Method of calibrating an atomic-functioning apparatus

Abstract

This method makes it possible to carry out the auto-calibration of the electromagnetic coils of an apparatus such as an atomic clock, a magnetometer or a gyroscope by injecting successive currents into the coils and measuring the magnetic fields induced in order to calculate the transfer coefficients (field/current) of each of the coils and the real angles that they form with very great precision.

Description

The subject of the invention is a method of calibrating an atomic-functioning apparatus, that is to say one using optical pumping of a light beam, and also comprising electromagnetic coils; it may be an atomic clock, a magnetometer or a gyroscope.

Electromagnetic coils are used in various ways in such apparatus, for making magnetic measurements or producing artificial magnetic fields. Very precise control of the coils is often necessary in order to obtain correct results. Calibration of the coils may encompass the transfer function, that is to say the ratio of the magnetic field that is induced therein as a function of the current passing through them, and the direction of this field with respect to a reference direction.

A calibration method is proposed here in which no additional apparatus is used for calibrating the characteristics of the magnetic fields induced by the coils, which thus carry out auto-calibration.

In a general form, the invention concerns a method of calibrating an apparatus comprising three electromagnetic coils and a light beam subjecting a gaseous medium to optical pumping, the method comprising a determination of gain coefficients of the coils and angles between the coils, characterised in that it comprises:

• • a cell filled with a measuring medium;
  • an optical beam subjecting the measuring medium to optical pumping;
  • three electromagnetic coils each generating a magnetic field axial to the cell;

the method being characterised in that it comprises the following steps:

• • successive injection of at least one current (i_x, i_y, i_z) for each coil;
  • measuring the magnetic field generated by each of the coils;
  • determining a gain coefficient for each coil (T);
  • simultaneous injection of currents in the coils in each pair of coils, including two currents in one of the coils in the pair and at least one current in the other coil in the pair, in accordance with at least two different states of combination of said currents (i_x i_y; −i_x i_y; −i_x −i_y; i_x −i_y);
  • measuring the magnetic fields generated by said combinations of currents;
  • determining at least one angle (α, θ, γ) between the coils.

This general form of the invention is applicable when the ambient magnetic field is disregarded, stopped by shielding or compensated for by the creation of an artificial or opposing field. A method for compensating for the field may comprise the following steps:

• • generation of two orthogonal oscillating fields;
  • generation of an essentially static magnetic field by each of the coils;
  • compensation for the ambient field by said essentially static field by magnetic adjustment generated for each of the coils.

If the field is not negligible, the method preferably comprises a successive injection of two currents for each coil for determining the gain coefficient and a simultaneous injection of currents in the coils in each pair of coils, including two currents in each of the coils in the pair according to four different states of combination of said currents.

Advantageously, the magnetic fields are measured by an application of radio frequencies and a search for resonance of the gaseous medium at a Larmor frequency.

The invention will now be described in the detail of its various aspects in relation to the figures, among which:

FIG. 1 is general view of the apparatus,

FIG. 2 is a definition of the angles measured between the coils.

FIG. 1 is an example of implementation of the method of the invention. This FIG. 1 illustrates a vector magnetometer that comprises a laser (1) emitting a beam (2). The beam (2) passes successively through a rectilinear polariser (3) and through a cell (4) filled with helium 4 that it subjects to optical pumping. A photodiode (5) collects the light restored by the content of the cell (4). The device comprises three coils (6, 7 and 8) supplying magnetic fields in three nominally orthogonal axes, a radio-frequency generator (9) and a electronic card (10) for controlling the laser (1) and the current passing through the coils (6, 7 and 8) according to the requirements of the method, for calibration or other reasons (ambient field compensation, creation of stabilisation field etc). Apparatus or methods where the ambient magnetic field is corrected effectively by application of an artificial static magnetic field of suitable intensity and direction are described in the documents FR-A-2 924 826 and 2 924 827, to which reference is made since the teaching thereof can be repeated here. Several currents may pass alternately through each of the coils (6, 7 and 8) at different frequencies. The radio-frequency generator (9) emits electromagnetic waves at the Larmor frequency in the invention, again by means of the coils (6, 7 and 8).

FIG. 2 illustrates the parameters used in the remainder of the disclosure. It is the three angles α, γ, θ between the axes ({right arrow over (e)}_x, {right arrow over (e)}_y, {right arrow over (e)}_z) of the coils (6, 7, 8) defined with respect to a direct orthonormed reference frame ({right arrow over (u)}_x, {right arrow over (u)}_y, {right arrow over (u)}_z). The relative orientation of the reference frame ({right arrow over (e)}_x, {right arrow over (e)}_y, {right arrow over (e)}_z) of the instrument, defined by the directions of these axes with respect to the orthonormed reference frame, is defined as follows: {right arrow over (e)}_x coincides with {right arrow over (u)}_x, {right arrow over (e)}_y is contained in the plane ({right arrow over (u)}_x, {right arrow over (u)}_z) and forms an angle α with {right arrow over (u)}, and the projections of {right arrow over (e)}_z on the planes ({right arrow over (u)}_x, {right arrow over (u)}_z) and ({right arrow over (u)}_y, {right arrow over (u)}_z) form the angles γ and θ with {right arrow over (u)}_z. The base vectors of the reference frame of the coils ({right arrow over (e)}_x, {right arrow over (e)}_y, {right arrow over (e)}_z) are therefore expressed in the orthonormed reference frame ({right arrow over (u)}_x, {right arrow over (u)}_y, {right arrow over (u)}_z) in accordance with the following equations:

${\stackrel{->}{e}}_x={\stackrel{->}{u}}_x;$ ${\stackrel{->}{e}}_y=-\sin \alpha *{\stackrel{->}{u}}_x+\cos \alpha *{\stackrel{->}{u}}_y;$ ${\stackrel{->}{e}}_z=\frac{\tan \theta }{\sqrt{1+{\tan }^2\theta +{\tan }^{2\gamma }}}*{\stackrel{->}{u}}_x+\frac{\tan \gamma }{\sqrt{1+{\tan }^2\theta +{\tan }^{2\gamma }}}*{\stackrel{->}{u}}_y+\frac{\tan 1}{\sqrt{1+{\tan }^2\theta +{\tan }^{2\gamma }}}*{\stackrel{->}{u}}_z.$

The purpose of the calibration procedure is to make it possible to precisely evaluate firstly transfer functions (gain coefficients) between the magnetic field and the current, denoted T_x, T_y, T_z for the coils (6, 7 and 8) and secondly the differences in orthogonality between the respective directions thereof. To do this, scalar magnetic measurements are generated with the atomic instrument by injecting currents successively in each of the coils (6, 7 and 8) and then simultaneously in several of them. This choice also guarantees that, for each series of measurements, the amplitude of the resonance signals of the light of the beam 2 at the Larmor frequency are sufficient to obtain a signal to noise ratio compatible with the required precision. The choice of the order of magnitude of the scalar fields to be generated stems from the expected precision for the calibration of the coils (6, 7 and 8). By way of example for guaranteeing absolute precision of 0.1 nT in a field the modulus of which may be as much as 5 μT, it is necessary to determine the current transfer functions on magnetic field at better than 2.10⁻⁵ and the orthogonality differences to within a millidegree. It will be assumed hereinafter that the polarisation direction {right arrow over (e)}₀* of the photons in the case of rectilinear polarisation is colinear with {right arrow over (e)}_x.

The procedure in the example of FIG. 1 will be detailed in the case of the coil (6) of axis {right arrow over (e)}_x, the determination of the transfer functions of the two other coils (7 and 8) being at every point similar.

The artificial field being created in the direction {right arrow over (e)}_x, the radio-frequency field by virtue of which a resonance will be induced according to the spin frequency or the Larmor frequency could be applied indifferently in the other two axes {right arrow over (e)}_y or {right arrow over (e)}_z.

A current i_x is injected successively positive and negative in the coil (6) and in each case the resonant frequency F of the magnetometer operating in scalar mode is measured. The modulus of the magnetic field is then equal to

$B\pm i_x=\frac{B}{y}\pm \mathrm{ixTx},$ where B_x is the component of the local field in the direction of the axis {right arrow over (e)}_x, ignoring the components of the local field in the perpendicular plane. It should be noted that these components can always be compensated for in advance by virtue of known vector measurement methods, disclosed in particular in the thesis by Gravarand et al “On the calibration of a vectorial 4He pumped magnetometer” which appeared in Earth, Planets and Space, 2001, volume 53, no 10, pages 949 to 958, the state thesis of Dupont-Roc of 1972 “Study of some effects relating to optical pumping in weak field” and the applications FR-A-2 924 826 and 2 924 827 already mentioned. For calibration precision of 10⁻⁶, it is necessary for the transverse magnetic field to be less that 10⁻³ times the magnetic field ixT_x generated by the coil concerned.

The following system of equations is derived from this:

$\hspace{1em}\{\begin{array}{c}{B}_{+i_x}=B_x+i_x{T}_x\\ B_{-i_x}=B_x-i_x{T}_x\end{array}$

by virtue of which T_x is determined immediately, i_x being known.

The effects of the fluctuations of the ambient magnetic field during this calibration phase can be attenuated by repeating these alternating measurements on several occasions. Where applicable, a drift of the projection B_x of the ambient field can be estimated in order to improve the precision of the calculations.

As mentioned, the same method is repeated for the other coils (7 and 8).

It is described below how to determine the orthogonality differences of the three coils (6, 7 and 8). As before, sequences of measurements of scalar fields corresponding to artificial fields created by the coils (6, 7 and 8) are proceeded with, using here combinations of pairs of coils chosen so as guarantee a signal amplitude allowing measurement with the required resolution.

For determining the angle α between {right arrow over (u)}_y and {right arrow over (e)}_y, currents are injected into the coils (6 and 7) of axes {right arrow over (e)}_x and {right arrow over (e)}_y according to the sequence (+i_x+i_y; −i_x+i_y; −i_x−i_y; +i_x−i_y), combining the two current values of each of the active coils according to the four possible states. The currents i_x and i_y are chosen so that the modulus of the field is 70% of the current applied during the phases of calibrating the transfer functions of the coils.

The following system of equations is then obtained:

$\hspace{1em}\{\begin{array}{c}{\left(B_{\mathrm{ix}+\mathrm{iy}}\right)}^2-{\left(B_{-\mathrm{ix}+\mathrm{iy}}\right)}^2=4i_x{T}_x\left(B_x-i_y{T}_y\sin \alpha \right)\\ {\left(B_{\mathrm{ix}-\mathrm{iy}}\right)}^2-{\left(B_{-\mathrm{ix}-\mathrm{iy}}\right)}^2=4i_x{T}_x\left(B_x+i_y{T}_y\sin \alpha \right)\end{array}$

from which sin α is immediately derived.

For determining the angles θ and γ that characterise the direction of {right arrow over (e)}_z in the orthonormed reference frame ({right arrow over (u)}_x, {right arrow over (u)}_y, {right arrow over (u)}_z), conjoint measurements are necessary. The following two four-state sequences {i_x+i_z/2; −i_x+i_z/2; −i_x−i_z/2; +i_x−i_z/2} and {i_y+i_z/2; −i_y+i_z/2; −i_y−i_z/2; +i_y−i_z/2} are then successively effected, according to notations similar to those above. The currents i_x, i_y and i_z will be chosen so that the modulus of the fields that they create is approximately 50 μT, giving i_x, i_y close to 80% of the currents supplied during the phases of calibration of the transfer functions of the coils (6, 7 and 8) and i_z/2 close to 40% of this value.

The following systems of equations are then obtained:

$\begin{array}{cc}\{\begin{array}{c}{\left(B_{\mathrm{ix}+\mathrm{iz}/2}\right)}^2-{\left(B_{-\mathrm{ix}+\mathrm{iz}/2}\right)}^2=4i_x{T}_x\left(B_x+\Theta *i_z{T}_z\right)\\ {\left(B_{\mathrm{ix}-\mathrm{iz}/2}\right)}^2-{\left(B_{-\mathrm{ix}-\mathrm{iz}/2}\right)}^2=4i_x{T}_x\left(B_x-\Theta *i_z{T}_z\right)\end{array}& \left(I\right)\\ \{\begin{array}{c}\begin{array}{c}{\left(B_{\mathrm{iy}+\mathrm{iz}/2}\right)}^2-{\left(B_{-\mathrm{iy}+\mathrm{iz}/2}\right)}^2=4i_y{T}_y\{\cos \alpha \left[B_y+\Gamma *i_z{T}_z\right]-\\ \sin \alpha \left[B_x+\Theta *i_z{T}_z\right]\}\end{array}\\ \begin{array}{c}{\left(B_{\mathrm{iy}-\mathrm{iz}/2}\right)}^2-{\left(B_{-\mathrm{iy}-\mathrm{iz}/2}\right)}^2=4i_y{T}_y\{\cos \alpha \left[B_y-\Gamma *i_z{T}_z\right]-\\ \sin \alpha \left[B_x-\Theta *i_z{T}_z\right]\}\end{array}\end{array}\mathrm{where}\Theta =\frac{\tan \theta }{2\sqrt{1+{\tan }^2\theta +\mathrm{ta}}}\mathrm{and}\Gamma =\frac{\tan \gamma }{2\sqrt{1+{\tan }^2\theta +{\tan }^{2\gamma }}}& \left(\mathrm{II}\right)\end{array}$

Θ and Γ and then θ and γ are then successively derived, which closes the calibration phase.

In a similar manner to the above, the component of the local magnetic field that is orthogonal to the magnetic field generated for the coils is ignored or compensated for, and the radio-frequency field is orthogonal to this field (it may be applied by the third coil, here 8, of axis {right arrow over (e)}_z).

The invention thus makes it possible to mitigate the drifts in the transfer functions of atomic instruments and to calibrate the coils of these instruments in a weak magnetic field

$\left(B<\frac{2\pi ·f_R}{\gamma }\right)$ in which the scalar measurements are not possible if the relaxation frequency time f_R of the spins is greater than the Larmor frequency f

$\left(f=B·\frac{\gamma }{2\pi }\mathrm{and}\frac{\gamma }{2\pi }=28.04\mathrm{Hz}\text{/}\mu T\mathrm{for}\mathrm{helium}4\right).$ The invention makes it possible not to use any additional instrument and to obtain the reference measurement as close as possible to the instrument to be calibrated. It thus improves the quality of the calibration. It makes it possible to use atomic instruments as a magnetometer since, in addition to the estimation of the current transfer function on a field of the coil in question, a measurement of the value of the component B_x, B_y or B_z of the local field in this direction is also obtained.

The method can be applied to gases of the alkaline or helium type. The polarisation of the light and the detection method must be adapted to the direction of the magnetic field imposed by the coils. It should be noted that the method can also be applied in circular polarisation provided that they have alternately, according to the direction of the magnetic field, different operating modes of the magnetometers in order to guarantee magnetic measurement conditions in accordance with table II given below. Finally, the invention makes it possible not to be dependent on the amplitude or direction of the ambient magnetic field.

The method has up until now been explained for an apparatus immersed in a relatively strong ambient magnetic field (B_x,y,z≠0 in the equations). It is simplified if the field is negligible or zero, stopped by shielding or compensated for by one of the methods indicated above. It is then possible to consider that B_x,y,z≈0, and then a single current (for example i_x or −i_x, instead of i_x and −i_x) suffices to determine the gain coefficient (T_x) of the associated coil, here (6), by simply solving an equation, such as B_+ix=i_xT_x, instead of a system with a pair of equations. Likewise, the angles between coils can be determined by single equations instead of pairs of equations. For example, (B_ix+iy)²−(B_−ix+iy)²=4_ixT_x (−i_yT_y−i_x) suffices to give α, that is to say it suffices to inject a current (i_y) into one of the coils and two currents (i_x, −i_x) in the other one of the coils, according to two combination states (i_xi_y; −i_xi_y). The same remarks apply to the determinations of θ and γ.

Table I establishes various techniques that can be used for each functional block.

TABLE I
		Means of			Measurement/
Flow of		generating			control of
polarised	Gas/	a magnetic	Detec-		magnetic
photons	Spins	field	tion	Controls	field
Laser	He4 (a	Generated	Photo-	Digital	Scalar, via
diode +	helium	by one	detector	signal	the Larmor
polariser	plasma	coil, at		processor	frequency
	must be	least
	created)
Lamp +	Alkaline			Computer	Vectorial,
polariser	gas (Cs,				via a zero
	Rb, K,				magnetic
	etc)				field
					control
VCSEL +				Micro-	Vectorial,
polariser				processor	via the
					Larmor
					frequency

The first line of this table corresponds in reality to the example described before. The laser (1) may be a diode of the InGaAs type with a wavelength of 1083 nm and a power of 1 mW. The cell (4) is filled with He⁴ at 1 torr. It is cylindrical, made from Pyrex, and has volume of 10 cm³. Two electrodes are placed up against it and are connected to the generator (9) creating the helium plasma; the radio-frequency waves are around 25 MHz in frequency and 100 mW in power. The vectorial coils (6, 7 and 8) also produce two magnetic fields at low frequency H_Ω cos Ωt and H_Ω cos ωt where

$\frac{\Omega }{2\pi }=3\mathrm{kHz},$ B_Ω=50 nT,

$\frac{\omega }{2\pi }=20\mathrm{kHz},$ B_ω=1000 nT, as well as the compensation fields for the ambient magnetic field B. Finally, the coils (6, 7 and 8) also produce the radio-frequency magnetic field at the Larmor frequency, which induces the magnetic resonance in the gaseous medium subjected to optical pumping. The exact frequency depends on the amplitude of the magnetic field to be measured

$(f=B·\frac{\gamma }{2\pi }$ where f is the frequency of the radio-frequency field and

$\frac{\gamma }{2\pi }=28.04$ Hz/nT as seen).

Table II indicates four techniques, known by the names M_x, M_z, Bell & Bloom and CPT in the art, which also make it possible to exploit the invention with a circular polarisation of light.

			Bell &
Technique	Mx	Mz	Bloom	CPT
Angle	45°	0°	90°	0°
between
static
magnetic
field and
direction
of laser
Radio-	Angle of 90°	Angle of 90°	Absent	Absent
frequency	with the	with the
magnetic	laser and	laser and
field	static	static
	magnetic	magnetic
	field	field
Comment			Laser	Laser
			intensity	intensity
			modulation	modulation

Claims

1. A method of calibrating a measurement apparatus with optical pumping functioning, said method comprising: successive injection of at least one current for each of three electromagnetic coils each generating a magnetic field axial to a cell filled with a measuring medium;measuring the magnetic field generated by each of the coils;determining a gain coefficient for each coil, which is a ratio between the magnetic field generated and the current injected;grouping the coils in pairs;simultaneous injection of currents in the coils in each pair of coils, including two currents in one of the coils in the pair and at least one current in the other coil in the pair, in accordance with at least two different states of combination of said currents;measuring the magnetic fields generated by said combinations of currents, an ambient magnetic field being ignored, stopped by shielding or compensated for; anddetermining at least one angle between the coils according to the magnetic fields measured, the gain coefficients and the currents injected.

2. The method according to claim 1, comprising a step of successive injection of two currents of different values and a step of subtracting two corresponding measurements of magnetic fields colinear with the coil.

3. The method according to claim 1 or 2, in which a step of compensating for the ambient magnetic field comprises the following steps: generation of two orthogonal oscillating fields in order to adjust for direction and intensity the field generated by each of the coils;generation of a essentially static magnetic field by each of the coils;compensation for the ambient field by said essentially static field by adjusting the essentially static magnetic field generated for each of the coils.

4. The method according to claim 1, in which there is provided a step of successive repetition of measurements of the magnetic field generated by each coil.

5. The method according to claim 1, wherein the magnetic fields are measured by an application of radio frequency and a search for a resonance of the measuring medium linked to a Larmor frequency.

6. A method of calibrating a measuring apparatus with optical pumping functioning, said method comprising: successive injection of at least one current for each of three electromagnetic coils each generating a magnetic field axial to a cell filled with a measuring medium;measuring the magnetic field generated by each of the coils;determining a gain coefficient for each coil, which is a ratio between the magnetic field generated and the current injected by a successive injection of two currents for each coil;grouping the coils in pairs;simultaneous injection of currents in the coils in each pair of coils, including two currents in one of the coils in the pair, in accordance with at least four different states of combination of said currents;measuring the magnetic fields generated by said combinations of currents; anddetermining at least one angle between the coils according to the magnetic fields measured, the gain coefficients and the currents injected.

7. The method according to claim 6, in which there is provided a step of successive repetition of measurements of the magnetic field generated by each coil.

8. The method according to claim 6, wherein the magnetic fields are measured by an application of radio frequency and a search for a resonance of the measuring medium linked to a Larmor frequency.