Apparatus and method for reflectometry by measuring optical interferometric noise signal

Abstract

The present invention relates to a method and apparatus for reflectometry (100) by measuring optical interferometric noise signal. According to the invention, the reflectometry apparatus comprises a light source (1), an optical beam splitter (4) spatially separating the modulated source beam (10) between an optical measuring arm (5) and a reference optical arm (7), a detection system (30) adapted to acquire, as a function of time and as a function of a sweep of the modulation frequency over several modulation periods, a power of an interferometric signal of the reflected beam (23) in at least one temporal superposition zone (25) between a pulse of a first reflected beam (21) at a reflection point (16) of the optical measuring arm and a pulse of a second reflected beam (22) by reflection of the second portion of source beam (12) on the optical reflector (8) of the reference optical arm. Figure for abstract: Figure 1

Description

Title of the invention: Apparatus and method for reflectometry by measuring optical interferometric noise signal Technical field

[0001] The present invention relates to the technical field of apparatus and methods for metrology of an optical device by reflectometry. The present invention finds applications in the detection of the position and amplitude of optical reflections distributed in an optical device. Prior art

[0002] There are already a number of methods for carrying out measurements of parasitic reflections in an optical device to be tested (or DUT acronym for Device Under Test in English).

[0003] Optical Time Domain Reflectometry (OTDR) consists of sending a light pulse of ultra-fine duration (generally between 10ns and 50ns) to one end of the optical device to be tested and continuously measuring parasitic reflections as a function of time. The duration between the emission of the light pulse and the reception of a reflection indicates the position of the reflection point (in optical length). To carry out several measurements, the OTDR emits several successive pulses, the time interval between two successive pulses being greater than the duration of a round trip in the DUT. The greater the optical length of the DUT, the lower the acquisition rate. The OTDR allows measurement over very long distances, such as in optical fibers.However, this method suffers from a blind zone at the beginning of the DUT and has temporal resolution limitations that translate into spatial resolution limitations of the position of parasitic reflections. The OTDR device and method have a fairly low cost.

[0004] White light interferometry is based on the use of a source whose coherence length, denoted Lco, is very low, of the order of 50pm to 500pm. The white light interferometer comprises a measuring arm in which the optical device to be tested is placed and a reference arm of variable length, for example based on a mirror that can move in translation. The optical return of the optical device is made to interfere with that of the reference arm. The measurement range is restricted by the amplitude of the moving reference arm. However, white light interferometry makes it possible to obtain very precise mechanical positioning accuracy. The cost of such an interferometer is very high.

[0005] Optical frequency domain reflectometry (or OFDR for Optical Frequency Domain Reflectometry (OFDR) is based on the use of a single-frequency source with a wavelength that can be tuned and has a long coherence length, the wavelength of which is scanned linearly. In OFDR, the return beam from the DUT is generally made to interfere with a part of the source beam. The position difference between the emitted and reflected beams is transformed into an optical frequency difference and generates a beat by interferometry at a frequency corresponding to the difference in optical frequencies. It is then necessary to perform a harmonic time analysis of the FFT type (acronym for Fast Fourier Transform) on a large sample with very rapid acquisition. OFDR is an expensive technique that allows access to a large measurement range, a little beyond 100m of optical length and with a good resolution around 50pm.

[0006] It is desirable to provide a method and system for performing measurements of parasitic reflections in an optical device to be tested, which is accurate over a large measurement range, without blind zone and inexpensive. Statement of the invention

[0007] In order to overcome the aforementioned drawbacks of the state of the art, the present invention proposes an apparatus for reflectometry by measuring optical interferometric noise signal, comprising: a light source capable of emitting a monochromatic source beam having a spectral width of less than 20 MHz; a modulation system adapted to temporally and periodically modulate the source beam and form a modulated source beam composed of source pulses modulated in power and wavelength at a given modulation frequency defining a modulation period, the modulation system being capable of varying the modulation frequency;an optical beam splitter arranged and configured to spatially separate the modulated source beam between a first portion of source beam directed towards an optical measurement arm comprising an optical device to be measured in reflection and a second portion of source beam directed towards a reference optical arm, the reference optical arm having a determined optical length L, L being greater than or equal to 20 meters and the optical device having an optical length less than L / 2, the reference optical arm comprising an optical reflector arranged at a distal end of the reference optical arm relative to the optical beam splitter, a difference in optical length between the optical measurement arm and the reference optical arm being greater than L / 2;the optical beam splitter being capable of optically recombining a first reflected beam formed by reflection of the first portion of source beam at at least one reflection point of the optical measuring arm and a second reflected beam formed by reflection of the second; portion of source beam on the optical reflector of the reference optical arm so as to form a reflected beam, the optical beam splitter being capable of transmitting the reflected beam to a detection system, the detection system being adapted to acquire, as a function of time and as a function of a sweep of the modulation frequency over several modulation periods, a power of an interferometric signal of the reflected beam in at least one temporal superposition zone between a pulse of the first reflected beam and a pulse of the second reflected beam, an electronic system for demodulating the power of the acquired interferometric signal, the electronic system being adapted to extract an optical interferometric noise signal based on a standard deviation of a power difference at the modulation period as a function of the modulation frequency,and a processing system comprising a computer adapted to process the optical interferometric noise signal as a function of the modulation frequency and to extract a reflectometry measurement as a function of an optical distance between at least one reflection point in the optical device to be measured and the optical reflector of the reference optical arm.

[0008] The device makes it possible, at low cost, to measure distributed optical reflections using a modulated optical source, the separation of the signal into two with an arm towards the DUT and a reference arm using a mirror at the end thus making it possible to recombine the signal reflected by the DUT and that reflected by the reference arm.

[0009] Unlike the white light interferometer, the return of the reference arm is time-shifted to avoid direct interference. This shift is achieved by using a very long reference arm. Similarly, the reference arm is fixed, thus greatly reducing the cost of the device, and no mechanical element is used in the proposal to determine the position of the parasitic reflection points in the DUT.

[0010] According to a particular and advantageous aspect, the calculator is adapted to extract from the reflectometry measurement a reflection amplitude of at least one reflection point in the optical device to be measured.

[0011] According to another particular and advantageous aspect, the light source comprises a distributed feedback laser diode.

[0012] According to one embodiment, the modulation system comprises means for modulating an electrical current supply to the light source.

[0013] Advantageously, the modulation system is adapted to perform a sweep of the modulation frequency with a frequency resolution of at least Ippm.

[0014] According to an exemplary embodiment, the modulation frequency varies from 650 kHz to 800 kHz with a variation step between 1 Hz and 100 Hz.

[0015] Advantageously, the modulation period has a duty cycle of between 5% and 90%, for example 25%.

[0016] According to a particular and advantageous aspect, the modulated source pulses have a power that is constant or temporally variable in a monotonic, exponential, linear or quasi-linear manner, over each modulation period and / or in which the modulated source pulses have a wavelength that is temporally variable in a continuous and monotonic, exponential, linear or quasi-linear manner, over each modulation period.

[0017] According to another particular and advantageous aspect, the detection system is adapted to sample the power of the interferometric signal at a sampling frequency higher than the modulation frequency, so as to obtain a number N of measurement points of the interferometric noise signal at each modulation period over a number M of successive modulation periods, with N ranging from 1 to 100 and M greater than 1000.

[0018] According to an exemplary embodiment, the detection system comprises an electronic low-pass filter having a cut-off frequency greater than or equal to the spectral width of the source.

[0019] Advantageously, the reference optical arm comprises a coil of optical fiber.

[0020] The invention also relates to a method of reflectometry by measuring optical interferometric noise signal, comprising the following steps: - emission by a light source of a monochromatic source beam having a spectral width less than 20 MHz; - periodic temporal modulation of the source beam to generate a modulated source beam composed of source pulses modulated in power and wavelength at a given modulation frequency defining a modulation period, the modulation frequency being variable; - optical separation of the modulated source beam between a first portion of source beam directed towards an optical measuring arm comprising an optical device to be measured in reflection and a second portion of source beam directed towards a reference optical arm of determined optical length L, L being greater than or equal to 20 meters, the reference optical arm comprising an optical reflector arranged at a distal end of the reference optical arm, the optical device having an optical length less than L / 2; - optical recombination of a first beam reflected by the optical measuring arm and a second beam reflected by the optical reference arm, so as to form a reflected beam; - detection as a function of time and as a function of a sweep of the modulation frequency over several modulation periods, of a power of an interference signal- metric of the reflected beam in at least one zone of temporal overlap between a pulse of the first reflected beam and a pulse of the second reflected beam, - demodulating the acquired interferometric signal power to extract an optical interferometric noise signal based on a standard deviation of a power difference at the modulation period as a function of the modulation frequency, - processing the optical interferometric noise signal as a function of the modulation frequency to extract a reflectometry measurement as a function of an optical distance between at least one reflection point in the optical device to be measured and the optical reflector of the reference optical arm.

[0021] Of course, the various features, variants and embodiments of the invention may be combined with each other in various combinations provided that they are not incompatible or mutually exclusive. Brief description of the drawings

[0022] Furthermore, various other characteristics of the invention emerge from the appended description given with reference to the drawings which illustrate a non-limiting form of embodiment of the invention and where:

[0023] [Fig.l] is a schematic view of a reflectometry device according to an exemplary embodiment;

[0024] [Fig.2] is a diagram of a power curve of a modulated source beam as a function of time;

[0025] [Fig.3] represents, as a function of time, a curve, in continuous lines, of the power of a modulated beam reflected by the reference optical arm and a curve, in dashes, of the power of a modulated beam formed by reflection in the optical measuring arm of a reflectometry device, the hatched zones indicating the zones of temporal superposition between these two modulated reflected beams;

[0026] [Fig.4] represents, as a function of time, an example of measurement of the power of the signal detected by the detection system over a modulation period;

[0027] [Fig.5] is an example of periodic sampling of the received signal;

[0028] [Fig.6] is another example of periodic sampling of the received signal;

[0029] [Fig.7] is an example of variation in wavelength of the source beam;

[0030] [Fig.8] is an example of wavelength variations respectively for a part of a pulse reflected in the optical measuring arm and a part of a pulse reflected in the optical reference arm;

[0031] [Fig.9] represents, as a function of time, a solid line curve of periodic modulation in power of a source beam, and respectively, a dashed line curve of periodic modulation in wavelength of the same source beam;

[0032] [Fig. 10] represents, as a function of time, a curve, in a continuous line, of power of a beam reflected by the reference optical arm, a curve, in dashes, of the power of a beam formed by reflection in the measuring optical arm of a reflectometry device and indicates the corresponding modulation period numbers for each reflected beam, in the case of a partial temporal overlap between the pulses reflected in the two arms;

[0033] [Fig. 11] represents, as a function of time, a curve, in a continuous line, of power of a beam reflected by the reference optical arm and a curve, in dashes, of the power of a beam formed by reflection in the measuring optical arm of a reflectometry device in the case where there is no temporal overlap between the pulses of the two arms;

[0034] [Fig. 12] schematically represents, as a function of time, a curve, in line continuous, of power of a beam reflected by the reference optical arm and a curve, in dashes, of power of a beam formed by reflection in the measuring optical arm in the case of perfect temporal overlap between the pulses of the two arms;

[0035] [Fig. 13] schematically represents, as a function of time, the power signal reflected electronic detection in the case of perfect time overlap between the pulses of the two arms;

[0036] [Fig. 14] schematically represents, as a function of time, an example of a signal of reflected power detected over three periods after electronic low-pass filtering in the case of perfect time overlap between the pulses of the two arms;

[0037] [Fig. 15] represents, as a function of time, a continuous line curve of a signal of reflected power before electronic low-pass filtering, in the case of partial time overlap between the pulses of the two arms;

[0038] [Fig. 16] represents, as a function of time, a power curve, in a continuous line, of a reflected power signal detected after electronic low-pass filtering, in the case of a partial time overlap between the pulses of the two arms;

[0039] [Fig. 17] represents two measurement curves of a signal representative of a deviation of reflected power at the modulation period as a function of the optical distance from the optical reflector of the reference arm: a dashed curve with, in the measuring arm, a cleaved section of optical fiber, and a solid line curve with, in the measuring arm, a complex optical device comprising several interfaces;

[0040] [Fig. 18] represents a zoom on a part of the solid line curve of [Fig. 17].

[0041] It should be noted that in these figures the common structural and / or functional elements different variants may have the same references.

[0042] Of course, the various characteristics, variants and embodiments of the invention may be combined with each other in various combinations provided that they are not incompatible or mutually exclusive. Detailed description

[0043] [Fig.l] schematically represents a reflectometry apparatus 100 according to an exemplary embodiment. The reflectometry apparatus 100 is based on a Michelson interferometer architecture, while operating outside of the usual interferometry conditions. As detailed below, the reflectometry apparatus 100 does not measure interference but rather an interferometric noise signal.

[0044] It is sought to detect at least one reflection point in an optical device 6 or DUT. The optical device 6 is capable of propagating light along at least one optical propagation axis. The optical device 6 comprises, for example, a waveguide-based device, such as one or more sections of optical fibers, one or more integrated optical components (also called PLC for Planar Lighwave Component), free-field devices consisting of an assembly of lenses and / or mirrors or any other optical component operating in transmission or reflection, the presence of interfaces or defects of which causes parasitic or voluntary reflections. The optical device 6 thus comprises one optical component or several optical components arranged in series. A reflection point corresponds, for example, to an optical defect or to an interface between two materials inside the optical device 6.A reflection point is likely to retroreflect at least part of a light beam propagating in the DUT along the optical axis of propagation, in the opposite direction to the incident light beam. We seek to detect the existence of such reflection points and, if so, to determine the position of each reflection point along the optical axis of propagation in the DUT. We also seek to measure the amplitude of the reflection on each reflection point.

[0045] The reflectometry apparatus 100 comprises a light source 1, also called a light source, a modulation system 2, an optical beam splitter 4, a detection system 30, an electronic system 40 for demodulating the signal and a processing system 50. The demodulation defined here is the subtraction between the signal detected at time t and the signal detected at time tT, T being a modulation period of the source modulation. The reflectometry apparatus 100 also comprises an electronic synchronization system 3 comprising, for example, a clock common to the source / modulation system and to the detection and demodulation system. Alternatively, the electronic synchronization system 3 comprises a system for detecting the control edges of the modulation system to define the acquisition instant(s) of the detection system.For example, the detection system includes a standard PIN photodiode type detector associated with an amplifier. transimpedance generator.

[0046] The light source 1 is adapted to emit a single-frequency (or monochromatic) source beam. The light source 1 comprises, for example, a DFB type laser diode or distributed feedback laser diode (Distributed FeedBack laser). The light source emits a source beam having a certain spectral width, generally between 1 MHz and 10 MHz, therefore with a coherence length of the order of a few tens of meters in air. The wavelength of the emitted light is chosen, for example, around 1550 nm to allow the use of standard telecom components. Unlike a white light interferometer, a broad spectral band light source is not used. Unlike an optical frequency domain reflectometer (OFDR), an extremely coherent light source does not necessarily have to be used, i.e. having a coherence length generally much greater than one km.

[0047] The modulation system 2 comprises an optical modulator external to the light source 1 or a system capable of directly modulating the light source 1. The modulation system 2 is adapted to temporally modulate the source beam and form a modulated source beam 10 composed of source pulses modulated in power at a modulation frequency F corresponding to a modulation period T. In addition, the modulation system 2 is here capable of varying the modulation frequency F. In one example, the light source 1 comprises a single-frequency DFB laser diode and the modulation system 2 is configured to periodically modulate the source of electrical current supplying the DFB laser diode so as to generate the source pulses. In another example, the light source 1 emits a continuous light beam and the modulation system 2 comprises an optical modulator, external to the source, arranged downstream of the light source 1.In this case, the optical modulator forms the source pulses.

[0048] The optical beam splitter 4 is connected to four branches: a first input branch connected to the light source 1, a second input-output branch connected to an optical measuring arm 5 comprising the optical device 6 to be measured in reflection, a third input-output branch connected to a reference optical arm 7 and a fourth output branch connected to the detection system 30. The optical beam splitter 4 is for example based on a 2x2 optical fiber coupler. In the forward direction of the modulated source beam, the optical beam splitter 4 divides the optical power received on the first input branch between the second input-output branch and the third input-output branch. In the return direction, after reflection of the split beams in the measuring arm and in the reference arm, the optical beam splitter 4 operates as a beam coupler or combiner.

[0049] The reference optical arm 7 comprises an optical reflector 8 arranged at a distal end of the reference optical arm 7 relative to the optical beam splitter 4. The reference optical arm 7 and the measuring optical arm 5 thus form the two arms of a Michelson-type interferometer.

[0050] The reference optical arm 7 has a determined optical length L, the optical length being taken between the optical beam splitter 4 and the optical reflector 8. The optical length L is greater than or equal to 20 meters and preferably greater than or equal to 100 meters, for example equal to 150 m. Advantageously, the reference optical arm 7 comprises a coil of optical fiber 9, which makes it possible to obtain such an optical length L in a reduced size, the coil of optical fiber having for example a diameter of approximately 5 cm. By way of example, the optical reflector 8 comprises a plane mirror. According to the present disclosure, the optical length L of the reference optical arm 7 is fixed. In addition, the reference optical arm 7 is manufactured so as to have no parasitic reflection between the optical beam splitter 4 and the optical reflector 8.Particularly advantageously, the reference arm includes a thermal regulation device to prevent thermal drift.

[0051] The optical measuring arm 5 comprising the optical device 6 has an optical length, denoted L1, taken between the optical beam splitter 4 and the distal end of the optical device 6. The optical length L1 is much less than the optical length L of the reference arm. Advantageously, L1 is less than L / 2. The difference in optical length between the optical measuring arm 5 and the optical reference arm 7 is preferably greater than 100 m, and thus generally greater than the coherence length of the light source. This configuration prevents the formation of stable interference between the beams reflected by the measuring arm 5 and by the reference arm 7.However, in this configuration, a source with a significant coherence length can be retained if the source is modulated directly, the modulation periodically applying a condition of zero emitted power and making it possible to eliminate the coherence between two successive pulses.

[0052] The detection system 30 comprises a detector adapted to acquire, as a function of time, the optical power of the beam propagating from the fourth branch of the optical beam splitter 4. The detector is for example a PIN photodiode associated with a transimpedance amplifier making it possible to convert the photodiode current into voltage. In a particularly advantageous manner, the detection system 30 comprises a low-pass electronic filtering device, having an adjustable cut-off frequency.

[0053] The electronic demodulation system 40 is arranged to receive the power signal acquired by the detection system 30, and preferably filtered. In particular, the electronic demodulation system 40 is adapted to convert the signal from power acquired by sampling it at the modulation period and to subtract the values ​​obtained at two successive periods. The electronic demodulation system thus makes it possible to obtain a periodic demodulation signal synchronized with the modulation of the source whose standard deviation over time defines the “interferometric noise signal”.

[0054] The processing system 50 comprises a computer adapted to process the optical interferometric noise signal as a function of the modulation frequency and thus form a reflectometry signal as a function of the optical distance (x) between at least one reflection point 16 in the optical device 6 to be measured and the optical reflector 8 of the reference optical arm.

[0055] We will now describe in detail the operation of the reflectometry device illustrated in [Fig.l].

[0056] [Fig. 2] shows an example of the power, denoted P(t) of the modulated source beam 10 as a function of time, denoted t. The duration of a source pulse is generally between 0.05 microseconds (ps) and 1 ps and the modulation frequency F between 100kHz and 2MHz, in other words the modulation period Ts between 0.5ps and 10ps. The shape of the power modulated pulses is for example rectangular, with a high level and a low level, the low level preferably being zero power. Other forms of power modulation are also considered, such as sawtooth or sinusoidal pulses, generally with a low level, preferably zero power. The duty cycle of the power modulation is generally between 5% and 50%.Unlike an optical time domain reflectometer (OTDR), here we do not use a single pulse of ultra-fine duration while looking for the delay of the optical returns of this initial pulse, but a series of periodic pulses whose juxtaposition to the return allows the measurement of parasitic reflections.

[0057] The optical beam splitter 4 receives the modulated source beam 10 on its first input branch. The optical beam splitter 4 spatially separates the modulated source beam 10 between a first portion of source beam 11 directed towards the optical measurement arm 5 and a second portion of source beam 12 directed towards the reference optical arm 7. The optical beam splitter 4 is advantageously configured to transmit approximately 50% of the modulated source power on each of the two optical arms 5, 7, especially since the beams pass twice through this optical beam splitter 4. Alternatively, to raise the return level of the DUT, the coupler can be unbalanced to transmit for example 10% of the modulated source power towards the reference optical arm 7 and 90% of the modulated source power towards the optical measurement arm 5 but this configuration is not considered optimal.

[0058] The first portion of source beam 11 is transmitted to the optical measuring arm 5 and propagates through the optical device 6 to be tested. Let us consider a reflection point 16 located in the optical device 6 at an optical distance x from the optical reflector 8 of the reference arm 7. For the sake of clarity, a single reflection point 16 is considered here. However, as described below, the apparatus is adapted to detect at the same time several reflection points located in the optical device 6 at different optical distances. Here, reflection point 16 is understood to mean a surface, an interface, an optical defect or any element capable of generating a light reflection. The reflection point 16 is at an optical distance equal to Lx from the optical beam splitter 4.This reflection point 16 receives the first portion of source beam 11 and contributes to forming a first reflected beam 21 propagating towards the optical beam splitter 4, in the opposite direction to the first portion of source beam 11. Thus, the first portion of source beam 11 travels the optical length Lx before reaching the reflection point 16 and the first reflected beam 21 also travels the optical length Lx before reaching the optical beam splitter 4. As indicated previously, the optical length (Lx) is here much less than the optical length L. .

[0059] The second portion of source beam 12 is transmitted through the reference optical arm 7 to the optical reflector 8 and forms a second reflected beam 22, propagating in the opposite direction to the second portion of source beam 12, towards the optical beam splitter 4. As indicated previously, the reference optical arm 7 is free from any parasitic reflection. Thus, the second portion of source beam 12 travels the optical length L before reaching the optical reflector 8 and the second reflected beam 22 also travels the optical length L before reaching the optical beam splitter 4.

[0060] The optical beam splitter 4 receives the first reflected beam 21 by the optical measuring arm 5 having traveled the optical length 2(Lx) and the second reflected beam 22 by the reference optical arm 7 having traveled the optical length 2L. The optical beam splitter 4 is capable of optically / spatially recombining / superimposing the first reflected beam 21 and the second reflected beam 22 to form a reflected beam 23 propagating from the fourth output branch towards the detection system 30.

[0061] [Fig. 3] represents as a function of time, the optical power of the first beam reflected 21 by the optical measuring arm 5 and respectively of the second beam reflected 22 by the reference optical arm 7, for a modulated source beam 10 as illustrated in [Fig. 2]. The two reflected beams 21, 22 are also modulated temporally in the form of pulses at the modulation frequency F. The first beam reflected 21 by the optical measuring arm 5 comprises for example im successive pulses 211, 212, 213 at the modulation period T. Similarly, the second beam reflected 22 by the reference optical arm 7 comprises successive pulses 221, 222, 223 with the same modulation period T. By performing a sweep of the modulation frequency F, the temporal overlap between the pulses 211, 212, 213 reflected by the measuring arm 5 and the pulses 221, 222, 223 reflected by the reference arm 7 is changed. The pulses reflected by the two arms exactly overlap when the modulation period T of the modulated source beam 10 satisfies the following equation, where c represents the speed of light in vacuum.

[0062] [Math.l] 2x - cT

[0063] Pulse 21 emitted at time t overlaps with pulse 22 emitted at time tT.

[0064] However, due to the large difference in optical length between the two arms and / or the return to 0 of the source power when it is modulated directly with zero low power, the reflected pulses which are temporally superimposed are not coherent with each other in terms of optical phase.

[0065] In the case of a partial temporal superposition of the pulses reflected by the two arms illustrated in [Fig. 3], a pulse 211 of the first reflected beam 21 and a pulse 221 of the second reflected beam 22 only partially superimpose in a temporal superposition zone 25 of duration less than the duration of the high level of the modulation. In a portion 24 of each modulation period T, only a portion of the pulse 221 of the second reflected beam 22 is received by the optical beam splitter 4. In another portion 26 of each modulation period T, only a portion of the pulse 211 of the first reflected beam 21 is received by the optical beam splitter 4. In yet another portion 27 of each modulation period T, no portion of the pulse 211 of the first reflected beam 21 nor any portion of the pulse 221 of the second reflected beam 22 is received by the optical beam splitter 4.

[0066] The detection system 30 receives the reflected beam 23 formed by temporal superposition of the first reflected beam 21 and the second reflected beam 22. As indicated above, the superimposed pulses coming from the two reflected beams 21, 22 are not coherent with each other and are generally temporally offset from each other.

[0067] The reflectometry measurement according to the present disclosure is very different from a white light interferometric measurement. Indeed, in white light interferometry, the measurement point of the DUT is made to be at the same distance optical length of the coupler-splitter as the reflection mirror of the reference arm (which would be equivalent to using an optical length of L close to Lx in the diagram of [Fig.l]). The white light interferometer uses for this purpose a mechanical displacement of the reference mirror and the spatial resolution is given by the coherence length of the source. When the two arms of the white light interferometer are of the same optical length, such an interferometer provides a direct interferometry measurement.

[0068] On the contrary, according to the present disclosure, the lengths of the two arms 5, 7 are fixed and very different from each other.

[0069] The coherence length of a single-frequency laser source is represented in a simplified manner as a series of sections of optical length Lco in which the phase is constant, and which exhibit a random variation of this phase from one section to the next. In the reflectometry apparatus and method 100 described here, in the case of a DFB type source, the sections which are superimposed on the return from the two arms have no coherence between them but interfere by giving a stable bias over a maximum time tco, corresponding to the length Lco. In the case of a source of long coherence length, it is the direct modulation of the source, with a low modulation level of zero power, which causes the loss of optical phase coherence between two successive pulses.

[0070] [Fig.4] represents, as a function of time, an example of a power measurement of the signal received by the detection system over a modulation period by modulating the constant source signal by an external modulator and balancing the power sent in the two arms. To provide a visually speaking example, the signal representing the return of the DUT is here much higher than it is generally in real use cases. Several curves have been represented over the same modulation period T, here equal to approximately 16.5 microseconds (ps): a curve with dotted lines indicating the average detected power or mean value, noted 31, a solid line curve indicating the standard deviation, noted 32, of the power measurement. This average is taken over 1 ms, therefore on average over 60 periods, with an acquisition step of 1 ns. The standard deviation is representative of the amplitude of the noise, also called in this document interferometric noise signal 32.During the portion 27 of the modulation period T, the detection system 30 receives no pulse portion of the first reflected beam 21 nor any pulse portion of the second reflected beam 22, so that the power average 31 of the detected signal is zero and the standard deviation 32 of the detected power is zero. During the portion 24 of the modulation period T, the detection system 30 receives a pulse portion of the first reflected beam 21 and no pulse portion of the second reflected beam 22, so that the power average 31 of the detected signal is non-zero. During the . portion 26 of the modulation period T, the detection system 30 receives a pulse portion of the second reflected beam 22 and no pulse portion of the first reflected beam 21, so that the power average 31 of the detected signal is non-zero. However, on the portions 24 and 26 respectively, the standard deviation 32, in other words the noise signal, remains low and characteristic of the intrinsic noise of the source. On the other hand, in the temporal superposition zone 25 of the modulation period T, the detection system 30 receives a pulse portion of the first reflected beam 21 and a pulse portion of the second reflected beam 22, so that the signals are added, the power average 31 of the detected signal is non-zero, and the standard deviation is very high: a non-zero interferometric noise signal 32 is observed.In other words, in the temporal superposition zone 25, the measured signal consists of a series of sections of interferometric signals each having a duration tco and which present a random variation of this interferometric signal from one section to the next.

[0071] The detection system 30 and the demodulation system 40 are adapted and configured to extract a representative measurement of the interferometric noise signal 32 as a function of time, for example based on the standard deviation or an equivalent of the standard deviation at the modulation period.

[0072] In the example illustrated in [Fig.4], an optical modulator external to the light source 1 is used and the wavelength is identical at every point of each power-modulated source pulse. Furthermore, the low power is not here perfectly zero as can be obtained by direct modulation of the source but the difference between the two arms is sufficient to reject any direct interferometry as can be seen approximately in the figure by the fact that the sum of the average powers of each beam given respectively at 24 and 26 is equal to the average power of part 25.

[0073] The reflectometry apparatus 100 is adapted and configured to detect as a function of time, and as a function of a sweep of the modulation frequency, a signal representative of the standard deviation of the power difference detected in particular in the temporal superposition zone 25 between successive modulation periods.

[0074] For this purpose, an electronic demodulation system 40 is used with the detection system 30 and the modulation system 2. The measurement of the interferometric noise signal 32 and the modulation of the source 1 are perfectly synchronized via the electronic synchronization system 3 based either on a common clock or on the detection of the edges of the modulation command to define the acquisition instant(s).

[0075] [Fig.5] represents an example of demodulation based on periodic sampling of the power of the detected signal as a function of time, over several periods modulation, for example here three modulation periods, to extract the interferometric noise signal. The detected signal here has any form. In this example, a measurement point is sampled at each modulation period T, each measurement point being noted respectively i at a modulation period, i-1 at the previous modulation period, and i+1 at the following modulation period. In other words, in this example, the acquisition frequency is equal to the modulation frequency. The calculation of the standard deviation of the demodulation is written as a sum according to the following equation:

[0076] [Math.2] J? — ■" y- (' / j- - / ----

[0077] where Pi i represents the instantaneous power of the signal detected at time i-1, P; the instantaneous power of the signal detected at time i and Pi+1 the instantaneous power of the signal detected at time i+1 and B represents interferometric noise. On a signal without interference, such demodulation gives a constant and very low value B representing the intrinsic noise of the source, which is low.

[0078] In the equation Math2, we obtain the value B. To extract the interferometric noise signal 32, we calculate for example the standard deviation of the interferometric noise, based on the square root of the value of B divided by the number of cumulative values.

[0079] Advantageously, a low-pass electronic filter is used on the signal which preserves the variations on the coherence time tco of the source, which amounts to having a cut-off frequency beyond the spectral width of the source which is less than 10 MHz, without however exceeding this value to limit the noise at higher frequencies. In practice, the length of the pulse being from O.lps to a few ps, it is useful to oversample the duration of the pulse which is generally greater than the coherence time of the light source 1.

[0080] Furthermore, in order to avoid having to precisely define the measurement point in relation to the shape of the pulse, we choose to make the measurement uniformly over the entire modulation period.

[0081] [Fig.6] illustrates another example of demodulation based on periodic oversampling of the signal received at each modulation period, to extract a value B representing the interferometric noise. In [Fig.6], only a part of the measurement points are shown, which are sampled at an acquisition frequency generally equal to a multiple of the modulation frequency, the multiple being equal for example from 10 to 100, which is accessible in current analog-digital conversion components. In this example, several measurement points are sampled at each modulation period T, each measurement point being denoted respectively i, j, k at a modulation period, i-1, j-1, k-1 at the previous modulation period, and i+1, j+1, k+1 to the next modulation period. The consecutive measurement points are spaced temporally by the sampling period noted TE which is generally equal to an integer fraction of the modulation period T. The demodulation calculation is written as a cumulative sum according to the following equation:

[0082] [Math.3] ^ =... + / vi / + + C^ j - / / / + - a /

[0083] where P; represents the instantaneous power of the signal detected at time i, Pj the instantaneous power of the signal detected at time j and Pk the instantaneous power of the signal detected at time k... More compactly, the above equation is written as follows:

[0084] [Math.4] B = - Pfidt - r))2

[0085] where P(idt) represents the power sampled at time idt and P(idt-T) the power sampled at time idt-T, N being the number of accumulation points, clocked by the user sampling frequency (for example 100Hz) considering each time an integer number of modulation periods. In the example given in [Fig.4], the number of periods on which the average and the standard deviation were calculated is 60 for practical reasons but it is generally around 1000 to 10000.

[0086] In the Math3 or Math4 equation, differences are calculated at the modulation period T. Consequently, the value of B makes it possible to find the only source / electronic noise assumed to be constant (which is easily verified) when there is no interference, i.e. outside the temporal superposition zone 25. Another advantage of the oversampling technique is to avoid searching for the optimal point of the measurement over the period T since all the points are measured.

[0087] In the equations Math3 or Math4, we obtain the value B representing the interferometric noise. To extract the interferometric noise signal 32, we calculate for example the standard deviation of the power difference at the period, by dividing B by the number of measurement points and taking the square root of this result.

[0088] The interferometric noise signal 32 makes it possible to obtain good dynamics for measuring the optical reflection at the measured point of the DUT, as for interferometric devices using direct interference since the interference measures the reflection in amplitude and not in power.

[0089] We will now describe how to obtain good spatial resolution in taking advantage of a wavelength modulation of the modulated source beam 10. For example, a DFB laser type light source is used. If the two reflected beams 21, 22 which are added comprise two wave trains superimposed in time while being shifted in wavelength, the signal detected over the duration tco is modulated at a beat frequency Af corresponding to the optical frequency difference of the two contributions.

[0090] Let us consider the two reflected beams 21, 22 having at a time t a different wavelength, that is to say a different optical frequency. The application to the detection system of a low-pass electronic filter whose cut-off frequency is lower than the beat frequency Af makes it possible to eliminate the noise generated by interference. The difference or deviation in optical frequency Af is expressed as a function of the difference or deviation in wavelength AX as follows:

[0091] [Math.5]

[0092] With an optical frequency deviation Af of 10 MHz and an average wavelength of 1550 nm, a wavelength deviation AX of 0.08 pm is deduced from the equation Math5. It can therefore be considered that the interferometric noise signal is not detected for wavelength deviations AX greater than 0.08 pm if electronic filtering is carried out above 10 MHz. For this purpose, the detection system 30 or the demodulation system 40 comprises an electronic low-pass filter. The electronic low-pass filtering thus makes it possible to improve the spatial resolution when the wavelength changes periodically over the duration of the pulse.

[0093] An example of wavelength variation of the power-modulated source pulses is described below. In the example illustrated in connection with FIGS. 7 and 8, a modulated source beam 10 composed of source pulses modulated in power and in wavelength at the modulation frequency is used. [Fig. 7] schematically represents the wavelength of a modulated source beam 10 as a function of time. In this example, the source pulses are power-modulated so as to have a non-zero power during a portion 28 of each modulation period and a zero power during another portion 29 of each modulation period. The sum of the portions 28 and 29 is here equal to the modulation period T. In addition, as illustrated in [Fig. 7], the source pulses are wavelength-modulated at least over the portion 28 of each modulation period where the power is non-zero.Advantageously, the wavelength modulation has a linear form in periodic slots at the power modulation period. In the example of [Fig.7], the wavelength modulation is linearly increasing on each portion 28. of modulation period T. In the portion 29 where the power of the source pulses is zero, which appears in gray on [Fig.7], the shape of the variation in wavelength is arbitrary.

[0094] [Fig. 8] represents as a function of time the wavelength variations respectively of the first beam reflected 21 by the optical measuring arm 5 and of the second beam reflected 22 by the optical reference arm 7, for a beam of source pulses modulated in power and wavelength as described in connection with [Fig. 7]. In the time overlap zone, the pulses reflected by the two arms therefore have a wavelength shift when they do not correspond exactly to the position defined by the modulation frequency in accordance with the Mathl equation. The wavelength variation illustrated in Figures 7 and 8 is linear with a slope equal to dX / dt.If the two pulses of the beams 21, 22 which interfere are separated by an optical distance AL (modulo the modulation period T), the wavelength difference between the first reflected beam 21 and the second reflected beam 22 is constant at any point in the temporal superposition zone, at each modulation period. We obtain the following relationship where c represents the speed of light in a vacuum: .

[0095] [Math.6]

[0096] For example, a DFB laser source emitting around 1550nm with a spectral width of 10 MHz and an electronic low-pass filter, the cut-off frequency of which is also chosen to be around 10 MHz, is used. In this case, the interferometric noise signal tends towards zero if the following relationship is respected:

[0097] [Math.7] 0 ÛSwm -------OÂ

[0098] In an exemplary embodiment, a wavelength variation dX / dt of 100 pm / ps is used for each pulse. By applying the Math7 equation, the resolution limit in optical distance x is approximately 24 cm in a single pass along the optical axis of the DUT. To obtain a resolution / precision on the optical distance x of 1 cm, it is necessary to apply a wavelength variation of the order of 2400 pm / ps. It is assumed here that the wavelength modulation (or chirp in English) continuously changes the phase of the source pulse thus modulated in wavelength, which is in practice the case, in particular for a current-modulated DFB laser source as detailed below. The wavelength variation is clas strictly controlled by the temperature and / or the electric current of the laser source. The chromatic dispersion of the optical fiber of the reference arm has little impact here in the choice of configuration.

[0099] Different examples of such power and wavelength modulation are obtained with a DFB laser source. The DFB laser source is supplied with electric current by a current source.

[0100] According to a first exemplary embodiment, a linear current ramp is applied, for example increasing, over the duration of the portion 28 of each modulation period where the power is non-zero. This gives a power varying linearly at each modulation period and a monotonic variation in the wavelength at each modulation period. For example, a sawtooth modulated electric current is used directly. A current modulation amplitude of the order of 20 mA at a modulation frequency of 3 MHz makes it possible to achieve a variation in the wavelength of approximately 800 pm, or of the order of 2400 pm / ps. Such sawtooth current modulation thus provides the variation necessary to obtain the resolution of 1 cm.

[0101] In [Fig.9] a second example of implementing modulation in power and wavelength is shown. This example is based on the observation that a rectangular square wave modulation of the current supplying the DFB laser source simultaneously produces a square wave modulation of the power of the source pulses reproducing the shape of the square waves of the electric current and a modulation of the wavelength which increases until it reaches an asymptotic value - depending on the length of the pulse - at each source pulse. Indeed, the variation in current instantly induces a variation in the delivered optical power. An external cooling device thermally regulates the temperature of the DFB laser source. However, the response time of the external cooling device is much longer than the variation in temperature of the DFB laser source induced by the variation in current for modulation frequencies exceeding 100 kHz.The wavelength variation comes from temperature variations of the DFB source induced by the current variations and which are faster than the thermal regulation applied externally to the DFB source. The wavelength variation is monotonic and continuous, for example increasing, during the portion 28 of each modulation period where the power is non-zero. In the portion 29 where the power of the source pulses is zero, the wavelength decreases rapidly. By reducing the duration of the portion 28 of each modulation period where the power is non-zero, that is to say by reducing the modulation period T and / or the modulation duty cycle, the wavelength variation tends to be linear or quasi-linear over the portion 28 of each modulation period. A slope of is easily obtained. wavelength variation dX / dt greater than or equal to 100 pm / s by modulating a DFB laser diode with current between 0mA and 100mA of current with a modulation period ranging from 0.5ps to 100ps and a modulation duty cycle ranging from 5% to 50%.

[0102] Note in these two examples of modulation in power and wavelength, that the amplitude of variation in wavelength is sufficiently low to preserve a monotonic and continuous variation on each pulse and periodic. The stability of the periodic response in wavelength is important, but required only over the duration of each noise measurement, between 1 ms and 10 ms.

[0103] We will now describe more precisely how to determine the position x of a reflection point 16 in the optical device 6 to be tested.

[0104] As disclosed above, in particular in connection with [Fig. 4], a signal representative of the interferometric noise is measured, based on the standard deviation of the power of the interferometric signal detected by the detection system 30. In other words, the value of the measured signal is the value of the noise, the spatial resolution in x being achieved by the wavelength shift between the two reflected beams. The aim is to measure the absolute or relative position in x of the reflection point(s) measured with respect to the reflector 8 of the reference arm 7.

[0105] Here we use advances in electronic components that make it possible to have, for example, low-cost programmable PLL (phase-locked loop) components, enabling a clock to be controlled with variations smaller than ppm and an absolute value that is just as precise. The clock makes it possible to vary the modulation frequency of the modulated source beam with very good resolution and very high precision.

[0106] The position deviation between the optical length L of the reference arm and the optical length, Lx, at the measuring point 16 starting from the optical beam splitter coupler 4 is given simply by the Mathl equation, indicated previously. In other words, the deviation x in optical distance is given by the following equation:

[0107] [Math.8]

[0108] As the modulation frequency F is controlled with very high precision by the PLL, the determination of the optical distance x is also obtained with very high precision. The optical length L of the reference optical arm 7 is fixed, and varies only with the temperature. A thermal regulation device of the reference arm, for example based on an optical fiber coil, makes it possible to obtain a measurement of the optical distance x with very high precision. Of course, the distance optical x depends on the optical index of refraction of the DUT 6 material.

[0109] We will now describe the operating range of the reflectometry apparatus, as a function of a sweep of the modulation frequency F. An upper limit of the modulation frequency is set by half the optical length of the reference arm. Indeed, when the modulation frequency corresponds to half the length L of the reference arm 7, the detection system 30 receives the superposition of two pulses in comparison: on the one hand, a pulse reflected from the part starting at the length of the half-reference (between the DUT reflection and the first pulse which follows) and on the other hand, a pulse reflected from the start of the DUT (between the DUT reflection and the following pulse), which is difficult to discriminate. The upper limit, denoted Fmax, of the modulation frequency is therefore equal to twice the characteristic frequency of the length of the reference, that is to say Fmax=2c / 2L=c / L.We therefore consider a modulation frequency range less than or equal to Fmax.

[0110] In summary, the optical interferometric noise signal measurement reflectometry apparatus operates under the following conditions. The modulation of the source pulse beam is configured to produce a linear or quasi-linear variation of the wavelength in each pulse, naturally or by current control, preferably by limiting the power variations in each source pulse. The scanning of the measurement range is done by changing the frequency of the modulation of the source pulse beam. The sampling is preferably carried out over the entire modulation period. Advantageously, the sampling is preferably carried out with a sampling period TE close to the coherence length of the source and equal to an integer multiple of the modulation period. The demodulation of the received power is representative of the standard deviation of the power at the modulation period.The detection system 30 includes an electronic low-pass filter having a cutoff frequency greater than or equal to the spectral width of the source, preferably close to the spectral width of the source, for example of the order of 10 MHz.

[0111] Figures 10 to 16 allow the modulation / demodulation method to be described in a more visual manner. [Fig. 10] represents, as a function of time, the power, denoted Ptest, of the first beam reflected 21 by the optical measurement arm 5 and respectively the power, denoted Pr, of the second beam reflected 22 by the reference optical arm 7, for a modulated source beam 10 as illustrated in [Fig. 9]. The two reflected beams 21, 22 are also temporally modulated in the form of pulses at the modulation frequency F. During a reflection at a reflection point of the DUT 6, a reflected pulse is added to the reflected pulse from the reference arm. For clarity of the presentation, a single reflection point is considered here in the measuring arm 5. The successive pulses are numbered N=0, 1, 2. Depending on the modulation frequency, one of the three cases illustrated in Figures 10, 11 and 12 occurs: the optical path difference between the reflection point of the DUT and the reference arm is so large that there is no temporal overlap between the pulses of the two arms ([Fig. 11]); the optical path difference between the reflection point of the measuring arm and the reference arm is exactly equal to the half modulation period (equation Math 1) and interferometric noise is observed in the bandwidth of the detection system ([Fig. 12]); or the optical path difference between the measuring arm and the reference arm is slightly different from the half modulation period and interferometric noise is observed outside the bandwidth of the detection system ([Fig. 10]).

[0112] In [Fig. 11], the case where there is no temporal overlap between the pulses of the two arms is shown schematically. In this case, we see that the power difference between two successive periods is zero, because the periodicity of the successive pulses remains perfect. The signal representing the interferometric noise based on a standard deviation of the power difference received between two successive periods is then zero.

[0113] In [Fig. 12], the case where there is perfect temporal overlap between the pulses of the two arms is shown schematically. In this case, as illustrated in [Fig. 13], a randomly varying detected power level is observed. The variation in the power level is measured and is expressed as follows: 4^'

[0114] where Pr represents the power of the second beam reflected 22 by the reference optical arm 7, and respectively Ptest represents the power of the first beam reflected 21 by the return of the measuring optical arm 5. This variation in the power level is added to the sum of the two powers according to the classical rules of interferometry. However, since there is no coherence between two successive pulses, the phase shift between the pulses differs between two successive modulation periods and the signal appears in time rather as in [Fig. 14].

[0115] In [Fig. 14], the mean value of the detected power, denoted Pm, averaged over several portions 28 of successive modulation periods, is shown in dashed lines, and, in solid lines, the mean power calculated over each portion 28 of each modulation period assuming that the time tco is of the same order of magnitude as the pulse time in this case, with the same time overlap between the pulses of the two arms as illustrated in Figures 12-13. It is observed that the difference in mean power between two successive modulation periods produces here a detectable interferometric noise signal. The measurement of this interference noise signal- metric allows to determine an amplitude which is proportional to the value of the amplitude reflection in the optical device 6 to be tested. When the modulation frequency satisfies the Math8 equation, the method described here allows to measure the reflection of the measuring branch 5 on the so-called synchronous reflection point, with respect to the modulation frequency. In practice, the power levels present a series of variations whose length is the effective coherence length and which voluntarily corresponds to the detection bandwidth.

[0116] As indicated above, the measurement is carried out by sweeping the modulation frequency. [Fig. 15] represents the power received on the detector in the case where the time shift between the pulses of the two arms is not synchronous with the modulation frequency (as illustrated in Figures 3 and 10).Different time zones or portions of each modulation period T are distinguished: portion 24, where only a part of the pulse 221 of the second reflected beam 22 is received on the detector; the time superposition zone 25 where a part of a pulse 211 of the first reflected beam 21 and a part of a pulse 221 of the second reflected beam 22 are superimposed and are received on the detector; portion 26, where only a part of the pulse 211 of the first reflected beam 21 is received on the detector; and portion 27, no part of the pulse 211 of the first reflected beam 21 nor any part of the pulse 221 of the second reflected beam 22 is received on the detector. In the temporal superposition zone 25, due to the wavelength modulation (see description in connection with [Fig.8]), the wavelengths between the pulses 211 and 221 are sufficiently far apart to generate a very high cadence beat.

[0117] [Fig. 16] represents the representative power curve of the inter- noise signal ferometric in the configuration of [Fig. 15], after electronic filtering of the detected power. The electronic filter here is an electronic low-pass filter, having a cut-off frequency above the modulation frequency, the cut-off frequency being greater than or equal to the spectral width of the source, preferably close to the spectral width of the source.

[0118] In [Fig. 16], we observe that the low-pass electronic filter has removed the high-frequency beats. We observe a periodic pattern and the difference at the modulation period T is zero. In this configuration of time shift between the two received pulses, the optical frequency difference being greater than the bandwidth of the electronic system, the interferometric noise signal is then zero.

[0119] The apparatus and method of the present disclosure thus make it possible to selectively detect a reflection point in the measuring arm 5, by measuring a non-zero interferometric noise signal at a specific modulation frequency. The apparatus of reflectometry by optical interferometric noise signal measurement has several advantages, including very low cost, being relatively compact and without moving parts for very interesting performances in terms of measurement range, longitudinal spatial resolution, position accuracy of the reflection point(s) in the DUT and measurement dynamics without blind zone in the optical device 6 to be tested. Indeed, the adjustment of the modulation frequency makes it possible to detect a reflection point at any location in the DUT.

[0120] Let us now consider the case where the optical device 6 to be tested in the measuring arm 5 has several reflection points at different distances x from the optical reflector 8 of the reference arm 7. Since the pulses reflected by the different reflection points are spaced apart in time by more than the effective coherence length of the modulated DFB source, their contributions add linearly with the following rule: only the synchronous part (between the measuring arm and the reference arm) of the different reflected beams is detected, thanks to the detection system and the electronic low-pass filtering system.

[0121] The physical phenomena involved in the present disclosure are described below by a simplified representation. P0(t) denotes the optical power of the source beam emitted by the light source 1. The optical beam splitter 4 separates the modulated source beam by dividing, for example, by two the power injected on each arm 5, 7. On the return, the optical beam splitter 4 superimposes the first beam reflected 21 by the optical measurement arm 5 and the second beam reflected 22 by the optical reference arm 7 to form a reflected beam 23 propagating towards the detection system 30. It is assumed here that the polarization of the beams is preserved, which allows a scalar addition, and that the losses in the fibers are negligible. The power detected by the detection system is written as follows:

[0122] [Math.9] P = 0.25^®^' T = û.25Pq(1. 4- J? + 2r case(Atù i + $>))

[0123] Where cp is a random phase deviation in time, co represents the pulsation (2p times the optical frequency) of the source beam, r represents a reflection coefficient in amplitude on a reflection point 16 in the DUT 6 and R its value in power (i.e. R=r2), Aco the difference in pulsation due to the wavelength modulation at each modulation period described above in connection with [Fig.8]. We simplify by assuming the wavelength deviation AX to be linear and therefore Aco constant.

[0124] The first step is to filter with an electronic low-pass filter. By choosing a high-order filter, we can consider that the average value of the cross term (cos) is zero above the cutoff frequency and does not affect the signal below the cutoff frequency. When the two signals are sufficiently close (modulo the modulation period T), we obtain a variable power on each period and a difference at the period, that is to say a power difference AP between two successive periods, which is written as follows, where q> and q>i represent the phase shifts on the two successive periods:

[0125] [Math. 10] AP = 0.5rP8(cos(â&> (t 4- T) — œs(A&* t +• ^))

[0126] Since the two phases q> and q>i are random with respect to each other and variable at each period, the term function of time (which hardly varies over the time of the pulse in this case) no longer has any consequence. We can therefore keep the generic expression with two random phases:

[0127] [Math. 11] AP = 0,SrPa(ços

[0128] It follows that the amplitude reflection coefficient, r, at the reflection point 16 in the DUT 6 is proportional to the standard deviation of AP measured at the modulation period T. The method also makes it possible to carry out a calibration of the apparatus using a reference optical device instead of the DUT.

[0129] A processing system 50 receives from the electronic power demodulation system 40 the interferometric noise signal 32 detected by the power deviation measurement at the modulation period, as a function of the modulation frequency sweep. The processing system 50 comprises a computer adapted to process the signal thus obtained as a function of the modulation frequency and to extract an optical distance deviation x between a reflection point 16 in the optical device 6 to be measured and the optical reflector 8 of the reference optical arm 7. The modulation frequency sweep is translated into a distance sweep x by applying the formula Math8.

[0130] Figures 17 and 18 illustrate an example of reflectometry measurement results by optical interferometric noise signal measurement as a function of the optical distance x between the entry point in the DUT and the measurement point in this same arm, which corresponds approximately to the “Lx” of [Fig.l]. For these measurements, a device 100 is used as illustrated in [Fig.l]. The optical fiber 9 on the reference optical arm 7 is wound over a length of 150 m (220 m in optical length) which corresponds approximately to the length L of [Fig.l]. The light source 1 is a DFB laser source emitting around 1550 nm and providing a power of approximately 1 mW, which is current modulated via the application of current slots of 60 mA amplitude. The source beam has a wavelength of 1550 nm. Frequency modulation is done over a range from 650 kHz to 800 kHz with a step sweep variation between 1Hz and 100Hz, which corresponds to an x-ray resolution in steps of less than 1mm to a few cm. The modulation duty cycle is around 25% and the wavelength variation is a few tens of pm over each modulation period.

[0131] First, a measurement is made (dashed curve in [Fig. 17]) of the parasitic return with, in the measuring arm, for DUT a section of optical fiber which ends straight cleaved (with a reflection coefficient R of 4%). Then a measurement is made (solid curve in Figures 17 and 18) with, in the measuring arm, a much more complex optical device 6, with different interfaces distributed along the longitudinal propagation axis, that is to say at different optical distances x from the optical reflector 8 of the reference arm 7. The distance is in optical length. It is noted in [Fig. 17] that the measurement dynamics here is 60 dB with a resolution in optical distance x of the order of 2 cm.

[0132] The dashed curve in [Fig. 17] shows a very intense reflection on the cleaved end of the optical fiber section and makes it possible to determine the optical distance between the cleaved interface of the optical fiber section and the optical reflector, which is of the order of 6 m.

[0133] The solid line curve in [Fig. 17] shows a series of low-intensity reflections on the different interfaces of the complex optical device 6, each interface forming a reflection point. The main reflection points are indicated in [Fig. 17] by arrows at the respective optical distances of approximately: 27 m, 32 m, 35 m and 37 m.

[0134] [Fig. 18] shows a zoom on the solid line curve of [Fig. 17], around the reflection point at the optical distance of about 37 m. Here we observe the good spatial resolution in optical distance.

[0135] Different variants of the reflectometry apparatus by optical interferometric noise signal measurement are envisaged here. Naturally, it is important to ensure that the reference arm does not have any parasitic reflection point outside the reflection of the mirror 8. According to a first variant, the optical fiber coil 9 is moved from the reference arm 7 to the measuring arm 5 where it is placed in series with the DUT 6 (the coil being closer to the splitter 4 than the DUT), the optical reflector being arranged at the output of the optical beam splitter 4. According to a second variant, the DUT 6, the optical fiber coil 9 and the optical reflector 8 are arranged in series on the measuring arm 5.

[0136] Unlike the white light interferometer, the return of the reference arm is time-shifted to avoid direct interference. This shift is achieved by using a very long reference arm. Similarly, the reference arm is fixed, thus greatly reducing the cost of the device, and no mechanical element is used in the proposal to determine the position parasitic reflection points in the DUT.

[0137] The periodicity of the signal and the use of high modulation frequencies make it possible to consider that the properties of this signal are identical between two successive periods, with the important exception of the optical phase, the value of which changes randomly in response to the decoherence obtained by the time difference between each pulse. The present disclosure uses the signal resulting from the superposition of two successive periods which will cross on return, the first coming from the reference and the second coming from the DUT, due to the difference in propagation length between the arm of the DUT and that of the reference.The phase decorrelation between the reference signal and the return signal of the DUT no longer allows stable interference to be obtained, but the coherence length of the source is sufficient for the superposition of the two beams to cause the power to vary randomly, but in a measurable manner, when the time difference reaches the modulation period of the signal. We thus observe interference noise whose amplitude characteristics make it possible to obtain, like direct interferometry used for example in white light interferometry, a very high measurement dynamic range.

[0138] As with other existing techniques, the aim here is to carry out distributed reflection measurements along the DUT and the resolution is a major point in the interest of the proposed device. This resolution is greatly increased by the addition of a variation in the optical wavelength of the modulated source signal, this variation being periodic and synchronized with the power variation. Unlike OFDR which uses wavelength variations (over a range of nearly 100nm) for the purpose of measuring the beat between the reference and the reflection points of the DUT, the invention proposed here uses wavelength variations over a very reduced range to increase selectivity and reject contributions close to the measurement point. This allows for selectivity of the order of cm.The low wavelength variation range required in the proposed device allows the use of standard DFB type laser sources with very small footprint and low cost, which is a significant advantage over OFDR which generally requires the use of tunable laser cavities which are very expensive. The proposal described here therefore makes it possible to avoid any mechanical moving element also at the source level.

[0139] After dynamics and resolution, the third important element in the realization of a device for measuring distributed optical returns is the determination of the position of these returns. The proposal described here does not call for the temporal detection of the returns as for the OTDR, nor for the determination of the beat frequency between the return signal of the DUT and that of the reference as in the case of the OFDR. As for white light interferometry, we does not determine the position by analyzing the measurements but the measurement is made on the DUT at a chosen position, which gives greater measurement flexibility. Indeed, the position of the reference mechanical arm in the white light interferometer gives the measurement on the equivalent point of the DUT. In the case of the proposed invention, it is the choice of the modulation period of the source which will provide the identification of the measurement point on the DUT and it is the quality of the clock used by the electronic card which makes it possible to precisely determine this absolute position. The evolution in quality, cost and control possibilities of electronic clocks allows absolute precision to the ppm in the definition of the modulation frequency which gives millimetric positioning of the reflections of the DUT while maintaining a very large modulation frequency range which results in a measurement range which can go well beyond 100m.

[0140] The apparatus and method of the present disclosure therefore make it possible to achieve significant dynamics, a resolution around the cm with a measurement range well above 100m and positioning to the mm, while remaining very low cost with the notable absence of any mechanical part, whether at the source or the reference arm.

[0141] The use of the modulation frequency to choose the measurement point only differentiates the different measurement zones in the DUT by the choice of the frequency on a variation of a maximum factor of 2 of the value of this frequency. The measurement is therefore very homogeneous in its realization for all the points of the DUT and there is no notion of blind zone of the measurement. In addition, as for the white light interferometer, but without the weight of the mechanics, it is possible to carry out focused measurements on a reduced range of the DUT thus allowing a potentially very rapid temporal monitoring of the evolution of the reflection in the zone concerned.

[0142] Of course, various other modifications may be made to the invention within the scope of the appended claims.

Claims

Claims

1. Reflectometry apparatus (100) by measuring optical interferometric noise signal, comprising: - a light source (1) capable of emitting a monochromatic source beam having a spectral width of less than 20 MHz; - a modulation system (2) adapted to temporally and periodically modulate the source beam and form a modulated source beam (10) composed of source pulses modulated in power and wavelength at a given modulation frequency defining a modulation period, the modulation system (2) being capable of varying the modulation frequency; - an optical beam splitter (4) arranged and configured to spatially separate the modulated source beam (10) between a first source beam portion (11) directed towards an optical measuring arm (5) comprising an optical device (6) to be measured in reflection and a second source beam portion (12) directed towards a reference optical arm (7), the reference optical arm (7) having a determined optical length L, L being greater than or equal to 20 meters and the optical device (6) having an optical length less than L / 2, the reference optical arm (7) comprising an optical reflector (8) arranged at a distal end of the reference optical arm (7) relative to the optical beam splitter, a difference in optical length between the optical measuring arm (5) and the reference optical arm (7) being greater than L / 2; - the optical beam splitter (4) being capable of optically recombining a first reflected beam (21) formed by reflection of the first portion of source beam (11) at at least one reflection point (16) of the optical measuring arm (5) and a second reflected beam (22) formed by reflection of the second portion of source beam (12) on the optical reflector (8) of the reference optical arm (7) so as to form a reflected beam (23), the optical beam splitter (4) being capable of transmitting the reflected beam (23) to a detection system (30), - the detection system (30) being adapted to acquire, as a function of time and as a function of a sweep of the modulation frequency over several modulation periods, a power of an interferometric signal of the reflected beam (23) in at least one super- temporal position (25) between a pulse of the first reflected beam (21) and a pulse of the second reflected beam (22), - an electronic system (40) for demodulating the power of the acquired interferometric signal, the electronic system (40) being adapted to extract an optical interferometric noise signal (32) based on a standard deviation of a power difference at the modulation period as a function of the modulation frequency, and - a processing system (50) comprising a computer adapted to process the optical interferometric noise signal (32) as a function of the modulation frequency and to extract a reflectometry measurement as a function of an optical distance (x) between at least one reflection point (16) in the optical device (6) to be measured and the optical reflector (8) of the reference optical arm (7).

2. Apparatus according to claim 1 wherein the calculator is adapted to extract from the reflectometry measurement a reflection amplitude of at least one reflection point in the optical device (6) to be measured.

3. Apparatus according to claim 1 or 2 wherein the light source (1) comprises a distributed feedback laser diode.

4. Apparatus according to one of claims 1 to 3 wherein the modulation system (2) comprises means for modulating an electrical current supply to the light source (1).

5. Apparatus according to one of claims 1 to 4 wherein the modulation system (2) is adapted to perform a sweep of the modulation frequency with a frequency resolution of at least Ippm.

6. Apparatus according to claim 5 wherein the modulation frequency varies from 650 kHz to 800 kHz with a variation step between 1 Hz and 100 Hz.

7. Apparatus according to one of claims 1 to 6 in which the modulation period has a duty cycle of between 5% and 90%, for example 25%.

8. Apparatus according to one of claims 1 to 7 in which the modulated source pulses have a constant or temporally variable power in a monotonic, exponential, linear or quasi-linear manner, over each modulation period and / or in which the modulated source pulses have a wavelength that is temporally variable in a continuous and monotonic, exponential, linear or quasi-linear manner, over each modulation period.

9. Apparatus according to one of claims 1 to 8 in which the detection system (30) is adapted to sample the power of the interferometric signal at a sampling frequency higher than the modulation frequency, so as to obtain a number N of measurement points of the interferometric noise signal at each modulation period over a number M of successive modulation periods, with N ranging from 1 to 100 and M greater than 1000.

10. Apparatus according to one of claims 1 to 9 wherein the detection system (30) comprises an electronic low-pass filter having a cut-off frequency greater than or equal to the spectral width of the source.

11. Apparatus according to one of claims 1 to 10 in which the reference optical arm (7) comprises a coil of optical fiber (9).

12. Method of reflectometry by measuring optical interferometric noise signal, comprising the following steps: - emission by a light source (1) of a monochromatic source beam having a spectral width less than 20 MHz; - periodic temporal modulation of the source beam to generate a modulated source beam (10) composed of source pulses modulated in power and wavelength at a given modulation frequency defining a modulation period, the modulation frequency being variable;- optical separation of the modulated source beam (10) between a first portion of source beam (11) directed towards an optical measuring arm (5) comprising an optical device (6) to be measured in reflection and a second portion of source beam (12) directed towards a reference optical arm (7) of determined optical length L, L being greater than or equal to 20 meters, the reference optical arm (7) comprising an optical reflector (8) arranged at a distal end of the reference optical arm (7), the optical device (6) having an optical length less than L / 2; - optical recombination of a first beam reflected (21) by the optical measuring arm (5) and a second beam reflected (22) by the reference optical arm (7), so as to form a reflected beam (23);- detection, as a function of time and as a function of a sweep of the modulation frequency over several modulation periods, of a power of an interferometric signal of the reflected beam (23) in at least one temporal superposition zone (25) between a pulse; of the first reflected beam (21) and a pulse of the second reflected beam (22), - demodulating the acquired interferometric signal power to extract an optical interferometric noise signal (32) based on a standard deviation of a power difference at the modulation period as a function of the modulation frequency, - processing the optical interferometric noise signal (32) as a function of the modulation frequency to extract a reflectometry measurement as a function of an optical distance (x) between at least one reflection point (16) in the optical device (6) to be measured and the optical reflector (8) of the reference optical arm (7).

Images