GB2292603A

GB2292603A - Testing systems for counting contaminant particles

Info

Publication number: GB2292603A
Application number: GB9417088A
Authority: GB
Inventors: Andrew John Spencer; Karen Kent
Original assignee: Texas Instruments Ltd
Current assignee: Texas Instruments Ltd
Priority date: 1994-08-24
Filing date: 1994-08-24
Publication date: 1996-02-28
Also published as: GB9417088D0

Abstract

A silicon test wafer (1) for testing a machine for counting contaminant particles bears a known regular pattern of marks (4) made by evaporating the silicon of the wafer with laser pulses to form pits. Those marks scatter light and thus contaminant particles to the machine. The machine may also be tested with the wafer to ascertain whether it is detecting the correct number of particles at the correct locations on the wafer. The pattern of marks identifies the location of the flat (2) on the circumference of the wafer. <IMAGE>

Description

TEST METHOD AND MATERIALS FOR USE THEREIN The present invention relates to the testing of systems for counting contaminant particles, especially systems used in clean rooms for the fabrication of semiconductor microcircuits.

Particulate contaminants include ordinary dust, bits of broken wafers, material produced by the wearing of bearings, SiO2 particles from reactors, and other reactants, reaction products and photoresist left over from wet or dry processing.

A measure of the performance of a semiconductor fabrication plant is the count of contaminant particles that become deposited on a wafer surface by the fabrication processes and by its exposure to the atmosphere in its clean rooms. The count is usually made by a machine.

If the count becomes too high as a result of a fault in the fabrication equipment or the clean room air filtration equipment, the yield of a fabrication plant will be adversely affected. A purpose of counting is to enable the increase in wafer surface particles to be detected before the yield is adversely affected.

One type of machine operates as follows. A wafer is mounted in the machine and is illuminated over all of its surface with a broad parallel beam of light. Particulates on the wafer scatter light from the incident beam. An image of the scattered light is formed on a detector array. The detector array is mounted to receive only light scattered at 20 from the reflected beam. The detected image is processed by a computer to derive a count of the particles on the wafer, which count is displayed on a screen. A map of the detected particles is also derived and is displayed on the screen with an 'x' marking the position of each detected contaminant particle.

Other machines raster scan a spot of light, sometimes from a laser, across the wafer and have a static detector that can detect scattered light from all points on the wafer.

To ensure the information such machines produce is valid, the machines are from time to time tested and adjusted if necessary. Such adjustment becomes necessary, for example, when the lamp used to illuminate the wafer is changed, to ensure that the machine still detects particles of the desired range of sizes. The method employed at present to test the size of the particles detected is to sprinkle latex spheres of a predetermined size onto a wafer that has been cleaned. The wafer is placed in the particle counter which is then, if necessary, brought into adjustment so that it counts the spheres. The counter should then count similarly sized contaminant particles. The latex sphere test is, however, limited.

The present invention provides, in a first aspect, a method of testing apparatus for counting contaminant particles on a sample, which comprises causing the apparatus to examine the appearance of a test sample carrying a known display of permanent marks that resemble, to the apparatus, the contaminant particles, comparing the results of the examination with the known display, and, if desired or required, adjusting the apparatus to conform the examination results and the known display.

The examination may detect the number of marks on the test sample and that number may be compared with the known number.

Knowing the number of marks on the sample means that the sample can be used to test whether the counter is counting the correct number and hence whether the counter is not detecting some marks or is detecting marks falsely. In the latex sphere test, the number of latex spheres on a sample is not controllable and so a latex sphere sample can only be used to test whether the counter is counting the spheres or not unless a special test were to be made to count the actual number of spheres.

The examination may detect the positions of the marks and those positions may be compared with the known positions.

Knowing the positions of the marks allows a particle counting machine to be tested to determine whether it detects each individual mark, for example, by making a visual comparison of a map of the marks on the test sample to the map produced by the particle counting machine from the test sample.

The examination may detect the size or distribution of sizes of the marks and that may be compared with the known size or distribution of sizes.

The known number, position, size or distribution of sizes may be input to the apparatus which may then make a comparison with the results of the examination. The apparatus may then perform self diagnosis or self adjustment in response to the comparison.

A plurality of test samples bearing different displays of marks may be examined.

The apparatus tested by the method may be one that examines samples optically.

The present invention provides, in a second aspect, a test sample bearing permanent marks that each resemble a contaminant particle on the surface of a sample to an apparatus for counting contaminant particles.

The permanency of the marks means that the sample may be cleaned of any actual contaminants resting on it and then reused, in contrast to the sample with the latex spheres because those spheres do not adhere to the sample and it is necessary to repeat the application of the spheres to the sample whenever the sample is cleaned.

The number of the marks may be recorded on the test sample or on a separate data carrier.

The positions of the marks may be recorded on the test sample or on a separate data carrier.

The present invention provides, in a third aspect, a test sample bearing marks that each resemble a contaminant particle on the surface of a sample to an apparatus for counting contaminant particles, on which, or together with a separate data carrier on which, is recorded the number of the marks.

The present invention provides, in a fourth aspect, a test sample bearing marks that each resemble a contaminant particle on the surface of a sample to an apparatus for counting contaminant particles, on which, or together with a separate data carrier on which, is recorded the positions of the marks.

In the third and fourth aspects the marks may be permanent.

The marks may be in a regular pattern, facilitating visual distinction between the marks and any contaminant particles on the test sample.

The test sample may have the shape of a semiconductor wafer and may be a semiconductor wafer.

Semiconductor wafers are usually generally circular in plan with one or more segments ground away, which features are commonly known as flats.

If the sample has a flat, then the positions of the marks with respect to a flat on the circumference of the sample may be recorded on the test sample or on a separate data carrier.

The marks may resemble dust particles.

The marks may be on the surface of the test sample.

The marks may be covered by a transparent layer of material.

The marks may be depressions in, or in a layer of, the test sample.

The separate data carrier may be paper.

The number of, or the positions of, or positions of with respect to the flat or the size or distribution of sizes of, the marks may be recorded in a machine readable form.

Where data is recorded on a separate carrier the test sample may bear a reference, machine readable or otherwise, linking it to that data. A machine readable reference would be of use to a particle counting machine storing data relating to several test samples for identifying which data should be compared with the marks detected.

The marks on the test sample may be made with a laser.

The laser may be focussed to a spot on the test sample with a lens. The lens may be comprised in a microscope. The focus of the microscope may be adjusted with an automatic focussing device, or by observing a spot of light on the sample produced by a beam of light passing through the microscope, or by bringing into focus an image, produced by the microscope, of a mark on the wafer already made by the laser. Those methods adjust the focus of the laser spot on the test sample.

Two or more wafers according to the present invention having different sizes or distributions of sizes of marks may make up a set. Such a set may be used to ascertain whether a contaminant particle counting apparatus counts effectively particles of different sizes.

There will now be described, by way of example only, a test sample and a method according to the present invention, with reference to the accompanying drawings, of which: FIGURE 1 is a plan view of a test wafer.

FIGURE 2 is an enlarged schematic cross section of part of the test wafer shown in Figure 1.

FIGURE 3 shows an apparatus for making the test wafer.

Figure 1 shows in plan view a semiconductor wafer 1 having a flat 2 on its circumference 3. The wafer bears a number of marks 4 that, to a contaminant particle counting machine, resemble contaminant particles in that such a machine would detect them as contaminant particles. The marks are quite small and so they are each diagrammatically marked by an 'x'. The figure is therefore similar to the display that a correctly functioning contaminant particle counting machine would produce if the wafer were to be tested in such a machine.

Each mark is produced by locally evaporating material from the wafer with laser pulses. The laser pulses produce a crater in the surface of the wafer.

Several laser pulses may be required to obtain a crater of the desired size.

Figure 2 shows schematically a cross section of part of a wafer 1 having a crater 5. The craters scatter incident light, various parts of their surfaces being oriented, like those of particles, at different angles.

To carry out a test of a particle counting machine a test wafer is first cleaned to remove any surface contamination that the wafer might have acquired during storage. The wafer is then placed in the machine which is set to scan the wafer. The craters are detected by the particle counting machine because they scatter incident light from the beam directed by the machine onto the test wafer.

The count of the particles detected and the map of their locations produced by the machine are then compared to the number and pattern of marks known to have been made on the wafer. Errors in the output of the machine detected by this method can be used to diagnose faults in the machine or the test process. Once the machine has been adjusted, the test can be performed again to confirm that the machine is working properly.

The pattern of marks on the wafer illustrated in figure 1 has several advantageous features.

Firstly, the pattern is a regular one. That enables the person performing the test to identify quickly, from the map displayed by the machine, which marks are not being detected and which detected marks are false. Such missing or extra marks might result from contamination on some part of the optical system of the machine.

Secondly, in each quadrant of the wafer there is a row of marks 6 near the edge of the wafer. That enables a test of whether the machine is scanning the entire surface of the wafer to be made. If, for example, the beam of light illuminating the wafer were slightly misaligned and so failed to illuminate a crescent shaped area to one side of the wafer then the end marks of one of those rows, or, in a severe case, the whole of that row, would not be detected. Such a fault would not be apparent from an ordinary contaminated wafer or from one prepared with latex spheres; it would simply appear that there were no particles in that narrow area.

Thirdly, the marks 7 of the row near the flat 4 on the wafer are more closely spaced than the marks of the other rows. That enables the position of the flat 2 in the map produced by the machine to be identified allowing the position of any missing mark or extra mark in the map to be related easily to locations within the machine.

It is possible to produce laser marks of different sizes. By increasing the size of the laser spot formed on the wafer and the number and energy of the pulses larger craters may be produced. A test wafer, or a series of test wafers, having marks of different sizes may be used to test the performance of a particle counting machine as a function of particle size. That is of interest because the performance of clean rooms is often measured against particle size which is done because some sizes of particle cause more critical defects than others in semiconductor circuit fabrication. Such tests using a series of wafers overcome the problem suffered by the latex sphere test, which is that the latex spheres stick together in an uncontrollable way making particles that are bigger than single spheres.

The sizes of the tiny laser marks may be measured with a scanning electron microscope. Those sizes are not necessarily simply related to the true sizes of contaminant particulates that appear in a particle counting machine to be the same size as the laser marks, as particulates are convex and the laser marks are concave. In establishing the relationship, a scanning electron microscope may also be used to measure the sizes of contaminant particles. Such measurements are not, however, critical with the test wafers of the present invention. Since the marks are permanent, the performance of a semiconductor fabrication plant can be directly related to the results obtained using particular test wafers.

The preferred method of forming the marks on a test wafer is as follows.

The starting substrate is a normal polished silicon wafer. No preprocessing is performed as that might introduce unwanted defects on the surface of the wafer.

The wafer 1 is mounted on the stage 8 of a x500 microscope 9 as is shown in Figure 3. The stage 8 is used to move the wafer to a position at which it is desired to make a mark and five pulses from a YAG laser 10 having a wavelength of 532 nm are focussed through lenses of the microscope onto the wafer. Each pulse has an energy of roughly 0.1 MJ. Craters so produced were measured in a scanning electron microscope to have a diameter of 3 to 4 micrometres. The laser pulses pass through a x10 eye piece 11, a pair of prisms 12 and a x50 objective lens 13. The prisms 12 allow an observer 14 to view the sample via a pair of x10 eye pieces 15. For simplicity only one of the eye pieces 15 is shown in Figure 3. To prevent eye damage, a filter 16 protects the observer from receiving reflected laser light. The filter passes much of the visible spectrum which allows the sample to be viewed.

The microscope is focussed onto the wafer with an automatic focussing device 17 before the laser pulses are fired, which device uses infrared or ultrasonic waves to measure the distance between the microscope and the sample. That ensures that when the laser is fired the light is focussed to a small spot. The wafer is then moved a short distance sideways using the stage. A distance of 20 mm does not usually affect the focus substantially. In moving the wafer over longer distances, unevenness of the wafer or imperfections in the microscope stage may alter the distance between the lens and the wafer surface necessitating re-focussing.

To build up a pattern of marks, the stage is moved and the laser fired repeatedly. Usually the marks are made at intervals of 5 mm. It is not normally necessary to focus the microscope before each mark is made; focussing the microscope every second or third mark has been found to be sufficient.

Smaller craters may be produced by increasing the magnification of the objective lens and by reducing the size of the aperture 18 at the output of the laser 10.

The silicon evaporated by the laser can redeposit on the surface creating unwanted marks. Those deposits are not strongly adherent to the wafer and can usually be cleaned off.

The process may be automated using a stepper to move the wafer.

As an alternative to automatic focussing the focusing of the microscope may be carried out manually by the observer.

The wafer 1 is a featureless mirror and so it is difficult to judge whether it is in focus.

One method of manual focussing is to make use of a beam of ordinary light provided by a lamp in the laser enclosure, which beam has similar beam characteristics to those of its laser light. The spot of light produced on the wafer 1 by the ordinary light beam can be seen through the microscope 10 by the observer 14, who may then adjust the focus of the microscope until the spot size is a minimum.

Another method is for the observer to focus the microscope on the edge of the wafer; in another he may wait for a dust particle to settle on the wafer and focus on that.

Once a mark has been made by the laser the observer may focus the microscope on that mark; he may then move the stage a few millimetres, without adversely affecting the focus, before firing the laser to make another mark.

Claims

CLAIMS:

1. A method of testing apparatus for counting contaminant particles on a sample, which comprises causing the apparatus to examine the appearance of a test sample carrying a known display of permanent marks that resemble, to the apparatus, the contaminant particles, comparing the results of the examination with the known display, and, if desired or required, adjusting the apparatus to conform the examination results and the known display.

2. A method as claimed in claim 1, wherein the examination detects the number of marks on the test sample and that number is compared with the known number.

3. A method as claimed in claim 1 or claim 2, wherein the examination detects the positions of the marks and those positions are compared with the known positions.

4. A method as claimed in any one of claims 1 to 3, wherein the examination detects the size or distribution of sizes of the marks and that is compared with the known size or distribution of sizes.

5. A method as claimed in any one of claims 1 to 4, wherein the known number or positions or size or distribution of sizes are input to the apparatus which then makes a comparison with the results of the examination.

6. A method as claimed in claim 5, wherein the apparatus performs self diagnosis or self adjustment in response to the results of the comparison.

7. A method as claimed in any preceding claim, wherein a plurality of samples bearing different displays of marks are examined.

8. A method a claimed in any preceding claim, wherein the apparatus is one that counts contaminant particles optically.

9. A test sample bearing permanent marks that each resemble a contaminant particle on the surface of a sample to an apparatus for counting contaminant particles.

10. A test sample as claimed in claim 9, on which, or together with a separate data carrier on which, is recorded the number of the marks.

11. A test sample as claimed in claims 9 or 10, on which, or together with a separate data carrier on which, is recorded the positions of the marks.

12. A test sample bearing marks that each resemble a contaminant particle on the surface of a sample to an apparatus for counting contaminant particles, on which sample, or together with a separate data carrier on which, is recorded the number of the marks.

13. A test sample bearing marks that each resemble a contaminant particle on the surface of a sample to an apparatus for counting contaminant particles, on which sample, or together with a separate data carrier on which, is recorded the positions of the marks.

14. A test sample as claimed in claim 12 or claim 13, wherein the marks are permanent.

15. A test sample as claimed in any one of claims 9 to 14, wherein the marks are in a regular pattern.

16. A test sample as claimed in any one of claims 9 to 15, wherein the test sample has the shape of a semiconductor wafer.

17. A test sample as claimed in claim 16 of generally circular plan and having a flat on its circumference, on which sample, or together with a separate data carrier on which, is recorded the positions of the marks with respect to a flat on the circumference of the sample.

18. A test sample as claimed in claim 16 or claim 17, wherein the test sample is a semiconductor wafer.

19. A test sample as claimed in any one of claims 9 to 18, wherein the marks resemble dust particles to an apparatus for counting contaminant particles.

20. A test sample as claimed in any one of claims 9 to 19, on which, or together with a separate data carrier on which, is recorded the size or the distribution of sizes of the marks.

21. A test sample as claimed in any one of claims 9 to 20, wherein the marks are on the surface of the test sample.

22. A test sample as claimed in any one of claims 9 to 20, wherein the marks are covered by a transparent layer of material.

23. A test sample as claimed in any one of claims 9 to 22, wherein the marks are depressions in, or in a layer of, the test sample.

24. A test sample as claimed in any one of claims 10, 11, 12, 13, 17 or 20 or in a claim dependent on those claims, wherein the data carrier is paper.

25. A test sample as claimed in any one of claims 10, 11, 12, 13, 17, or 20, or in a claim dependent on those claims, wherein the number of, or the positions of, or positions of with respect to the flat, or the size or distribution of sizes of the marks is or are recorded in a machine readable form.

26. A method of making a test sample as claimed in any one of claims 9 to 25, wherein the method comprises making the marks with a laser.

27. A method as claimed in claim 26, wherein the laser is focussed to a spot on the test sample with a lens or lenses.

28. A method as claimed in claim 27, wherein the lens or lenses are comprised in a microscope.

29. A method as claimed in claim 28, wherein the microscope is focussed with an automatic focussing device.

30. A method as claimed in claim 28 or claim 29, wherein the microscope is focussed by an observer observing a spot of light on the test sample produced by a beam of light passing through the microscope.

31. A method as claimed in any one of claims 28 to 30, wherein the focus of the microscope is adjusted by bringing into focus an image, formed by the microscope, of a mark on the wafer already made by the laser.

32. A set of test samples, each test sample being a test sample as claimed in any one of claims 9 to 25, or made by the method of any one of claims 26 to 31.

33. A set of samples as claimed in claim 32, wherein the marks of a test sample of the set are of a selected size or have a selected distribution of sizes and that size or distribution of sizes differs from that of another sample of the set.

34. A test sample substantially as herein described with reference to, and as illustrated by, Figures 1 and 2 of the accompanying drawings.

35. A method of making a test sample, wherein the method is substantially as herein described.

36. A method of testing an apparatus for counting contaminant particles, wherein the method is substantially as herein described.