1978: spectrum obtained by comparison with an independent cavity we can

1979: infer a laser linewidth lower than 20\,Hz, and more than 90\,\% of

1980: the optical power in the carrier.\cite{Ferrari2004,Poli2006} The

1981: RC is acoustically isolated, though we do not keep it under

1982: vacuum.\cite{UltraStableCavity} The optical table is actively

1983: isolated from seismic noise with pneumatic legs. The frequency

1984: drifts of the cavity are compensated by the servo to the atomic

1985: signal which acts on the PZT of the high finesse cavity with a

1986: 20\,Hz bandwidth. In the frequency measurement experiment

1987: described in section \ref{FreqencyMeasure} the Doppler free

1988: saturated fluorescence on a thermal strontium beam provides the

1989: signal for cavity stabilization on the atomic line. The thermal

1990: beam source has a similar design as described in

1991: \ref{VacuumSystem} and it is pumped by a 20 l/s ion pump. However,

1992: when using the stable 689\,nm laser for second stage cooling and

1993: trapping, as described in section \ref{RedMOT}, we employ an

1994: heat-pipe kept at $\sim 750$\,K instead of the thermal beam, thus

1995: obtaining a larger signal and a more robust lock.

1996: 

1997: 

1998: 

1999: 

2000: 

2001: 

2002: \subsection{Infrared laser sources for dipole trapping} \label{IRlaser}

2003: 

2004: 

2005: 

2006: After the production of an ultracold sample in double stage

2007: magneto-optical trapping, we transfer the atoms in an optical

2008: dipole trap made of two infrared laser beams crossing each other

2009: near the waist. The two beams, respectively aligned along the

2010: horizontal and vertical direction, are produced with two

2011: independent TA injected with light coming from the same infrared

2012: source used for producing the blue light (see figure

2013: \ref{FORTsources}), that is close enough to the ``magic wavelength'' for the $^1$S$_0$-$^3$P$_{1}$

2014: transition.

2015: Typically 300\,mW of

2016: light coming from that source are coupled into a fiber and sent to the two amplifiers.

2017: The injection of the two TAs it is regulated by two AOMs which are

2018: used to shift the frequency of the two beams (avoiding

2019: interference at the center of the dipole trap) and to apply a fast

2020: control to the output TA optical power. For mode cleaning and

2021: delivering the output beams from the TAs to the atoms, we use two

2022: independent single mode fibers. Typically we obtain about 650\,mW

2023: and 440\,mW at the fiber output for the horizontal and vertical

2024: beams, respectively. We finally focus the beams at the MOT center,

2025: with 1/e$^2$ radii of 15 $\mu$m and 35 $\mu$m respectively.

2026: 

2027: 

2028: 

2029: \begin{figure}[t]

2030: \centerline{\psfig{file=laserFORT,width=7cm}} %\vspace*{8pt}

2031: \caption{The 922\,nm laser source for the optical dipole trap.

2032: }\label{FORTsources}

2033: \end{figure}

2034: 

2035: 

2036: 

2037: 

2038: 

2039: 

2040: 

2041: 

2042: 

2043: 

2044: \begin{thebibliography}{0}

2045: 

2046: 

2047: 

2048: 

2049: 

2050: 

2051: \bibitem{KatoriMOT1999}

2052: H. Katori, T. Ido, Y. Isoya and M. Kuwata-Gonokami, {\it Phys. Rev.

2053: Lett.} {\bf 82}, 1116 (1999).

2054: 

2055: 

2056: 

2057: 

2058: \bibitem{Ferrari2006}

2059: G. Ferrari, N. Poli, R. E. Drullinger, F. Sorrentino and G. Tino,

2060: {\it Phys. Rev. A.} {\bf 73}, 023408 (2006).

2061: 

2062: 

2063: 

2064: 

2065: \bibitem{Mukaiyama2003}

2066: T. Mukaiyama, H. Katori, T. Ido, Y. Li and M. Kuwata-Gonokami,

2067: {\it Phys. Rev. Lett.} {\bf 90}, 113002 (2003).

2068: 

2069: 

2070: 

2071: \bibitem{Xu2003}

2072: X. Xu, T. H. Loftus, J. W. Dunn, C. H. Greene, J. L. Hall, A.

2073: Gallagher and J. Ye, {\it Phys. Rev. Lett.} {\bf 90}, 193002

2074: (2003).

2075: 

2076: 

2077: 

2078: \bibitem{Loftus2004}

2079: T. H. Loftus, T. Ido, A. D. Ludlow, M. M. Boyd and J. Ye, {\it

2080: Phys. Rev. Lett.} {\bf 93}, 073003 (2004).

2081: 

2082: 

2083: 

2084: \bibitem{Katori2001}

2085: H. Katori, T. Ido, Y. Isoya and M. Kuwata-Gonokami, in {\it Atomic

2086: Physics 17}, p. 382, edited by E. Arimondo, P. De Natale, and M.

2087: Inguscio (AIP, New York, 2001).

2088: 

2089: 

2090: 

2091: \bibitem{Takamoto2003}

2092: M. Takamoto and H. Katori, {\it Phys. Rev. Lett.} {\bf 91}, 223001

2093: (2003).

2094: 

2095: 

2096: 

2097: \bibitem{Courtillotclock}

2098: I. Courtillot, A. Quessada, R. P. Kovacich, A. Brusch, D. Kolker,

2099: J.-J. Zondy, G. D. Rovera and P. Lemonde, {\it Phys. Rev. A} {\bf

2100: 68}, 030501 (R) (2003).

2101: 

2102: 

2103: 

2104: 

2105: \bibitem{Ferrari2003}

2106: G. Ferrari, P. Cancio, R. Drullinger, G. Giusfredi, N. Poli, M.

2107: Prevedelli, C. Toninelli and G. M. Tino, {\it Phys. Rev. Lett.}

2108: {\bf 91}, 243002 (2003).

2109: 

2110: 

2111: 

2112: \bibitem{Ido2005}

2113: T. Ido, T. H. Loftus, M. M. Boyd, A. D. Ludlow, K. W. Holman, and

2114: J. Ye, {\it Phys. Rev. Lett.} {\bf 94}, 153001 (2005).

2115: 

2116: 

2117: 

2118: 

2119: 

2120: \bibitem{Bidel2002}

2121: Y. Bidel, B. Klappauf, J. C. Bernard, D. Delande, G. Labeyrie, C.

2122: Miniatura, D. Wilkowski and R. Kaiser, {\it Phys. Rev. Lett.} {\bf

2123: 88}, 203902 (2002).

2124: 

2125: 

2126: 

2127: \bibitem{Derevianko2003}

2128: A. Derevianko, S. G. Porsev, S. Kotochigova, E. Tiesinga and P. S.

2129: Julienne, {\it Phys. Rev. Lett.} {\bf 90}, 063002 (2003).

2130: 

2131: 

2132: 

2133: 

2134: \bibitem{Raab1987}

2135: E. L. Raab, M. Prentiss, A. Cable, S. Chu and D. Pritchard, {\it

2136: Phys. Rev. Lett.} {\bf 59}, 2631 (1987).

2137: 

2138: 

2139: 

2140: \bibitem{Kurosu1990}

2141: T. Kurosu and F. Shimizu, {\it Jpn. J. Appl. Phys.} {\bf 29},

2142: L2127 (1990).

2143: 

2144: 

2145: 

2146: 

2147: \bibitem{Chanelière2004}

2148: T. Chaneli\'ere, D. Wilkowski, Y. Bidel, R. Kaiser and C.

2149: Miniatura, {\it Phys. Rev. E.} {\bf 70}, 036602 (2004)

2150: 

2151: 

2152: 

2153: \bibitem{FerrariBloch}

2154: G. Ferrari, N. Poli, F. Sorrentino and G. Tino, {\it Phys. Rev.

2155: Lett.} {\bf 97}, 060402 (2006).

2156: 

2157: 

2158: 

2159: 

2160: \bibitem{KatoriFORT1999}

2161: H. Katori, T. Ido and M. K.-Gonokami, {\it J. Phys. Soc. Jpn.}

2162: {\bf 68}, 2479 (1999).

2163: 

2164: 

2165: 

2166: \bibitem{Ido2000}

2167: T. Ido, Y. Isoya and H. Katori, {\it Phys. Rev. A} {\bf 61},

2168: 061403 (2000).

2169: 

2170: 

2171: 

2172: 

2173: \bibitem{Takamoto2005}

2174: M. Takamoto, F.-L. Hong, R. Higashi and H. Katori, {\it Nature}

2175: {\bf 435}, 321 (2005).

2176: 

2177: 

2178: 

2179: \bibitem{Katori2006}

2180: T. Kishimoto, H. Hachisu, J. Fujiki, K. Nagato, M. Yasuda, and H.

2181: Katori, {\it Phys. Rev. Lett.} {\bf 96}, 123001 (2006).

2182: 

2183: 

2184: 

2185: 

2186: 

2187: \bibitem{Lemonde2005}

2188: P. Lemonde and P. Wolf, {\it Phys. Rev. A} {\bf 72}, 033409

2189: (2005).

2190: 

2191: 

2192: 

2193: \bibitem{Targat2006}

2194: R. Le Targat, X. Baillard, M. Fouch\'e, A. Brusch, O. Tcherbakoff,

2195: G. D. Rovera and P. Lemonde, arXiv:physics/0605200 (2006).

2196: 

2197: 

2198: 

2199: 

2200: \bibitem{Nagel2005}

2201: S. B. Nagel, P. G. Mickelson, A. D. Saenz, Y. N. Martinez, Y. C.

2202: Chen, T. C. Killian, P. Pellegrini and R. C\^ot\'e,  {\it Phys.

2203: Rev. Lett.} {\bf 94}, 083004 (2005).

2204: 

2205: 

2206: 

2207: \bibitem{Mickelson2005}

2208: P. G. Mickelson, Y. N. Martinez, A. D. Saenz, S. B. Nagel, Y. C.

2209: Chen, T. C. Killian, P. Pellegrini and R. C\^ot\'e, {\it Phys.

2210: Rev. Lett.} {\bf 95}, 223002 (2005).

2211: 

2212: 

2213: 

2214: 

2215: \bibitem{Chanelière2005}

2216: T. Chaneli\'ere, J.-L. Meunier, R. Kaiser, C. Miniatura and D.

2217: Wilkowski, arXiv:physics/0412119v1 (2004).

2218: 

2219: 

2220: 

2221: \bibitem{Neuman1995}

2222: J. A. Neuman, P. Wang and A. Gallagher, {\it Rev. Sci. Instrum.}

2223: {\bf 66}, 3021 (1995)

2224: 

2225: 

2226: 

2227: \bibitem{LibroFisicaAtomica}

2228: A. Corney, {\it Atomic and laser spectroscopy}, Oxford, New York:

2229: Clarendon Press (1977).

2230: 

2231: 

2232: 

2233: \bibitem{Grunert}

2234: J. Gr\"unert and A. Hemmerich, {\it Phys. Rev. A} {\bf 65}, 041401

2235: (2002).

2236: 

2237: 

2238: 

2239: \bibitem{Oates1999}

2240: C.W. Oates, F. Bondu and L. Hollberg, {\it Eur. Phys. J. D} {\bf

2241: 7}, 449 (1999).

2242: 

2243: 

2244: 

2245: \bibitem{Prodan1982}

2246: J. V. Prodan, W. D. Phillips and H. Metcalf, {\it Phys. Rev.

2247: Lett.} {\bf 49}, 1149 (1982).

2248: 

2249: 

2250: 

2251: \bibitem{Shimizu}

2252: F. Shimizu, K. Shimizu and H. Takuma, {\it Chem. Phys.} {\bf 145},

2253: 327 (1990).

2254: 

2255: 

2256: 

2257: \bibitem{Rasel1999}

2258: E. Rasel, F. Pereira Dos Santos, F. S. Pavone, F. Perales, C. S.

2259: Unnikrishnan and M. Leduc, {\it Eur. Phys. J. D} {\bf 7}, 311

2260: (1999).

2261: 

2262: 

2263: 

2264: \bibitem{Vogel1999}

2265: T. P. Dinnen, K. R. Vogel, E. Arimondo, J. L. Hall and A.

2266: Gallagher, {\it Phys. Rev. A} {\bf 59}, 1216 (1999).

2267: 

2268: 

2269: 

2270: \bibitem{KatoriFermion}

2271: T. Mukaiyama, H. Katori, T. Ido, Y. Li and M. Kuwata-Gonokami,

2272: {\it Phys. Rev. Lett.} {\bf 90}, 113002 (2003).

2273: 

2274: 

2275: 

2276: \bibitem{Binnewies2001}

2277: T. Binnewies, G. Wilpers, U. Sterr, F. Rihele, J. Helmke, T.E.

2278: Mehlst\"aubler, E.M. Rasel and W. Ertmer, {\it Phys. Rev. Lett.}

2279: {\bf 87}, 123002 (2001).

2280: 

2281: 

2282: 

2283: \bibitem{Santos1995}

2284: M. S. Santos, P. Nussenzveig, L. G. Marcassa, K. Helmerson, J.

2285: Flemming, S. C. Zilio and V. S. Bagnato, {\it Phys. Rev. A} {\bf

2286: 52}, R4340 (1995).

2287: 

2288: 

2289: 

2290: \bibitem{Modugno2001}

2291: G. Modugno, G. Ferrari, G. Roati, R. J. Brecha, A. Simoni, M.

2292: Inguscio, {\it Science} {\bf 294}, 1320 (2001).

2293: 

2294: 

2295: \bibitem{Murdrich2002}

2296: M. Murdrich, S. Kraft, K. Singer, R. Grimm, A. Mosk and M.

2297: Weidemüller, {\it Phys. Rev. Lett.} {\bf 88}, 253001 (2002).

2298: 

2299: 

2300: 

2301: 

2302: \bibitem{Mewesl999}

2303: M.-O. Mewes, G. Ferrari, F. Schreck, A. Sinatra and C. Salomon,

2304: {\it Phys. Rev. A} {\bf 61}, 011403(R) (1999).

2305: 

2306: 

2307: \bibitem{Loftus2001}

2308: T. Loftus, J. R. Bochinski and T. W. Mossberg, {\it Phys. Rev. A}

2309: {\bf 63}, 053401 (2001).

2310: 

2311: 

2312: 

2313: \bibitem{Honda2002}

2314: K. Honda, Y. Takasu, T. Kuwamoto, M. Kumakura, Y. Takahashi and T.

2315: Yabuzaki, {\it Phys. Rev. A} {\bf 66}, 021401(R) (2002).

2316: 

2317: 

2318: 

2319: \bibitem{Stas2004}

2320: R. J. W. Stas, J. M. McNamara, W. Hogervorst and W. Vassen, {\it

2321: Phys. Rev. Lett.} {\bf 93}, 053001 (2004)

2322: 

2323: 

2324: 

2325: \bibitem{Schreck2001}

2326: F. Schreck, G. Ferrari, K. L. Corwin, J. Cubizolles, L.

2327: Khaykovich, M.-O. Mewes and C. Salomon, {\it Phys. Rev. A} {\bf

2328: 64}, 011402 (2001).

2329: 

2330: 

2331: 

2332: \bibitem{Khaykovich2001}

2333: F. Schreck, L. Khaykovich, K. L. Corwin, G. Ferrari, T. Bourdel,

2334: J. Cubizolles and C. Salomon, {\it Phys. Rev. Lett.} {\bf 87},

2335: 080403 (2001).

2336: 

2337: 

2338: 

2339: 

2340: \bibitem{Suptitz1994}

2341: W. Suptitz, G. Wokurka, F. Strauch, P. Kohns and W. Ertmer, {\it

2342: Opt. Lett.} {\bf 19}, 1571 (1994).

2343: 

2344: 

2345: 

2346: \bibitem{Stuhler2001}

2347: J. Stuhler, P. O. Schmidt, S. Hensler, J. Werner, J. Mlynek, and

2348: T. Pfau, {\it Phys. Rev. A} {\bf 64}, 031405(R) (2001).

2349: 

2350: 

2351: 

2352: \bibitem{Nagel2003}

2353: S. B. Nagel, C. E. Simien, S. Laha, P. Gupta, V. S. Ashoka and T.

2354: C. Killian, {\it Phys. Rev. A} {\bf 67}, 011401(R) (2003).

2355: 

2356: 

2357: 

2358: \bibitem{Poli2005}

2359: N. Poli, R. E. Drullinger, G. Ferrari, J. L\'eonard,F. Sorrentino

2360: and G. Tino, {\it Phys. Rev. A} {\bf 71}, 061403(R) (2005).

2361: 

2362: 

2363: 

2364: \bibitem{opticaltrapping}

2365: R. Grimm, M. Weidemüller and Y. B. Ovchinnikov,  {\it Adv. At.

2366: Mol. Opt. Phys.} {\bf 42}, 95 (2000).

2367: 

2368: 

2369: 

2370: \bibitem{Dicke}

2371: R. H. Dicke, {\it  Phys. Rev.} {\bf 89}, 472 (1953).

2372: 

2373: 

2374: 

2375: \bibitem{Ido2003}

2376: T. Ido and H. Katori, {\it Phys. Rev. Lett.} {\bf 91}, 053001

2377: (2003).

2378: 

2379: 

2380: 

2381: \bibitem{Fedichev1996}

2382: P. O. Fedichev, M. W. Reynolds and G. V. Shlyapnikov {\it Phys.

2383: Rev. Lett.} {\bf 77}, 2921 (1996).

2384: 

2385: 

2386: 

2387: \bibitem{Delannoy2001}

2388: G. Delannoy, S. G. Murdoch, V. Boyer, V. Jossel, P. Bouyer and A.

2389: Aspect, {\it Phys. Rev. A} {\bf 63}, 051602(R) (2001).

2390: 

2391: 

2392: 

2393: \bibitem{Yamashita2003}

2394: M. Yamashita and T. Mukai, {\it Phys. Rev. A} {\bf 68}, 063601

2395: (2003).

2396: 

2397: 

2398: 

2399: \bibitem{Monroe1993}

2400: Tc. R. Monroe, E. A. Cornell, C. A. Sackett, C. J. Myatt and C. E.

2401: Wieman, {\it Phys. Rev. Lett.} {\bf 70}, 414 (1993).

2402: 

2403: 

2404: 

2405: \bibitem{Arndt1997}

2406: M. Arndt, M. Ben Dahan, D. Gu\'ery-Odelin, M. W. Reynolds and J.

2407: Dalibard, {\it Phys. Rev. Lett.} {\bf 79}, 625 (1997).

2408: 

2409: 

2410: 

2411: \bibitem{Hopkins2000}

2412: S. A. Hopkins, S. Webster, J. Arlt, P. Bance, S. Cornish, O.

2413: Marago and C. J. Foot, {\it Phys. Rev. A} {\bf 61}, 032707-1

2414: (2000).

2415: 

2416: 

2417: 

2418: \bibitem{Schmidt2003}

2419: P. O. Schmidt, S. Hensler, J. Werner, A. Griesmaier, A.

2420: {G\"orlitz}, T. Pfau and A. Simoni, {\it Phys. Rev. Lett.} {\bf

2421: 91}, 193201 (2003).

2422: 

2423: 

2424: 

2425: \bibitem{Odelin1999}

2426: D. Gu\'ery-Odelin, F. Zabelli, J. Dalibard and S. Stringari, {\it

2427: Phys. Rev. A} {\bf 60}, 4851 (1999).

2428: 

2429: 

2430: \bibitem{Gensemer2001}

2431: S. D. Gensemer and D. S. Jin, {\it Phys. Rev. Lett.} {\bf 87},

2432: 173201 (2001).

2433: 

2434: 

2435: 

2436: 

2437: \bibitem{Kavoulakis2000}

2438: G. M. Kavoulakis, C. J. Pethick and H. Smith, {\it Phys. Rev. A}

2439: {\bf 61}, 053603 (2000).

2440: 

2441: 

2442: 

2443: \bibitem{yasuda2004}

2444: M. Yasuda, T. Kishimoto, M. Takamoto and H. Katori, {\it Phys.

2445: Rev. A} {\bf 73}, 011403(R) (2006).

2446: 

2447: 

2448: 

2449: \bibitem{Goldwin}

2450: J. Goldwin, S. Inouye, M. L. Olsen, B. Newman, B. D. DePaola and

2451: D. S. Jin, {\it Phys. Rev. A} {\bf 70}, 021601 (2004).

2452: 

2453: 

2454: 

2455: \bibitem{K-Rb}

2456: G. Ferrari, M. Inguscio, W. Jastrzebski, G. Modugno, G. Roati and

2457: A. Simoni, {\it Phys. Rev. Lett.} {\bf 89}, 053202 (2002).

2458: 

2459: 

2460: 

2461: \bibitem{Mosk}

2462: A. Mosk, S. Kraft, M. Murdrich, K. Singer, W. Wohlleben, R. Grimm

2463: and M. {Weidem\"uller}, {\it Appl. Phys. B} {\bf 73}, 791 (2004).

2464: 

2465: 

2466: 

2467: \bibitem{Delannoy}

2468: G. Delannoy, S. G. Murdoch, V. Boyer, V. Josse, P. Bouyer and A.

2469: Aspect, {\it Phys. Rev A} {\bf 63}, 051602(R) (2001).

2470: 

2471: 

2472: 

2473: \bibitem{Burt}

2474: E. A. Burt, R. W. Ghrist, C. J. Myatt, M. J. Holland, E. A.

2475: Cornell and C. E. Wieman, {\it Phys. Rev. Lett.} {\bf 79}, 337

2476: (1997).

2477: 

2478: 

2479: \bibitem{Hess1986}

2480: H. F. Hess, {\it Phys. Rev. B} {\bf 34}, (R)3476 (1986).

2481: 

2482: 

2483: 

2484: \bibitem{Takasu2003}

2485: Y. Takasu, K. Maki, K. Komori, T. Takano, K. Honda, M. Kumakura,

2486: T. Yabuzaki and Y. Takahashi, {\it Phys. Rev. Lett.} {\bf 91},

2487: 040404 (2003).

2488: 

2489: 

2490: 

2491: \bibitem{Weber2002}

2492: T. Weber, J. Herbig, M. Mark, H.-C. Naegerl and R. Grimm, {\it

2493: Science} {\bf 299}, 232 (2002).

2494: 

2495: 

2496: 

2497: 

2498: \bibitem{Weiss2003}

2499: T. Kinoshita, T. Wenger and D. S. Weiss, {\it Phys. Rev. A} {\bf

2500: 71}, 011602(R) (2005).

2501: 

2502: 

2503: 

2504: \bibitem{Peters1999}

2505: A. Peters, K. Y. Chung and S. Chu, {\it Nature} {\bf 400}, 849

2506: (1999).

2507: 

2508: 

2509: 

2510: \bibitem{Gustavson2000}

2511: T. L. Gustavson, A. Landragin and M. A. Kasevich, {\it Class.

2512: Quantum Grav.} {\bf 17}, 2385 (2000).

2513: 

2514: 

2515: 

2516: \bibitem{Wicht2002}

2517: A. Wicht, J. M. Hensley, E. Sarajlic and S. Chu, {\it Physica

2518: Scripta} {\bf 102}, 82 (2002).

2519: 

2520: 

2521: 

2522: \bibitem{Clade2006}

2523: P. Clad\'e,  E. de Mirandes, M. Cadoret, S. Guellati-Kh\'elifa, C.

2524: Schwob, F. Nez, L. Julien and F. Biraben1, {\it Phys. Rev. Lett.}

2525: {\bf 96}, 033001 (2006).

2526: 

2527: 

2528: 

2529: \bibitem{Stuhler2003}

2530: J. Stuhler, M. Fattori, T. Petelski and G. M. Tino, {\it J. Opt.

2531: B: Quantum Semiclass. Opt.} {\bf 5}, S75 (2003).

2532: 

2533: 

2534: 

2535: 

2536: 

2537: \bibitem{Fray2004}

2538: S. Fray, C. A. Diez, T. W. H\"ansch and M. Weitz, {\it Phys. Rev.

2539: Lett.} {\bf 93}, 240404 (2004).

2540: 

2541: 

2542: 

2543: 

2544: 

2545: \bibitem{Antezza2005}

2546: M. Antezza, L. P. Pitaevskii and S. Stringari, {\it Phys. Rev.

2547: Lett.} {\bf 95}, 113202 (2005).

2548: 

2549: 

2550: 

2551: \bibitem{Long2003}

2552: J. C. Long, H. W. Chan, A. B. Churnside, E. A. Gulbis, M. C. M.

2553: Varney and J. C. Price, {\it Nature} {\bf 421} (2003).

2554: 

2555: 

2556: \bibitem{Dimopoulos2003}

2557: S. Dimopoulos and A. A. Geraci, {\it Phys. Rev. D} {\bf 68},

2558: 124021 (2003).

2559: 

2560: 

2561: \bibitem{Samullin2005}

2562: S. J. Samullin, A. A. Geraci, D. M. Weld, J. Chiaverini, S. Holmes

2563: and A. Kapitulnik, {\it Phys. Rev. D} {\bf 72}, 122001 (2005).

2564: 

2565: 

2566: 

2567: \bibitem{Bloch2005}

2568: I. Bloch, {\it Nature Phys.} {\bf 1}, 23 (2005).

2569: 

2570: 

2571: 

2572: \bibitem{Bloch1929}

2573: F. Bloch, {\it Z. Phys.} {\bf 52}, 555 (1929).

2574: 

2575: 

2576: 

2577: \bibitem{Raizen1997}

2578: M. Raizen, C. Salomon  and Q. Niu, {\it Physics Today} {\bf 50}

2579: No. 7, 30 (1997).

2580: 

2581: 

2582: 

2583: \bibitem{Anderson1998}

2584: B. P. Anderson, M. A. Kasevich,{\it Science} {\bf 282}, 1686

2585: (1998).

2586: 

2587: 

2588: 

2589: \bibitem{Roati2004}

2590: G. Roati,  E. de Mirandes, F. Ferlaino, H. Ott, G. Modugno and M.

2591: Inguscio, {\it Phys. Rev. Lett.} {\bf 92}, 230402 (2004).

2592: 

2593: 

2594: 

2595: %\bibitem{Petelski}

2596: %T. Petelski, {\it Ph.D. Thesis, Fienze} (2004).

2597: 

2598: 

2599: 

2600: \bibitem{Udem1999}

2601: Th. Udem, J. Reichert, R. Holzwarth and T.W. Ha¨nsch, {\it Opt.

2602: Lett.} {\bf 24}, 881 (1999).

2603: 

2604: 

2605: 

2606: \bibitem{Diddams2000}

2607: S. A. Diddams, D. J. Jones, J. Ye, S. T. Cundiff, J. L. Hall, J.

2608: K. Ranka, R. S. Windeler, R. Holzwarth, T. Udem and T. W. H\"ansch

2609: {\it Phys. Rev. Lett.} {\bf 84}, 5102 (2000).

2610: 

2611: 

2612: 

2613: \bibitem{Udem2002}

2614: Th. Udem, R. Holzwarth and T.W. H\"ansch, {\it Nature} (London)

2615: {\bf 416}, 233 (2002).

2616: 

2617: 

2618: 

2619: \bibitem{Udem2001}

2620: T. Udem, S. A. Diddams, K. R. Vogel, C. W. Oates, E. A. Curtis, W.

2621: D. Lee, W. M. Itano, R. E. Drullinger, J. C. Bergquist and L.

2622: Hollberg, {\it Phys. Rev. Lett.}  {\bf 86}, 4996 (2001) and

2623: references therein.

2624: 

2625: 

2626: 

2627: \bibitem{Gill2005}

2628: P. Gill, {\it Metrologia} {\bf 42}, S125 (2005).

2629: 

2630: 

2631: 

2632: \bibitem{Katori2003}

2633: H. Katori, M. Takamoto, V. G. Pal'chikov and V. D. Ovsiannikov,

2634: {\it Phys. Rev. Lett.} {\bf 91}, 173005 (2003).

2635: 

2636: 

2637: 

2638: \bibitem{Hall1989}

2639: J. L. Hall, M. Zhu and P. Buch, {\it J. Opt. Soc. Am. B} {\bf 6},

2640: 2194 (1989).

2641: 

2642: 

2643: 

2644: \bibitem{Tino1992}

2645: G. M. Tino, M. Barsanti, M. de Angelis, L. Gianfrani and M.

2646: Inguscio, {\it Appl. Phys. B} {\bf 55}, 397 (1992).

2647: 

2648: 

2649: 

2650: 

2651: \bibitem{Ludlow2005}

2652: A. Ludlow, M. M. Boyd, T. Zelevinsky, S. M Foreman, S. Blatt, M.

2653: Notcutt, T. Ido and J. Ye, arXiv:physics/0508041 (2005).

2654: 

2655: 

2656: 

2657: \bibitem{Hoyt2005}

2658: C. W. Hoyt, Z. W. Barber, C. W. Oates, T. M. Fortier, S. A.

2659: Diddams and L. Hollberg, arXiv:physics/0503240 (2005).

2660: 

2661: 

2662: 

2663: \bibitem{Santra2005}

2664: R. Santra, E. Arimondo, T. Ido, C. H. Greene and J. Ye, {\it Phys.

2665: Rev. Lett.} {\bf 94}, 173002 (2005).

2666: 

2667: 

2668: 

2669: \bibitem{Hong2005}

2670: T. Hong, C. Cramer, E. Cook, W. Naugourney and E. N. Fortson, {\it

2671: Phys. Rev. Lett.} {\bf 94}, 050801 (2005).

2672: 

2673: 

2674: 

2675: \bibitem{Taichenachev2005}

2676: A. V. Taichenachev, V. I. Yudin, C. W. Oates, C. W. Hoyt, Z. W.

2677: Barber and L. Hollberg, {\it Phys. Rev. Lett.} {\bf 96}, 083001

2678: (2006).

2679: 

2680: 

2681: 

2682: \bibitem{Barber2005}

2683: Z. W. Barber, C. W. Hoyt, C. W. Oates, L. Hollberg, A. V.

2684: Taichenachev and V. I. Yudin, {\it Phys. Rev. Lett.} {\bf 96},

2685: 083002  (2006).

2686: 

2687: 

2688: 

2689: 

2690: 

2691: \bibitem{Cundiff2001}

2692: S. T. Cundiff, J. Ye, and J. L. Hall, {\it Rev. Sci. Instr.} {\bf

2693: 72}, 3749 (2001).

2694: 

2695: 

2696: 

2697: %\bibitem{Minardi1999}

2698: %F. Minardi et al., {\it Phys. Rev. A} {\bf 60}, 4164 (1999); M.

2699: %Artoni, I. Carusotto and F. Minardi, {\it Phys. Rev. A} {\bf 62},

2700: %023402 (2000).

2701: 

2702: 

2703: 

2704: %\bibitem{Hall1976}

2705: %J. L. Hall and C. J. Borde´, {\it Appl. Phys. Lett.} {\bf 29}, 788

2706: %(1976).

2707: 

2708: 

2709: 

2710: %\bibitem{Crane1994}

2711: %J. K. Crane, M. J. Shaw and R.W. Presta, {\it Phys. Rev. A} {\bf

2712: %49}, 1666 (1994).

2713: 

2714: 

2715: 

2716: \bibitem{Buchinger1985}

2717: F. Buchinger, R. Corriveau, E. B. Ramsay, D. Berdichevsky and D.

2718: W. L. Sprung, {\it Phys. Rev. C} {\bf 32}, 2058 (1985).

2719: 

2720: 

2721: 

2722: %\bibitem{Young1999}

2723: %B. C. Young, F. C. Cruz, W. M. Itano and J. C. Bergquist, {\it

2724: % Phys. Rev. Lett.} {\bf 82}, 3799 (1999).

2725: 

2726: 

2727: 

2728: \bibitem{Courtillot2003} I. Courtillot, A. Quessada, R. P.

2729: Kovacich, J. J. Zondy, A. Landragin, A. Clairon and P. Lemonde,

2730: {\it Opt. Lett.} {\bf 28}, 468 (2003).

2731: 

2732: 

2733: 

2734: \bibitem{WindowSeal}

2735: A. Noble and M. Kasevich, {\it Rev. Sci. Instr.} {\bf 65}, 3042

2736: (1994).

2737: 

2738: 

2739: 

2740: \bibitem{Littrow1991}

2741: C. E. Wieman and L. Hollberg,{\it Rev. Sci. Instr.} {\bf 62}, 1

2742: (1991).

2743: 

2744: 

2745: 

2746: \bibitem{Couillaud81}

2747: B. Couillaud and T.W. H\"{a}nsch, {\it Opt. Comm.} {\bf 35}, 441

2748: (1981).

2749: 

2750: 

2751: 

2752: \bibitem{Pound-Drever-Hall}

2753: R. W. P. Drever, J. L. Hall, F. V. Kowalski, J. Hough, G. M. Ford,

2754: A. J. Munley and H. Ward, {\it App. Phys B} {\bf 31}, 97 (1983).

2755: 

2756: 

2757: \bibitem{Ferrari2004}

2758: G. Ferrari, T. Brzozowski, R. E. Drullinger, N. Poli, M.

2759: Prevedelli, C. Toninelli and G. M. Tino, {\it Proc. SPIE } {\bf

2760: 5478}, 5.3 (2004).

2761: 

2762: 

2763: \bibitem{Poli2006}

2764: N. Poli, G. Ferrari, M. Prevedelli, F. Sorrentino, R. E.

2765: Drullinger and G. M. Tino, {\it Spectrochim. Acta Part A} {\bf

2766: 63}, 981 (2006).

2767: 

2768: 

2769: %\bibitem{ClaironStabilization}

2770: %L.  Hilico, D. Touahri, F. Nez and A. Clairon, {\it Rev. Sci.

2771: %Instr.}{\bf 65}, 3628 (1994).

2772: 

2773: \bibitem{UltraStableCavity}

2774: B. C. Young, F. C. Cruz, W. M. Itano and J. C. Bergquist, {\it

2775: Phys. Rev. Lett.} {\bf 82}, 3799 (1999).

2776: 

2777: 

2778: 

2779: 

2780: 

2781: %\bibitem{ProceedingStPetersburg}

2782: %G. Ferrari, T. Brzozowski, R. E. Drullinger, N. Poli, M.

2783: %Prevedelli, C. Toninelli and G. M. Tino, {\it Laser Optics 2003:

2784: %Solid State Lasers and Nonlinear Frequency Conversion}, V. I.

2785: %Ustugov ed., proceedings of SPIE Vol. 5478, p. 210 (SPIE,

2786: %Bellingham, WA).

2787: 

2788: 

2789: 

2790: 

2791: 

2792: 

2793: 

2794: 

2795: 

2796: \end{thebibliography}

2797: 

2798: 

2799: 

2800: \end{document}

2801: