WO2011020983A2

WO2011020983A2 - Measurement of fluid concentration

Info

Publication number: WO2011020983A2
Application number: PCT/GB2010/001265
Authority: WO
Inventors: Karl Howlett; Stathis Louridas; Leigh Henebury; Stuart Murray
Original assignee: Marshalls Of Cambridge Aerospace Limited
Priority date: 2009-06-30
Filing date: 2010-06-30
Publication date: 2011-02-24
Also published as: WO2011020983A3

Abstract

A system for operating an opto-luminescent chemical transducer which is adapted to generate, in response to an input of excitation light, an output radiation capable of absorption by an analyte, thereby to enable a concentration level of the analyte to be established, comprises: a source of cyclically-varying incident excitation light directed upon the transducer; a detector adapted to generate an output signal whose amplitude varies with time in accordance with intensity of output radiation generated by the transducer as a result of its opto-luminescence; and a phase detection module for determining phase of the output signal within a cycle of variation of the excitation light.

Description

MEASUREMENT OF FLUID CONCENTRATION

BACKGROUND OF THE INVENTION

1. FIELD OF THE INVENTION

The present invention relates to the monitoring of fluids, such as, for example, the monitoring of oxygen or some other fluid using electromagnetic radiation, such as light in the visible or near-visible spectrum. One application, though by no means the only application, of such a technique is the monitoring of the presence of oxygen within the fuel tank of an aeroplane.

2. DESCRIPTION OF RELATED ART

The sensing of oxygen partial pressure within fuel tanks and mechanisms for inerting such tanks in the event that significant levels of oxygen are detected are disclosed in US2004083793, US2005270525, US2003116679 and US2004035461.

SUMMARY OF THE INVENTION

A first aspect of the present invention provides a method of operating a fluorophore chemical transducer to establish a concentration level of an analyte fluid, comprising the steps of:

directing cyclically-varying incident excitation light onto the transducer, thereby periodically to excite it into opto-luminescence;

generating, using an optical detector, an output signal corresponding to the transducer's opto-luminescence; and

measuring the phase of the output signal in a cycle of variation of the excitation light.

A further aspect of the present invention provides a system for operating a fluorophore chemical transducer to establish a concentration level of an analyte, comprising:

a source of cyclically-varying incident excitation light directed upon the transducer;

an optical detector adapted to generate an output signal whose amplitude varies in accordance with the intensity of light generated as a result of opto-luminescence of the transducer; and an analyser for determining the phase of the output signal within a cycle of variation of the excitation light.

BRIEF DESCRIPTION OF DRAWINGS

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

Fig. 1 is a simplified, schematic representation of elements of a system according to an embodiment of the present invention;

Fig. 2 is a detail of an element of the system shown in Fig. 1 ;

Figs. 3A to 3D are graphs illustrating aspects of operation of the system of Figs. 1 and 2;

Fig. 4 is a graph illustrating the changing response of the transducer with varying oxygen concentration;

Fig. 5 is a graph illustrating a characteristic of a fluorophore chemical transducer in Fig.

1;

Fig. 6 is a schematic illustration of a measuring system for measuring oxygen concentration and temperature according to an embodiment of the present invention;

Fig. 7 is a schematic illustration of phase detection in the system of Fig. 6;

Figs.δA to 8C are graphs illustrating operation of the system of Fig 7;

Fig. 9 is a graph illustrating temperature sensitivity characteristics of the fluorophore transducer;

Figs. 1OA to 1OD are graphs illustrating the characteristics shown in Fig. 9 in more detail;

Fig. 11 is a schematic illustration of a system in a practical context;

Fig. 12 shows a detail of the design of the system of Fig. 11 ; and

Figs. 13A and 13B are illustrations of a further detail of the system of Fig. 11.

DESCRIPTION OF PREFERRED EMBODIMENTS

Referring now to Fig. 1 , a light source, in this example a diode 10, emits light. Light from the diode 10 is incident upon a chemical transducer 20 which comprises a fluorophore.

The light therefore causes the transducer to exhibit opto-luminescent characteristics, absorbing photons emitted from the diode 10 to undergo atomic excitation and, upon relaxation, the emission of lower wavelength photons (the incident light therefore being known as an 'excitation input'). The chemical transducer 20 is located inside a vessel 30 which also contains, or at least has a likelihood of containing, a fluid or 'analyte' 40 whose presence it is desired to detect. Preferably, detection is performed in a manner which enables a quantitative assessment of that quantity of analyte 40 present. Accordingly, the characteristics of the analyte are matched to the opto-luminescent characteristics of the fluorophore, with the result that the presence of the analyte in the vessel acts to 'quench' the opto-luminescence of the transducer. It follows that the extent of opto-luminescence exhibited by the transducer 20 can provide an indication of the quantity of analyte present: the shorter the opto-luminescence lifetime, the greater the quenching and, in turn, therefore, the greater the quantity of analyte 40 which is present. In the present example, the vessel is a fuel tank within a vehicle - typically though by no means exclusively an aircraft - and the object fluid is oxygen gas, since the presence of oxygen in a fuel tank is potentially explosive. The diode 10 is driven by a suitable power supply 200 to provide an excitation input comprising regular bursts of light of wavelength at the appropriate wavelength to initiate excitation of the transducer, say for this application 470nm, whose intensity varies cyclically at a frequency of 40KHz. In one embodiment, this is a supply which provides an amplitude shift-keying (AFK) modulated signal, i.e. a sequence of what is effectively square waves, but modulated sinusoidally. In a preferred embodiment, the 'duty cycle' of the excitation input, that is to say the proportion of a complete cycle over which the excitation input is in a state capable of generating an excitation, is in the region of 20% light source. A duty cycle in this range has been found to provide sufficient excitation to enable appropriate measurement without unduly causing degradation in the form of photo-bleaching of the sensor fluorescing material due to continual excitation and relaxation. This excitation input is transmitted to the transducer 20 via a TX optical fibre 30. The TX fibre 30 couples with a sensor fibre 50, upon the distal end of which the chemical transducer 20 is located. Referring now additionally to Fig. 2, in the present embodiment, the fluorophore in the transducer 20 is provided by a matrix comprising of a ruthenium compound 22 (or, alternatively, a suitable compound sensitive to the analyte being measured) within a sol gel substrate 24, coated onto the distal end 34 of fibre 50. The 470nm wavelength output light transmitted along the transmission fibre 30 passes into the sensor fibre 50 and is then incident upon the transducer 20. The 470nm output light interacts with the ruthenium compound to cause fluorescence as a result firstly of the absorption of 470nm photons to result in the excitation of the ruthenium compound to a higher energy state and, upon its subsequent relaxation, the emission of 620nm wavelength photons. The fluorescence of ruthenium compound and the properties of oxygen are known to be matched such that the presence of particular oxygen molecules which demonstrated to have a quenching effect on the extent of the fluorescence of ruthenium. The term quenching is ordinarily used to refer to a 'reduction' in opto-luminescence; in the present example this is fluorescence. It is understood a variety of 'Quenching' processes can decrease fluorescence intensity of a given substance. Other forms of opto-luminescence, such as phosphorescence, may equally be of utility. In this instance the category for quenching is related to a nonradioactive transfer of energy "fluorescence resonance energy transfer". This process causes the fluorescing material to relax by dissipation of excitation energy (collision) into a solvent such as molecular oxygen (O₂). O₂ is found to be an effective quencher because of its unusual triplet ground state. Collision quenching occurs when molecules of (say) oxygen collide (collide means 'in proximity', typically less than 10nm) with the excited ruthenium compound, whereupon there is an energy transfer - which therefore enables the ruthenium to lose energy without the emission of (in this example) 620nm output photons. Collision quenching is, therefore, one form of quenching that reduces fluorescence by actually inhibiting it. However, since this embodiment of the present invention is concerned with the effect on observable phenomena, when used in the present specification, the term is intended to be interpreted 'macroscopically', so that it additionally encompasses a model where fluorescence actually occurs but a locally- present quencher then attenuates the output photons emitted upon relaxation, which therefore results in a reduction in observable fluorescence. Quenching is therefore intended to encompass any process which operates to 'reduce' opto-luminescence, such as fluorescence, whether this reduction is simply a reduction in the observable fluorescence intensity of a substance or, whether it is a result of mechanisms that operating to inhibit subsequent relaxation (or even initial excitation).

Referring now to Figs. 3, Fig. 3A illustrates a variation with time of the intensity of the light incident upon the transducer when the laser diode 10 is driven to provide an excitation input modulated with a 40Khz on/off signal. Figs. 3B - 3D illustrate the variation with time in observable fluorescence occurring at the transducer as a result of the incident light variation illustrated in Fig. 3A for differing concentrations of O₂: Fig. 3B being the output observed for the lowest concentration of O₂ (i.e. little quenching of the fluorescence due to attenuation of the output fluorescent light), Fig. 3C a higher concentration than Fig. 3B (and therefore more quenching) and Fig. 3D the highest concentration. It will be observed that the amplitude of the output fluorescent light and its lifetime, decreases with increasing concentrations of O₂. In a complication which presents practical difficulties (the consequences of which are addressed subsequently), the fluorescence lifetime similarly decreases with increasing temperature.

It follows that, for a given level of excitation input from the diode, the extent, that is to say the intensity and lifetime, of the fluorescence detected will indicate the concentration of O₂ present; the lower the intensity and lifetime, the greater the quenching and thus the higher the concentration of O₂ present. Emitted light of wavelength 620nm from the fluorescing transducer passes back down the sensor fibre 50 and couples with an RX fibre 60, via which it is then directed onto an RX detector 80, which in the present example is an avalanche photo diode (¹APD') whose voltage output corresponds to the intensity of 620nm light incident thereon. Accordingly, the amplitude of the voltage signal from the RX detector 80 corresponds to the intensity of the incident 620nm light and, therefore, the concentration of O₂ in the tank. Measurement of amplitude value will therefore provide a measure of the extent of quenching and thus O₂ concentration. However, investigations conducted when seeking to employ an amplitude measurement as a way of generating a meaningful, accurate value of O₂ concentration reveal that low signal to noise ratios result in reduced reliability and accuracy. This is, in part, a result of the desire to measure, accurately, relatively low O₂ concentrations, coupled with such factors as microbend loss and attenuation in the fibres. These latter factors are particularly acute for fibres which are installed in constrained environments and require complex routing paths such as those found in an aircraft, which can cause losses of up to >60% of light initially transmitted along a fibre.

One aspect of an embodiment of the present invention provides an appreciation on the part of the inventors that, although, for a given intensity of input light via the TX fibre 30 from laser diode 10, the voltage amplitude signal from the RX detector 80 correlates directly with the concentration of O₂ in the tank, there exists an alternative parameter which enables measurement of O₂ concentration. As can be seen from Figs 3, quenching of fluorescence results not only in a reduction in output fluorescing intensity but also operates to increase its rate of decay, or shorten its 'lifetime' and thus produce a phase shift between the trailing edge of excitation input and that of the RX output at detector 80. This is illustrated in the graph of Fig.4 which shows the change in decay lifetime across the time period of a single pulse, with varying O₂ concentration (a different concentration for each of the three curves illustrated).

The extent of the variation in phase shift between the RX output pulses representing the variation in output amplitude of the transducer and the pulse of the subsequent excitation input - in other words the magnitude of the change in decay lifetime - has been found to vary with varying O₂ concentration substantially as illustrated in Fig. 5. Beneficially, it can be seen that, at lower O₂ concentrations, a given change in O₂ concentration causes the phase angle to alter more sharply, thus providing an opportunity for increasing precision at lower O₂ levels. Embodiments of the present invention relating to measurement of O₂ concentration in both civilian and military aircraft fuel tanks, are principally concerned with O₂ concentrations above 4%, more preferably above 7% and preferably between 9 and 12%. At concentrations in the region of 9%, it has been found that a change in O₂ concentration of approximately 1% causes a phase shift of the order of 0.75° to 1.25° The relative phase of the input and output pulses is a parameter which varies with varying concentration of O₂ and can, therefore, be used to measure O₂ concentration, once other factors causing a variation in the decay lifetime (and therefore the phase shift), such as temperature, have been accounted for.

By contrast, measurement of amplitude variation as a mechanism to reveal O₂ concentration requires the attribution of a precise, numerical value which signifies amplitude and, when comparing two amplitudes of low value, the need to discriminate numerically between two such low values. It is this which renders the use of amplitude measurement vulnerable to inaccuracies which arise as a result of the noise and loss phenomena discussed above. All that is required in order to measure a phase shift of the kind which the graphs of Figs 3B to 3D illustrate occurs with varying O₂ concentration, however, is to establish the mere presence of an amplitude in excess of some predetermined threshold or gating value (i.e. a value at which, logically, a signal whose phase is to be measured is deemed to exist). Embodiments of the present invention therefore provide measurement of quenching and thereby O₂ concentration using phase rather than amplitude. Referring initially to Fig. 6, an oxygen sensing system 100 is configured as part of an overall sensing system for both temperature and O₂ (which therefore also permits sensing of other environmental parameters in addition to, or substitution of O₂ or temperature. The reason for this that, as referred to previously, a further aspect of the present invention lies in an appreciation that, even where O₂ concentration remains constant, a change in the intensity and lifetime of the output fluorescent light occurs with changing temperature. The system 100 includes an RX detector provided by an avalanche photodiode 80, which detects flourescence from the fluorescence transducer 20. The RX detector is connected, via a multiplexer 70 to a phase detection module 82. The module 82 receives a phase reference input 84 and, on the basis of this reference input and the RX output of detector 80, generates two digital output signals XRX and XRY which are used in trigonometric calculation performed by a micro-controller 92 to generate a phase angle of the output RX detector relative to the excitation input. The phase angle is converted to an output indicating the O₂ concentration using suitable algorithms linking quenching to concentration. Alternatively, where sufficient processing and memory are available, empirically-derived lookup tables linking phase angle to concentration (in a given environmental context) may be employed.

Referring now to Figs. 6 and 7, the phase detector module 82 receives a reference signal 102 from the CPLD 90, this being the reference used for the signal generator 200 which drives the LED 10 to produce the 40Khz modulated excitation input. The output of the signal generator 200 is passed, via a phase control module 104, to a pair of phase- sensitive detectors 110, 112. The phase detector 110, receives an unconditioned signal, the 'X' Reference signal; the phase detector 112 receives a signal via a phase-shift module 116, which shifts the phase of its output back by 90" to produce a 'Y' reference signal. The 'X' and 'Y' reference signals therefore have a quadrature relationship relative to each other. The phase detection module 82 additionally receives an RX input from the detector 80. This passes through a band pass filter 120 and is then sent to each of the phase detectors 110, 112.

Each phase detector 110, 112, in conjunction with its respective signal conditioning module 140, 142 is adapted to perform as number of functions upon the various signals it receives. The DC component of, respectively, the Ref X and Ref Y reference signals, and the RX signal is removed. In addition, all three signals are effectively 'gated', for example using threshold detectors, to turn them into what are, effectively, binary signals of equal amplitude (i.e. above a given amplitude threshold a signal is a logical "1" and below that threshold it is a logical "-1" - with no interstitial values allowed). Each phase detector 110, 112 and its associated signal conditioning circuit 140, 142 generates an output, XRx and Yrx respectively, being the product of X and Rx or Y and Rx as the case may be, each of which is then averaged over many cycles, and the average value is converted to a digital value by an A/D converter (not shown). The ultimate output of the phase detection module is, therefore, digital values of XRx and YRx averaged over a single cycle, and which are subsequently used in the micro-controller 92 to generate a phase angle.

Referring now to Figs. 8A to 8C, three examples are shown of the manner of processing and using the signals to generate a phase angle. Referring to Fig. 8A, it can be seen that reference signal Ref X leads Ref Y by a phase angle of 90^°. The output signal RX from the Rx detector is, in this example, entirely in phase with Ref Y. The product of Ref X and RX, XRx, is thus an output which has twice the frequency of Ref X; while YRx₁ which is the product of RX and Ref Y produces a constant DC output. The average of XRx is zero; the average of YRx is 1. The phase angle of the RX is given by the ArcTan of the two averages YRx/XRx which, (for the purpose of simple illustration, skating over difficulties which arise as a result of division by zero being forbidden) is °°, the ArcTan of which gives a phase angle of 90'. Fig. 8B shows the inverse - where the phase angle is zero. Referring to Fig. 8C, where RX has a phase of 45 ' relative to RX it can be seen that the products XRx and YRx each have an average value per cycle of 6/2, i.e. 3. This means that phase angle, being the arctan of 3/3, or 1, is 45^°, which is indeed the case.

These examples illustrate the principle of how a phase shift in RX is established numerically. In these examples, however, it can be seen that, for the purposes of simplicity of illustration, RX has been illustrated as having a constant lifetime. In practice, the shift in the phase of the trailing edge of the RX signal which it is sought to measure occurs as a result of a decrease in the lifetime of the RX signal; though the manner of processing the signal to generate a phase shift value remains the same as illustrated in connection with Figs. 8 above. Compensation of variations in decay lifetime which arise as a result of changes in temperature is achieved by observing the changes in the same parameter as that used to measure O₂ concentration, namely the phase shift which results from changes in the fluorescence lifetime. In order to isolate any fluorescence lifetime changes which arise from temperature from those which may arise as a result of changes in O₂ concentration - the parameter which it is desired to measure - a fluorophore transducer in the form of a ruthenium compound coated tip fibre is encapsulated within a hermetically-sealed environment which is defined as 'fully quenched' which, in the present example where, it is sought to establish O₂ concentration, is an O₂ concentration of 20.8%, equal to that in normal air and, therefore, the maximum concentration that is likely ever to be present in an aircraft fuel tank. Because the container is hermetically sealed, it follows that any change in fluorescence lifetime of the sensor within the container will arise only due to temperature changes, since the concentration of O₂ is held constant. The temperature sensing transducer is connected to a similar detection and processing system as that provided for the O₂ sensing and illustrated with reference to Figs. 6 and 7 but which, for brevity, will not be illustrated further. The value of phase shift derived as a result of changes temperature can then be used to correct the O₂ sensing output value to derive a pure value for O₂ concentration. Theoretically, this can be done by simple subtraction of the phase shift value derived from the temperature sensing transducer (this being constrained by exposure to a constant concentration of O₂ and therefore representing a 'pure' temperature-base phase shift) from that derived from the O₂ sensing transducer, the phase shift output from which is as a result of both quenching and phase shift phenomena. However, it has been found that the difference in the values of phase shift derived from the temperature sensing transducer and that of the O₂ sensing transducer (which, in fact, is in effect actually an O₂ AND temperature sensing transducer) is not linear with different values of temperature. That is to say that the extent of the influence of temperature on the quenching phenomenon changes for changing O₂ concentrations at different temperatures, meaning that a simple, linear subtraction of one phase shift value from another will not yield an accurate result for the value of the phase shift which occurs solely as a result of quenching.

This non-linear response is accounted for by calibration which then enables the values of phase shift derived from the temperature sensing transducer and that derived from the O₂ sensing transducer to be used to obtain a single, temperature-compensated value of phase shift attributable solely to quenching.

Referring now to Fig. 9, the response to variation in O₂ concentration at three different temperatures is illustrated. Due to variations in transducer response It is necessary to calibrate each before any accurate measurement can be made (i.e. exposure to differing concentrations of O₂ at a set of differing temperatures) to generate a set of response curves for differing temperatures. The initial calibration curve is generated by measurements made of its fluorescing response over a range of oxygen concentrations, all at a temperature T₁. Based on empirical evidence, an assumption is then made that the transducer will respond in accordance with this characteristic curve at different temperatures. Accordingly, further curves are then generated at different ambient temperatures but based on significantly fewer readings of operating lifetime. Interpolation between calibration points then enables the phase output to be adjusted to account for temperature changes.

The phase output of the O₂ sensor is a function of both a change in the O₂ concentration (causing a change in the extent of quenching of the transducer), also expressed as partial pressure (ppθ₂) and Temperature (T). To calibrate the sensor a 'surface', that is to say a three dimensional path which passes through known values of temperature and ppθ₂ has to be fitted; moreover, to be of utility in establishing, for a known value of temperature, say, the resulting change in value of ppθ₂that path is preferably capable of expression with precision - such as by mathematical expression. To fit a surface to (x,y) data points a two stage process is used. The first stage uses a polynomial curve to fit the oxygen phase shift against ppθ₂ for each temperature range measured. A 3^rd order polynomial gives the best fit and lead to the following equation: f(ppO₂ ,τ )= a (T) (ppO₂)³ + b (τ )(ppO₂)² + c (T)(ppO₂)+ d (T ) Where:

f() .oxygen sensor output, which is a function of the partial pressure of Oxygen and temperature -, expressed as the phase-shift derived from the output of the oxygen sensing probe.

O₃... α_o() : Polynomial coefficients which are a function of temperature. b₃...b₀() : Polynomial coefficients which are a function of temperature.

C₃... C₀Q : Polynomial coefficients which are a function of temperature. d₃... do() : Polynomial coefficients which are a function of temperature. ppθ₂ : Partial pressure of Oxygen - the parameter whose value it is sought to establish

T₀...T₃ : Temperature data points in ⁰C

Values of ppθ₂ and f are obtained for different temperature ranges in laboratory conditions by cycling:

a), at a constant temperature, different, known values of ppθ₂ and

b) at the mean of three different known values of temperature, different known values of PpO₂. This yields, for each temperature range a unique equation: f(ppO₂J₀) = α₀(T₀)(ppO₂y +b_Q(T₀)(pp0₂)² + c₀{T₀)(pp0₂)+ J₀(T₀)

/(PPO₂ J₁) = O₁[T₁) (ppOj+b, (T_x)(pp0₂ )² + c_x (T_x)(ppO₂)+ d_x(T_x)

/(PPO₂J₁) = 0₂(T₂)(pp0₂γ + b₂(T₂)(pp0₂)² + c₂(T₂)(pp0₂)+ d₂(T₂)

/(pp0₂J₃)= α₃(T₃)(pp0₂Y +b₃(T₃)(_PP0₂)² + c₃(T₃)(pp0₂)+ d₃(T₃)

The subsequent stage in the process then uses a polynomial curve to fit each coefficient a, b, c, d separately against the temperature data points. A 3^rd order polynomial gives the best fit and leads to the following form: α(τ) = k₃ T' +k₂ T² +k_xT + k₀

b(τ) = k₁ Tⁱ +k₆ T² + k_sT + k₄

c(τ) = k_n T³ +k_X0 T² +k₉T + k_s

d(τ)= k_X5 T' + k_H T² + k_X3T + k_X2

Where:

k₁₅...ko() : Polynomial coefficients.

To demonstrate the calibration process a polynomial in step one is fitted against a set of temperature data points. Whereas Fig 9 illustrates the polynomial curves fitted to represent discrete temperatures. The second step of the calibration process fits the polynomials against a, b, c, and d coefficients and this is illustrated in Figs 10a, 10b, 10c and 10d respectively. Once established for each temperature range, the coefficients can then be used to generate the 'surface' - i.e. a three dimensional map, expressed mathematically, of the variation in f with variation in the value of ppθ₂ within a given range of Temperature T. Each such surface then enables a value of ppθ₂ to be derived for values of T and f at which no laboratory measurement had been made by what amounts, in effect, to 'interpolation' using the surface to obtain a value for ppθ₂ on the basis of measured values of T and f which, almost inevitably, will not correspond to any of the values which were obtained under laboratory conditions to create the surface initially.

The output value of ppθ_2l once generated, will then preferably be corrected by using ideal gas law equations for variations in pressure, though this can take place subsequently using a simple pressure measurement on the basis of a look-up table, for example.

In this application the calibration data is stored within the micro-controller 92 and used to calculate the actual O₂ concentration based on inputs from the O₂, temperature and pressure transducers. Similarly the calibration data could be used to generate a series of lookup tables so that an O₂ measurement could be made by cross-referencing the transducer outputs rather than performing calculations continuously. The use of phase to measure (in this example) O₂ concentration has a number of advantages. Firstly, measurement of amplitude requires a precise value to be attributed to a signal parameter that is small and whose variations are small compared to its absolute value and noise. This is, innately, apt to produce inaccuracies. By contrast, the measurement of phase requires only a precise assessment of the instant in time at which a step change in amplitude value occurs. While it is true that such a step change is small in magnitude (since, as we have said above, the absolute value of amplitude is small), provided that the step change is greater than a predetermined level, the value of the amplitude is of no consequence. A further factor which we have established militates against the use of amplitude is that the transducer's performance deteriorates over time, resulting in a continuing reduction in the extent of output fluorescence for any given excitation input. Were a measurement of amplitude required this performance deterioration would have to be corrected for in every measurement, while the reducing amplitude would further also reduce the output signal amplitude for any given input level and, consequently, therefore further reduce the ratio of signal to noise. Such changes in performance can, when using phase as a basis for measurement of the extent of quenching, be accounted for more easily by suitable adjustment of threshold gating values and a restriction on the operating life of the transducer. By contrast, it would be significantly more complicated to take account of performance deterioration to a sufficiently precise degree to enable the generation of a sufficiently precise amplitude value to be employed as a basis for measuring the extent of any quenching.

Measurement of temperature using the same sensing mechanism which is used to sense changes in decay lifetime due to changes in O₂ concentration provides the advantages of intrinsic safety, as electrical connections are not required, and of simplicity of the interfacing system, which can simply be a duplicate of the oxygen side. Further, it enables measurement of temperature which avoids the insertion of any electrical elements into the fuel tank (such as would be required were, for example, a thermocouple to be used) and also enables, once calibration has occurred, as subtraction of the phase outputs derived from the temperature sensing and the O₂sensing transducer to be used.

Aspects of a practical system used on an aircraft will now be described with reference to Fig 11. Referring now to Fig. 11 in conjunction with Fig. 1, a system schematically- illustrated for use in an aircraft wing which contains a fuel tank 300 is illustrated. Four sensing fibres 50, each tipped with a fluorophore transducer of a ruthenium compound held in a matrix of sol gel project, via corresponding apertures (not shown in Fig. 10), into the fuel tank in order that the transducer of each fibre is exposed to the interior of the tank. The fibres 50 are spaced apart within a single tank to seek to ensure that local, varying concentrations of O₂ (which may still be sufficient to present a combustion risk) can be detected. Each sensing fibre 50 is coupled to two fibres: a TX fibre, which is connected at its remote end to a laser diode and therefore provides the source of 470nm excitation light at varying increasing intensities thus compensating for decay in fluorescence emission strength due to photo-bleaching effects; and an RX fibre, along which light generated by the fluorescence of the ruthenium fluorophore in the transducer 20 is conducted towards the photo detector 70. Each of the four sensor fibres 50 is polled sequentially at intervals of around 10 seconds (since, in this example, O₂ concentration is unlikely to change more rapidly).

As described previously, temperature is also sensed using a transducer in the form of a ruthenium compound coated fibre tip which runs parallel with the oxygen sensing transducer - as illustrated schematically in Fig. 11. Referring now to Fig. 12, the two fibres, that is to say the O₂ sensing fibre 50 and the temperature sensing fibre 50' are preferably both housed in a single, substantially cylindrical housing 410, with the fibres running parallel with each other and the axis A of the housing 410. The temperature sensing fibre 50' is exposed to a compartment 420 which is hermetically sealed, of air to enable 'full quenching¹ as described above. The exposed end 5OA of the O₂ sensing fibre 50 is chamfered flush with the housing at an angle of 45° to the axis A, this facilitating a wider local diffusion of the light emitted from the O₂ sensing fibre 50. Preferably, the housing 410 is made of a conductor, such as copper, since this helps ensure that both fibres are at the same temperature - which will therefore mean that the temperature sensed by the temperature sensor fibre 50' is as close as possible to that experienced by the O₂ sensing fibre 50; in an alternative embodiment, the housing will be made of stainless steel which has been proven to survive in harsh environments, such as those seen in aircraft fuel tanks.

Phase is, in effect, a timing measurement and in this embodiment what is being measured is the time interval between the excitation input and the increase in output fluorescence intensity of the transducer. As referred to briefly above, this therefore requires a consideration of certain practical issues relating to differing optical path lengths. Because the reference signal is derived from the signal generator, a correction to the timing is required in order to compensate for the time required for the excitation input to travel outwardly along the fibre and for the fluorescing output to travel back along the fibre. This will simply be Uc where L is the total length of fibre and c is the speed of light. Additionally, the known phenomenon of jitter, which operates to create an inherent error in the phase of the signal which drives the laser diode 10 must be accounted for. The output of the above-described measurements is a value for the concentration, or partial pressure of O₂ within a fuel tank. Where it is sought to establish the propensity for fuel ignition as a result of the presence of O₂, the key parameter is the absolute amount of O₂ - and whether this is sufficient to support combustion. Accordingly, it will be necessary to covert a partial pressure or concentration value to an absolute value - a trivial exercise given knowledge of the size of the fuel tank and the quantity of fuel present in it (which provides an indication on the volume of 'vacant' space). Preferably, along with temperature, total pressure within the fuel tank is also measured as one of the environmental parameters, which may then be used to assist with calculating the total O₂ percentage within the tank.

Evidently, no two transducers are entirely alike. Variations in the concentration of (in this example) ruthenium and the manner of its dispersion within the sol gel matrix - each of which will operate to alter the fluorescing response of each transducer to a standard excitation input - are inevitable from one transducer to another. Accordingly, it is necessary to calibrate each transducer individually before installation and use. Given the variation in fluorescence, it is clearly important when selecting the threshold level at which effectively to gate the RX output signal (i.e. the level at which that signal then logically assumes a positive binary value) to ensure that this continues to occur as oxygen concentration increases and signal detector will continue to trigger at higher concentrations of O₂ (i.e. at lower output signal levels from RX detector 80). This is further complicated by the fact that the extent of the output response of a transducer deteriorates during its lifetime due to photo-bleaching effects. Thus, the amplitude of RX reduces during the lifetime and, in order to avoid introducing errors as a result of this, account must be taken of this phenomenon. This is achieved by a continual monitoring of the absolute amplitude of the RX signal over time and, as this decreases steadily, making corresponding adjustments in the gain of processing and/or signal conditioning circuitry to compensate for this. It will be apparent that various of the methods and systems disclosed in the present application require computation in order, for example, to derive a value of ppθ₂ from measured values of f and T. These computations can be performed by a suitable computer, having random access memory, a central processing unit, where appropriate suitable data storage, networking cards and possibly other peripheral elements as desired, all interconnected by a suitable bus or buses.

As mentioned above, because the signal level for fluorescence from the transducer is extremely low, it is evidently desirable to ensure that as much signal as possible is transmitted to the detector 70. Referring now to Figs. 13A and 13B, conventionally, as illustrated in Fig. 13A, when two fibres are coupled to a single fibre, the coupling mechanism will operate to centre the pair of fibres relative to the single fibre. The consequences of such an arrangement are that equal proportions of light are lost in transmitting to or from each of the pair of fibres to the single, larger fibre. According to a further, independent aspect of the present invention, coupling is performed differently for each fibre of the pair. Referring to Fig. 13B, the RX fibre is positioned so that its coupling end lies entirely in register with, that is to say within the periphery of the sensor fibre 50. A consequence of this is that the coupling end of the TX fibre is substantially displaced off centre with respect to the sensor fibre 50. The consequences of this arrangement are that significantly less light is lost in coupling from the sensor fibre 50 to the RX fibre; but a significantly greater proportion of light is lost in coupling from the TX fibre to the sensor fibre 50. Such an arrangement is beneficial because losses of light in coupling from the TX fibre to the sensor fibre 50 can be compensated for by a simple increase in the output power of the laser diode 10 which therefore result in the same absolute power level of excitation input light. By contrast, light coupled from the sensor fibre 50 to the RX fibre is entirely 'signal' and therefore cannot be boosted. Accordingly, a coupling arrangement which preferentially preserves this signal (by contrast to the input signal which can be boosted by an increase in input power) is advantageous and a further aspect of the present invention comprises an optical coupling between a bidirectional optical fibre and two unidirectional fibres, one of which is an input fibre and the other an output fibre, the diameter of the bidirectional fibre is smaller than the sums of the diameters of the unidirectional fibres and wherein the ends of the unidirectional fibres are positioned asymmetrically with respect to the end of the bidirectional fibre, with the perimeter of the end of the output unidirectional fibre lying more within the perimeter of the input fibre. Preferably, where input power and relative dimensions permit, the perimeter of the end of the output fibre will lie entirely within the perimeter of the bidirectional fibre.

Claims

1. A system for operating an opto-luminescent chemical transducer which is adapted to generate, in response to an input of excitation light, an output radiation capable of absorption by an analyte, thereby to enable a concentration level of the analyte to be established, the system comprising:

a detector adapted to generate an output signal whose amplitude varies with time in accordance with intensity of output radiation generated by the transducer as a result of its opto-luminescence; and

a phase detection module for determining phase of the output signal within a cycle of variation of the excitation light.

2. A system according to claim 1, wherein the source of cyclically-varying incident excitation light includes a diode coupled to an optical fibre.

3. A system according to claim 2 wherein the transducer is located on a distal end of the optical fibre.

4. A system according to claim 1 comprising an optical fibre for transmitting light resulting from fluorescence of the transducer to the detector.

5. A system according to claim 1 wherein the opto-luminescence of the transducer is fluorescence.

6. A system according to claim 1 wherein the phase detection module generates first and second output values, each output value being generated by deriving the product of: (a) a reference signal which has a phase and frequency corresponding to cyclic variation of incident light; and (b) an RX signal having a phase and frequency corresponding to cyclic variation of the output signal.

7. A system according to claim 6 wherein the first output value is derived using a reference signal having a first phase angle and the second output value is derived using a reference signal having a phase angle offset by a predetermined angle relative to the reference signal used in generating the first output value.

8. A system according to claim 7 wherein each of the first and second output values is the average value of the product of the respective reference signals and the RX signal over a single cycle.

9. A system according to claim 8 wherein a phase angle is derived from the ratio of the first and second output values.

10. A system according to claim 1 wherein the transducer is provided by ruthenium- coated optical fibre.

11. A system according to claim 10 wherein the incident excitation light is conveyed to the transducer along an optical fibre and light emitted from the transducer as a result of opto-luminescence is directed to the optical detector along a fibre.

12. A system according to claim 11 as dependent upon claim 9, wherein the reference signal is derived directly from a signal used to drive a light source to generate the cyclic variation in incident light and wherein the phase of the signal from the output detector is corrected by a time related to L/c where L is the total length of fibre travelled by incident and opto-luminescent-generated light and c is the speed of light.

13. A system according to claim 1 , wherein amplitude and lifetime of the transducer's detectable output radiation vary in dependence upon concentration levels of analyte and transducer temperature, the system further comprising a further transducer located within a constant concentration of analyte.

14. A system according to claim 13 further comprising a detector adapted to generate an output signal whose amplitude varies with time in accordance with intensity of output radiation generated by the further transducer as a result of its opto- luminescence; and a phase detection module for determining phase of the output signal within a cycle of variation of the excitation light.

15. A system according to claim 13 wherein the transducer and further transducer are located within a single, conducting housing.

16. A system according to claim 13 further comprising a temperature interpolation module adapted to generate, for a given phase value, a value of temperature of the further transducer.

17. A method of operating a fluorophore chemical transducer to establish a concentration level of an analyte fluid, comprising the steps of:

directing cyclical Iy- vary ing incident excitation light onto the transducer, thereby periodically to excite it into opto-luminescence;

18. A method according to claim 17 wherein the analyte fluid quenches opto- luminescence of the transducer, with quenching magnitude being dependent upon the concentration of the analyte fluid; and wherein the phase angle of output signal increases with increased quenching.

19. A method according to claim 17 wherein the cyclically varying incident excitation light is directed onto the transducer using at least one optical fibre.

20. A method according to claim 17 wherein the step of measuring the phase comprises the step of generating first and second output values, each output value being generated by deriving the product of: (a) a reference signal which has a phase and frequency corresponding to cyclic variation of incident light; and (b) an RX signal having a phase and frequency corresponding to cyclic variation of the output signal.

21. A method according to claim 20 wherein the first output value is derived using a reference signal having a first phase angle and the second output value is derived using a reference signal having a phase angle offset by a predetermined angle relative to the reference signal used in generating the first output value.

22. A method according to claim 21 wherein each of the first and second output values is the average value of the product of the respective reference signals and the RX signal over a single cycle.

23. A method according to claim 24 comprising the step of deriving the phase angle from the ratio of the first and second output values.

24. A method according to claim 23 further comprising the steps of:

deriving the reference signal directly from a signal used to drive a light source to generate the cyclic variation in incident light; and wherein

the phase of the signal from the output detector is corrected by a time related to L/c where L is the total length of fibre travelled by incident and opto-luminescent- generated light and c is the speed of light.