GB2036360A - The assessment of colour in diamonds and other gems - Google Patents

Abstract

A method of assessing the colour of a diamond or other gem is provided which comprises aligning the table of the gem normal to the axis of light projected from a source, and making the axis of the light co-linear with an axis of symmetry of the gem which is normal to the table, projecting the light of a given single wavelength at the gem and determining the proportion of incident light transmitted by the gem, repeating this determination for different wavelengths of light over the visible wavelength range so as to obtain a transmission spectrum and calculating the chromaticity co-ordinates therefrom. Using the method it is possible to obtain a wholly objective assessment of the colour of the gem.

Description

SPECIFICATION The assessment of colour in diamonds and other gems This invention relates to the assessment of colour in diamonds and other cut gems.

A perfect diamond crystal would be expected to be completely coiourless and would, therefore, transmit equally well all radiation with wavelengths from 380 nanometres (nm) to 750 nm i.e. the wavelengths constituting "white" light. The majority of diamonds, however, display some colour due to the presence of defects. These defects may be due to an atom missing from its proper site in the crystal i.e. a vacancy, an atom occupying a position in the crystal other than its correct one i.e. an interstitial, or a foreign atom which does not properly belong in the crystal i.e. an impurity. Each type of defect causes the diamond to absorb selectively at one or more regions within the visible wavelength range thus causing an imbalance in the mixture of wavelengths which make up the "white" light incident upon it and producing a sensation of colour.A defect which imparts colour to the host crystal in this manner is called a "colour centre". The type of defect i.e. the origin of the colour in the diamond is of no significance in relation to the present invention.

Traditionally, the colour of a cut and polished diamond, especially one in the range from colourless to yellow, known as a "Cape Series" diamond, has been estimated visually by comparing the stone in question with a preselected set of stones of increasing depth of yellowness. Various systems of standard colour samples have evolved in different parts of the world, for example the GIA (Gemological Institute of America) and AGS (American Gemological Society) systems in the U.S.A. as well as the CIBJO (Confederation Internationale de la Bijouterie, Joaillerie, Orfevrerie des Diamantos, Peries et Pierres) and the German RAL Standard.

The visual method of colour grading is clearly open to criticism since it is a subjective test which is likely to vary from person to person depending on such factors as genetic characteristics, experience in colour assessment, age and sex. Furthermore, the ability of any individual to grade consistently from day to day is influenced by physical factors such as the nature of the ambient illumination, and by physiological factors such as the state of the dark, light and colour adaption of the eye, and state of health. In consequence, therefore, inconsistencies inevitably arise. These factors are well recognised.

In order, therefore, for colour grading to become a more accurate science, there is a need for a truly objective measurement of colour, in the past, various techniques, including electron spin resonance (e.s.r.) have been attempted but satisfactory results have, so far, been unobtainable.

It has now been found, according to the present invention, that a truly objective and reproducible colour measurement can be obtained by making use of direct colorimetry rather than indirectly as with e.s.r. According to the present invention there is provided a method of assessing the colour of a diamond or other gem which comprises aligning the table of the gem normal to the axis of light projected from a source, and making the axis of the light co-linear with an axis of symmetry of the gem which is normal to the table, projecting the light of a given single wavelength at the gem and determining the proportion of incident light transmitted by the gem, repeating this determination for different wavelengths of light over the visible wavelength range so as to obtain a transmission spectrum and calculating the chromaticity co-ordinates therefrom.The aligned gem is generally placed in an integrating sphere prior to projecting the light of a given single wavelength at it.

Accordingly, the assessment of the colour of the gem by the method of this invention is determined by obtaining the chromacitity co-ordinates. According to Grassman's laws, it is possible for a normal observer to match any colour by mixing together the correct amounts of any three real lights, called primary stimuli, of sufficiently different colours. These three quantities are called the tri-stimulus values of the test colour. If the three lights are referred to as X, Y and Z it can be seen that it is possible to measure experimentally, as a function of wavelength, and throughout the visible range, the amount x (A), V (A) and z (A) of the primary lights to match spectral lights emitting equal energies at all wavelengths.

The perception of colour in transparent substances such as diamond involves three distinct factors, namely a source of light (the relative spectral energy distribution of which is denoted S (A)) which illuminates the object, the spectral transmittance thereof which is denoted T (A) and finally the observer. It can be seen that the product S(A)T(A), the relative spectral energy distribution reaching the observer from the object, is the stimulus which is perceived as a colour by the observer.

At a meeting of the CIE (Commission Internationale de L'Eclairage) in 1 931, agreement was reached regarding standard illuminants and the numerical values to be assigned to the functions R (A), v (A) and z (A). The standard illuminants are referred to as A, B and C simulating, respectively, incandescent light, noon sunlight and overcast North sky daylight. Tabulations of their spectral energy distributions can be found in standard textbooks on colour science, for example Hardy A. C., Handbook of Colorimetry (Cambridge, Massachusetts), MIT Technology Press, 1936 and Wyszecki G. and Stiles W. S., Color Science, New York, (Wiley) 1967.

In order to describe the colour or energy distribution in terms of its tri-stimulus values, the product S(A)T(A) must be weighted in turn by each of the tri-stimulus values for the equal energy spectrum x (A), y (A) and z (A) and each product integrated over all wavelengths in the visible region.The tri stimulus values X, Y and Z of a test object in terms of the primary lights are thus given by a set of equations as follows: X = k' 0 S(A)T(A)R(A) dA, 1 (a) Y = k' i'S(A)T(A)y(A) dA, 1 (b) and Z = k' JS(A)T(A)z(A) dA, 1 (c) The normalising constant k' is chosen so that

In this way the tri-stimulus values are expressed on a scale such that Y = 100 for an object which has T(A) = 1 for all wavelengths. In practice, the integrals of equation 1 are approximated by sums of the type: x = k0S(A)T(A)x(A)A A, 2 with k0= 100/#(#)x(#)##, with analogous sums for Y and Z.Tabies listing the products S(#)x(#)##, S(#)#(#)## and S(#)z(#)## for various C.I.E. illuminants and for different values of ## are given in standard textbooks such as those mentioned above, for use with experimentally determined transmission spectra T(A). The most commonly used tables are those for AA = 10 nm, these values being adequate when T(A) is a slowly varying function of A.

From the tri-stimulus values X, Y and Z of a colour, the chromaticity co-ordinates x, y and z may be calculated from: X x = X + Y + Z 3 Y y = X + Y + Z and 3 z = X + Y + Z In fact, the chromaticity co-ordinates x and y specify the colour in terms of the attributes of hue and saturation i.e. depth of colour, respectively. (The z co-ordinate provides no additional information since z = I-x-y). From the chromaticity co-ordinates x and y it is also possible to determine the "dominant wavelength", AD, which is the wavelength at which the line joining the chromaticity coordinates of test colour and illuminant cuts the spectrum locus on the CIE x-y chromaticity diagram, this being related to the hue. Figure 1 of the accompanying drawings illustrates this diagram; S marks the position of the sample co-ordinates.The "excitation purity", pet is obtained from the fraction a a+b where a is the distance on the chromaticity diagram from the chromaticity co-ordinates of the sample to those of the illuminant while b is the distance from those of the sample to the point where the line joining the illuminant with the sample cuts the spectrum locus (at the dominant wavelength).

There remains one attribute or aspect of colour, namely lightness, about which mention must be made. The Y tristimulus value is expressed on a scale from 0 to 100. This value expresses the lightness of the colour, irrespective of hue. This property of the Y value arises because the CIE y(A) function was deliberately chosen to be equal to the relative luminous efficiency function which expresses the relative brightness of equal energy spectral colours as a function of wavelength throughout the visible region.

Clearly, a colour may equally well be specified by listing the two co-ordinates x and y or by listing dominant wavelength, excitation purity and the value of the Y tristimulus function.

It can be seen that since there are tables listing the products S(A)R(A)A,S(A)Y(A)AR and S(A)z(A)A,l for various CIE illuminants and for different values of A.l, by determining the transmission spectrum over the visible wavelength range it is possible to determine the chromaticity co-ordinates and, therefore, obtain an objective measurement of colour.

In fact the CIE method has been used for many years to obtain colour indices of transparent substances such as coloured glasses and clear plastics from measurements of their transmission spectra. It has also been used for measuring opaque materials such as painted surfaces, glazed ceramics and the like.

While attempts have been made to apply the CIE method to the measurement of the colour of a diamond it has not been possible to obtain satisfactory reproducible results. It is believed that this has been due to the fact that whereas the measurement of a plain parallel-sided slab of transparent material is straightforward and whiie the measurement of a reflection spectrum for an opaque surface is somewhat more complicated, because the reflected light travels in all directions from the specimen surface, a brilliant cut gem is anomalous in that whilst it is transparent it is deliberately cut in such a way that most of the light in a beam which falls upon the table of the diamond and then enters the stone leaves again as a number of well-defined beams travelling more or less in the opposite direction to the incident beam, namely via the table and crown facets.Geometrically, therefore, the situation is more akin to reflection than transmission. Clearly, however, with the large number of different surfaces at different angles in a cut gem it is of fundamental importance to devise a system which will enable a diamond to be set up and tested in such a manner that its position relative to the light source is so well defined that it is possible to remove the gem and then on replacing it in the system and adjusting its position until it complies with the definition that highly reproducible results can be obtained.

The aim of the method of this invention is to shine monochromatic light at the specimen and then to measure the amount of light which is transmitted by the gem. As indicated above, light which has been internally reflected in the gem travels in all directions from the specimen surface and, moreover, the intensity may vary with direction. In order to obviate these difficulties, it is known to employ a device referred to as an "integrating sphere". This consists of an enclosure, usually but not necessarily spherical, whose internal surface is ideally a perfectly diffuse reflector at all wavelengths. This is well approximated in practice by coating the inside of the "sphere" with specially prepared barium sulphate or magnesium oxide paint. The sphere has three apertures.The first aliows entrance of the light beam; by means of the second, diametrically opposite the first, it is possible to place the sample to be measured. At the third is placed a photodetector. The sphere entraps all the light coming from the specimens surface and multiple reflections at the wall render the light perfectly diffuse before it is sampled by the detector. The intensity of reflected light is then measured as a function of wavelength relative to a surface made of compressed barium sulphate or magnesium oxide, respectively, these two being universally accepted white standards.

It is then necessary to devise a way of setting up as small an object as a brilliant cut diamond in the integrating sphere in such a way that a truly reproducible result can be obtained.

It has been found, according to the present invention, that reproducible results can be obtained if the table of the gem is accurately aligned normal to the axis of the light incident upon it such that there is an axis of symmetry normal to the table of the gem, this axis being made co-linear with the axis of the light. One of skill in the art will appreciate that once it has been realised what the requirements are it is a relatively simple matter to devise an arrangement such that these requirements are met. Naturally, a holder for the gem is required, the holder being provided with means enabling one to adjust the gem with respect to the light source, Also the gem must be mounted in such a way that the mounting means do not affect the results in any way or, at least, if they do, they do so in a manner which can be quantitatively compensated for.

Clearly, it is necessary for the light source to be monochromatic but for the wavelength of the light to be adjustable. Commercially available spectrophotometers satisfy these requirements. It will be appreciated that monochromatic laser beams could also be used i.e. using a dye laser tunable over the entire visible wavelength range. In fact, experimental work has been carried out on a Beckman Model 25 double-beam instrument equipped with a 6 inch diameter integrating sphere.

It will be appreciated that although in a "Cape Series" diamond yellowness is the observed sensation this may well not be due purely to absorption at 415 nm caused by the so-called N3 centre.

Thus it is not infrequently the case that an absorption in another part of the spectrum will contribute to the final sensation of colour. Thus it is believed that the 1 b centre, and the H3 centre, which has a zero-phonon line at 503 nm and a phonon-assisted band at lower wavelengths, may also contribute to the final colour. Accordingly, examination of only a part of the spectrum, for example the 41 5 nm absorption region, cannot give rise to a reliable set of results.

The beam of monochromatic light which passes from the optics of the spectrophotometer into the integrating sphere is diffuse and slightly divergent, suitable, as intended by the manufacturers, for the measurement of spectra of relatively large (about 1 inch square) flat specimens. Naturally, for gem stones, it is necessary to concentrate and direct the light beam onto the stone using a converging lens.

It is envisaged that the use of a lens would be unnecessary in the case of a laser beam which could be produced sufficiently small for concentration to be unnecessary.

The present invention will now be described, merely by way of Example, with reference to the accompanying drawings, in which: Figure 2 illustrates, diagrammatically, a typical diamond holder arrangement equipped with appropriate adjustment means: Figure 3 illustrates, diagrammatically, a typical optical bench suitable for use in aligning the gem in the holder.

The holder shown in Figure 2 consists essentially of three parts A, B and C. A plate C carries a goniometer G. In turn, G carries, at the end of a small rod, X, a cup, S, for example of aluminium, filled with compressed barium sulphate powder into which is pressed the pavilion of a diamond, Z. It will be realised that by mounting the diamond in this way, no extraneous colour is introduced by the "mounting".

By means of the goniometer G, it is possible to translate the diamond (in the cup) by, say, a centimetre along each of two directions at right angles in the plane of the plate C, and also to rotate the stone by up to, say, 100 about each of two orthogonal axes para!lel to the translation directions. Plate C is detachable from a plate B, to which it is attached by means of two screws. Two locating pins on B, with matching holes in C, ensure the accurate location of C on B.

The plate B is spring-loaded and may be moved smoothly towards and away from plate A (in a direction perpendicular to A) on guide rods R by means of micrometer screw M. This endows the diamond with a third degree of translational freedom, perpendicular to the two provided by the goniometer.

The plate A carries the lens in a lens-holder L. In A there is a hole H machined so that the lens may be made approximately concentric with H by swivelling the rod S carrying the lens-holder. Plate A is capable of being screwed onto a similar sized plate A' (not shown) which is permanently fixed to the outside of the integrating sphere. A locates accurately on A' by means of two locating pins. Plate A' has a hole H' which is accurately concentric with hole H in A, and approximately concentric with the sample (or second) aperture machined in the integrating sphere and with the aperture which allows entrance of the light beam (i.e. the first aperture which is diametrically opposite the sample aperture). Thus when A is affixed to A' the lens and diamond protrude into the sphere.It will be appreciated that it is necessary for the diamond to protrude into the sphere to ensure that all light transmitted is detected. Furthermore, the plane of the matching surfaces of A and A' is machined to be approximately tangential to the sphere, so that when A is screwed to A', the lens and the light beam are approximately co-axial and can be made so by means of an alignment procedure described below. The diamond can be made co-axial with the last two by means of small movements of the goniometer arcs.

The lens is an achromatic doublet with a focal length of 2.6 mm. The achromat is employed in order to cut down dispersion effects to a minimum.

It will be appreciated that if a laser source were employed it would be unnecessary to provide the diamond with the third degree of translational freedom since a lens could be dispensed with. If, however, a lens is used it is necessary to ensure that the table of the gem is at the focus.

The optical bench, shown in Figure 3, carries a lamp L (behind a condenser lens in a lamp housing); a large whitened screen S with a small circular aperture near its centre; and a post with an adaptor plate A", identical to plate A' on the integrating sphere, to which the plate A of the specimen holder may be attached by means of the same two screws used to attach it to A'.-A green gelatin filter, F, is interposed between L and S in order to reduce dispersion effects when the light beam is used to align the diamond. These components are slidably mounted on a rod, with a horizontal -- and - vertical traverse carriage. The distance between L and S is conveniently about 70 cm, and the distance from S to the specimen holder lens is initially arranged to be about 1 5 cm.

Alignment of the Diamond in Preparation for Colour Measurement The critical alignment of the diamond in the holder with respect to the lens, prior to making a colour measurement, can be accomplished outside the sphere on this optical bench; subsequently the holder is transferred to the sphere (by screwing plate A onto plate A') for measurement of the spectrum of the stone.

A typical alignment procedure for the diamond will now be described on the assumption that the condenser lens in L, the aperture in S and the specimen holder lens are already strictly co-axial. (The procedure for achieving this condition is outlined later).

Firstly, plate C is detached from B. The diamond is then mounted centrally in the cup of barium sulphate powder by pressing a flat object, for example a flat glass plate, against the table facet until the pavilion and girdle are buried beneath the level of the surrounding barium sulphate surface, and the table is approximately parallel to that surface. Barium sulphate is used for holding the stone because it is an almost perfect white and makes no contribution to the colour of the stone. Adhesion between the barium sulphate and the pavilion facets is sufficient to retain the stone firmly in position such that the cup can be manipulated without altering the position of the diamond therein. Plate C is then attached to B, and the holder mounted on the optical bench by screwing plate A to adaptor plate A".The table facet is then adjusted by eye, using the rotational movements of the goniometer, to be appróximately perpendicular to the axis of the lens. On darkening the room and switching on lamp L, a (divergent) beam of light defined by the aperture in S strikes the lens and is made to converge to form an image some 2.5 cm behind the lens. By means of the translational movements of the goniometer, the table of the stone is now brought to intercept the light beam. At this stage an image of the light beam specularly reflected from the table facet should appear on the screen in the vicinity of the aperture. If not, it is necessary to alter the distance between S and the specimen-holder lens. This image is now made concentric with the aperture by using the rotational movements of the goniometer.By this means it is possible to ensure that the table is accurately normal to the axis of the iens.

Provided that the light beam is falling near the centre of the table, a pattern of spots comprising eight major and many more minor ones (in the case of a brilliant cut diamond), showing approximate eight-fold symmetry and centred on the aperture in S, will now be seen upon the screen. This pattern is due to divergent light beams, arising from reflection of the incident beam on the pavilion facets, escaping through the table facet. The degree of divergence varies somewhat from stone to stone, and it may be necessary to shorten the distance between specimen holder lens and screen S at this stage in order to intercept the spot pattern completely. The eight major spots of the pattern are now adjusted so as to be of approximately equal intensity, as judged by the eye, by using the translational movements of the goniometer again.By this means one can ensure both that the table of the stone is normal to the lens axis and that the eight-fold axis of the stone (in the case of a brilliant cut diamond of course it might be a different axis with a different cut giving rise to a different spot pattern) and the lens axis are collinear. Clearly, this orientation is a unique one, capable of being obtained repeatedly after successive remountings of the stone and/or after rotation of the stone about its axis.

The final adjustment which needs to be made is to move the stone alongs its axis, using the micrometer screw M, to that point where its table will be at the position of focus of the image formed by the specimen-holder lens, of the spectrophotometer slit inside the sphere. This can be accomplished by referring the position of the table to a datum plane (the back of a small screen D carried on a short rod which can be inserted into a hole in A diametrically opposite the holder carrying the rod R); whose position is known relative to the position of focus of the spectrophotometer slit (determined visually).

With this last adjustment, the stone is so arranged that when the stone is inside the sphere, that part of the light beam which does not penetrate the stone, i.e. the light specularly reflected from the table surface, travels back along its original path and leaves the sphere via the entrance aperture. In this way the specularly reflected beam, which carries no colour formation, is consistently eliminated from the sphere irrespective of stone size (with the proviso, of course, that the dimensions of the image of the slit are less than the table dimensions), and hence does not, by admixture, affect the colour of the diffuse light (which originates inside the body of the stone and carries the desired colour information) sampled by photocell.

Thus the holder is now ready for transfer to plate A' on the sphere. To within a good approximation the light beam will always follow the same path through any particular stone. It is essential to achieve this condition, because the exact shape of the transmission spectrum, on which the calculation of colour indices (chromaticity co-ordinates) is based, depends not only on the concentration of colour centres in a stone, but also on the distance the light travels within the diamond.

With practice, mounting and alignment of the stone may be achieved in a few (three or four) minutes.

As indicated this is on the assumption that the condenser lens in L, the aperture in S and the specimen holder lens are strictly co-axial. A typical procedure for achieving this condition will now be described.

Alignment of the Components of the Optical Bench (i) The first step is to adjust the lamp housing so as to make the axis of the light beam parallel to the optical bench. To achieve this, a screen, on which are drawn a number of concentric circles, is affixed to a horizontal-and-vertical traverse carriage and placed on the bench in front of the lamp. When the lamp is switched on and the iris diaphragm incorporated in the lamp housing immediately in front of the condenser lens is fully stopped down, a circle of light, which is an image of the rear surface of the lens, appears on the screen. The position of the screen is adjusted until the circle of light falls on one of the concentric circles on the screen. The screen is now traversed along the bench and the relative positions of the circle of light and the concentric circles are observed.Should the circle of light not remain concentric with the set of circles, the lamp is tilted in the appropriate direction. The position of the screen is again adjusted for concentricity of the circle of light and the set of circles on the screen.

The screen is again traversed, and the lamp again tilted if necessary; this procedure is repeated until the circle of light remains concentric with the circles on the screen when the screen is moved towards and away from the lamp. When this condition is finally achieved, the axis of the light beam is parallel to the optical bench.

(ii) Second, the screen S must be aligned so that the axis of the light beam passes through the aperture in S. This is achieved by bringing a pointer up to the centre of the concentric circles on the screen used in step (i), removing this screen, placing screen S on the optical bench and adjusting the position of S so that the aperture is in line with the tip of the pointer.

(iii) Next, one must ensure that the plate A" is perpendicular to the light beam. A plane mirror is held flush against the plate so that the beam passing through the aperture in S is reflected back so as to fall on S. Plate A" is then tilted until the reflected light spot is concentric with the aperture in S.

(iv) Plate A, carrying the specimen-holder lens, is attached to A", and the screen S is placed approximately 10 cm from the lens. Two circles of light will now be seen on the screen corresponding to the reflections of the light beam from the front and back surfaces of the lens. The horizontal and vertical positions of A" are adjusted so that the two reflected images are concentric with the aperture in screen S.

(v) In order to check that the lens on the specimen-holder is co-axial with the light beam, a plane mirror is held behind the lens flush against plate A. The reflection of the light beam in the mirror should be concentric with the aperture in screen S (and thus, by inference), also with the two circles of light seen as described in step (iv)). If this is not the case, the rod carrying the lens should be very slightly and carefully bent in the appropriate direction.

(vi) Repeat steps (iv) and (v) until all three circles of light are concentric.

Alignment of the Spectrophotometer Light Beam It is necessary that the alignment of the components of the optical bench has been accomplished before proceeding. Once this has been done, one can proceed as follows: (i) The first step is to align the optical bench so that it is accurately perpendicular to plate A' in the integrating sphere.

Remove all components except lamp L from the bench, and arrange the bench to be approximately perpendicular to A', with the lamp at the end of the bench furthest from A', and facing the sphere. Using a pointer, as described in step (ii), arrange a screen with an aperture, for example an iris diaphragm, between L and A' so that the aperture is exactly on the axis of the light beam from the lamp. Hold a plane mirror flush against A', and by lateral and/or vertical movement of the bench as a whole, arrange matters so that the beam of light reflected from the mirror falls on the screen (on the side nearest the mirror) so as to be concentric with the aperture. The bench is then strictly perpendicular to A', and must be kept in this position until the rest of the alignment procedure has been completed.

Remove the screen from the bench.

(ii) Next, it is necessary to arrange a screen, whitened on both sides and having a small aperture (e.g. 5 mm diameter), on the optical bench so that the aperture will fall on the axis of the lens of the specimen holder when the holder is attached to plate A'.

Remove plate C from the specimen holder, and attach the holder by means of plate A to the plate A" on the post, and place post on the optical bench so that the lens faces towards the lamp L. Switch on the lamp and open completely the iris diaphragm in the lamp housing. A convergent beam of light of circular cross-section can now be seen behind the specimen holder lens. Bring the tip of a pointer, carried on a horizontal-and-vertical traverse carriage and parallel to the length of the bench, to the position of the centre of this circle. Note the horizontal and vertical co-ordinates, x and y, of this point on the scales of the traverse carriage. Then, again using the pointer, find the horizontal and vertical coordinates x' and y' (which in general will be close to, but slightly different from x and y) of the centre of the hole H in A.Now remove the post carrying A" from the bench.

Again using the pointer, find the co-ordinates of the centre of the hole H' in A'. Suppose these to be x" and y". Now place the whitened screen on the bench so that the aperture falls at the tip of the pointer set so that the scales on the horizontal and vertical traverses read [x" + (x' - x)] and [y" + (y' y)j. Since holes H" and H are machined to be accurately concentric, with the screen in this position, the aperture falls on the axis of the specimen holder lens when the holder is attached (by plate A) to plate A' on the sphere.

(iii) Next, the light beam in the spectrophotometer must be aligned so that it and the lens in the specimen holder will be co-axial when the specimen holder is in position on the sphere.

Remove the lamp and pointer from the bench, replacing lamp L by the post carrying plate A", and attach a plane mirror flush against the face of A". Switch on the spectrophotometer lamp and adjust the wavelength of the light to about 550 nm (in the green region of the spectrum). Darken the room completely. The light beam will now be passing into the sphere through the entrance aperture and will fall on the whitened screen, on the side nearest the sphere, to form an image of the lamp filament. By adjusting the mirrors which guide the light beams into the sphere, arrange matters so that this image falls centrally on the aperture. Some light now passes through the aperture, is reflected by the mirror on A", and is directed back to fall on the whitened screen on the side remote from the spectrophotmeter.

Alignment now consists of adjusting the mirrors until the images on the two sides of the whitened screen are simultaneously symmetrical with respect to the aperture on the screen. When this condition has been achieved, the desired alignment has been attained. The optical bench may now be removed.

(iv) We have found empirically that a circular aperture of diameter 0.5 cm placed at the entrance port of the sphere improves reproducibility of the measurements. The last step in the alignment is to position this aperture so that it lies on the axis of the specimen holder lens and light beam inside the sphere.

Replace the components L, S and A" on the optical bench and check their alignment. Attach plate C to plate B on the specimen holder, and attach the holder by means of plate A to plate A". Replace the cup which holds the barium sulphate and diamond by a small (0.5 cm square) front-surface-silvered mirror and align the mirror normal to the axis of the lens (in the same way as a diamond is aligned).

Transfer the specimen holder to plate A' on the sphere.

The plate containing the aperture is painted with white barium sulphate paint on the side which abuts onto the sphere. When this is placed at the entrance port of the sphere so that part of the light beam enters the sphere, a beam reflected by the front-siivered mirror becomes visible as a circular spot near the entrance aperture. Position the mirror along the axis of the lens (using micrometer screen M) until the spot diameter becomes greater than the diameter of the aperture (viewing through the exit port of the reference beam). The plate must now be positioned so that aperture and circular spot become concentric.

With the apparatus set up and aligned correctly it is possible to measure the colour of the gem.

As indicated above the light source is conveniently from a spectrophotometer, the principal component of which is a monochromator, usually a diffraction grating, which provides a beam of light of a single wavelength. This wavelength may be continuously varied over a predetermined range; for the purposes of the present invention, the wavelength range of interest is in the visible, from about 740 nm (immediately above which lies the infra-red region of the electromagnetic spectrum) down to about 380 nm (below which lies the ultraviolet). The monochromatic beam is made to fall upon the gem in its holder inside the integrating sphere. At each wavelength the gem absorbs a certain proportion (depending on the colour of the stone) of the light falling upon it and transmits the remainder.A photocell measures the percentage transmitted as a function of wavelength thus providing a transmission spectrum. The spectrum of a typical, fairly yellow Cape Series stone is shown in Figure 4 giving the wavelength in nanometres along the abscissa and % transmission along the ordinate. In the Beckman spectrophotometer, the light beam coming from the monochromator is limited by a slit system, after which it is guided into the sphere by three independently adjustable mirrors.

The output of the spectrophotometer can be represented pictorially by a strip-chart recorder although this is not essential; however, it is useful in that to an experienced eye it provides, at a glance, a means of checking whether or not there is anything unusual about the stone under investigation. The output must however be recorded numerically. In view of the large number of computations implicit in equations 2, leading to the tristimulus values X, Y and Z it is convenient for this to be done with the aid of a computer. Conveniently the computer is programmed to calculate chromaticity co-ordinates directly from a tape cassette on which it is filed, into the computer memory.This programme is written in such a way that the first action of the computer on pressing the "RUN" key is to load into its memory, the standard functions S(A)x(A)AA, S(A)y(A)AA and S(A)z(A)hA for C.l.E. illuminant "C" for example, and for wavelength intervals AA of 1 nm and the reflectance measurements W(A) of a standard white specimen made of optically-pure magnesium oxide (measured and stored prior to any measurements on stones). The computer also has stored in its memory the functions for the determination of the chromaticity co-ordinates. A Hewlett-Packard 9825 A desk-top computer was found to be particularly suitable.C.l.E. Illuminant C was chosen since it approximates the light traditionally used for visual grading but, of course, the other illuminants can be used since the appropriate values of the standard functions are known for these as well.

To measure the colour of the stone it is convenient to set the wavelength scanner of the spectrophotometer to one end of the visible wavelength range, for example 740 nm. The wavelength scan is then started (by pressing the "START" key on the spectrophotometer). This actuates the automatic scanning of the wavelength from 740 nm down to 380 nm. At 1 nm intervals beginning at 740 nm and ending at 380 nm, the computer automatically reads a percentage transmission T'(A). Each T'(A) is corrected by multiplication by a factor 1 OO/W(A), to give T(A).The T(A) are then percentage transmittances relative to the reflectance of the standard white specimen taken as 100%. At each A, the T(A) are multiplied by each of S(A)x(A)AA, S(A)y(A)AA and S(Az(A)AA in turn, and at the end of the scan, the summations given by equations 2 are evaluated instantaneousiy, giving X, Y and Z. The chromaticity co-ordinates x, y and z are also computed from equations 3. The computer then prints out X, Y, Z, x, y and z. With the particular instruments used the scan from 740 down to 380 nm takes about 1821 minutes and, once the bench has been aligned, the entire procedure, from mounting the specimen to obtaining a colour index, can be achieved in half an hour or less.Indeed, by sacrificing some degree of accuracy (which is generally acceptable) it is possible to speed up the instruments so as to obtain a complete scan in about 4 minutes. Naturally for strictly comparable results the same speed should be used for all gems.

In order to attain good reproducibility certain precautions should be taken. For example it is desirable that the spectrophotometer is kept on permanently, in order to maintain electronic stability.

Second, it is desirable to work at constant temperature, for example by using a thermostatically controlled room, because thermal expansions and contractions affect the metal components upon which the beam alignment mirrors are mounted. Also, when using a double-beam instrument, such as a Beckman spectrophotometer the light beam should be mechanically "chopped" in order to make phasesensitive detection possible. The light beam therefore spends half a cycle falling on the sample, and then half a cycle falling on a reference white specimen, for example a pressed magnesium oxide surface.

However the optically pure magnesium oxide deteriorates sufficiently rapidly on exposure to the atmosphere to make necessary the preparation of a fresh magnesium oxide surface at the beginning of each day desirable, although this is quick and simple.

It should be pointed out that if a suitable laser source were available the alignment procedure and, indeed, the taking of measurements would be very much simpler. Thus one would not need a lens to concentrate the light beam onto the gem; further it should be possible to dispense with separate optical bench and spectrophotometer arrangements. The alignment of the diamond and the measurement of its colour indices could then be carried out on one piece of apparatus, with the laser at one end and the gem in the integrating sphere at the other.

Various Examples have been carried out to demonstrate the viability and accuracy of the method of this invention.

EXAMPLE 1 A set of seven visually graded stones denoted by 1 A to 7A, (in order of increasing yellowness) and another set of five visually graded stones denoted by 1 B to 5B were provided. The measured chromaticity co-ordinates in terms of C.l.E. illuminant C are presented in tabular form in Tables 1 and 2, and graphically, for an overall view, in Figure 5, which represents a portion of the C.I.E. xy chromaticity diagram. Four measurements were made on each stone, with the diamond being rotated about its axis by 900 between each measurement. The tables also present the mean values of the x and y chromaticity co-ordinates; with the dominant wavelength for each stone calculated from the means.

The following points emerge: (i) It is difficult to describe the colours of the stones, but in order to give some idea of the subtle differences involved, it can be pointed out that the untutored eye cannot distinguish between 1 B and 3B (let alone between 1 B and 2B) and can only with difficulty distinguish a colour difference between 1 B and 4B. Furthermore, a comparison of the scales of Figures 1 and 5 will make it evident that the colours being measured occupy only a very small region of the C.l.E. xy colour space near the centre of the diagram.

(ii) The four readings from each individual stone group together with a scatter which depends on the individual. It is believed that the amount of scatter depends on the fine details of cut resulting in slight path-length differences for the light within the stone as the stone is rotated to different positions about its axis. The basis for this belief is that, for any individual, the same trend is apparent in the values of the chromaticity co-ordinates on repeated rotations of the stone through a full 3600. However the differences in cut are generally so small, from stone to stone, that it is not possible even by close examination prior to measurement to predict the extent of scatter for any individual.

(iii) It is clear that the B specimens form an excellent standard colour sample; the individual stones are well separated from one another as regards their xy co-ordinates, the amount of scatter for any individual being much less than the distances separating the members of the set. Indeed, it is clear from the graph that, if one wished, one could easily distinguish instrumentally stones falling between the B grades. In fact, one could easily define one, possible even two, distinct grades between the present ones.

(iv) The A set of stones are very much less well graded. Members graded 1 A and 2A are barely distinguishable instrumentally, and the same holds true for stones 4A and 5A, and 6A and 7A. The indistinguishability is reflected in attempts at visual colour estimation, shown in Table 3. This shows the sets of stones listed in terms of increasing yellowness as determined instrumentally, and the contradictions between visual orderings assigned by three individual, experienced colour graders.

(v) The straight line in Figure 5 is the best line, found by a least-squares linear regression, which can be drawn through all the points representing stones of pure Cape Series colour. This line, if extended, would cut the spectrum locus at 571.3 nm. A glance at the tables shows that the dominant wavelengths calculated for each individual stone agree well with this value of 571.3 nm for the dominant wavelength. (The dominant wavelengths calculated for stones lying close to the "white" point, for example those of stones up to and including 2B, should be treated with caution since in this region small inaccuracies in measurement of chromaticity co-ordinates give widely different dominant wavelengths).

(vi) When measurements of chromaticity co-ordinates are made after demounting and remounting stones, and also when measurements are made on the same stone at different times, no noticeable differences emerged.

EXAMPLE 2 Three stones showing off-colour tints were examined. Two of these have different degrees of "Cape Series yellowness" in combination with a very slight (imperceptible to the untutored eye) brownness. The results are presented in Table 4, and are also plotted in Figure 5. It is apparent that a feature of such stones is that they lie "off the line" representing pure Cape Series stones, and in fact show a slight but definite shift in dominant wavelength (which for both stones is 573.5 nm) towards the orange region of the spectrum.

The third stone, marked W2/MAUVE in Table 4, and also plotted in Figure 5, is of special interest.

This stone was submitted as being of very good colour, roughly comparable with 1 B or 1 A. On running a spectrum and computing chromaticity co-ordinates for this stone, it became apparent that the diamond was not a true member of the Cape Series in that the chromaticity co-ordinates indicated a mauve or lilac colour (with reference to the colour chromaticity diagram), albeit very faint. This stone was then resubmitted to visual graders for regrading; they, after careful reappraisal, and without prompting, affirmed that the stone "draw a faint lilac tint".

This Example illustrates that the method of this invention can be used to assess not only commercially important Cape Series yellowness, but also (both qualitatively and quantitatively) faint offcolour hues.

EXAMPLE 3 In order to assess whether the method was independent of stone size, two sets of five diamond brilliants of two different (Cape Series) colours, with as large a variation of size as was possible within each colour grade were supplied. The results of measurements on these sets are given in Table 5 and in Figure 6, which again is a portion of the C.l.E. xy chromaticity diagram. The size variation for the group of better (i.e. whiter) colour is from 1.1 8 ct down to 0.20 ct, whilst that of the yellower category is from 1.23 ct down to 0.25 ct. It is apparent from Figure 6 that stones of the same colour (as judged visually) but of different size cluster together in the xy colour space, showing that, at least for stones in the range 0.20 ct to 1.25 ct, the machine measures colour independently of size.There are however two points to note: (i) In the group showing better colour the clustering is less tight than in the yellower group. it seems likely that the explanation for this is that the visual matching of colours becomes increasingly difficult with decreasing saturation of hue.

(ii) The smallest stone in the category of better colour (0.20 ct) shows an exceptionally large scatter on rotation of the stone relative to the scatter for other members of the group. It seems likely that the method of the present invention becomes less reliable as the stone size is reduced further.

EXAMPLE 4 All the diamonds examined luminesce to a greater or lesser degree under the stimulation of radiation from an ultraviolet lamp. Most exhibit only the blue N3 luminescence (see Collins, "Industrial Diamond Review", April 1974, page 131), but several also showed an admixture of yellowish or pinkish luminescence. Since there is an ultraviolet component in daylight which will cause some degree of luminescence, and since the method does not take into account the effect on the eye of possible luminescence, the method may lead to discrepancies between machine and visual colour grading.

However it is believed that the effect of the ultraviolet component in daylight in causing luminescence is in nearly all cases negligible. This is because most of the stones examined for luminescence have been graded by assigning a number which serves as a measure of the relative strength of luminescence under an ultraviolet lamp. The results are shown in Table 6. No systematic discrepancies between machine and visual gradings which can be explained in terns of parallel variations in strength of luminescence have been observed. Accordingly it is believed that the machine will parallel visual gradings in the majority of cases. However, it is necessary to bear in mind that luminescence, if exceptionally strong, could lead to anomalous results.Such might well be the case with those stones which fall into the category "Overblue", i.e. display an excessive luminescence under ultraviolet stimulation resulting in a fluorescent appearance even in daylight. In any event, it is a simple matter to test for excessive luminescence prior to attempting a colour measurement, and thus to be prepared for possible anomalies.

It will be appreciated that although the discussion and Examples have been directed especially to brilliant cut diamonds, the principles of the present invention could also be applied to diamonds of other cuts such as the oval brilliant and the marquise as well as other (substantially transparent) gems such as emeralds provided the stone possesses an axis of symmetry normal to the table of the stone. Similar comments apply to colour; attention has been directed principally at yellowness although some tests on off-colour tints such as a brownish tint and a lilac tint have also been carried out; there is no reason why the method should not be applicable to coloured stones and stones with other coloured tints.

It is envisaged that by means of the method of this invention it will be possible to devise new objective sets of grades, each grade corresponding to a particular range of values for the x-y coordinates. As Example 1 reveals such sets might include more different grades than an existing set, thus making grading more specific (apart from being accurate and objective) than at present.

Alternatively the method can be used to obtain a selected set of stones, graded with reference to their chromaticity or-ordinates; this set could then be used as a "standard" against which other stones could be assessed, for example with the naked eye.

TABLE 1 DATA FOR THE "A" COLOUR SAMPLES

SPECIMEN MEASURED x MEASURED y MEAN x MEAN y MASS (ct) #D(nm) .30996 .31625 .31005 .31635 0.95 .31007 .31639 1A .31010 .31643 .31010 .31648 .31005 .31631 .31039 .31659 .31036 .31661 0.96 572.2 2A .31030 .31658 .31031 .31661 .31044 .31669 .31171 31879 .31176 .31885 0.82 571.6 3A .31194 .31901 .31182 .31895 .31158 .31866 .31301 .32087 .31294 .32080 0.81 571.3 4A .31284 .32069 .31291 .32082 .31301 .32081 .31331 .32123 .31316 .32107 0.87 571.5 5A .31313 .32104 .31317 .32096 .31303 .32105 .31450 .32343 .31453 .32334 0.95 571.3 6A .31456 .32336 .31458 .32330 .31447 .32325 .31482 .32347 .31467 .32335 0.80 571.7 7A .31467 .32336 .31441 .32324 .31477 .32348 TABLE 2 DATA FOR THE "B" COLOUR SAMPLES

SPECIMEN MEASURED MEASURED MEAN MEAN MASS(ct) (nm) x Y x y .31024 .31652 .31025 .3t653 0.28 (568.9) 1B .31019 .31656 .31030 .31653 .31028 .31652 .31097 .31798 .31106 .31807 0.28 569.4 2B .31097 .31800 .31126 .31822 .31103 .31809 .31268 .32049 .31268 .32045 0.29 571.t .31269 .32047 3B .31264 .32036 .31270 .32046 .31465 .32353 .31463 .32356 0.23 571.2 4B .31461 .32362 .31458 .32342 .31468 .32363 .31623 .32623 .31639 .32650 0.25 571.1 5B .31630 .32654 .31662 .32671 .31640 .32652 TABLE 3 ORDERING OF THE "A" and "B" GRADED COLOUR SAMPLES BY MACHINE AND BY THREE DIFFERENT VISUAL GRADERS.

SPECIMEN MACHINE GRADER 1 GRADER 2 GRADER 3 GRADER 3 (1st ATTEMPT) (2nd ATTEMPT) 1A 1 1 1 1 2 1B 2 3 3 3 3 2A 3 2 2 2 1 2B 4 4 4 4 5 3A 5 5 5 4 4 3B 6 6 7 6 6 4A 7 7 6 7 7 5A 8 8 8 8 8 6A 9 9 10 9 10 7A 9 11 11 11 11 4B 9 10 9 9 9 5B 12 12 12 12 12 TABLE 4 DATA FOR OFF-TINTED DIAMONDS

SPECIMEN MEASURED MEASURED MEAN MEAN MASS (ct) #D(nm) x y x y DE BEERS .317163 .32681 .31770 .32656 0.76 573.5 Y1/BROWN .31761 .32638 .31744 .32607 .31792 .32696 DE BEERS .31410 .32157 .31409 .32163 0.81 573.5 T4/BROWN .31388 .32145 .31412 .32168 .31427 .32181 DE BEERS .30933 .31426 .30952 .31456 0.26 - W2/MAUVE .30955 .31456 .30961 .31463 .30960 .31479 TABLE 5 DATA FOR THE TWO SETS MATCHED FOR COLOUR.

SPECIMEN MEASURED MEASURED MEAN MEAN MASS x y Y x y (ct) SET 1 .31222 .31906 .31184 .31873 0.20 (0.20 ct) .31194 .31831 .31159 .31847 .31159 .31858 SET 1 .31160 .31834 .31156 .31832 0.31 (0.31 ct) .31158 .31842 .31144 .31830 .31161 .31823 SET 1 .31108 .31772 .31106 .31769 0.63 (0.63 ct) .31103 .31766 .31110 .31770 .31104 .31768 SET 1 .31157 .31837 .31140 .31823 0.85 (0.85 ct) .31130 .31809 .31129 .31809 .31142 .31835 SET 1 .31144 .31835 .31137 .31829 1.18 (1.18 ct) .31140 .31832 .31131 .31822 .31132 .31828 SET 2 .31301 .32126 .31309 .32134 0.25 (0.25 ct) .31320 .32142 .31305 .32134 .31308 .32135 SET 2 .31286 .32126 .31301 .32145 0.32 (0.32 ct) .31304 .32151 .3132t' .32177 .31287 .32124 SET 2 .31284 .32162 .31286 .32166 0.50 (0.50 dt) .31305 .32189 .31288 .32173 .31267 .32140 SET 2 .31349 .32172 .31347 .32168 0.75 (0.75 ct) .31343 .32164 .31353 .32172 .31343- .32165 SET 2 .31329 .32176 .31331 .32173 1.25 (1.25 ct) .31331 .32172 .31331 .32173 .31332 .32171 TABLE 6 RELATIVE LUMINESCENCES OF DIAMONDS.

RELATIVE SPECIMEN STRENGTH OF COLOUR OF LUMINESCENCE LUMINESCENCE 1B 1 Blue 2B 4 Blue 3B 8 Blue 4B 6 Blue 5B 10 Blue 1A 2 Blue 2A 4 Blue 3A 2 Yellow 4A 4 Blue 5A 2 Yellow 6A 2 Blue plus yellow tint 7A 2 Blue plus yellow tint SET 1 (0.20 ct) 1 1 Blue plus pink tint SET 1(0.31 ct) 6 Blue SET 1(0.63 ct) 2 Blue SET 1 (0.85 ct) 2 Blue SET 1 (1.18 ct) 1 Blue plus yellow tint SET 2 (0.25 ct) 2- Blue SET 2 (0.32 ct) 8 Blue SET 2 (0.50 ct) 4 Blue SET 2 (0.75 ct) 2 Blue plus pink tint SET 2 11.23 ct) 6 Blue

Claims (10)

1. A method of assessing the colour of a diamond or other gem which comprises aligning the table of the gem normal to the axis of light projected from a source, and making the axis of the light co-linear with an axis of symmetry of the gem which is normal to the table, projecting the light of a given single wavelength at the gem and determining the proportion of incident light transmitted by the gem, repeating this determination for different wavelengths of light over the visible wavelength range so as to obtain a transmission spectrum and calculating the chromaticity co-ordinates therefrom.

2. A method according to claim 1 in which the gem is a diamond.

3. A method according to claim 2 in which the gem is a brilliant cut diamond.

4. A method according to claim 3 in which the gem is embedded in a universally accepted white standard solid up to its girdle.

5. A method according to claim 4 in which the solid is barium sulphate.

6. A method according to any one of the preceding claims in which the transmission spectrum is obtained by making determination for at least 10 different wavelengths.

7. A method according to claim 6 in which a determination is made at 1 mm intervals throughout the visible wavelength range.

8. A method according to any one of the preceding claims in which the aligned gem is positioned in an integrating sphere prior to projecting the light at it.

9. A method according to claim 1 substantially as described in any one of the Examples.

10. A set of diamonds or other gems for use as a comparative standard for colour assessment, in which set the diamonds or other gems have been graded with reference to their chromaticity coordinates, said co-ordinates determined by a method as claimed in any one of the preceding claims.