CA2139537C - Method and apparatus for the classification of matter - Google Patents

Abstract

One aspect of the invention concerns a method of classifying particles according to their composition. In the method the particles (12, 68, 200) are irradiated with electromagnetic radiation, typically X-radiation, at respective first and second energy levels. First and second values are derived which are representative of the attenuation of the radiation by each particle. A third value is then derived as the difference between or ratio of the first and second values, and the particles are classified according to whether the third value is indicative of the presence in the particles of a particular substance. In one application of the method it can be used to classify diamondiferous kimberlite into a fraction consisting of kimberlite particles containing diamond inclusions and a fraction consisting of barren kimberlite particles.

Description

BACKGROUND TO THE INVENTION
THIS invention relates to a method and apparatus for classifying particulate matter.
There are several modes of interaction between certain forms of electromagnetic radiation and matter. Examples are photoelectric absorption, Compton scattering, Rayleigh scattering and pair production in nuclear and electron fields. All of these can contribute to the attenuation of radiation such as X-radiation or gamma-radiation, depending on radiation energy level and frequency. Attenuation can be referred to as the process of removal of photons from an incident beam as a result of one or other of the modes of interaction between the beam and the particle.
The invention has particular application to the detection of diamonds within host kimberlite particles. In practice in diamond recovery operations, it would be highly desirable to detect kimberlite particles that are host to internal diamond inclusions since it would then be possible to reject those kimberlite particles which are barren and to continue with processing of only those particles known to include diamonds. With barren particles rejected at an early stage, the load on, and capacity required of, the downstream processing equipment would be vastly reduced.
The present invention seeks to provide a method and means whereby diamonds in kimberlite particles can be detected using an attenuation-based technique.

SUMMARY OF THE INVENTION
According to a first aspect of the present invention there is provided a method of classifying particles according to composition, the method comprising the steps of irradiating the particles with electromagnetic radiation at respective first and second energy levels, deriving respective first and second values representative of the attenuation of the radiation by each particle, deriving a third value as the difference between or ratio of the first and second values, and classifying each particle according to whether the third value is indicative of the presence in the particle of a particular substance.
In a preferred form of the invention, the particles are irradiated, and transmitted radiation is detected and analysed, in a CAT-scanning or laminography technique. Preferably also, the electromagnetic radiation is in the X-ray, gamma ray or microwave part of the electromagnetic spectrum.
In a preferred application, the method summarised above is used to classify kimberlite particles according to whether or not they have diamond inclusions. In this case, in one embodiment of the invention, the first energy level is selected such that Compton scattering effects are dominant in causing attenuation and the second energy level is selected such that neither Compton scattering nor photoelectric absorption effects are dominant but attenuation attributable to Compton scattering is nevertheless comparable to that at the first energy level, so that when the first image is subtracted from the second image, the Compton scattering effects are largely eliminated.

The first energy level is typically in the range 120keV to 150keV and the second energy level is in the range 60keV to 80keV.
Further according to the invention, there is provided apparatus for classifying particles according to composition, the apparatus comprising means for irradiating the particles with electromagnetic radiation at respective first and second energy levels, means for detecting radiation transmitted by each particle at each energy level, means for deriving respective first and second values representative of the attenuation of the radiation by each particle, means for deriving a third value as the difference between, or ratio of, the first and second values, and means for classifying each particle according to whether the third value is indicative of the presence in the particle of a particular substance.
In the preferred application the method is used in a process in which kimberlite particles containing diamond inclusions are sorted from kimberlite particles which have no diamond inclusions.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will now be described in more detail, by way of example only, with reference to the accompanying drawings in which:
Figure 1 diagrammatically illustrates an apparatus of the invention in side view;

-s-Figure 2 shows a diagrammatic, cross-sectional view of the apparatus of Figure 1;
Figure 3 shows a graphical comparison between attenuation values for diamond and kimberlite;
Figure 4 diagrammatically illustrates another apparatus of the invention;
Figure 5 diagrammatically illustrates the principles underlying a third apparatus of the invention;
Figure 6 diagrammatically illustrates the third apparatus;
and Figure 7 diagrammatically illustrates a further apparatus of the invention.
DESCRIPTION OF PREFERRED EMBODIMENTS
The invention is particularly concerned with electromagnetic radiation in the microwave, X-ray and gamma ray regions of the spectrum. In the case of X-rays, attenuation of incident radiation by dense matter such as rock particles is attributable mainly to two effects, namely the photoelectric effect and the Compton scattering effect. The Compton scattering effect arises from the scattering of free electrons by incident photons, whereas the photoelectric effect arises as a result of liberation of bonded electrons. Which of these two effects will dominate in a particular situation is dependent on the energy level of the incident radiation.
At lower energy levels, both effects are found, in a diamond/kimberlite system, to contribute substantially to the overall attenuation of incident radiation. However at higher energies attenuation other than from the Compton effect is negligible.
Thus by forming separate images related to radiation transmission through a particle at respectively higher and lower energy levels, and then arithmetically subtracting one image from the other, a resultant image can be obtained in which the Compton effect is substantially eliminated and which is attributable almost entirely to the photoelectric effect. In the case of diamond in kimberlite, a good contrast can be achieved between the diamond inclusion and the host kimberlite.
Referring to Figures 1 and 2, particles 12 which are to be classified and sorted are deposited from a hopper 20 onto the upper run of a conveyor belt 16 passing around a driven head roller 18. The conveyor belt 16 transports the particles 12 through a scanning zone 14.
In this embodiment of invention, a CAT scan of each particle is carried out in the scanning zone 14. The CAT scanning apparatus at the zone 14 includes an X-ray source 24 mounted on an annular gantry 22 surrounding the belt 16. The gantry 22 also carries a series of angularly spaced detectors 26. The gantry is rotatable about its axis as the particles 12 pass through, so that scanning on a slice-by-slice basis of each particle takes place, with the particle being irradiated from all circumferential positions.
Typically, the detectors 26 are provided by scintillator screens or image intensifiers which convert the incident X-ray photons into visible light.
The light is viewed by a CCD (charge coupled device) camera (not illustrated) which converts the light signals into electrical pulses. It will be understood that these pulses are related to the degree of attenuation undergone by the X-rays transmitted through each particle.
The electrical pulses are fed to and analysed by a processing unit 28 to form a series of picture elements (pixels) taken at various angular positions in each "slice" of the path travelled by the X-ray source 24. The numerical information from each slice is formed into a tomogram and the attenuation contrast for individual pixel locations is determined.
On the basis of the attenuation values measured in the various pixel locations, contrast values for groups or areas of pixels can be derived.
The processor 28 analyses these values for attenuation patterns indicative of the presence of diamond in the particle undergoing analysis, such patterns being recognisable in terms of pre-programmed criteria.
When the processor 28 recognises a pattern indicative of the presence of a diamond particle, it sends an activating signal to an ejector unit 30.
The particles 12 are projected in free flight from the conveyor belt 16 after passing over the head pulley 18. After a predetermined time delay, the unit 30 issues a short duration fluid blast which deflects relevant particles, i.e. those suspected of containing diamonds, out of the normal _g_ trajectory and into a bin 32. The other particles are allowed to continue in normal free flight into a tailings bin 36 separated from the bin 32 by a splitter plate 34.
The detection parameters need to be optimised according to the nature of the particulate material being classified so as to be most sensitive to a selected substance. Impartant amongst these parameters is the X-ray energy level. Where, as in the present example, the substance being detected is diamond in a kimberlite host, the energy of the radiation must be high enough for the radiation to penetrate the particle while at the same being suitable for any diamond inclusion to provide an adequate discernible attenuation contrast for a reliable classification and sort to be made.
In Figure 3, the line labelled "Kimberlite" represents attenuation values for kimberlite particles for the range of X-ray energies indicated on the horizontal axis. The line labelled "diamond" represents corresponding values for diamond, and the line labelled "diamond/kimberlite"
represents the ratio of the diamond value to the kimberlite value at the various energy values. It will be seen that the diamond/kimberlite ratio is furthest from unity at low energy levels.
Thus the greatest attenuation contrast between diamond and kimberlite can be expected at low energy levels. However, at these levels, the penetrative power of the radiation is decreased and results in a poor signal to noise ratio. Such energy levels are therefore not optimal.
In one version of the invention, a subtraction technique is used to counter the background noise problem. The particle is irradiated with radiation at two distinct energy levels. An experiment carried out on a kimberlite sample confirmed that diamond appeared as a relatively low absorption structure at X-ray energy levels of 70keV or lower.
A first energy level of 70keV was selected as a compromise between penetrative power and attenuation contrast. The second energy value selected was 120keV, at which the diamond/kimberlite ratio is close to unity and the diamond/kimberlite contrast in the derived CAT-scan image can be expected to be poor, but nevertheless at which good penetration is achieved.
At the lower energy level, the CAT-scan image showed an area of pixels having characteristic diamond attenuation values. However, the contrast was poor and was accompanied by a low signal to noise ratio. As explained previously, at this energy level the attenuation attributable to photoelectric absorption is about equal to that attributable to Compton scattering, i.e. both phenomena contributed roughly equally to the tomogram image.
The tomogram derived from the higher energy radiation also showed a low absorption area indicative of the presence of diamond. At this energy level, the attenuation attributable to photoelectric absorption is approximately four times lower than that attributable to Compton scattering. Hence the derived tomogram was due chiefly to Compton scattering effects. However, the attenuation due to Compton scattering is roughly the same at both energy levels so that, when values from one image were subtracted arithmetically from the corresponding values of .. 2139537 - to -the other image, the Compton contribution was substantially eliminated or at least greatly reduced. The resulting, third image was thus attributable mainly to photoelectric absorption effects and exhibited improved attenuation contrast between diamond and kimberlite.
A second version of the invention uses a division or ratio technique. As in the subtractive technique described above, the sample is scanned at two different energy levels, typically the same as those previously mentioned. In this case, a ratio between values derived from the respective scans is derived.
While in the dual-energy subtraction technique described above the aim is to isolate the effect attributable to photoelectric linear attenuation, i.e.
to isolate a linear attenuation coefficient attributable to the photoelectric effect, the aim in the dual energy division or ratio technique is to isolate only the material dependent component of the photoelectric effect coefficient. As stated previously, the linear attenuation coefficient at the relevant X-ray energy levels has a contribution both from photoelectric and Compton scattering effects. Each of these effects has a material dependent component and an energy dependent component. In the photoelectric effect, the material dependent component is the effective atomic number of the material while in the Compton scattering effect, the material dependent component is related to the electron density.
The division or ratio technique effectively isolates the contribution made by the material dependent component of the photoelectric effect from the contribution made by the Compton scattering material dependent component and hence gives a value related to atomic number. Thus in -m-the division technique, the ratio of linear attenuation coefficients or images which is derived is related to the effective atomic number of the particular material under investigation. In the case of diamond in kimberlite, the atomic number of diamond is readily distinguished compared to the other kimberlite inclusions, enabling a sort to be made.
Figure 4 illustrates a modified CAT-scanning apparatus. In this case, a parallel plate collimator 32 is provided for the X-ray source 24, in the form of an ultralong, single anode scanning electron beam X-ray tube.
The parallel plate collimator associated with the X-ray source 24 split the X-ray beam into a series of parallel beams, enabling multiple slice scanning to be carried out simultaneously. The ability to perform multiple slice scanning greatly reduced processing time and increasing throughput.
In Figure 4, it will be noted that the X-ray source 24 and opposing detector 26 are carried on a structure 50 which can rotate about an axis 52.
Figure 6 illustrates a further scanning apparatus within the scope of the invention. In contrast to the. CAT scanning apparatuses described above, this apparatus relies on a laminography technique. The principles underlying this technique are explained with reference to Figure 5, in which a source of radiation is indicated with the reference numeral 60 and an image receptor with the reference numeral 62. The radiation source 60 produces X-radiation at energy levels similar to those described above. The image receptor 62 may be similar in nature to the detectors 26.
The source 60 and receptor are connected to one another by means of a rigid rod 64 (Figure 6). When the rod is rotated about an axis passing through a point M lying in a plane 66 through a particle 68, the source 60 moves along an arcuate path from a location Tl to a location T2.
Simultaneously, the rotation of the rod causes the receptor 62 to move in the opposite direction from the location P1 to the location P2. The X-ray beam 70 produced by the source 60 undergoes a corresponding "pivotal" change in orientation. Those X-rays 72 within the beam 70 that pass through the point M impinge on the receptor 62 at the same point at all times during the change of orientation of the beam. When the source 60 is at position T1, the impingement point is M1 and when it is at position T2, the impingement point is M2. M2 is at exactly the same position on the receptor as M1. In similar fashion, all other X-rays within the beam 70 that pass through other points lying in the plane 66 will also always impinge on exactly the same position on the receptor 62.
However, those rays in the beam 70 which pass through points lying in planes which are parallel to but spaced from the plane 66 impinge on the receptor 62 at positions which vary constantly. By way of example, when the source 60 is at position Tl, the X-rays 76 which pass through a point S lying in a plane 78 strike the receptor 62 at position S1.
However, when the source is moved to position T2, rays 80 which pass through the same point S strike the receptor at position S2. Position S2 is clearly spaced away from the position S1 on the receptor. A similar analysis can be carried for rays passing through a point A lying in a plane 82, with the impingement point tracking across the receptor from a position A1 to a position A2 as the source moves from position Tl to position T2.
Thus in the example seen in Figure S, points lying in the plane 66 are clearly imaged by the receptor while points lying in other planes produce blurred images.
With suitable changes in the positions of the source and receptor relative to the particle 68, scans can be carried out in different, parallel planes.
In practice, to re-set the geometry each time a new scan is to take place would render the apparatus impractical for high volume particle sorting.
It is therefore proposed to present the individual particles on a conveyor belt 90 which moves them past respective source/receptor combinations 92, 94, as seen in Figure 6. Each source/receptor combination is set up to scan different planes, parallel to one another and to the plane in which the particles move on the belt 90.
Radiation at different energy levels, and a subtraction or ratio technique as described above, is used in each scan. If the scan indicates the presence of a diamond inclusion 100 in a kimberlite particle 68, that particle can be removed immediately under computer control, as indicated schematically in Figure 6 by a trapdoor 102 leading to a diamond bin 104. Those particles for which a scan does not indicate the presence of diamond pass on to the next scanning station where a scan is performed in a different plane, and so on. As an alternative to immediate removal of relevant particles, it would be possible to allow each particle to pass through all the scanning stations and only then to separate selected the particles.

In the case of a diamond in kimberlite application a typical particle size may for instance be 50mm. In such a case it is proposed to provide seventeen to twenty scanning stations at different positions along the length of the conveyor belt.The scanning stations are set up to conduct scans in planes spaced apart by approximately 2,5 to 3mm so that substantially all diamonds of this dimension would be detected.
Figure 7 illustrates another apparatus of the invention which relies on laminography technique. In this case, there is a single X-ray source 200 which irradiates diamondiferous particles 202 moved through the irradiation zone on a conveyor 204 which passes over a detector array including a series of spaced apart X-ray detectors 206. The outputs of the detectors are fed to a computer 208 which performs an image rearrangement process to provide a complete scan of each particle. Thus in this embodiment there is no movement of the X-ray source and no need for a plurality of X-ray sources.
In each of the described embodiments, irradiation at two different energy levels can be achieved in various different ways. For instance, it would be possible to use appropriate filters to produce different energies from an X-ray source, or to switch the energy level of the X-ray source.
While the subtraction arid ratio techniques used in the various embodiments described above result in some degradation of the signal to noise ratio, the overall sensitivity for diamond is seen to improve compared to a single energy level CAT-scan. The advantage of either of the dual energy techniques over a single energy technique is that the difference in values, or the ratio thereof, discards or at least decreases the contribution to linear attenuation coefficient made by Compton scattering. Since Compton scattering is heavily dependent on the density of the materials themselves, and bearing in mind that the densities can vary widely, choosing a single energy at which there is a sufficient signal-to-noise ratio to enable accurate differentiation of diamond from other kimberlite inclusions becomes extremely difficult if not impossible.
Furthermore, the linear attenuation coefficient of diamond is equal to that of different inclusions at certain energy levels. At these energy levels, there is therefore an insufficient signal-to-noise ratio for accurate differentiation of diamond from the other kimberlite inclusions.
As a practical matter it is not feasible to select a single energy level at which the signal-to-noise ratio is sufficient to differentiate diamond from other possible inclusions. However, with a dual energy technique it is possible to select appropriate energy levels at which the signal-to-noise ratio is sufficient for differentiation of diamonds from other inclusions.
Added to this, with a single energy level scan, the low signal-to-noise ratio necessitates an unacceptably long analysis time unsuitable for on-line processing of diamondiferous ore.
The embodiments of the apparatus described may be modified for field geology use, for example to enable the analysis of drill core samples.
The drill core may be moved in an axial direction through a scanning zone and the subtraction and ratio techniques described above applied.
The results would enable an exploration geologist swiftly to determine the micro-diamond or other mineral content in the sample.

Claims (12)

1. A method of classifying particles according to whether or not they are host to diamond inclusions, the method comprising the steps, for each particle, of irradiating the particle with electromagnetic radiation at respective first and second energy levels, deriving respective first and second values which are representative of the attenuation of the radiation by the particle at the first and second energy levels, deriving a third value as the difference between or ratio of the first and second values, and classifying the particle according to whether the third value is indicative of the presence in the particle of a diamond inclusion, wherein the first and second energy levels are selected such that the Compton scattering contribution to the third value is at least substantially reduced as compared to the Compton scattering contribution to the first and second values thereby allowing for detectable contrast between diamond and non-diamond material.

2.A method according to claim 1 wherein a CAT-scan of each particle is obtained at each of the first and second energy levels.

3. A method according to claim 2 wherein the particles are irradiated by an irradiation source which is rotated about the particles.

4. A method according to claim 2 wherein the particles are irradiated in multiple, parallel slices by an irradiation source.

5. A method according to claim 1 wherein a laminographic scan of each particles is obtained at each of the first and second energy levels.

6. A method according to claim 5 wherein the particles are sequentially irradiated by a plurality of irradiation sources which are rotated about the particles.

7. A method according to claim 5 wherein the particles are irradiated by a single irradiation source while moving past a series of spaced apart radiation detectors.

8. A method according to claim 1 wherein the electromagnetic radiation is in the X-ray, gamma ray or microwave part of the electromagnetic spectrum.

9. A method according to claim 8 wherein the particles are kimberlite particles which are classified according to whether or not they have diamond inclusions.

10. A method according to claim 9 wherein:
- the particles are irradiated with X-radiation, - the first energy level is selected such that Compton scattering effects are dominant in causing attenuation, - the second energy level is selected such that neither Compton scattering nor photoelectric absorption effects are dominant in causing attenuation but at which attenuation attributable to Compton scattering is comparable to that at the first energy level, and - the first image is subtracted from the second image so that Compton scattering effects are substantially eliminated.

11. A method according to claim 10 wherein the first energy level is in the range 120keV to 150keV and the second energy level is in the range 60keV to 80keV.

12. A method according to any one of claims 9 to 11 wherein the method is used in a process for sorting kimberlite particles which have diamond inclusions from particles which have no diamond inclusions.