A prototype industrial laser system for cold atom inertial sensing in space
Romain Caldani, Sébastien Merlet, Franck Pereira Dos Santos, Guillaume Stern, Anne-Sophie Martin, Bruno Desruelle, Vincent Ménoret
EEPJ manuscript No. (will be inserted by the editor)
A prototype industrial laser system for cold atom inertial sensingin space
Romain Caldani , S´ebastien Merlet , Franck Pereira Dos Santos and Guillaume Stern , Anne-Sophie Martin ,Bruno Desruelle , Vincent M´enoret LNE-SYRTE, Observatoire de Paris, Universit´e PSL, CNRS, Sorbonne Universit´e, 61 avenue de l’Observatoire, 75014 Paris,France MUQUANS, Institut d’Optique d’Aquitaine, rue Fran¸cois Mitterrand, 33400 Talence, FranceReceived: date / Revised version: date
Abstract.
We present the design, realization, characterization and testing of an industrial prototype of alaser system, which is based on frequency doubling of telecom lasers and features all key functionalitiesto drive a cold atom space gradiometer based on the architecture proposed in [1]. Testing was performedby implementing the laser system onto a ground based atomic sensor currently under development. Thesystem reaches a Technology Readiness Level (TRL) of 4, corresponding to an operational validation ina controlled environment. The optical architecture of the system can be adapted to other space missionscenarios.
PACS.
PACS-key discribing text of that key – PACS-key discribing text of that key
Cold atom inertial sensors have demonstrated excellentperformances in terms of sensitivity, stability and accu-racy [2,3,4,5,6,7,8,9]. In space, cold atom clouds in free-fall have a null velocity with respect to the science cham-ber, meaning that the interrogation time 2 T is not limitedby the size of the setup as it is in ground-based instru-ments, but rather by the residual expansion of the clouddue to its finite temperature. Sensitivity, which scales like T in most cold atom inertial sensors, can therefore begreatly improved in microgravity environments [10,11,12].This makes atomic sensors interesting alternatives to clas-sical instruments for future space missions dedicated tofundamental physics or geodesy [13,14,15,16,17,18].Most demonstrations of high performance atomic in-ertial sensing are laboratory-based experiments that donot meet the high level of qualification and testing re-quired for an operation in space. Over the last few years,significant efforts have been conducted to develop mobileinstruments that can be operated outside the lab [10,11,19,20]. Each of these demonstrations required to improvethe compacity and reliability of the laser system, whichis a key subsystem of a cold atom sensor [21,22,23,24,25,26,27]. In the perspective of future space missions, lasersystems will require further improvements, especially interms of reliability and resistance to external conditionssuch as temperature, vibrations and radiations. a E-mail: [email protected]
In this paper, we present the design, realization andcharacterization of a prototype industrial laser system (ILS)designed to meet the scientific goals of a space-based grav-ity gradiometry mission using cold Rubidium atoms [1,13]. The system is based on frequency doubling of telecomlasers operating at 1560 nm. This technology is capable ofmeeting both the scientific and operational specificationsof the mission.We first briefly recall the specifications of the laserbased on the mission concept study, and describe the lasersystem focusing both on scientific performance and ontechnological choices that make it suited to a future spacemission. We then show the results of optical characteriza-tions. Finally, the laser was evaluated on a ground-basedatomic gravity sensor.
This work uses as a baseline the Cold Atom Interferome-ter (CAI) gradiometry mission originally proposed in [13].In this space mission scenario, cold atom interferometersare used in a differential configuration to enable high sen-sitivity measurements of the Earth’s gravity gradient. Thespecifications of the laser system were derived from thatof the atom interferometry sequence [1]. Namely, it shouldfeature several frequency-stabilized fibered outputs usedfor Rubidium cooling, detection, Bloch elevator and Ra- a r X i v : . [ a s t r o - ph . I M ] A ug R. Caldani et al.: A prototype industrial laser system for cold atom inertial sensing in space
Table 1.
Main specifications for the laser system. Raman de-tuning is from the | F (cid:48) = 1 (cid:105) level.Cooling Bloch RamanWavelength (nm) 780.24 780.24 780.24Power (mW) 200 200 45Detuning (MHz) [ − − < < < man interrogation (Table 1). In addition, the Raman out-put should have a narrow linewidth (10 kHz), and a highshort-term power stability (0.1% over 2 s).As proposed in [1], we selected frequency-doubled tele-com lasers to build the ILS. This approach relies on theuse of laser diodes and amplifiers operating in the telecomC-band around 1560 nm and frequency-doubled to 780 nmin nonlinear crystals [28]. It benefits from the high matu-rity and reliability of telecom components, most of whichare qualified according to the Telcordia standard. It hasbeen used successfully in laboratory and mobile cold atomsensors over the last few years, meeting both optical andenvironmental constraints [20,24,29]. A significant advan-tage of this technology is the fact that it uses individualfibered components that are fusion spliced to one anotherto build the system. It is therefore convenient to replacea component by another when required and to include re-dundancy in the system by using fiber couplers to link thespare component to the rest of the system. Furthermore,this technology offers a significant level of space qualifica-tion as most of the optical components have either beenqualified specifically or have qualified alternatives [24]. The optical architecture of the ILS is based on a master-slave architecture (Fig. 1). A master laser is frequency-locked to an atomic transition with a known absolute fre-quency. Several slave lasers are in turn offset-locked on thismaster laser. Their frequency can be adjusted by tuningthe setpoint of the lock. In the following paragraphs, wedescribe each of the functions of the laser system.The master laser is a fibered laser diode emitting ap-proximately 10 mW at 1560 nm (RIO Planex), current-modulated at 5 MHz. It is frequency-doubled in a waveg-uide PPLN crystal (NEL WH-0780) and the resulting780 nm light is sent through a Rubidium spectroscopycell in a double-pass configuration. The resulting saturatedabsorption spectroscopy signal is demodulated and usedto lock the diode on the | F = 3 (cid:105) → | F (cid:48) = 3 (cid:105) / | F = 4 (cid:105) crossover of Rb by retroacting on its current. A smallfraction of the 1560 nm light is used to lock the slavelasers.The cooling and detection laser is used to cool theatoms in a MOT followed by far-detuned optical molasses,and to detect them at the output of the interferometer. ARIO laser diode is offset-locked to the master laser witha tunable setpoint that can be adjusted in order to set
Iso/Tap PPLN-WG Rb C o u p l e r CouplerIso/Tap Ph-mod EDFA PPLN-WG AOM C o u p l e r CouplerIso/Tap P h - m o d EDFA PPLN-WG Coupler AOMAOM PMUXCouplerCouplerIso/Tap EDFA PPLN-WGIso EDFA PPLN-WG PMUX Coupler AOM
MASTERCOOLINGDETECTIONBL / DKCRAMAN 1RAMAN 2 COOLINGDETECTIONBLOCHELEVATORDELTA KICK COOLINGRAMAN
Fig. 1.
Optical architecture of the laser system. Iso/Tap: op-tical isolator with tap coupler, PPLN-WG: waveguide PPLNcrystal, Rb: Rubidium cell, Ph-mod: phase modulator, EDFA:Erbium-Doped Fiber Amplifier, AOM: Acousto-Optic Modu-lator, PMUX: polarization multiplexer. the detuning of the laser with respect to the | F = 2 (cid:105) →| F (cid:48) = 3 (cid:105) transition of Rb from 0 to 120 MHz. Lightthen goes through a phase modulator (Ixblue MPZ-LN-10) that generates sidebands at approximately 6.5 GHz,one of which being tuned on the | F = 1 (cid:105) → | F (cid:48) = 2 (cid:105) tran-sition for repumping. We then amplify the signal to ap-proximately 1 W in a custom-made Erbium-Doped FiberAmplifier (EDFA) with active output power stabilization,and frequency-double it in a fiber-coupled PPLN crystal.At this stage, the power at 780 nm is 500 mW. Finally,a custom-made fibered Acousto-Optic Modulator (AOM)module is used to set the output optical power. At thefiber output, the power is 260 mW.The Bloch Elevator laser needs to be detuned fromthe master laser by approximately 100 GHz at 780 nm.We achieve this detuning by using a 1560 nm phase mod-ulator driven with a 7 GHz RF signal with a power ofapproximately 25 dBm. We then select the 7th sidebandto lock the laser with an offset of 49 GHz before frequency-doubling, resulting in a 98 GHz detuning at 780 nm. Ini-tially, it was planned to realize the delta-kick collimationof the atom source using an optical dipole trap [30], ratherthan the CAI mission atom chip itself [31,32]. This mo-tivated the generation of a laser beam with a few wattsof optical power, and our optical architecture combinesthis function with the Bloch Lattice. The 49 GHz detunedlaser is amplified to 5 W in a high-power EDFA. 4 W wouldthus have been used for the dipole trap, and the remainingpower is frequency-doubled and used for the Bloch eleva-tor. This light is split in two paths and sent through twoAOM modules that have adjustable frequencies, to createthe required optical frequencies for the two counterpropa-gating Bloch lattices. The two beams are then recombined . Caldani et al.: A prototype industrial laser system for cold atom inertial sensing in space 3 with orthogonal polarizations using a fibered polarizationmultiplexer. The output powers in the two beams are 100and 156 mW.Finally, we use two independant RIO laser diodes todrive the Raman transitions. With this configuration, wemake sure that there are no spurious sidebands during theinterferometer [33]. The Raman 1 laser is offset-locked onthe master laser with a Raman detuning of 3.4 GHz fromthe | F (cid:48) = 1 (cid:105) level to minimize spontaneous emission. It isamplified to 1 W and frequency doubled. The Raman 2laser is offset phase locked to the Raman 1 laser directlyat 780 nm. This architecture is used to suppress any phasefluctuations due to physical path separation. The two Ra-man lasers are combined in the same fiber with orthog-onal polarizations, and an AOM module is used to drivethe Raman pulses. At the output, each Raman laser hasa power of more than 400 mW.The control electronics for the ILS were designed tominimize human supervision. All the lasers can be lockedautomatically through a dedicated software, which alsochecks that the laser remains stable on the long term. Thesoftware also improves the long-term stability of the lockby adding a digital integrator stage [24]. After integration, all the outputs of the ILS were testedindividually in order to verify that they met the specifica-tions required for the CAI mission. In this section we onlydescribe the optical characterizations performed on theRaman output, which is the most critical of the system.The linewidth of the Raman 1 laser was measured byrecording a beatnote between this laser and another fullyindependant system, built with similar specifications andlocked using its own saturated absorption spectroscopysetup [20]. The wings of the spectrum are fitted witha Lorentzian function, whose FWHM is the sum of thetwo individual widths. Preliminary characterizations haveshown that the external laser used for the tests has alinewidth higher than 20 kHz. According to its specifi-cations, the Raman 1 laser should have a linewidth lowerthan 10 kHz. The beatnote has a FWHM of 49 kHz (Fig.2a), which is compatible with the specifications of the twodiodes.By using the same setup and recording the center fre-quency of the beatnote over time, we evaluated that thelong-term frequency stability of the laser was better than75 kHz rms over several days (Fig. 2 b). This is well withinthe specifications, and shows that the laser system can re-main locked over long periods of time without drifting,and without requiring human supervision. Since the Ra-man 1 laser is phase locked onto the master laser, thesemeasured frequency fluctuations originate from the lockon saturated absorption peaks. The Raman 2 output wasnot tested with this setup. Indeed, being phase-locked tothe Raman 1, it has the same frequency behaviour.Finally, we monitored the power in the two Raman out-puts over several hours, to characterize the power stabil-ity of the system (Fig. 3). Lasers were free-running during -10000 -5000 0 5000 10000-60-40-200 N o r m a li z ed po w e r ( d B ) Frequency (kHz) a ) b ) B ea t no t e f r equen cy ( k H z ) Time (days)
Fig. 2.
Beatnote between Raman 1 laser and a similar inde-pendant laser. a) Beatnote recorded on a spectrum analyzerwith a 9.1 kHz resolution bandwidth. A Lorentzian fit of thewings of the distributions gives a FWHM of 49 kHz. b) Cen-ter frequency of the beatnote recorded on a frequency counter.The standard deviation over more than 3.5 days is lower than75 kHz.
Normalized optical power
T i m e ( h o u r s )
Fig. 3.
Normalized power of the Raman 1 (red) and Raman 2(blue) outputs over several hours. Average powers are 423 mWfor Raman 1 and 415 mW for Raman 2. The two datasets wererecorded at different times. this measurement because the Raman 2 output can notbe measured on its own if it is locked. The power of bothlasers has a long term stability below 2%. Raman 2 laserhas a short-term noise (below 0.35% rms in a 100 s win-dow) which is higher than Raman 1 (0.05% rms in 100 s).The two lasers being built with identical architectures, weattribute this difference either to a higher intrinsic noisein one of the EDFA pump diodes or to different settingsin the servo-loop that controls the output power of theEDFAs. In both cases the noise is limited by residual low-frequency fluctuations and is compliant with the missionspecifications: in a 2 s window, fluctuations are well belowthe level measured in 100 s.Before using the ILS to cool and manipulate atoms,we have verified that all the other outputs have a similarbehaviour to what we have presented here.
R. Caldani et al.: A prototype industrial laser system for cold atom inertial sensing in space
The laser system was tested on a ground-based laboratorysensor under development at SYRTE. This sensor is de-signed for measurements of gravity and its gradient, basedon simultaneous atom interferometers performed on twovertically separated atomic sources [34]. At the time of thetest campaign, the setup was partly functional, with thebottom interferometer under operation only [35].We only briefly describe this setup and the measure-ment sequence here, more details can be found in [35,34].At first, we trap about 10 cold Rb atoms in 480 ms in a3D mirror magneto-optical trap (MOT), which is loadedfrom the cold beam of a 2D MOT. We then cool the atomsin a far detuned optical molasses before releasing them infree fall in the | F = 2 (cid:105) hyperfine ground state and select-ing them in the | F = 1 , m F = 0 (cid:105) state with a combinationof microwave and pusher pulses. An atom interferometeris then realized with a sequence of three counterpropagat-ing Raman pulses, equally separated in time, and alignedvertically. The interferometer phase shift is given by kgT ,where k is the effective Raman wavevector, g is the grav-itational acceleration and T is the time interval betweenconsecutive pulses. When the atoms are simply releasedafter cooling, the duration of the interferometer 2 T is lim-ited to about 160 ms by the size of the vacuum cham-ber. After the Raman interrogation, the populations inthe two output ports of the interferometer are detectedusing state selective fluorescence, out of which the inter-ferometer phase is deduced.To drive this sensor, we use the compact and simplehome-made laser system (HMLS) described in [23]. Thislaser system, based on two semiconductor diode lasers andone amplifier, generates all the laser beams required forrunning a cold atom interferometer based on Raman tran-sitions, namely laser cooling, interferometer and detectionbeams, and has been used for a number of different atominterferometer experiments [36,37,35,34].The availability of this well characterized HMLS al-lowed to validate individually each functionality of theILS, and compare the performances obtained with the twosystems. This was done by operating the sensor with theHMLS, with the laser beam of a given functionality ex-changed for the corresponding one of the ILS. We reportbelow the results of the validation and test campaign ofthe different functionalities. We first tested the detection by exchanging the detectionfiber of the HMLS with the cooling/detection output ofthe ILS. The frequency of the ILS laser was set so as tooptimize the fluorescence signal, with the phase modula-tion turned off (as there is no need for repumping). Wecontrol the MOT parameters with the HMLS.Figure 4a displays the fluorescence signal collected as afunction of the optical power in the detection light beams. a ) b )
10 10010100 N t o t ( M a t ) MOT loading (ms) T o F s i gna l ( no r m a li z ed ) P det (mW) Fig. 4.
Characterization of the detection laser. a) Detectionsignal obtained for different optical powers in the detectionlight sheets. b) Total atom number detected by the ILS (blacksquares) and HMLS (red circles) as a function of the MOTloading duration. The optical power in the detection lightsheets is set to 2.3 mW.
To compare this detection efficiency with the HMLS, wemeasured the total atom number detected with respectto the MOT loading time with both laser systems for thesame optical power in the detection of 2.3 mW. We ob-tained very similar atom numbers as displayed in Fig. 4b.In a second series of experiments, we measured the de-tection noise as a function of the detected atom number.For that purpose, atoms are prepared in an equal super-position of | F = 1 , m F = 0 (cid:105) and | F = 2 , m F = 0 (cid:105) states,using a combination of microwave and pusher pulses. Fig-ure 5 displays the Allan standard deviation of the mea-sured transition probabilities P as a function of the atomnumber. The atom number was varied by adjusting theparameters of the microwave selection sequence.The results obtained with the nominal detection powerof 2.3 mW are displayed on Fig. 5a and compared to theresults we obtained with HMLS in the same conditions,showing identical behaviours, suggesting that noise contri-butions arising from intensity and frequency fluctuationsof the lasers do not play a major role here.Such measurements were performed at different detec-tion powers of the ILS. Figure 5b displays the results ob-tained with a detection power of 4 mW, for which thenoise lies less than a factor of two above the quantumprojection noise limit, for atom numbers in the 10 − range. We actually found that for larger laser powers, thenoise on the transition probability does decrease at lowatom number, as expected, but degrades at high numberof atoms. To evaluate the laser cooling functionality, we connectedthe same cooling fiber from the ILS to the MOT fiber split-ter which distributes the beams to the 2D and 3D MOTs.Here the amplitude of phase modulation is adjusted suchas to put a few percents of power in resonance with the re-pumping line. The computer control system of the sensorwas used to trigger the DDS frequency sysnthesizer thatcontrols the frequencies (cooling and repumper) of the ILScooling laser, so as to modify the frequency of the cooling . Caldani et al.: A prototype industrial laser system for cold atom inertial sensing in space 5 a ) b )
100 100010 -3 ! P N at (kat)
100 100010 -3 ! P N at (kat) Fig. 5.
Characterization of the detection noise. a) Allan de-viations of the transition probability P as a function of thetotal number of atoms for identical optical power in the de-tection light sheets of 2.3 mW, for both the ILS (red circles)and HMLS (black squares). b) Same with the ILS only, with4mW power. The solid line displays the quantum projectionnoise limit. a ) b ) N a t ( M a t ) Detuning ! ( " unit)
100 120 140 160300040005000 1.901.952.002.05 N a t ( k a t ) Power (mW) D e t un i ng ! ( " un i t ) Fig. 6. a) Number of trapped atoms as a function of the de-tuning, expressed in units of Γ, for three different laser coolingpowers : 95mW (grey), 117mW (red) and 158mW (black). b)Optimal number of atoms (black) and corresponding optimaldetunings (grey) as a function the optical power. light during the cycle synchronously with the rest of thesequence.Figure 6 represents the number of atoms detected asa function of the detuning of the cooling line for threedifferent cooling optical powers, after a loading time of480 ms. For the same MOT parameters (frequency andoptical power corresponding to the grey squares in Fig.6 a) that we use on the SYRTE experiment with HMLS,we obtained here the same number of atoms, and twiceas much for an optical power of 158 mW. As expected,the optimal detuning is found to increase with the laserpower.To validate the capability of the ILS to cool the atomsin the far detuned molasses phase down to the tempera-ture limit of sub-Doppler cooling, we measured the tem-perature of the atoms after their release from the coolingbeams using Raman velocimetry. Using the Raman func-tionnality of the HMLS, we applied a 80 µ s long Raman π pulse after a free fall delay of about 70 ms, and recordedthe transition probability as a function of the Raman fre-quency (Fig. 7). The spectrum then corresponds to the dis-tribution of Doppler shifts, and thus to the atomic velocitydistribution. One observes as well two side resonances cor-responding to parasitic Zeeman shifted conterpropagatingtransitions. The 1 /e width of the spectrum of 70 kHzcorresponds to a temperature as low as 2 µ K. Transition probability
R a m a n d e t u n i n g ( k H z )
Fig. 7.
Raman spectrum measured after the molasses se-quence. The Gaussian fit of the main peak gives a 1 /e widthof 70 kHz, corresponding to a molasses temperature of 2 µ K. a ) b ) ⌫ t~g ~g R a t i o Chirp duration (ms)
Chirp duration C h i r p Fig. 8.
Launching the atoms using a Bloch elevator. a) Fre-quency difference between the two optical beams during thelaunch. The adiabatic loading and release phases are 200 µ slong. In between, the length of the acceleration phase can bemodified. b) Efficiency of the launch with BO for different ac-celeration phase durations. The direct measurements of thefraction of launched atoms are displayed as filled black squares.These ratios are then corrected to account for the change in thevelocity of the atoms in the detection sheets and in the cloudexpansion. Corrected data are displayed as empty squares. Next, we connected the Bloch elevator fiber output of theILS to the Raman collimator instead of the Raman beamfiber, in order to generate a Bloch lattice in the verticaldirection. In addition, a polarization beamsplitting cubewas inserted before the retroreflecting optics in order tosuppress one of the two Bloch beams before the retrore-flection so as to obtain one lattice only, instead of twolattices moving in opposite directions as proposed in [1].The frequencies of the lattice beams are controlled byadjusting the frequencies of the ILS AOMs using pro-grammable DDSs. This leads to the sequence displayedon Fig. 8a: first, the adiabatic loading of the lattice, for200 µ s, where the frequency difference between the Blochbeams matches gravity acceleration, second, the launchphase, where the chirp on the frequency difference is re-versed and much increased in order to accelerate the atomsupwards and third, the adiabatic release where the fre-quency sweep matches gravity acceleration. R. Caldani et al.: A prototype industrial laser system for cold atom inertial sensing in space
We measured the fraction of launched atoms as a func-tion of the duration of the acceleration phase for a fixednumber of 100 Bloch Oscillations (BO). Figure 8b displaysthe results as a function of the chirp duration. The bestresult is obtained for a chirp of 24 ms, which correspondsto a modest acceleration of 5 g .The measurements shown in Fig. 8b are corrected inorder to account for the finite size of the detection sheetsand the difference between launched and dropped atoms interms of velocity at the detection and ballistic expansion.Indeed, the faster the atoms, the shorter the time theyspend in the detection light sheets and the smaller thefluorescence signal.These results are comparable in terms of fraction oflaunched atoms to the ones we obtained in [35]. This frac-tion corresponds more or less to the fraction of the velocitydistribution fitting into the first Brillouin zone (of width2v r ). In [35], we obtained a corrected ratio of up to 40percent for 1.8 m . s − launch velocity, and an accelerationduration of 2 ms, owing to the higher laser power (300mW per beam instead of 100 mW) and smaller detuning(50 GHz instead 100 GHz). We finally connected the Raman fiber of the ILS to theRaman collimator. A frequency chirp α generated by oneof SYRTE’s DDS is applied to the phase lock loop of theILS Raman lasers. The output optical power of each of thetwo Raman beams was controled independently by adjust-ing the output level of the EDFAs with the ILS controlsoftware. The total power of the Raman output was con-troled by changing the RF power on the Raman AOM,using a controlled voltage provided by SYRTEs controlsystem. We modified the experiment in order to optimisethe performances for continuous gravity measurements.Indeed, as the phase difference between the Raman lasersis linked to the position of the mirror, fluctuations of thisposition due to ground vibrations can induce significantinterferometer phase noise, washing out the interferom-eter fringes. We thus used a passive isolation plateformand recorded the remaining vibration noise with a low-noise seismometer (Guralp 40T). We then post-correctedthe interferometer phase from the effect of these residualvibrations, which improves the sensitivity of the measure-ment [38,39]. In order to measure the mirror vibrations asaccurately as possible, this mirror is directly fixed on theseismometer that we placed at the top of the experiment.The distance from the atomic sample’s initial position tothe detection sheets allowed for performing interferom-eters with durations of up to 2 T = 160 ms. The totalcycling time was 460 ms.We first measured the short term phase noise of the in-terferometer as a function of the interferometer duration2 T . Results are presented on Fig. 9, where the measurednoise with and without the correction from vibration noiseis displayed with black filled and empty squares respec-tively. We observe as expected a significant increase of thephase noise with T due to the increasing impact of the (cid:1) f (rad) Fig. 9.
Interferometer phase noise as a function of the interfer-ometer duration 2 T , with (black squares) and without (opensquares) vibration correction. Table 2.
Interferometer phase noise measurements for differ-ent AI measurement conditions.Configuration 2
T τ ( π/ C σ P σ Φ /shot(ms) ( µ s) % ( × − ) (mrad)Counter-prop 160 4 26 1 . . . . vibration noise. This increase is efficiently mitigated bythe correction. Also, we find an asymptotic level of noiseat low 2 T , ie free from the impact of vibration noise, oforder of 50-100 mrad per shot. This is well above the typi-cal level of detection noise, which for our parameters, con-tributes up to σ Φ = 2 σ P,det /C = 8 mrad per shot, with σ P,det = 10 − the detection noise on the transition prob-ability and C the contrast of the interferometer ( C =26%for 2 T = 160 ms).To investigate the source of excess noise, we also per-formed AI measurements with co-propagating Raman beams,and for different interferometer durations 2 T and Ramanpulse durations τ , and compared the phase noise for co-and counter-propagating Raman beams.Table 2 presents some of these sensitivity measure-ments. The most relevant set of parameters for gravitymeasurements is the first one: it results in a sensitivity of8 . − g at 1s (82.6 µ Gal), close to the one obtained in acompact gravimeter based on a similar laser architecture[20]. It is about one order of magnitude higher than thebest sensitivity obtained with the state of the art SYRTECold Atom Gravimeter (CAG) [3] which operates witha shorter cycling time (380ms), a better isolation plate-form, a more linear and lower noise seismometer and withan anti-acoustic enclosure.The phase sensitivity in co-propagating measurementsis more than twice better for the same duration of Ramanpulses ( τ π/ ), irrespective of the interferometer duration2 T . This sensitivity improves when increasing the dura-tion of the Raman laser pulses. This indicates that the . Caldani et al.: A prototype industrial laser system for cold atom inertial sensing in space 7 (cid:1) g (cid:1) (cid:2) (cid:13) (cid:4) (cid:6) (cid:10) (cid:3) M J D (cid:1) (cid:5) (cid:8) (cid:11) (cid:9) (cid:7)(cid:12)(cid:6) (cid:10) (cid:1) (cid:2) (cid:13) (cid:4) (cid:6) (cid:10) (cid:3)
Fig. 10.
Gravity measurements (black dots), uncorrected fromtidal effects (red line). The data are averaged over 184s. Thedifference is represented in grey. noise is dominated by the contribution of the laser phasenoise at high Fourier frequency. These results are com-patible with the expected phase noise, calculated out ofthe power spectral density of Raman phase fluctuations.This is not a fundamental limitation of the system, phasenoise can be improved by optimizing the electronics andby using a longer pulse duration.We performed gravity measurements using the fourconfigurations measurement protocol we routinely imple-ment for the CAG measurement to reject most of thesystematic effects which affect the gravity measurement.This protocol, based on interleaved measurements withopposite Raman wavevector orientations k ↑ and k ↓ andtwo different Raman laser intensities, is described in de-tail in [40]. The results, displayed in Fig. 10, follow theexpected gravity variations due to tides. The differencewith the local tide model is represented in grey. The Al-lan standard deviation of this difference is displayed asa black line in Fig. 11. As discussed in [40], the correc-tion of the two-photon light shift leads to a degradationof the short-term sensitivity with respect to, for instance,an average over the four measurement configurations sim-ply corrected from tides, which is displayed as a blue line.Here, the Raman power of the ILS is sufficiently stable,meaning that fluctuations of two-photon light shifts arenot a limitation to the long-term stability of the measure-ment. Therefore, using the complete protocol alternatingover four configurations is not necessary in this case. Abasic alternation of k ↑ and k ↓ is sufficient for measure-ments that are both stable and precise, provided that thetwo-photon light shift has been previously measured. Thiswould reduce the effective cycling time and improve theshort term sensitivity. We have described a prototype industrial laser system de-signed for atom interferometry measurements, in the per-spective of future space missions with quantum inertialsensors. We identified frequency doubled telecom lasers asthe most mature technological solution. It benefits from (cid:1) g (cid:2) (cid:2) (cid:3) (cid:1) (cid:2) (cid:7) (cid:4) (cid:5) (cid:6) (cid:3) (cid:2)(cid:1) ( s ) Fig. 11.
Allan deviations of gravity fluctuations. The blackline corresponds to the residuals of Fig. 10, which are correctedfrom tides and two-photon light shift. The blue line displaysthe average over the four measurement configurations, which isalso corrected from tides, but not from two-photon light shift. the availability of a wide range of components from sev-eral suppliers, most of which are qualified to the Telcordiastandard. With this study, we reach a Technology Readi-ness Level (TRL) of 4, corresponding to the validation ofa propotype in a laboratory environment [41].The prototype industrial system presented here wasfully characterized and all of its functions were testedquantitatively on a ground-based cold atom interferom-eter. The system was found to comply with all of the mis-sion specifications, and had performances comparable tothat of a home-made laboratory laser system. Moreover,the system was operated during several weeks without re-quiring human supervision, and used to study differentialphase extraction in the gradiometer experiment [34].Most of the components used in this prototype havebeen validated individually to a higher TRL, or have acommercial alternative available which is qualified [24].The only major exception are the PPLN waveguide mod-ules used for frequency doubling. These devices will re-quire additionnal hardening before they can be used inspace. A way towards this goal is the use of micro-opticalassembly techniques to package the crystal in a device thatis compatible with space environment.Further improvement of the TRL will require testing inenvironments representative of a space mission. These in-clude vibrations, shocks, radiations, thermal cycling andoperation under vacuum. We believe that meeting envi-ronment requirements can be achieved with a system sim-ilar to the one presented here with only minor adapta-tions, for example regarding the integration of the fre-quency doubling modules.This work is a first step towards a space-qualified lasersystem for cold atom interferometry missions. We havevalidated the technological choice with a fully functionaloptical architecture in the frame of a case study for agravity gradiometry mission. The optical architecture caneasily be adapted to meet the requirements of differentmission scenarios, for example in the field of fundamentalphysics [14,15,42].
R. Caldani et al.: A prototype industrial laser system for cold atom inertial sensing in space
Acknowledgements
This work is supported by contract 4000116740/16/NL/MPfrom the European Space Agency. R.C. thanks the supportfrom LABEX Cluster of Excellence FIRST-TF(ANR-10-LABX-48-01), within the Program Investissements d’Aveniroperated by the French National Research Agency (ANR).The authors acknowledge a major contribution from theMuquans team in the integration and optimization of theILS.
Author contributions
The ILS was designed by VM, GS, BD, SM and FPDS.ASM, GS and VM (resp. RC, SM and FPDS) performedthe optical (resp. functional) characterizations. All theauthors contributed to the data analysis and manuscriptpreparation. All the authors have read and approved thefinal manuscript.
References
1. A. Trimeche, B. Battelier, D. Becker, A. Bertoldi,P. Bouyer, C. Braxmaier, E. Charron, R. Corgier, M. Cor-nelius, K. Douch, arXiv e-prints arXiv:1903.09828 (2019),
2. Z.K. Hu, B.L. Sun, X.C. Duan, M.K. Zhou, L.L. Chen,S. Zhan, Q.Z. Zhang, J. Luo, Physical Review A (2013)3. P. Gillot, O. Francis, A. Landragin, F. Pereira Dos Santos,S. Merlet, Metrologia , L15 (2014)4. F. Sorrentino, Q. Bodart, L. Cacciapuoti, Y.H. Lien,M. Prevedelli, G. Rosi, L. Salvi, G.M. Tino, Phys. Rev.A (2014)5. C. Freier, M. Hauth, V. Schkolnik, B. Leykauf,M. Schilling, H. Wziontek, H.G. Scherneck, J. Mller, A. Pe-ters, Journal of Physics: Conference Series , 012050(2016)6. D. Savoie, M. Altorio, B. Fang, L.A. Sidorenkov, R. Geiger,A. Landragin, Science Advances , eaau7948 (2018)7. R. Karcher, A. Imanaliev, S. Merlet, F.P.D. Santos, NewJournal of Physics , 113041 (2018)8. P. Asenbaum, C. Overstreet, T. Kovachy, D.D. Brown,J.M. Hogan, M.A. Kasevich, Physical Review Letters (2017)9. C. Overstreet, P. Asenbaum, T. Kovachy, R. Notermans,J.M. Hogan, M.A. Kasevich, Physical Review Letters (2018)10. B. Barrett, L. Antoni-Micollier, L. Chichet, B. Battelier,T. L´ev`eque, A. Landragin, P. Bouyer, Nature Communi-cations , 13786 (2016)11. D. Becker, M.D. Lachmann, S.T. Seidel, H. Ahlers, A.N.Dinkelaker, J. Grosse, O. Hellmig, H. M¨untinga, V. Schkol-nik, T. Wendrich et al., Nature , 391 (2018)12. E.R. Elliott, M.C. Krutzik, J.R. Williams, R.J. Thompson,D.C. Aveline, npj Microgravity , 16 (2018)13. O. Carraz, C. Siemes, L. Massotti, R. Haagmans, P. Sil-vestrin, Microgravity Science and Technology , 139(2014) 14. J.M. Hogan, D.M.S. Johnson, S. Dickerson, T. Kovachy,A. Sugarbaker, S.w. Chiow, P.W. Graham, M.A. Kase-vich, B. Saif, S. Rajendran et al., General Relativity andGravitation , 1953 (2011)15. B. Altschul, Q.G. Bailey, L. Blanchet, K. Bongs, P. Bouyer,L. Cacciapuoti, S. Capozziello, N. Gaaloul, D. Giulini,J. Hartwig et al., Advances in Space Research , 501(2015)16. S.w. Chiow, J. Williams, N. Yu, Phys. Rev. A , 063613(2015)17. J.M. Hogan, M.A. Kasevich, Phys. Rev. A , 033632(2016)18. G.M. Tino, A. Bassi, G. Bianco, K. Bongs, P. Bouyer,L. Cacciapuoti, S. Capozziello, X. Chen, M.L. Chiofalo,A. Derevianko et al., arXiv e-prints arXiv:1907.03867(2019),
19. Y. Bidel, N. Zahzam, C. Blanchard, A. Bonnin,M. Cadoret, A. Bresson, D. Rouxel, M.F. Lequentrec-Lalancette, Nature Communications , 627 (2018)20. V. M´enoret, P. Vermeulen, N. Le Moigne, S. Bonvalot,P. Bouyer, A. Landragin, B. Desruelle, Scientific Reports , 123000 (2018)21. P. Cheinet, F. Pereira Dos Santos, T. Petelski, J. Le Gou¨et,J. Kim, K. Therkildsen, A. Clairon, A. Landragin, AppliedPhysics B , 643 (2006)22. M. Schmidt, M. Prevedelli, A. Giorgini, G.M. Tino, A. Pe-ters, Applied Physics B , 11 (2011)23. S. Merlet, L. Volodimer, M. Lours, F. Pereira Dos Santos,Applied Physics B , 749 (2014)24. T. L´ev`eque, L. Antoni-Micollier, B. Faure, J. Berthon, Ap-plied Physics B , 997 (2014)25. V. Schkolnik, O. Hellmig, A. Wenzlawski, J. Grosse,A. Kohfeldt, K. D¨oringshoff, A. Wicht, P. Windpassinger,K. Sengstock, C. Braxmaier et al., Applied Physics B ,217 (2016)26. Y. Cheng, K. Zhang, L.L. Chen, W.J. Xu, Q. Luo, M.K.Zhou, Z.K. Hu, AIP Advances , 095211 (2017)27. X. Zhang, J. Zhong, B. Tang, X. Chen, L. Zhu, P. Huang,J. Wang, M. Zhan, Applied Optics , 6545 (2018)28. R.J. Thompson, M. Tu, D.C. Aveline, N. Lundblad,L. Maleki, Opt. Express , 1709 (2003)29. F. Theron, Y. Bidel, E. Dieu, N. Zahzam, M. Cadoret,A. Bresson, Optics Communications , 152 (2017)30. T. Kovachy, J.M. Hogan, A. Sugarbaker, S.M. Dickerson,C.A. Donnelly, C. Overstreet, M.A. Kasevich, Phys. Rev.Lett. , 143004 (2015)31. H. M¨untinga, H. Ahlers, M. Krutzik, A. Wenzlawski,S. Arnold, D. Becker, K. Bongs, H. Dittus, H. Duncker,N. Gaaloul et al., Phys. Rev. Lett. , 093602 (2013)32. R. Corgier, S. Amri, W. Herr, H. Ahlers, J. Rudolph,D. Gu´ery-Odelin, E.M. Rasel, E. Charron, N. Gaaloul,New Journal of Physics , 055002 (2018)33. O. Carraz, R. Charri`ere, M. Cadoret, N. Zahzam, Y. Bidel,A. Bresson, Phys. Rev. A , 033605 (2012)34. R. Caldani, K.X. Weng, S. Merlet, F. Pereira Dos Santos,Physical Review A , 033601 (2019)35. M. Langlois, R. Caldani, A. Trimeche, S. Merlet,F. Pereira dos Santos, Physical Review A (2017)36. J. Lautier, L. Volodimer, T. Hardin, S. Merlet, M. Lours,F.P.D. Santos, A. Landragin, Appl. Phys. Lett. ,144102 (2014)37. A. Trimeche, M. Langlois, S. Merlet, F. Pereira Dos Santos,Physical Review Applied (2017). Caldani et al.: A prototype industrial laser system for cold atom inertial sensing in space 938. J. Le Gou¨et, T. Mehlsta¨ubler, J. Kim, S. Merlet, A. Cla-iron, A. Landragin, F. Pereira Dos Santos, Applied PhysicsB , 133 (2008)39. S. Merlet, J. Le Gou¨et, Q. Bodart, A. Clairon, A. Landra-gin, F. Pereira Dos Santos, P. Rouchon, Metrologia , 87(2009)40. A. Louchet-Chauvet, T. Farah, Q. Bodart, A. Clairon,A. Landragin, S. Merlet, F.P. Dos Santos, New Journalof Physics , 065025 (2011)41. European Space Agency, Technology readinesslevels handbook for space applications (2008), https://artes.esa.int/sites/default/files/TRL Handbook.pdf
42. W. Ertmer, C. Schubert, T. Wendrich, M. Gilowski,M. Zaiser, T.v. Zoest, E. Rasel, C.J. Bord´e, A. Clairon,Landragin et al., Experimental Astronomy23