EP0731908A1

EP0731908A1 - Methods and apparatus for indentation, scratch or tribological testing

Info

Publication number: EP0731908A1
Application number: EP95932849A
Authority: EP
Inventors: Gillies David Pitt; Paul Hayward Ian; Paul Centre Suisse d'Electronique et ALERS
Original assignee: Centre Suisse dElectronique et Microtechnique SA CSEM; Renishaw PLC
Current assignee: Centre Suisse dElectronique et Microtechnique SA CSEM; Renishaw PLC
Priority date: 1994-09-30
Filing date: 1995-10-02
Publication date: 1996-09-18
Also published as: WO1996010737A1

Abstract

A scratch tester is described, in which a film or coating on a sample (10) is scratched by a stylus (14) in order to examine qualities such as its cohesion or adhesion to the substrate. Simultaneously or subsequently, the scratched surface is illuminated by laser light, e.g. through a microscope (42), in order to generate Raman scattered light. The Raman scattered light passes back through the microscope (42) and is analysed to determine stresses or strains present in the scratched region. This enables the determination of quantitative information about the quality of the film or coating. Similar Raman analysis methods may be used in indentation testing and tribological testing.

Description

METHODS AND APPARATUS FOR INDENTATION, SCRATCH OR TRIBOLOGICAL TESTING

Field of the Invention

This invention relates to methods and apparatus for analysing a sample, in which an indentation or scratch is made into the sample surface, or to tribological testing.

Description of Prior Art

To measure the hardness of a sample, it is very well known to make an indentation into its surface, for example in a Rockwell hardness test using a diamond tip. Where it is desired to measure the hardness of a coating on the sample, the depth of the indentation must be substantially less than the coating thickness (raicrohardness test) . For very thin films and coatings (e.g. below 2 μm) the indentation must be very small and can be difficult to. measure with a conventional optical microscope. For such measurements, the ultra raicrohardness test is available (also called nanohardness test or nano-indentation test) . An ultra microhardness tester is available from Centre Suisse d'Electronique et de Microtechnique SA (CSEM) , Maladiere 71, CH-2007, Neuchatel , Switzerland. This device measures the loading force and the penetration depth during a loading and unloading cycle with a specially shaped indentor, and provides information about the plastic and elastic deformation of the sample.

Another type of known test which is useful for testing coatings and films is the scratch test and micro scratch test. Suitable test machines are again available from CSEM under the trademarks Revetest and MST. In these machines, the surface of the sample is scratched and the coating/substrate interface is deformed by relative movement between the sample and a diamond point. The load applied to the diamond point may increase continuously as it travels along the surface. Critical points along the scratch may be determined by monitoring the load force (normal to the sample surface) against the frictional force (in the direction of the scratch) . A breakdown in the cohesion or adhesion of the film or coating is indicated by a sudden increase in the frictional force. Alternatively or additionally, the machine has an acoustic emission detector, which monitors the acoustic emission produced during the scratching process. Again, breakdowns in the coating or film are accompanied by sudden increases in the acoustic emission.

The CSEM micro scratch testing machine also has a built-in optical microscope, with which the scratch can be inspected visually after the test has been performed. It is possible to view the position along the scratch at which visible damage started to occur.

A review of the above indentation and scratch techniques is provided by H.E. Hintermann, "Characterisation Of Surface Coatings By The Scratch Adhesion Test And By Indentation Measurements", Fresenius Journal of Analytical Chemistry (1993) 346:45-52.

Raman spectroscopy is an entirely separate analytical technique, in which a sample is irradiated with light at an exciting frequency, e.g. from a laser. As a result of interactions with molecular vibrations in the sample, a spectrum of Raman scattered light is produced and can be analysed. Many substances have characteristic Raman spectra. It is also known that strain or stress can have an effect on the Raman spectrum of a sample, for example by shifting a peak of the spectrum to a different wavenumber, or by broadening the width of a peak.

Apparatus and methods for Raman spectroscopy are described in, for example, European Patent Application No 543578. This describes a Raman system which is commercially available from Renishaw pic, Old Town, Wotton-Under-Edge, Gloucestershire, GL12 7DH, United Kingdom.

Summary of the Invention

In at least some aspects, the present invention relates to a novel method of analysis, in which an indentation, scratch or tribological test is performed, and the resulting indentation or scratch is examined by Raman spectroscopy. The Raman examination may be performed either subsequently to the formation of the indentation or scratch, or during it. The present invention also relates to apparatus for performing the method. We have found that the Raman analysis may show up strain in and around the area of the indentation or scratch, giving information which is not available from the existing measurement methods used on such machines.

In a further aspect, the present invention provides a probe for use in an indentation, scratch or tribological test, in which Raman scattered light is detected simultaneously with said test.

Brief Description of the Drawings

Preferred embodiments of the invention will now be described by way of example, and with reference to the accompanying drawings, wherein:

Fig 1 is a schematic diagram of a scratch testing machine,

Fig 2 is a schematic diagram showing details of parts of the machine of Fig 1,

Fig 3 is a schematic diagram of a Raman spectrometer,

Fig 4 is a schematic diagram showing apparatus in which elements from Fig 1 and Fig 4 have been combined.

Fig 5 is a schematic plan view of a scratch in a sample,

Fig 6 is an enlargement of part of Fig 5, Figs 7 and 8 are graphs of results obtained from the use of the apparatus of Figs 1-3, in accordance with a preferred method of the present invention,

Fig 9 is a schematic diagram of a further Raman detection apparatus;

Fig 10 is a partly sectional view of a probe for use in scratch or indentation testing;

Fig 11 is a plan view of apparatus for tribological testing; and Fig 12 shows a modification of Fig 10.

Description of Preferred Embodiments

Figs 1 and 2 show a scratch testing machine available from CSEM and similar to those described in the above-referenced paper by Hintermann. A sample 10 may be mounted on a table 12 which is traversable in a direction x. As the sample 10 is moved in the direction x, it can be scratched by a needle 14 having a diamond tip of known characteristics, e.g. a Rockwell C diamond indentor having a hemispherical tip of radius 200μm. As the sample is moved at a constant rate dx/dt, force is applied to the needle 14 from a force applying device 16. This increases at a rate dL/dt. Once the scratch has been made in the sample 10, the table 12 can move it underneath a microscope 18, where it can be examined visually. However, this microscope is not essential.

This apparatus is shown in more detail in Fig 2. A motor 20 operates the table 12, with servo feedback from a suitable scale (not shown) to a control unit 22 and computer 24. These enable precise control of the speed dx/dt of the table 12 and also precise positioning of the sample 10 relative to the needle 14 and microscope 18. A motor 26 applies the gradually increasing force dL/dt to the needle 14, by any suitable mechanism, shown schematically as a lever 28. The scratch tester has an acoustic emission detector AE on the needle 14. The table 12 is also provided with a sensor for measuring the frictional force between the needle 14 and the sample 10 as the motor 20 is driven. This sensor may for example measure the driving force of the motor 20. Both the frictional force Ft and the acoustic emission from the sensor AE are monitored by the computer 24 during the scratch test. Breakdowns in the cohesion or adhesion of a film or coating on the sample 10 are accompanied by increases in the frictional force and the acoustic emission, which can be plotted on a graph against the applied load L by the computer 24. The x position at which such effects occur can also be determined by the computer. Such x positions may also be correlated with the visual image of the scratch seen in the microscope 18, knowing the offset distance in the x direction by which the sample 10 is moved on the table 12 to position it under the microscope 18.

Fig 3 shows a Raman spectrometer of the type described in more detail in the above-referenced European Patent Application No. 543578. Light from a laser source 30 is directed to a sample 10 via mirrors 32,34,36 and a holographic notch filter 38. It is focused on the sample 10 by an objective 40 of a microscope 42. Raman scattered light passes back through the objective 40 and the holographic notch filter 38, which rejects light (e.g. reflected and Rayleigh scattered light) having the same frequency as the laser source 30. A further such holographic notch filter 44 may be provided to improve the rejection of the reflected and Rayleigh light if required. A rugate filter or filters may be used in place of the filters 38,44. The scattered Raman spectrum is then analysed by an analyser 46, and is focused by a lens 48 onto a detector such as a cooled charge coupled device (CCD) 50. The analyser 46 is suitably a diffraction grating, which disperses the spectrum across the CCD detector 50. However, it is also possible to use a tunable filter, in which case it is possible to image an area of the sample 10 onto the CCD 50, in light of a selected Raman wavenumber. With the diffraction grating, the system is focused such that only a point on the surface of the sample 10 is examined, whereas if an area is to be examined the laser source 30 can be de-focused. The data from the CCD is acquired by a computer 54 for further analysis. This computer may also control the Raman analyser 46.

The Raman spectrometer has a table 52 for mounting the sample 10 under the microscope 42. This table may be motorised with scales and servo control in the same way as the table 12 in Fig 1, except that desirably it is motorised in the y direction (normal to the plane of Fig 3) as well as the x direction. There is also advantage if it is similarly movable in the z direction (vertically) , since this obviates the need to level the sample when it is mounted. The table 52 may be controlled by the computer 54.

A novel method according to the invention proceeds as follows. First, a sample is scratched on the scratch testing device shown in Figs 1 and 2. The sample may consist, for example, of a substrate having a thin coating or film, the cohesion or adhesion of which is to be tested. One or more scratches may be made in the normal way, and examined visually under the microscope 18 if desired.

Having performed a scratch test upon the sample, the sample is transferred to the Raman spectrometer shown in Fig 3.

With the scratched sample 10 mounted on the table 52 of Fig 3, first of all the sample is viewed under the microscope 42, in order to discover the alignment of the scratch on the sample surface relative to the x and y axes of movement of the table 52. This is necessary because it is difficult or impossible to align the sample in exactly the same way as it was aligned on the table 12 of the scratch tester. Either the mirror 36 can be removed for this alignment, so that the alignment takes place using the microscope in the normal manner, illuminating the sample 10 with white light. Alternatively, the filters 38,44 and Raman analyser 46 may be temporarily removed from the optical path, again enabling an image of the sample 10 to be viewed in white light on the CCD 50. In either case, the xy table 52 is adjusted so that a datum point (e.g. the centre of a graticule) is aligned on a point on the scratch. The xy coordinates of this point are noted from scales of the table 52. This is repeated for another point on the scratch. This gives the alignment of the scratch relative to the xy coordinate system of the table 52, and it is now possible for the computer 54 of the device to move the table 52 in such a manner that the point of focus of the laser source 30 moves in alignment with the scratch.

We have found that the best method of such alignment is if the two points which are taken are the visible start and end of the scratch. The data produced by the computer 24 of the scratch tester (Figs 1 and 2) easily enables correlation between the end point of the scratch (i.e. the point at which the table 12 was commanded to stop movement) and the sudden increases in the frictional force Ft or the acoustic emission. It is found in practice that the visible start of the scratch seen on the microscope 42 of the Raman spectrometer may not correspond exactly with the start of the scratch as indicated by the data from the computer 24 of the scratch tester. This is because at the start of the scratch, the scratching force is very low and so the scratch may not be visible. Nevertheless, from the data from the computer 24 the length of the scratch is known. This known length may be used in conjunction with the visible end of the scratch seen in the microscope 42, and the alignment of the scratch as determined above, to determine in the xy coordinates of the table 52 the actual start of the scratch. If desired, it would be possible to move the xy table 52 of the Raman spectrometer directly to points on the scratch, taking Raman spectra at such points. Useful results could be obtained merely by programming the computer 54 to move the table 52 such that the focus point of the laser traverses along the length of the scratch, for example at the mid point of its width. During such a traverse along the scratch, at each of a plurality of points the Raman analyser 46 and CCD 50 are used to take a Raman spectrum. Preferably the diffraction grating is used for the Raman analyser 46. From these Raman spectra, strain of the material of the film or coating may be determined, by examining the width and/or the Raman shift of a peak characteristic of the material of the film or coating. It is also possible simultaneously to measure the strain (if any) in the substrate of the sample, underneath the film, by examining a Raman peak characteristic of the material of the substrate in the same way.

However, we have found that more interesting analysis can be performed by the method which follows, illustrated in Figs 5 and 6. At each of a plurality of points along the scratch 60, a lateral scan is made, as indicated by the lines 62 transverse to the scratch 60 in Fig 5. Fig 6 is an enlargement of one such lateral scan from Fig 5. At a large number of points 64 along the lateral scan line 62, the table 52 is stopped and a spectrum is acquired by the CCD 50, using the diffraction grating as the Raman analyser. The computer 54 then moves the table 52 to the next acquisition point 64 along the lateral scan line 62, and another spectrum is acquired.

If lower spectral resolution can be tolerated, it would be possible to perform this method using the tunable filter instead of the diffraction grating. This would have the advantage that an area around one or several of the lines 62 can be imaged, and data acquired simultaneously from numerous points 64. The computer 54 may then plot the results of such scans either as shown in Fig 7 or as shown in Fig 8. These show results from an actual test, taken on a sample consisting of a silicon carbide (SiC) coating on a silicon (Si) substrate, using the diffraction grating. The graphs of Figs 7 and 8 are derived by analysing changes in the characteristic peak of the silicon substrate at 520cm^'1, though similar graphs may also be produced by analysing a peak characteristic of the SiC coating if desired. In Fig 7, each plotted line represents the half width at half maximum of the Si peak. It will be seen that near the start of the scratch track (the lower lines plotted in Fig 7) this half width changes little with the lateral position relative to the centre line of the scratch. In contrast, at 2500μm from the visible start of the track of the scratch, substantial changes in the half width of this peak can be seen at certain lateral positions. This indicates the presence of strain in the silicon substrate at these positions. By examining the various lines, it may be discerned that first signs of damage occur at about 1250μm from the visible start. This is attributed to plastic deformation of the Si substrate.

Of course, it is easy to correlate this distance of 1250μm along the track with the exact position in the data acquired by the computer 24 of the scratch tester, e.g. by correlating it with the end point of the scratch. From the data acquired by the computer 24, it is possible to relate this distance of 1250μm with the positions (and the corresponding instantaneous scratching loads) at which visible signs of damage are seen in the microscope 18, and at which increases are seen in the frictional force Ft and the acoustic emission AE.

Fig 8 is a similar graph to Fig 7, except that instead of showing how the half width of the Si peak changes, it shows the Raman shift of that peak away from 520cm^"1. Again, clearly there is a very visible shift at distances such as 2000 or 2500μm from the visible start of the scratch. However, by examining the various lines in Fig 8, such a shift can be discerned even at about 750μm from the visible track start. This Raman shift is attributed to stressing and/or relaxing of the Si beneath SiC coating.

If desired, similar plots could be made for a characteristic peak of the SiC coating material.

When the x positions determined from Figs 7 and 8 are correlated with the distances and the corresponding instantaneous scratching loads recorded by the computer 24 of the scratch tester, it is found in practice that strain effects can be detected by the Raman analysis at much lower loads (i.e. earlier along the scratch) than can be detected using the frictional force or acoustic emission data.

Furthermore, the Raman data is much less dependent on experimental conditions that the friction or acoustic emission data. The Raman data therefore gives definitive numerical results which characterise the coating/substrate system. This is an improvement over friction or acoustic emission data, which normally give results which can only be compared qualitatively with results from similar tests performed under the same experimental conditions.

So far, an example of a method according to the invention has been described which uses two separate instruments, the scratch tester and the Raman spectrometer. These instruments may located a great distance apart from each other, and the sample may be tested on each instrument at different times. The only modification required for these instruments is software in the computer 54, which enables automation of the scanning process shown in Figs 5 and 6, and which facilitates the acquisition of the start and end points of the scratch for alignment purposes. However, Fig 4 shows an" improved apparatus, which combines both the scratch tester and the Raman spectrometer. Here, a single xy motorised table 12 is provided, upon which the sample 10 is mounted. (As mentioned above, the table may also advantageously be motorised in the z direction.) The force applying device 16 of Fig 1 and needle 14 are provided in exactly the same way as in Fig 1. In place of the microscope 18 of Fig 1, we provide the microscope 42 of the Raman spectrometer shown in Fig 3. Thus, the Raman analysis may be performed upon a newly scratched sample 10 merely by moving the table 12 so that the sample 10 is located under the microscope 42 of the Raman spectrometer. This apparatus has the advantage that the angular alignment of the sample is assured, and it is not necessary to measure the visible start and end points of the scratch provided that the offset between the needle 14 and the microscope 42 is known. It also simplifies the lateral scanning procedure illustrated in Figs 5 and 6, since it is assured that the scratch is in the x direction of the table and thus the lateral scans may merely proceed in the y direction (i.e. normal to the plane of Fig 4) . Results may be obtained from this apparatus in an otherwise similar manner to that explained above.

The examples above have made use of a scratch testing machine. However, this may if desired be replaced by an indentation or nano-indentation machine of the type described in the introduction, to obtain Raman information about the strains introduced by the indentation process.

Fig 9 shows a Raman detection device of a further embodiment of the invention. Most components are the same as in Fig 3, so they need not be described further. However, in place of the microscope 42 in Fig 3, the light from the laser source 30 is directed via the holographic notch filter 38 down an optical fibre 70 towards the sample. Raman scattered light from the sample passes back through the optical fibre 70 and is transmitted by the holographic notch filter 38, which rejects reflected and Rayleigh scattered light as previously.

Fig 10 shows one possible arrangement at the other end of the optical fibre 70 of Fig 9. The equipment shown in Fig 10 is basically a scratch or indentation or nano- indentation tester, as in Figs 1 and 2.

As shown schematically in Fig 10, an indentation is made into the surface of a sample 10 placed on a table 12, by means of a diamond tip 74 at the end of a stylus 14. The load L which produces the scratch or indentation is variable with time, and is applied by a motor (not shown) via a lever 28 having a fulcrum 29. Thus, in the case of scratch tester, the sample 10 is moved horizontally under the tip 74 by means of the table 12, and a gradually increasing force is applied to scratch the surface. The tip may have any desired shape, e.g. conical, hemi¬ spherical or pyramid-shaped. Instead of diamond, it may b made of another suitable material such as A1₂0₃ or sapphire The scratch or indentation test is especially useful when the sample 10 has a very thin coating or surface film (not shown) , as explained in the above paper by Hintermann.

The commercially available scratch or indentation tester i modified in that the stylus 14 is tubular, having the optical fibre 70 passing down it. The tip 74 has a flat, polished upper surface 75. The laser light from the laser source 30 is delivered by the optical fibre 70 and focused by a lens 72 through the surface 75 onto the tip 74. Preferably it is focused to the point 76 which is the interface between the tip 74 and the sample 10. Raman scattered light from this same point is then collected by the lens 72 and passes back along the optical fibre 70 and through the holographic filter 38, to be analysed by the Raman analyser 46 and detected by the CCD 50. Corresponding data is acquired by the computer 54. In use, Raman spectra are acquired by the computer 54 at numerous sample times during the course of a normal scratch or indentation test on a sample 10. It is preferable that the scratch or indentation tests should be controlled by the same computer 54 which controls the Raman detection, to facilitate the correlation of the times at which the various Raman spectra are acquired with the corresponding times and/or positions on the sample 10 of data normally provided during the scratch or indentation test. For example, it is desirable to correlate the different Raman spectra with the instantaneous applied load L. The Raman data may be analysed to determine instantaneous values of strain or stress present either in the sample 10, or in the diamond tip 74, or both. To do this, use is made of a preselected Raman peak or peaks characteristic of the material of the sample 10 or of the diamond of the tip 74. In particular, strain/stress dependent properties of such a peak or peaks are monitored, such as an increase in the width of the peak or a shift of the peak to another wavenumber when strain is present. Of course, the tip 74 may be made from another material rather than diamond, with a suitable strain/stress-sensitive Raman peak which is monitored.

The computer 54 is then able to output information, e.g. in graphical form, showing how the strains and stresses induced in the sample 10 vary during the cycle of an indentation or scratch test. As discussed in relation to Figs 1-8, changes in this data indicate critical points in the behaviour of the sample. For example, where the sample comprises a coating or film upon a substrate, the stress/strain data derived may indicate critical points at which there is a breakdown in the cohesion or adhesion of the film or coating. Effects such as plastic deformation of the material of the film or coating may also be determined. Advantageously, the Raman spectrometer in Figs 9 and 10 may make use of confocal techniques. This may involve the use of a pinhole as described by G J Puppels et al, Nature, Vol. 347, (20 Sept 1990) pp 301-303, or may be as described in International Patent Application No. WO 92/22793. The Raman scattered light which is analysed and detected may thereby be restricted to a small depth of field, say about lOμm in the region of the point 76. This ensures that the spectrometer is sensitive only to the stressed or strained material of the tip 74 and/or of the sample 10 within this small region. Similar confocal techniques may be used in Figs 3 and 4 if desired.

In Figs 9 and 10, where the sample 10 comprises a thin film or coating on a substrate, it is possible to determine not only stress/strain data for the tip 74 and for the material of the coating, but also for the underlying substrate material, by monitoring suitable Raman peaks characteristic of the coating and the substrate respectively. This yields still further information about the properties of the interface between the coating and the substrate.

If desired, it is possible to determine strain only of the tip 74, or only of the sample 10, by restricting the analysis to corresponding peaks in the Raman spectrum. In these cases, the lens 72 may be focused at a point within the tip 74, or within the sample 10, rather than at the interface point 76.

The stress/strain data for the material of the tip 74 is potentially useful because it gives a very direct indication of the instantaneous load applied at the point 76 to the sample 10.

Fig 11 is a plan view of apparatus for tribological testing of a sample 80 which is in the form of a disc. The sample 80 is mounted on a turntable 82 and rotated in the direction indicated by an arrow 84. A pin or needle 86 is mounted vertically at the end of a horizontal lever arm 88, and has a tip which presses down on the surface of the sample 80 as it rotates, producing a wear track 90. The lever arm 88 is mounted about a pivot 92, and the drag between the pin 86 and the sample 80 may be determined by measuring lateral forces on the lever 88 e.g. as indicated by arrows 94.

The tip at the lower end of the pin 86 may be of diamond or another suitable material, shaped in any conventional manner as commonly used in such pin-on-disc tribological testing. However the pin 86 is constructed in a similar manner to the stylus 14 in Fig 10. Thus, it is tubular and has the optical fibre 70 passing down it. Light from the laser source 30 (Fig 9) is directed to the tip of the pin 86 down the optical fibre 70, and Raman scattered light passes back up the optical fibre to the Raman analyser 46 and CCD detector 50, as previously. It is also possible to use confocal techniques as described previously.

Analysis of the resulting Raman data yields information about the strains and stresses experienced either in the tip of the pin 86, or in the material of the sample 80, or both. It may also give data on the composition of wear debris in the track 90, or on the breakdown of any lubricant which may be used, by monitoring appropriate parts of the Raman spectrum.

Raman data may also be obtained in the same way from a pin similar to the pin 86, but mounted in tribological apparatus of the type in which the pin undergoes reciprocating motion relative to the sample, rather than rotary motion.

Fig 12 shows a modified probe, generally similar to that of Fig 10. However, in place of a single optical fibre 70, there are two optical fibres 70A,70B. In practice, a single optical fibre cable having multiple cores may be used.

The optical fibre 70A carries the exciting laser light from a suitable source, without the need for it to be injected into the optical path by the filter 38 (Fig 9) . However, within the stylus 14, a filter 38A is provided on the end of the fibre 70A. This is a narrow passband filter which passes only the light of the laser frequency, rejecting all other frequencies. In particular, it rejects Raman-shifted light and fluorescence generated within the fibre 70A itself as a result of interaction between the laser light and the material of the fibre. Such unwanted Raman and fluorescence signals could otherwise cause spurious signals which mask the Raman effects which occur in the stylus tip 74 and sample 10.

As previously, the Raman scattered light from the tip 74 and sample 10 is collected by the lens 72. It is taken to a Raman analysis section according to Fig 9 via the optical fibre 70B. The fibre 70B also has a filter 38B at its end, within the stylus 14. The filter 38B is a narrow notch filter, designed to reject light having the same frequency as the laser. It therefore rejects Rayleigh scattered and reflected laser light, while passing the desired Raman signals.

The filter 38B provides further rejection of the strong signal of the laser frequency, assisting the filter 38 (Fig 9) in this respect. Furthermore, it minimises any tendency for spurious Raman and fluorescence signals to be generated within the fibre 70B by interaction between the laser frequency signal and the material of the fibre 70B.

For further details of optical arrangements in optical fibre probes for Raman purposes, which can be used with the present invention, reference is made to U.S. Patents 5,112,127 (Carrabba) and 5,377,004 (Owen et al) . Both these patents describe optical arrangements which may be incorporated within the stylus 14, in place of the filters 38A,38B and the lens 72. Alternatively, the probes of these U.S. patents may be mounted adjacent the probe seen in Fig 10, the objective lenses of the probes shown in the US patents being linked to the probe of Fig 10 by a short length of an optical fibre. By keeping this linking fibre short, any tendency to generate spurious Raman or fluorescence signals within the fibre will be minimised.

In place of the above arrangements, a single filter may be placed on the end of the fibre 70 within the stylus 14 in Fig 10. This should be a notch filter, tuned to the specific Raman line created by interaction between the exciting laser light and the material of the fibre. If there is more than one such Raman line, then the filter may have a corresponding number of notches at the appropriate wavenumbers to remove them. This can be achieved by having two or more filters in series, or by providing a single filter tailored to have the appropriate notches. In any case, the function of such a filter is to pass both the exciting laser light from the fibre 70 to the tip 74 and sample 10, and to pass the required Raman signals from the sample 10 and tip 74 back to the fibre 70, while removing the spurious Raman signals generated within the fibre. A similar filter or filters to remove spurious signals may be provided at the spectrometer end of the fibre 70, if required.

Claims

1. A method of analysing a sample, comprising the steps of: making an indentation or scratch or track in a surface of the sample, illuminating said surface with light, thereby causing the generation of Raman scattered light, and detecting the Raman scattered light.

2. A method according to claim 1 including a step of analysing the Raman scattered light, and wherein said analysing step includes a determination of stress or strain present in the sample as a result of the indentation or scratch or track.

3. A method according to claim 1 or claim 2, wherein said indentation or scratch or track is an elongate scratch, and Raman scattered light is analysed and detected from a plurality of locations along the scratch.

4. A method according to any one of claims 1 to 3, wherein Raman scattered light is analysed and detected from a plurality of locations extending across the indentation or scratch or track.

5. A method according to any one of the preceding claims, wherein the Raman scattered light is generated and detected simultaneously with the formation of the indentation or scratch or track.

6. Apparatus for analysing a sample, comprising: a needle, pin or stylus, having a tip for engaging a surface of a sample, a device for applying a force to the needle, pin or stylus, thereby causing the tip to form an indentation or scratch or track in the surface of the sample, a light source for illuminating the surface of the sample, thereby causing the generation of Raman scattered light, and a detector for detecting said Raman scattered light.

7. Apparatus according to claim 6, including an analyser for analysing the Raman scattered light and determining stress or strain present in the sample as a result of the indentation or scratch or track.

8. Apparatus according to claim 6 or claim 7, including a table for holding the sample which is movable relative to the needle, pin or stylus, whereby said tip forms an elongate scratch in the surface of the sample as said force is applied to the needle, pin or stylus; and whereby Raman scattered light can be detected from a plurality of locations along the track.

9. Apparatus according to claim 8, wherein said table is also movable transversely relative to the scratch, whereby

Raman scattered light can be detected from a plurality of locations extending across the scratch.

10. Apparatus according to any one claims 6 to 9, wherein the Raman scattered light is detected and generated simultaneously with the formation of the indentation or scratch or track.

11. Apparatus for analysing a sample, comprising: a needle, pin or stylus, having a tip for engaging a surface of a sample, a device for applying a force to the needle, pin or stylus, thereby causing the tip to form an indentation or scratch or track in the surface of the sample, a light source for illuminating said tip or the sample thereby causing the generation of Raman scattered light simultaneously with the formation of the indentation or scratch or track, and a detector for detecting the Raman scattered light.

12. Apparatus according to claim 11, wherein the sample is illuminated through the tip.

13. Apparatus according to claim 11 or claim 12, wherein Raman scattered light from the sample is passed to the detector through the tip.

14. Apparatus according to any one of claims 11 to 13, wherein the detector detects Raman scattered light generated in the tip, and including an analyser for analysing the Raman scattered light from the tip and determining stress or strain present in the tip during the formation of the indentation or scratch of track.