JP3476742B2 - Method and apparatus for identifying material of particles generated on the surface of a substrate - Google Patents

Description

Detailed Description of the Invention

[0001]

The present invention relates to optical inspection of flat surfaces of substrates such as silicon wafers, computer disks, computer flat display screens and the like.

[0002]

Optical inspection methods are often used to inspect the quality of flat surfaces of substrates such as silicon wafers and computer disks. In most such inspection systems, a beam of laser light is incident on the surface and the light scattered and reflected from the surface is collected and converted into an electrical signal. These electrical signals are analyzed to estimate the presence and size of some defects on the surface. At least when optically inspecting silicon wafers, which are used as starting materials for making integrated circuit chips, the main concern about the type of defect is the contamination of the surface by particles.

Particles on the wafer surface can interfere with the lithographic process. This lithographic process forms lines of electrically conductive material on the surface. In general, any particle whose diameter is greater than half the width of the electrical line laid on the surface results in unacceptable defects. If there are too many such particles, the wafer should be removed.
Currently, integrated circuits are made with line widths as thin as 0.25 μm (250 nm), so the diameter that appears on the wafer surface is 12
Particles larger than 5 nm cause the wafer to be removed, while particles smaller than 125 nm are acceptable. The semiconductor industry is rapidly moving towards the production of circuits consisting of 0.18 μm then 0.15 μm lines, which means that much smaller particles will soon become of interest.

Wafer inspection systems must be calibrated to function properly and to accurately determine particle diameter. In general, calibration is performed by intentionally placing a plurality of particles of various known diameters on the wafer surface and inspecting the wafer with an inspection device so that scattered light made from particles of different sizes is Brightness can be corrected for particle size. Usually these calibration particles are
It has a spherical shape made from polystyrene latex (PSL).

[0005]

One problem encountered in the wafer inspection process is that the same size particles of different material types can produce significantly different scattered light intensities. Viewed from different angles, two particles made of different materials and having significantly different diameters can produce nearly the same measured scattered light intensity. For example, it has been found that silicon particles of a given diameter produce much greater scattered brightness than PSL particles of the same diameter. In fact, of the various types of materials that commonly appear on the wafer surface in forming particles, PSL particles tend to be one of the lowest sources of light scattering. This tends to overestimate the diameter of silicon particles and many other materials after they have been calibrated with PSL particles. Therefore, the wafer is removed as if there were particles larger than half the line width, even though the particles were actually smaller than half the line width. Thus, if there are known materials for the particles, the accuracy of sizing the particles with scattered light can be greatly enhanced.

Another advantage to the semiconductor industry in being able to identify the material of the particles is that this information provides a strong clue as to the source of contamination. The rapid detection and removal of pollution sources is economically important because particle contamination must be reduced to a level where useful products can be made.

For light of a given wavelength, all materials have a refractive index n that indicates how much the speed of light is reduced in the material and an absorption coefficient that usually indicates how opaque the material is to light. have k and. The combination of n and k, known as material constants, is unique for each different material. The combination of n and k can be roughly divided into four groups. That is,
(1) Dielectrics such as PSL, SiO ₂ and Al ₂ O ₃ (n is small, k = 0), (2) Semiconductors such as silicon (n is large, k is small), (3) Tungsten, etc. Gray metal (n is large and k is large), and (4) a good conductor such as silver (n is small and k is large).

The combination of particle material and the n and k values of the substrate surface, together with the particle material, completely and uniquely define the pattern of light scattered by the particle for any given light source. Furthermore, for particles whose average diameter is less than about one fifth of the wavelength of the illumination light, the particle shape does not play a significant role in determining the scattering pattern. Thus, for visible light and particles below about 100 nm, knowledge of the average particle diameter and various material constants is sufficient to calculate the scattering pattern for a given scattering shape. This fact has allowed the development of scattering models that predict the scattering pattern for a given set of conditions. These models have been confirmed experimentally and the results have been published.

What is desired is a system and method for solving more difficult inverse problems. That is,
It is desirable to be able to determine the material of the particles and the average particle diameter from knowledge of the scattering pattern. So far, methods have been developed to determine the average particle diameter by analyzing the scattering pattern, as described, for example, in US Pat. No. 5,712,701. This patent is incorporated herein by reference. However, as mentioned above, the accuracy of such methods depends on the calibration of the system, which must now be done using the PSL sphere. The material constants of PSL spheres are significantly different from some other materials that may appear as particles on the wafer.

Methods have been proposed for identifying particulate material. For example, US Pat. No. 5,037,202 to Batchelder et al. Describes two parallel beams of light that are initially coherent with each other but have different polarizations are focused at a focal plane (such as the surface of a wafer) such that they are at each other at the focal plane. Disclosed are methods and devices that are biased away from. After the beam is reflected from the surface, another optical system blocks and combines the beams, and the particle-induced phase shift in one beam is manifested by a change in the elliptically polarized light of the combined beam. To be First detector is first
A second detector for sensing the intensity of the combined beam along the polarization axis of the second detector to produce a first output and a second detector for sensing the intensity of the combined beam along the second polarization axis for the second detector. Produces the output of. These first and second outputs are summed to provide an extinction signal and subtracted to provide a phase shift signal. It is said that the phase shift and extinction signals are correlated with the index of refraction of the particulate material, so that the material's identity can be determined based on the phase shift and extinction values. It is said that the size of a particle can be estimated from its position on the curve of absorbance vs. phase shift. Therefore, in the Batchelder system and method, information about the particles is estimated by analyzing the specularly reflected beam. The disadvantage of this method is that the reflected light is not very sensitive to changes in the properties of the particles, so small particles (eg particles on the order of 100 nm or less)
In the specularly reflected beam, very small changes that are difficult to measure accurately occur. For this reason, Bachelda's method is not optimal for identifying small size particles that begin to be a problem in integrated circuit manufacturing.

US Pat. No. 5,515,163 to Kupershmidt et al. Discloses a polarized laser beam intensity modulated at a first frequency and split into two orthogonal polarized beams, the two beams being a second beam. Method and apparatus are phase shifted with respect to each other at a frequency of. The two phase-shifted beams are directed onto the surface to be inspected and the light scattered by the particles at an angle to the two beams is detected. The detected light is simultaneously demodulated to determine the amplitude of the scattered light at the frequency of the intensity modulation and the amplitude and phase of the scattered light at the frequency of the phase modulation. It is said that these quantities can be correlated to particle size and refractive index to identify particles. The Cooper-Schmidt method involves complex calculations and the measurement requires a large number of modulation cycles in order to obtain an accurate measurement for a given scan portion of the surface to be inspected. Therefore, sampling the entire surface tends to be relatively time consuming.

[0012]

SUMMARY OF THE INVENTION We use the appropriate source and angle of incidence of the illuminating beam, and with the correct collector geometry to measure the scattering pattern in sufficient detail. We have found that we can solve a more difficult inverse problem for determining the material and dimensions of a material from knowledge of the scattering pattern.

According to the invention, a P-polarized beam is incident on the surface of the substrate at an oblique angle of incidence, and of a hemispherical space above the surface, including at least the forward and backward positions with respect to the illuminated portion of the substrate surface. Scattered light is collected at multiple spaced locations. The brightness of the scattered light at the front and rear positions is measured and converted into a signal that defines the scattering pattern generated by the particles. The material of the particle is determined by comparing the detected forward and backward scatter signal magnitudes with a plurality of predetermined scatter patterns defined by the forward and backward scatter signal magnitudes for the plurality of materials. . The predetermined scattering pattern that most closely matches the detected scattering pattern is identified, and the material to which this predetermined scattering pattern corresponds is identified as the material of the particles. The substrate is preferably moved with respect to the incident light and the incident beam on the surface is scanned to detect and identify particles that occur everywhere on the substrate. The significant advantages of the present invention are:
Its ability to quickly identify particulate material with a single scan of the surface by the incident beam.

It is preferred to use a wide angle collector to collect light that is forward scattered over a range that is remote from but just before the reflected light. Width is about 20 ° to 50
It is preferred to collect the forward scattered light in a region at least about 5 ° away from the reflected light. Even more preferably, it surrounds the reflected beam and has a diameter of at least about 1
The forward scattered light is collected over an annular region having an inner circumference of 0 ° and an outer circumference of about 20 ° to 50 ° in diameter. However, other collector geometric arrangements and positions are desired, including positions off the plane of incidence or away from one side of the reflected beam towards a vertical plane that passes through the illuminated portion of the surface. If you want, you can use it for the front collector.

A width of about 20 ° to 45 °,
It is preferred to collect the backscattered light. It is preferable to be able to collect the backscattered light over a generally circular region centered about 45 ° to 75 ° away from the vertical plane passing through the region of the plane illuminated by the incident beam. If desired, the backscattered light can be collected over an area off the plane of incidence formed by the plane perpendicular to the incident beam, as shown in FIG.

According to a preferred embodiment of the present invention, the first material identification parameter is calculated as the ratio of the backscatter signal magnitude to the forward scatter signal magnitude. Determining the particulate material was performed by comparing the back / forward ratio and the magnitude of the backscatter signal with the correlation of the back / forward ratios to the magnitude of the backscatter signal for the materials and measured. It consists of identifying the material that has the closest correlation to the point defined by the back / forward ratio and the magnitude of the backscatter signal. These correlations are preferably stored in a data storage unit and preferably comprise a comparator which compares the measured data with the correlation stored in this data storage unit.

In a further preferred embodiment of the present invention, once the particle material is determined, the particle size of the particle is determined from the backscatter signal magnitude based on the correlation between the particle material, diameter and backscatter signal magnitude. The average diameter is determined. Multiple correlations of particle diameters with backscatter signal magnitudes corresponding to multiple materials are stored in a data storage unit and a comparator searches for correlations corresponding to the identified materials and based on the retrieved correlations. It is preferable to be able to determine the particle diameter from the magnitude of the backscatter signal.

For some particle materials and diameters, the backward / forward ratio correlation is directed to more than one material, and the particle material is trusted from knowledge of the magnitude of the backscatter signal and the backward / forward ratio. It may not be able to be identified easily.
Therefore, according to another preferred embodiment of the invention, scattered light is collected from the central region of the space closest to the vertical plane passing through the region of the illuminated surface and the brightness of the collected light is measured. At the same time, it is converted into a central scattered signal having a magnitude indicating this brightness. A second material identification parameter is calculated as the ratio of the backscatter signal magnitude to the central scatter signal magnitude. The particulate material is determined based on both the correlation of the particulate material with the backscatter signal magnitude and the back / forward ratio and the second correlation of the particulate material with the backscatter signal magnitude and the back / center ratio. . More specifically, if the correlation based on the back / forward ratio does not provide good material discrimination, the back / center ratio and the magnitude of the backscatter signal are compared to the back / center to the magnitude of the backscatter signal for multiple materials. The particulate material is determined by identifying the material that has the closest correlation to the point defined by the measured back / center ratio and the magnitude of the backscatter signal, as compared to the second set of correlations. . This second correlation helps to provide a proper correlation in the region where the first correlation by the backward-forward signal ratio does not make a good discrimination. In this way, the overall discriminating ability of the method is improved.

The foregoing and other objects, features and advantages of the present invention will become more apparent from the following description of some preferred embodiments when considered in conjunction with the accompanying drawings.

[0020]

DETAILED DESCRIPTION OF THE INVENTION The present invention will now be described in more detail with reference to the accompanying drawings. These drawings show preferred embodiments of the invention. However,
The present invention can be embodied in many different ways,
It should not be construed as limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. Like numbers refer to like elements throughout the specification.

Referring to FIGS. 1 and 2, a device 20 according to a preferred embodiment of the present invention is shown. The apparatus 20 includes a wafer transfer device 22 adapted to hold a substrate or wafer W to be inspected. For the purposes of explaining the present invention, it is assumed that the wafer W is held such that the surface S to be inspected is horizontal, but it is understood that this wafer need not be oriented horizontally. See. Device 20 has a strong P
It comprises a light source 24 capable of producing a narrow beam of light with a type-polarized element. The light source 24 is preferably a laser that emits a light beam in the visible spectrum. Shorter wavelengths tend to generate larger particle scattering signals, thus contributing to improved measurement accuracy. On the other hand, longer wavelengths tend to increase the maximum particle diameter with which the material can be identified, which
The range of particle diameters for which the method and apparatus are useful is expanded.
If desired, the beam of radiation can pass through a beam expander 26, which is then directed to one or more mirrors such as mirror 28 and / or one such as lens 30.
It is redirected by one or more lenses to impinge on the surface S of the wafer at an oblique angle of incidence θi measured from a vertical plane passing through an area of the surface S illuminated by the incident beam. The incident angle θi is, as described further below,
To emphasize the difference in scattered intensity for different locations in the hemispherical space above the surface S, for example about 45 ° -80 °.
It is preferable that the angle is considerably large.

As shown schematically in FIG. 3, the incident beam is incident on the surface S at an incident angle θi, from which it is reflected like a mirror at a reflection angle θr, and in many directions due to defects on the surface. Scattered. The plane of incidence is defined by the plane perpendicular to the incident beam, shown as the Z-axis.
When the incident beam strikes a particle on the substrate, the scattered light is scattered from the incident plane, as shown by scattered light θs. The polar angle that the scattered light makes with the plane of incidence is defined as φs.

According to the invention, the scattered light is collected and analyzed by a plurality of collectors in order to determine the material from which the particles detected on the surface S are made. Theoretical calculations have been performed using scattering models to predict scattering patterns for many different particle sizes and materials.
The scattering model used is based on a separate illuminant scheme and has been experimentally validated as described in the published literature.

In FIG. 4, four different particles (each of which is a dielectric material, as described above), which are arranged on a silicon substrate and are incident with an incident beam of P-polarized light having a wavelength of 488 nm and an incident angle of 65 degrees. , Particles from the groups of semiconductors, gray metals, and good conductors) are presented for the results of theoretical calculations performed by the scattering model. The following table shows the optical constants of four different materials along with the particle diameters corresponding to FIG.

[0025]

[Table 1]

The diameter of the particles is chosen so that the brightness of the scattered light in the rear region of the space above the surface of the substrate (ie, the scattering angle θs is less than about −45 °) is about the same for all four particles. Was done. It can be clearly seen that the brightness of scattered light in the forward region of space (ie, positive scattering angles above about 20 °) is significantly different for different particles. For example, the brightness of the light scattered by the silicon particles is significantly greater than the brightness of the light scattered by the PSL particles, even if the silicon particles are quite small. Therefore, if the particle sizing system is calibrated with PSL particles based on the intensity of scattered light in the back region, the silicon particle diameter will be severely overestimated.

According to the invention, the material from which the particles are made is
The influence of the material on the scattered brightness of the light is automatically taken into account and determined so that a more accurate evaluation of the particle size can be made. To this end, referring to FIGS. 1 and 2, a plurality of scattered light collectors 40 are placed at various locations with respect to the illuminated area of the substrate surface. The rear collector 42 is arranged in the rear region of the space on the surface S in order to collect the light scattered in the rear direction (that is, the direction of the negative scattering angle θs). As shown in FIG. 2, the rear collector 42 can be offset from the plane of incidence at a polar angle φs if desired. The central collector 44 is arranged in the central region of the space on the surface S closest to the vertical plane. The front collector 46 is located in the front region of the space closest to the specularly reflected beam. These collectors 42, 44, 46 collect the scattered light, and collect the collected light into corresponding detectors 42a,
Formed by lenses and / or mirrors that focus on 44a, 46a. These detectors produce a signal indicative of the intensity of the scattered light that is collected and focused on the detector.

The collectors 42, 44 and 46 have a scattering angle θs.
Of the circle, and preferably has a substantially circular shape. However, other collector shapes can be used. The lenses or mirrors of the collector efficiently integrate the collected scattered light so that the magnitude of the signal generated by the corresponding detector indicates the integral of the light scattered over the area of the collector.

In FIG. 4, there is a large depression in each intensity distribution near the position of the vertical plane and a pair of "shoulders" are defined in the intensity distribution in the front and rear positions with respect to the vertical plane. Please note that. The extent and position of the depression and the size and position of the shoulder are
Also note that it is different for different particles. The present invention utilizes the shifting of the intensity distribution of this scatter to identify the particulate material. The dips in the intensity distribution tend to be created when using a P-polarized light source as in FIG. 4, or when the light source has a strong P-polarized element. Light sources dominated by S-polarized light tend to create smaller dips or no dips in the intensity distribution, which may not work well. However, in some applications it may produce acceptable results.

In FIG. 4, the intensity of the scattered light is shown only in the plane of incidence, but in reality the light is scattered over the entire reflecting hemisphere. Running a theoretical model, the scattered light received over the area corresponding to the area of each collector 42, 44, 46 is integrated to produce the backward, center,
And the full collector signal to the front collector can be provided. FIG. 5 shows the integrated total rear collector signal over a range of different particle diameters for the particle material of FIG. Thus, FIG. 5 can be used to determine the particle diameter from the aft collector signal when the material is known. Note that the three non-PSL materials are tightly lumped when compared to the response of the PSL material. PSL materials have similar index values to dielectrics that may be found on wafers such as SiO ₂ and Al ₂ O ₃ . Thus, if the low index dielectric can be separated from the non-dielectric material, it means that a tight fit of the non-PSL material can make significant improvements in particle sizing.

In FIG. 6, the results of the model for the same particles as in FIG. 5 show the ratio of the magnitude of the backward collector signal to the magnitude of the backward collector signal with respect to the magnitude of the backward collector signal (hereinafter referred to as " Rear / front ratio "). It can be seen that these curves for various materials are clearly distinct from each other over at least a major range of magnitudes of the rear collector signal.
Thus, this back / front ratio includes material identification parameters that can allow the material of the particles to be identified. Based on these results, the following methods can be used to determine particle size and material.

An incident beam is incident on the surface S of the substrate and collectors 42, 44, 46 collect light scattered by all particles on the illuminated area of the substrate, and
The collected light is focused on the detectors 42a, 44a, 46a. These detectors generate a signal indicative of the intensity of the scattered light. The rear / front ratio is calculated, such as by the CPU 48 of the processor 50 connected to the detector. Backward /
The magnitude of the front ratio and the rear collector signal is compared with the correlation between the rear / front ratio of the particulate material and the magnitude of the rear collector signal shown in FIG. 6 by the comparator 52 of the processor 50 or the like. The correlation between the particle material rear / front ratio and the rear collector signal magnitude is determined by processor 50.
Are preferably stored in the data storage unit 54 of, and are accessible to the comparator 52. The measured rear collector signal magnitude and the calculated rear / forward ratio define the points on the graph of FIG.
If the scatter is due to one of the four materials, this point approaches one of the curves. This determines the material of the particles. The particle size is then determined by using the curve of FIG. This curve of FIG. 5 is consistent with the identified material and determines the particle diameter from the magnitude of the aft collector signal.

The apparatus 20 comprises a scanning system for moving the wafer W and the incident beam relative to each other,
The incident beam is preferably scanned across the surface S in order to detect and identify particles originating anywhere on the surface. The scanning system can include a carrier 22 for translating and / or rotating the wafer W, and a beam scanning system including a rotating mirror (not shown) that linearly scans the incident beam across the wafer surface. Can also be provided.

Once the particle material is known, the magnitude of the aft collector signal usually indicates a generally reliable particle diameter. However, since the scattered light intensity generally increases with increasing particle diameter at most places around the hemisphere, the particle diameter can be scaled based on the scattered light intensity measured at other locations. It will be appreciated that this can be inferred. For example, the magnitude of the front collector signal can be used instead of the magnitude of the rear collector signal. The size of the front collector signal is about 100 in diameter.
It may be particularly useful for determining the size of particles larger than nm, at least for the particular collector geometry of FIG. However, if the forward collector 46 is located near the specularly reflected beam, scattered light due to the effects of the specularly reflected beam and the surface roughness of the surface S will not be detected by the detector. As such, care must be taken in the design of the front collector. Light scattered by surface roughness tends to closely surround the specularly reflected beam. Thus, as shown in FIG. 2, the collector 46 is shaped so that the small circular area that contains and surrounds the specularly reflected beam is not collected or prevented from reaching the detector 46a. Is preferably a ring. The half-angle α of this undetected region is at least about 5
It is preferable that the angle is °. In other words, the inner circumference of the annular collection area of the front collector 46 preferably has a diameter of at least about 10 ° for a specularly reflected beam passing through the center of the undetected circular area. However, it will be appreciated that other collector geometrical arrangements and positions may be used for this front collector. For example, a generally circular front collector placed symmetrically across the entrance plane,
It may be spaced on one side of the mirror-like reflected beam between the vertical surface and the reflected beam.

The front collector 46 has a diameter of approximately 20 °.
It is preferable to have a substantially circular outer circumference of 50 °.
Each center collector 44 and rear collector 4
2 is substantially circular and preferably has a diameter of about 20 ° to 45 °.

Note that there are some potential problems with FIG. The PSL and tungsten lines are close together at the small grain end, and the silicon line loops back across the other line. Locations where all overlapping points or lines are very close to each other indicate locations where the material is not reliably identified. This condition is improved by using a second material identification parameter based on the signal from the central collector 44. More specifically, FIG. 7 shows the calculated results of the model for the four particle material of FIG. In this figure, the ratio of the magnitude of the rear collector signal to the magnitude of the center collector signal ("rear / center ratio") is plotted against the magnitude of the rear collector signal. Although there are still points of intersection, the PSL and tungsten particles are further apart at the smaller diameter end and the silicon is improved at the larger diameter end. Therefore, by using both a first material identification parameter based on the magnitudes of the rear and front collector signals and a second material identification parameter based on the magnitudes of the rear and center collector signals. The overall material identification capability of the can be improved. It is to be understood that this invention is not limited to the particular embodiments described above, and that variations and other embodiments are intended to be within the scope of the appended claims. Although specific terms have been used herein, they have been used in a generic and descriptive sense only and not for purposes of limitation.

[0037]

Many variations and other embodiments of the invention, which have the advantages disclosed in the foregoing description and the associated drawings, will suggest themselves to those skilled in the art to which this invention belongs. Let's do it. For example, although the present invention has been described in the context of testing silicon wafers, the method and apparatus of the present invention may also be used to test various other non-silicon substrates including computer disks, flat panel displays, and others. Equally suitable. Further, while it is preferred to perform the method of the present invention by calculating the ratio of the magnitude of the collector signal to the front to the rear, and to the center to the rear, the method of the present invention is not limited to the reciprocal of the described ratio, (desired. It will be appreciated that other cases can be performed in other cases by using other material identification parameters, such as the difference between the collector signals (normalized by one of the signals) and other parameters. Moreover, it is possible to identify the material of the particles without explicitly calculating any material identification parameters.
For example, the predetermined correlation of the particle material with the signal magnitudes of various collectors can be stored in a data storage device 54 as a multidimensional array.
Which can be stored in
It can be accessed and rewritten or manipulated by the comparator 52 to identify the material of the particles having the stored scatter pattern that most closely matches the measured scatter pattern. Moreover, although the preferred geometry for the collector has been described, other types and geometries of collectors can be used to detect the scattering pattern produced by the particles.

[Brief description of drawings]

1 is a schematic side view of an apparatus according to a preferred embodiment of the present invention.

FIG. 2 is a schematic plan view of the device of FIG.

FIG. 3 is a schematic perspective view showing a substrate on which an incident beam is incident and light reflected and scattered from the substrate.

FIG. 4 shows theoretically calculated scattered light intensities for angular positions of particles of four diameters and different materials on a silicon substrate when the particles are illuminated by a P-polarized beam of oblique incidence. It is a graph.

5 is a graph showing theoretically calculated results plotted for the four materials shown in FIG. 4 as the average particle diameter versus the magnitude of the brightness of the scattered light at the back collector position.

FIG. 6 shows four materials shown in FIG. 4 and FIG.
6 is a graph showing theoretically calculated results plotted as a ratio of aft-forward detector signal magnitude to aft detector signal magnitude.

FIG. 7: Theoretical calculated results plotted as ratio of aft-center detector signal magnitude to aft detector signal magnitude for the four materials shown in FIGS. 4-6. It is a graph which shows.

[Explanation of symbols]

20 devices 22 Wafer carrier 24 light sources 26 Beam Expander 28 mirror 30 lenses 42 Rear collector 42a detector 44 center collector 44a detector 46 Front collector 46a detector 50 Processor θi incident angle W Wafer

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Sonpin Gao 01772, Southborough, Ted Lane, Massachusetts, USA 16 (72) Inventor Michael E. Fossey, California 91367, Woodland Hills, Victory Boulevard 22216 , Inventor C320 (72) Inventor Lee Dante Clementi 29710, South Carolina, United States, Carroll Cove, Lake Willie 311 (56) Bibliography JOURNAL OF THE IN STITUTION OF ENVIRON MENTAL SCIENCES (Thomars, Thor Larmers) et.al.) "Towerd Classic. ATION OF PARTICLE PROPERTIES USING LIGHT SCATTERING TECHNIQUES "MAY / JUNE 1997, PP15-21 OPTICS COMMUNICA CUTS IONS ART FILTER GONZ ▲ a ▼ lez et al. lts. shadowing effects "Vol. 137, 1997, pp359-366 (58) Fields investigated (Int.Cl. ⁷ , DB name) H01L 21/66 F Term System (JPO) JOIS

Claims (9)

(57) [Claims] 1. A method of inspecting the surface of a substrate to identify the material from which the particles on the substrate are made, wherein all particles present on the surface portion cause light to be in a hemispherical space above said surface. Directing an incident beam of light onto the surface portion at an oblique angle of incidence so as to scatter in a plurality of directions of; collecting light scattered in each of a plurality of regions of the space; Measuring the intensity of the collected light for each region to produce a set of scatter data representing a sample of the generated scatter pattern; and two data measurements of the intensity of the light collected from each of the regions.
Calculating a material identification parameter by dividing the data by applying the material identification parameter to a plurality of predetermined scatter patterns corresponding to a plurality of materials to obtain the same state as the measured scatter data. Determining the material of the particles by identifying the material of the scatter pattern that is closest to them, wherein an incident beam of strongly P-polarized light is incident on the substrate portion. 2. The method of claim 1, wherein the measuring step measures the intensity of light collected in the anterior region of the space, the forward scatter signal having a magnitude indicative of the intensity. And using a front detector to generate a backscatter signal for measuring the intensity of light collected in a rear region of the space to generate a backscatter signal having a magnitude indicative of the intensity. Using a dedicated detector, and determining the material of the particles such that the magnitudes of the forward and backscatter signals match the predetermined scattering pattern most closely with the measured signal. Comparing the material to a predetermined scatter pattern defined by the magnitudes of the forward and backscatter signals to identify the material to be processed. 3. The method of claim 2, wherein a magnitude of one of the forward and backscatter signals corresponds to the material identified in the material determining step and the forward and backscatter signals. The method further comprising the step of determining a particle size by comparing to a predetermined correlation relating to the diameter of the particle to said one size. 4. The method of claim 2, wherein the numerator correlated with the magnitude of the backscattered signal is divided by the denominator correlated with the magnitude of the forward scattered signal. Calculating a first material identification parameter; identifying the particulate material based on a correlation with the first material identification parameter of the particulate material and one magnitude of the backscatter signal and the forward scatter signal And a method further comprising: 5. The method of claim 4, wherein the step of calculating the first material identification parameter comprises dividing the magnitude of the backscatter signal by the magnitude of the forward scatter signal. Determining the particulate material, including the step of calculating a posterior to anterior ratio,
Comparing the back-to-front ratio and the backscatter signal magnitude to a plurality of scatter patterns defined by the back-to-front ratio to the backscatter signal magnitude for a plurality of materials; and A method comprising identifying the material whose scatter pattern is closest to the measured back-to-front ratio and magnitude of the backscatter signal. 6. The method of claim 2, wherein light collected in the central region of the space closest to the vertical plane passing through the illuminated region of the surface is collected and the collected light is collected. The step of using a central detector to generate a central scatter signal having a magnitude indicative of the intensity by measuring the intensity of the, the magnitude of the forward, central, and backscatter signals, the scatter pattern of which was measured. Comparing to a predetermined scatter pattern defined by the magnitudes of the forward, center, and backscatter signals to identify the material that most closely matches the signal. 7. The method of claim 2, wherein the step of collecting the forward scattered light includes scattering the light scattered over a region spaced to the side of the reflected beam or an annular region surrounding the reflected beam. A method comprising: using a wide-angle collector to collect. 8. The method of claim 2, wherein the step of collecting the backscattered light comprises 45 ° from a vertical plane passing through a portion of the plane illuminated by the incident beam.
A method comprising collecting light on a substantially circular area that is concentrated 75 ° apart. 9. The method of claim 2, wherein the step of collecting the backscattered light includes the incident beam and its
Collecting light on an area offset from the plane of incidence formed by the vertical plane.