Laser source for a cold-atom inertial sensor

Abstract

A laser-source assembly that is configured to illuminate a vacuum chamber containing atoms in the gaseous state so as to implement a cold-atom inertial sensor, the atoms having at least two fundamental levels that are separated by a fundamental frequency difference comprised between 1 and a few gigahertz, the assembly comprises: a master laser that emits a beam having a master frequency; a first control loop that is configured to stabilize the master frequency of the master laser on a frequency corresponding to half a set frequency of an atomic transition between a fundamental level and an excited level of the atoms; a slave laser that has a slave frequency; and a second control loop that is configured to stabilize the slave frequency of the slave laser with respect to the master frequency, the slave frequency being offset with respect to the master frequency successively, over time, by a first preset offset value, a second preset offset value, and a third preset offset value, the offset values being comprised in an interval equal to half the fundamental frequency difference plus or minus a few hundred MHz.

Claims

1. A laser-source assembly that is configured to illuminate a vacuum chamber (E) containing atoms (At) in the gaseous state so as to implement a cold-atom inertial sensor, said atoms having at least two fundamental levels that are separated by a fundamental frequency difference (f.sub.0) comprised between 1 and a few gigahertz, said assembly comprising: a master laser (Lm) that emits a beam having a master frequency (fm); a first control loop (BA1) that is configured to stabilize the master frequency of the master laser on a frequency corresponding to half a set frequency of an atomic transition between a fundamental level and an excited level of said atoms; a slave laser (Le) that has a slave frequency (fe); and a second control loop (BA2) that is configured to stabilize the slave frequency of the slave laser with respect to the master frequency (fm), said second control loop (BA2) being made up of an optical phase-locked loop (OPLL) that is configured to achieve the stabilization on the basis of an error signal () that is dependent on a phase difference between, on the one hand, a beat (fm-fe) between the master frequency and the slave frequency, and, on the other hand, a reference signal having a reference frequency defined beforehand on the basis of a preset offset value, the master frequency (fm) is set so that it corresponds to half a re-pumping frequency (f.sub.Repomp) of said atoms; the slave frequency (fe) is offset with respect to the master frequency (fm) successively, over time, by a first preset offset value (f1), a second preset offset value (f2), and a third preset offset value (f3), said offset values being comprised in an interval equal to half the fundamental frequency difference (f.sub.0) plus or minus a few hundred MHz, the first, second and third successive slave frequency values (fe1, fe2, fe3), corresponding to the master frequency (fm) offset by the first offset value (f1), second offset value (f2) and third offset value (f3), respectively, are set so that they correspond to half a cooling frequency (f.sub.Refroid), to half an optical pumping frequency (f.sub.Pomp) and to half a detection frequency (f.sub.det) of said atoms during the implementation of an inertial sensor, respectively; and both master laser and slave laser being intended to illuminate the vacuum chamber containing atoms used for a measurement with the sensor.

2. The laser-source assembly as claimed in claim 1, wherein the optical phase-locked loop comprises: a photodiode (Ph) that is configured to detect the beat (fm-fe) between the master frequency and the slave frequency; a mixer (M) that is configured to deliver, on the basis of said beat, a converted signal of lower frequency compatible with an operating range of a phase comparator (Comp); a reference oscillator (Oref) that is configured to deliver a radio reference signal having a radio reference frequency (frefr) that is defined beforehand on the basis of a preset offset value; a phase comparator (Comp) that is configured to deliver an error signal () that is dependent on the phase difference between the converted signal and the radio reference signal; and an electronic feedback device (ER) that is configured to act on the slave frequency of the slave laser so as to minimize said error signal.

3. The laser-source assembly as claimed in claim 1, wherein the first control loop (BA1) comprises a phase modulator (Mod ), an amplifier, a frequency-doubling component and a saturable absorption cell comprising the same species of atoms (At) as those used for the implementation of an inertial sensor.

4. The laser-source assembly as claimed in claim 1, wherein the slave and master lasers are distributed feedback laser diodes.

5. The laser-source assembly as claimed in claim 1, furthermore comprising a beam-forming stage comprising: a plurality of optical amplifiers (A) that are configured to amplify the beams generated by the frequency-stabilized lasers; and a plurality of frequency-doubling components (D) that are configured to double the frequency of said amplified beams.

6. The laser-source assembly as claimed in claim 1, wherein the laser sources emit a wavelength of 1560 nm and the amplifiers are semiconductor optical amplifiers.

7. The laser-source assembly as claimed in claim 1, comprising, on a silicon substrate, a first integrated photonic circuit (PIC1) comprising the second control loop (BA2) and some of the optical components required to generate the optical beams for the implementation of said sensor.

8. The laser-source assembly as claimed in claim 7, wherein the first integrated photonic circuit furthermore comprises the phase modulator (Mod ) of the first control loop (BA1).

9. The laser-source assembly as claimed in claim 1, comprising, on an InP substrate, a second integrated photonic circuit (PIC2) comprising the amplifiers and/or comprising a third integrated photonic circuit (PIC3) comprising the frequency-doubling components, which take the form of waveguides made of PPLN.

10. The laser-source assembly as claimed in claim 7, wherein the first, and where appropriate the second and where appropriate the third integrated photonic circuit are hybridized so as to be integrated into the same chip.

11. The laser-source assembly as claimed in claim 1, comprising a source integrated photonic circuit (PICs) comprising the master laser, the slave laser and associated isolators.

12. The laser-source assembly as claimed in claim 1, wherein the source integrated photonic circuit (PICs), the first integrated photonic circuit (PIC1), the second integrated photonic circuit (PIC2) and the third integrated photonic circuit (PIC3) are hybridized so as to be integrated into the same chip.

13. A cold-atom inertial sensor comprising a laser-source assembly as claimed in claim 1 and further comprising the vacuum chamber (E) containing the atoms (At) in the gaseous state.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Other features, aims and advantages of the present invention will become apparent on reading the following detailed description with reference to the appended drawings, which are given by way of nonlimiting example and in which:

(2) FIG. 1 illustrates the main atomic levels of interest of rubidium.

(3) FIG. 2 illustrates the illumination frequencies used in the three phases of implementation of a cold-atom inertial sensor.

(4) FIG. 3 schematically shows a laser-source assembly for an inertial sensor according to the prior art.

(5) FIG. 4 schematically shows a laser-source assembly according to the invention.

(6) FIG. 5 describes the various frequencies and offset values for the implementation of a cold-atom sensor based on rubidium atoms.

(7) FIG. 6 illustrates the operating mode of a second control loop using an optical phase-locked loop.

(8) FIG. 7 illustrates an example laser source for a cold-atom sensor using an integrated optics technology.

(9) FIG. 8 illustrates an example of a cold-atom inertial sensor according to the invention.

DETAILED DESCRIPTION OF THE INVENTION

(10) FIG. 4 schematically shows a laser-source assembly 10 according to the invention for implementation of a cold-atom inertial sensor. The inertial sensor uses atoms in the gaseous state At trapped in a vacuum chamber E. The atoms At have at least two fundamental levels that are separated by a fundamental frequency difference f.sub.0 comprised between 1 and a few gigahertz.

(11) Preferably, the atoms are rubidium 87 atoms, for which f.sub.0=6.834 GHz. However, various alkali atoms such as rubidium 85, cesium, sodium or potassium 40 may also be used.

(12) The assembly 10 comprises a master laser Lm that emits a beam having a master frequency fm that is stabilized by a first control loop BA1 to a frequency corresponding to half a set frequency of an atomic transition between a fundamental level and an excited level of the atoms At. For the case of rubidium, the atoms pass from a fundamental level F=1 or F=2 to an excited level when they are optically excited with a wavelength of 780 nm. The first control loop BA1 comprises a rubidium cell Cell1, in order to be able to stabilize the master laser on half an atomic frequency (a frequency doubling system is used before transmission of the portion of the laser serving in the stabilization on the atoms). This control loop BA1 uses a phase modulator and the absorption signal generated by the atomic cell to carry out a Pound-Drever-Hall stabilization.

(13) Half an atomic frequency is considered rather than the atomic frequency because the emission of the lasers is preferably substantially equal to a telecom wavelength (band C), for example for a rubidium-87 sensor the telecom wavelength being 1560 nm, which is then doubled in frequency to 780 nm before illuminating the sensor (in order to correspond to the optical transitions of the rubidium).

(14) The assembly 10 furthermore comprises a slave laser Le, having a slave frequency fe that is stabilized with respect to the master frequency fm via a second control loop BA2. The slave frequency fe is offset with respect to the master frequency fm successively over time by a first preset offset value f1, a second preset offset value f2, and a third preset offset value f3.

(15) These offset values are comprised in an interval equal to half the fundamental frequency difference (f.sub.0) plus or minus a few hundred MHz.

(16) Typically f1 f2, f13 are comprised between [f.sub.0/2400 MHz; f.sub.0/2+400 MHz].

(17) Typically the frequency of the slave laser is lower than the frequency of the master laser.

(18) These preset values must be obtained with a high precision, typically of about 1 to a few MHz.

(19) For example for rubidium:

(20) f.sub.0=6.836 GHz

(21) 2.Math.f1, 2.Math.f2, 2.Math.f3 are comprised between [f.sub.0300 MHz; f.sub.0+300 MHz] (see FIG. 5 below).

(22) Half the fundamental frequency difference f.sub.0 is considered for the same reason as explained above.

(23) Contrary to the laser-source assembly of the inertial sensor of the prior art, in which each of the frequencies f1 and f2 is stabilized on an atomic transition, using an atomic-gas cell Cell1, Cell2, the laser assembly according to the invention comprises a single atomic cell Cell1 in the first control loop BA1, the frequency of the slave laser fe being stabilized with respect to the frequency fm of the master laser, which remains fixed, and not with respect to an absolute, i.e. to a second atomic transition using a second cell. Thus, the frequency of the slave laser fe is adjusted with respect to fm to obtain the right values for the various phases of implementation of the sensor.

(24) During the implementation of the inertial sensor, the master frequency fm of the master laser is set so that it corresponds to half the repumping frequency f.sub.Repomp of the atoms:
fm=f.sub.Repomp/2

(25) Since this frequency is fixed, the successive frequency values of the slave laser are set as follows:

(26) Let fe1 be the first slave frequency value corresponding to the frequency applied to the slave laser during the cooling phase.

(27) Let fe2 be the second slave frequency value corresponding to the frequency applied to the slave laser during the optical pumping phase.

(28) Let fe3 be the third slave frequency value corresponding to the frequency applied to the slave laser during the detecting phase.

(29) These successive frequency values correspond to the master frequency fm offset by the first offset value f1, by the second offset value f2, and by the third offset value f3, respectively:
fe1=fmf1,
fe2=fmf2,
fe3=fmf3

(30) The offset value f1 is set so that fe1 corresponds to half the cooling frequency f.sub.Refroid of the atoms At:
fe1=f.sub.Refroid/2

(31) The offset value f2 is set so that fe2 corresponds to half the optical pumping frequency f.sub.Pomp of the atoms At:
fe2=f.sub.Pomp/2

(32) For example, for rubidium, the optical pumping is carried out to the atomic level |2,2>.

(33) The offset value f3 is set so that fe3 corresponds to half the detecting frequency frequency f.sub.det of the atoms At:
fe3=f.sub.det/2.

(34) For example, for rubidium, the detection is carried out on the atomic transition from |2, m.sub.F> to |3, m.sub.F>.

(35) It may clearly be seen in FIG. 2 that the various frequencies fe1 (f.sub.Refroid/2), fe2 (f.sub.Pomp/2) and fe3 (f.sub.det/2) are offset with respect to the frequency of the master laser fm (f.sub.Repomp/2) by a offset value close to f.sub.0/2 to a few hundred MHz, this value needing to be obtained with a high precision. The repumping frequency depopulates level F=1, whilst the frequencies fe1, fe2 and fe3 excite the atoms from the level F=2.

(36) It will be noted in the invention that the two, master and slave, lasers are used to generate the various beams necessary for the implementation of the sensor. It is in this case necessary for each of the beams generated by these two lasers to be phase locked (see below) in order to produce, for example, Raman transitions that serve to interact with the atoms during the interrogation of the sensor.

(37) On account of the values of the atomic transitions of the atoms usable for the implementation of a cold-atom inertial sensor, use is preferably made of:
fei=fmfi. where i=1,2,3
and where fi is comprised in the interval [f.sub.0/2800 MHz; f.sub.0/2+800 MHz].

(38) It will be noted that the cooling, which consists in forming a three-dimensional magneto-optical trap in the chamber, typically requires 3 beams, which are commonly denoted 3-D MOT X1, X2 and H, and which may be obtained from a single beam (see below).

(39) It will also be noted that the objective of the steps of pumping and detecting is to get the atoms into the right state before the interferometric sequence, in the implementation of an inertial sensor. Since the pump and detection beams are not used at the same time in the measurement cycle of the sensor, it is possible to use the same output to illuminate the chamber.

(40) Preferably, the master laser Lm and the slave laser Le of the laser-source assembly 10 are grating laser diodes, or a strip of laser diodes, of distributed-feed-back (DFB) or distributed-Bragg-reflector (DBR) type emitting a telecom wavelength, advantageously 1560 nm.

(41) The slave laser Le is stabilized to the master laser by detecting a beat between the two lasers using a photodiode, then by carrying out a stabilization based on the principle of an optical phase-locked loop (OPLL) such as described below.

(42) The phase locking of the master and slave beams makes it possible to obtain both a coherence between the two beams and predefined values of the frequency differences between master laser and slave laser.

(43) According to one preferred embodiment, the laser-source assembly 10 according to the invention also comprises a beam-forming stage 20, which is also illustrated in FIG. 4 and which comprises: A plurality of optical amplifiers A campaign that are configured to amplify the beams generated by the frequency-stabilized lasers. Advantageously, the amplifiers are semiconductor optical amplifiers (SOA). A plurality of frequency-doubling components D that are configured to double the frequency of the amplified beams. In the preceding example of an emission at 1560 nm, the doubled beam has a wavelength of 780 nm compatible with the excitation of the aforementioned atoms for the implementation of a cold-atom inertial sensor.

(44) The laser-source assembly also comprises various components for the manipulation of the beams generated by the master and slave lasers: these beams generated by the lasers are split, attenuated and recombined to form beams that are intended to be amplified then doubled so as to generate beams that are configured to illuminate the chamber of a cold-atom sensor so as to implement a cooling phase 1, a pumping phase 2 and a detecting phase 3 during a measurement carried out during the implementation of the sensor.

(45) In the cooling phase 1, the beams at the top of FIG. 4 are cut, and the cell is illuminated with the beams at the bottom of FIG. 4. In the pumping phase 2, the beams at the bottom of FIG. 4 are cut, and the cell is illuminated with the beams at the top of FIG. 4.

(46) In the detecting phase 3, the beams at the bottom of FIG. 4 are cut, as is the beam at the top of the figure, which originates from the master laser, which nevertheless remains turned on in order to carry out the frequency stabilization of fe3 with respect to fm.

(47) A nonlimiting example of offset values Oft f2, f3 to be achieved with the loop BA2 is illustrated in FIG. 5 for rubidium. The master frequency fm is given in THz and corresponds to an optical wavelength of about 780 nm, as described above, with a f.sub.0 between the fundamental state F=1 and the fundamental state F=2 of 6834 GHz. The frequency of the master laser is regulated to half the what is called repumping frequency f.sub.Repomp such as illustrated in FIG. 2.

(48) In the cooling phase, the frequency fe1 corresponds to the frequency of the master laser fm decreased by f1 with:
2.Math.f1 varies between f.sub.0267 MHz and f.sub.0167 MHz

(49) In the pumping phase, the frequency fe2 corresponds to the frequency of the master laser fm decreased by f2 with:
2.Math.f2=f.sub.02 MHz

(50) In the detecting phase, the frequency fe3 corresponds to the frequency of the master laser fm decreased by f3 with:
2.Math.f3=f.sub.0267 MHz

(51) These offset values are determined on the basis of the known values of the atomic energy levels (see FIGS. 1 and 2) and of the physics of the cold-atom sensors. The desired precision in the slave-frequency value, and therefore in the offset values, is preferably about 1 to a few MHz.

(52) Preferably, the first control loop BA1 comprises a phase modulator Mod , an amplifier, a frequency-doubling component and a saturable absorption cell Cell1 containing atoms At of the same material as that used in the chamber of the sensor, such as described above.

(53) According to one preferred embodiment, the second control loop BA2 is based on the principle of phase-locked loops PLLs, and is able to act on the relative phase between two optical beams by transposition to an electrical signal. This method is referred to as an optical phase-locked loop or OPLL.

(54) The principle of such a stabilization is illustrated in FIG. 6.

(55) It will be recalled that the frequency of a signal f is proportional to the derivative of the phase of the signal with respect to time. Making the phase difference between the slave signal to be stabilized and a reference signal, zero for example, allows a slave frequency stabilized (made equal) to the reference frequency to be obtained.

(56) In other words, the second control loop BA2 is formed on the basis of an optical phase-locked loop OPLL that is configured to carry out the stabilization on the basis of an error signal c that is dependent on a phase difference between, on the one hand, the beat fm-fe between the master frequency and the slave frequency, and, on the other hand, a reference signal having a reference frequency fref defined beforehand on the basis of a preset offset value f1, f2 or f3.

(57) According to one preferred embodiment, the phase-locked loop OPLL comprises a photodiode Ph that detects the optical beams generated by the slave and master laser, and more particularly the beat signal between these frequencies, frequency fm-fe. The frequency of the slave laser fe is stabilized on the frequency of the master laser fm (itself directly stabilized on an atomic reference), on the basis of this beat fm-fe, which is adjusted to a value close to f0/2 (3.418 GHz for rubidium), and which must be modifiable by a few hundred MHz, for example between 150 MHz and +50 MHz, repeatedly and in a time of about 1 millisecond.

(58) Phase comparison between signals of the order of a gigahertz being very complex to implement, a conversion of the beat frequency fm-fe to a converted signal of lower frequency fm-fefd is carried out using a mixer M, this operation being referred to as a down conversion. The aim is to make the frequency of the signal to be stabilized compatible with the operating range of the phase comparator Comp. The obtained signal of frequency fm-fefd and of phase =2(fm-fefd)t+1 typically has a frequency comprised between 1 and 500 MHz. Preferably, a filter is added after the mixer in order to select only the signal of interest of frequency fm-fefd in the desired range.

(59) The OPLL loop also comprises a reference oscillator Oref that is configured to generate a radio reference signal having a radio reference frequency frefr and a reference phase ref=2.Math.frefr.Math.t+2, the frequency frefr being defined beforehand on the basis of a desired preset offset value either f1 or f2 or f3. The frequency of the radio reference signal frefr is typically in the radio frequency domain, comprised between 1 and 500 MHz.

(60) Next, the phase comparator Comp generates an error signal c that is dependent on the phase difference ref between the converted signal and the radio reference signal.

(61) Finally, a feedback electronic device ER generates a correction signal and acts on the slave frequency fe of the slave laser so as to minimize the error signal . In the case of a DFB laser diode typically the supply current of the laser is acted on, the optical frequency being a function of this current.

(62) Typically, on turn-on, after a certain time, the frequency of the converted beat signal locks on the chosen frequency frefr of the oscillator.

(63) The advantage of using an OPLL loop is that it is compatible with the orders of magnitudes of the offsets to be applied and with an integrated photonic implementation, this allowing the bulk and overall cost of a cold-atom inertial sensor to be drastically decreased.

(64) The invention also relates to a cold-atom inertial sensor 100, also illustrated in FIG. 4, comprising a laser-source assembly 10 and furthermore comprising a vacuum chamber E containing trapped atoms At in the gaseous state, said chamber being illuminated by the beams generated by the complete laser-source assembly.

(65) In the schematic of FIG. 4, a splitter S, for example a cube corner is used, on the already amplified beam, to split the beam into three for the formation of the magneto-optical trap. Alternatively, the splitting is carried out upstream of the amplifiers.

(66) An example of a cold-atom-sensor laser source, and the sensor 100 according to the invention, are illustrated in FIG. 7. The assembly of FIG. 7, with the exception of the chamber E and the saturable absorption cell Cell1, may be integrated to various degrees using integrated photonic components.

(67) According to one preferred variant, the laser-source assembly 10 comprises, on a silicon substrate, a first integrated photonic circuit PIC1, comprising at least the second control loop BA2 and some of the optical components necessary to generate the various optical beams for the implementation of the sensor. Optionally, the circuit PIC1 also comprises some of the loop BA1, except the saturated absorption cell Cell1.

(68) Preferably, the laser-source assembly 10 comprises, on an InP substrate, a second integrated photonic circuit PIC 2 comprising the optical amplifiers.

(69) Advantageously, the laser-source assembly 10 comprises a third integrated photonic circuit PIC3 comprising the doubling components, which take the form of waveguides made of PPLN.

(70) Preferably, the first integrated photonic circuit PIC1 comprises the phase modulator of the first locking loop BA1 and/or the second integrated photonic component PIC2 comprises the amplifier of the first locking loop and/or the third integrated photonic circuit PIC3 comprises the doubling component of the first locking loop.

(71) In one embodiment, the first, second and/or third integrated photonic circuits are hybridized with the adjacent integrated photonic circuits so as to be integrated into the same chip. In a completely integrated version, a hybridized matrix-array of PPLN doublers is used in the photonic circuit.

(72) These various integrated photonic circuits replace all or some of the complex optical system of the prior art required to implement a cold-atom sensor.

(73) According to one embodiment, the laser-source assembly 10 according to the invention comprises a source integrated photonic circuit PICs comprising the lasers Le and Lm and associated isolators ISOe and ISOm. According to another embodiment, the lasers and their isolator are connected to the circuit PIC1 with optical fibers.

(74) According to one all-integrated embodiment, the source integrated photonic circuit PICs and the first, second and third integrated photonic circuits PIC1, PIC2 and PIC3 are hybridized so as to be integrated into the same chip PIC.

(75) According to one variant, a strip of free-space or fiber-based isolators is used to protect the diodes in the case where the laser diodes are remote from the integrated circuit. In the case where the laser sources are hybridized in the circuit, according to one variant isolators that are directly integrated into the photonic circuit PIC are used, these isolators for example employing layers of Faraday materials or even creating nonreciprocal Brillouin scattering in order to make a guide transparent in one direction only.

(76) The components listed below are compatible with an integration into a silicon chip, such as schematically shown in FIG. 7, comprising integrated photonic components for creating and manipulating the various beams required to manipulate the cold atoms and to stabilize the various laser diodes:

(77) 1/99 (1 to 2) and 33/33/33 (1 to 3) splitters, static 2 to 1 and 2 to 3 (or 2 to 6) couplers and crossover between guides produced in CMOS technology;

(78) Attenuator adjustable from 0 to 10 dB without constraint on the passband.

(79) These variable attenuators may be produced using a Mach-Zehnder approach, or variable splitters an output of which serves to divert the undesired power. For the manipulation of cold atoms, the extinction ratio of the lasers must be relatively high and to this end a plurality of attenuators may be used one after the other. On/Off: for example 2 to 2 couplers the assignation of the outputs of which may be changed in a time of the order of a microsecond. With an extinction ratio of 20 dB. A photodiode PhD at 1560 nm for the stabilization of the lasers with a passband of 3.5 GHz. Fiber-based input-output, for example with grating coupling. Phase modulator Mod (achieved for example by carrier injection). Phase-locked loop in integrated photonic technology (for example in CMOS electronics technology). Isolator using non-reciprocal effects obtained, for example, with thin layers of Faraday-effect materials or even by means of Brillouin scattering induced transparency (BSIT).

(80) FIG. 8 illustrates an example of a cold-atom inertial sensor according to the invention comprising an all-integrated laser-source assembly integrated into a single photonic circuit PIC. The integrated photonic circuit PIC contains its own master and slave lasers, which are formed (amplified and doubled) then coupled to optical fibers FO.sub.1 and FO.sub.23 at the output of the photonic circuit. These fibers are placed nearest the vacuum chamber E and the beams output from these fibers serve to cool, pump and detect the atoms. In the schematic, a magneto-optical trap that comprises a cube corner CC and that requires only a single initial beam for the cooling is used by way of nonlimiting example.