JPS61104243A - Method and apparatus for detecting foreign matter - Google Patents

Abstract

PURPOSE:To detect fine foreign matter stably with high sensitivity, by detecting the ratio of scattering light signals based on illumination having change in scattering effect. CONSTITUTION:A polarization beam splitter 150 has a characteristic for reflecting a P-polarized beam component and permitting the passage of an S-polarized beam component. The output ratio VP/VS of the output VP of a photoelectric converter 7L and the output VS from a photoelectric converter 7H is operated by an analogue ratio comparing circuit 100 and binarized on the basis of a threshold value (m) in a binarization circuit 101 to enable the detection of small foreign matter. If a solid image pick-up array is used in place of the photoelectric converters 7H, 7L, the enhancement of detection sensitivity is attained. In this case, a plurality of analogue comparing circuits 100 and binarization circuits 101 are used to simultaneously perform analogue comparison in parallel.

Description

Detailed Description of the Invention [Field of Application of the Invention] The present invention relates to semiconductor LSI wafers or masks, particularly LSI
The present invention relates to a detection method and apparatus suitable for foreign particle inspection that detects microscopic foreign particles on patterned wafers, etc. at high speed and with high sensitivity during intermediate manufacturing steps.

[Background of the invention]

Conventional foreign particle inspection equipment on wafers uses (I) a combination of one-dimensional high-speed scanning of the laser beam and low-speed translational movement of the sample;
) The entire surface of the sample was scanned and detected using linear scanning, which is a combination of high-speed rotation and low-speed translation of the sample.

Furthermore, in Japanese Patent Application Laid-Open No. 57-80546 (Public Patent Application 1), scanning equivalent to the above (I) is realized by combining electrical scanning of a self-scanning type dimensional photoelectric conversion element array and low-speed movement of the sample.

Furthermore, A, D, Gara: Automatic Mic
rocircuit and Wafer In5pec
tion, Electronics Te5t, Vol.
, 4°No, 5. May 1981. pp, 6O-70
In (Known Example 2), a self-scanning dimensional photoelectric conversion element array is arranged at a radial position of the sample wafer, and this is combined with rotational movement of the sample to realize scanning equivalent to the above (n).

However, in the methods of Known Examples 1 and 2, it is impossible to avoid "missing one foreign object" when scanning a foreign object due to the dead zone existing in the adjacent portion of each photoelectric conversion element picture element.This is strictly avoided. In order to achieve this, it is necessary to install multiple photoelectric element arrays overlappingly to cover the dead zone.This increases the amount of signal processing circuitry more than necessary and causes a decrease in reliability.However, , when the size of the foreign object to be detected is sufficiently large compared to the above dead zone width even if the photoelectric element arrays are not overlapped, or when the total dead zone width is small enough to be ignored compared to the total width of the photoelectric conversion element picture elements. In this case, the above-mentioned ``overlook'' is not a big problem. From this point of view, in the methods of Known Examples 1 and 2, "1 miss" due to the dead zone is ignored.

[Detection of foreign matter on patterned wafer]

Inspection of foreign substances on patterned wafers during the intermediate process of LSI manufacturing is essential for improving product yield and reliability. The automation of this work is disclosed in Japanese Patent Application Laid-Open No. 55-149829.
JP-A-54-101390, JP-A-55-94145,
This is realized by a detection method using polarized light, as shown in a series of patents such as Japanese Patent Application Laid-Open No. 56-30630. This principle will be explained using FIGS. 43 to 50.

As shown in FIG. 43, if only the illumination light 4 is radiated at an inclination angle φ to the surface of the wafer 1, reflected light and scattered light 5°6 will be generated from the pattern 2 and the foreign matter 3 at the same time. It is not possible to distinguish and detect only the foreign object 3. Therefore, a device was devised to detect the foreign matter 3 by using polarized laser light as the illumination light 4.

As shown in FIG. 44(a), a pattern 2 existing on a wafer 1 is irradiated with S-polarized laser light 4. (Here, when the electric vector 10 of the laser beam 4 is parallel to the wafer surface, it is called S-polarized laser illumination.) In general, the surface irregularities of the pattern 20 are microscopically small enough compared to the wavelength of the illumination light, and optically Since it is clear, the reflected light 5 also has an S polarization component 11.
is maintained. Therefore, if the analyzer 13 for blocking S-polarized light is installed in the optical path of the reflected light 5, the reflected light 5 will be blocked and will not reach the photoelectric conversion element 7. On the other hand, as shown in FIG. 44(b), the scattered light 6 from the foreign object 3 includes a P-polarized light component 12 in addition to the S-polarized light component 11.

This is because the surface of the foreign object 3 is rough and the polarization is canceled.
This is because the polarized light component 12 is generated. Therefore, analyzer 1
If the photoelectric conversion element 7 detects the P-polarized light component 14 passing through the foreign object 3, the foreign object 3 can be detected.

Here, as shown in FIG. 43, when the angle between the pattern 2 and the longitudinal direction of the laser beam 4 is at right angles, the reflected light 5 is completely blocked by the analyzer 13. , if this angle is different from the right angle, the light will not be completely blocked. This discussion is published in the Proceedings of the Society of Instrument and Control Engineers, Mol.
17, No, 2. p232 4-p242
．． 1981. This is because only the reflected light from the pattern whose angle is within ±30' from the right angle enters the objective lens installed above the wafer. Although the light is not blocked, its intensity is small enough to be distinguished from the 2-3 μm foreign object scattered light, so it does not pose a practical problem.

Here, the inclination angle φ of the polarized laser beam 4 is set to about 1'' to !. This is for the reason shown below. In the Eiken test shown in FIG.
The intensity Vs of the component 140 of the foreign body scattered light passing through the analyzer 13 and the intensity evp of the component passing the analyzer of the pattern reflected light 5 were measured using an objective L/7 lens 9 (magnification 40X, N, A=0.55). The experimental results are shown in FIG.

This plotted the foreign matter/pattern discrimination ratio VS/V with the laser inclination angle φ as the horizontal axis. From the same figure, the inclination angle φ is! ;' Since Vs can be easily distinguished from Vp in the following cases, stable foreign matter detection becomes possible. Further, considering design matters, φ=1° to 36 is optimal. (Refer to Japanese Unexamined Patent Publication No. 56-30630) The reason why two laser light sources 15 are used from the left and right is to enable stable detection of foreign objects that generate anisotropic scattered light. Next, a foreign substance inspection method using this detection principle will be explained with reference to FIGS. 47 to 50.

As shown in FIG. 47(a), a slit 8 is provided on the sample imaging surface to limit the detection range. This allows slit 8
Since only the scattered light within the projected area 8a of the aperture onto the sample is detected at a time, the foreign object scattered light P component 14d is smaller than the integrated intensity 14F of the pattern reflected light P component within this area. If it is large enough, foreign matter 3 can be detected stably. Therefore, this area 8a is the size of the foreign object to be detected.
If the size is about the same as 3 μm), the detection sensitivity will be optimal, but the number of scans will be increased as shown in FIG. 50(b), and the inspection time will be long. Conversely, if the aperture area 8a is increased, inspection can be performed in a shorter time, but the detection sensitivity will deteriorate. - Taking this into account, currently the area 8a is set to IOX 200 μm, and foreign particles of 2 to 3 μm are inspected in about 2 minutes (150 (in the case of mφ wafer)). This situation is shown in Figures 48 and 49. I will explain using

First, FIG. 48 shows a plan view (a) and a cross-sectional view (b) of the wafer surface. The pattern 2 has (1) a slight depression in the pattern and (II) a part having an angle other than perpendicular to the irradiation direction of the laser beam 4, and a small amount of scattered light P component 14p is generated from each of these parts. .

On the other hand, small foreign matter 3a with a size of about 0.5 to 2 μm and 2μ
From the large foreign matter 3b with a size of m or more, the above (I) and (II)
A P component 14d having a larger intensity than each of the locations is generated.

FIG. 49 shows the signal output of the photoelectric conversion element 7 when the aperture 8a scans over the sample. In the same figure (a), P component 1
The distribution of 4p (circuit pattern) and 14d (foreign matter) on the sample is shown. When the aperture 8a scans this distribution, the same figure (b)
) is obtained, and when it is binarized, the defect signal shown in (C) of the same figure is obtained. In this example, small foreign object 3
Since the output from edge a and pattern 2 are the same,
Since the threshold indicated by the broken line must be set at a higher position than this output, as a result, detection is limited to only large foreign objects.

However, in the manufacture of highly integrated LSIs such as the 256 Kbit memory I'lI, the presence of foreign matter of 1 μm in size greatly affects the product yield, so a detection sensitivity of 1 μm of foreign matter is required.

If the aperture 8a is limited to 5×5 μm2 or less in the apparatus shown in FIG. 47, the scattered light P of (1) and (It) can be
Since the cumulative effect of the components is reduced compared to the case where the aperture 8a is 10 x 200 μm, it is possible to detect a foreign object of 1 μm. However, in this case, the inspection time is about 40 minutes.
This makes it difficult to synchronize with the manufacturing throughput, which poses a problem in practical application.

[Purpose of the invention]

An object of the present invention is to provide a method for distinguishing minute foreign objects from patterns and inspecting them at high speed.

[Summary of the invention]

In the present invention, in addition to polarized laser illumination that has a large scattering effect on foreign objects, dual illumination is performed with illumination l that has a small scattering effect. Focusing on the fact that the scattered light from the former illumination is likely to be generated by foreign objects, and the scattered light by the latter illumination is likely to be generated by patterns, by detecting the ratio of the two scattered light signals, it is possible to detect minute foreign objects more stably and with high sensitivity. The reason is that it can be detected.

Furthermore, in the present invention, the size of the light receiving portion of each of the two wheels is 5×5μ.
By using multiple photoelectric conversion solid-state imaging devices with a size of less than m (calculated on the sample surface) and simultaneously performing parallel comparison processing on the output signals from each device, foreign particles can be detected with high sensitivity without deteriorating high speed. The purpose is to conduct an inspection.

[Embodiments of the invention]

Embodiments of the present invention will be described in detail with reference to FIGS. 1 to 13.

FIG. 1 shows how a solid-state image sensor array 20 is used in place of the slit 8 of the conventional example shown in FIG. 50. FIG. 2 illustrates an example of the solid-state image sensor array 20. The light receiving section 20a is a silicon photodiode or a GaAsP photodiode, and among these, a PIN junction type photodiode has characteristics of high speed response and high sensitivity, and is most suitable for the use of the present invention. The width of each light receiving part 20a (pixel) is 500μ
m, and there is a dead zone with a width of 50 μm between adjacent pixels. When the number of pixels is 40, for example, if the total magnification of the detection system is 100x (objective lens 90x magnification and relay lens (not shown) magnification 2.5x), one pixel The size is 5 x 5 μm on the sample surface, and in the end, a range of 5 x 220 μm is scanned while being detected (this results in an inspection speed comparable to that of the conventional method).

The effects of this solid-state image sensor array 20 will be explained with reference to FIG. For comparison, the conventional example shown in Fig. 50 is shown on the right (d) of the same figure.
+e) and (f). Figure left (a), (b)
l, (C); For simplicity, the number of pixels is set to 5 (" +
/ ,' + 1+ -). (a) shows the state in which the solid-state image sensor array 20 scans, and (b)
is the video signal (SLl, Si1, 8
Al, 81-1. Sm1) and (C) C M value VTH
Convert the signal (Sr1.S72.SA2゜S, E
2. S? FL2) is shown. If the output signal 3A of the pixel 1 is binarized using the threshold value VTH, the binarized signal SA2 will be the small foreign object 3.
Even a becomes '1'', and sensitivity is improved compared to the conventional method.

FIG. 4 shows a signal processing method for each pixel of the solid-state image sensor array 20. The output of each pixel L-% is simultaneously binarized in parallel by the binarization circuit 21, and a binarized signal (11
“) is led to the OR circuit 22, and if a foreign object is detected in at least one of the two wheels, the output of the OR circuit becomes 11” and is input to the foreign object memory 23. With this method, the outputs of 40 pixels are simultaneously processed in parallel, and inspection speed and detection sensitivity can be significantly improved compared to when a self-scanning image sensor is used.

However, the dead zone 204 of the solid-state image sensor array 20
gives rise to the disadvantages explained below. This solution is the fifth
It is shown in FIGS.

As shown in FIGS. 5 and 7, the solid-state image sensor array 20
When the arrangement direction and the scanning direction are perpendicular, and the relationship between the dead zone 204 between pixels and j and the small foreign object 3c is as shown in the figure, the small foreign object 3c will be overlooked. (In the case of Figure 5 (
Scanned from a) to (b). ) Therefore, as shown in FIGS. 6 and 8, if the arrangement direction of the solid-state image sensor array 20 and the scanning direction are set at an appropriate angle (for example, 45 degrees), the above-mentioned oversight can be avoided.

(In the case of FIG. 6, scanning is performed from (a) to (b) and from (b) to (C).) This angle does not necessarily need to be 4 da when the shape of both wheels 20a is rectangular.

In FIG. 8, the small foreign object 3C is detected redundantly by pixels 1 and A, resulting in double counting. However,
As a way to avoid this double counting, JP-A-56-1
32549, JP-A-56-118187, JP-A-57-
66345, JP-A-56-126747, JP-A-56-
118647 may be used.

FIG. 9 shows an example of application of the invention in the case of spiral scanning.

FIG. 10 shows the overall configuration of the embodiment. The wafer 1 is attracted to the wafer chuck 40 by the vacuum chamber 41 and
The stage 46 and the Y stage 49 move in the X and Y directions. Foreign object information detected by the solid-state image sensor array 20 passes through a binarization circuit 21 and an OR circuit 22 to a control circuit 32 including a foreign object memory 23, and is displayed on a display device 33.

In the present invention, the pixel size is set to about 5×5 μm or less, so if a defocus occurs during calibration due to undulations on the wafer surface, the foreign matter detection sensitivity will be significantly reduced. Therefore, it is essential to use a configuration in which the automatic focus detection section 30 detects the amount of defocus during the inspection and feeds it back to the driver 31 of the focusing mechanism mortar 43. The principle of this automatic focus function was announced on pages 223 to 224 of the 22nd 8ICB Academic Conference Preprint Collection, and
70540, but as shown in Fig. 11~
This principle will be explained using FIG. 13. This method is ideal for the present invention because it is characterized by stable automatic focusing without being affected by the pattern on the sample.

FIG. 11 shows the main parts of the automatic focus detection section 3o.

Striped pattern Striped pattern on the crow board 60a, 604
are respectively projected onto the sample by the objective lens 9, but the respective focal point positions are set slightly above and slightly below the focal point of the imaging element array 2o. The images of the respective striped patterns 60a and 604 on the sample are magnified by the objective lens 9, reflected by the semi-transmissive mirrors 34 and 62, and imaged onto the image pickup device 61.

FIG. 12(a) shows a projection pattern that is imaged on the image sensor 61 when the wafer is too far down (Z(0)).
FIG. 2(d) shows the video signal waveform detected by the image sensor 61 in the case shown in FIG. 12(a). Figure ″12 (
b) shows the projection pattern imaged on the image sensor 61 in the case of the focused position (2=0), and FIG.
2 shows a video signal waveform detected by the image sensor 61 in the case shown in FIG. 2(b). FIG. 12(e) shows the projected fringe pattern imaged on the image sensor 61 when the wafer is too high (Zoo), and FIG. 12(f) shows the image sensor in the case shown in FIG. 12(C). 61 shows a video signal waveform detected at 61.

Therefore, the detection signal of the image sensor 61 is the same as that of the image sensor array 20.
When is the focused point, the stripe patterns 60a and 604 are equal in location, so the difference signal between the two becomes zero.

On the other hand, in the case of too much rise (or too much fall), the deviation from the in-focus point of the image sensor 610 corresponds to the magnitude of the output of the difference signal, so that the servo signal shown in FIG. 13 is obtained.

The figure shows an example of actual measurement of the difference signal when the sample surface is an aluminum surface and when the sample surface has a complex pattern (memory cell surface). This allows focusing within ±0.5 μm, so
When the objective lens has a magnification of 90x and a magnification of 40x, stable foreign object detection is possible. The autofocus mechanism has a simple configuration using a motor 43, a slope 45, a ball 44, and a leaf spring 42, as shown in FIG. 10, for example.

Next, explanations corresponding to the scope of claims of the present invention will be given in Figures 14 to 2.
This will be explained using FIG.

First, the basic principle will be described using FIGS. 14 and 15. In the method of FIG. 47, the analyzer 13 is removed,
When the polarizing beam splitter 150 is installed, FIG. 14 is obtained. Here, the slits 8H98L detect the same point on the sample. Since the polarizing beam splitter 150 has a characteristic of reflecting the P-polarized ash and passing the S-polarized component, the output Vp of the photoelectric conversion element 7L is as shown in FIG. 47 (FIG. 49).
is the same as This is shown in Figures 15(a) and 1(b). On the other hand, the output Vs from the photoelectric conversion element 7H after being scanned as shown in FIG. 15(C) becomes as shown in FIG. 15(d). (a).

Comparing (b) with (C) and (d), (a),
In (b), the output of the foreign object is higher than that of the pattern, and in (C) and (d), the output of the pattern is higher than that of the pattern.
00 (as shown in FIG. 15(e)), the binarization circuit 1
It can be seen that when 01 is binarized using the threshold value m (as shown in FIG. 15(f)), it becomes possible to detect the small foreign matter 3a (which could not be detected by the binarization in FIG. 49).

Furthermore, if the solid-state image sensor arrays 20H and 20L described above are used in place of the photoelectric conversion elements 7H and 7L, detection sensitivity can be improved. In this case, it is necessary to use a plurality of analog comparison circuits 100 and binarization circuits 101 to simultaneously perform analog comparison in parallel. (See Figure 16).

The above illumination/detection method is called (I[) type illumination.

Next, type (I) illumination will be explained using FIGS. 17 and 18. For example, as shown in FIG. 17, by using the output characteristics of the foreign matter and the pattern due to the inclination angle φ shown in FIG. Illuminating the same sample point with wavelength λ15H, color separation branching prism and analyzer 151H, 151
In this method, only the P component in L is detected and compared using arrays 20H and 20L.

The outputs of the arrays 20H and 20L and the binarization method are shown in FIG.

FIG. 18(a) shows a case in which a laser beam is irradiated from an obliquely lower side of the SL where foreign objects 3a and 3b are present, for example. Figure (b) shows the output signal VL in that case, and Figure (b) shows the output signal VL in that case.
C) shows the binary signal. FIG. 18(d) shows, for example, S
The case where the laser beam is irradiated onto L from diagonally above is shown. FIG. 4(e) shows the output signal VH in that case.

L FIG. 18(f) shows the signal waveform of τ7, and FIG. 18(g) shows its binary signal.

The above illumination and analysis conditions are modeled in Figure 30 (q). (I) type illumination (in C, the use is not limited to the above-mentioned S-polarized illumination lights 15L and 15H;
(h), various illumination lights shown in FIGS. 31(a) to (dl)
It is also possible to use conditions for light analysis. Among these, the illumination/analysis conditions (L) that emphasize foreign matter include either (a) analysis of P-polarized ash with S-polarized illumination, or (o) analysis of P-polarized components with P-polarized illumination. The following conditions are used. The reason for this will be explained in detail later.

On the other hand, the illumination-analysis condition H that emphasizes the pattern is
Polarized light does not necessarily need to be used, as any cases other than (a) and (0) above may be used. (In other words, incoherent light such as an ordinary halogen lamp may be used, and this is shown by the symbol 8+P in FIGS. 31(a) to 31(d).) - ・1
Providing a dichroic prism (or mirror, -) as described in 49829 and JP-A-56-43539,
A light branching prism (semi-transmissive mirror) and a color filter or an interference filter may be used in combination.

In addition, the illumination lights 15H and 15L are He-Ne laser (
λ=6.328A), GaAlAs L/-the diode (λ=7.800~8.300A), and InGaAsP
L/-the diode (select two different types from λ=n,
For example, since the lens system 56L narrows the focus at the sample surface, the ill light is removed and the detection becomes more stable.

As mentioned above, in type (1) illumination, foreign object highlighting illumination (L)
The pattern emphasizing illumination (H) and the foreign object emphasizing illumination (L) are of different wavelengths (λ1.
λ2) is an essential condition.

Here, instead of color separation, if a method is used that performs time-division detection and compares the detection outputs in (L) and (H), as shown in Japanese Patent Laid-Open No. 57-66345, illumination and analysis can be performed using information. Parts become easier. In this case the illumination is one pole and the array 20
However, the analyzer must include an optical element such as a Bockel cell that has a high-speed analysis characteristic switching function.

Next, the present invention will be explained in detail with reference to FIGS. 19 to 28.

FIG. 19 shows details of the signal processing circuit shown in FIG. 16.

This is the same method as in Figure 4 Hatake.

Figures 20 to 22 are similar to WJ10 and Figure 20, but the differences are the illumination light 15H, lens system 15HJ, light branching prism 1501, color filters 151H, 151L.
.

This is because an array 20H and an analog comparison circuit 100 are added.

Modify the analog comparison method using Figures 23 to 28 FC
Explain in detail.

23 and 24 show the results of an experiment using the conditions shown in FIG. 17. ! At t&, regarding the output 14p of pattern 2, pattern outputs VL and VH were measured while rotating pattern 2 by one angle from a right angle to the direction of projection of illumination light 4H (4L) onto the wafer surface. On the other hand, foreign matter is 0.7
, 1.2 μm standard particles (in this case, rotation is not necessary), and VL and VH were measured. The measured values are shown in FIG.

From this, it can be seen that at any angle of the pattern, the output ratio (white circle) ML/VH from the pattern is smaller than m (the reciprocal of the slope of the broken line in the figure).

On the other hand, the output ratio VL/VH of foreign matter indicated by a black circle is larger than m.

A method of discriminating between the foreign matter and the pattern using an electric circuit in consideration of the output characteristics of the foreign matter and the pattern shown in FIG. 24 will be explained with reference to FIGS. 25 to 28.

FIGS. 25 and 26 show an example using an analog division circuit. The output ratio VL/Vl (is calculated by the analog divider circuit 100, and when VL/VH>m, the binarization circuit tot K
The awn is output. It means cc-c.

This means that in the output ratio calculation, when vH is small, the calculation error becomes large and the calculation result becomes unstable (for example, when VH is zero, VL/VH=■). As a way to avoid this, the binarization circuit 104
Only in the case of vTH (vTH is the value of Vt corresponding to a foreign object of about 0.511 m) (11"), the calculation result of VL/VH may be set as valid '1". This is realized by the binarization circuit 104 and the AND circuit 103.

FIG. 27 and FIG. 28 show an example using the analog subtraction circuit 105. In this case, it is important to adjust the intensity of the H or L illumination light and the gain of the output amplifier (not shown) of the array 20H or 2OL so that "= 1 (45 degree inclination).Analog. The same goes for decreasing.

Instead of the above-mentioned analog division/subtraction, the output may be A/D converted and calculations made with digital values. FIG. 29 shows an (l[I) type illumination. In this case, the illumination lights 15H and 15L each have a wavelength of λ1. λ2, and uses, for example, a light branching prism 150 and color filters 151H and 151L as a color separation optical system. Here, the semi-transparent mirror 15C of the illumination system is used to combine the illumination lights 15H and 15L.

Above, the illumination of (Il, (I[), (I[)W)
Regarding the light analysis conditions, it is essential to use the conditions K) and (0) above as the condition (L) for highlighting foreign substances. On the other hand, the condition (H) for emphasizing the pattern is shown in Fig. 30(a)~
Various methods can be considered as shown in FIGS. 31(h) to 31(a) to 31(d). Table 1 below shows (1) (If) (I [[) type conditions (a) to (h) in Figure 30 to (a
) to (d) are applicable.

Table 1 O applicable Next, consider the condition (L) that emphasizes foreign matter.

1. Analysis of pattern reflected light In this chapter, we will analyze the polarization characteristics of pattern reflected light. M4
", I will explain the symbols necessary for analysis.Here, the components of the vector are the XYZ coordinate components. Also, the normal to the pattern 20, the incident light 4, the reflected light 5, the light passing through the objective lens 90, and the polarization of the incident light. Each of the 10 vectors is normalized.

Pattern 20 normal vector:! N = (α, β, γ) (I
N+ = 1') (1) Pattern 20 incident light 4 vectors: A = [λ, μ, S(+
A+ = 1) (2) 2 reflected light 5 vector 2B = [ρ, δ, τ] (IB
l=1) (3) Light passes through vs. $#7 and 9/l/: B'= (0
, 0, 1:] (lB'1=1) (4) Polarization component 10 of the incident person: 'E = (Ex, By, Ez) (I
El=1) (5) Polarization component of reflected light B: R= (R
x, Ry, Z) (6) Polarized light component 14 of the light passing through the objective lens: Z = [Melon, Rshi, R'Z)
(7).

Angle between the incident light beam and the normal line: φ (8) Angle between the normal line N and the XY plane 6, (9) Angle between the longitudinal direction of the pattern and the X axis: η (lO) Reflected light beams B and Z Angle with the axis: ψ (11) X of reflected light B
Angle between the projected component on the Y plane and the X axis: θ
(12) Angle between the analyzer's analysis axis and the ,do.

FIG. 32 shows the relationship between incident light A and reflected light B with respect to the normal vector N of the pattern edge. Here, the following equation is obtained using the inner product of vectors.

σ””B = −n (14
)=-(OA+AC) (15)=-
(-A+ 2 (A-Nl)Nl) (1g)8
Mu-2 (A 11N) N (17)
Here, the angle φ between the incident light A and the normal N is obtained from the following equation.

MueN= IAI ・INII cos(π-φ)
(18)=-■φ (19)
Substituting equations (11, (2) into equation (17), vector β
The following equation is obtained as the component of According to Fresnel's law of reflection, as shown in Figure 33, the polarization component (electric field vector)
The polarization components IEp parallel to the IEs are each reflected by the BS
, (ktp.

Here, the coefficients S(φ) and p(φ) are the refractive index n of the pattern
It is given by the following formula using .

FIG. 34 shows an example of calculating the values of S(φ) and p(φ) in the case of n=1.5. FIG. 5(a) shows the relationship between the incident angle φ and the intensity, and FIG. 6(b) shows the relationship between the incident angle φ and the amplitude.

Here, as shown in (b), S(φ) becomes negative in the range of φ=(f to 9σ, but p(φ) changes from positive to negative after φ=φ. This Rp== The incident angle φ that becomes O is pola
It is called the rization angle. Figure 33 (
al l (A comparison is shown between the case where φ is smaller than φ and the case where φ>φ.Here, note that the relationship between Ep and Rp is the same as the process that led to equation (17), and the incident The polarization component E of the light and the polarization component distribution of the reflected light satisfy the following relational expression.First, the polarization component E of the incident light is decomposed into an S component and a p component.

Concave === Es+112p(211) = (Esx, BS)'+Esz) + (Epx+Ep
y+Epz ) (29) Next, the polarization component B of the reflected light is decomposed into an S component and a p component.

B=fusion+lRp (34)
= (Rsx, Sy, R3z) + (Rx, R,,y
, Rpz) (35)=-p(φ)(α+ in Epx-2
Epy-2 /, Epz 2Kr) (38) Here = Epxα + F'pyβ + Epzγ (=lEp・N
) (39) Substituting equations (36) and (38) into equation (35), the following equation is obtained.

Note: In general, the pattern profile exhibits a smooth shape as shown in FIG. Therefore, the modulus @N of a pattern having an angle η formed between the Y axis and its longitudinal direction changes along the profile of the pattern. Calculate the reflected light β for the incident light A along the line and obtain Figure 36. From this, 1.)
For example, as shown in FIG. 4(a), N @ A = 0.
When the objective lens 9 of 55 is placed above the pattern 2,
Among these reflected lights B, a line segment R within the entrance pupil of the objective lens
Only the reflected light that reaches Q enters the objective lens. Here, when the locus of the line segment RQ is determined while increasing the angle η, it can be seen that when the angle tf v >3σ, the reflected light no longer enters the objective lens, as shown in FIG. 3(b).

The reflected light B incident on the objective lens enters the objective lens and is refracted in parallel to the Z-axis direction as shown in FIGS. 37(a) and 37(b). At this time, the polarization fraction changes due to refraction, which will be analyzed below. Here, the refracted light is assumed to be B', and its polarized light component is assumed to be R'. First, the reflected polarized light component is decomposed into a radiation component Kr and a tangential component Rθ with respect to the optical axis Z.

R= IFLr + B6
(43) The projection components Rr and Rθ of the polarization component T and Rθ on the XY plane can be obtained from the following equation using coordinate transformation.

Here, the polarization component Rθ does not change due to refraction, but the component B
, is the component 几 rotated by an angle ψ with θ as the rotation axis? becomes. Here, the angle ψ is the angle between the reflected light B and the two axes.

h'r=C-ψ■θRr, II! cfsInθRr,
0) (4B) From equations (37) and (40), the polarization component R' of the refracted light B' can be obtained from the following equation.

ba; omitted + pottery = [plum, plum, 0) (49
)=[Formula ψ(2)θ'f%r−amθRθ,wcfsi
na Rr+0θRθ, O) (50) &=(1w
! cψ02θ−5n2θ)Rx+xsθsinθ(−11) % (st)Yuzu=resistant−θ(s!Icψ−1)R
x+(−10 2θ)R, (52) where, polarization component B
It is confirmed below that the sizes of and 8' do not change.

Reflected light B has an angle θ.

The following components are obtained using η.

113 = (-xθ x blood ψ, - θ x manifestation ψ, coaψ)
(sa) Since the reflected light B and its polarized light component B are direct, the following equation is obtained.

IB @ 1R = (2) θ−ψ uniform θ苅ψ Yuutoge■ψBonz=
The magnitude of the O(54) polarization component (2) can be obtained using equation (54) as follows.

1d=Rei+YujuR: (5!i) On the other hand, the magnitude of the polarization component I·' is obtained from the following equation using equations (51) and (52).

+i+2=42+4 (58) ((
! ec? Jθ 10 ships H duty – θ (formula ψ – 1) ■) + (c
rsasmθ (ki ψ−1) Rx+(−f”” 10”)
Ry)2=(CIIecfOθ+stnθ+2sec−
7 fcm θ θ) + (2) 2 θ 20 (→-2 ψ + 1
)) Number ((2-h god 4θ ten straw ψ 5 in 2θ) + Q) 5
2asin2θ(→-2secψ+1 month RF,)-2(
cosθsinθ(formula・ψa)82θ+5in2θ)(
9ecψ−1)+xsθ−(−−1)(secpsin
2θ+”2a))RxRy (
so)=(sL112θ(sin!θ10)l”θ)−
) - Jψtsu (Jθ God 2θ)) Tsuzuriju (Uke (5ixF a God 2θ) + 4Jθ (Jθ God 2θ))
)θ(secψ−1)(gecψ+1) )tJLY
(61) = (sin”θ10 seconds”f cos
”θ)g+(.′θ+Jψ工′θ) Ry−)−sin
2f) tarFψRxRy
(62) It was confirmed that equation (58) and equation (62) are the same, and the magnitudes of a and R' do not change. Furthermore, it is confirmed by the following equation that the polarization component IFLr and the thickness of the signal do not change.

l IFLrl = K, + 'R,, (63) =
(,RrcoIIθ)2)-(RrsLrIθ)21-
R2 (64) = ccs”θ
(Y θ melts θR a) 2-) -tangaa θ to 1 θ%
> 2 = (1 + difficult ψ) (c - 1 - θ 凰 + 萅 θ ~) 2 = Jψ (i
~ 100 θRy)2(67) = l [Rr12(68) Reflection JIB from a pattern having an arbitrary angle θ
The magnitude of the polarized light component when the light passes through the objective lens was expressed using the magnitude of the polarized light component of the incident light.

(3) Calculation when the incident light is horizontal In this section, we will calculate the polarization component when the incident light A coincides with the Y axis. This is the case where the tilt angle is 00, and the specularly reflected light from the wafer does not enter the objective lens, so it is suitable for detecting foreign objects. That is, the incident light is expressed by the following equation.

A = (0, -1, O) (69) Formula (
17) For the reflected light B, the following equation is obtained using the component to the normal vector.

B=[2αβ, 2β2−1, 2βγ) (7
0) From equation (54), the angles ψ and θ are expressed as follows using the normal vector fraction.

Below, the polarization vector of the incident light is ■l113=(1,0
Calculate separately for the case of ゜0] and the case of ■IB=(0, 0, 1). In this book, the former is called 8-polarized illumination, and the latter is called P-polarized illumination.

■ In the case of S-polarized illumination (E = (1, 0, 0)) Equation (30
) to (33), calculate the 311p component for the pattern of the polarization component E. (FIG. 38) The polarization component N of the reflected light is expressed by the following equation using equations (40) to (42).

The magnitude of the polarization component H is determined by the following equation using equation (58).

1B12= (S(φ))2γ2//(α2+r)+(
p(φ))2αsi(αhalflo) (78)■ In the case of P-polarized illumination (""LO+0+1:l) Similarly, the following equation is obtained.

Ihl = (s(φ))2α2/(αmuro+(bori)
2γ2/(α2+γ2) (84) ■ Comparison of S-polarized light and P-polarized illumination In the cases of (1) and (2) above, the polarization components of the dot on the spherical surface are compared as follows. This means that if we consider cases in which the analysis axis of the analyzer placed on the objective lens is parallel to the X-axis and perpendicular to the X-axis, the polarized light components passing through the analyzer will be 4 in each case. do. At this time, the intensity of the passing light is 42 in each case.

In Fig. 39, the refractive index of the pattern n = Z, 45 (S+0
The intensity distribution of passing light is shown in the case of a refractive index of 2). In the figure, the objective lens entrance pupil is indicated by a broken line. In either case, since the incident light coincides with the Y-axis, the distribution is symmetrical with respect to the Y-axis. In addition, the polarization component of the incident light (electric field vector V)
When E and the analysis axis are perpendicular to each other (i.e., when the S-polarized illumination is used and the P-polarized illumination is used), the value becomes zero on the Y-axis. Figure (a),
(b) shows the intensity distribution y2.4s of passing light under S-polarized illumination ([lE!=(1,0,0)). Also, the same figure (C)
, (d) shows the case of P-polarized illumination (β-(0,0,1)). Comparing the former case with the latter case, it can be seen that the ratio of the maximum transmitted light intensity within the entrance pupil is about 1=10, and the pattern reflected light intensity is the lowest in the former case. Furthermore, when the analysis axis is perpendicular to the Y-axis with P-polarized illumination, the reflected light intensity KRy becomes zero in the arc-shaped portion within the entrance pupil, which is interesting.

FIG. 40 shows the calculation results for the field of refractive index n w4.2 (refractive index of Po1y-8i). Figure (a) l (b
) indicates the case where E = (1°0.0], and (C) in the same figure,
Gd) shows the case where E=(0°0.1).Compared to FIG. 39, the reflected light intensity is higher due to the difference in the refractive index, but the tendency of the distribution is the same.

From the above, the intensity of the pattern reflected light passing through the objective lens is S
It can be seen that the value is lowest when the analysis axis is perpendicular to X@ with polarized illumination. On the other hand, since the scattered light from the foreign object under these conditions is depolarized, it is not blocked by the analyzer, and as a result, it is expected that the discrimination ratio will be the highest. Next, we will examine this through experiments.

2. Experimental Study An outline of the No. 411 NK experimental method is presented. 5mW H for lighting
3 on the wafer using an e-Ne laser, a 25× beam expander, and an fθ lens (f = 300*x).
A laser spot of 0 μmφ − was formed. The irradiation angle is from horizontal! And so. The objective lens has a magnification of 40X (NA
-0,55), IJ Silanes 2.5x was used, and a Carnicon (2/311) image pickup tube was used as the detector. An analyzer (polarizing plate) was arranged above the objective lens so as to be rotatable about the optical axis. The wafer sample used had both Poly-8i and 5i02 patterns, and 0.4 μm thick Poly-8i and 5i patterns were formed on the surface.
The intensity N of reflected light from the 02 pattern edge was measured. On the other hand, polystyrene standard particles (n
= 1.59) using diameters of 0.7 μmφ and 1 μmφ,
The scattered light intensity S was measured.

The output of the pattern and foreign matter rotates the wafer and the rotation angle η
N was measured in the range of θ° to 18C.

Figure 42 shows the measurement results. In the measurement, S-polarized illumination (a)
・In each case of (b) and P-polarized illumination (C1・(d),
The analysis axis was parallel and perpendicular to the Y axis. In the figure, the vertical axis shows the signal output S-N, and the horizontal axis shows the pattern rotation angle η. 2% to avoid saturation of the G' A transmission ND filter was used.

From the figure, the pattern reflected light from SiO2 becomes zero as the rotation angle η increases as calculated in Fig. 36, but Po1y-8
The reflected light from i is generated up to around η=9σ. As shown in the 8EM observation image shown in Figure 41, this is true for Po1y-8i.
This is because minute irregularities exist on the pattern edges. In either case, the pattern reflected light is shown in Figure 39 (Figure 40).
However, the foreign object scattered light is not completely depolarized and has the same tendency as the pattern. However, when comparing the maximum output N of the 1 μm foreign object scattered light output and the poly-8i pattern reflected light, it was found that a high S/N ratio can be obtained in cases (b) and (d). When considering actual foreign matter inspection, the inspection time can be reduced by using a detector with high-speed detection such as an optical 1, a child multiplier, or a PIN-8i photodiode array rather than a TV camera. It becomes possible. Taking this into consideration, when comparing cases (b) and (d), case (d) has a 10 times larger output, allowing for stable inspection against electrical noise and ambient light. I understand that.

From the above, when the analysis axis is perpendicular to XsK with P-polarized illumination (P-polarized light component detection), the most stable and highly sensitive inspection is possible, and it is also possible to detect foreign objects of 1 μm or less. Also, when the analysis axis is perpendicular to X@ with S-polarized illumination (P-polarized component detection), a high 87N ratio can be obtained for the pattern if the illumination light is sufficiently strong.

Furthermore, the present invention is not limited to wafers as described above, but can also be applied to the inspection of other products such as photomasks and reticles.

Also, the limit on the size of both wheels is loXloμms degrees,
It has been confirmed through experiments that there is no practical problem in detecting foreign matter of 1.5 μm to 2 μm.

〔Effect of the invention〕

As explained above, according to the present invention, it is possible to detect minute foreign objects present on a target object with high sensitivity and stability while maintaining high speed of foreign object detection.

[Brief explanation of drawings]

Fig. 1 is a configuration diagram showing an embodiment of the foreign object detection method of the present invention, Fig. 2 is a perspective view showing details of the solid-state image sensor shown in Fig. 1, and Fig. 3 is a comparison between the present invention and a conventional example. FIG. 4 is a diagram showing the signal processing circuit of the solid-state image sensor shown in FIG. 1, FIG. 5 is a diagram showing the positional relationship between the dead zone and a foreign object, and FIG. A diagram showing the positional relationship between the dead zone and a foreign object, FIG. 7 is a diagram showing the relative scanning direction of the solid-state image sensor with respect to the wafer in FIG. 5, and FIG. FIG. 9 is a diagram showing the relative spiral scanning of the solid-state image sensor with respect to the wafer. FIG. 10 is a configuration diagram showing the embodiment shown in FIG. 1 in more detail. Fig. 11 is a perspective view showing the automatic focus detection section shown in Fig. 10, and Fig. 12 is a perspective view for explaining automatic focus detection. Figure 14 is a perspective view showing the basic principle of the present invention, and Figure 15 is a diagram showing the relationship with out-of-focus.
Fig. 4 shows the state in which the wafer is scanned and the output signals obtained in the apparatus shown in Fig. 14, and Fig. 16 shows the state in which the wafer is scanned and the output signals obtained in the apparatus shown in Fig. 14. FIG. 17 is a perspective view showing the case of (I) type illumination, and FIG. 18 is a perspective view showing the case of (I) type illumination.
19 is a diagram showing the analog comparison circuit in detail, and FIG.
A perspective view showing a further embodiment of the embodiment shown in FIG.
1 is a front view of FIG. 20, FIG. 22 is a diagram showing the signal processing circuit not shown in FIG. 21 as a block diagram, and FIG. 23 is a diagram showing the circuit pattern and the state of reflected light from foreign objects. Fig. 24 is a diagram showing the experimental relationship data between the outputs VH and VL obtained based on this, Fig. 25 is a diagram showing the results of VH/VL, and Fig. 26 is a diagram showing the VH/VL shown in Fig. 25. Figure 27 shows an analog divider circuit for V
A diagram showing the results of L-VH, Figure 28 is the V shown in Figure 27.
FIG. 41 is a perspective view of the experimental equipment, and FIG. 42 is a diagram showing an analog subtraction circuit for realizing r, -VH.
A graph of experimental data showing the N ratio, FIG. 43 is a cross-sectional view of the wafer, FIG. 44 is a diagram showing the state of reflection from the circuit pattern on the wafer and foreign matter with respect to the irradiated laser beam, and FIG. 45 is the conventional FIG. 46 is a schematic perspective view showing a first example of the foreign object detection method; FIG. 46 is a graph showing measurement data of the output ratio Vs/Vp when the inclination angle φ is changed in FIG.
The figure is a schematic perspective view showing a second example of the conventional foreign object detection method, FIG. 48 is a diagram showing the circuit pattern on the wafer and the state of reflection from the foreign object, and FIG. FIG. 50, which is a diagram showing the relationship between video signals obtained by relatively scanning a wafer, is a schematic perspective view showing a conventional foreign object detection method similar to the second example shown in FIG. 47. Figure 29 is a perspective view showing the (I[) type illumination, Figures 30 and 31 are diagrams showing various combinations of illumination and analysis, and Figure 3
Figure 2 is a perspective view showing the incident light A and reflected light B of the pattern, Figure 33 is a perspective view showing the polarization of the incident light and reflected light, and Figure 34 is an example of calculating the coefficients S(φ) and p(φ). Graph showing, 3rd
5 is a perspective view showing the pattern profile, FIG. 36 is a perspective view showing the direction of pattern reflected light B and the locus of line segment RQ, and FIG. 37 is a perspective view showing the relationship between reflected light B and the objective lens. Figure 38 shows the case of S-polarized illumination (incident light A(0, -
1°0], E(1,0,0)), and FIGS. 39 and 40 are plan views showing polarized intensity distributions (n = 1.0). 45 and n=4.2). [Explanation of symbols] 1... Uniha 2... Pattern 3... Foreign matter 4... Illumination light 5... Reflected light 6.
...Scattered light 7...Photoelectric conversion element 9...Objective lens 10...S polarized light 11...S polarized light component 12.
14...P polarized light component 13...Analyzer 15...Polarized laser light source 20...Solid-state image sensor array for photoelectric conversion 20a...
Light receiving section 20b...Dead zone 21...Binarization circuit 22...OR circuit 23...Foreign object memory 30...
・Automatic focus detection unit 150...Polarizing beam splitter (
or color separation prism, dichroic mirror, or light branching prism) 100...analog comparison circuit 101.104,105...binarization circuit 103...
AND circuit 151H, 151L...color filter (or interference filter, or analyzer)

Claims (1)

[Claims] 1. The foreign matter on the target object is emphasized and detected by the first photoelectric conversion element, the background on the target object is emphasized and detected by the second photoelectric conversion element, and the 1. A foreign object detection method, comprising: emphasizing and detecting a foreign object detection signal obtained from a photoelectric conversion element with a detection signal obtained from a second photoelectric conversion element. 2. An optical device that emphasizes a foreign substance on a target object and detects it with a first photoelectric conversion element, and emphasizes a background on the target object and detects it with a second photoelectric conversion element, and the first photoelectric conversion element. 1. A foreign object detection device comprising: comparison means for emphasizing a foreign object detection signal obtained from the element with a detection signal obtained from a second photoelectric conversion element. 3. The above optical device includes an illumination optical system that illuminates the target object with different wavelengths at different inclination angles, and a system that separates the light reflected from the target object into first and second photoelectric conversion elements. 3. The foreign object detection device according to claim 2, characterized in that it is constructed of a color separation/branching optical element that branches. 4. The foreign object detection device according to claim 3, wherein the color separation/branching optical element is composed of a light branching prism or a semi-transparent mirror and a color filter. 5. The foreign object detection device according to claim 3, wherein the color separation/branching optical element is constituted by a dichroic prism or a dichroic mirror. 6. The foreign object detection device according to claim 4 or 5, wherein the color separation/branching optical element further includes an analyzer. 7. The optical device includes a linearly polarized illumination optical system that illuminates a target object with linearly polarized light having an inclination angle, and a first and second photoelectric conversion system that polarizes and decomposes the light reflected from the target object. 3. The foreign object detection device according to claim 2, characterized in that it is constituted by a polarization splitting/branching optical element that splits into elements. 8. The foreign object detection device according to claim 7, wherein the polarization separation/splitting optical element is constituted by a polarization beam splitter. 9. The foreign object detection device according to claim 7, wherein the polarization separation/branching optical element is composed of a light branching prism or a semi-transmissive mirror and an analyzer. 10. The above optical device includes an illumination optical system that illuminates a target object with different wavelengths at the same tilt angle, and a first and second optical conversion element that separates the light reflected from the target object into colors. 3. The foreign object detection device according to claim 2, characterized in that it is constructed of a color separation/branching optical element that branches. 11. The foreign object detection device according to claim 10, wherein the color separation/branching optical element is composed of a light branching prism or a semi-transparent mirror and a color filter. 12. The foreign object detection device according to claim 10, wherein the color separation/branching optical element is constituted by a dichroic prism or a dichroic mirror. 13. Claim 11 or 1, wherein the color separation/branching optical element further includes an analyzer.
The foreign object detection device according to item 2. 14. The foreign object detection device according to claim 2, wherein the comparing means is equipped with a dividing means. 15. The foreign object detection device according to claim 14, wherein the output of the dividing means is ignored when the signal obtained from the second photoelectric conversion element is near zero. 16. The foreign object detection device according to claim 2, wherein the comparison means is equipped with a subtraction means. 17. The foreign object detection device according to claim 16, wherein the output of the subtraction means is ignored when the signal obtained from the second photoelectric conversion element is near zero.