WO2012140411A1

WO2012140411A1 - Optical sensor

Info

Publication number: WO2012140411A1
Application number: PCT/GB2012/050765
Authority: WO
Inventors: Arnold Peter Roscoe Harpin
Original assignee: Oxsensis Ltd
Priority date: 2011-04-12
Filing date: 2012-04-04
Publication date: 2012-10-18
Also published as: GB201106204D0; GB201319933D0; GB2505114A

Abstract

There is disclosed an optical sensor using a sensing interferometer, and a processing interferometer coupled to and matchable to the sensing interferometer. The processing interferometer is formed using waveguides integrated onto a substrate, and has first and second arms. A path length adjuster is formed using a first depletion modulator in one arm, and a second depletion modulator in the other arm, such that modulation of light is dominated by p-type carriers in the first depletion modulated by n-type carriers depletion in the second depletion modulator.

Description

OPTICAL SENSOR

The present invention relates to an optical sensor, a sensing

interferometer and a processing interferometer. Such a sensor may be described as using "white light interferometry" to interrogate the sensing interferometer, and may use a broadband optical source. In particular, but not exclusively, the invention relates to such an optical sensor in which at least the processing interferometer is implemented using waveguides integrated onto a substrate, such as a silicon substrate, for example as a Mach Zehnder interferometer.

Introduction

W098/22775 and WO99/60341 describe the use of white light

interferometry in which a sensing interferometer is interrogated using a processing interferometer and a broadband light source such as a

superluminescent diode. Such an arrangement is illustrated in figure 1 . A superluminescent diode 10, a photodiode detector 12 and a Mach Zehnder processing interferometer 14 are integrated onto a single silicon substrate 16, with optical connections between the components being made by ridge waveguides or similar. The superluminescent diode 10 delivers broadband light via a coupling optical fibre 18 to a Fabry Perot sensing interferometer 20 having an optical path difference OPD_s. The modified broadband light reflected back from the sensing interferometer 20 is delivered along the coupling optical fibre to the substrate 16 where an optical coupler y-junction 22 directs the light through the processing interferometer 14 to the photodiode detector 12.

In the arrangement of figure 1 , the processing interferometer 14 is shown as a Mach Zehnder interferometer having two arms. A phase modulator 24 located in each arm is used to adjust the optical path difference OPD_p between the two arms. When this optical path difference OPD_p matches the optical path difference OPD_s of the sensing interferometer, the intensity detected by the photodiode detector reaches a maximum. As the optical path differences diverge from each other, the intensity at the detector varies as an approximately sinusoidal pattern of fringes 26 under a Gaussian-like fringe envelope 28, as illustrated in figure 2. Ideally, both the peak of the fringe envelope and the peak of the corresponding central fringe occur when OPD_p=OPD_s.

The spacing between fringes in figure 2 corresponds to the central wavelength of the broadband light source, which typically might be around 1550 nm. A broader band light source leads to increased definition of the central fringe, and more rapid decay of the fringe envelope away from this point, but in practice the bandwidth of the light source is limited by the availability of suitable device types. Superluminescent diodes typically have a bandwidth of a few tens of nanometers. A requirement of the white light interferometry technique is that the coherence length of the light source must be significantly smaller than the optical path difference OPD_s of the sensing interferometer, so that when using a superluminescent diode broadband source a sensing interferometer with an OPD_s of a few tens to a few hundreds of micrometers is practical. The high accuracy of available techniques in fabricating the processing interferometer 14 on a silicon substrate facilitates accurate matching of the two interferometers.

It should be noted that, in the arrangement of figure 1 , the processing interferometer 14 could instead be placed between the broadband light source 10 and the sensing interferometer 20, or using a reflective processing interferometer, on a fourth branch from the coupler 22, with the same effect. Other types of processing interferometer can be used, such as a Michelson interferometer.

Instead of using a single detector device at the combined output of the two arms of the processing interferometer, separate detector devices may be used, with subsequent processing electronics combining the resulting signals, for example in quadrature. The sensing interferometer may conveniently be implemented using a Fabry Perot sensor, for example as set out in detail in WO99/60341 , but a Bragg grating or other interferometer type could also be used. For some types of sensing interferometer, a separate optical fibre coupling light transmitted through the sensing interferometer and on to the detector could be used instead of requiring the y-coupler junction 22. Although a photodiode detector 12 is shown, other types of detector could be used.

WO98/22775 describes how an arrangement such as that of figure 1 may be operated by locking the phase difference OPD_s - OPD_p to a fixed value, such as zero. If OPD_p is adjusted to keep the phase difference at zero using a control feedback loop 25 as illustrated in figure 1 , then the current value of OPD_p provides the current value of OPD_s, which represents a parameter to be measured by the sensing interferometer such as temperature or pressure. The measurement range then depends on the range of OPD_p which can be provided by the phase modulators 24.

The invention addresses problems and limitations of the related prior art.

Summary of the Invention

It has been found by the inventors that the range of OPD_p which can be scanned using a processing interferometer integrated onto a substrate, such as a semiconductor substrate, is often limited by degradation of the fringe visibility at larger phase offsets, and that one way in which this can be addressed is by better balancing the optical loss in the two arms of the processing interferometer.

This can be achieved using an integrated semiconductor interferometer arranged to provide a controlled optical path difference comprising: first and second waveguide arms; a p-type depletion modulator disposed on the first waveguide arm; and an n-type depletion modulator disposed on the second waveguide arm. This scheme is effective because the same change in density of p-type carriers can be arranged to be some four to five times more effective at changing the refractive index and hence the modulation than n-type carriers, while giving rise to about the same level of optical loss in both waveguide arms.

In particular, the n-type depletion modulator may be arranged to balance between the waveguide arms at least some of the optical loss caused by the p- type depletion modulator in the first arm. The optical loss in the arms of the interferometer may therefore be dominated by p-type carriers in the p-type depletion modulator, and by n-type carriers in the n-type depletion modulator. For example, the optical loss in the p-type and n-type depletion modulators may be balanced to within 20%, or more preferably to within 10%, during operation of the device. Each depletion modulator may be formed, for example, using a one-sided p-n junction, with the optical modes of the respective waveguide arm intersecting the depletion region of each respective p-n junction.

The interferometer may be a Mach Zehnder interferometer, although other interferometer types may be used. The interferometer may be implemented, for example, on a silicon substrate, or on a silicon-on-insulator substrate, or using other semiconductor materials such as indium phosphide (InP) or gallium arsenide (GaAs) .

As well as use as a processing interferometer in an optical sensor, such an interferometer may be used, for example, to implement an optical switch or an optical multiplexer.

When this interferometer is implemented as part of an optical sensor, the invention provides an optical sensor comprising: a sensing interferometer; an optical source coupled to deliver light to the sensing interferometer; a

photodetector coupled to receive the light from the sensing interferometer; and a processing interferometer coupled and matchable, in the sense of a matched filter or a matched optical path difference, to the sensing interferometer, the processing interferometer comprising waveguides integrated onto a substrate, wherein the processing interferometer comprises first and second waveguide interferometer arms and at least one path length adjuster to control a path length difference between the arms, each path length adjuster comprising a first depletion modulator in one of the interferometer arms and a second depletion modulator in the other of the interferometer arms, such that modulation is dominated by changes in density or depletion of p-type carriers in the first (p- type) depletion modulator, and is dominated by changes in density or depletion of n-type carriers in the second (n-type) depletion modulator.

In such a sensor, the signal detected by the photodetector as a function of path length mismatch between the sensing interferometer and the processing interferometer may comprise a series of fringes. The sensor may comprise two complementary path length adjusters with p-type and n-type depletion modulators in opposite arms. The optical sensor may then be arranged to repeatedly drive a first of these path length adjusters across at least a portion of the fringes using one or more predetermined dither signals, and to drive a second of these path length adjusters based on the resulting signal received at the photodetector, suitably processed such that the second path length adjuster causes the processing interferometer to move towards or track a feature in the fringes, such as a peak of a fringe envelope, or a peak of an individual fringe.

More generally, the invention provides an optical sensor arranged to scan across multiple fringes present in the detector signal as the mismatch between the path length differences of the sensing and processing interferometers is varied. The general form or envelope within which multiple fringes are contained may be used within this process. Typically, this envelope of multiple fringes is single peaked, and the envelope peak is sought or tracked to provide a start position for tracking a single fringe using the second mode, or to provide an output more directly without tracking of a single fringe.

In this light, the invention also provides an optical sensor comprising: a sensing interferometer; an optical source coupled to the sensing interferometer; a photodetector coupled to the sensing interferometer; a processing interferometer integrated onto a substrate, such as a silicon or other semiconductor substrate for example InP or GaAs, and matchable, in the sense of a matched filter or matched optical path differences, to the sensing interferometer, the processing interferometer being arranged such that the signal detected by the photodetector as a function of optical path difference mismatch between the sensing

interferometer and the processing interferometer comprises a series of fringes within an envelope; and a controller arranged to control the processing

interferometer to scan across a plurality of said fringes. A parameter to be detected can then be derived from photodetector signal and properties of the scanning process, for example by detecting a fringe or envelope peak or other feature, the position of the feature corresponding to a said parameter to be detected at the sensing interferometer.

The controller may be arranged to operate in first and second control modes, in which the processing interferometer is scanned across a smaller range, for example less than the width of one fringe in the first control mode, and is scanned across a plurality of said fringes in the second control mode. The optical sensor may be arranged to control the processing

interferometer responsive to a form, shape or other aspect of a fringe in the first mode, and responsive to a form, shape or other aspect of multiple fringes, and in particular an envelope of said fringes in the second mode. Conveniently, the optical sensor may be arranged to seek and/or track a peak in a said fringe or envelope.

The controller may be a feedback controller arranged to combine a dither signal configured for scanning the processing interferometer across, for example, a portion of one fringe in the first mode and across a plurality of said fringes in the second mode, with a signal received from the photodetector, to process the combined signal, and to use the processed combined signal to provide a servo signal arranged to drive the processing interferometer responsive to said fringe in the first mode and responsive to the envelope in the second mode. The servo signal may then be used to provide a sensor output signal. This output signal may be suitably processed to represent one or more physical parameters, such as temperature, pressure and vibration at the sensing interferometer.

The signal received from the photodetector is rectified in the second mode. When smoothed or low pass filtered this has the effect of providing a shape of the envelope instead of individual fringes, enabling the sensor system to track towards a peak in the envelope. The controller may be arranged to adjust the processing interferometer so that a peak of an individual fringe is sought and/or tracked in the first mode, and so that a peak of the envelope is sought and/or tracked in the second mode. The controller may be arranged to scan the processing interferometer across less than one fringe in the first mode, and over a plurality of fringes in the second mode.

The processing interferometer may be provided in various ways, in particular using first and second arms implemented using waveguides integrated onto the substrate, the arms being provided with at least one path length adjuster arranged for control of the optical path difference of the processing interferometer by the controller. The at least said one path length adjuster may implemented as one or more PIN diodes, with the intrinsic region of the one or more PIN diodes intersecting said waveguide for injection of charge carriers into the waveguide, or using a depletion modulator arrangement as discussed above, or a combination of these and/or other modulator types, for example including one or more thermal modulators.

The optical sensor may be arranged to interleave periods of operation according to the first and second modes, for example applying a dither signal to scan over at least a portion of the fringe envelope a number of times, periodically between periods during which the first mode is used, to check that the correct fringe is being tracked. For example, the individual fringe to be tracked, and for example the peak of that fringe may coincide with the peak of the envelope. The sensor may then use the second mode to find the approximate position of the individual fringe to be tracked, and then use the first mode to track the fringe itself.

The optical sensor may be arranged to operate according to the second mode following turning on of the sensor or following detection of an error condition.

If the sensing interferometer described herein provides a plurality of sensing cavities, the controller may be arranged to use the second mode to lock onto different determined ones of said sensing cavities from time to time, following which the first mode may be used to track an individual fringe of the selected sensing cavity. More specifically, the invention may provide an optical sensor comprising: a sensing interferometer having a plurality of optical cavities, such as Fabry Perot cavities, of different lengths; an optical source coupled to the sensing interferometer; a photodetector coupled to the sensing interferometer; a processing interferometer matchable to each of the optical cavities of the sensing interferometer, the processing interferometer being arranged such that the signal detected by the photodetector as a function of optical path difference mismatch between each optical cavity of the sensing interferometer and the processing interferometer comprises a different series of fringes; and a controller arranged to control the processing interferometer so as to detect properties of the series of fringes corresponding to any of the selected cavities. The sensor may thereby be arranged to interrogate any of the optical cavities of the sensing interferometer, to thereby detect multiple physical parameters of the sensing interferometer such as temperature, pressure, vibration and combinations of such parameters, for example subject to different timescales of variation. Interrogation of the optical cavities may use any of the arrangements and techniques described herein.

An optical sensor as discussed herein may be arranged to detect a physical parameter at the sensing interferometer by analysis of broadband light received at the detector. Physical parameters which may be detected can include one or more of temperature, pressure, and acoustic vibration, depending on the design of the sensing interferometer and other aspects of the sensor system. Multiple different physical parameters, and functional combinations of such parameters for example for subsequent deconvolution, may be detected by such an optical sensor, especially using a sensing interferometer with multiple optical cavities.

The sensing interferometer may be, for example, a Fabry Perot interferometer, for example a Fabry Perot interferometer constructed from sapphire for use at high temperatures.

The invention also provides methods corresponding to the above, for example a method of operating an optical sensor which comprises a sensing interferometer, an optical source coupled to deliver light to the sensing

interferometer, a photodetector coupled to receive the light from the sensing interferometer, and a processing interferometer coupled and matchable to the sensing interferometer, the processing interferometer being formed using waveguides integrated onto a substrate and comprising p-type and n-type depletion modulators in respective waveguide arms of the interferometer, the method comprising: driving the p-type and n-type modulators to provide a variable optical path difference of the processing interferometer, wherein optical loss in the p-type depletion modulator is at least partially balanced by optical loss in the n-type depletion modulator.

In this and similar methods the optical sensor may be arranged such that the signal detected by the photodetector as a function of optical path difference mismatch between the sensing interferometer and the processing interferometer comprises a series of fringes, and may comprise driving the p-type and n-type modulators to scan the optical path difference mismatch over a plurality of said fringes.

In another aspect, the invention provides a method of operating an optical sensor comprising a sensing interferometer, an optical source coupled to the sensing interferometer, a photodetector coupled to the sensing interferometer, a processing interferometer integrated onto a substrate and coupled and matchable to the sensing interferometer, the processing interferometer being arranged such that the signal detected by the photodetector as a function of optical path difference mismatch between the sensing interferometer and the processing interferometer comprises a series of fringes within an envelope, the method comprising:

controlling the processing interferometer in first and second modes, in which the processing interferometer is controlled to repeatedly scan across a portion of a said fringe in the first mode, and to scan across a plurality of said fringes in said second mode.

The processing interferometer may be controlled using a feedback loop based on a signal at the photodetector.

The light source or optical source mentioned above may be, for example a broadband light source, such as a superluminescent diode. However, other arrangements such as a scanned narrow band source, for example a tunable laser, swept through a suitable frequency range, at a suitably high repetition rate, may be used.

In the above arrangements, the processing interferometer may be coupled between the optical source and the detector, for example between the source and the sensing interferometer or between the sensing interferometer and the detector, or may lie on a branch from this path. An advantage of coupling the processing interferometer between the broadband optical source and the sensing interferometer may be that it is easier to ensure a stable or constant polarisation of the light in the processing interferometer, which can be important for stability for example in a Mach Zehnder or other interferometer implemented using waveguides integrated onto a substrate such as ridge waveguides. Note that the light returning from the sensing interferometer may have a less satisfactory, stable or predictable polarisation, but that this will not generally have any adverse effect on detection by the photodetector.

Brief summary of the drawings

Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings, of which:

Figure 1 shows an optical sensor arrangement using white light

interferometry including a sensing interferometer and a processing

interferometer;

Figure 2 is a graph of fringes detected using the arrangement of figure 1 as the mismatch between the sensing and processing interferometers is varied about a zero matching point;

Figures 3a and 3b show an optical sensors embodying the invention using dual and single modes of a dither signal;

Figure 4 shows the arrangement of figure 3b with details of a feedback control scheme;

Figures 5a to 5c illustrate operation of the feedback control scheme of figure 4;

Figure 6 is a graph of multiple groups of fringes caused by multiple cavities at the sensing interferometer;

Figure 7 shows an optical sensor arranged to scan any of the multiple groups of fringes of figure 6 to interrogate any selected one of the cavities;

Figures 8a to 8c show different configurations of path length adjusters within a processing interferometer;

Figures 9a to 9c show different configurations of path length adjusters implemented using p-type and n-type depletion modulators;

Figure 10 illustrates in cross section a way of implementing a depletion modulator of figures 9a - 9c.

Detailed description of embodiments

The measurement range of a sensor system such as that shown in figure 1 depends, in addition to the response characteristics of the sensing interferometer, on factors such as the available range of the optical path difference OPD_p of the processing interferometer. In a sensor system where the anticipated measurement range is sufficiently small, the two interferometers may be constructed so that during use the signal at the photodetector always lies somewhere on a single fringe 26 as shown in figure 2, for example the fringe centred about OPD_s = OPD_p. To increase the measurement range while still fulfilling this condition, the sensor interferometer can be designed to be less sensitive to the parameter to be measured, so that over the full measurement range the change in OPD_s is still less than about one wavelength of the broadband light, but this will lead to reduced absolute accuracy of the sensing system. For example, in many applications, such as the use of a Fabry Perot based acoustic sensing interferometer for high temperature applications such as gas turbine and internal combustion engine use, the measurement of a small acoustic signal over a wide range of static pressures may require a dynamic range of the order of 1 x10⁶. This places requirements on the level of the noise floor which are not practical to meet over a single fringe, for example due to the relative intensity noise of broadband sources such as SLDs.

In many cases it may also be difficult to design a sensing interferometer subject to the dual requirements that OPD_s is significantly larger than the coherence length of the light source and the change in OPD_s over the full measurement range is less than about one wavelength.

Figure 3a illustrates a sensor system which is designed for operation of the processing interferometer over changes in OPD_p which are larger than the width of one fringe, typically tracking corresponding changes in OPD_s. Some aspects of the operation of this sensor system can be understood from the discussion of figure 1 above, and similar elements are generally indicated using the same reference numerals in both figures. The broadband optical source 30 may be a superluminescent diode, or some other suitable light source, and may be for example a combination of two or more discrete light sources such as two or more superluminescent diodes. It is not necessary, for example, for the broadband light to be spectrally evenly distributed or single peaked in spectral power, and a source emitting, for example, over two or more spectral peaks could be used.

Instead of a broadband source, a narrow band source such as a tunable laser, swept over a suitable frequency range, may be used. It may be necessary, however, to select a frequency sweep repetition rate which is sufficiently high to cooperate with other aspects of the apparatus, for example at a significantly higher rate than the first clock signal 60 delivered to the path length adjuster in the dither arm 62 discussed below in connection with figure 4.

The light from the broadband source 30 is passed through a depolariser element 32 to a processing interferometer 34, which in this case is a Mach

Zehnder interferometer comprising first and second arms and implemented using waveguides integrated on a semiconductor substrate, which could be a

conventional semiconductor substrate, or a semiconductor-on-insulator substrate or similar if required. The path lengths of one or both arms of the Mach Zehnder interferometer are controlled using one or more path length adjusters 36. Each such path length adjuster 36 can be provided by at least one phase modulator, such as a PIN diode arrangement having an intrinsic region arranged to coincide with the waveguide. Injection of free carriers into the intrinsic region by applying an appropriate current to the PIN diode is then used to control the refractive index in the waveguide. Such phase modulators are described, for example, in US 5,757,986. Other phase modulators may be advantageously used in the path length adjusters, for example as discussed below.

The output of the processing interferometer 34 passes through a 3dB coupler 38, or similar non-reciprocal component such as an optical circulator, to an optical fibre 40 for delivery to a sensing interferometer 42 which in this case is provided by a Fabry Perot cavity, for example as discussed in WO2009/077727. Light reflected back from the sensing interferometer 42 along the optical fibre 40 and arriving at the 3dB coupler 38 is directed to a photodetector 44, which may conveniently be provided by a photodiode. No particular spectral discrimination is required at the photodetector, which responds to the overall intensity of arriving light. The optical path differences of the sensing 42 and processing 34 interferometers are similar, but the optical path difference of the processing interferometer OPD_p is controllable, such that if the difference OPD_s - OPD_p is scanned over a suitably large range, the signal from the photodetector appears as a number of fringes peaking within a broader envelope, for example as already illustrated in figure 2. In the arrangement of figure 3 the processing interferometer is controlled by a feedback controller 50 which delivers control signals to the path length adjusters 36 of the processing interferometer 34 based on the signal received at the feedback controller 50 from the photodetector 44.

The feedback controller operates in two control modes. According to a first control mode, illustrated by control input 52, the feedback controller 50 sends control signals to the path length adjusters 36 to respond to the shape of a local fringe 26 in the photodetector output. According to a second control mode, illustrated by control input 54, the feedback controller 50 sends control signals to the path length adjusters 36 to respond to the shape of the envelope 28. More particularly, the feedback controller 50 may be arranged to adjust the control signals delivered to the path length adjusters 36 in order to cause the

photodetector output to seek and/or track either the peak of a fringe 26 or a peak of the envelope 28, depending on the current mode of operation.

Figure 3b shows a variation of the arrangement of figure 3a, in which the first mode mentioned above is not used, and instead the feedback controller operates in just a single mode in which control signals are sent to the path adjusters 36 to respond to the shape of the envelope 28 in order to track a peak of the envelope 28.

The feedback controller 50 of either figure 3a or figure 3b may operate in hardware, in software, or a combination of both. For a sensor system where the processing interferometer is integrated on a semiconductor substrate, the path length adjusters 36 may easily be controlled at frequencies above 1 MHz, and for some types of phase modulators at frequencies above 1 GHz, and to function at corresponding speeds, it would be desirable for the feedback controller 50 to operate largely using a hardware arrangement. Figure 4 shows how the feedback controller 50 of figure 3a may be so implemented, noting that the features used to effect the second control mode may also be used to implement the arrangement of figure 3b. Generally, one arm of the processing interferometer 34 is designated a servo arm 63, and the path length adjuster in this arm is driven with a slowly varying signal so that OPD_p generally seeks or tracks OPD_s. A second arm of the processing interferometer is designated as a dither arm 62, and the path length adjuster in this arm is driven with a more rapidly varying signal, with the corresponding changes in intensity at the photodetector being used to adjust the slower control of the servo arm.

Referring now to figure 4 in more detail, in the first mode of operation, a first clock signal 60 is delivered to the path length adjuster in the dither arm 62 of the processing interferometer 34 using a dither driver 64. This first clock signal is shown in the figure as a square wave, but may instead be a sinusoidal, triangular or other form, and is of a frequency at which the path length adjusters 36 can easily respond, for example 100 MHz. Using the dither driver 64, the first clock signal is delivered to the path length adjuster 36 of the dither arm with a magnitude sufficient for consequent oscillations in the photodetector output to be well within a single fringe 26. In the second mode of operation, a second clock signal 66 is delivered to the path length adjuster in the dither arm 62 of the processing interferometer 34. The second clock signal 66 is preferably a sawtooth or triangular wave, and may be of lower frequency than the first clock signal, for example around 10 MHz. The second clock signal is delivered to the path length adjuster of the dither arm 62 with a magnitude sufficient for the consequent photodetector output to cover several fringes 26, for example more than 5 and optionally more than 10 fringes, thereby covering a substantial part of the envelope 28. The physical response frequency of the sensing interferometer to external parameters such as pressure, temperature and vibration will at best be orders of magnitude below 10 MHz, so that both the first and second clock signals effectively provide a scan of OPD_p across respective parts of a single fringe 26 and the envelope 28, as OPD_s is held approximately constant.

The effects of the first clock signal on the photodetector output under three different circumstances are shown in figures 5a - 5c, in which photodetector output 70 is plotted against the relative optical path differences of the sensing and processing interferometers. In figure 5a, the first clock signal 60 causes the photodetector output 70 to oscillate between two points at either side of the peak of the central fringe 26. The photodetector output is therefore approximately level, and the feedback controller has established the position of the centre of the fringe 26. In figure 5b, the first clock signal 60 causes the photodetector output 70 to oscillate between a lower value to the left of the fringe peak, and a higher value to the right of the fringe peak. The photodetector output 70 therefore oscillates in phase with the first clock signal 60, indicative that the midpoint of the first clock signal represents a point to the left of the fringe peak. In figure 5c, the first clock signal 60 causes the photodetector output to oscillate between a higher value to the left of the fringe peak, and a lower value to the right of the fringe peak. The photodetector output therefore oscillates out of phase with the first clock signal 60, indicative that the midpoint of the first clock signal represents a point to the right of the fringe peak.

Returning to figure 4, in the first mode of operation both the output 70 of the photodetector 44 and the first clock signal 60 are passed to a multiplier 72 where they are combined, to provide a different result depending on whether variations in the photodetector output and first clock signal are in phase, out of phase, or the photodetector signal variations are small, as shown in figures 5a- 5c. The results of this combination are passed through a low pass filter 74 and an amplifier 76, the amplifier output being passed to a servo driver 65 to drive a path length adjuster 36 located in the servo arm 63 of the processing interferometer. If the photodetector output and first clock signal are in phase or out of phase as shown in figures 5b or 5c then the low pass filter 74 and amplifier 76 drive the path length adjuster of the servo arm 63 in one or other direction until the photodetector output 70 ceases to oscillate and adopts the pattern shown in figure 5a. In this way, the signal applied to the servo arm by the amplifier 76 is that required for the optical path difference of the processing interferometer 34 to track the optical path difference of the sensing interferometer, such that OPD_s- OPD_p is a fixed value, such as zero if the central fringe is tracked. The signal output by the amplifier 76 therefore also provides the output of the feedback controller 50, and may be passed on to data processing elements for further processing, for example to derive a temperature, pressure, or acoustic variation to which the sensing interferometer is subjected.

When the second clock signal 66 is delivered to the path length adjuster in the dither arm 62 of the processing interrogator, in the second mode of operation, the output from the photodetector is coupled to the multiplier 72 through an additional rectifier 78. The output of the multiplier 72 is still coupled through the low pass filter 74 and amplifier 76 to the path length adjuster 36 of the servo arm 63, but the parameters of the low pass filter 74 and amplifier 76 may be altered, for example as illustrated by parameter sets 80, 82, where the first parameter set 80 is applied to the low pass filter and amplifier in the first mode of operation, and the second parameter set 82 is applied to the low pass filter and amplifier in the second mode of operation.

When the second clock signal 64 is applied to the path length adjuster in the dither arm 62, such that the photodetector scans across multiple fringes 26, the effect of the rectifier 78 in combination with the low pass filter 74 is that the servo arm 63 of the processing interferometer is driven towards a peak of the envelope 28 instead of a peak in a local fringe.

Instead of operating using a feedback controller as described above, a sensor system according to the invention may be implemented in other ways. For example, instead of tracking a peak in the detector output, scanning of OPD_p may be used to derive a shape and a peak detected from the shape, whether in the first or second mode of operation. Hardware and/or software functions could be used in various ways to find and/or track the central or biggest fringe, or to provide a more complex single or multimode analysis of the fringes.

A sensor system may be operated to use the first and second control modes described above in a number of ways. On startup of the sensor system, or following a power interruption or error condition, a sensor system may

automatically operate in the second control mode for a fixed period or until a stability condition is met, and then switch to the first control mode. The sensor system may additionally or alternatively interleave use of the first and second control modes, for example operating for a predetermined period or sample count in one mode, and a predetermined period or sample count in the other mode, and/or switch modes according to one or more stability conditions. The feedback controller 50 may also operate in other modes additional to the first and second control modes.

The arrangements discussed above may also be used to provide a sensor system which is capable of interrogating a sensing interferometer having two or more different concurrent values of OPD_s, corresponding to different sensing interferometer cavities or path length differences. For example, figure 4 of WO2009/077727 shows an optical element which provides three separate Fabry Perot cavities, each of a different length, and each with different response characteristics to temperature and pressure at the optical element. As illustrated in figure 6, such an optical element will exhibit three separate groups of fringes 71 , 72, 73 in photodetector intensity as the value of OPD_p is scanned across the OPD_s of each sensing interferometer cavity in turn. The arrangements described above may be used to interrogate each of a plurality of such sensing

interferometer cavities, for example by first adjusting OPD_p to be at an

approximate predetermined position for a selected sensor cavity, refining the value of OPD_p responsive to the fringe envelope for the selected sensor cavity, and then using an individual fringe to enable the value of OPD_p to accurately track the value of OPD_s.

Figure 7 illustrates a sensor system arranged to interrogate multiple cavities in a sensing interferometer. The sensing interferometer 42 is shown as having three cavities A, B and C. A predetermined approximate offset value for OPDp corresponding to each of the three cavities is provided as OA, OB and Oc. The offset value for the sensor cavity currently selected for interrogation, using selector element 74, which may be provided as part of software control of the sensor and/or a hardware function, is provided to the feedback controller 50 which uses this to set the path length controllers 36 of the processing

interferometer 34 to provide a value of OPD_p which is close enough to that of the expected OPD_s value for the selected cavity for use of the second mode of operation of the feedback controller to lock on to the correct group of fringes 71 , 72 or 73 using second control mode and second clock signal 54. More refined tracking of a particular fringe within the group can then be achieved using the first control mode and first clock signal 52 as generally described above, or a single mode of control may be used for example as discussed in connection with figure 3b.

An alternative way to implement interrogation of multiple sensing cavities of the sensing interferometer is to provide multiple parallel groups of processing interferometer, photodetector and feedback controller, each group being arranged to interrogate and provide an output for a corresponding one of the sensor cavities. In such an arrangement, the multiple processing interferometers may conveniently be implemented on a single substrate.

Although figures 3a, 3b and 4 show a processing interferometer with a path length adjuster 36 in each of the servo 63 and dither 62 arms, the same functionality can be delivered using a path length adjuster 36 located in just one arm of the processing interferometer by suitable summing of the dither and servo signals as illustrated in figure 8a. Similarly, a path length adjuster 36 common to both arms of the processing interferometer and driven by a summed signal as shown in figure 8b, or separate path length adjusters each common to both arms but separately driven by the servo and dither signals as shown in figure 8c, may be used, in which cases referring to a particular interferometer arm as either a dither arm or a servo arm may not be appropriate.

A path length adjuster 36, as discussed above, can be constructed using a

PIN diode arranged to inject free carriers into an intrinsic region to provide a phase modulator. Depending on the semiconductor material and construction of the PIN diode, the density of carriers in the intrinsic region and waveguide of the phase modulator 34 leads to changes in the refractive index, and therefore OPD_p. For a typical silicon carrier injection phase modulator the change in refractive index may be substantially a linear function of the current applied to the phase modulator, over normal operating ranges, for example see Soref and Bennett, "Electro optical effects in Silicon", IEEE Journal of Quantum Electronics, vol 23, issue 1 , 1987, 123 - 129. An issue with the use of such phase modulators, however, is that the optical loss of the broadband light passing through the phase modulator is also a strong function of the free carrier density in the waveguide. Similar loss effects with carrier density may be seen in other semiconductor materials such as InP and GaAs.

In order to change the value of OPD_p significantly using such a phase modulator, the optical losses in the two arms of the processing interferometer will diverge significantly, heavily distorting the shape of the fringe envelope 28 illustrated in figure 2, possibly to the extent that fringes across part of the envelope become indistinguishable from the background noise. In order to better scan over several fringes or the whole shape of the envelope 28, for example as desirable in putting into effect the second control mode discussed above, or between sets of fringes resulting from different optical cavities in the sensing interferometer as illustrated in figure 6, it would be desirable to provide a processing interferometer 34 in which the optical loss in the two arms was approximately balanced over a wide range of OPD_p, with the result that the symmetry of the envelope and relative visibilities of the separate fringes is maintained. This can be achieved using phase modulators with minimum levels of optical loss, for example thermal phase modulators. However, a thermal phase modulator has a very slow response time compared with a semiconductor charge carrier based modulator, so is not useful in a sensor where dithering of a path length controller or similar at frequencies, say above 100 KHz, or above 1 MHz, may be required.

A processing interferometer 34 which may be used to provide fast modulation with balanced optical loss over multiple fringes is illustrated in figure 9a. Such a processing interferometer 34 may be used in any of the sensor system arrangements described herein, and in other applications such as optical switches and (de)multiplexers. The processing interferometer 34 of figure 9a comprises a path length adjuster 36 comprising a depletion modulator in each of the two arms of the interferometer to form a transmissive Mach Zehnder interferometer, although other types of interferometer may be used such as a Michelson interferometer, a reflective Mach Zehnder interferometer, and others. A first one of these depletion modulators 90 is a p-type depletion modulator, characterised in that the phase modulation of the light is dominated by the presence and/or depletion of p-type carriers (holes). A second one of these depletion modulators 92 is an n-type depletion modulator, characterised in that the phase modulation of the light is dominated by the presence and/or depletion of n-type carriers (electrons). The optical losses caused in a depletion modulator by p-type and n-type carriers of the same density is about the same, but the phase modulation effect of the p-type carriers is typically about four to five times greater than that of the n-type carriers for the same carrier density in silicon.

The p-type depletion phase modulator 90 and n-type depletion phase modulator 92 may therefore be driven to provide similar optical loss in the two interferometer arms, while yielding a net change in path difference for the processing interferometer 34 as a whole.

The following equations, which can be derived from the Soref and Bennett reference above, describe the refractive index change and loss for each type of charge carrier in silicon, where χ denotes refractive index change, σ denotes loss per cm of waveguide, n and p denote the electron and hole concentrations per cm^"3. At light wavelength of 1 .31 μιη:

A_Z = A_Ze + A_Zh = -{6.2χ10-²² · Δ« + 6.0χ10-¹⁸ · (Δρ)^{0 8} } (1 ) Aa = Aa_e + Aa_h = 6.0 xlO^"18 · An + 4.0xl0^"18 · Ap (2) At light wavelength of 1 .55 μιη:

A_Z = A_Ze + A_Zh = -{8.8χ10-²² · Δη + 8.5 χ10-¹⁸ · (Δρ)^{0 8} } (3) Aa = A _e + A _h = 8.5 x lO^"18 · An + 6.0x l0^"18 · Ap (4) It is apparent from these equations that p-type (hole) carriers are more effective at perturbing refractive index than n-type (electron) carriers, whereas the optical losses are quite similar.

An optical loss balanced path length adjuster 36 can therefore be constructed using complementary p-type and n-type depletion modulators in first and second arms of the interferometer, in which the p-type depletion modulator is arranged in the first arm to provide more than the required phase modulation, and the n-type depletion modulator is arranged in the second arm to at least partially balance, in the second arm, the optical loss caused by the p-type carriers in the first arm, while only partly offsetting the phase modulation in the first arm.

Desirably, the depletion modulators in the first and second arms may be controlled so that the optical losses caused by the two depletion modulators are substantially the same, or at least matched to within, say, about 10%. Using a path length adjuster of this type in the sensor arrangements described above provides the advantage that the visibility of the fringes 26 is not degraded by differing optical losses in the two arms of the processing interferometer, although there may be a penalty in that the optical loss for the whole device and therefore the noise in the sensor output increases.

In the arrangement of figure 9b two path length adjusters 36 similar to the path length adjuster 36 of figure 9a are shown. The path length adjuster to the left side in figure 9b provides a p-type depletion phase modulator 90 in the upper arm of the interferometer and an n-type depletion phase modulator 92 in the lower arm, whereas the path length adjuster to the right side provides a p-type depletion phase modulator 90 in the lower arm, and an n-type depletion phase modulator in the upper arm. In such an example, a dither signal as discussed above could be applied to the left side path length adjuster as illustrated by the connection of the dither driver 64, and a servo signal to the right side path length adjuster as illustrated by the connection of the servo driver 65, or vice versa.

In the arrangement of figure 9c, the path length adjuster to the left side is provided by a p-type depletion phase modulator in a first arm and an n-type depletion phase modulator in a second arm of the processing interferometer, and this path length adjuster may be used to provide the rapid and wide ranging adjustments to OPD_p driven by the dither signal as described above. The path length adjuster to the right side is provided by a thermal type phase modulator in which the temperature of the semiconductor material of the waveguide is varied to thereby control the refractive index. The thermal type phase modulator has a slow response time making it unsuitable for applying the dither signal to the interferometer, but may be adequate to apply the much more slowly varying servo signal in some circumstances. Figure 10 illustrates a way in which a depletion phase modulator discussed above may be constructed such that the modulation in the waveguide is dominated by the density of either p-type or n-type carriers as desired, according to details of the particular design. A ridge structure 100 formed in a layer 102 of silicon (or another semiconductor such as InP or GaAs), and an insulator layer 104 formed on substrate 106 underneath the ridge structure 100 define a waveguide structure for constraining propagation of light modes 108. To one side of the centre of the ridge structure the layer 102 is p-doped and to the other side the layer 102 is n-doped to form a p-n junction within the waveguide structure. The depletion region in a typical p-n junction is about 0.3 μιτι across, so that the ridge structure 100 should be of a similar size, say between about 0.1 and 1 .5 μιτι across for the depletion region to dominate the area in which the light modes 106 propagate, depending on details of the particular design. To make the p-n junction asymmetric or one sided, either the p-type or n-type doping is made much stronger than the other. This has the effect of providing a higher net concentration of either n-type or p-type carriers, respectively, close to the p-n junction at the centre of the ridge structure where the light modes 108 propagate. Applying an appropriate reverse bias to the junction using contacts 1 10 then depletes the concentration of the dominant carrier, providing the desired modulation effect. The structure can be optimised or adjusted in various ways, for example by positioning the centre of the p-n junction off centre within the ridge 100, and/or by adjusting the abruptness and depth of the junction. In particular, it may be desirable to position the p-n junction a little off centre from the

propagation modes 108, so that the propagation modes lie more within the heavily doped side of the junction where depletion will dominate, than the lighter doped side.

For an abrupt p-n junction, the width of the depletion junction may be given by:

2s, {N

W = bi (5)

I(N_AN_D ) where ε₃ is the dielectric coefficient, N_a and N_d are the acceptor and donor doping concentrations respectively, q is the electronic charge, and V_bl is the built-in potential, where V_bi is given by:

_T7 kT . N_A.N_D

V_bi =— Ι^η^γ^ (6)

where Ze is Boltzmann's constant, Tis temperature in Kelvin and Π_/ is the intrinsic carrier density.

The width of this depletion region can be varied by application of a reverse bias voltage, and the density of free carrier charges in the waveguide ridge 100 varies as the depletion region width is varied. To make such a junction

asymmetric or one sided, the doping concentration is made much higher on one side than on the other. In such a case it can be shown that the width of the depletion region can be simplified from the above expression, and also that the depletion layer width 1Vdepends on the applied voltage bias as follows:

2*, (V„ bi -V)

(7)

' B where V is the applied bias voltage, and N_B is the concentration of the lighter doping type. A depletion phase modulator using such a one-sided junction, under reverse bias, will largely deplete carriers of the lower doping concentration type in the area where the light modes 106 propagate. If the lower doping concentration is p-type then this will provide a p-type depletion modulator. Similarly, if the lower doping concentration is n-type then this will provide an n-type depletion modulator. By way of numerical example, an p-type depletion modulator may be constructed where the n-type doping has a concentration of about 1 x10¹⁷ cm^"3, and the p-type doping has a concentration of about 1 x10¹⁶ cm^"3, so that at full depletion one could obtain an n-type carrier concentration change of the order of 1 x10¹⁷ cm^"3. From equation (3) above this would provide an optical path length change of about 4 μιη for each length of 1 cm of p-type depletion modulator waveguide. If such a p-type depletion modulator is used in a first arm of the processing interferometer 34, a similar n-type depletion modulator may be used in the second arm with a similar optical loss, but an optical path length change of only about 1 μιτι for each length of 1 cm of n-type depletion modulator

waveguide. These optical path length changes apply to silicon but the loss matching technique may be applied to other semiconductor materials such as InP and GaAs.

An integrated waveguide processing interferometer as described above using depletion modulators may also be used in other applications, in particular in any application where it is required to at least partly balance the optical loss between two arms of the interferometer. Such applications may include, for example, multiplexers and optical switches, and may be implemented as transmissive, reflective, Mach Zehnder, Michelson and other interferometer types as required.

Although specific detailed embodiments of the invention have been described, the skilled person will appreciate that modifications and variations on these can be carried out without departing from the scope of the invention as defined in the appended claims.

Where terms such as left, right, top, bottom have been used, these are only exemplary and not limiting unless the context demands otherwise. The described sensor arrangements may be used for a variety of purposes, but may in particular be useful for temperature, pressure and/or acoustic sensing in high temperature environments such as gas turbines and internal combustion engines. Although described embodiments use Mach Zehnder and Fabry Perot

interferometers, other sensing and processing interferometer types may be used.

Claims

CLAIMS:

1 . An optical sensor comprising:

a sensing interferometer;

an optical source coupled to deliver light to the sensing interferometer; a photodetector coupled to receive the light from the sensing

interferometer; and

a processing interferometer coupled to and matchable to the sensing interferometer, the processing interferometer being formed using waveguides integrated onto a substrate,

wherein the processing interferometer comprises first and second interferometer arms and at least one path length adjuster to control a path length difference between the arms, each path length adjuster comprising a first depletion modulator in one of the first and second interferometer arms and a second depletion modulator in the other of the first and second interferometer arms, such that modulation of the light is dominated by p-type carriers in the first depletion modulator, and is dominated by n-type carrier depletion in the second depletion modulator.

2. The optical sensor of claim 1 wherein optical loss of said light in the processing interferometer is dominated by loss due to p-type carriers in the first depletion modulator, and by loss due to n-type carriers in the second depletion modulator.

3. The optical sensor of claim 2 wherein the second depletion modulator is arranged to balance between the interferometer arms at least some of the optical loss caused by the first depletion modulator.

4. The optical sensor of claim 3 wherein the optical sensor is configured such that optical loss of said light in the first and second depletion modulators is balanced to within 10% during operation.

5. The optical sensor of any preceding claim wherein each depletion modulator comprises a one-sided p-n junction providing a depletion region, at least part of said light passing along the corresponding interferometer arm passing through said depletion region.

6. The optical sensor of any preceding claim arranged such that the signal detected by the photodetector as a function of path length mismatch between the sensing interferometer and the processing interferometer comprises a series of fringes.

7. The optical sensor of any claim 6 comprising a first said path length adjuster comprising a said first depletion modulator in the first interferometer arm and a said second depletion modulator in the second interferometer arm, and second said path length adjuster comprising a said first depletion modulator in the second interferometer arm and a said second depletion modulator in the first interferometer arm.

8. The optical sensor of claim 7 arranged to repeatedly drive the first said path length adjuster across at least a portion of said fringes using one or more predetermined dither signals, and to drive the second said path length adjuster based on a resulting signal received at the photodetector.

9. An optical sensor comprising:

a sensing interferometer;

an optical source coupled to the sensing interferometer;

a photodetector coupled to the sensing interferometer;

a processing interferometer integrated onto a substrate and matchable to the sensing interferometer, the processing interferometer being arranged such that the signal detected by the photodetector as a function of optical path difference mismatch between the sensing interferometer and the processing interferometer comprises a series of fringes within an envelope; and a controller arranged to control the processing interferometer to scan across a plurality of said fringes.

10. The optical sensor of claim 9 wherein the controller is arranged to control the processing interferometer to scan across less than one said fringe in a first control mode, and across a plurality of said fringes in a second control mode.

1 1 . The optical sensor of claim 10 wherein the controller is a feedback controller arranged to combine a dither signal configured for scanning the processing interferometer across a portion of one fringe in the first control mode and across a plurality of said fringes in the second control mode, with a signal received from the photodetector, to process the combined signal, and to use the processed combined signal to provide a servo signal arranged to drive the processing interferometer responsive to said fringe in the first control mode and responsive to the envelope in the second control mode.

12. The optical sensor of claim 1 1 wherein the signal received from the photodetector is rectified in the second mode.

13. The optical sensor of claim 1 1 or 12 wherein the processing the combined signal comprises applying a low pass filter to the combined signal.

14. The optical sensor of any of claims 10 to 13 wherein the controller is arranged to adjust the processing interferometer so that a peak of an individual fringe is tracked in the first control mode, and so that a peak of the envelope is tracked in the second control mode.

15. The optical sensor of claim 14 wherein the peak of the envelope coincides with the individual fringe, and the optical sensor is arranged to operate in the second control mode to find the individual fringe in advance of operating in the first control mode to track the peak of the individual fringe.

16. The optical sensor of any of claims 1 1 to 15 wherein the controller is arranged to scan the processing interferometer using the dither signal across less than one fringe in the first control mode, and over a plurality of fringes in the second control mode.

17. The optical sensor of any of claims 9 to 16 wherein the processing interferometer comprises first and second arms implemented using waveguides integrated onto the substrate, the arms being provided with at least one path length adjuster arranged for control of the optical path difference of the

processing interferometer by the controller.

18. The optical sensor of claim 17 wherein at least said one path length adjuster is implemented as one or more PIN diodes, with the intrinsic region of the one or more PIN diodes intersecting said waveguide for injection of charge carriers into the waveguide.

19. The optical sensor of claim 17 wherein at least one said path length adjuster is implemented as one or more depletion modulators, with the depletion region of the one or more depletion modulators intersecting said waveguide for depletion of charge carriers from said waveguide to thereby control optical path length in the path length adjuster.

20. The optical sensor of claim 19 wherein the at least one said path length adjuster comprises a p-type depletion modulator in the first arm and an n-type depletion modulator in the second arm.

21 . The optical sensor of any of claims 10 to 20 arranged to interleave periods of operation according to the first and second control modes.

22. The optical sensor of any of claims 10 to 21 arranged to operate according to the second control mode following turning on of the optical sensor or following detection of an error condition, so as to establish the position of an individual fringe to be scanned across in subsequent operation according to the first control mode.

23. The optical sensor of any of claims 9 to 22 wherein the sensing

interferometer provides a plurality of sensing cavities, and the controller is arranged to use the second control mode to lock onto any selected one of said sensing cavities.

24. The optical sensor of claim 23 wherein the controller is arranged to track an individual fringe of the selected sensing cavity using the first control mode.

25. An optical sensor comprising:

a sensing interferometer having a plurality of optical cavities of different lengths;

an optical source coupled to the sensing interferometer;

a photodetector coupled to the sensing interferometer;

a processing interferometer matchable to each of the optical cavities of the sensing interferometer, the processing interferometer being arranged such that the signal detected by the photodetector as a function of optical path difference mismatch between each optical cavity of the sensing interferometer and the processing interferometer comprises a different series of fringes; and

a controller arranged to control the processing interferometer so as to enable the sensor to detect one or more properties of the series of fringes corresponding to any selected one of the optical cavities.

26. The optical sensor of claim 25 further comprising a selector arranged for selection of any one of the optical cavities.

27. An optical sensor according to any of claims 1 to 26 arranged to detect one or more physical parameters at the sensing interferometer by analysis of broadband light received at the detector.

28. The optical sensor of claim 27 wherein the one or more physical parameters include at least one of temperature, pressure, and acoustic vibration.

29. The optical sensor of any preceding claim wherein the sensing

interferometer comprises a Fabry Perot interferometer.

30. The optical sensor of any preceding claim wherein the processing interferometer comprises a Mach Zehnder interferometer.

31 . The optical sensor of claim 30 wherein the Mach Zehnder interferometer is integrated onto a silicon substrate.

32. A method of operating an optical sensor which comprises a sensing interferometer, an optical source coupled to deliver light to the sensing

interferometer, a photodetector coupled to receive the light from the sensing interferometer, and a processing interferometer coupled and matchable to the sensing interferometer, the processing interferometer being formed using waveguides integrated onto a substrate and comprising p-type and n-type depletion modulators in respective waveguide arms of the interferometer, the method comprising:

driving the p-type and n-type depletion modulators to provide a variable optical path difference of the processing interferometer, wherein optical loss in the p-type depletion modulator is at least partially balanced by optical loss in the n-type depletion modulator.

33. The method of claim 32 where in the optical sensor is arranged such that the signal detected by the photodetector as a function of path length mismatch between the sensing interferometer and the processing interferometer comprises a series of fringes, and the method comprises driving the p-type and n-type modulators to scan the path length mismatch over a plurality of said fringes.

34. A method of operating an optical sensor comprising a sensing

interferometer, an optical source coupled to the sensing interferometer, a photodetector coupled to the sensing interferometer, a processing interferometer integrated onto a substrate and coupled and matchable to the sensing

interferometer, the processing interferometer being arranged such that the signal detected by the photodetector as a function of optical path difference mismatch between the sensing interferometer and the processing interferometer comprises a series of fringes within an envelope, the method comprising:

controlling the processing interferometer to scan across a plurality of said fringes.

35. The method of claim 34 further comprising operating the sensor in first and second control modes, in which the processing interferometer is controlled to repeatedly scan across less than a singlefringe in the first control mode, and to scan across a plurality of said fringes in said second control mode.

36. The method of claim 35 wherein the processing interferometer is controlled using a feedback loop based on a signal at the photodetector.

37. A method of operating an optical sensor comprising a sensing

interferometer having a plurality of optical cavities of different lengths, an optical source coupled to the sensing interferometer, a photodetector coupled to the sensing interferometer, and a processing interferometer arranged such that the signal detected by the photodetector as a function of optical path difference mismatch between each optical cavity of the sensing interferometer and the processing interferometer comprises a different series of fringes, comprising: selecting one of the optical cavities;

controlling the processing interferometer responsive to one or more properties of the fringes corresponding to the selected optical cavity; and

detecting a property of the selected cavity from the signal detected by the photodetector.

38. The method of claim 37 wherein the property of the selected cavity is detected from control of the processing interferometer under a feedback loop including the signal detected by the photodetector.

39. The invention of any preceding claim wherein the optical source is a broadband optical source.

40. An integrated semiconductor interferometer arranged to provide a controlled optical path difference comprising:

first and second waveguide arms;

a p-type depletion modulator disposed on the first waveguide arm; and an n-type depletion modulator disposed on the second waveguide arm.

41 . The integrated semiconductor interferometer of claim 40 wherein the n- type depletion modulator is arranged to balance between the waveguide arms at least some of the optical loss caused by the p-type depletion modulator in the first arm.

42. The integrated semiconductor interferometer of claim 40 or 41 wherein each depletion modulator is formed using a one-sided p-n junction, with the optical modes of the respective waveguide arm intersecting the depletion region of each respective p-n junction.

43. An optical switch comprising the integrated interferometer of any of claims 40 to 42.