A cold atom interferometric inertial sensor - ESAsci2.esa.int/hyper/HyperSymposium1/HYPERCD/hyper-sympLandragin.pdfA cold atom interferometric inertial sensor André ... Patrick Cheinet

Observatoire de Paris BNM-SYRTE

A cold atom interferometric inertial sensor

André ClaironNoël DimarcqDavid HollevilleArnaud Landragin

Patrick CheinetJérôme FilsFlorence Yver

Philippe Bouyer, Laboratoire Charles Fabry de l’I.O.Christian Bordé, Laboratoire de Physique des Lasers

DGA, CNES, BNM, CNRS, SAGEM


Out line

• Introduction

•Presentation of our atomic inertial sensor

• Measure of the Raman laser phase noise

• Test of rejection of acceleration noise

• Conclusion


Atom interometry as long term stableinertial sensors

� Inertial NavigationSub-marine, satellites, …

� Test of fundamental physicsLense -Thirring, equivalence principle, �

� Geophysics measurementfluctuations of the rotation of the earth (global, tidal, seismic... )


An atomic interferometer

Caesium Atomic source

detection

State selection

MF = 0Interrogation zone

time 2T

Stimulated Raman TransitionRecoil velocity �keff/m

0,0

0,5

1,0

-15 -10 -5 0 5 100,0

0,5

1,0

Phase shift

Sensitive to rotation �effrot k2�� LV effk2�2L�

2TLV

effacc k�� a2T�Sensitive to acceleration


The atomic inertial sensor

Difference of the two signals� Rotation

Sum of the two signals� Acceleration


Important paramater: flicker floor

1 0 0 1 0 1 1 0 2 1 0 3 1 0 4 1 0 5 1 0 61 0 -1 0

1 0 -9

1 0 -8

1 0 -7

1 0 -6

1 0 -5

Stab

ility

(u.a

.)

Integration time �� (a.u.)

1/��

Compromise sensitivity vs stability

Requirements for stability:

• Low coupling with outside world

• Good scaling factor

• Good rejection of the acceleration sensitivity

Raman transition

Time domain


Fullfilling the stability requirements

Space domain :

Time domain : Parameters T et Vx

L

V

V

T

Parameters L et Vx

2rot 2 TV keff ��

Rotation Acceleration

Rotation Acceleration

VL

keff2

rot 2 ��

2 T.akeffacc ��

2

2

VL

.akeffacc ��

Atomic beam source

Cold atomic source

Time domain with cold atoms :Better control of T than LBetter control of V

Better control of the scaling factorsBetter rejection of acceleration


Scheme of our inertial sensor�

��/2�/2

Gyroscope designed for compactness and good long-term stability

• One pair of large Raman beams, switchedon and off 3 times

• Parabolic trajectories in earth gravity environment

� multi-axes sensor


Inertial sensor setup

50 c

mdetection

Cs atomic sources

Raman interaction zone

Launching velocity : V = 2,6 m.s-1

� = 8°

x

z

Interrogation zone : 2L = 30 mm2T = 100 ms area = 6 mm2

Vx = 30 cm.s-1

2L

Expected sensitivity of 30 nrad.s-1.Hz-1/2

Atomic selection


• hight frequency acceleration or rotation fluctuations (aliasing effect)

• Raman laser noise

�Raman lasers wave front distortion

�AC stark shift

�Raman laser phase noise


Raman laser phase noiseAtomic interferometer , �/2 – � – �/2 sequence

Time fluctuation of the differenceof phase ��t��between the two Raman lasers

�� 1(t) – 2�2 �t+T) + �3 �t+2T)= ��t��t+T) + ��t+2T�

Raman laser phase noise appears as an acceleration not as rotation

For gyroscope: �� << 2�

For accelerometer: �� < 10-3

(Signal/noise)

��/2 ��/2�

�1 �3�22T


Scheme of the Raman lasers

Vacuum windows

ECLD�

� + 9,2 GHzECLD

MOPA

SD

1st Order

Polarization maintaining fibre

Phase lock�� = 9.2 GHz

P

Frequency reference 9.2 GHz

-+�

Spectral analyser

Switch of the Raman pulsesAOM


Raman laser beatnote

9,188 9,190 9,192

-60

-40

-20BW 10 kHz

1,5 MHz

Ram

an la

ser b

eatn

ote

(dBm

)

Frequency (GHz)


Laser phase noise• �� sequence

Transfert function for the noise power density:F�(w) = 8 sin4(�T/2) (total interrogation time 2T)

• Low frequency filter: (laser pulses of duration )Transfert function for the noise power density:G�� = 1/(1+(f/fc)2) with fc = 1/2�for square pulses (about 20 kHz)

Raman laser phase difference

Time

T

�� t��t+T) + ��t+2T�


Residual laser phase noise

Weighted by the transfert function � 1.5 mrad rmsDominated by low frequency fibre noiseDoes not limit the gyroscope sensitivity

Main contributions:• Fibre noise• 50 Hz and harmonics

Frequency (Hz)

PSD

�(r

ad2 /H

z)


Problem of acceleration noiseHigh frequency acceleration � reduction of the fringes contrast

��acc << 2� (�a<1µg rms)

0,01 0,1 1 101E-4

1E-3

0,01

0,1

1

frequency (1/T)

rela

tive

acce

lera

tion

sens

itivi

ty

1/f2 Pre-stabilise� isolation plateform

Cancel acceleration by Raman phase shift: generate laser phase noise opposite to the accelerationnoise

Increase the dynamics ofthe gyroscope

�


Test of rejection of accelerationPhase correction

1

0

Ref. 80 MHz�/2

PECLD AOM

�/2

VCO80 MHz

��+-

��

++

��

��

PZT

Modulation 90 Hz

G+- ��

P

A1

�/4

mirrors

��

Spectralanalyser

+-

A2

�/4


Test of rejection of acceleration

1 10 1001E-11

1E-9

1E-7

1E-5

1E-3

0,1

PSD

� (r

ad2 /H

z)

Frequency (Hz)

accelerometer noise floor Phase noise without rejection Phase noise with rejection

80 90 100 110 120

1E-9

1E-7

1E-5

1E-3

0,1

rejection ~ 35 dB


Prelimary results of rejection of acceleration

Experimental data

Modeling

-0.5 0.5 1 1.5 2

-40

-30

-20

-10

Frequency (Hz)10 100

Acc

eler

atio

n re

ject

ion

in d

B1

Present limitations:� accelerometer bandwidth (0.3 Hz)� need of more specific filter for the integrator


Conclusion

• Compact atom inertial sensor design for long term stability

• Suitable Raman laser phase stabilisation for groundgyroscope application

�Need to be improve for accelerometer applications�Need a better frequency reference (9.2 GHz)

• Possibility of increasing the dynamics by rejection of vibration (strapdown experiment, HYPER…)

Documents

A cold atom interferometric inertial sensor - ESAsci2.esa.int/hyper/HyperSymposium1/HYPERCD/hyper-sympLandragin.pdfA cold atom interferometric inertial sensor André ... Patrick Cheinet