JPH01224692A - Foreign object detection method and device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention provides a foreign matter detection method suitable for highly sensitive detection of minute foreign matter on a sample surface in an intermediate process of manufacturing semiconductor LSI Quano 1 and TPT. Regarding equipment.

[Conventional technology]

2. Description of the Related Art As semiconductor devices become more highly integrated and patterns become increasingly finer, the line width of circuit patterns has become approximately 11×n or less. In order to manufacture such semiconductor devices with high yield, it is important to manage the differences.

Conventional foreign particle inspection equipment on wafers uses (1) a combination of one-dimensional high-speed scanning of a laser beam and low-speed translational movement of the sample, or (1ii) high-speed rotation and low-speed translational movement of the sample as shown in JP-A-55-135551. In combination with this, line scanning was used to scan and detect the entire surface of the sample.

Furthermore, in Japanese Patent Application Laid-open No. 57-80546, a scanning equivalent to the above (1) is realized by combining the electric scanning of a self-scanning f-dimensional photoelectric conversion probe array and the low-speed movement of the sample.

In addition, automatic microcircuits and
Wafer Inspection Electronics Test, Volume 4, Issue 5, (Automatic Microcirc
cuit and wafer inspection
ElectronicsTeat, VoL 4.
NCL 5) In May 1981, in pB60-74, a self-scanner-dimensional photoelectric conversion element array was placed at a radial position to the sample wafer, and this was combined with rotational movement of the sample to perform scanning equivalent to the above (11). has been realized.

However, with the above method, it is not possible to avoid "missing" the foreign object when the dead zone existing in the adjacent area of each solid-state imaging photoelectric conversion element is completely occupied by the foreign object, or the cautious signal from the foreign object is degraded. It ends up.

However, when the size of the foreign object to be detected is sufficiently large compared to the dead zone width of the solid-state imaging photoelectric variable element, and when the dead zone width is negligibly small compared to the pixel width of the solid-state imaging photoelectric variable element, However, the above 6 "overlooked" is not a big problem.

In the above NrE method, the possibility of "missing one point due to a dead zone from such a continuation point" is ignored and not discussed.

[The problem that the invention aims to address]

- As LSIs become more highly integrated with IM and 4 Mbits, the presence of foreign matter of 1iim size or smaller has a large effect on product yield in the manufacturing process. Detection sensitivity is required.

However, it is difficult to distinguish this from a shiver turn due to a foreign object being missed in the dead zone adjacent to the solid-state d&image photoelectric conversion element pixel or due to a decrease in the detection signal.

SUMMARY OF THE INVENTION An object of the present invention is to provide a foreign matter inspection method and apparatus that can distinguish minute foreign matter from patterns and detect foreign matter with high sensitivity.

[Means for solving the problem] The size of the light receiving area of each pixel is 20 x 20 μmm.
Use a solid-state imaging photoelectric conversion element having a plurality of t- or less (calculated on the sample surface) to cover the dead zone existing in the adjacent part of the solid-state imaging photoelectric conversion element pixel.

Forty-one solid-state imaging photoelectric conversion elements are installed overlappingly, and foreign objects are detected through processing of the output output of each solid-state imaging element.

For work] 1) By reducing the entire detection area (imaging work) of the solid-state imaging photoelectric conversion element, the entire pattern detection area can be divided, so the scattered light detection signal from the pattern within one pixel can be reduced.

2) By installing a plurality of solid-state imaging light beams so as to cover the area adjacent to the solid-state imaging photoelectric conversion confectionery pixel, it is possible to prevent foreign objects from being missed and the detection signal to deteriorate. This makes it possible to detect foreign substances with high sensitivity.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows the main steps in TFT fabrication applied to one embodiment of the present invention.

This is a conductor (chrome [gate] 2
, insulator (Si-N) 5 and amorphous silicon (a-
8i) The circuit pattern t-sequentially is formed by 4.

When the sample is a thin film transistor substrate 5, the configuration is a combination of circuit patterns in the X; The spatial filter includes a light shielding section 7. It is formed by the transparent part 8.

FIG. 2 shows the pattern of this example and the distribution of scattered light from foreign objects.

The same figure (C1 shows the distribution of diffracted light on the Fourier transform surface F of the objective lens 11 when the pattern linear part 9a is irradiated with laser light. There is a narrow diffracted light in the center of the Fourier transform surface F. The width W of this diffracted light is determined by the width of the laser beam spot on the sample in the X direction.

Figures (d) to (f) show the state of the diffracted light on the 7-layer conversion surface F when the pattern corner 9b is irradiated with a laser beam, and the width W of the diffracted light is wider than the turn straight portion 94. Become.

Does this mean that the diffracted light of the laser beam at pattern corner 9bvc is directional? I need it because I want to lose it.

Figures (1) to (L) show cases in which the pattern 9C is irradiated with a laser, but the diffracted light is not detected in this case because it does not enter the objective lens 11.

The figures (C to (-')) show the case where the foreign object 10 is irradiated with a laser beam, and the diffraction source spreads over the entire Fourier transform surface F.

This is because the scattered light from the foreign object 10 has no directionality.

As described above, the scattered light from the shield turn corner 9b is completely blocked by the spatial filter 6 having the wAW of the light blocking portion 7. Also, since pattern 9C does not enter the detection system, foreign matter 1
It is sufficient to consider only the pattern corner 9b that affects the detection of zero.

According to the thin film transistor manufacturing process, the Shiba turn 9 is of type (
The scattered light distribution is different because the shape is different from #family). Therefore, light-shielding spatial filters 6aa, 6ba, and 6C that are optimal for each pattern process and pattern shape are required.

Next, FIG. 5 shows the difference in the scattered light distribution depending on the pattern shape and the shape of the light shielding part 7 of the light shielding spatial filter 6.

The same figure (43, (b)% (d) shows the diffracted light distribution (tb) when the roundness of the corner of the pattern corner 9b is small as shown in (d)). Since the width W of the diffracted light of F is narrow, the light shielding portion 7 can completely shield the light with W, as shown in (8).

However, tea, (f) ~ (h) i (L),
(1) ~ (The pattern corner 9b is like a force (d)
If it is larger than the case shown in , the diffracted light ((fJ, shown in −)) of the 7-Lier transformation plane F will spread out, and accordingly (e
), (%-Wst of shading l1S7 as shown in jJ
-Light cannot be blocked unless it is wide.

Therefore, in the pattern shape shown in (h), (-'1), it is necessary to widen IlgW of the light shielding part 7 of the spatial filter 6.

Next, detection signals by the solid-state imaging photoelectric conversion element will be explained with reference to FIGS. 4 to 7.

FIG. 4 (aJ is a solid-state imaging photoelectric conversion element 12A with a foreign object 1)
0 is detected by setting each pixel size H in the X direction on a sample, and FIG.

The detection signal obtained from each pixel (L, j, A, t, m) of the solid-state imaging photoelectric conversion element 12A at the position S is shown, and (c) of the same figure shows the detection signal obtained by binarizing this detection signal using the threshold value of the VTR. FIG.

FIG. 5 (-) shows that the solid-state imaging photoelectric conversion element 12B is exposed to a foreign object 1.
] Q, R, S by pixel dimensions H in the X direction from a position shifted in the X direction by 172 of the pixel dimension H
6t- is shown, and in the figure [bJ is each pixel of the solid-state imaging photoelectric conversion element 12B (Lv Jl#'B
The detection signal obtained from e 11 mm, ) is shown, and the same figure (
CJ is a diagram showing a binarized signal binarized using the threshold value of this detection signal ftvTH.

FIG. 6(a) shows the solid-state imaging photoelectric conversion element 12Ct--from a position shifted in one direction by 172 of the pixel dimension H on the sample having foreign matter, by the pixel dimension H in one direction.

Indicates the state of scanning and detecting the position of R"e 3t",
In the same figure (bJ is each image X (↓2
*JHm Age],.

-), and in the same figure (CJ is a diagram showing a binarized signal obtained by binarizing this detection signal with the threshold value of v'rit.

FIG. 7(a) shows a solid-state imaging photoelectric conversion element 120.12B.
is shifted by 1/2 of the pixel size H in the -X and 1 directions, and a specimen containing 10 tons of foreign matter is detected by being shifted in the X direction? The figure (b) shows the trough pixels ('se 1st
ABm lBp In6-”4e J4m '4
a 14a In4) A color 4b detector number is shown, and (C) of the same figure shows the detection signals obtained by the solid-state imaging photoelectric conversion elements 12 and 12g, which are each binarized using a threshold value VTII and subjected to zero processing. FIG. 3 is a diagram showing a value signal.

In addition, in FIGS. 4 to 7, in order to simplify the explanation, the number of pixels is 5 ("m J e 'a Ia m" J@-f
fj@・L,・・・−・↓,・・m,・export・m4
).

In the detection in the solid-state imaging photoelectric conversion element 128 12a12C, each pixel (↓...m a Ll...ml
L,...m,) and the respective threshold values Vr
When m is used as a binary signal, pixels j and A (or A and j,
L and j.

I will miss the small difference *1 degree between L and j, which is α5~1 in the gap 13.

Therefore, at the magnified imaging position of the sample surface as shown in FIG. In particular, the pixels of the solid-state imaging photoelectric conversion element 121), and j, or! , the insensitivity f between and -
The small foreign object 10 located at 13 is the solid-state imaging photoelectric conversion element 12E.
This detection signal is binarized using the threshold value of VTH, and the solid-state imaging photoelectric conversion element 1
By ORing and processing the 2x12g signals, the binary signal becomes "1" even with a small foreign object 10, and an improvement in sensitivity of 9 is obtained.

Next, FIG. 8 shows a signal processing method of the solid-state fj & image-to-electrical conversion element.

Pixels of solid-state imaging photoelectric conversion element 12c, 12E...m
,.

The output of each town is converted into a digital signal (IT number by the NΦ converter 14) and binarized by the binarization circuit 15 according to the VTR instruction of the threshold value to produce a binarized signal (1"). is stored in the foreign object memory 16 and guided to wr17 08 times. This is then displayed on the display monitor 18. By this method, the width detection sensitivity can be improved.

Next, the relationship between the size of the detection pixel of the detector and the pattern size will be explained using FIGS. 9 and 10.

- A multilayer pattern as shown in FIG. 9(b) is a pattern 90
The difference in level is too large and pattern 9 is multi-layered, so the same figure (a
), the pixel U of the detector is placed in a field of 601m"l/), receives the shadows *' of multiple (two) pattern corners 9b, and
The detection output of U increases. So, lI! If the pixel size is set to 2 and 1, as in the case of the Li nest 2, the influence of the pattern corner 9b can be made smaller compared to the VM element U.

The pattern on the thin film transistor sample used in the example has a width of 2
If the multi-pixel dimension is set to about 0 to 60 μm or less, the pattern corner 9b of 1 cad in one pixel will not be included.

FIG. 10 (Tsukasa is 1
Since only the scattered light from one pattern corner 9b enters the pixel, the composition output of the pattern corner 9b is smaller than the detection output of the small difference '@10, as shown in FIG.

Since the ridges are set to 20" per pixel, it is possible to distinguish between Meji Kotenmon IQ and Pattern 9 within one pixel of pattern corner 9b.

- If the sample is a thin film transistor sample, the pattern 15 is a combination of circuit patterns in the
I can't tell it apart from the different viJ10.

However, if the image size is increased to 60 squares as shown in (1) of the same figure, multiple pattern corners 9b are included in one pixel, so the detection output from the pattern corner 9b is The detection output of the small foreign object 10 becomes larger.

In this way, when the TPT sample surface is scanned with large pixels, scattered light from a plurality of pattern corners 9b is detected, so that the foreign object detection output is buried in the scattered light. Therefore, it is necessary to reduce the number of pixels to 20 points or less.

Next, an example in which the present invention is applied to a device for detecting foreign matter on a thin film transistor substrate will be described with reference to FIGS. 11 and 12.

The foreign matter inspection device of this embodiment includes an optical illumination section 19, a TPT
Substrate 5t - x (y) stage 21 for scanning,
(22) a θ stage 23 that rotates in the θ direction; a stage drive unit 24 consisting of an automatic focusing mechanism (not shown); A processing circuit 29 that processes the detection signal detected by the detection unit 28;
It is made up of many things.

Further, the constituent elements will be explained.

After setting and positioning the thin film transistor substrate 5 on the upper surface of the θ stage 25, the stage driving unit 24 fixes the thin film transistor substrate 5 by vacuum suction. Encoders 30α, 30 of X stage 21 and X stage 22
At step b, coordinate position information is sequentially input to the CPU 31.

The irradiation unit 27 focuses the semiconductor laser 25 through a lens 26 and irradiates the surface of the thin film transistor substrate 5 with the focused light. In order to inspect the entire surface of the thin film transistor substrate 5, X-Y quantification is performed. At this time, if the height of the surface of the thin film transistor substrate 5 changes.

Since the illumination position by the semiconductor laser 25 changes and the foreign object detection performance deteriorates, an automatic focusing mechanism (not shown (included in the stage drive section 24)) is required.

For this purpose, a projection stripe pattern contrast extraction method such as that disclosed in No. 58-70540 is used.

In the detection s28, the scattered light reflected from the thorny surface of the thin-layer transistor substrate 5 is sent to the objective lens 32, and the semi-transparent mirror 3
After branching by 3, a Wi image is formed on the detector 12m 12E. Here, solid-state imaging photoelectric conversion confectionery 121112Et-
use

In addition, the detection 928 includes scattered light kjJ1 of pattern 9ρ).
A spatial filter 6''#6b*6C is provided at the beginning.

Detector 12mu 12i! The detection signal of J is processed by the processing circuit 29. The output from Yusaku Tani in the detection 912 episode 12E is converted into a digital code by the AID converter 14, and in the 211f conversion, it is binarized at 1615 and converted into a binary signal (
11') is stored in the foreign object memory 16 and sent to the ORIIg circuit 1.
7 and displayed pixel by pixel on the display monitor 18.

The coordinate position is determined by encoder 5otL and ρCPU by 50b.
The threshold level according to the coordinate position input to 61 is 110C.
It is set within the PU51.

Furthermore, the present invention is not limited to thin film transistors, but can also be applied to the calibration of other products such as wafers, photomasks, and reticles.

〔Effect of the invention〕

As explained above, according to the present invention, it is possible to realize an automatic foreign matter measuring device that can detect minute foreign matter with high sensitivity and stability.

[Brief explanation of the drawing]

Figure 1 is a diagram showing the main process of TF'T lining that is applicable to one embodiment of this A-mei. Figure 2 is a diagram showing the distribution of scattered light from the pattern and foreign matter. Figure 3 is the scattered light due to the pattern shape. A diagram showing the difference in the distribution and the shape of the spatial filter light shielding part, Figures 4 to 7 are diagrams showing the detection method and detection signal using solid-state imaging light, and Figure 8 is a diagram showing the solid-state imaging photoelectric conversion element. FIG. 9 and FIG. 10 are diagrams showing the relationship between the pattern in the detection pixels of the detector and a foreign object. FIG. 11 is an overview of a foreign object inspection device according to an embodiment of the present invention. FIG. 12 is a schematic diagram showing a block diagram of a foreign matter inspection apparatus according to an embodiment of the present invention. Explanation of symbols 5...Thin film transistor, 6...Spatial filter, 9...-pattern, 10...Foreign object, 11.1 objective lens, 12...Detector, 13...Dead zone, 14...
・NΦ converter, 15... Binarization circuit, 16...
Foreign object memory, 17...-circuit, 18...Display monitor, 19...Optical illumination Ij 11 section, 21...X stage, 22-Y stage, 23...θ stage, 24...
- Stage drive section, 27--Irradiation section, 2B--Detection section, 29--Processing circuit, 31-CPU. 1st Prisoner ■ oO I (Cn (1, (i)
0-Tour & Mitsushige /θ-・Juyao 4I21 60th 6th Kousuke 70 弗B 目/2D Ginμ Ko 1 (α) (b) 10th prisoner

Claims (1)

[Claims] 1. A plurality of solid-state imaging photoelectric conversion elements that convert scattered light from the sample generated by an illumination means into electrical signals are arranged at enlarged image positions on the sample surface, and each of the solid-state imaging photoelectric conversion elements A method for detecting foreign matter, characterized by detecting foreign matter on the surface of a sample using an output signal. 2. The above-mentioned plurality of solid-state imaging photoelectric conversion elements are arranged in an overlapping manner in the relative scanning direction between the solid-state imaging photoelectric conversion element and the sample to avoid overlooking foreign objects due to dead zones existing in adjacent areas of individual detection areas. 2. The method of detecting foreign matter according to claim 1, wherein: 3. An illumination means for illuminating the sample with light, a condensing optical system for condensing the scattered light from the sample illuminated by the illumination means, and a condensing optical system for receiving the light condensed by the condensing optical system to generate electricity. A foreign object detection device comprising: a solid-state imaging photoelectric conversion element having a plurality of pixels to be converted into signals; and a detection means for detecting a foreign object based on the signals obtained from the pixels of the solid-state imaging photoelectric conversion element. 4. The size of the above pixel is converted to 20 x 20 on the sample surface.
4. The foreign object detection device according to claim 3, wherein the particle diameter is .mu.m^3 or less. 5. The foreign object detection device according to claim 3, wherein the condensing optical system is provided with a spatial filter that blocks scattered light in a predetermined direction. 6. The foreign object detection device according to claim 5, wherein a plurality of types of spatial filters are provided and configured to be switchable. 7. The foreign object detection apparatus according to claim 3, wherein the plurality of solid-state imaging photoelectric conversion elements are arranged in an overlapping manner in a relative scanning direction between the solid-state imaging photoelectric conversion element and the sample.