Field of the Invention

The present invention relates to an improved apparatus and method for making measurements. In particular, the present invention relates to a measurement system using optical fibres and two-photon fluorescence.

Background to the Invention

There is a growing need for sensors that are capable of very remote measurement and measurement at many points. There is also a need for sensors that are impervious to electromagnetic interference, operational over a wide range of temperatures and are unobtrusive. Optical fibre based sensing systems show great promise in meeting these criteria.

Optical fibre sensing systems can be divided into extrinsic and intrinsic types. In extrinsic sensors, the optical fibre is used purely to transfer the data from a sensing site to a measurement site. In intrinsic sensors, the optical fibre is used as both the measurement means and the information carrier. Examples of extrinsic sensors include those using a miniature Fabry-Perot etalon or a bead of temperature sensitive material. These sensors commonly employ high information capacity multi-modal fibres to transport signals between the measuring and detecting devices. This is because the variable quantity in these systems is generally a change in light intensity or wavelength, which is undistorted in these fibres. In contrast, intrinsic sensors generally rely on variation of the properties of the fibre and as such generally employ single mode fibres having predictably varying properties and wave vector directions. These considerations are well known and are summarised for example in "Fiber Optic Sensor Technology: An Overview", Sensors and Actuators, VoI 82 (2000) p40- 61.

Fibre optic sensors can be further grouped by measurement capacity into point, quasi-distributed and distributed sensors. Point sensors make readings at one location and include most of the extrinsic sensors discussed above and some types of intrinsic fibre sensors. Quasi-distributed sensors are sensors that can measure a property at several discrete sites and are essentially a string of point sensors connected in series. The most versatile type is a fully distributed sensor, for which a measurement can be taken at any point along the length of the fibre. By definition these are always intrinsic systems.

A sensor system commonly used in both point and quasi- distributed systems is that based on the use of fibre Bragg gratings (FBGs) . This uses a fibre core containing periodic sections of varying index of refraction to produce a Bragg grating that reflects light of a specific wavelength that is dependant on the spacing of the periodic sections. As the fibre expands and contracts under thermal action or strain, the spacing of the periodic sections of the FBG and thus the wavelength of the reflected light changes correspondingly. Thus, larger measurement ranges require larger variations in wavelength. Several FBGs, each with its own distinct range of reflecting wavelengths, may be incorporated in one fibre to achieve a multiplicity of measurement sites. A disadvantage of this type of sensor is that to prevent overlap of signals as the number of FBGs on a single fibre increases, the possible range of measurement decreases, thus limiting the number of sensing locations. This effect is overcome in fully distributed sensors such as those based on the use of Raman or Brillouin scattering.

Light travelling down a fibre optic cable is partially reflected due to Rayleigh, Mie, Brillouin and Raman (Stokes/Anti-Stokes) scattering. The ratio of the intensity of the Stokes to the anti-Stokes component of the reflected light is dependant on temperature. Hence, injecting a pulse of light into a fibre optic cable and analysing this ratio relative to time can be used to obtain distributed temperature readings. However, techniques relying on Raman interactions suffer from the fact that the resulting signals are weak relative to other backscattered effects. This leads to a poor signal to noise ratio, long integration times and commonly the need for some form of signal amplifying technique.

Brillouin scattering involves the interaction of an optical wave (pump) with phonons in the fibre to create backscattered light, which varies in wavelength from the pump wave by an amount known as the Brillouin shift. The magnitude of the Brillouin shift is temperature and strain dependant. Thus it is possible to employ this effect to produce a remote temperature sensor. A common approach involves use of stimulated Brillouin scattering, which involves passing both a pump light wave and a counter-propagating light pulse (probe) through the system. If the frequency of the probe pulse is less than the frequency of the pump wave by an amount exactly equal to the Brillouin shift, this leads to an enhancement of the phonon population and thus an increase in intensity of the probe pulse through Brillouin scattering. Either the time domain waveform or the time delay and intensity of a varying frequency pulse can be analysed to obtain the value and the position of the measurement and thus achieve fully distributed sensing.

There have been several attempts to implement sensors using Brillouin scattering such as those described in GB2289331 or in "Simple Distributed Fiber Sensor Based on Brillouin Gain Spectrum Analysis", Optics Letters, V.21, No. 10 (1996) p758 or in "Technique for Measuring the Dynamic Strain on an Optical Fiber Based on Brillouin Ring Amplification", Journal of Lightwave Technology, V17 (1999), p234. Spatial resolutions of 10m in lOOkms and temperature resolutions of ±1°C have been claimed for methods based on Brillouin effects. However, the various known implementations suffer to a greater or lesser extent from a range of problems. These include a poor signal to noise ratio as a result of the weakness of Brillouin effects, attenuation due to interaction of the probe pulse at several points of similar temperature and strain and the need for stable light generation as these methods are very sensitive to input light intensity.

Measurement using single photon fluorescence in optical fibres (as described in "Sapphire-Ruby Single-Crystal Fibre for Application in High Temperature Optical Fibre Thermometers up to 15000C", Measurement Science & Technology, V.12 (2001) p.981 or patent US4223226) overcomes many of these limitations. This technique involves using light to stimulate a fluorescing material whose rate of fluorescent light decay and fluorescence spectrum upon removal of the stimulus is dependent on the temperature. This method is less sensitive to the intensity of input light than Brillouin methods and is capable of operating up to very high temperatures due to the single crystal fibres used (maximum operating temperature «2000°C compared with «1000°C for the traditional silica fibres used in other methods) . However, the measurement devices using single photon fluorescence described to date suffer the limitation of being able to take measurements at only a few discreet points, rather than the fully-distributed sensing achieved through use of the Brillouin or Raman methods.

An object of the present invention is to overcome one or more of the problems associated with the above prior art.

Summary of Invention The present invention utilises two-photon absorption and fluorescence (TPAF) to achieve measurement of various properties. TPAF involves the simultaneous absorption of two photons by an optically active ion. This leads to the promotion of an electron from a lower energy state of the ion directly to a higher energy state having an energy that is greater than that of the lower state by an amount equal to the sum of the energies of the photons. This step is achieved without the involvement of an intermediate energy level. Some relaxation from the higher energy level down to an intermediate level may or may not take place. The exited electron then relaxes back to a terminating energy level from an emitting level through emission of a fluoresced photon whose energy is equal to the difference in energy between the emitting and terminating energy levels. The overall rate of relaxation is characterised by the upper state decay time, τ, which is dependent on the nature of the ions, the host in which they are embedded and factors in the external environment such as temperature.

According to a first aspect of the present invention there is provided a method for measurement of a physical parameter, such as temperature or strain, the method involving injecting pulses of light into opposite ends of a fibre made of a fluorescent material, the wavelengths of the injected light pulses being such that two photon fluorescence occurs through absorption of a photon from each pulse, detecting the fluoresced light and determining a value of the parameter using a property of the fluoresced light.

Preferably, the fluorescent material used is such that there are no intermediate levels between the lower and upper energy levels between which TPAF takes place.

The method may further involve varying the timing of the injection of the two counter-propagating pulses relative to each other to control the point at which the pulses meet and thus where the measurement is taken. Preferably, the process can be repeated with a wide range of pulse timings to obtain measurements at as many points along the fibre as is required.

The two counter-propagating light pulses may have different wavelengths. v'

The property of the fluoresced light may be for example, the decay rate of fluoresced light, the spacing of spectral lines of the fluoresced light or the intensity ratios of the spectral lines. The parameter may be temperature or strain or fluid flow or magnetic field strength or electric field strength.

The method may further involve providing calibration data and using that calibration data in the step of determining thereby to identify the value of the parameter that is being measured.

According to a second aspect of the present invention, there is provided an apparatus for distributed measurement of a physical parameter, the system having means for injecting pulses of coherent light into opposite ends of a fibre optic cable made of a fluorescent material, the wavelengths of the injected light pulses being such that two photon fluorescence occurs through absorption of a photon from each pulse, means for detecting the resulting fluoresced light and means for using the detected light to determine a value of the physical parameter.

The means for injecting pulses of coherent light may be one or more pulsed lasers. Examples of suitable pulsed lasers include mode-locked lasers and gain switched diode lasers. Preferably, the pulsed lasers have an electronic means of producing time delay between their drivers.

The fibre optic cable may include either ion-doped glass or an ion-doped single crystal.

The detector may be a phase sensitive detector. The apparatus may include at least one beam splitter for isolating fluoresced light and diverting it towards the means of detection.

The fibres may either be single mode or multi-mode fibres.

Description of the Drawings

Various aspects of the invention will now be described by way of example only and with reference to the accompanying drawings of which:

Figure 1 is a schematic of the apparatus of the present invention;

Figure 2 is an energy level diagram for a two-photon fluorescence process in a doped single crystal fibre;

Figure 3 is an energy level diagram for a two-photon fluorescence process in a doped glass fibre;

Figure 4 shows the results of a computational model of the time development of fluorescence intensity of light emitted as a result of a two-photon fluorescence process with time; and

Figure 5 is a graph showing a typical example of the decay in fluorescence intensity of light emitted as a result of a two-photon fluorescence process with time.

Specific Description The present invention provides a system and method for determining a physical parameter such as temperature or strain using two-photon absorption and fluorescence (TPAF) . This involves the simultaneous absorption of two photons to excite an electron directly from a lower energy level of an ion to a higher energy level of an ion without the involvement of an intermediate state. The exited ion then relaxes from an emitting level back to a terminating energy level via emission of a fluoresced photon with an energy equal to the difference in energy between the emitting and terminating energy levels. In use, to achieve a fluoresced light intensity of a measurable magnitude, pulses of many photons may be used, resulting in multiple excitations and fluoresced photons . Several properties of the fluoresced light, such as the rate of decay in the overall intensity of fluoresced light, the spacing of spectral lines or the relative intensities of spectral lines, are dependant on external factors including temperature and strain. Hence, measuring these provides a mechanism for measuring environmental parameters such as temperature and strain.

Figure 1 shows a measurement system having two lasers 5 and 10 connected to opposite ends of a fibre optic cable 15. The lasers 5, 10 of Figure 1 are operable to inject light pulses 20, 25 into opposing ends of the cable. In order to provide control, the lasers 5 and 10 are operable to vary the delay between pulses. The lasers may be pulsed lasers. Examples of suitable lasers are gain switched diode lasers. Preferably, variable electronic delay means are provided between the laser drivers, so that the position at which the counter propagating pulses meet can be varied.

The fibre optic cable 15 is made of a fluorescent material, preferably a fluorescent material with no energy bands that may result in single photon absorption of individual photons 30, 35 having energy less than the energy of the transition 40. This minimises the attenuation and thus maximises the range of the device.

The fibre optic cable 15 may be a single mode fibre optic cable. Single mode fibres provide increased spatial resolution due to the lack of modal dispersion. Alternatively, the fibre optic cable 15 may be a multimode fibre optic cable. Using a multi-mode fibre permits the use of ion doped single crystal fibres, which have discrete energy levels and can be used up to high temperatures (2000°C) .

Preferably lanthanide or transition metal ions are used as the dopants in the fibre material. Praseodymium, neodymium, erbium, thulium and ytterbium, are examples of suitable dopants. Conveniently, they possess absorption bands in the visible or near infrared, which can be excited by two photons from readily available diode lasers.

At an end of the fibre of Figure 1 is a beam splitter that is arranged so that the input light pulses 20, 25 can pass into the fibre optic cable 15 but any fluoresced light 45 is separated out to at least one detector 50. The beam splitter 55 may be any device that allows the separation of fluoresced light 45 from other light that may be present in the fibre optic cable 15, such as input light 20, 25 or backscattered light. Examples of suitable beam splitters 55 include fibre couplers or filters or one-way semi-reflective mirrors such as dichroic mirrors. As noted previously, the beam splitter 55 is positioned so as to divert the fluoresced light out to a detector 50. This detector 50 may use the amplitude and phase of the detected signal to increase the signal to noise ratio and to allow selective detection of the fluoresced light 45, thus increasing the signal to noise ratio. Examples include phase sensitive detection with lock-in amplification, photon counting and boxcar integration.

In order to stimulate TPAF, the wavelength of the light injected into the ends of the cable has to be carefully selected, depending primarily on the energy profile of the fluorescent material of the fibre. As an example, Figure 2 shows an energy level diagram for the TPAF process in ion doped single crystals. The energy of an excited state 65 in ion doped single crystals is usually discrete and sharply defined. Thus to achieve TPAF, the wavelengths of the input pulses 20, 25 have to be such that the sum of the energies E, E', of photons 30, 35 from both pulses 20, 25 is exactly equal to the energy of a possible transition 40 from a lower energy level 60 to a higher energy level 65 in the ion doped single crystal fibre 15. If the wavelengths of the two pulses 20, 25 are different, i.e. the energy of the individual photons 30, 35 is not exactly half the energy required for the transition 40 to occur, simultaneous absorption of two photons from the same pulse will not result in TPAF. In addition, certain ion-doped glass fibres 15 also have energy levels that allow application of this method. Possible examples for this mode of operation in ion doped single crystal fibre 15 or ion doped glass fibre 15 include the 3H6-VG4 transition in thulium using pulses

20, 25 of wavelength 905nm and 980nm, and the 7F,-»5D, transition in terbium with pulses 20, 25 of wavelength 905nm and 1064nm. Since the only absorption of photons 30, 35 occurs through two-photon absorption of a photon from each of the counter propagating pulses 20, 25, absorption of photons 30, 35 only occurs at the measurement point 70, thereby minimising the attenuation and increasing the signal to noise ratio.

Figure 3 shows an energy level schematic of the TPAF process in ion-doped glass. Exited energy levels 65 in ion-doped glass are such that there is a broad range of allowed excited energy levels smeared into a band. Thus, in this case, the sum of the energies of photons 30, 35 from two input pulses can be equivalent to or slightly greater than the energy difference 40 between the terminating energy level 60 and the emitting energy level 65 by an amount equal to the range of allowed energies of the exited states, whilst still producing TPAF. Thus, single photon fluorescence by absorption of an individual photon 30 or 35 from either pulse 20 or 25 will not occur. Where the photons 30 have a higher energy than photons 35, TPAF will not arise due to the combination of two lower energy photons 35, since the sum of the energies of these photons 35 would be insufficient to facilitate the transition 40. However, the combination of two higher energy photons 30 may lead to a background fluorescence effect.

The background fluorescence effect may be characterized theoretically for rectangular light pulses 20, 25 whose duration is short in comparison with the dopants' fluorescence decay times. In this theoretical treatment, a pulse of duration Δt containing N photons of energy hυ is incident on an element of fiber of cross-sectional area A, dopant number density N1and length dl . An upper state population density N2(t) is generated by two-photon absorption, which decays exponentially with a decay time x, except at the sensing location where the decay time is τH (H = "Hot") . As a result, the excited element of fiber emits a fluorescence power dp(t)TPAF given by

dP(t)TPAF = hυ. (equation 1]

The two photon absorption cross-section, σ , is the absorption coefficient per ground state ion per unit volume per incident photon flux and hυ2 is the energy of an emitted fluorescence photon.

In the most general model of the sensor, the two counter- propagating pulses 20 and 25 separately produce background TPAF. Integration of equation 1 ' along the length of the fiber gives the temporal development of the power of the background TPAF, which is: P«(tU = h«2!≡£exp(- %)exp(A%) )- exp(- %)]

- ΔT < t (equation2)

whereNA is the number of photons in pulse 20 and T is the fibre transit time for a propagation speed v . In equation 2, the time origin has been shifted to correspond to the instant at which this pulse 20 meets an oppositely travelling pulse 25 delayed or advanced byΔT .

A similar expression, equation 3, is obtained - for the background TPAF power caused by pulse 25, but detected at the opposite end of the fibre, i.e. along with the TPAF from pulse 20. i \ σNNB v PB OVTPAF = hυ2 -*- - AΔt eχp(" Vx) t1 - eχp(~ % t > O (equation 3

Again, the origin of time is the instant of overlap, and to distinguish between the pulses, the number of photons in the counter-propagating pulse has been identified as NB . The TPAF flash, ΔPTPF(t), produced at the overlap of the counter-propagating pulses is given by the integral of equation 1 along the length of the fibre excited by the second order cross-correlation function of the pulses, i.e.

t > 0 (equation 4;

K is a factor of order unity which depends on the pulse shape. The total TPAF signal detected, P(t)TPAF, is the sum of the components represented by equations 2 (TPAF background due to pulse 20) , 3 (TPAF background due to pulse 25) and 4 (TPAF due to combination of both pulses 20 and 25) . When normalized against the background fluorescence from pulse 20 at t = 0, the instant of overlap, is given by

t≥O (equation 5) The quantity represented by equation 5 is the temporal dependence of the detected fluorescence 45, usually after the pulses have already exited the fiber since the decay times involved are normally longer than the transit time (T « 5μsfor a length of 1km) . The contrast of the TPAF flash from the overlap region relative to the background is maximized in shorter fibers and smaller decay times. This is because the TPAF power is proportional to the number of photons and the number of illuminated ions but is inversely dependent on the decay time. In the case of the TPAF flash from the overlap region, the number of illuminated ions is determined by the durations of the pulses, but since the temperature may be considerably higher than ambient, the decay time is reduced. Conversely, the background TPAF power is not only inversely dependent on the longer decay time throughout the rest of the fiber, but is proportional to the large number of ions encountered during a complete transit. Considerations of this sort determine the type of environment for which the sensor is suited - long or short sensing paths, high or near ambient temperatures. Also since the power of the TPAF flash depends on the number of ions within the overlap region of the fiber 15, longer pulses 20, 25 of lower intensity can be used thus avoiding the complexity of a femtosecond laser system. They also prevent the occurrence of competing nonlinear effects such as self-phase modulation and continuum generation, and stimulated scattering, and are readily available from semiconductor diode lasers 5, 10 synchronized and mutually delayed by electronic means.

Background fluorescence may be minimised by having a higher flux of the non-resonant pulse 25 than pulse 20. Since the magnitude of the TPAF effect is dependant on the total flux at the intersection point 70, this will suppress the background noise relative to the signal. The background may be further minimised by recording a baseline. This can be done by sending only pulse 20 around the fibre optic cable 15 and subtracting this baseline data from the data taken when both pulses 20 and 25 pulses are present. This baseline process may be repeated automatically to allow the sensor to respond rapidly to changes in the environment.

Figure 4 shows the results of a computational model, based on equation 5, of the time development of a normalized fluorescence signal 45 due to counter- propagating pulses 20, 25 overlapping at the mid-point of a fiber 15 of length 50m. The time origin is the instant of overlap, and the fluorescence decay time of the dopant is assumed to be lms, except for the position of overlap where it is 500μs, 250μs and lOOμs due to the local environment. The pulse durations are Ins, the ratio of their photon fluxes is:

% ■ 10 and their wavelengths are such that only the less intense 20 cause TPAF to occur on their own, i.e. the third term in equation 5 is effectively zero.

Typical values for the variable parameters in equations 2-5 are σ«10"52 cm4.s.ion"1.photon"1 , N1 » 2 x 1016mm~3 (which corresponds to a doping concentration of lOOOppm) , A = 4μm2 -» lmm2, T = 5ns -» 5μs (length L = m -» km) , τ and τH = ms -> μs , NB / NA = 1 —> 10, and Δt « lOOps -> Ins (equivalent to a spatial resolution of 20 -> 200mm) . The maximum pulse repetition rate is linked to the decay rate, as the fluorescence must have decayed after excitation before the next pair of pulses 20, 25 arrive. At room temperature, this corresponds to 1-10OkHz depending on the specific combination of dopant and host. For pulses 20, 25 with an average wavelength of say lμm , duration Ins and peak power 1OW, the number of TPAF photons emitted in the flash at the overlap of a single pair of pulses in a fiber of core diameter 5μm is

approximately 106 , determined using equation 4. This corresponds to the generation of > 109 - 10" TPAF photons 45 per second for pulse repetition rates of 1-10OkHz. Even if only 1% of these photons 45 are detected, this is sufficient to allow the fluorescence decay time to be measured rapidly with high precision. For a dopant with a short fluorescence decay time or as the fluorescence decay time decreases with increasing temperature, the pulse repetition frequency may be increased to offset any temperature dependent non-radiative decay mechanism present.

Figure 5 shows the variation in fluoresced light intensity when two pulses 20, 25, both of lOOps duration, are injected into either end of a 50m long, ion doped glass fibre optic cable 15, with no delay between injection of the two pulses 20, 25 (i.e. they meet at the mid point of the fibre optic cable) . The flux of pulse 25 is ten times that of pulse 20 and the wavelength of pulse 20 is greater than that of pulse 25 so that only pulse 20 gives rise to background TPAF by itself. The origin of time is the instant of overlap of the pulses 20, 25. It can be seen from Figure 5 that the intensity of fluoresced light decays exponentially. The rate of decay is proportional to the rate of relaxation of ions from the emitting level 65 back to the terminating level 60. The rate of this relaxation depends on certain properties such as temperature and strain and so the rate of decay in the fluoresced light can be measured and used to determine these properties with reference to calibration data or known properties of the material. For example, in the case of thulium doped silica fibre, the fluorescence lifetime decreases from c.86μs at 100°G to c.lβμs at 1000°C for the 3F4-V5H6 transition at 1.9μm. Alternatively or additionally, the separation or relative intensities of spectral lines in the fluoresced light may be used to determine the required property.

To carry out the required measurement, other factors affecting the TPAF decay rate may be held constant. Alternatively, at least one other fibre measurement apparatus may be arranged in such a way as to measure the non-constant properties affecting TPAF apart from the property that is required. The change in fluoresced light due to other factors may then be subtracted from the total measurement to obtain measurement of the required property.

In addition to temperature and strain, other physical parameters can be measured or determined directly or indirectly using TPAF. For example, fluid flow may be determined indirectly by measuring the temperature change and linking this to fluid flow by thermal conduction. This may be achieved by having the fibre 15 at a different temperature to the fluid and recording the temperature variation with time, and then determining the time taken for the fibre to come into equilibrium with the fluid temperature. Since this depends on the flow of the fluid, a measurement of this flow can be determined. This may be done by comparing the measured data with calibration data obtained using the same fluid and measurement system to obtain the fluid flow.

Electric and magnetic field strengths may be measured through the effect of splitting of the exited energy level 65 by the Zeeman and Stark effects. When the split energy levels are coherently exited, the fluorescence is subjected to temporal modulation with the period of modulation dependant on the splitting. Thus the period of modulation of the fluoresced light can be measured and used to derive the external field, again with reference to calibration data.

To determine the value of a property for a system that is not already fully characterised or to obtain the most accurate determination, a calibration procedure may be carried out. Calibration data for a given set of experimental conditions may be obtained in a variety of ways depending on the property under investigation. For example, to calibrate for temperature, a section of the fibre optic cable 15 may be placed in a suitably large, sealed oven and heated to a selection of temperatures, recording the TPAF decay time at each temperature. In another example, calibration for strain may be made by placing the fibre optic cable 15 in a tensilometer and recording the TPAF decay time at a range of strains.

In order to measure a parameter at a selected measurement point 70, the injection of one input pulse 25 may be delayed relative to the other 20 so that the two pulses intercept at that selected position 70 in the fibre optic cable 15. For example, to make a measurement at a point on the fibre optic cable 15 closer to laser 10 than laser 5, the pulse 25 from laser 10 will be injected after the pulse 20 from laser 5 by a time dependant on the position of the measurement point and the speed of the light pulses in the fibre.

In order to achieve fully distributed measurement, the injection of the two intercepting pulses 20, 25 may be carried out many times with varying delays in injection of the intercepting pulses relative to each other. For example, if one measurement is required at each of two positions along a fibre optic cable, then a pair of pulses 20, 25 are injected with the delay between the pulses being such that they meet at the first measurement point. The intensity of fluoresced light is then recorded until it has decayed to zero before the next pair of pulses is injected. The process is repeated with a second pair of pulses, having a delay between the pulses such that they meet at the second measurement point. The decay of fluoresced light from each measurement point may be compared with calibration or known data to obtain the measurement at that point. This procedure may be repeated many times to monitor the change of the measured property with time. Variations on this example with many pairs of pulses, having varying delays between the pulses in each pair, may be carried out to achieve sensing at any multiple of points, limited only by the length of the fibre optic cable 15 and the length of the pulses 20, 25.

The length of the light pulses 20, 25 may be selected to suit the required application. Since the dimensions of the measurement point 70 depend only on the size of the pulses 20, 25, minimisation of the pulse length can lead to a measurement with very high spatial resolution. For example, if the speed of the light pulses in the fibre optic cable is 2XlO^s"1 then a pulse length of lOOps would be equivalent to a spatial resolution of 20mm. This effect makes the method and apparatus of the present invention particularly suitable for monitoring or detecting localised or small-scale effects such as determining possible fracture locations in structures or detecting localised heating in industrial plant.

Increasing the pulse length can improve the signal strength by increasing the number of photons available for TPAF. This approach would be particularly suitable for measurement over a very long distance and where spatial resolution is less important, as the increased signal compensates for attenuation effects.

To measure temporal changes in a property, measurements at a selected point 70 may be taken repeatedly to determine any changes over time. The frequency of this measurement is limited only by time taken for the fluorescence to decay. As temperature increases, the time taken for the fluorescence to decay decreases, allowing the measurement frequency to also be increased. As an example, for Cr: Sapphire single crystal fibres, the time for the fluorescence to decay is of the order of 3ms at room temperature. This decreases by 12μs per I0C increase in temperature. Similarly, if a rapid measurement frequency is required then a doped optical fibre material having a short decay time may be used.

In practice, a number of pairs of pulses may be used for each measurement to improve the signal to noise. With a decay time of ms, one thousand pairs of counter- propagating pulses would give a resolution of 1 second.

The method and apparatus of the present invention may be used advantageously to take measurements in inaccessible places due to the compact and flexible nature of the fibre optic cables 15 used for measurement. Equally, the method and apparatus of the present invention may be used to take measurements over a long distance or at locations far removed from the sensing apparatus due to the low attenuation and high signal to noise ratio of the present invention.

Since the detection electronics 50 can be widely separated from the measurement site 70 and the fibre optic cables 15 are electrical insulators, the method and apparatus of the present invention is particularly advantageous for taking measurements in areas of high electrical fields whilst minimising electrical interference with the detection apparatus 50. Also, this advantageously allows measurements to be taken in an environment where use of such electrical signals or electronics might be considered dangerous.

As a still further advantage, variation in the relative timing of the injection of each counter propagating pulse 20, 25 in the method of the present invention allows measurement at any point 70 on the fibre optic cable 15, thus resulting in continuously distributed sensing. Also, because the fibre-optic measurement section 15 has no moving parts, is reasonably inert and requires no routine maintenance it can be permanently embedded in structures such as buildings. Furthermore, use of appropriate single crystal fibres 15 can enable the method and apparatus of the present invention to be used to provide distributed measurement at temperatures up to 20000C.

A skilled person will appreciate that variations of the disclosed arrangements are possible without departing from the invention. For example, whilst calibration data is described as being used to determine the value of the parameter that is being measured, it will be appreciated that previously known data, for example data published in journals or books, may be used instead. Accordingly the above description of the specific embodiment is made by way of example only and not for the purposes of limitation. It will be clear to the skilled person that minor modifications may be made without significant changes to the operation described.