WO2009050437A1

WO2009050437A1 - A system and method for infrared imaging

Info

Publication number: WO2009050437A1
Application number: PCT/GB2008/003440
Authority: WO
Inventors: Christopher Clement Phillips; Hemmel Amrania
Original assignee: Imperial Innovations Limited
Priority date: 2007-10-15
Filing date: 2008-10-10
Publication date: 2009-04-23
Also published as: GB0720097D0

Abstract

The present invention relates generally to infrared imaging. Embodiments of the present invention provide an infrared imaging system (100) for multispectral infrared imaging of a sample (104), wherein infrared radiation of one or more narrow frequency ranges is/are directed towards the sample; and one or more images of said sample at the one or more narrow frequency ranges is/are detected by an infrared imaging detector (103).

Description

A SYSTEM AND METHOD FOR INFRARED IMAGING

The present invention relates generally to infrared imaging. Embodiments of the present invention relate to an infrared imaging system and method for multi-spectral infrared imaging of a sample as well as an image processing method for multi-spectral infrared imaging of a sample.

In the field of medical analysis, it is oftentimes required to examine a biological sample, i.e. a tissue sample from a patient, and determine its structure and composition, e.g. in order to arrive at a patient's diagnosis. This is typically done with a biopsy whereby a tissue sample is removed from the patient for examination.

Previously, in order to analyze a tissue sample, it was necessary to prepare a thin layer of the tissue sample and stain it using a stain and fixative. The fixative causes the stain to affix itself to a particular chemical compound thereby distinguishing the compound from its surroundings and allowing it to be identified. An histologist, when viewing the stained tissue sample layer through an optical microscope, would not only be able to determine the presence and location of the compound but also, by assessing the colour density of the stain, a qualitative estimation of the concentration of the compound. Also, the shape of the stain allows the histologist to determine the distribution of the compound in the sample.

By using a variety of different fixatives associated with different chemical compounds along with different coloured stains, it is possible to analyse the structure and composition of the tissue sample. Furthermore, by selecting a specific combination of stains and fixatives corresponding to a specific combination of chemical compounds that are found in a biological molecule, such as a nucleic acid, amino acid, protein or lipid; then the presence, concentration and distribution of the biological molecule within the tissue sample can be qualitatively estimated. A well trained histologist would be able to arrive at a diagnosis based on the colour, shape, degree of staining, and pattern of the stains.

Problems which arise with such an analytical technique include the complicated process of preparing the thin layer of stained tissue sample, which requires training to do, and destroys the biological sample. Furthermore, only qualitative, not quantitative, measurements can be taken.

An alternative technique for biological sample analysis, which does not require the chemical staining of samples under analysis, relies on infrared microspectrometry. This provides a method for identifying the chemistry of biologically active material in a biological sample in a non-destructive manner by analysing the spectra of the sample in the infrared part of the electromagnetic spectrum. The technique can equally well be applied in the fields of physical chemistry, solid state physics, material science and biology. Most biological molecules have vibrational modes with wavelengths which lie in the mid- infrared spectral range between 3 μm to about 16 μm. The positions, width and strength of the vibrational modes vary with composition and structure of the molecule. Identification of vibrational modes of major biological molecules, such as proteins, lipids and nucleic acids, can be determined by Fourier transform infrared spectroscopy.

Infrared radiation directed at a biological sample, e.g. a tissue sample, is variously absorbed or transmitted depending on the biological material present, i.e. compounds and functional groups present in the sample, as well as the concentration and distribution of the material in the sample. The sample's infrared spectrum exhibits characterising spectral features such as absorption bands of characteristic shape and size at characteristic frequencies. These characterising spectral features act as "fingerprints" by which to identify uniquely the presence of a particular functional group; moreover the presence of a certain functional group is indicative of a certain biological molecule. As an example, an absorption peak at a wavenumber of 1,240 cm^"1 relates to phosphodiester groups and is indicative of a nucleic acid. Likewise absorption peaks at 1,545 cm^"1 and 1,650 cm^"1 relate to amide groups which are indicative of proteins. Peaks at 2,850 - 2,960 cm^"1 relate to methylene and methyl groups found in proteins and lipids.

Accordingly, by measuring and analysing the infrared spectrum of a sample at a specific point on the sample using microspectrometry, one can uniquely identify the presence of certain biological molecules, such as: proteins, nucleic acid and lipids, at that specific point by their characteristic spectral feature fingerprints. Thus, using microspectrometry to measure and analyse the spectra of a sample at points over the entire surface of the sample, e.g. by raster scanning, one can compile a distribution map of the biological molecules in the sample. From knowledge of the distribution and concentration of such biological molecules, an histologist would be able to determine whether a biological sample was cancerous or not. However, since a large number of spectra are required to build up the distribution map, the time taken to generate the map is inordinately long, i.e. of the order of minutes or hours. Due to the time consuming process of preparing a distribution map, it is not possible to map the sample in real time. Accordingly, only stable, inert and stationary objects can be accurately mapped. Thus, such a measurement technique would be unsuitable for mapping e.g. live biological samples.

Previously, in order to measure a sample's infrared spectra, a continuous or full spectral source of infrared radiation is required. Typically, a 'glow bar' is used, i.e. a thermal source at a temperature of between IOOOK and 1500K. An alternative source of infrared radiation for infrared spectrometry for biological sample analysis, which provides a much higher radiant flux, is synchrotron radiation [see "Highly Resolved Chemical Imaging of Living Cells by Using Synchrotron Infrared Microspectrometry", by Nadege Jamin et al., Proceedings of the National Academy of Sciences of the United States of America, Vol. 95, No. 9, 28 April 1998, 4837-4840)]. A synchrotron source provides a radiant flux which is a thousand times that of thermal sources. However, notwithstanding the immense cost and complexity of a synchrotron, infrared radiation from a synchrotron has its drawbacks. Due to the low spectral radiant flux of synchrotron sources, the exposure time for each sample point is of the order of minutes. Furthermore, long exposure times themselves create further disadvantages such as causing the sample to overheat.

In order to improve the spectral radiant flux density (WHz^" W²) of infrared radiation generated from a synchrotron source, the infrared radiation is focussed to a region on the sample. The maximum size of the focussed region is dependent on the minimum spectral radiant flux density necessary to achieve an acceptable signal to noise ratio in the measurement of the spectrum. Accordingly there is a trade off between the size of the focussed region, i.e. the size of the region of the sample that is able to be sufficiently illuminated and thus analysed during a single exposure, and the spectral radiant flux density. The signal to noise ratio is improved by focussing the infrared radiation to a smaller region, thus reducing the sample region exposure time. However more exposures would be required to obtain spectra for the entirety of the sample, i.e. the sample would need to be split up in to a greater number of individual sample points which increases overall exposure time for analysing the entire surface of a sample. Thus, there is a further trade off between the resolution of a resultant distribution map and the length of time required to generate the map. For example, it would take a large amount of time to obtain all the many spectra for a sample when only a small region is exposed per spectra measurement, which would be necessary to compile a distribution map of a relatively high resolution. Typically, the synchrotron microspectroscopy technique uses a con-focal arrangement, which helps filter out undesired frequencies. However, it also means that the incident infrared synchrotron radiation is focused to a spot on the sample and a spectrum from the spot is measured. The spot is then scanned along the sample in a raster fashion and stopped at each new spot for the required exposure time. Since typical exposure time per spot is about one minute it can take hours to obtain spectra for the whole sample. Thus the total time involved in measuring the spectra of the whole sample is increased.

Furthermore, due to the long exposure time, it is not possible to image the sample in real time. Accordingly, only stable, inert and stationary objects can be accurately imaged. Thus, such a measurement technique would be unsuitable for mapping e.g. live biological samples.

The invention is set out in the attached claims.

Embodiments of the present invention seek to alleviate certain problems of the prior art and provide a system and method for multi-spectral infrared imaging, i.e. directly imaging a sample just at one or more discrete, i.e. non- continuous, frequencies in the infrared. Further embodiments provide a system and method for direct multi-spectral infrared imaging of a sample only at one or more specific discrete infrared frequencies, e.g. frequencies corresponding to the frequencies of a biological molecule's fingerprint thereby enabling the biological molecule to be directly imaged just at said frequencies. By producing and combining images of the sample at several specific frequencies associated with specific organic compounds, embodiments provide a useful adjunct tool to image the distribution of selected organic compounds. Such images can assist an histologist in determining the location of, for example, proteins, nucleic acids and lipids, and assist in diagnosing a tissue sample, e.g. to determine the presence or absence of cancerous tissue in a tissue sample.

The sample can be imaged across its entire extent instantaneously providing an improvement over previous devices, such as those using microspectrometry. Importantly, with previous devices, it was not possible to directly image the whole of a sample. Moreover, it was not possible to directly image the sample just at a particular desired frequency or narrow frequency band or at a set of desired frequencies or narrow frequency bands, e.g. at the frequencies of the characteristic spectral features associated with one or more particular biological molecules. This was because, where synchrotron sources were used, they were not able to generate infrared radiation of a sufficiently high spectral radiant flux to enable the entire surface of the sample to be illuminated with infrared radiation of a sufficiently high spectral radiant flux density at the particular desired frequency or frequencies.

Moreover, as will be discussed in greater detail below, embodiments of the invention allow multi-spectral imaging of a patient's tissue sample to occur in situ e.g. during surgery. Also, embodiments provide the ability to image the tissue sample with a short exposure period thereby enabling real time images and even real time video of a sample to be obtained which greatly facilitates the rapid diagnosis and treatment of the tissue sample.

Embodiments of the invention will now be described, by way of non- limiting examples only, with reference to the accompanying drawings, in which:

Figure 1 is a schematic cross-sectional view of an infrared imaging system according to an embodiment of the present invention;

Figure 2 is a schematic cross-sectional view of a reflective mode arrangement of an embodiment of the infrared imaging system; Figure 3 A is a schematic view of an optical parametric generator;

Figure 3 B is a schematic view a two-pass arrangement of an optical parametric generator;

Figure 4A is a plot representative of an angular tuning curve for a nonlinear crystal; Figure 4B is a plot representative of an emission output from an optical parametric generator;

Figure 5 is a schematic view of a further optical parametric generator;

Figure 6 is a schematic cross-sectional view of a further embodiment of an infrared imaging system;

Figure 7 is a schematic cross-sectional view of a yet further embodiment of an infrared imaging system;

Figure 8A is a plot representative of an absorption spectrum from a biological sample; Figure 8B is a plot representative of a transmission spectrum from a biological sample;

Figure 8C is a composite plot representative of the absorption spectrum from a biological sample at three specific frequencies;

Figure 9A is a schematic image of a biological sample imaged over the range of infrared frequencies ;

Figure 9B is a schematic image of a biological sample imaged at a first frequency;

Figure 9C is a schematic image of a biological sample imaged at a second frequency; and Figure 9D is a schematic image of a biological sample imaged at a third frequency.

The invention can be broadly understood with the following summary of the embodiment as shown in figure 1. The infrared imaging system 100 comprises a generator or source of infrared radiation 101. A source of mid- infrared radiation is used, preferably of a wavelength between 1 - 25μm or 3 - 20μm.

One suitable source is glow bar, which generates a source of continuous infrared radiation i.e. a broad band emission of infrared radiation, and a narrow band pass filter which filters the continuous infrared radiation emitted from the glow bar such that only a discrete narrow frequency range or narrow bandwidth of infrared radiation is emitted, for example having a narrow frequency range of up to: 10 THz, 5 THz, or 1 THz. Preferably, the filter is arranged such that it only allows the transmission there through of substantially just a single frequency and the narrow frequency range is up to: 10%, 5%, 3% or 1% of the single frequency. Yet more preferably, the filter is tunable such that the frequency which is transmitted there through can be selectively altered such that a predetermined frequency can be used. An example of such a selectively tunable filter is a Fabry-Perot interferometer or etalon.

An alternative source of infrared radiation is, an optical parametric generator, whose output is able to be tuned such that emits infrared radiation of a specific tuned frequency or narrow frequency range having the aforementioned frequency characteristics. The discrete narrow frequency range infrared radiation or frequency tuned infrared radiation is directed towards a sample 104 which, depending on its structure and composition, variously transmits or absorbs the incident discrete narrow frequency range of infrared radiation. The infrared radiation which passes through the sample is focussed by a focussing element 102, such as a mirror or lens arrangement, onto a high resolution infrared array detector 103, for example having resolution of 256 x 320, providing high spatial resolution images. Thus, an image of the sample is captured at just the narrow frequency range on infrared radiation. It is possible for an image of the sample to be produced whose spatial resolution is limited only by diffraction. With the glow bar and filter configuration as discussed above, the filter can be disposed between the glow bar and the sample such that it filters the continuous broad spectral infrared radiation emitted from the glow bar to a narrow frequency range. This filtered infrared radiation of a narrow frequency range is then directed to the sample for imaging by the detector so that an image of the sample is detected only at the narrow frequency range. Alternatively, the filter can be disposed between the sample and the detector (not shown) so that the continuous broad spectral source of infrared radiation from the glow bar is incident on the sample, but this is then filtered prior to its detection such that an image of the sample is detected only at a narrow frequency range. With either configuration, the input infrared radiation is transformed into infrared radiation having a narrow frequency range.

In order to analyse biological molecules/compounds, one is only interested in the spectral features obtained at predefined characteristic frequencies associated with certain biological molecules/compounds. Accordingly, to image a biological sample just at these predefined and separate characteristic frequencies, one requires infrared radiation of a sufficient intensity at these frequencies, i.e. a high spectral radiant flux (WHz^"1 which is the power emitted at a particular frequency). This way, an acceptable signal to noise ratio is achieved and good quality images at the desired frequencies can be obtained with short exposure times. Having short exposure times means that there is little time to accrue noise, e.g. due to background infrared radiation. Thus good quality images can be taken even in noisy background.

By tuning the generated output to different frequencies, the radiant flux is concentrated about the desired tuned frequency and thus a high spectral radiant flux is achieved. In contrast, previous generators of infrared radiation such as a synchrotron emit a non discrete continuum or 'white' infrared source, i.e. the radiant energy is equally distributed over the breadth of the infrared part of the electromagnetic spectrum. Consequently, previous infrared sources' spectral radiant flux were not particularly high. By providing a high spectral radiant flux, the entirety of a sample can be illuminated whilst still maintaining a sufficiently high spectral radiant flux density (WHz^" W²) over the entire surface of the sample. Accordingly, a shorter exposure time is required to capture an image of the sample at the desired frequency than was necessary for prior devices. Furthermore, the whole of the sample can be imaged with a single exposure compared to the multiple exposures per sample which was required in prior devices.

Embodiments of the infrared imaging system are able to image in fractions of a second as opposed to hours and thus can image non-stable objects and objects in motion such as biological samples. A further advantage achieved by the high spectral radiant flux is that embodiments of the imaging system can operate in a reflective mode as shown in figure 2 as well as the traditional transmission mode of figure 1.

In the reflective mode of figure 2, the discrete narrow frequency range infrared radiation reflected from the sample is directed to the detector via radiation guide element such as a semi transparent mirror 201.

Figures 3A and 3B show two possible arrangements of an optical parametric generator. Optical parametric generators provide coherent sources of continuously frequency tunable radiation and are discussed in "Parametric generation of tunable infrared radiation in ZnGeP₂ and GaSe pumped at 3 μm", K. L. Vodopyanov, Vol. 10, No. 9, September 1993, J. Opt. Soc. Am. B. and also "Mid- Infrared optical parametric generator extra- wide (3 - 19 μm) tuneability: applications for spectroscopy of two-dimensional electrons in quantum wells", K. L. Vodopyanov, Vol. 16, No. 9, September 1999, J. Opt. Soc. Am. B. both of which are incorporated herein by reference.

Figure 3 A shows a schematic diagram of an optical parametric generator 101 which comprises an infrared laser 301, such as an Er₅Cr: YSGG laser which provides 100 picoseconds long pulses of infrared radiation at a wavelength of 2.8 μm. This infrared radiation passes through a non-linear optical medium 302, such as an optically anisotropic nonlinear crystal, e.g. ZnGeP₂, CdSe or GaSe. These crystals are transparent to the incident infrared radiation. However, due to their nonlinear optical nature, a percentage of the incident photons from the infrared laser are divided into two photons, i.e. the input infrared radiation is transformed into infrared radiation having one or more separate narrow frequency ranges. The mechanism by which this occurs is via one of two types of processes known as Type I or Type II. Phase matching conditions determine which process takes place. Preferably, the Type II production mechanism is used as it provides signal photons with well defined discrete frequency range of infrared radiation, i.e. a sharper narrow band frequency bandwidth of infrared radiation. The photons emitted comprise two new light beams known as the signal wave, comprising signal photons and the idler wave, comprising idler photons.

The sum of the energy of the signal and idler photons is equal to the energy of the pump photon. Thus the signal and idler photons must satisfy the following equations:

where: ω = frequency; λ = wavelength p, _Sj and j refer to the pump, signal and idler photons respectively. The frequency of the signal and idler waves are determined by phase matching conditions which are changed when the angle between the incident pump laser beam and the optical axis of the crystal is changed (the optical axis is defined as an axis perpendicular to the direction of the incident infrared radiation from the laser) or when the temperature of the crystal is altered. The wavelengths, and correspondingly the frequency, of the signal and the idler waves, can therefore be tuned by changing the phase matching conditions, i.e. by rotating the non-linear optical crystal about its axis or altering the temperature of the crystal. The optical parametric generator thus provides a tunable output with a relatively narrow bandwidth line of infrared radiation, i.e. it provides a source of infrared radiation at a discrete narrow frequency range. The generated infrared radiation has a high spectral radiant flux that is readily and rapidly tunable to a desired frequency in the infrared part of the electromagnetic spectrum. If the laser provides pulses whose duration lasts lOOps, the tuned frequency infrared radiation is correspondingly also pulsed with lOOps long pulses, in effect providing a flash bulb to illuminate the sample to capture its image. The optical parametric generator can generate a beam of infrared radiation with a radiant flux of 100 MW. This beam can be expanded so as to spread the power over a wider area, for example a 1 cm², giving a radiant flux density of lOOMWcm^"2. The short duration of the pulses and the corresponding short exposure time in capturing an image of the sample means that in spite of the high radiant flux of the generated tuned frequency infrared radiation, the sample is not overly heated. Thus biological samples are not denatured. Furthermore, the extremely short exposure time provides a very high temporal resolution.

ZnGeP₂ and GaSe crystals are particularly suitable for the present invention due to their wide transmission range in the specific frequencies under consideration. For example, ZnGeP₂ is transparent to infrared radiation of 0.74-12 μm and GaSe is transparent to infrared radiation in the range 0.65-18 μm. Furthermore, by rotating the crystal about its axis, continuously tunable radiation in the range of 4 to 10 μm and 3.5 to 18 μm is possible for crystals of ZnGeP₂ and GaSe respectively. Figure 3 B shows an optical parametric generator which has been configured such that the infrared radiation from the laser 301 has two opportunities to pass through the non-linear optical crystal using the dichroic mirror 305 and mirror 306. Since the quantum efficiency of a pump photon being split into a signal and an idler photon is in the region of 10% depending on the nonlinear crystal used and the phase matching conditions present, by passing the pump photons through the nonlinear crystal twice, this greatly increases the chances that the pump photon will be split into a signal and an idler photon. Any generator of infrared radiation at a discrete narrow frequency range or any generator of tunable infrared radiation would be suitable. One possible alternative frequency tunable infrared radiation generator that could be used is an optical parametric oscillator, such as discussed in the two referenced papers by K.L. Vodopyanov. This operates in a similar manner the parametric generator of Figure 3B in that the pump source is arranged to pass through the nonlinear optical crystal as many times as possible, with the crystal effectively acting like a gain medium as in a conventional laser (with the difference being that the pump source is a coherent source, the infrared laser, as opposed to a non coherent pump source as typically found in a conventional laser). As with the optical parametric generator, the optical parametric oscillator can be frequency tuned by altering the phase matching conditions such as rotating the nonlinear optical crystal.

Embodiments of the invention, such as those employing an optical parametric generator or and optical parametric oscillator, provide an infrared generator that is tunable such that it is able to emit infrared radiation at a specific desired tuned frequency, there is no waste of radiant flux at frequencies that are of no concern. Furthermore, by being able to tune the output of the generator to a desired frequency or desired separate frequencies, there is no requirement for a complex arrangement of optical components to remove or filter out the undesired frequencies which may otherwise interfere with the imaging.

Figure 4A is a plot of an exemplary tuning curve for a non-linear optical crystal in an optical parametric generator. This shows how the wavelength of the optical parametric generator's output signal and idler photons varies with the rotation of the nonlinear optical crystal. The signal photons correspond to the lower half of the curve below the dotted line, having a shorter wavelength and higher frequency than the idler photons. Figure 4B shows a plot of the spectrum of a typical output of an optical parametric generator which emits infrared radiation within a discrete frequency range with a narrow bandwidth 402 centred about a specific tuned frequency 401. This frequency can be tuned to the desired infrared frequency.

Figure 5 shows an alternative configuration 500 of an optical parametric generator in which two or more non- linear crystals 302 and 503 are arranged such that the optical parametric generator can output infrared radiation comprising two or more specific desired discrete or separate frequencies tuned by the appropriate rotation of the non-linear crystals about their axis 303 or 304. The infrared radiation from the infrared laser 301 is directed by a first beam splitter 501 partly to non-linear crystal 302 and partly to a second beam splitter 502 which directs part of the infrared radiation a second non-linear crystal 503. If further separate narrow frequency ranges of infrared radiation are desired, then further beam splitters and non-linear crystals could be suitably arranged (not shown) thereby transforming the input infrared radiation to infrared radiation having one or more separate narrow frequency ranges..

Figure 6 shows a configuration 600 similar to that shown in Figure 5 but where a glow bar 601 and narrow band filters 604 and 607 are arranged so as to output infrared radiation comprising two specific desired discrete or separate frequencies having a narrow frequency range. Again, if further specific separate narrow bandwidth frequencies of infrared radiation are desired, additional beam splitters and narrow band filters could be suitable arranged. The infrared radiation from the glow bar 601 is partly directed along a first optical path 603 by a first beam splitter 602 to a first narrow band infrared frequency filter 604. The first narrow band infrared frequency filter only allows infrared radiation of a predetermined first discrete and narrow frequency range to pass there through. The infrared radiation of the first narrow frequency range is then directed to the sample and imaged by the detector (not shown). The infrared radiation from the glow bar 601 is also partly directly along a second optical path 606 by a second beam splitter 605 to a second narrow band infrared frequency filter 607. The infrared radiation of the second narrow frequency range, which is different to the first, is then directed to the sample and imaged by the detector (not shown). Such a configuration enables the input infrared radiation from the glow bar to be transformed into infrared radiation having one or more separate narrow frequency ranges. Optionally, shutters 608 and 609 are provided on the first and second optical paths respectively to gate the infrared radiation in order to control selectively the passage of infrared radiation there through such that an image of the sample just at the first or second frequency can be detected separately. Alternatively, an image of the sample can be detected at both the first and second frequencies simultaneously.

Preferably, the two or more desired narrow frequency ranges of infrared radiation outputted by the embodiments shown in figures 5 and 6 correspond to frequency ranges that are different, separate and distinct, i.e. are not overlapping frequency ranges. Yet more preferably, the frequencies correspond to frequencies at which spectral characteristics associated with a certain biological molecule occur. The advantage of such arrangements is that they allow a single infrared source to produce simultaneously two or more desired frequencies having a narrow frequency range as opposed to only producing a single desired frequency and then having to retune the narrow frequency range infrared radiation generator (i.e. for an optical parametric generator - rotating the non- linear crystal between each image capture) or for a glowbar and narrow band filter - changing the narrow band frequency at which the narrow band filter transmits) to produce a second desired frequency. Accordingly, in a single exposure, a multi-spectral image can be captured at two or more specific desired discrete narrow frequency ranges thereby further reducing the time which would otherwise have been required to capture two or more separate images at two or more differing narrow frequency ranges.

Figure 7 shows a further embodiment of infrared imaging system comprising a tunable generator of narrow frequency range infrared radiation

101 which emits a beam of frequency tuned infrared radiation that is expanded by beam expander 701. The beam can be expanded to illuminate the entire sample. The expanded beam then passes through a rotating diffuser 702 which advantageously diffuses the highly coherent frequency tuned infrared radiation from the optical parametric generator to remove speckle, which can interfere with the imaging of the sample. The expanded and diffused beam of frequency tuned infrared radiation is focussed by focussing element 703 onto the sample 104. Infrared radiation from the sample is then focused by focussing element

102 onto the detector 103. Optionally, a selective filter 704 can be provided to block or filter infrared radiation incident on the detector. Alternatively, the registration period of the detector can be gated in time by electronic means, i.e. the detector itself effectively acts as a gate. However, since such gating can only be achieved at timescales of microseconds compared to the 100 picoseconds variable output from the optical parametric generator, it is preferred to use the time profile of the laser pulse to define the time interval over which the image is acquired.

Figure 8A shows an exemplary absorption spectrum obtained from a biological cell. The peaks 801, 802 and 803 correspond to spectral features that are associated with a particular functional group. These spectral features make up the fingerprint which indicates the presence of a particular biological molecule in the cell. Accordingly, one can use such spectral features to identify the chemical composition of the cell such as identifying proteins, nucleic acids and lipids. Figure 8B shows the corresponding transmittance spectrum with transmittance troughs 801', 802' and 803' which correspond to the absorption peaks 801, 802 and 803. Where the infrared radiation at frequencies corresponding to features 801, 802 and 803 are strongly absorbed, i.e. have a low transmittance, this indicates the present of a particular functional group associated with the spectral features.

Figure 8C shows an exemplary composite spectrum that would be obtainable with embodiments of the present invention where the detector comprises a spectrometer. Since, in embodiments of the present invention only infrared radiation of a specific infrared frequency corresponding to a narrow bandwidth of infrared radiation is detected, the spectrum obtainable is likewise constricted to the bandwidth of the tuned infrared radiation. Accordingly, one obtains spectrum measurements in only the desired frequency bandwidth which encompasses the bandwidth of the spectral feature under analysis. If more than one frequency bandwidth is desired to be analysed, the imaging system can be tuned to the further desired frequency and a further measurement can be taken, as shown by the dotted and dashed spectrum lines. By combining the spectra obtained from separate exposures, or by imaging at two or more separate narrow ranges of frequencies simultaneously, one can create a composite spectrum or image from which one could determine whether or not the associated functional group is present and thus can use the spectral fingerprint to specify the proteins, nucleic acids and lipids of the sample.

Figure 9A to 9D show schematic images of a biological cell captured at different frequencies. The image of 9A is an exemplary image that would have been obtained by imaging with a bright, non-discrete, full spectral continuum infrared radiation source. Here, all chemical compositions which absorb any frequency of the incident infrared radiation appear opaque. This makes is difficult to identify and determine the shapes of the individual biological molecules 901, 902 and 903, especially where the molecules overlap one another.

By contrast, as shown in figures 9B to 9D which relate to images obtained from embodiments of the present invention taken at three different tuned frequencies each frequency being a characteristic frequency corresponding to absorption peaks of particular associated biological molecule, 901, 902 and 903 respectively. Only the biological molecule which absorbs infrared radiation of its specific respective frequency is visible as it is opaque to the frequency tuned infrared radiation whilst the other material in the cell is substantially transparent at the specific frequency. Thus the shape and position of the respective biological molecules 901, 902 and 903 are more easily determined and distinguished from other material in the sample in each image 9B, 9C and 9D respectively. Thus, embodiments of the invention provide the beneficial effect of enabling the distinguishment of various imaged features.

The tables below show the frequencies of spectral features characteristic of certain functional groups which themselves are indicative of certain biological molecules. Preferably the imaging system is tunable to the frequencies at which the following spectral features occur:

Each image 9B to 9D can be thought equivalent to an image of a stained biological sample. Where previously a biological sample was stained with a stain and fixative which latches onto a certain chemical compound associated with a biological molecule, thereby readily distinguishing the molecule from the rest of the sample, embodiments of the present invention, by imaging a biological sample at certain specific fingerprint frequencies associated with a biological molecule, readily distinguishes the molecule from the rest of the biological sample. This is due to the fact that the molecule is particularly absorbent at its fingerprint frequencies, and is more absorbent than the rest of biological material in the sample at those frequencies. Thus the molecule would appear more opaque than the rest of sample. A "digitally stained" image can thus be created when the images are loaded onto a computer and a composite image is created from several images, each captured a specific frequency.

In one simplified use of the invention, the images in figure 9B, 9C and 9D could correspond to images taken at frequencies corresponding to spectral features 801, 802 and 803 respectively of figure 8A. However, in a more realistic situation it may be that, for example, where all the fingerprint spectral features 801, 802 and 803 are present, this indicates the presence of a biological molecule, e.g. protein. Accordingly, if figure 9B were to correspond to an image of the cell taken at the three respective fingerprint frequencies 801, 802 and 803, the shapes would correspond to the position of that molecule within the cell. Likewise, if an alternative set of images were combined at different fingerprint frequencies which correspond to frequencies associated with fingerprint spectral features of another biological molecule, e.g. a nucleic acid, then the resultant composite image would map the presence of the nucleic acid in the cell. As an example, if one desired to map the distribution of a certain protein within a cell, and one knew that the protein produced a characteristic fingerprint corresponding to specific spectral features which are associated with a functional group which make up the protein, then one would image the cell at the frequencies where the spectral features are present and combine the images to form a composite image. If the chemical species which exhibit the characteristic spectral features are present, then they will absorb the infrared radiation at these frequencies. Accordingly, the absence or presence of the chemical species would be clearly identifiable in the composite image. Accordingly, the absence or presence of the biological molecule which comprises the chemical species would likewise be readily apparent. By taking a series of images at the requisite frequencies to identify a particular molecule such as a protein, nucleic acid or lipid, one can readily determine the presence and morphological shape of the same. An histologist, when viewing such images would be able to determine not only a qualitative estimation of the distribution and amount of particular molecule in a sample but also, since the images can be directly digitally processed, manipulated, and the image characteristics such as shape and intensity of a feature in an image can be accurately measured, a quantitative determination and measurement of the amount of a particular molecule in a sample can be obtained. Such information can be used by an histologist to greatly facilitate the diagnosis of a condition.

During an operation to remove a cancerous tumour from a patient, not only is the tumour itself removed but also material surrounding the tumour. This is done to ensure that all of the cancerous parts of the body have been removed. However, during such an operation, it is difficult for a surgeon to distinguish between healthy tissue and cancerous tissue. In particular, it is difficult for the surgeon to determine the point at which the tissue being removed from the body is no longer cancerous. Accordingly, a surgeon would err on the side of caution and remove an excess of healthy tissue rather than risk leaving behind any cancerous tissue. In order to minimise the amount of healthy tissue that is unnecessarily removed, it would be beneficial to be able to perform substantially real-time analysis of a tissue sample, preferably in situ, to determine whether or not the tissue is cancerous and if further tissue needs to be removed. Using embodiments of the present invention, a user could combine detected images captured at various frequencies to create and view a composite image, e.g. using a conventional computer. More preferably, the system would enable the user to manipulate or selectively control the combination of such images, for example by: assigning a particular colour to each image and/or adjusting the relative weighting of each image (i.e. by adjusting its transparency or degree of opacity) to create an adjustable false colour composite image. Such an adjustable composite image can be though of as a "digitally stained" image of the sample, which can be fine tuned to highlight the desired biological molecules. This provides a useful adjunct tool that would assist an histologist in analysing a sample and arriving at a diagnosis and appropriate treatment. Thus, embodiments of the present invention provide an improved system and method of analysing chemical and biological samples.

Of course, the multi-spectral infrared imaging system is not limited to imaging of just biological samples, and non biological samples can equally well be imaged.

The invention is not restricted to the features of the described embodiments. It will be readily apparent to those skilled in the art that it is possible to embody and implement the invention in specific forms other than those of the preferred embodiments described above. The invention is defined by the following claims.

Claims

1. An infrared imaging system (100) for multi-spectral infrared imaging of a sample comprising: a source of infrared radiation (101, 301, 601); means for transforming the infrared radiation to infrared radiation having one or more narrow frequency ranges (101,302,604); and an infrared imaging detector (103) for detecting one or more images of said sample at the one or more narrow frequency ranges.

2. An infrared imaging system as claimed in claim 1 wherein each of the one or more narrow frequency ranges is a separate narrow frequency range.

3. An infrared imaging system as claimed in any previous claim wherein each of the one or more narrow frequency ranges has a frequency range of up to: 10 THz, or preferably 5 THz, or yet more preferably 1 THz.

4. An infrared imaging system as claimed in any previous claim wherein each of the one or more narrow frequency ranges comprises substantially a single frequency and further wherein each of the one or more narrow frequency ranges has a frequency range of up to: 10%, 5% or preferably 3%, or yet more preferably 1% of the single frequency.

5. An infrared imaging system as claimed in any previous claim wherein the means for transforming the infrared radiation to infrared radiation having one or more narrow frequency ranges comprises means for selectively transforming the infrared radiation to one or more predetermined narrow frequency ranges.

6. An infrared imaging system as claimed in any previous claim wherein the means for transforming the infrared radiation comprises one or more narrow band infrared filters.

7. An infrared imaging system as claimed in claim 6 wherein each of the one or more narrow band infrared filters is arranged such that only infrared radiation of substantially a single frequency passes there through.

8. An infrared imaging system as claimed in any of previous claims 1 to 5 wherein the means for transforming the infrared radiation comprises one or more nonlinear optical media (302).

9. An infrared imaging system as claimed in claim 8 wherein each of the one or more nonlinear optical media comprises an optically anisotropic nonlinear optical medium, preferably comprising a crystal of one of: ZnGeP₂, CdSe or GaSe.

10. An infrared imaging system as claimed in claim 8 or 9 wherein each of the one or more nonlinear optical media is arranged so as to be rotatable about an optical axis (304, 303).

11. An infrared imaging system as claimed in any previous claims further comprising means to control selectively the passage of infrared radiation there through.

12. An infrared imaging system as claimed in any previous claim wherein the system is arranged such that infrared radiation from the sample region is directed onto the infrared imaging detector in at least one of: a reflection mode or a transmission mode.

13. An infrared imaging system as claimed in any previous claim wherein the infrared radiation is mid-infrared radiation.

14. An infrared imaging system as claimed in any previous claim wherein at least one of the one or more narrow frequency ranges of infrared radiation encompasses at least one of:

28.9 THz, 29.9 THz, 37.2 THz, 46.3 THz, 52.2 THz, 49.5, THz, 85.4 - 88.7, THz or 99.0 THz.

15. An infrared imaging system as claimed in any previous claim wherein the infrared radiation has a wavelength between 1 - 25μm or, more preferably, 3 - 20μm.

16. An infrared imaging system as claimed in any previous claim further comprising means to capture and store the one or more images detected at the one or more narrow frequency ranges and combining the images to produce a composite image.

17. An infrared imaging system (100) as claimed in claim 16 wherein the manner in which the images are combined is determined by an algorithm.

18. A method for multi-spectral infrared imaging of a sample comprising the steps of: generating a source of infrared radiation; transforming the infrared radiation to infrared radiation having one or more narrow frequency ranges; and detecting one or more images of said sample at the one or more narrow frequency ranges.

19. A method as claimed in claim 18 wherein each of the one or more narrow frequency ranges is a separate narrow frequency range.

20. A method as claimed in any of claims 18 to 19 further comprising the step of combining the one or more images to produce a composite image of the sample at the one or more narrow frequency ranges.

21. A method as claimed in any of claims 18 to 20 wherein the infrared radiation is mid- infrared radiation or has a wavelength of 1 - 25 μm or more preferably 3 - 20μm.

22. A method as claimed in any of claims 18 to 21 wherein the infrared radiation has a wavelength of at least one of: 10.37 μm, 10.05 μm, 8.06 μm, 6.47 μm, 5.75 μm, 6.06 μm, 3.51

- 3.38 μm or 3.03μm

23. A method of analysing a biological sample comprising the method as claimed in any of previous claims 18 to 22.

24. A biological sample analyser comprising a multi-spectral infrared imaging system as claimed in any of claim 1 to 17.

25. A computer program arranged to selectively control the combination of the one or images of any previous claim.

26. A computer program as claimed in claim 25 wherein the selective control of the combination of the images comprises: allowing a user to select a weighting that is to be applied to at least one of the images.

27. An infrared imaging system (100) for imaging a sample comprising: a frequency tunable infrared radiation generator (101); and a detector (103) for detecting infrared radiation.

28. A method for imaging a sample comprising the steps of: tuning a frequency tunable infrared generator; directing the tuned frequency infrared radiation at a sample region; detecting the infrared radiation from the sample region; and producing an image of the sample region.