EP2498150A1 - Atomic clock - Google Patents

Abstract

The method involves sending a linearly polarized laser beam (11) from a laser source e.g. laser diode (1) such as vertical-cavity surface-emitting laser (VCSEL) diode, across a cell (3). A beat frequency obtained between the beam output by the source and transmitted through the cell and a beam induced by Raman Effect within atoms of the cell is detected. Optimal laser injection current is found in open-loop mode of the atomic clock. The clock is operated in a closed-loop mode by returning a microwave signal received as output of the cell to control the injection current for priming the clock. An independent claim is also included for an atomic clock.

Description

Introduction

The present invention relates to a method of transmitting a time reference signal by an atomic clock as well as an atomic clock as such using such a method.

State of the art

Miniature atomic clocks (volume of one cm ³ or less), with low power consumption (below Watt) and which allow portable applications are devices made possible by the combination of the physical principles CPT (coherent entrapment of population) or Raman with an atomic clock architecture based on a gas absorption cell. These two physical principles do not require a microwave cavity to interrogate the reference atoms (typically Rubidium or Cesium) and thus eliminate the volume constraint associated with conventional cell-type atomic clocks. The physical part of the clock, which consists of the light source, the optical elements, the gas cell, the photodetector and all functions such as heating and magnetic field generation, will be the subject of considerations that follow. The implementation of technologies such as vertical cavity surface-emitting laser (VCSEL), microfabrication techniques for gas cells and vacuum encapsulation have made it possible to Massively reduce the volume and power consumption of these atomic clocks. VCSEL lasers offer the possibility of combining the optical pumping function and the microwave interrogation of the reference atoms. This type of laser offers the following advantages: modulation of the injection current possible up to several gigahertz, low power consumption, wavelength compatible with standard reference atoms (Rubidium or Cesium), excellent service life, high temperature operation, low cost and ideally adapted optical power. Silicon microstructuring technologies coupled to the bonding / welding processes of a glass substrate (typically pyrex or quartz) on a silicon substrate make it possible to produce gas cells that are much smaller in size than is possible. achieve with the traditional technique of blowing and forming of glass tube. The reduction of the dimensions of the gas cell is also accompanied by a reduction in the consumption required to heat the gas cell.

Different arrangements of the physical part of such a clock have been made. The majority of the arrangements are based on a single pass of the laser beam through the cell (see S. Knappe, MEMS atomic docks, Book chapter in Comprehensive Microsystems, vol. 3, p. 571 (2008), Ed. Elsevier ), others take advantage of gas cells with mirrors inside the cell or allowing a double pass of the laser beam through the cell (see documents US7064835 and EP0550240 ). Arrangements with double passage of light through the cell have the advantage of doubling the effective optical length of the cell and thus improving the performance of the atomic clock (in terms of power consumption and / or stability of the cell. frequency). Nevertheless, these double-pass arrangements have not been implemented for reasons of instability of the device and in particular because of laser disturbances caused by the light reflected back by the mirrors on the laser.

The documents US7064835 (Symmetricom) US5340986 (Wong ) and US2009 / 128820 (Seiko, Fig. 6 ) describe the use of a separator element to direct the reflected beam to the photodetector. The light emitted by the laser is polarized linearly, converted into circular polarization by a quarter-wave plate before passing through the cell, reflection on the mirror, second passage in the cell, and detection on a photodetector.

The configurations described above have disadvantages for producing a CPT oscillator. Indeed, a detector can be placed before the passage of light in the cell and another after the double passage in the cell, but no photodetector can be positioned after a single passage of light in the cell. This additional detector makes it possible to obtain a signal additional to that of the detector placed after the double pass. This additional signal is useful for measuring and controlling clock parameters such as the temperature of the cell or the frequency of the laser source, for example. In addition, the configurations described above are not applicable in a configuration of a Raman oscillator because the servocontrolling of the frequency of the laser source is performed by the same detector ensuring the detection of the return laser beam of the cell. .

However, the existing atomic clocks remain insufficient and the invention seeks a solution for an atomic clock that represents a better compromise between accuracy and consumption.

Brief description of the invention

• une première phase de recherche du courant d'injection optimal du laser en boucle ouverte de l'horloge atomique,
• une seconde phase d'allumage de l'horloge comprenant la mise en boucle fermée de l'horloge par le retour du signal microonde reçu en sortie de la cellule pour le contrôle du courant d'injection du laser.

For this purpose, the invention is based on a solution for transmitting a time reference signal by an atomic clock, notably comprising a method for igniting the atomic clock comprising the following steps:

• a first phase of research of the optimal injection current of the open-loop laser of the atomic clock,
• a second phase of ignition of the clock comprising the closed loop of the clock by the return of the microwave signal received at the output of the cell for the control of the injection current of the laser.

The invention is more precisely defined by the claims.

Brief description of the figures

• La figure 1 représente un schéma de principe d'une horloge atomique selon un mode de réalisation de l'invention.
• La figure 2 représente un schéma fonctionnel de l'horloge atomique selon un mode de réalisation de l'invention.
• La figure 3 représente un schéma électrique équivalent d'un système de détection optoélectronique selon un mode de réalisation de la présente invention.
• La figure 4 représente un schéma électrique équivalent d'un détecteur optoélectronique selon un autre mode de réalisation de la présente invention.
• La figure 5 représente schématiquement les courbes de gain g en fonction de la fréquence ω, les deux axes étant logarithmiques, pour un amplificateur à transimpédance classique (trait plein), un amplificateur à transimpédance doté d'un élément permettant d'augmenter la bande passante (« inductor peaking » ou « high frequency gain boosting », traitillé), et un système de détection selon l'invention (pointillé).
• La figure 6 représente le spectre d'absorption d'un gaz en fonction du balayage en courant d'injection laser avec l'horloge atomique en boucle ouverte.
• La figure 7 représente un premier mode de réalisation d'une horloge atomique à double passage.
• La figure 8 représente un second mode de réalisation d'une horloge atomique à double passage.
• La figure 9 représente un troisième mode de réalisation d'une horloge atomique à double passage.
• La figure 10 représente une vue schématique éclatée d'une horloge atomique basée sur le second mode de réalisation à double passage et une géométrie à angle droit.
• La figure 11 représente une vue schématique éclatée d'une horloge atomique basée sur le second mode de réalisation à double passage et une géométrie droite.
• La figure 12 représente une vue schématique d'une horloge atomique basée sur le premier mode de réalisation à double passage.
• La figure 13 représente une vue schématique d'une horloge atomique basée sur le premier mode de réalisation à double passage avec géométrie à angle droit.
• La figure 14 représente une vue schématique d'une horloge atomique basée sur le troisième mode de réalisation à double passage.

These objects, features and advantages of the present invention will be set forth in detail in the following description of particular embodiments given as a non-limiting example in relation to the appended figures among which:

• The figure 1 represents a block diagram of an atomic clock according to one embodiment of the invention.
• The figure 2 represents a block diagram of the atomic clock according to one embodiment of the invention.
• The figure 3 represents an equivalent circuit diagram of an optoelectronic detection system according to an embodiment of the present invention.
• The figure 4 represents an equivalent circuit diagram of an optoelectronic detector according to another embodiment of the present invention.
• The figure 5 schematically represents the gain curves g as a function of the frequency ω, the two axes being logarithmic, for a conventional transimpedance amplifier (solid line), a transimpedance amplifier provided with an element making it possible to increase the bandwidth ("inductor peaking "or" high frequency gain boosting ", etched), and a detection system according to the invention (dotted).
• The figure 6 represents the absorption spectrum of a gas as a function of laser injection current scanning with the open-loop atomic clock.
• The figure 7 represents a first embodiment of a double-pass atomic clock.
• The figure 8 represents a second embodiment of a double-pass atomic clock.
• The figure 9 represents a third embodiment of a double-pass atomic clock.
• The figure 10 is an exploded schematic view of an atomic clock based on the second double-pass embodiment and a right-angle geometry.
• The figure 11 represents an exploded schematic view of an atomic clock based on the second double-pass embodiment and a right geometry.
• The figure 12 is a schematic view of an atomic clock based on the first double pass embodiment.
• The figure 13 is a schematic view of an atomic clock based on the first dual pass embodiment with right angle geometry.
• The figure 14 is a schematic view of an atomic clock based on the third double pass embodiment.

The chosen solution is based on the use of an atomic oscillator based on the Raman effect, which is based on the irradiation of reference atoms at an optical resonant frequency which induces the emission of photons with an optical frequency shifted by the hyperfine frequency of the reference atom. The combination of the two resulting signals provides a detectable beat, the signal frequency of which serves as a time base for the clock.

The figure 1 schematically illustrates the optical part of a Raman atomic clock according to one embodiment of the invention. It comprises a laser diode 1, which can be low-consumption and of the VCSEL type, which emits a linearly polarized beam 11, a quarter-wave plate 2 which polarises the light coming from the laser in a circular polarization incident beam 12. beam 12 passes through a cell 3 comprising selected atoms, such as cesium or rubidium with a buffer gas, optionally placed in a magnetic field B. At the output of this cell 3, the incident signal 12 is combined with the second signal 13 generated by the Raman effect, as explained above. The combination of the two signals is detected by a photodetector 4 which allows the recovery of the signal comprising the atomic time base, originating from the cesium or rubidium atoms. This output signal 14 is analyzed by an electronic signal processing device of the microwave frequency divider type 5 to generate the signal frequency necessary for the time base. Exit 15 finally represents this time base, exploited by a clock as will be explained later. An optional radio frequency amplifier 6 is positioned at the output of the photodetector 4.

As a remark, optionally but advantageously, a part of the output signal 14 is used to modulate the injection current of the laser, by a microwave injection at the level of the laser 1, represented by the arrow 7. This makes it possible to achieve a signal-on-noise output 14 of better quality and easier to operate. This principle is equivalent to an amplitude modulation of the laser.

As a remark, the cell 3 has been positioned within a magnetic field B, which makes it possible to remove the degeneracy of the Zeeman sub-states of the atoms. Alternatively, it could be in a zero magnetic field, to obtain a superposition of energy levels, and a high signal, and a simplified clock.

The figure 2 functionally represents a Raman atomic clock according to one embodiment of the invention. It comprises a power supply device and DC / DC converter 21, a processing center 23 which can be a processor or a low-power electronics, the main functions of which comprise all or some of the following functions: setting the operating frequency of the laser 1 and its injection current, control of the temperature of the cell 3 and the laser 1, management of the intermittent mode of the laser, correction of the frequency of the atomic clock as a function of the temperature, setting a clock additional low precision as quartz based. The implementation of these functions will be detailed later. The clock then comprises a DC 24 power source for the laser 1, a DC current source 25 for heating the laser 1, a current source 26 of the solenoid for the generation of the magnetic field B 36, a current source 27 for heating the cell 3, which cooperates with an associated heater 37, to which a temperature sensor can be additionally added.

These various components allow the operation of the laser 1 which acts on the optical device 10 of the clock, a simplified representation has been presented with reference to the figure 1 . In this embodiment, the assembly formed by the optional magnetic field generator B 36, the heater 37 and the cell 3 is positioned in an enclosure for reaching their magnetic shielding. Alternatively, only part of these components can be integrated within this shield. In another variant, this magnetic field may be zero and the clock simplified, as explained above. At the output, a fast photodetector 4 comprises a DC output for returning a signal proportional to the light intensity received to the processing center 23. It furthermore comprises an RF output for a signal which is first amplified by a signal chain. amplification 32 then a delay line and phase shifter 33 to be fed back to a diplexer 34 (bias tee) which combines the RF signal with the DC laser injection current from the current source 24. Part of the amplified RF signal is processed by a frequency divider 5 before returning to the processing center 23. At the output of this processing center, a user frequency signal 22 is obtained (for example 32 kHz, or 1 pulse per second, etc. .). Finally, the implementation of this clock is preferably made from low power components.

As a reminder, the CPT type atomic clocks all use a complex architecture and include a device for correcting the local clock, called by its English name "Voltage Controlled Oscillator" (VCO), as well as an electronic control system. the clock, representing a high total power consumption. The Raman atomic clock described previously the advantage of greater simplicity for a greatly reduced consumption.

In such a Raman oscillator, a laser beam incident at a first frequency interacts with an atom vapor, thereby stimulating, by light-atom interaction, the emission of a second Raman beam having a second frequency. As mentioned, the beat between the first frequency and the second frequency produces a third frequency: the beat frequency, which is used as a time base. In the case where the vapor comprises, for example, Rubidium-85 and the laser is of the vertical cavity and surface-emitting semiconductor type emitting a beam of light at a wavelength of around 780 nm or 794nm, the beat frequency is of the order of 3GHz with a bandwidth around a hundred kHz. This beat frequency is generally very low and has a very low spectral content. The detection of these beat frequencies at the output of the clock is a delicate technical problem, in particular to limit consumption.

To answer this technical problem, it is proposed a detection system of a signal (i _PD ) high frequency (ω _C ) narrow band, said system having a low power consumption. The system includes a generator for providing the signal (i _PD ) as a current, and a parallel resonance circuit for varying the impedance of the generator output as a function of the frequency of the generated signal and for converting the current in tension. The system further includes an amplification stage to further increase the gain by minimally degrading the system noise to allow the detection of a signal of very low amplitude. The generator is the photodetector 4 mentioned above, stimulated by electromagnetic radiation.

According to an embodiment of the detection system, represented on the figure 3 a simple inductor L1 is included in the realization of the parallel resonance circuit and the photodetector is of the PD photodiode type. The photodiode PD is polarized through the inductor L1 connected to a voltage source. This makes it possible to maintain the photodiode PD at a desired voltage by supplying the current necessary for the photodiode PD to work properly. It should be noted that the signal to be detected has a spectral content centered around a predetermined frequency ω _C which is of the order of a few gigahertz and very narrow (of the order of 10 ^-4 × ω _C ).

The signal to be detected i _PD appears in the form of a current on a node N) which connects the inductance L1 to the photodiode PD. This node N is electrically coupled to the input of the MAMP amplifier and the amplified signal appears at the output of the MAMP amplifier. N node thus configured has a parasitic capacitance C _IN . This parasitic capacitance C _IN forms with the inductor L1 the parallel resonance circuit. The value of the inductance is determined so that its inductive reactance at the frequency of the signal to be detected is equal to the capacitive reactance of the parasitic capacitance C _IN . In other words ω _C × L1 = 1 / (ω _C × C _IN ). Under these conditions, there is a bandpass filter with Q quality factor and a half-height width of 1 / Q. With an inductor L1 integrated in the circuit, a quality factor Q of about 10 is reached, whereas a quality factor Q of about 50 with an inductance L1 external to the circuit is obtained. The equivalent parallel resistance Rp has a value of ω × L × Q. Thanks to a high quality factor Q, you can achieve a significant gain without the consumption that would normally be associated with it. Without the present invention, a broadband transimpedance amplifier with 10 GHz bandwidth would be used in place of the proposed one. Typically, this kind of amplifier consumes around a watt, while the amplifier proposed above consumes less than two milliwatts.

The figure 5 shows the difference in gain versus frequency for both types of amplifier. A broadband transimpedance amplifier of the state of the art can cover a large frequency range, but leads to high consumption and noise comparatively high, since the noise is all the more important that the bandwidth is wide. Unlike the broadband transimpedance amplifier, the proposed solution selects with a resonant element a signal centered around a central frequency that is significantly lower than the cutoff frequency typical of the photodetector technology used. The gain characteristic shows a very narrow bandwidth, compatible with the narrow spectral content of the signal (of the order of 10 ^-4 × ω _C ), which greatly reduces the noise compared to a transimpedance amplifier. Consumption is very low because the system has no active elements.

Since node N has a very high impedance, it is sufficient to use a simple low noise common source MOS type amplifier to further increase the gain by minimally degrading the system noise to allow the detection of a signal. very low amplitude. In one embodiment, the amplifier has a resistive load on the output. In another embodiment, taking advantage of the fact that the signal to be detected has a very low spectral content, which may be of a single unmodulated frequency, the load at the output of the amplifier is provided by a second inductor L2. whose value is chosen to maximize the gain for a signal at the predetermined frequency ω _C.

The input of the amplifier can be coupled in AC mode with the node N, ie, with a DC coupling capacitance, and the input of the amplifier can therefore be biased by a voltage source Vb through a resistor Rb so that the input of the amplifier is at an optimum voltage. In the manufacture of a circuit according to the present invention, it is possible that the value of the parasitic capacitance C _IN or the value of the inductance L1 varies from one batch to another or from one room to another. This would have the effect of moving the resonance frequency of the resonance circuit out of the appropriate frequency band to detect a signal at the predetermined frequency. For this reason, it is proposed to act on the value of the capacity of the node N by adjusting this capacity. This can be done in various ways, for example by using an adjustable capacitor (trim capacitor) or by the use of several capacities that can be connected or disconnected at the node N, for example by the targeted deposition of metal during manufacture. This can also be accomplished by a laser-trimming system where the node (N) is connected to a capacity whose value is adjusted by laser ablation at the time of testing the assembly.

According to another embodiment of the present invention, the resonance circuit comprises an electromechanical resonator of the bulk wave resonator type or Bulk Acoustic Wave (BAW), as illustrated by FIG. figure 4 . The volume wave resonator (BAW) allows an even more selective filtering and it has, at the anti-resonance, a high real impedance while allowing to neutralize the parasitic capacitance C _IN of the node N. According to one embodiment, the electromechanical resonator makes it possible to achieve a quality factor greater than 300. In this embodiment, the photodiode is polarized by means of an adaptive circuit whose output stage is a source of current CCS controlled so as to guarantee a fixed polarization voltage on the diode at low frequency.

Another technical problem encountered for the implementation of the Raman clock is to achieve sufficient stability, allowing it to operate accurately over a satisfactory period. This problem is solved by the operation described above in relation to the figure 1 and functionally represented by the figure 2 .

A feedback of the RF signal detected on the optical frequency of the laser, in order to control the emission frequency of the laser, is always recommended in the state of the art to obtain a stable and high-precision atomic oscillator, in particular for the atomic clocks of type CPT. In the present case, it has been found that it is almost impossible to control the operation of the Raman oscillator in a repeatable and reliable closed-loop manner with respect to the optical frequency of the laser. Synchronous detection for frequency stabilization of a laser is not adequate in the case of a closed-loop Raman oscillator.

Surprisingly, it was possible to operate the Raman oscillator without servocontrolling the optical frequency of the laser, that is to say with zero frequency control, or in other words without active control of the optical frequency of the laser. either an open loop operation with respect to the frequency of the laser.

Stability tests were performed according to the previous principle, which demonstrated a high frequency stability of the atomic clock. At a temperature of 87.5 ° C, the Raman oscillator varies by one second every 160 years and runs steadily for several days at least continuously.

The temperature of the cell, with an active length of 5 mm, was also lowered to below the melting temperature of Rubidium (39.3 ° C). A temperature drop of 90 ° C to 35 ° C corresponds to a decrease in the saturation vapor pressure by two orders of magnitude (~ 10 ^-4 torr at 10 ^-6 torr). The stability depends on the temperature of the cell but remains acceptable up to a temperature of 35 ° C. Indeed, at a temperature of 40 ° C, the Raman oscillator always works to satisfaction with a stability of one second every 16 years, which is remarkable. At 35 ° C, the Raman signal was still present and sufficiently stable. This Unexpected finding allows to consider an atomic oscillator without heating the cell, operating for example only when the temperature around the cell is sufficient, for example around 35 ° C, preferably around 40 ° C. It is also possible to reduce the operating temperature by using Cs instead of Rb in the cell, the cesium melting temperature being even lower than that of Rubidium (28.5 ° C instead of 39.3 ° C).

An additional technical problem is encountered when starting the clock. Indeed, the solution explained above shows how to obtain stable and efficient operation of the clock when it is in cruise mode, from the devices described in connection with the Figures 1 and 2 . A fully open loop operation, that is to say without the return 7 of the figure 1 , would be an alternative embodiment feasible less efficient because the signal obtained would be relatively weak and spectrally less pure.

For this, it has been found that there is a reduced range of laser injection current, ie a corresponding frequency range, close to the optical absorption peak of the cell gas, which allows when starting a laser irradiation on the open-loop cell and then going into closed mode as described above to obtain a resonance of the clock to achieve the optimum operating regime described above. Thus, by the judicious choice of the injection current of the laser during the ignition of the laser then the closed circuit with respect to the injection current of the laser as explained above, the clock reaches of course its optimal operating regime. This phenomenon allows a self-ignition of the clock, and makes it possible to use it intermittently.

This operating range is more accurately illustrated by the figure 6 for the case of natural Rubidium. This figure shows the optical absorption curve 50 of the Rubidium, by the signal obtained on the photodiode 6, as a function of the injection current of the laser. The favorable current range is located in zone 52, which represents a portion of the peak of greater absorption 51, at a distance from the two maximum values Vmax and Vmin of this peak. By choosing a reduced range [V1; V2], sufficiently far from these values, we deduce a favorable current range [i1, i2]. The preceding considerations allow the implementation of the following ignition method of a Raman oscillator, which is part of the method of transmitting a time signal per atomic clock according to the invention.

• mise en boucle ouverte de l'horloge à effet Raman ;
• balayage de la fréquence du laser et identification du point Vmax d'absorption maximale et du courant d'injection Imax correspondant ainsi que du point Vmin d'absorption minimale du pic 51 associé et du courant d'injection Imin correspondant ;
• détermination d'un courant d'injection ILD entre i1 et i2 en ajoutant une certaine valeur seuil delta à Imin ou en la retranchant à Imax. Par exemple, une valeur proche de i1 peut être choisie.

A first phase consists of a search for the optimal injection current i of the laser, ie the range i1 to i2. This first phase includes the following steps:

• open looping of the Raman clock;
• scanning the laser frequency and identifying the maximum absorption point Vmax and the corresponding Imax injection current and the minimum absorption point Vmin of the associated peak 51 and the corresponding Imin injection current;
• determining an ILD injection current between i1 and i2 by adding a certain delta threshold value to Imin or subtracting it from Imax. For example, a value close to i1 can be chosen.

For example, for the Rubidium and the VCSEL laser used for the experiments, the laser injection current must be chosen between 2.25760 mA and 2.25824 mA. V1 is 15% Vmax-Vmin above Vmin and V2 at 67% Vmax-Vmin above Vmin.

This first phase of the ignition process can be performed before each ignition of the clock, in order to obtain the greatest accuracy possible, which allows to modify the previous values in time according to a possible drift of the device or conditions of measurement. Alternatively, this phase is performed only once to calibrate the device and the data are stored to be repeated at each ignition.

• mise en boucle fermée de l'horloge, en ajoutant la rétroaction 7 explicitée précédemment ;
• ajustement du courant d'injection du laser à la valeur ILD identifiée par la première phase ;
• vérification de l'obtention du phénomène de résonance de l'horloge en sortie ;
• en cas de non résonance, légère modification de la valeur du courant d'injection ILD dans la plage [i1 ; i2] selon un pas prédéfini, et répétition de cette étape jusqu'à obtenir le phénomène de résonance.

The ignition process further implements the following steps of concrete ignition of the laser and the clock;

• closed looping of the clock, adding feedback 7 explained previously;
• adjusting the laser injection current to the ILD value identified by the first phase;
• verifying the obtaining of the resonance phenomenon of the output clock;
• in case of non-resonance, slight modification of the value of the injection current ILD in the range [i1; i2] in a predefined step, and repetition of this step until the resonance phenomenon.

According to an advantageous variant, this method comprises a preliminary step of measuring the optical power of the laser, since the frequency of the clock may depend on the optical power interacting with the atoms. This can be done by measuring the optical power by means of a photodiode of the device and comparing the photovoltage thus generated with a reference stable voltage source. Adjusting the laser injection current and the laser temperature then makes it possible to obtain the nominal optical power and the optical frequency of the clock.

According to another advantageous embodiment, this method comprises a preliminary step of heating the gas cell and the laser, since the operation of the clock depends on the temperature, as has been previously mentioned. There is a correlation between the frequency of the closed-loop Raman clock and the temperature of the cell. This property makes it possible to control the frequency during the phases of running and stopping of the clock by the sole measurement of the temperature.

Thus, according to the embodiment chosen, the Raman effect clock comprises a temperature control. For this purpose, it comprises a temperature sensor, which may be a photodiode, and a heating device for increasing the temperature if it is under a set temperature.

The previously described steps of the ignition process are automatically managed by the clock, on the basis of the hardware and software means of the aforementioned processing center 23, in particular under the control of the microprocessor.

According to a first atomic clock realization, the Raman oscillator is used intermittently, in addition to a conventional oscillator of the state of the art, for example quartz. In this embodiment, the atomic oscillator transmits a time base which allows the calibration of the quartz oscillator, its correction, and allows to greatly increase its accuracy over time. This intermittent operation of the atomic oscillator has the advantage of controlled consumption. The ignition period of the atomic oscillator is chosen according to a compromise between the consumption and the accuracy of the clock: the more this oscillator is used, the more accurate the clock but the higher the consumption will be. When the additional oscillator of less precision is corrected by the atomic oscillator, the latter is off.

According to a second atomic clock realization, the Raman oscillator is used alone, as a single time base, and therefore according to a permanent operation. In this embodiment, the highest accuracy is obtained, but through greater energy consumption.

The atomic clock described above is furthermore produced in a compact and space-saving structure. The Figures 7 to 14 thus describe several embodiments of the optical part of the atomic oscillator. For this, all these achievements are based on a double pass of the laser beam in the cell, which allows to achieve a large total length of the laser beam in a small volume.

The Figures 7 to 9 illustrate three different embodiments making it possible to simultaneously perform a double pass in the gas cell 106 and a protection of the laser source 102 towards the reflections. A common feature of these various embodiments is the presence of a semitransparent mirror 107 which passes a portion of the laser beam passed through the gas cell 106 to reach a photodetector 109 for servocontrol of the cell temperature. Alternatively, these embodiments could be simplified by removing this photodetector 109 and using a non-transparent mirror.

These three embodiments differ in the means used to direct the beam towards the cell and the photodetectors, and in the means used to prevent the beam reflected by the mirror from disturbing the laser source.

The figure 7 illustrates the first embodiment of the invention. The laser source 102 produces a linearly polarized laser beam which is directed towards a polarizer 103, the transmission axis of which is oriented so as to let the laser beam pass, then to a separator 101 whose percentage of separation is predefined. Part of the beam is thus transmitted to an optional photodetector 108b. The separator reflects the other part of the beam to a quarter-wave plate 105. Polarization linear is noted "P" for the part parallel to the axis of transmission of the polarizer (transmitted part) and "S" for the part perpendicular to the axis of transmission of the polarizer (part absorbed by the polarizer). In the figures, the part "P" is symbolized by solid circles and the part "S" by lines. The role of the blade 105 is to change the linear polarization of the laser beam into a circular polarization and this blade is oriented relative to the polarizer so as to generate a circular polarization. Indeed, the interaction between the light and the atoms of the gas cell 106 is optimal when it is performed with a circular polarization beam. A portion of the beam leaving the gas cell 106 is then reflected by a mirror 107, which reverses the direction of its circular polarization, and thus passes through the gas cell 106 a second time. When leaving the gas cell 106, the beam reaches the quarter wave plate 105. According to the predefined separation percentage of the separator 101, this beam is then partially transmitted and reaches the photodetector 108a. Another part of this beam is deflected by the separator 101 and is strongly attenuated by the polarizer 103 because its polarization is perpendicular to that of the transmission axis of the polarizer 103, the laser source 102 thus being protected from retro-reflections. A small portion of the beam passed through the gas cell 106 is transmitted by the mirror 107 and picked up by the photodetector 109.

The figure 8 illustrates the second embodiment. It differs from the first mode described above by the use of a separator 101 which reflects the beam in a first polarization and passes the beam in a second polarization. Thus the beam leaving the laser source 102 is separated according to its polarization and the same principle applies to the reflected beam. It is thus not necessary to place a polarizer between the separator 101 and the laser source because the reflected beam is fully transmitted to the photodetector 108a. The linear polarization is denoted "P" for the part parallel to the polarization axis of the separator (part transmitted in the right angle configuration of the figure 8 ) and "S" for the part perpendicular to the polarization axis of the separator (part deflected at 90 °). In the figure 10 , the part "P" is symbolized by lines and the part "S" by solid circles. A small portion of the beam passed through the gas cell 106 is transmitted by the mirror 107 and picked up by the photodetector 109.

The figure 9 illustrates the third embodiment of the invention. In this figure, the deflection of the laser beam is provided by the semi-transparent mirror 107 which is disposed at an angle not perpendicular to the axis of the laser beam. Thus, the reflected beam does not reach the laser source 102 but is directed directly on the photodetector 108a. Advantageously, the mirror 107 is of concave shape for focusing the light beam reflected on the photodetector 108a. A small portion of the beam having passed through the gas cell 106 is transmitted by the mirror 107 and picked up by the photodetector 109. This concave shape of the mirror can also be used on the two embodiments of the Figures 7 and 8 providing the benefits described above.

A more complete exemplary embodiment corresponding to the second embodiment is illustrated in FIG. figure 10 . The separator 101 is in the form of a polarizing beam splitter cube (PBSC). This cube makes it possible to implement a double crossing of the gas cell 106, which doubles the interaction between the laser light and the atomic medium. We obtain a better atomic signal and thus a better stability of the frequency of the atomic clock.

On the figure 10 , the optical assembly is based on a miniature separator cube 101 whose sides are preferably less than or equal to 1 mm, the cube 101 acting as a separator. In a standard mode, the volume of the cube is typically 1 mm ³ . The light beam of the laser diode 102 arrives on one of the sides of the cube 101. According to one embodiment, the The laser diode is of the Vertical Cavity and Surface Emission Semiconductor (VCSEL) type emitting a diverging beam of light at 795 nm. In other embodiments, other types of laser diodes having wavelengths typically ranging from 780 nm to 894 nm may be used for a gas cell 106 containing Rubidium or cesium. This choice is dictated by the atomic composition of the gas cell. According to one embodiment, a collimating lens may be added in front of the laser diode to produce a non-diverging laser beam.

According to a standard mode, the light produced 112 by the laser 102 has a linear polarization and is attenuated by an absorbent neutral filter 104a. A different type of filter can be used in other embodiments. The presence of this filter is not necessary for the invention. A half wave plate 104b may be used to change the angle of the linear polarization of the laser source. In combination with the miniature cube 101, the half wave plate 104b plays the role of a variable attenuator. In other embodiments, the use of the half-wave plate 104b may be omitted and the light intensity ratio between the beams transmitted and reflected by the cube 101 is adjusted by an appropriate orientation of the polarization axis. linear light emitted by the laser relative to the separator cube. A quarter-wave plate 105 is placed at the cube outlet against the face from which the laser beam deflected by the separator 101, or at right angles from the incident beam to the cube. The fast axis of the quarter wave plate 105 is oriented so that the incident linear polarization 113 is changed to a circular polarization 114 in a first direction of rotation. In other embodiments, the quarter-wave plate 105 is oriented such that the incident linear polarization 113 is changed to a circular polarization in a reverse direction of rotation to the first. The circular polarization laser beam 114 passes through the gas cell 106 and reaches the mirror 107. The latter only returns the ray partially and a portion of the ray passes through the mirror 107 to go towards the photodetector 109. According to a standard mode, the gas cell is made of glass-silicon-glass by MEMS (electromechanical microsystem) techniques with an interior volume of typically 1 mm ³ and filled with an absorbent medium of the atomic vapor type of alkali metal (Rubidium or Cesium), and a buffer gas mixture. In a standard mode, the gas cell is filled with natural Rubidium and a mixture of nitrogen and argon as a buffer gas. In other embodiments, other types of cells may be filled with different buffer gases. In a particular embodiment, a cylindrical miniature cell can be used. According to another particular embodiment, the gas cell can be integrated in the PBSC 101. The cell 106 can be filled with other types of alkaline metal vapor (rubidium-85, rubidium-87, cesium-133 for example) and other types of buffer gas (Xe, Ne for example).

The figure 11 illustrates a double-pass optical design based on the second embodiment corresponding to the figure 8 , with a straight geometry that is very similar to the right-angle and double-pass design depicted on the figure 10 . The main difference lies in the position of the entity "gas cell 206, quarter-wave plate 205, semitransparent mirror 207 and photodetector 209" and photo-detector 208b. In the model of figure 11 , the gas cell 206 is placed above the PBSC 201 and is therefore located vis-à-vis the laser 202. In this way, the polarization light beam P 213 transmitted by the PBSC and then modified circular polarization beam by the quarter-wave plate 205 interacts with the atomic medium. The S-polarization light beam 217 is reflected by the PBSC 201 and the right-angle photodetector 208b is used to measure the laser power. Apart from these differences, the operating principle of this embodiment is the same as for the previous model.

The figure 12 illustrates the schematic representation of the double-pass straight geometry case of the embodiment of the Raman clock according to the first embodiment, corresponding to the figure 7 . The digital coding starts at 201 for this design, keeping the same tens and units as those used on Figures 7 to 9 for the same elements. A separator cube 201 is used, whose percentage of separation is predefined so as to have a minority reflection and a majority transmission, of about 2% and 98% respectively (+/- 2%). The retro-reflected beam 216 is then mainly deflected towards the photodetector 208a. In this model, the gas cell entity 206 is placed above the separator cube 201 and is therefore located with respect to the laser 202. The photodetector 208b is placed at right angles, where the light beam 212 emitted by the laser 202 is reflected 218 by the separator cube 201 and is used for example for the measurement of laser power. The operating principle of this design remains similar to the previous descriptions.

The figure 13 illustrates a device according to the first embodiment and right angle geometry. The separation percentage of the separator 101 is predefined so as to have a minority transmission and a majority reflection of about 2% and 98% respectively (+/- 2%). After its interaction with the alkali metal vapor atoms, the incident light beam 114a and the stimulated Raman scattered light beam (called the Raman beam) 114b are reflected by a mirror 107. In a standard Raman embodiment the mirror 107 is coated with silver, it is inclined (typically 2 to 20 degrees) and / or eccentric with respect to its axis of symmetry and the axis defined by the incident laser beam and is concave with a chosen focal length to focus the retro-reflected light beams 115 (incident beams and Raman) on the photodetector 108a. Mirror 107 has a typical transmission of a few percent. These percent of transmitted light reaching the surface of the photodetector 109 are used to measure the absorption spectrum. In a different embodiment Raman, the exit window of the gas cell 106 is concave, coated with silver (or other metal, such as gold) and acts as a reflector. In other embodiments, the coating of the exit window of the mirror can be made of dielectric layers.

The retro-reflected light beams 115 (incident and Raman) pass through and interact a second time with the atomic medium (double pass). The quarter-wave plate 105 transforms these circularly polarized light beams into linear polarization light beams 116. These light beams are mainly deflected 119 (incident and Raman) and reach the first photodetector 108a which records the frequency beat between the beam. incident beam and the Raman beam. In a standard Raman embodiment, the first photodetector 108a is a high-speed semiconductor photodetector (silicon or gallium arsenide) which is positioned at the focus of the concave mirror 107. In other embodiments Raman, different types of high-speed photodetectors may be used. The second photodetector 108b records the light 118 coming directly from the laser 102 and initially transmitted by the miniature divider cube 101. In this manner, the output power of the laser diode 102 can be measured and adjusted. Optionally, the photodetector 121 records the retro-reflected beam 117 transmitted by the separator 101. The diaphragms 110 and 111 are used to prevent unwanted light from reaching the photodetectors if their dimensions are larger than those of the miniature separator cube 101 .

The figure 14 illustrates the third embodiment of the Raman clock, not based on a splitter cube but on a simple double-pass geometry. The light emitted by the laser source 102 is linearly polarized, converted into circular polarization by a quarter-wave plate 105 before passing through the cell 106, reflection on the mirror 107, second passage in the cell, and detection on a first photodetector 108a. The mirror 107 is semi-transparent, with a second photodetector 109 placed behind the mirror. This use of the semi-transparent mirror 107 allows the detection of light having interacted with the atoms of the cell by the photodetector 109. In order to prevent the beams retro-reflected by the mirror from disturbing the laser source 102, it is also advantageous to place a polarizer 103 in front of the laser source 102 and with a transmission axis parallel to the polarization of the beam emitted by the laser source 102.

• un filtre neutre 104 placé entre la source laser 102 et la lame quart d'onde 105 afin d'ajuster la puissance du faisceau laser
• un filtre réflectif incliné 104 placé entre la source laser 102 et la lame quart d'onde 105 afin de réfléchir une partie du faisceau laser et d'ajuster sa puissance
• un troisième photodétecteur 108b placé de manière à enregistrer la lumière réfléchie par le filtre réflectif incliné 104 pour l'asservissement de la puissance optique du laser 102.

Optionally, you can also use the following elements:

• a neutral filter 104 placed between the laser source 102 and the quarter wave plate 105 in order to adjust the power of the laser beam
• an inclined reflective filter 104 placed between the laser source 102 and the quarter-wave plate 105 to reflect a portion of the laser beam and adjust its power
• a third photodetector 108b positioned to record the light reflected by the inclined reflective filter 104 for servocontrolling the optical power of the laser 102.

In note, in these achievements described in connection with the Figures 7 to 14 , the photodetector 108a, 208a serves to detect the beat induced by the Raman effect of the gas present in the cell 106, 206, and is therefore a photodetector adapted for the detection of microwaves. The first photodetector 108a has a very narrow bandwidth and centered around the resonant frequency of the atoms to maximize its signal detection efficiency. The high atomic resonance frequency (typically> 1GHz) has the consequence of having a small photodetector 108a. This specification is not compatible with a signal detection that interacted with the atoms of the cell to adjust for example the temperature of the cell, which is implemented by the photodetector 109, 209 and / or the photodetector 108b, 208b. For the latter, a low cut-off frequency (typically <100 kHz), or even a DC operation are indicated. This is why it is preferable to have at least two detectors, one 108a serving for the detection of the clock signal, the other 109 for controlling the temperature of the cell. The ideal way to achieve this second detection of a signal having interacted with the atoms of the cell is to use a semi-transparent mirror 107 for reflection and to place behind this mirror a photodetector 109, as has been illustrated. It is also advantageous if the mirror 107 is of concave shape, as illustrated in FIG. figure 14 , the concave shape being intended to focus the light beam reflected on the photodetector 108a. As a note, these last photodetectors are optional.

Claims (19)

A method of transmitting a time signal by an atomic clock, comprising the steps of: sending a laser beam (11; 112; 212) from a laser source through a cell (3; 106; 206); detecting a frequency of beats obtained between the beam originating from the laser source and transmitted within the cell and the beam induced by the Raman effect within the atoms of the cell (3; 106; 206);
and characterized in that it comprises a method of igniting the atomic clock comprising the following steps: a first phase of searching for the optimum injection current of the open-loop laser of the atomic clock, - A second phase of ignition of the clock comprising the closed loop of the clock by the return of the microwave signal received at the output of the cell for the control of the injection current of the laser. A method of transmitting a time signal by an atomic clock according to the preceding claim, characterized in that the first phase of searching for the optimum laser injection current of the atomic clock ignition process comprises the following steps : - Open looping of the atomic clock; - Scanning of the laser frequency and identification of the maximum absorption point (Vmax) and the corresponding injection current (Imax) as well as the minimum absorption point (Vmin) of the absorption peak (51) associated with the point (Vmax) maximum absorption and injection current (Imin) corresponding; - Determination of an initial injection current (ILD) by adding a certain delta threshold value to (Imin) or subtracting it from (Imax), in order to be in the interval [Imin; Imax], far from the limits (Imin, Imax). A method of transmitting a time signal by an atomic clock according to the preceding claim, characterized in that the second ignition phase of the atomic clock ignition process comprises the following steps: - Closed loop of the clock by the return of the microwave signal received at the output of the cell for controlling the injection current of the laser; - Adjusting the laser injection current to the predetermined value (ILD); - Verification of obtaining the resonance phenomenon of the output clock; - In case of non-resonance of the clock, slight modification of the value of the injection current (ILD) according to a predefined step, and repetition of this step until the resonance phenomenon. A method of transmitting a time signal by an atomic clock according to one of the preceding claims, characterized in that it comprises a step of adjusting the power of the laser. A method of transmitting a time signal by an atomic clock according to one of the preceding claims, characterized in that it comprises a step of returning the microwave signal received at the output of the cell for controlling the injection current of the laser and that it comprises a zero control of the frequency of the laser. A method of transmitting a time signal by an atomic clock according to one of the preceding claims, characterized in that it comprises a measurement of the temperature of the atomic clock for correcting the time signal emitted by the clock according to the temperature. A method of transmitting a time signal by an atomic clock according to one of the preceding claims, characterized in that it comprises a temperature control of the atomic clock. A method for transmitting a time signal by an atomic clock according to the preceding claim, characterized in that it comprises an operation of the atomic clock at a temperature of less than or equal to 40 ° C, or operation at a temperature less than or equal to 35 ° C. A method of transmitting a time signal by an atomic clock according to one of the preceding claims, characterized in that it comprises the operation of the atomic clock in a zero magnetic field. Atomic clock, comprising a laser source (1; 102; 202), a cell (3; 106; 206) comprising an atom vapor and arranged to receive a laser beam (11; 112; 212) from the source laser, and a frequency detection system arranged to receive the laser beam from the cell (3; 106; 206) to detect a beat frequency obtained between the beam from the laser source and transmitted within the cell and the Raman-induced beam within the atoms of the cell (3; 106; 206) characterized in that it comprises a processing center (23) which implements the method of transmitting a time signal by atomic clock according to one of the preceding claims. Atomic clock according to the preceding claim, characterized in that it comprises a zero control of the frequency of the laser. Atomic clock according to one of claims 10 or 11, characterized in that the beat frequency detection system comprises a photodetector (4; 108a; 208a) and an amplifier. Atomic clock according to one of claims 10 to 12, characterized in that it comprises a current source (24) for the laser, a diplexer (34), and a return link from the frequency detection system to the diplexer (34) for combining the signal detected by the detection system with the current source (24) for controlling the injection current of the laser. Atomic clock according to claim 12 or 13, characterized in that the frequency detection system is a signal detection system (i _PD ) corresponding to Raman-induced beats, with a narrow spectral content centered around a center frequency (ω _C ), comprising at least a first inductive element (L1) which is connected to the photodetector (4; 108a; 208a) and a parasitic capacitance (C _IN ) parallel to the photodetector, forming a resonance circuit for selecting the signal to be detected, said resonance circuit having a resonant frequency which corresponds to the center frequency (ω _C ). Atomic clock according to one of claims 10 to 14, characterized in that it comprises at least one second photodetector (108b, 109; 208b, 209). Atomic clock according to one of claims 10 to 15, characterized in that it comprises at least one mirror (107; 207) for reflecting the laser beam and enabling it to pass at least a second passage through a cell (3; 106; 206 ) before reaching the frequency detection system. Atomic clock according to the preceding claim, characterized in that it comprises a semi-transparent mirror (107; 207) and a second photodetector (109) placed behind the mirror (107). Atomic clock according to one of claims 10 to 17, characterized in that it comprises a cell (3; 106; 206) comprising Cesium or Rubidium. Atomic clock according to one of claims 10 to 18, characterized in that it comprises a heating device.

Images