WO2004044563A1

WO2004044563A1 - A method of analysing a sample in order to evaluate at least one property of the sample

Info

Publication number: WO2004044563A1
Application number: PCT/GB2003/004871
Authority: WO
Inventors: Carlos Alberto Rega; Christopher James Lloyd; David John Clarke
Original assignee: Multiplex Photonics Limited
Priority date: 2002-11-12
Filing date: 2003-11-11
Publication date: 2004-05-27
Also published as: GB0226356D0; AU2003301993A1

Abstract

A method of analysing a sample in order to evaluate at least one property of the sample, which method comprises: (a) excitating the sample by excitation means; (b) detecting emission from the sample by detector means; (c) signal processing a data stream in a time or frequency domain; and (d) analysing the processed signal. The method may include the step of comparing the detected emission with a predefined threshold and producing a binary signal when the threshold is exceeded.

Description

A METHOD OF ANALYSING A SAMPLE IN ORDER TO EVALUATE AT LEAST ONE PROPERTY OF THE SAMPLE

Field of the Invention

This invention relates to a method of analysing a sample in order to evaluate at least one property of the sample. The invention may be used to model, measure or characterise a sample or optimise design or production of a sample. The invention may be of particular use in the field of materials analysis and may be of specific use in the design, layout optimisation and production of logic circuitry and materials used to produce logic circuitry. The invention may allow deconvolution of signals that may also be of use in assays that are used to identify analytes.

Background of the Invention

In materials analysis there is an increasing requirement to analyse low quantities of impurities or dopants where conventional equipment cannot provide the required sensitivity. There is an increasing requirement to analyse with increased background noise, and the lack of specificity of coventional equipment is a disadvantage. There is an increasing requirement to analyse with increased background noise, and the lack of specificity of conventional equipment is a disadvantage. There is an increasing requirement to analyse samples more rapidly and the time taken to analyse by conventional equipment is a disadvantage. The signal from a sample may be a mixture of signals from multiple species and the inability of conventional equipment to deconvolute the signals is a disadvantage.

In the analysis of mixed signals in materials the analysis may be an ill-conditioned problem such that the problem is not solvable by conventional mathematics. The analysis may be an ill-conditioned and non-convergent problem such that the problem is not solvable by conventional mathematics.

In semiconductor analysis there is a requirement for real-time processing and the limitations of storage and offline analysis are a disadvantage.

In the design and production of logic circuits there is a requirement to operate at smaller aspect ratios and typically to have more transistors on each substrate. The limitations of yield of conventionally designed circuits is a disadvantage as many circuits may be of no use. The reduced aspect ratio increases heating and the limitations of the design software in modelling of heating is a disadvantage. The reduced aspect ratio may increase cross talk and the limitations of conventional software in modelling cross talk is a disadvantage. The cross talk may be a function of clock frequency and may only occur in specific switching stages of transistors and the inability of test equipment to analyse when and why cross talk occurs is a disadvantage.

In the design and production of logic circuits there is also a requirement to operate at higher frequencies. The increase in frequency limits the usefulness of previous models of how logic parts will behave and this is a disadvantage. The increase in frequency may lead to increased cross talk which conventional technology may be unable to model or measure and this gives a disadvantage. The increased frequency increases heating and which conventional technology may be unable to model or measure and this gives a disadvantage.

In the design and production of logic circuits there is also a requirement to use test ports. One such test port is known as JTAG. Test ports have known disadvantages. The test circuit consumes power which is a disadvantage. The test circuits consume transistors which is a disadvantage. The test port consumes input and output legs which is a disadvantage. The test area increases the time taken to design a circuit which is a disadvantage. The test area constrains the circuit design to a design which can be tested and this is a disadvantage. The test area makes de-engineering easier and facilitates theft of intellectual property which is a disadvantage.

In addition, the concept of testing a circuit using a circuit produced at the same time and in the same material may produce a test circuit which wrongly suggests that a circuit is operating correctly or incorrectly due to a flaw within the test area.

In areas including the design and production of logic circuits there is a requirement to operate with lower power consumption. One means of reducing power is producing circuits which do not require an oscillator, but the inability of modern equipment to properly model, measure or analyse clock-less circuits is a disadvantage. To produce faster or low power circuits, asynchronous designs may be used but these give the same disadvantages as those defined for clock-less circuits. In the testing of integrated circuits it is beneficial to carry out non- invasive and non-coupling measurements such that the circuit is unaffected by the test equipment. Most conventional measurements are invasive at this is a disadvantage. Methods of measurement based on electron microscopy require an electron stream to interact with the material and are thus invasive and this gives a disadvantage. Methods of measurement based of fluorescence require an excitation beam to be directed at the sample and thus fluorescence is invasive. Methods based on refractive index difference require a light beam to be passed through or reflected from the sample and this is invasive. Methods of electro-luminescence require an electrical tap be applied to the material and this is invasive. Methods that are sometimes as non-invasive such as analysis of capacitance effects in probes close to the surface of the sample interact with the fields on and within the sample and thus may be considered invasive when dealing with high frequencies or low voltage circuits.

In material testing and semiconductor testing there is a requirement for high bandwidth measurements and such measurements when requiring an electrical tap are bandwidth limited which is a disadvantage. In semiconductor testing there is also a requirement to find causality between two or more pixels or transistors to evaluate errors such as cross-talk as well as functionality. Where measurement occurs to an external electrical tap as with conventional equipment, then such measurements may be difficult or impractical to make and this is a disadvantage. In semiconductor testing there is also a requirement to find timing between pixels or transistors and where all measurements are made relative to an external electrical tap such measurements may be difficult or impractical to make and this is a disadvantage.

Where multiple pixels are analysed, there is a technique known where one time correlated single photon counting board is used to measure time of arrival and a position sensitive photon multiplier is used to define the spatial position and the information is stored for offline analysis. The requirement to store data is a disadvantage as this disallows very sparse errors to be evaluated. The requirement to store data is a disadvantage as this disallows real-time monitoring and the immediate reduction, increase or repeat of an experimental duration due to the data analysed. The requirement for a position sensitive detector is a disadvantage as this has low spatial resolution. The requirement for a position sensitive detector is a disadvantage as a large detector is required and this increases noise. The requirement for a position sensitive detector is a disadvantage as the bandwidth of the detector is reduced.

Whilst most logic chips currently operate with electrical pulses it is known in the art that chips operating with optical pulses will be beneficial. It is a disadvantage that the conventional test equipment requires an electrical tap as this limits the equipment's direct use to electrical circuits.

Where errors are sparse, conventional instruments often have difficulty in analysis as the error is lost beneath a noise floor and this is a disadvantage. In conventional analysis of multiple pixels the electrical taps provide a start pulse and the pixel defines a stop pulse. It is not possible to define one pixel to be a start pulse and another pixel to be a stop pulse and this is a disadvantage as timing is relative to an external tap which leads to time consuming and inaccurate measurements.

A method of signal analysis known as digital photon correlation has the disadvantage that for pseudo-random data, the processor must analyse a signal line continuously and thus can only analyse a single pixel and thus the method is unsuited to multiple pixel analysis.

The method of signal analysis known as digital photon correlation has the disadvantage that it is calculation intensive and complex to produce a device operating at a sampling frequency significantly greater than 100MHz.

It is known in the art that photon detector systems typically can only measure the presence or lack of a photon event and not the number of photons emitted such that multiple photon emissions are ignored in most test equipment. Ignoring multiple photon emissions loses information and may cause bias and is a disadvantage.

It is known in the art that when using a multiple pixel detector only one photon may be detected on one pixel at any one time and that a dead time follows each photon event. A multiple pixel detector will ignore photons on other pixels, which causes information loss and may cause bias and is a disadvantage.

In production environments it is beneficial to have rapid methods of measurement to optimise layout of a design. The lack of suitable tools particularly for circuit layout is a disadvantage. In production environments, it is beneficial to have rapid methods of measurement to check and control the process line. The lack of suitable tools, particularly in circuit production is a disadvantage. In manufacturing, it is beneficial to be able to fully test each device or a larger number of devices on each batch and to sort according to quality and the lack of suitable tools, particularly in circuit production is a disadvantage.

In many instances, assays must be performed to ascertain the presence or lack of one or more analytes. Assays may be carried out for reasons that include but are not limited to the detection of impurities in food water and air, the detection of viruses, the detection of bacteria, the detection of a property or function of DNA, RNA, proteins, antibodies and the detection of failure of a bodily function. Often the sample to be analysed is dirty and may produce erroneous signals. Often the assay is biochemical and relies on a natural product for the test. One example of such biochemical materials are antibodies which bind to predefined materials and may be used to bind analytes. One problem associated with antibodies is non-specific binding.

One problem of an assay is that a baseline analysis may be required as a test may never give a null reading and readings may vary depending on the age of the assay and the environment of the assay. A further problem with assays is a triggered positive test may be required as the magnitude of the signal may be affected by the age of the chemicals and impurities in the sample, leading to inaccurate quantification. A further problem of assays is that the assay only tests for a material and not a cause. By way of example cyanide may be detected in a water course but a significant number of other tests may be required to evaluate the form or activity of the cyanide and from what process it is liable to have originated.

It is known in the art that assays and particularly biochemical assays suffer from false positive results and false negative results and frequently numerous assays must be carried out to give a high degree of certainty in the final result which gives a disadvantage.

The requirement for multiple assays requires large amounts of the analyte which limits the assay and is a disadvantage. The requirement for multiple assays is costly which limits the overall test and is thus a disadvantage. The requirement for multiple assays is time consuming which limits the overall test and is thus a disadvantage. It would be preferable to mix the different assays such that a single test could be carried out to evaluate the lack or presence of an analyte with a high degree of certainty, but production of multiple signals from a single test is liable to produce cross talk and may produce analysis that is ill conditioned or non convergent thereby reducing or removing the benefit and this is a disadvantage.

In areas including mass screening, a vast number of samples must be analysed. A well-plate may contain between 100 and 100,000 wells, each containing an assay which must be analysed individually. A large number of plates require analysis. In mass screening the vast majority of the samples may have a null result but these may take as long as samples that have a positive result in primary screening such that measurement or analysis time is significant.

Brief Description of the Invention

It is an aim of the present invention to obviate or reduce the above mentioned problems.

Accordingly, in one non-limiting embodiment of the invention there is provided a method of analysing a sample in order to evaluate at least one property of the sample, which method comprises:

(a) excitating the sample by excitation means;

(b) detecting emission from the sample by detector means;

(c) signal processing a data stream in a time or frequency domain; and

(d) analysing the processed signal.

The invention may include the step of comparing the detected emission with a predefined threshold and producing a binary signal when the threshold is exceeded.

The invention may enable one or more high sensitivity analysis, high accuracy analysis, high specificity analysis, and rapid analysis.

The invention may enable one or more of a means of reducing power consumption, a means of increasing yield, a means of operating at high frequencies, a means of allowing more complex circuits and clock types to be analysed, and a means of allowing higher density circuits to be designed and produced.

The invention may enable one or more of restricting IP theft, allowing removal or reduction of test ports, using reduced storage, and a means of analysis whereby storage does not limit the analysis. The invention may allow real-time analysis, real-time multiple pixel analysis, real-time multiple confocal pixel analysis, analysis whereby no coupling between the test equipment and test piece occurs, analysis whereby the measurement is not limited by the bandwidth of an electrical tap, analysis whereby the measurement is not limited by the jitter of the electrical tap, analysis whereby the measurement is not affected by phase loss between a tap and two test points, analysis whereby the measurement is not affected by the jitter between a tap and two test points, and analysis whereby causality of cross-talk may be established.

The invention may allow optical chips to be analysed.

The invention may provide a means whereby the transfer function of a material, a gate or a set of gates may be analysed in a single measurement. The invention may allow analysis of asynchronous circuits, analysis of clock-less circuits, analysis of optimum duty cycle in addition to bandwidth of both clock pulses and data signals, analysis of cross talk in a circuit where coupling between transistors may occur only sparsely, and analysis of a plurality of pixels at one time and thus the production of an image. The invention may provide a means whereby any pixel may be assumed the start channel or the reference channel, and re-analysis with differing pixels acting as the start channel may occur.

The invention may enable the optimisation of circuit layout, and the optimising of process control. The invention may provide a means of analysing product grade. The invention may provide a means of measurement and analysis of signals from mixed species where the analysis would normally be considered ill conditioned.

The invention may provide a means of measurement and analysis of signals from mixed species where the analysis would normally be non- convergent. The invention may provide a means of allowing a species to be represented by a complex function without significantly increasing the complexity of the analysis. The invention may provide a means of allowing a species to be analysed such that a function of the complexity required by the analysis is produced.

The invention may allow assays to be performed with high sensitivity, more rapid primary screening, a reduction of false negatives and positives, and tests to occur where the result is based upon more than one assay rapidly and economically.

The method of the invention may include the step of excitation of a sample. The method may include the step of collection of quanta emitted from the sample by a detector. The detector may be followed by, or include, a threshold means. The method may include setting a threshold such that a trigger occurs if the emitted quanta produce a signal that is greater than the threshold value. The method may include the step of measurement of the time of the quanta emission.

The method may include time domain signal processing. The method may include analysis of the time domain data.

The method may include the use of a collection arrangement to measure a signal from a specific area or volume of the sample. The collection arrangement may include confocal arrangement. The method may include the use of one or more optical fibres as a collection device.

The method may include reflectors. The reflectors may be elliptical. The reflectors may be parabolic. The reflectors may be cones. The reflectors may be such that they transmit an excitation beam but reflect the emitted quanta. The reflectors may be mirrors.

The collection arrangement may contain devices or materials to equalise path-lengths of detected quanta regardless of what angle they are emitted by the sample. The collection arrangement may contain devices or materials to equalise path-lengths of detected quanta regardless of what position they are emitted by the sample.

The detector may be sensitive to individual quanta. The time measured may include the time from one threshold event to a subsequent threshold event. The number that represents the time between thresholds may be compressed in a non-linear form. The compression may be of a logarithmic nature over all or part of the range of compression. Logarithmic nature may be used to define functions which increase nonlinearly and thus provide a means of desensitising measurement at longer times. The compression may be in real-time. Real-time compression is defined such that it does not preclude the use of a buffer but that the processor may act continuously and the buffer does not overfill due to time taken to compress data provided that the instrument is used within its operating range.

The quanta may be photons. The emitted quanta may be reflected excitation quanta. The emitted quanta may be luminescence. The sample may be coated with a material to produce or increase the efficiency of production or collection a specific type of quanta. By way of example the sample may be coated with a material that changes colour when heated and the excitation may be a wide band source such that only a specific colour of light is emitted by the sample for a given temperature.

The detector may include a phosphor. The detector may include a dynode chain. The detector may include an intensifier. The detector may be or may include a solid state detector. The solid state detector may be or may include an avalanche photodiode. The solid state detector may be or may include a PIN diode. The detector may be a multiple pixel detector.

The excitation may include electromagnetic radiation. Electromagnetic radiation includes X-rays, hard UV, soft UV, visible light, infra red and deep infrared light, micro-waves and radio-waves. All electromagnetic radiation is composed of photons. The excitation may include pulsed excitation. The frequency of the pulses may be ramped during a measurement. The frequency of the pulses may be pseudo random and contain a specific weighting of a number of frequencies during one measurement. The duty time of the pulses may be varied during a measurement. The duty cycle may be ramped or the duty cycle may vary in a pseudo-random manner.

The excitation may be the pulses supplied to switch a logic circuit. The pulses may be electrical and the circuit may be an electrical circuit, at least in part. The pulses may be optical and the circuit may be optical, at least in part. The excitation may include excitation that is native to the sample. Excitation may be considered to be native if the excitation is within the sample or within the system within which the sample usually operates. The native excitation may be an oscillator. The oscillator may be on a different substrate to the sample but on the circuit that the sample will usually operate within. The native excitation could include chemically triggered fluorescence or enzyme produced fluorescence.

The processing may be such that there is no trigger pulse used in the processing to establish the excitation pulse. The processing may be digital. The processing may include generation of at least one pulse arrival distribution histogram. The processing may include at least one correlation. The correlations may include autocorrelations. The correlations may include cross correlations. The correlations may include pseudo autocorrelation. The processing may be cross correlation means of intersecting pixels to allow autocorrelation of an area smaller than the pixel to be analysed. The processing may include cross correlating a first pixel with a second pixel and then the second pixel with the first pixel. The processing may include cross correlation of one pixel with two or more further pixels.

The processing may include measurements to allow determination of baselines. The processing may include measurements to allow determination of full scale. The processing may include measurements to allow determination of the validity of other measurements.

A sample may be analysed by one or more pixels. The pixel may contain one or more transistors of a circuit. The pixel may contain a sub- circuit. The pixel may contain an entire device. The pixel may define spatial position. The pixel may define a colour from a specific spatial region. The pixel may define a polarisation of light from a specific spatial region.

The analysis may include the analysis of self jitter. The analysis may include the analysis of jitter between pixels. The analysis may include the analysis of causality. The analysis may include the analysis of cross-talk. The analysis may include the analysis of delay between pixels. The analysis may include the calculation of distance between pixels. The analysis may include the mapping of pixels by means of the distance calculated. The analysis may include referencing of a stored circuit design to establish which transistor or set of transistors may have given rise to a specific signal. The method of the invention may enable analysing a plurality of circuits of a similar overall nature to provide a means of optimising the characteristics and yield of circuits subsequently produced. The method may include a feedback loop where a first processing of the data is analysed and a decision made on the result of the analysis, and where the decision is positive, a further processing stage and analysis occurs where the method may still be considered as real-time according to the definition herein described. The method may allow a plurality of feedback loops to occur and produce a processing and analysis that occurs in a stepwise fashion, whereby an entire sample is analysed first and then more focused measurements occur if a signal of interest is apparent. Such a method may still be considered as real-time.

Measurements by cross correlation may occur between samples of interest and known samples. The reference samples may be on a completely different substrate or be a part of the substrate that contains the sample under analysis. The method may include measurement of at least one reference sample and the result of those measurements being used as a constraint in the analysis of the test sample. The method may include multiple scans on references samples where at least one scan is selected to preferentially reduce the emission of one or more of the species that may be present in the sample. The results for each scan of the reference samples are stored and used as constraint in the analysis of the multiple scans on the sample where all scans are used in the same error reduction process such that multiple scans add to the constraint. The scans may include scans by means of a change in the environment of the sample. The scans may include scans by means of a physical change in the sample. The physical change in the sample may be a function of the intensity or time that excitation has occurred. The scans may include scans by means of alteration of the excitation source. The scans may include scans by means of alteration of the detected light.

The analysis of the sample may be ill conditioned without the constraint.

The analysis of the sample may be non-convergent without the constraint.

The method allows for multiple photons to occur from a single spatial position or from multiple spatial positions, after each clock pulse. A single analysis unit may collect and analyse data from a multitude of pixels simultaneously.

The method may be used as a quality control tool with measurements being made on all or a number of the circuits produced and the output used as feedback to control the process. The method may be used as a selection tool with measurements being made on all or a number of a batch of products and the output used to define the final price of the product. The method may be used as a part of the product process with measurements being made on all or a number of the circuits produced and the output used to defined further processing.

The analysis may involve an iterative fitting step. The iterative fitting step may be constrained by background measurements. The background measurements may be made previously and stored.

The analysis may be ill-conditioned without the constraint. The analysis may be non-convergent without the constraint. The background measurements may include an iterative fitting step. The iteration may include a variation of the type of function used in fitting. The iteration may include a variation of the number of functions used in fitting. The iteration may include a variation in the magnitude of the second function defined to model the spread of the first function used in fitting. The iteration may include a variation of the type of the second function. The iteration may include a variation of the range that a predefined number of functions are spaced over. The iteration may produce a family of functions where a set of functions, parameters and relative weightings are defined for a specific species such that analysis of that species may occur more accurately but with no increase in the number of unknowns. The family may be used in a further iteration to allow analysis of the species present in a complex mixed sample where there is only 1 unknown per family that is assumed to be present. A family may be defined to compensate for background noise and be used as a constraint when analysing data to produce a family for a specific species. A family may be defined to compensate for background noise and be used with other species families when analysing a mixed sample.

The analysis may be used to allow an unknown first sample to be analysed and include the following steps of analysing at least one background sample by means of fitting a defined set of functions of fixed parameters but variable magnitude and varying the magnitudes in an iterative step until the optimum fit is found. If any magnitude is below a predefined value the process being repeated with a reduced number of functions until no parameters are below the predefined threshold. The parameter values and magnitudes then being stored but where the magnitudes are normalised to one magnitude value and so provide a family of parameters which is stored and where the defined family is then used in a second or further fitting processes on emission from the first sample by constraining the analysis such that it includes the pre-defined family with the calculated weighting such that only the magnitude of the family is an unknown in fitting as opposed to each family member being an unknown as would normally be the case.

The method may include the step of iterating the parameter values of the reduced family as well as the magnitudes to optimise the fit. The parameter values of the family and the magnitudes and the number of functions may be optimised in one step. The method may include the step of selecting exponentials as the function to fit. The method may include the step of scanning a plurality of reference samples where each scan is arranged to preferentially reduce the emission of at least one reference sample and storing the magnitude of the emission on each scan and then performing the same scans on at least one sample of interest which contains one or more of the background samples and then using knowledge that the sample is common for each scan and the scan magnitudes of the background samples to constrain a single fitting process and define which background materials were present. The method may include the step of quantification of the background samples present in the sample of interest. The method may include the step of calculating the probability that the analysis is correct and where the probability is below a defined value carrying out further scans to provide greater constraint.

The method may include the authentication of an article.

Brief Description of the Drawings

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

Figure 1 shows the method of analysis of the temporal characteristics of a sample;

Figure 2 shows a lens arrangement being used to collect a signal from the material and transfer it to a detector.

Figure 3 shows a lens arrangement that reduces the area of the material;

Figure 4 shows the use of a waveguide as a detection device;

Figure 5 shows a multiple core waveguide producing a multiple pixel detection means;

Figure 6 shows a photon detector;

Figure 7 shows timing between events of a trigger-less system;

Figure 8 shows non linear timing between events;

Figure 9 shows the effect of non linear timing on the analysis of fluorescent decay;

Figure 10 shows how the compression may be used to analyse signals not centred at highest frequency; Figure 11 shows an automated means of optimal programming of the compression;

Figure 12 shows a means of real time non-linear compression;

Figure 13 shows (a) excitation of a material (b);

Figure 14 shows a photon auto correlation means;

Figure 15 shows a cross correlation means;

Figure 16 shows pseudo autocorrelation;

Figure 17 shows a second psuedo autocorrelation arrangement;

Figure 18 shows the use of a detector bank;

Figure 19 shows a mass pixel instrument;

Figure 20 shows a measurement where no coupling occurs;

Figure 21 shows the autocorrelation analysis of self jitter;

Figure 22 shows the implementation of a subtraction mode;

Figure 23 shows the use of non linear timing to measure the characteristics of jitter of a pulse;

Figure 24 shows the analysis of a transistor by autocorrelation where the transistor is subjected to white noise;

Figure 25 shows that cross correlation AB of two pixels;

Figure 26 shows another cross correlation of two pixels A and B;

Figure 27 shows a plurality of traces (a, b, c, d, e) of cross correlation traces similar to that of Figure 21 ,_but taken at different clock frequencies;

Figure 28 shows how the invention may be used to improve circuit design; Figure 29 shows how the invention may be used to improve circuit layout;

Figure 30 shows a circuit which includes 3 transistors;

Figure 31 shows a means of improving material analysis;

Figure 32 shows use of a toggle in the invention;

Figure 33 shows a method of stepwise analysis;

Figure 34 shows a rapid stepwise method of analysis;

Figure 35 shows the invention used to analyse refractive index changes in a material;

Figure 36 shows how the method may be used to optimise duty cycle of a circuit;

Figure 37 shows application to assays;

Figure 38 shows a use of cross correlation in assays;

Figure 39 shows stepwise use of the invention applied to mass screening;

Figure 40 shows the use of cross correlation to analyse multiplex assays;

Figure 41 shows use of the invention with a database;

Figure 42 shows a means of accurately analysing a trace by means of the invention;

Figure 43 shows a means to allow better representing a species in a model;

Figure 44 shows how a family representing a species may be reduced; Figure 45 shows the use of distribution assumption in a model of the sample;

Figure 46 shows a means of reducing the effect of background;

Figure 47 shows a means of performing multiple scans on an assay by control of wavelength;

Figure 48 shows a means of performing multiple scans where the scans occur after a physical change in the sample;

Figure 49 shows a second means of performing multiple scan where scans occur a physical change in the sample due to excitation;

Figure 50 shows a means of altering the environment of a sample to perform multiple scans; and

Figure 51 shows a means of analysing errors of a semiconductor as a function of temperature.

Detailed Description of the Drawings

Figure 1 shows the method of analysis of the temporal characteristics of a sample. There is an excitation means (a) which excites the sample (b). The material may emit a signal during or following excitation (c) and the emission is measured by a detector (d). The output ofthe detector passes to a threshold means (e) which gives an output when the emission exceeds a predetermined value (f). The emission time is measured by a timing means (g) and this data is directed to a processing means (h) which carries out signal processing of the timing data. The output of the processing means may then be analysed (i,), allowing the temporal characteristics of the emission by the material to be evaluated. The detector may include a light detector. The detector may include a photon detector. The emission may include luminescence. The luminescence may include fluorescence. The processor may include a correlator. The analysis may include a constrained iteration according to previously measured samples. The sample may include a logic substrate. The sample may include a logic circuit. The sample may include an assay.

Figure 2 shows a lens arrangement (a) being used to collect a signal from the material (b) and transfer it to a detector (c). The lens may be physical lens made of glass or some other material. The lens may be an effective lens produced to deflect the emission. The lens may use electric or magnetic fields. A reflector may be used to collect and transfer emission in conjunction with or replacing the lens. The reflector may be a mirror.

Figure 3 shows a lens (a) arrangement that reduces the area of the material (b) that light is collected from and impinges on the detector (c). One or more lenses may be physical lenses. One or more lenses may be effective lenses. One or more lenses may be replaced by reflector arrangements.

Figure 4 shows the use of a waveguide (a) as a detection device. The waveguide may be a fibre. For high spatial resolution the waveguide core (b) should be small. For high collection efficiency the waveguide numerical aperture (c) should be large. The waveguide may be monomode, the waveguide may be a step index waveguide. The waveguide may be gradient index waveguide. The waveguide may be a polarisation maintaining waveguide. The waveguide may be a holey waveguide comprising many hollow fibres.

Figure 5 shows a multiple core waveguide producing a multiple pixel detection means. The core spacing (a) may be reduced using a multi-cored single waveguide structure.

Figure 6 shows a photon detector. The detector (a) has near null output (b) until a particle of light (c) is incident on the detector and a pulse occurs (d). Preferably the detector active area is small (e) as this reduces background noise. Preferably the detector active area is small as this reduces jitter. Preferably the total volume of the detector is small as this reduces jitter.

Figure 7 shows timing between events of a triggerless system. The events may be quanta, the quanta may be photons. The time interval may be re-zeroed on each event. The processing may use a first detected event to define the initial zero and thus there may be no conventional trigger.

Figure 8 shows non-linear timing between events. The timer increments in a non-linear manner following each event (a) and is reset to zero and the initial linearity on each event. The events may be photons. The non-linearity may be of a logarithmic form.

Figure 9 shows the effect of non-linear timing on the analysis of fluorescent decay. The non linear timing leading to less steep areas of the curve containing less data points. Since the gradient of the curve is a function of the information content this method allows a number of non- linear data points to represent the trace better than double that number of linear spaced data points. The result is often represented as in Figure 9 by conventional means but in conventional means, compression is not used to allow improved throughput of signal processing such that conventional means are restricted. The exact form of the compression may be a logarthim or other mathematical function such as the square of the delay. Where the function decreases in senstivity non-linearly with time from reset, the function will be defined as substantially a logarthimic form in terms of this invention.

Figure 10 shows how the compression may be used to analyse signals not centred at highest frequency. Figure 10 may represent the analysis of a gate switching triggered by an earlier gate switching and the pulse-width and shape may be an important analysis to study jitter and the correct operation of the transistors. The sensitivity of the non linear function may include highly insensitive regions (a), areas of the highest sensitivity in linear spacing (b) and areas where the sensitivity changes in a logarithmic nature (c).

Figure 11 shows an automated means of optimal programming of the compression. A first measurement is made (a) with linear channels set in the look up table (b). The area is analysed (c) and a decision made on the area of interest of the delay time and a non-linear spacing produced which is applied to the LUT for further measurements (d). The data from the first measurement may be combined with the second measurement to make best use of the data (e). Figure 12 shows a means of real time non-linear compression. The event stream (a) is directed at a timer means (b) which times between event pairs and resets on each event and reads time between events to controller or processor means (c). The event may be non-linearised by a means that includes a mathematical function. The events may be non-linearised by a means that includes a look up table (d). The processor means may allow the mathematical function to be re-programmable. The look up table values may be re-programmable. The non-linearised event may be directed to a processor means or an analysis means (e).

Figure 13 shows (a) excitation of a material (b). The excitation may be, or include radiation. The radiation may be electromagnetic. The radiation may include X-rays. The radiation may include ultraviolet light. The radiation may include visible light. The radiation may include infrared light. The radiation may include deep infrared light. The excitation may include a sonic wave. The excitation may include a magnetic field. The excitation may include an electrical field. The excitation may include one or a plurality of electrical pulses. The excitation may include one or more means to restrict the area excited (c). The excitation may be pulsed. The excitation pulses may be below 10ns in width. The excitation pulses may be below 1ns in length. The excitation pulses may be in the range of 0.1 - 40GHz in repetition range. The repetition rate may be scanned in a predetermined manner during an experiment. The repetition rate may be pseudo random and contain a fixed weighting of differing periods during a measurement. The excitation may include a direct contact probe. The probe may be an electrical wire. The area of restricted excitation may be a circuit within a substrate. The excitation means may be intrinsic to the substrate itself. The excitation may include an electrical oscillator.

Figure 14 shows a photon auto-correlation means. A sample (a) is excited by an excitation means (b). The excitation means may include a pulsed light source. The excitation may include electrical pulses. A stream of quanta is emitted by the substrate (c). The quanta may be photons. The quanta may be collected by a collimating means (d). The area of material that light is collected from may be restricted by a detection arrangement (e). The arrangement may include a pinhole. The arrangement may include a field lens. The arrangement may include a fibre. The emitted quanta may be focused (f). The emitted quanta are detected by detector means (g). The detector means includes an amplifier (h). The detector means includes a threshold device emitting a binary sequence. The binary sequence may be directed toward a timing and means which includes storage and compression and control means (i). The binary data is directed into an autocorrelator (j)- The output of the autocorrelator may be directed to an analysis means (k). The analysis means may include fitting. The fitting may be force fitting. The forced parameters may be read from a storage area (I). The force fitting parameters may be defined from previous measurements of reference samples. The output of the analysis means may be directed at one or more output devices (m). The output device may include a screen. The output device may include a printer. The output device may include a network connection. The sample may include a logic circuit. The means may perform a test of the operation of the logic circuit. The sample may include an assay and the process may perform the quantification of one or more analytes.

Figure 15 shows a cross correlation means. A sample (a) is excited by an excitation means (b). The excitation means may include a pulsed light source. The excitation may include electrical pulses. A stream of quanta is emitted by the substrate (c). The quanta may be photons. The quanta may be collected by a collimating means (d). The area of material that light is collected from may be restricted by a detection arrangement. The arrangement may include a pinhole. The arrangement may include a field lens. The arrangement may include a fibre (e). The emitted quanta may be focused. The emitted quanta are detected by detector means (f). The detector means includes an amplifier (g). The detector means includes a threshold device emitting a binary sequence (h). The binary sequence may be directed toward a timing and means which includes storage and compression and control means (i). The binary data is directed into an autocorrelator (j). The output of the autocorrelator may be directed to an analysis means (k). The analysis means may include fitting. The fitting may be force fitting. The forced parameters may be read from a storage area (I). The output of the analysis means may be directed at one or more output devices (m). The output device may include a screen. The output device may include a printer. The output device may include a network connection. The detected quanta may include photons. The detectors may be sensitive to individual quanta. The excitation may include electrical pulses. The material may include a logic circuit. The means may perform a test of the operation of the logic circuit. The control means and storage means may be produced so as to allow the cross correlation of a detector pair to occur in both direction without data loss by means of the compression. The control and storage means may be produced to allow at least one auto-correlation to be performed in addition to at least one cross correlation by means of the compression.

Figure 16 shows pseudo autocorrelation. This may be carried out for reasons that include increasing the bandwidth of the measurement. This may be carried for reasons that include analysis of higher spatial resolution. This may be carried out for reasons that include higher light collection efficiency. This may be carried out to reduce correlatable background noise from the detector. Emission is collected from the sample (a). The collection may include a lens means (b). The collection may include a fibre (c). The emission is split by a beam splitter (d). The beam splitter may comprise a fibre beam splitter. The beam splitter may comprise a cube beam splitter. The beam splitter may comprise a plate beam splitter. The beam splitter may comprise a pellicle beamsplitter. The preferred split ratio is fifty percent in each arm. The emission is detected by at least two detectors (e). Each detector or detector bank output (e) comprises an input to a cross correlator (f). The arrangement allows an effective autocorrelation to be performed with analysis that is higher bandwidth that would normally be allowable by the detectors. An emitted quantum will be detected on one but not both detectors and thus cross correlation of a quanta stream allows an autocorrelation to be performed where two quanta may be detected within the reset or stabilisation time of a detector.

Figure 17 shows a second pseudo autocorrelation arrangement where two separate collection devices (a) are used each collecting emission (b) from a sample (c) and directing it to a detector (d) where each detector is the input to a cross correlator. The arrangement allows higher collection efficiency that using a single collection device. The arrangement allows higher spatial resolution as the area light is collected from may not be fully confocal and a correlatable signal only occurs in the intersection of the two beams (e).

Figure 18 shows the use of a detector bank to provide a means of a detector that has a very high burst rate and allows a very high collection efficiency. Light (a) from a single area is separated between a number of detectors (b). Preferably the separation gives all detectors equal signal outputs. Partially reflective mirrors may be used to separate the light. A fibre beamsplitter may be used to separate the light. A fibre bundle may be used to separate the light. The light may be collimated across an area that includes all the detectors. The detectors may include a threshold means and pulse shaping means (c). Preferably the detectors have a circuit to reduce the pulse width of a high signal (d). The detectors are connected as a logical OR (e) such that the detectors act as a single device with a faster than normal detector reset time. The input light may be collected by many collection means, allowing higher efficiency.

Figure 19 shows a mass pixel instrument that may be programmed to perform a plurality of differing measurements. The instrument has a plurality of input pixels (a1, a2) each of which is split a plurality of ways equally (b1, b2). The input pixels may be formed by a lens means or a fibre means or a mirror means. The separation may be by fibre means or bulk optic means. The light on each path is then separated a plurality of ways according to its wavelength (d-4) and passed to a detector sensitive to the wavelength range (d1-8). When the detector goes high a logic circuit (e1-8) causes a pattern on a bus (f). One line of the bus acts purely as a trigger and when high causes a timing means (g) to output the time since last event along with the number of the pixel that causes this event and this number is passed to memory (h). The device may have a control means (i). When a memory slot has been filled then a second memory is filled whilst the first is processed. The processor (j) may be a correlator or a pulse arrival distribution circuit or some other device. The processor may perform a number of sequential calculations on the same data segment serially. By way of example the processor may analyse the times between any input on detectors (d1-4) and so perform an autocorrelation on the pixel regardless of colour. The processor may then perform a cross correlation where (d1) or (d3) triggers one channel and (d2) and (d4) trigger the second channel and thus analyse if the photons of different colours of light are due to a common process. The processor may then analyse where any detector (d1 to d4) triggers one channel and detectors (d4 to d8) trigger a second channel and thus look for correlations between pixels (a1) and (a2). It is to be understood these are only a few of the possible analyses that could be performed using the restricted circuit shown and that separation may also occur for polarisation. Preferably the timing means is not a rollover timing means and may be made to stop at a predefined time and thus clip the timing to a maximum value. Preferably the processor means allows compression prior to analysis of the data. The data may be further analysed (k) and sent to an output means (I).

Figure 20 shows a measurement where no coupling occurs. A logical circuit (a) includes an oscillator means (b). Some of the oscillator means may be on the substrate under test. The circuit is made to operate and gate switching leads (c) to photon emission (d). The photon emission is picked up by a detector (e) and processed (f) and analysed (g). No field is applied to the circuit by means of the measurement. The distance between the collection optic (h) and the substrate may be a number of millimetres and thus no coupling between the measurement and the measurement equipment occurs.

Figure 21 shows the autocorrelation analysis of self jitter of a pixel as figure 20 where the pixel contains one or more transistors and the analysis circuit does not contact the circuit and is not invasive. The analysis shows a peak at a delay equivalent to the clock frequency and every harmonic of the frequency. The width and shape of the peaks showing the jitter of the transistor relative to itself as opposed to an external line as would conventionally be the case. The peak heights will vary with the most significant peak being that at period equivalent to the duration of the average photon rate.

Figure 22 shows the implementation of a subtraction mode. The timer (a) outputs data to a memory store (b). There may be two memory stores used sequentially or one memory store used in ring mode. All or some of the data is input to an autocorrelator (c) and the data is analysed and the clock frequency measured (d). The memory may be rescanned or all new data is passed via a subtraction circuit (e). The circuit be programmed by the measured period (f). The subtraction circuit subtracts the period of the clock from the period measured until the period is less than the threshold period. This makes all harmonics degenerate and reduces the time base the autocorrelator must cover to analyse the signal. The subtraction mode may be implemented in numerous ways on the data before or after storage. The programming may occur on the timer itself and define when the .timer rolls over. One method of doing this is using a count down timer which is pre-set to count down from a time equivalent to the period of the clock.

Figure 23 shows the use of non-linear timing to measure the characteristics of jitter of a pulse.The circuit may include non-linearisation. Where non-linearsation occurs as well as subtraction it is preferable the subtraction occurs first. The Figure shows how non-linearisation allows an accurate determination of the extremes of the jitter of the pulse using only 16 correlator delay channels. Where the leading and trailing edges of the pulse need not be distingished the subtraction circuit may be used to mirro the pulse around its peak and reduce the number of channels required to 8.

Figure 24 shows the analysis of a transistor by autocorrelation where the transistor is subjected to white noise. The transistor will operate successfully to a certain frequency and then may not fire with a high probability at higher frequencies. The autocorrelation shows that the transistor is unstable when pulse spacing lower than t occur. The noise may be pseudo-random and defined. The noise may be pink noise and not truly white. Providing the function that would occur for an ideal circuit is known a variety of frequency distributions may be used. This allows a single measurement to analyse the transfer function of a pixel. The pixel may be a transistor. The pixel may be a number of transistors. The pixel may be an entire logic circuit. The pixel may include one or more entire substrates.

Figure 25 shows that cross correlation AB of two pixels where there is no causality between A and B and thus only a baseline of noise is measured.

Figure 26 shows another cross correlation of two pixels A and B but now as crosscorrelation BA where there is some causality between pixels B and A. The method allows the level of causality to be established by means of the height and volume of the peak compared to the baseline. The method allows the phase relationship of the pixels to be established by means of the delay to the peak. This may be used for circuit mapping as the distance to a peak may be used to establish which transistor gave rise to the peak. Alternatively the method may be used to validate the circuit map and measure the electrical distance between transistors compared with the design projection. Alternatively the method may be used to allow cross talk to be studied allowing discrimination of which transistors cross talk and in which directions (mono or bi-directional cross talk). This method is advantageous over conventional measurements as no storage of data is required and so long experiment duration may occur to see small cross talk signals. This method is advantageous as it allows very small cross talk signals to be analysed as other signals may form a baseline allowing clear delineation of the signal of interest.

Figure 27 shows a plurality of traces (a, b, c, d) of cross correlation traces similar to that of Figure 21 but taken at different clock frequencies or different mean frequencies of random noise of data similar to Figure. The level of cross talk differs between the traces such that the effective bit error rate due to cross talk may be defined for a specific clock or data frequency (e).

Figure 28 shows how the invention may be used to improve circuit design. The circuits requirement is defined (a) and a circuit formed (b) by application of design rules (c) using a processor (c). The circuit is the tested according to the invention (e) and transfer functions, cross talk and other characteristics are defined to the processor and update the design rules (f).

Figure 29 shows how the invention may be used to improve circuit layout. A series of transistors or transistor sets (a1-a3) is defined (b) on a substrate (c) by a circuit forming means (d) using design rules (e) where the design rules are modified by the processor (f) for each transistor or transistor set. All transistor sets are then tested according to the invention (g). The design rules which produced the optimum characteristics are then selected for use in subsequent circuits (h).This gives advantage as it improves the characteristics of a circuit and reduces the time to yield of any circuit design.

Figure 30 shows a circuit which includes three transistors which are closely spaced such that the probed area cannot be made to include only transistor c. By using two probed areas, the intersection of probe areas may be used to reduce the number of transistors that are common to both probe areas and improve spatial resolution. The method is advantageous as when all the transistors are made to trigger in an uncorrelated manner the cross correlation of the pixels defined will give a signal equivalent to the autocorrelation of transistor c alone and thus cross talk, transfer function jitter and other parameters may be analysed.

Figure 31 shows a means of improving material analysis. The means may be used to test a circuit. A transistor (a) within a substrate (b) is made to switch by a signal or oscillator means at its normal operating frequency such that the bit error rate is low. The emitted photons are analysed by the invention to measure the bit error rate. The transistor is heated by a heat source (f). The increased heating increases the bit error rate such that analysis occurs more rapidly that at normal operating temperature as faults are more prodigious. The analysis occurs for a number of different transistor temperatures. An extrapolation of errors versus temperature may be used to allow the error rate at normal operating temperature to be established. The heat source may be a laser. The laser may emit. infrared radiation. The infrared radiation may be absorbed by a two or three photon process. Using a multi-photon process is advantageous at it allows heating to be defined at a depth within the substrate. Using three photon or higher excitation is advantageous as it allows the spatial resolution of the instrument to be improved by means of the non-linear excitation field of a laser spot. Other fields may be directed at a material to alter the materials emission characteristics.

Figure 32 shows use of a toggle in the invention. For measurement of very high bandwidth signals, particularly when multiple signals are connected to the same processor the method may improves the temporal characteristics of the system. Light from one of more pixels (a) is separated (b) between two detector banks (c). The light may be shared by any means. Each detector bank may contain one or a plurality of detectors (d) each may have a threshold means (e). Each may have a pulse length reduction circuit (f). The circuit may contain a logical OR function (g). Each detector bank acts as an input to a toggle (h). The toggle may have a set line (i). The output of the toggle may pass to a delay line (j). Where a plurality of signal inputs and thus toggles exist each delay line may differ in value. The input optics may form part of all of the delay line. The output feeds to a circuit that defines ensures the output pulses are always the same length (k). Preferably this is the shortest length required by the logic. The output then leads to a circuit that produces a pattern on a bus (1). The pattern being specific to each toggle in the circuit. One line of the pattern may be used as a trigger and common to all toggles. The logic circuit may contain delays to ensure all signals directed at the bus are in phase. The circuit thus outputs to the bus after two photons have been detected. On average half the photons are rejected as hitting the same detector bank.

Figure 33 shows a method of stepwise analysis. Light (a) is collected from the total area of a sample (b). The sample may be a complete or partial slice of a semiconductor wafer. The wafer may be excited. The excitation may be by electrical connection to the semiconductor such that circuits are made to operate. The sample may contain a plurality of integrated circuits. The circuits may be mounted and each circuit excited individually. The sample may be a single mounted circuit. The collection device may include a mirror (c). The mirror may be elliptical. The mirror may be parabolic. The mirror may be a cone. The mirror may be formed such that light from the semiconductor to the detector travels a path where the delay between different path lengths is less than the period of excitation source. The light is detected by a detection means (d) and analysed (e). The arrangement may include further optics to compensate for different light path lengths. The collection device may include a fibre bundle. The collection device may include gradient index optics. The collection device may include a fibre taper. If errors are detected on the sample then only part of the sample is analysed to identify which area the errors occur in (f). The restriction may occur by restricted the area that is excited. The restriction may occur by restricting the area that light is detected from. The area restricted may increase in a stepwise manner until only a single circuit or chip or mask area is identified as being faulty. The restriction may utilise a mask. The mask may be programmable. The mask may include a liquid crystal display. The mask may contain an optically programmed plate. The restriction may occur by use of different pixel heads.

Figure 34 shows a rapid stepwise method of analysis. A production line (a) is made to move samples (b) in front of a light collection means (c). The light collection means detects light from the entire sample. There may be a plurality of collection means working in parallel to make analysis rapid (d). The light is detected and analysed (e) and the information used to control (f) further light collection means (g) which restrict light from a reduced area of the sample (b) which is analysed. This may also occur in parallel This data may be used to programme further collection means of greater restriction. There may be further production lines linked to the same system (i).

Figure 35 shows the invention used to analyse refractive index changes in a material. A material which is exposed to change in environment may alter refractive index and this allows a measure of the environment. The sample (a) is illuminated (b) by an optical source (c) in an environment that is lower refractive index than the sample. The light beam is reflected and detected (d) and processed and analysed. The light beam is set at angle at or near Brewster's angle (e) such that total internal reflection occurs. If the refractive index of the material changes due to excitation (f) then the reflection condition changes and light may be transmitted (f). The excitation may be pulsed and the autocorrelation ofthe signal at the detector (d) will therefore be a function of the excitation. The light source may be a laser. The light source and excitation source may be common. The excitation may be or cause heating of the sample and thus a change in refractive index. The quantity of heating may be for a specific level of excitation may be used to analyse the sample. The excitation may be scanned in frequency and the heating for a specific frequency used to evaluate the sample. The excitation source may use multiple photon absorption by at least a part of the sample to ensure a highly localised excitation. The excitation source may use multiple photon absorption by at least a part of the sample to ensure sub surface excitation. The excitation source may be scanned across the surface to evaluate the entire sample. The excitation source may illuminate the entire sample and multiple detectors allow imaging of the surface.

Figure 36 shows how the method may be used to optimise duty cycle of a circuit. A circuit is operated at numerous clock or data pulse duty cycles such that the percentage the signal is high level is varied for the same clock frequency (a, b, c). The circuit may be operated with a pseudo random or continuous clock pulses. The circuit output is tested by means of the invention and measurements taken on operational characteristics (d, e, f) for different duty cycles where at least one measurement is by means of the invention. The measurements may include analysis of the transfer function. The measurements may include measurements of the temperature. The measurements may include measurements of cross talk. The measurements may include measurements of power usage. The measurements are then analysed (g) and a decision made (h) on the optimum circuit. The method may be used to optimise the operation of an entire circuit or a part of the circuit. The measurement may be used to analyse number transistors during testing of a new fabrication process.

Figure 37 shows application to assays. The invention is described by means of a competition binding assay. A series of antibodies (a) are laid out in a wellplate. Each well having antibodies to a different analyte. Fluorescent species are added to each well and bind non specifically to the antibody. Preferably the binding sites are in excess. If the correct analyte (c) in the sample is present then the fluorophore is released and undergoes Brownian motion (d). The sample is illuminated by light which causes the fluorescent species to emit light (c) Fluorescent emitted light is directed (e) through a filter which removes the excitation light and onto a detector (g). There detector may be individual to that pixel or a multi-pixel detector may be used or a position sensitive detector nay be used. The output from the detectors is sent to the processing means (h) which uses correlation to determine the level of fluorescent binding in each well. The assay may be for a chemical or bacteria or virus. The assay may include analysis of DNA, RNA or proteins. Analysis by correlation may examine fluorescent lifetime or mobility.

Figure 38 shows a use of cross correlation in assays. A system according to Figure 32 or similar may be used to collect data from a plurality of well plates. There may be a plurality of tests carried out to check if the analysis is correct and to determine the baseline and sensitivity of the measurement. One or more cross correlations (a) between two wells may be used to check for a lack of correlation and ensure no false triggering has occurred. One or more cross correlation's between two wells may be used to check the range of the assay and calibrate any measurement and increase the sensitivity or accuracy of the measurement (b). One or more cross correlations may be used to validate the presence or quantity of an analyte (c). Using cross correlation instead of autocorrelation allows noise from other fluorescent mobile species to be reduced. The assay may be for a chemical or bacteria or virus. The assay may include analysis of DNA, RNA or proteins.

Figure 39 shows stepwise use of the invention applied to mass screening. A number of wells (a) are measured (b) simultaneously (a) while the sample is excited (c). The collection device may include fibres (d). The data is stored. The data is auto-correlated ignoring which pixel it originated from. The autocorrelation may be real-time. If the autocorrelation shows a positive result (e) then the data is kept and further analysis occurs. The instrument may continue scanning while secondary analysis occurs. If the result is negative then the area may be cleared. The analysis then occurs in a stepwise manner on the stored data. The data from half the pixels may be autocorrelated (a) and if the result from this is are positive if the probability of a positive result is low then the other half of the well may be ignored. This may be repeated (g) until the positive well is noted (h). Where the probability of two positive wells within the set is high the all wells within an area that tested positive may be analysed. Figure 40 shows the use of cross correlation to analyse multiplex assays (need to sort the bloody claim for this one). An assay is performed where the result may be positive in a plurality of ways as multiple analytes are being performed. In the example we assume three species of result may be present (a, b, c) in the assay sample (d). On the same or a different well plated there are samples of each result separately (d e f). Conventional autocorrelation of the assay (d) would require analysis which may be complex if impurities may also be present in the sample, in this method the cross-correlations G(e, d), G(f, d) and G (g, d) are performed such that the results are weighted towards signals present in both signals and environmental or material variations may be ignored.

Figure 41 shows use of the invention with a database. The invention may be used to measure a plurality of differing samples (a, b, c, d, e) where the set (f) includes all species that could be present in an unknown sample. The unknown sample (g) is then analysed and a fitting procedure (h) applied which allows analyses the optimum subset of the database which could produce the resulting signal where any number of subspecies may be present and in any relative quantity (q). The species may be represented traces from a signal processing of the species. The signal processing may be correlation. The processing may include decay analysis. The species may be specific types of transistors or specific errors of a transistor. The species may be impurities or dopants of a material. The species may be markers or tags. The species may be fluorescent. Figure 42 shows a means of accurately analysing a trace by means of the invention. The species (a-e) as defined in Figure 25 are measured a plurality of times (f) and on each scan a means of preferentially changing the signal of one or more of the species is used (g). The means may include filters. The means may include excitation signal magnitude. The means may include excitation frequency. The means may include chemical addition to the species. The means may include physical damage to species. The emission strength of each species is analysed for each species on each scan and relative magnitudes calculated. This produces a data table of scan versus species type of relative magnitudes (g). The unknown sample is then measured using the same scans giving rise to a data file (i). The data file may then be fitted using the relative magnitude file for each scan and producing a constraint for the analysis. This method allows more accurate determination of the relative magnitude of each species than would conventionally be the case. This method allows more specjes to be analysed in a single trace than would conventionally be the case. The relative magnitudes (g) may be a defined list of relative weightings of a family as herein defined.

Figure 43 shows a means to allow better representing a species in a model by means of definition of a family of parameters. In analysis it is common to represent a species with an approximate model. In the case of fluorescence this model is a decay of fixed lifetime. It is claimed that a decay is not a single fixed lifetime but a range of lifetimes. The lifetimes may be due to differing species in the sample. The species may include impurities. A measure is taken on the sample. The measurement may be allow a fluorescent decay (a) to be analysed. The measurement may be represented by a number of functions where the parameters of each function are defined (b) but where the relative magnitude of each function iterated to find the optimum solution. The solution may be constrained such that all parameters must be positive. The range the functions are fitted over may then be reduced in a stepwise fashion (c, d) until a defined number of functions are null at each extreme (e). The result may then be defined in an array or matrix where the decay parameter and relative magnitude is defined and a single parameter (g). This allows a complex representation of a sample to be achieved without significant increase in the complexity of further mixed species measurements as the number of unknowns in the second fitting has not increased.

Figure 44 shows how a family representing a species may be reduced. The method of Figure 43 or another method may be used to produce a model for each species that could be present in a sample. The family may be defined as a number of functions that is great enough to represent the most complex species but this leads to over resolution of measurements of the other species. The family is defined as Figure 43 or by other means where the parameters are constrained to be physically possible. One such constraint may be that the parameters are positive (a) . The fit is then repeated with no constraint that the results must be physical. If one parameter exhibits a non-physical result then the number of parameters used to represent the data is reduced in a step wise fashion without the range over which the parameters are selected over being effected (b). This is repeated until no parameter is non-physical (c). The stepwise reduction may include the iteration of the parameter positions as well as the magnitudes in order to find the optimum solution. An array may represent the data as shown in previous Figures.

Figure 45 shows the use of distribution assumption in a model of the sample. A sample may be analysed with one or more functions being used to define the value of the functions used in the fitting (a) and a second parameter (b) being iterated which defines the spread of the functions. The method may include a further parameter (c) which defines a function for the spread of first function (c).

Figure 46 shows a means of reducing the effect of background. The supporting matrix for the sample or the instrument without the sample of any other object that may cause a background signal is first measured according to the invention as described by previous 2 figures of by other means. The array for the background family is thus defined. The sample is then placed in the instrument and the process repeated but in this case the analysis is constrained that it includes the background family and the quantity of the background family is also fixed. Thus a means for analysis of a species such that it may be properly represented even in the presence of a background is defined.

Figure 47 shows a means of performing multiple scans on an assay by control of wavelength. An excitation source (a) contains a means of controlling the wavelength emitted (b). The excitation source may include a thermally produced light, or laser light . The light is directed at the assay (c) and fluorescence emission occurs. The emission is collected and passed through a filter means (d) to a detector means (e) which measures the magnitudes for a plurality of excitation and emission colours which are stored (f). The assay may include measurement of reference samples (g) to analyse backgrounds, full scale and check the validity of the analysis. The reference samples may be analysed serially or in parallel. The reference measurements may include measurements of the sample which forms the assay prior to the final assay step being complete. The reference samples may be used to normalise the data before or after an analysis step. When the measurements and any normalisations have been carried out a processor means analyses the data (h) and sends the information to an output means (i).

Figure 48 shows a means of performing multiple scans where the scans occur after a physical change in the sample due to an excitation. A test (a) based on chemiluminescent emission (b) is triggered by the addition of one or more substrates(c) such that one or more enzymes produce fluorescent product. The emitted light is detected and stored (d) relative to the time after excitation (e). The emission profile is analysed and one or more values are used as magnitudes in the analysis. If substrate is added in excess then the emission will rise as a function of the time it takes the enzyme to generate product. The maximum emission being reached when all enzymes cease to produce. An enzyme typically catalyses of the order of 2000 reactions prior to its cessation. Figure 49 shows a second means of performing multiple scan where scans occur a a physical change in the sample due to excitation. A sample (a) is illuminated by an excitation beam (b) and emits fluorescence. The fluorescence of the sample slowly bleaches with time and thus the emission decreases (d). Different materials will bleach at different intensities. Different materials will bleach differently using different wavelengths of light.

Figure 50 shows a means of altering the environment of a sample to perform multiple scans. A sample (a) is placed in a sealed box during measurement and different gasses (b) may be passed into the environment altering the emission of the material due to oxygen quenching or other means. Scans may be taken at different pressures of gas. The sample may be heated or cooled by temperature controlled means (c) and scans taken at different temperatures. Additions may be made to a buffer surrounding the sample and change for example the pH or dielectric properties of the buffer.

Figure 51 shows a semiconductor (a) which is excited by means of an electrical signal leading to the switching of transistors and thus the emission of light. The collection optics (b) collect light from a defined spatial area and pass the light to a refracting element (c). The element may be a prism. The emitted light is directed at a detector arrangement. The detector arrangement may comprise a plurality of detectors. The detector arrangement may comprise a position sensitive detector. The detectors are photon counting detectors. The signals are compared with a threshold and may be shaped (e) and sent to an OR logic function such that a photon on either detector generates a pulse and allows the luminescence emission to be analysed. The signals are also directed at counters (g). The counters note the photon emission per detector per measurement and are read out by the system. The relative magnitude of each detector defines the wavelength of emission. The wavelength of emission may thus be measured from a data stream which allows luminescence. The luminescence may allow the analysis of errors on the semiconductor and thus the errors and temperature may be analysed.

It is to be appreciated that the embodiments of the invention described above with reference to the accompanying drawings have been given by way of example and that modifications may be effected.

Claims

1. A method of analysing a sample in order to evaluate at least one property of the sample, which method comprises:

(a) excitating the sample by excitation means;

(b) detecting emission from the sample by detector means;

(c) signal processing a data stream in a time or frequency domain; and

(d) analysing the processed signal.

2. A method according to claim 1 and including comparing the detected emission with a predefined threshold and producing a binary signal when the threshold is exceeded.

3. A method as according to claim 1 or claim 2 and including collecting the emission from at least one defined area of the sample by collection means.

4. A method as according to any one of the preceding claims and including resetting the timer on each event such that the time between events is processed.

5. A method according to claim 4 in which a value representing the time between events is compressed in real-time.

6. A method according to claim 4 in which the compression includes a substantially logarithmic form of compression.

7. A method according to claim 5 or claim 6 and including a subtraction in the compression.

8. A method according to any one of claims 5 - 7 and including a first measurement which is analysed and the result used to define an optimum compression which is programmed into a storage area and used by the compression means in at least one subsequent measurement.

9. A method of any one of the preceding claims in which the compression occurs after a buffer step, and including storing a code that defines a pixel that caused an event along with the time to a preceding event from any pixel in a temporary data store and reading out the stored data through a mask such that when a pixel from a predefined set occurs an accumulator is read out but events from other pixels cause the accumulator to sum the stored times such that the time between events on a predefined set of pixels is output to a processing means where the mask is also to be reset to allow definition of a different set of pixels.

10. A method according to claim 9 in which the data store is read a plurality of times for different mask settings prior to the data store being refilled in real-time.

11. A method of claim 9 or claim 10 and including analysing a first set of processed data and making a decision according to the analysis which defines if a second or more sets of data will be processed with different mask settings.

12. A method according to any one of claim 7, 9, 10 or 11 in which at least part of the compression occurs after the mask function

13. A method according to any one of the preceding claims in which a threshold event is triggered by a single quanta.

14. A method according to any one of the preceding claims in which the excitation includes electromagnetic radiation.

15. A method according to any one of the preceding claims in which the excitation includes an electron beam.

16. A method according to any one of the preceding claims in which the excitation includes a flow of electricity.

17. A method according to any one of the preceding claims in which the excitation is pulsed.

18. A method according to any one of the preceding claims in which the frequency of the excitation pulses varies during a measurement.

19. A method according to claim 17 in which the frequency is ramped during a measurement.

20. A method according to claim 17 in which the excitation pulse spacings are random or pseudo-random in nature.

21. A method according to any one of the any preceding claims in which at least part of the excitation source is native to the sample, and non- coupling analysis is performed.

22. A method according to any one of the preceding claims in which the processing includes generating an event arrival distribution histogram.

23. A method according to any one of the preceding claims in which the processing includes autocorrelation.

24. A method according to any one of the preceding claims in which the processing includes cross correlation.

25. A method according to claim 23 in which at least one autocorrelation is a pseudo-autocorrelation.

26. A method according to claim 24 or claim 25 in which emission from two pixels which only partially overlap is processed by cross correlation.

27. A method according to claim 23 and including selecting at least one reference sample and cross correlating the data-stream from the reference samples with the data-stream if the sample to evaluate of the samples have similar characteristics.

28. A method according to any one of the preceding claims in which at least one reference sample is used to provide a means of establishing a baseline to calibrate the result of an analysis.

29. A method according to any one of the preceding claims in which at least one reference sample is used to provide a means of establishing a full scale value to calibrate the result of an analysis.

30. A method according to any one of the preceding claims in which at least one reference sample is included to check the validity of the measurement.

31. A method according to any one of claims 24 - 30 in which the analysis includes the analysis of a difference in phase in a plurality of spatial positions.

32. A method according to any one of claims 24 - 31 in which one spatial pixel is first cross-correlated with at least a second spatial pixel and then the second spatial pixel is cross correlated with a first spatial pixel to establish causality.

33. A method according to claim 30 or claim 31 in which the analysis includes the analysis of cross talk of at least a part of a logic circuit.

34. A method according to any one of the preceding claims in which the analysis includes the analysis of the jitter of at least a part of a logic circuit.

35. A method according to any one of the preceding claims in which the analysis includes the analysis of the characteristics of the excitation field applied to the sample.

36. A method according to any one of the preceding claims in which the analysis includes the analysis of the transfer function of at least one part the sample.

37. A method according to any one of claims 24 - 36 in which the analysis includes the analysis of the relative position of at least two emission points on the sample by means of analysis of the time of emission between them.

38. A method according to any one of the preceding claims in which the analysis includes measuring at least one reference sample, storing the result of the reference sample and using the result of the analysis of the reference sample to provide a constraint in the analysis ofthe sample.

39. A method according to claim 38 in which a plurality of reference samples are each measured with a plurality of scans, in which each scan is selected to preferentially reduce the detected signal of at least one reference sample, in which these magnitudes are stored, and in which a sample that may contain one or more of the reference signals is then analysed with similar scans and the magnitude stored, and then the magnitudes of the reference measurements are used as a constraint in analysis ofthe sample in a single iterative fitting step.

40. A method according to claim 39 and including calculating of the certainty of the output result where if the certainty is below a defined value further scans are taken and the process repeated until the certainty is above a defined value or until a predefined maximum number of scans have been taken.

41. A method according to claims 39 or claim 40 in which at least one scan is performed where a change in wavelength of the excitation is used to alter the emission magnitude.

42. A method according to claims 39 or claim 40 in which at least one scan is performed where a change in wavelength of the emission detected is used to alter the emission magnitude.

43. A method according to claims 39 or claim 40 in which at least one scan is performed where a change in the environment of the sample used to alter the emission magnitude.

44. A method according to claims 39 or claim 40 in which at least one scan is performed where a physical change in the sample due to an excitation field is used to alter the emission.

45. A method according to claims 38 - 44 in which analysis would be ill conditioned without the constraint.

46. A method of any one of claims 38 - 46 in which the analysis would be non-convergent without the constraint.

47. A method according to any one of the preceding claims for allowing the analysis of the impurities of dopants in a material.

48. A method according to any of the preceding claims for allowing the analysis of the environment of a material.

49. A method according to any one of the preceding claims for allowing the determination of the function of a logic circuit.

50. A method according to any one of the preceding claims for the analysis of circuits where there is no repetitive clock pulse that is common to the entire circuit.

51. A method according to claim 49 or claim 50 in which the result of at least one analysis is stored along with a definition of the sample or part of the sample that produced said result, and in which this stored data is made available to a design package such that the design package has greater surety of the design rules in subsequent designs and so provide a means of improved logic circuit design.

52. A method according to any one of claims 49 - 51 and including removing samples from a production line according to a sampling protocol, measuring the samples, and modifying operation of the production line according to the results of the said measurements.

53. A method according to any one of the preceding claims in which at least one sample of a batch of samples is analysed and the analysis used to define the grading of the sample, thus allowing differential pricing of the samples according to grade.

54. A method according to any one of the preceding claims for allowing the optimisation of a circuit or circuit part at a manufacturing plant to maximise yield and characteristics of a subsequently produced logic circuit.

55. A method according to any one ofthe preceding claims for allowing a means of assay.

56. A method according to any one of the preceding claims for allowing a means of analysis of a fluorescence assay.

57. A method according to any one of the preceding claims for allowing a means of analysis of a data stream from a dynamic light scattering signal.

58. A method according to any one of the preceding claims and including authenticating an object by means of analysis of a signal from a label comprising at least one luminescent material, where the producer has prior knowledge of the label and, by constraining the analysis of the emitted signal that it must contain specific species, the producer may iterate to find the relative quantities of each species and use this to validate the sample where analysis would be ill conditioned without knowledge of what dyes could be present.

59. A method according to claim 58 in which a species type and quantity are selected such as to produce an analysis which would be non-convergent without prior knowledge where without prior knowledge the label could not be analysed by conventional means and so provide a means of secure authentication of an article.

60. A method according to claim 58 or claim 59 in which dots of below 100μm are printed in an array and at least two dots may comprise a differing set of dyes.

61. A method according to any one of claims 58 - 60 in which the method is used to store information securely.

62. A method according to any one of claims 58 - 61 in which the method is used by a vendor of a product to protect against unwanted distribution of merchandise.