GB2273559A - Method and apparatus for examining an object - Google Patents

Abstract

In order to examine a gemstone 1, the stone is irradiated 3 and observed through a filter 5 at a first wavelength which is characteristic of a first class of gemstones. The filter 5 can be rocked from a setting normal to the optical axis to transmit radiation at at least two reference wavelengths. The intensities of radiation transmitted at the first and reference wavelengths are observed and compared, to classify the gemstone 1 as belonging to the first class or not. <IMAGE>

Description

METHOD AND APPARATUS FOR EXAMINING AN OBJECT Background of the Invention In general terms, the invention relates to examining or classifying an object by detecting the spectral properties of the object. The invention is particularly, but not exclusively, concerned with identifying gemstones such as diamonds, e.g.

distinguishing diamonds from diamond-like simulants and distinguishing natural diamonds from synthetic diamonds.

WO 86/07457 discloses a method for distinguishing diamond from diamond like simulant, by visually detecting the Raman signal emitted from a specimen which is irradiated with suitable exciting radiation. The Raman emission has two peaks, one on either side of the wavelength of the exciting radiation, termed the Stokes signal and the anti-Stokes signal. The Stokes signal is much stronger than the anti-Stokes signal, but it is still very weak. One of the problems is that if a diamond-like simulant luminesces, it is very hard to discern the appropriate Raman peak against the luminescent background.

Diamond simulant comprises dense non-diamond material (eg. metal oxides, particularly zirconium dioxide) which has similar refractive properties to diamond. Synthetic diamond comprises diamond material (ie. crystalline carbon) produced by an industrial process.

The technique disclosed by WO 86/07457 is only suitable for distinguishing diamond from diamond-like simulant.

All diamonds, natural or synthetic, show the Raman emission when irradiated with suitable exciting irradiation, and cannot be distinguished by this technique.

The Invention.

The invention provides a method of and apparatus for classifying an object as set out in Claim 1, 16, 21, or 22 Preferred and/or optional features are set out in Claims 2 to 15 and 17 to 20.

At least in its preferred forms, the invention enables diamonds and other suitable gemstones to be examined and classified by operators with little scientific or technical training.

According to a first embodiment of the invention, the object is irradiated with stimulating radiation and the emission/luminescence of the object is examined. If it is desired to distinguish diamond from diamond simulant, it is preferable to use a laser that will cause Raman activation in the visible spectrum; a suitable Raman wavelength is about 552. 4 nm which can be produced by an argon ion laser operating at 514. 5nm, and in general terms the laser wavelength may lie between 450 nm and 1064 nm, but may be outside this range.

As the band passed by the alterable filter is faltered, the signal detected will be markedly different, depending on whether the object or gemstone is say a diamond, or a diamond simulant which does not exhibit Raman emission at the correct wavelength, and does not luminesce, or a diamond simulant which luminesces. This is explained later with reference to the drawings.

Though it may be possible to use other methods, a simple method of altering the band passed is by tilting the filter about an axis normal to its optical axis. With narrow band pass filters, the cut-offs are most clearly defined when the filter is correctly orientated with its optical axis; as the filter is tilted, the centre of the band passed changes, and the band widens - this w n -lg is not essential to the preferred embodiment of the invention, but is an incidental effect.

According to a second embodiment of the invention, the object is irradiated with light in the longwave ultraviolet/visible part of the spectrum and the absorption spectrum of the object may be studied by measuring the intensity of light absorbed by the object.

The object may be illuminated with a lap running off a mains electricity supply. A change in the lamp supply voltage can alter the temperature of the emission source of the lamp and thus the spectral distribution of its output energy may vary. Provision should be made to detect this variation so that parameters such as ratios between transmitted wavelengths can be corrected for errors introduced by the spectral variation. By making more than two observations of the absorption spectrum of the object, any spectral shift due to lamp variation can be detected and compensated for.

If natural diamond is to be distinguished from synthetic diamond, the absorption (or, equivalently, transmission) spectrum may be observed by measuring the absorption at 415. 5nm and at least two slightly different reference wavelengths, say 410nm and 418. sum. The absorption should be measured at three wavelengths very close together, as this will help to clearly identify a characteristic absorption. 415. 5nm is a very strong absorption, characteristic of diamonds of type IaAB.

418. 5nm and 410nm fall outside the absorption peak for this characteristic absorption and so the absorption is relatively low. In general, diamonds of different types to IaAB do not show a very strong absorption at 415. Snm, and even if there is some absorption, it will not be very much different from the absorption at 410no and 418. 5nm.

Accordingly, diamonds of type IaAB can be positively distinguished, and as diamonds of this class of diamonds are effectively always natural in origin, the second embodiment of the first aspect of the invention allows all diamonds encountered to be classified as belonging to a class comprising definitely natural diamonds or a class containing diamonds which may or may not be natural. This will be explained further below.

The apparatus may be very simple to use and construct, as it only has a small number of components. The whole apparatus may only occupy a space of about 25 x 10 x 15 cm, being suitable for use on a bench top. The method does not require any great skill on the part of the operator and is suitable for producing an answer very quickly.

The Drawings.

The invention will be further described by way of example, with reference to the accompanying drawings, in which: Figure 1 is a schematic side view of apparatus in accordance with a first embodiment of the invention; Figure 2 illustrates the bands passed by the two filters in Figure 1; Figures 3a to 3c illustrate what occurs in figure 1 when a diamond with no luminescence is examined, showing the signal position, the no signal position, and the result of rocking; Figures 4a to 4c correspond, but show the situation for a diamond with luminescence; Figures 5a to Sc correspond, but show the situation with a diamond simulant without luminescence; and Figures 6a and 6c correspond, but show the situation with a diamond simulant with luminescence; Figure 7 shows an example of a portion of the absorption spectrum of type Ib diamond; Figure 8 shows an example of a portion of the absorption spectrum of type IaAB diamonds; Figure 9 shows a high resolution transmission spectrum for a type IaAB diamond, between 410nm and 420nm; Figure 10 shows an apparatus for observing a gemstone according to the second embodiment of the invention Figure 11 shows the filter of the apparatus of Figure 12 in first and second positions; Figure 12 shows the variation with angle of incidence of band pass characteristics of the filter of Figures 10 and 11; Figure 13 shows the use of three observations to fit a curve; ; Figure 14 is a schematic illustration of apparatus according to a third embodiment of the invention; Figure 15 shows a flow chart for use with the invention.

Figures 1 to 6 Figure 1 shows a stone 1 on a dop 2, the stone 1 being illuminated with a laser 3 which may be an argon ion laser operating at 514. 5nm (shown as 9 in Figure 2), thereby stimulating luminescence of the stone 1 if tne stone 1 is capable of it at the stimulating wavelength.

The stone 1 is viewed by eye through a laser blocking filter 4 which rejects light at laser wavelength (for example a Sehott OG 530 rejects the laser light at 514. 5nm) and a narrow band pass filter 5, for example a lnm FWHM 4 cavity design centred at 552. 4 nm, manufactured by the Andover Corporation in the USA. The narrow band pass filter 5 can be tilted about an axis 6 normal to its optical axis, and a drive 7 is shown for rocking or oscillating the filter 5.

The band passes of the narrow band pass filter 5 are shown in Figure 2, which shows graphs of transmissivity T against wavelength X for incident light parallel to and for incident light inclined to the optical axis of the filter 5. The continuous lines indicate the bands passed when the filter is set up properly, normal to the optical or viewing axis. The filter 5 passes a narrow band of width about 1 nm (measured at the height of half maximum transmissivity), centered on the Raman emission (Stokes) at 552.4 nm. It can be seen that with this set-up a narrow band of detectable radiation wavelengths can pass - in this case, detectable radiation is radiation that can be detected by the eye.

As the filter 5 is tilted about the axis 6, its optical characteristics alter in that its cut-on moves to lower wavelengths and the band passed widens - this is illustrated in the dashed lines in Fiqure 2.

The filter 5 can be set-up so that it is rocked from a first setting represented by the continuous line in Figure 2 to a second setting represented by the dashed line in Figure 2. This rocking can be done by hand, or, as indicated in Figure 1, by a motor drive 7.

Figures 3a-3c illustrate what occurs when the stone 1 is a non-luminescing diamond. Figures 3a and 3b are graphs of T against X, while Figure 3c is a graph of intensity i against time t. The Stokes emission (illustrated as the peak in the emission 8) is passed when the filter 5 is in its first setting, and thus there is a signal (Figure 3a). When the filter 5 is in its second setting, the Stokes emission 8 is blocked and there is no signal (Figure 3b). As the filter 5 is rocked between the first and second settings, there is a an approximately square-wave signal with no dc shift, i.e. the troughs are at zero intensity (Figure 3c).

If the stone 1 is a diamond with luminescence, there will be a signal when the filter 5 is at normal incidence (Figure 4a) comprising the Raman and luminescence background. When the filter 5 is tilted (Figure 4b) a signal corresponding to luminescence background is produced . As the filter 5 is rocked between the first and second settings, there is an approximately square-wave signal with the troughs at non-zero intensity (Figure 4c).

If the stone 1 is a diamond simulant with no luminescence, there is no signal at all (Figures 5a to 5c).

If the stone 1 is a diamond simulant with luminescence, in the first setting of the filter 5 there will be a signal. However, as the filter 5 is gradually moved from the first setting to the second setting, its band pass width slightly increases whilst its peak transmission drops slightly; in many cases, to a good first approximation the total integrated light transmission is unchanged and the total signal from the background luminescence spectrum is approximately constant. Thus, the signal is as in Figure 6c.

Instead of the human eye, a suitable photodetector 10 can be used. The photodetector can be associated with processing equipment to distinguish between the types of signals shown in Figures 3 to 6 inclusive.

Figures 7 t 9 One way of classifying diamonds is according to their spectroscopic properties. The absorption spectrum of a diamond in the visible region will determine its colour. To a certain extent, it is possible to associate each type of diamond with a range of structure, concentration and composition of impurity defects. An analysis of diamonds in this manner gives the following classification.

Type I This general type class is defined as the class of diamonds which have a measurable defect induced infra-red absorption in the 1-phonon region (below 1332cm 1). The defects result from the incorporation of nitrogen atoms into the crystal lattice substituting for carbon atoms during growth of the diamond. Natural type I diamonds will typically contain several hundred to a few thousand ppm of nitrogen. The content of nitrogen in synthetic diamonds can be controlled during the process of synthesising the diamonds. This gives a range of nitrogen atom content of a few hundred ppm to practically zero in synthetic diamonds.

The general class type I is divided into the following subtypes; Type Ib In this type of diamond single nitrogen atoms are substituted for single carbon atoms at random throughout the lattice. This gives rise to an optical absorption starting at about 600nm which continues with increasing strength into the longwave ultra-violet region (Figure 7). This gives rise to the so-called canary yellow colour shown by some diamonds. Type Ib diamonds represent a non-equilibrated form of diamond. Diamonds are formed at conditions of very high temperature and pressure, and if the diamond is maintained at these conditions impurity nitrogen atoms will tend to aggregate.Natural diamonds were usually maintained at these equilibrating conditions for geologically significant periods of time and accordingly type Ib diamonds are rare in nature (much less than 1% of all natural diamonds). On the other hand, synthetic diamonds are not maintained at equilibrating conditions and accordingly most synthetic diamonds are type Ib.

Type Ia This class comprises diamonds in which the nitrogen has migrated to form more complex defects. There are two pr;ncipal forms of nitrogen defect which are found in type Ia diamonds, the A form and the B form. The A form comprises pairs of nitrogen atoms on nearest-neighbour substitutional sites. The B form of nitrogen is believed to comprise a complex of four substitutional nitrogen atoms surrounding a vacancy. The ratio of the concentration of A type defects to B type defects varies continuously, the extreme ends of the sequence being labelled type IaA and type IaB. Pure type IaB diamonds are very rare. Synthetic type Ib diamonds can be converted to type IaA by a high-temperature and high-pressure treatment.

Type IaA diamonds have no absorption in the visible region of the spectrum so they are colourless. There is very little visible absorption associated with B centres, and as a result IaB diamonds are colourless.

Most natural diamonds contain both A and B centres and are known as type IaAB. In addition to the two principle forms of nitrogen defect, they contain two "by-products" of the nitrogen aggregation process: platelets and N3 centres. Platelets are interstitial planar defects, a few tens of nanometres in diameter lying on cube planes. These give rise to a peak in the infra-red spectrum. N3 centres comprise three co-planar nitrogen atoms probably surrounding a vacancy. N3 centres give rise to absorption between 490nm and 350nm with a sharp zero-phonon line at 415. 5nm. This absorption in the blue/violet region causes the so-called cape yellow colour exhibited to a greater or lesser extent by the vast majority of natural diamonds (Figure 8).Figure 9 is a high resolution transmission spectrum showing the 415. 5nm absorption of a type IaAB diamond in more detail. It can be seen that there is a strong decrease in transmission at about 415. 5nm, transmission being much higher at other wavelengths, for example 410nm.

Type IIa This class comprises diamond in which nitrogen is only present in trace amounts, of the order of 1 ppm. There is often a form of background absorption at the shorter wavelength end of the visible spectrum, giving some of these diamonds a generally brown colour. This near absence of nitrogen in diamonds rarely occurs in nature (less than 2% of natural diamonds are type IIa) but can be assurred in the production of synthetic diamonds.

Type IIb This is a very rare class of semiconducting diamonds in nature. The diamonds contain trace amounts of substitutional boron as semiconductor acceptor centres which give the diamonds a bluish colour due to the tail of the photoionization spectrum at the acceptor centre.

Type IIb diamonds are generally natural, but synthetic diamonds containing added boron can be produced.

In all, most natural diamonds are type IaAB and IaA, only about 2% being II, Ib or IaB.

Figure 10 Figure 10 is a schematic drawing a second embodiment of apparatus according to the invention, which is set up to classify a finished diamond as definitely natural or not definitely natural.

A diamond 12 is illuminated with radiation generated by a halogen lamp 13 of a suitable wavelength. The illuminating radiation is fed to the diamond via a fibre optic 14 and, in the case of a brilliant cut diamond, the light is fed in through the table of the diamond. A brilliant cut diamond is intended to be viewed through the table and is-so shaped that the maximum amount of light is reflected by the lower faces of the diamond back out of the table 15.

In order to study the absorption spectrum of the diamond, a second fibre optic 16 is provided to collect light leaving the diamond via the table 15. Transmitted light is fed via the fibre optic 16 into detector apparatus 17 which includes a filter 18. A photomultiplier tube or other photodetector 19 is provided to give a signal dependent upon the intensity of light passed by the filter 18, which signal is fed to an amplifier 20 and then to a microprocessor 21.

The filter 18 is rotatable about an axis 25 to transmit light at different wavelengths, being driven by a motor 22. The motor 22 can be controlled by the microprocessor 21, a transducer 23 comprising a shaft encoder or the like being provided to give a signal indicating the position of the filter 18. In order for the readings taken by the apparatus to be simply presented and easily understood, a visual display unit (VDU) 24 may be provided receiving signals from the microprocessor 21.

As shown in Figure 11, the filter 18 can be rotated about an axis 25 normal to its optical axis 26 into a tilted position (as shown at 18'). The band pass characteristics of the filter 18 vary with the angle 5 between the optical axis 26 of the filter and the direction of incident light 27. Figure 12 shows the band pass characteristics of a CWL = 418. 5nm filter at various values of e. It may be seen that as o increases, tbe maximum of the transmission moves to lower wavelength, the transmission maximum decreases in intensity and the width of the band passed increases.

The full width at half maximum for the filter where 6 = 0 is lnm. Such a filter is manufactured by Omega Optical Company in the USA.

Thus, if the filter 18 is tilted through a variety of angles o by the motor 22, a region of the absorption spectrum of the diamond 12 may be scanned and sampled.

The apparatus shown in Figure 10 can be used to classify a diamond as belonging to type IaAB or not. A filter 18 having the band pass characteristic shown in Figure 12 is used, so that a signal can be derived representive of the absorption of light at 415. 5nm. On its own, this signal does not give much useful information unless it is normalized, because the signal will vary with the size of the diamond. Furthermore, diamonds of type IaAB will vary greatly in the absorption co-efficient at 415. 5nm between themselves, and no positive range can be assigned to clearly identify a diamond of type IaAB on the basis of this one uncorrected absorption signal alone. Accordingly, a second measurement is taken at a reference wavelength of 410nm for example. This lies completely outside the absorption peak at 415. 5nm and is of higher energy.

The lamp 13 used to illuminate the diamond may be an halogen lamp, for example a 12 volt, 12 watt Thorn type M64 with a Spindler and Hoyer lens 063097. This form of lamp operates at about 3,000K with a peak towards the red end of the visible spectrum. The wavelengths to be observed lie on a steep part of the thermal radiation curve. Accordingly, if the temperature of the lamp shifts to, say, 3,200K due to a perturbation in the power supply, the shape of the curve will vary and the intensity of light at the wavelengths to be observed will vary quite markedly, the ratio of the intensities between the two irradiating wavelengths will vary, and so the reading based upon the intensity of radiation absorbed at these wavelengths by the diamond can be in error. In order to detect this error, a third measurement may be made at a wavelength of, for example 418. 5nm.

Preferably, a series of measurements are made in the region 418. 5 to 410nm, and the absorption results interpreted by a curve fitting technique, operated by the microprocessor 21 to detect if a 415.on absorption is in fact present.

The filter 18 is rotated at high speed (3,000 rpm) about its axis 25 and the absorption of light at various wavelengths (deducible from the angle e of the filter, measured by transducer 23) measured many times over and stored by the microprocessor. Thus instead of just three readings, a mass of data is obtained quickly and simply which can be analysed by a statistical technique to provide more accurate information on the absorption characteristics of the diamond. This improves the repeatability of the test and reduces the error.

The microprocessor 21 can be programmed to compare the readings directly and to produce a signal representative of whether the diamond is natural or should be tested further, or all the readings may be shown numerically, or graphically on the VDU 24.

Figure 13 shows how the three measurements at 410, 415. 5 and 418. 5nm are used in the microprocessor to plot a curve showing the absorption characteristics of the diamond in this region of the spectrum, so that an absorption at 415. sum can be clearly identified.

Three absorption curves are shown, for diamonds of type IaAB showing the cape yellow colour with varying degrees of intensity.

Instead of correcting the first and second wavelength measurements using a third wavelength and a curve fitting technique, the spectrum of the lamp 13 may be sampled directly, using a reference channel. A third fibre optic may be used, leading directly from the lamp 13 to a detector, which passes data to the microprocessor 21.

Being able to classify a diamond as type laAB or not will be useful to the jeweller or other craftsman in identifying natural diamonds, as the vast majority of natural diamonds belong to class IaAB and (because of the complexity of the defects and the fact that they take a long time to develop) synthetic diamonds are effectively never type IaAB. Thus the apparatus of the invention can be set up as above, to divide all diamonds into one of two classes: definitely natural; possibly natural, possibly synthetic; The number of natural diamonds classified in the second class by the apparatus of Figure 10 will be very small (about 2%), comprising type Ia, Ib, Ila, IIb and IaS. or IaB diamonds, which are all very rare.

The apparatus of Figure 10 can also be used to measure the colour of a cape diamond, by measuring the strength of the 415. 5nm absorption.

In the apparatus shown in Figure 10, the lamp 13 may be a 12 volt 12 watt halogen lamp manufactured by Thorn, type M64, using a lens 063097 manufactured by Spindler and Hoyer Limited. Suitable fibre optic cable is manufactured by Schott. The lenses shown in the detector 17 are, from left to right, a lens 063097 and lens 063045 respectively, both manufactured by Spindler and Hoyer. The photomultiplier tube 19 can be of the type manufactured by Hamamatsu KK.

Figure 14 Figure 14 shows a third embodiment of apparatus for classifying a diamond according to the invention. In this apparatus, radiation is produced by a source 28 and fed into the system via mirror 29 through a wheel 30 having a broad band pass filter 30a for excluding infra red radiation or aperture 30b with infra red filter for controlling dynamic range. The wheel 30 can be rotated by motor 31 to present a different aperture size 30c, also having an infra red filter. The radiation passes along a fibre optic system which has an input arm 32 and an output arm 33 in a similar manner to the arrangement of Figure 10.The diamond 34 is irradiated with radiation and the transmitted radiation is collected by the fibre optic system and fed to a photomultiplier tube 36 through a narrow band pass filter 35a that passes radiation of wavelength 415. 5nm. The signal from the photomultiplier tube 36 is amplified at 37 and fed to a microprocessor 38 which can operate with a visual display unit 39. Using the apparatus of Figure 14, instead of tilting the filter, second and third narrow band pass filters 35b and 35c of slightly different band pass characteristics, say 410nm and 418. 5nm, may be interposed between the diamond 34 and the photomultiplier tube 36 by rotating the filter wheel 35 using motor 31.The filter wheel 35 may be rotated at high speed (for example 3,000rpm) to obtain a large number of measurements as in the apparatus of figure 10. Thus in a similar manner to the apparatus of Figure 10, a signal representative of the absorption at the characteristic wavelength 415. 5nm and at a reference wavelengths 410nm and 418. 5nm can be obtained and compared. If the first signal is higher than the second and third signals, a strong absorption at 415. 5nm has been identified and the diamond is accordingly a type IaAB diamond and is classified as natural. Diamonds of other types do not give very different 415. 5nm, 418. Snm and 410nm signals and are classified as possibly natural or possibly synthetic.

The lamp 28, the fibre optic cable 32, 33 and the photomultiplier 36 used in the apparatus Figure 14 may be the same as those used in the specific embodiment shown in Figure 10. A Spindler and Hoyer mirror 29 may be used.

Figure 15 Figure 15 shows a flow chart for classifying a finished diamond using apparatus according to the first and second aspects of the invention. The diamond is first analysed in terms of its colour at 41. Two classes 42, 43 are produced, consisting of the following colour types (with their estimated occurence, as a percentage, derived from intake figures for +0.5cut rough diamonds): : Class 1 Tinted white to yellow (72%) Fancy yellow (less than 0. 1) Brown (approximately 1%) Green and yellow green (less than 0.1%) Pink (less than 0. 1%) Class 2 Colourless (27%) Blue (less than 0. 1%) Diamonds of Class 1 are subjected to the 415. 5nm test at 44 to produce a class of diamonds which are definitely natural (type IaAB) and a class of diamonds which are not definitely natural or 45 and 46. Class 2 diamonds are rejected as not definitely natural.

The present invention has been described above purely by way of example, and modifications can be made within the invention.

Claims (21)

Claims

1. A method of classifying an object, comprising irradiating the object; observing the object through a narrow band pass filter that passes radiation of a first wavelength corresponding substantially to a characteristic wavelength of a particular class of objects; cycling the wavelength passed by the filter a number of times through each wavelength of a set of wavelengths ncluding the characteristc wavelength and at least two reference wavelengths; making a plurality of observations of the object at each of the wavelengths of said set; and using the observations at the wavelengths of said set to classify the object as belonging to the particular class of objects or not.

2. The method of Claim 1, wherein the filter is tilted about an axis normal to its optical axis to alter the band of radiation passed by the filter.

3. The method of Claim 2, wherein the filter is tilted to and fro between a first position and a second position.

4. The method of Claim 3, wherein the first position corresponds to the filter being normal to the optical axis.

5. The method of Claim 1 to 4, wherein the intensity of radiation absorbed by the object at the first wavelength and at the reference wavelengths is observed, and wherein the reference wavelengths lie substantially outside the absorption peak corresponding to the characteristic wavelength in the absorption spectrum of an object of the particular class.

6. The method of Claim 5, wherein the object is classified as belonging to the particular class of objects if the intensity of radiation absorbed at the first wavelength is greater than at the reference wavelengths.

7. The method of any of the Claims 1 to 6, wherein the method is for examining diamonds.

8. The method of Claim 7, wherein the first wavelength is substantially equal to 415 nm, and the particular class of object comprises type IaAB diamonds.

9. The method of Claim 8, wherein one of the reference wavelengths is about 410nm.

10. The method of any of the preceding claims, further comprising correcting the observations at the first wavelength and at one of the reference wavelengths to allow for spectral variations in the radiation irradiating the object.

11. The method of any of claims 1 to 4, wherein the object is viewed through a broad band pass filter which passes radiation of the first wavelength and the reference wavelengths.

12. The method of Claim 11, wherein the broad band pass filter is a laser blocking filter, the object being irradiated using laser radiation.

13. The method of Claim 11 or 12, wherein the characteristic radiation is visible light.

14. The method of Claim 13, wherein the object is observed by eye.

15. The method of any of Claims 11 to 14, wherein the first wavelength corresponds to the Raman emission of a diamond.

16. Apparatus for classifying an object, comprising: means for irradiating the object; a filter which passes radiation at a first wavelength substantially corresponding to a characteristic wavelength of a particular class of objects; means for cyclically altering the band of radiation passed by the filter through each of a set of wavelengths including the first wavelength and at least two reference wavelengths different from the first wavelength; means for giving signals dependent upon the intensity of radiation passing through the filter, and means for classifying the object as belonging to the first class of objects or not on the basis of the signals.

17. The apparatus of Claim 16, wherein the filter is tilted about an axis normal to its optical axis to alter the band of radiation passed by the filter.

18. The apparatus of Claim 17, wherein the the filter passes radiation of the first wavelength when it is normal to the optical axis.

19. The apparatus of any of Claims 15 to 18, configured to carry out the method of any of Claims 2 to 15.

20. A method of examining an object, substantially as herein described with reference to Figures 1 to 6, 9 to 12, 13, 14 and 15 of the accompanying drawings.

21. Apparatus for examining an object, substantially as herein described with reference to Figures 1 to 6, 9 to 12, i3, 14 and 15 of the accompanying drawings.