JPS6231815A - Focus position detecting device - Google Patents

Abstract

PURPOSE:To make equal the influence of the sample surface reflecting ratio, to cancel the effect and to execute the high accuracy by preventing the interference of two detecting systems and detecting the projection pattern at the same place on the sample. CONSTITUTION:A projection pattern mask 3 is one piece, and one pattern projected on a sample 6 surface by image sensors 9 and 11 dislocated and arranged in the optical axis direction is detected commonly. As the result, the interference of two detecting systems is prevented, the sample surface reflecting ratio is also the same to detect the projection pattern of the same place on the sample 6, and as the result, the effect is cancelled. Thus, the point where the contrast of both detecting systems, the focusing target position, is equal, is not changed, nd the focusing position detection can be executed with a high accuracy.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a focus position detection device suitable for an automatic focusing device that automatically focuses on eight surfaces of an LSI.

[Background of the invention]

As a pattern projection type focal position detection device, two projection pattern masks are installed at different positions so that the image plane is shifted in front and behind the focal position of the sample surface observation optical system. An example of a device that projects images onto different locations on the sample surface so that they do not overlap and determines the direction of defocus from the contrast between the two projected patterns is disclosed in Japanese Patent Application No. 1983-
No. 168619.

Figure 8 shows its basic configuration, which includes two pattern masks 5.
g and 5b are shifted back and forth along the optical axis of the illumination optical system, and are arranged side by side on a plane perpendicular to one optical axis, as shown in FIG. The pattern of this mask is Half Mirror 4 (
(FIG. 8) Projected onto the sample 6 via the imaging lens 5.

The pattern projected onto the sample surface is captured by lens 5. half mirror
4 and 7, the image is formed on the light receiving surface of the image sensor 9. The image sensor 9 has photoelectric elements 91 arranged in an array as shown in FIG. The pitches are matched.

An example of the detection signal of the image sensor 9 is shown in FIG.

The left side is the projection pattern of the mask 5z, and the right side is the detection output of the projection pattern of 5h. Here, the contrast Cα, C
b changes to 5 as shown in FIG. 12 depending on the position of the sample in the optical axis direction since the imaging planes of the projection patterns do not coincide. Note that (?Nα, Cptb in 12th @ is the normalized contrast, and the brightness difference Ca, Cb of the detection signal shown in FIG.
is normalized by the average level AαAA. The reason for normalization is to prevent variations in the reflectance of the sample surface from affecting the contrast detection results. Note that this signal processing is performed by a contrast detection circuit 50 shown in FIG. In addition, the sample observation optical system (Fig. 8) K Ogero,
The focal position for the observation image detection image sensor 8 is
A point Zo in FIG. 12 where both normalized contrasts are equal is set. Thereby, the direction of defocus can be determined by comparing both normalized contrasts.

A problem with this device is that since two projected patterns are detected by one image sensor, mutual interference occurs due to reflected light from the inner surface of the incident window glass of the image sensor and the chip surface.

The fifteenth factor is that the image sensor 9 is shown in cross-section 1, in which the light transmitted through the incident window glass 95 is reflected on the surface of the sensor chip 940, and then reflected again on the lower surface of the glass 95, causing the light to pass through the incident window glass 95.
This shows how to irradiate. Note that the reflection on the chip surface 94 is greatly affected by the bulge phenomenon that occurs strongly in the gold-plated areas around the photoelectric element rows, as shown in Figure 14, due to the difference in reflectance due to the pattern on the sample surface. Because of this, 2
This is a case where the brightness of the two projection patterns is significantly different. In other words, even if a small amount of reflected light enters a part with low reflectance from the bright side, there will be an offset vo due to that K.
greatly disrupts the normalization operation described above. As a result of the average level being raised by Vo, the normalized contrast changes to the correct value c, as shown in Figure 15.
sh to CNb1, and the intersection of both contrast characteristic curves moves, causing an error.

Note that when the brightness of both projection patterns is equal, both of them are equal to the offset amount due to the diffused reflection, and the influence thereof is canceled.

Another problem is that because the positions of the masks 5a and 5h are misaligned, the dimensions of the two patterns on the light-receiving surface of the image sensor 9 are different, and the relationship with the photoelectric element is no longer the same, resulting in contrast detection. Differences occur in characteristics.

Details will be explained later with reference to FIG.

[Purpose of the invention]

An object of the present invention is to prevent the occurrence of focal position detection errors due to the above-mentioned problems and to provide a highly accurate focal position detection device.

[Summary of the invention]

In the present invention, in order to eliminate the interference caused by unnecessary reflection that occurs in the conventional method of detecting two projection patterns using one image sensor, the projection pattern mask 5 is made 1 m in length and 2.0 m in length, as shown in FIG. The image sensors 9 and 11 are arranged offset in the optical axis direction. In this method, the two image sensors 9 and 11 commonly detect only one pattern projected onto the sample surface. the result,
If the two detection systems can be prevented from interfering with each other, the projection pattern at the same location on the sample will be detected, so the influence of the sample surface reflectance will be the same, and as a result, the effects will be canceled out and higher accuracy will be achieved. It becomes possible.

[Embodiments of the invention]

The present invention will be explained in detail below. FIG. 1 shows the basic configuration of the present invention, in which a light source 1 projects the pattern of a mask 5 onto a sample 6.

It is also a light source for observing the sample 6. An image sensor 8 for observing the sample 6 is focused on the pattern imaging plane of the mask 5.

The image sensor 9.11 for detecting the projection pattern is attached to the mask 30 so that the focal plane is at a position shifted forward and backward along the optical axis across the projection pattern imaging plane of the mask 30. That is, as shown in FIG. 2, the image sensor 1 is focused 1f on the side closer to the lens 5 by centering pressure on 14, which is the pattern imaging surface of the mask 5 and the focusing surface of the image sensor 8.
Image sensor 9 is placed on the fi1a away from the lens 5.
The focal plane 16 of Further, the image sensors 9 and 11 are installed so as to detect a projected pattern at the same location on the sample surface.

The outputs of image sensors 9 and 11 are connected to contrast detection circuits 40α and 40b.

In the contrast detection times Q aoa and aob, the contrast is determined as the difference between the bright part and the dark part of the projection pattern detection signal, and this is normalized.

However, since the reflectance varies depending on the pattern on the sample surface, as shown in Figure 5, the average value Hi of the bright pattern level H4Hi+J of the bright pattern part of the detection pattern and the bright pattern levels H4Hi+J on both sides of the dark pattern part is calculated. Calculate using

Hi = 1 concave rainbow Therefore, the contrast Ci is Ci = Hi - Di Furthermore, a normalization operation is performed using the following equation.

Hi -Di CNi: Shigeko〇i When this operation is performed using an anarchograph circuit.

The circuit configuration is shown in FIG.

From the pattern detection signal amplified to an appropriate level by the amplifier 41, S sample holds 42, 45.44 light level Hi, light level Hi+1 . Dark 1/ Hair A' Dj
@Hold. The outputs of the sample holds 42 and 45 holding the bright levels H4 and H6-+s are connected to a series circuit of resistors 45 and 46 having equal resistance values, and the average value is output from the midpoint. This average value output is passed through the buffer to the °C subtractor 4.
8 and adder 49.

In addition, the dark level signal output from the sample hold 44 is also input to a subtracter 48 and an adder 49.

The outputs of the subtracter 4B and the adder 49 are input to a divider 50 to perform a normalization operation.

Another embodiment is shown in FIG. In this embodiment, a lens 12 is provided for adjusting the magnification of the projected pattern image formed on the image sensor 11. If this lens is not present, the distances from the lens 5 to the image sensors 9 and 11 are different, so the magnifications of the projected pattern images formed on both image sensors are different. As a result, as shown in FIG. 6, the locations detected by the two image sensors no longer match. In the figure, the photoelectric element Pa detects the pattern dark level in both sensors, but does not detect the P%I dark level of the image sensor 11. In such a state, one contrast cannot be detected correctly.

That is, the two detection systems correctly determine the contrast of the projected light pattern and compare the values, which enables accurate detection of the focal position. Therefore, it is preferable that the magnifications of the patterns imaged on the light receiving surfaces of the two image sensors are equal. The purpose of the lens 12 is to match the magnification for this purpose.

As mentioned above as a problem with the prior art, due to the inappropriateness of the temple point in pattern detection, the detection could be performed with confidence.
There is a method of reducing the photoelectric element size, that is, the pixel size, of an image sensor. FIG. 7 shows a pattern detection waveform detected by reducing the pixel size. In this case, since the pixel size is small, the peak of the light pattern and the pixel position are thick (there is no deviation).

〔Effect of the invention〕

According to Akira Kinoe, two detection systems with different focusing planes perform contrast detection without interfering with each other, and also commonly detect patterns projected at the same location on the sample surface, so the Even if the reflectance changes significantly, the rain detection system is affected equally.The effect is canceled out, and the point where the contrast of the rain detection system is equal, which is the focusing target position, does not change, making it possible to detect the focus position with high precision. can be done.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an embodiment of the present invention, FIGS. 2 and 5 are explanatory diagrams of contrast detection in the embodiment, FIG. 4 is a diagram showing a logic judgment circuit in the embodiment, and FIG. Figure 6 is a diagram showing another embodiment of the present invention when the magnification of the detection system is different.
Diagram is detection. Figure 7 is a pattern detection waveform diagram when the pixel size is small;
FIG. 8 is a block diagram of a conventional focal position detection device, FIG. 9 is a diagram explaining the arrangement of the pattern mask in FIG. 8, FIG. 10 is a diagram showing a conventional image sensor and an imaging pattern,
Figures 1 and 8112 are explanatory diagrams of contrast signals, @1
Figures 5 to 15 are illustrations of optical interference within the Senna and its effects. 1...Light source, 2,5.12...Lens, 5...Pattern mask, 4,7.10...Half mirror-16.
...Sample (Queha), 8,9.11...Image sensor, 50.40...Contrast detection circuit, 41...
・Amplifier, 42, 45, 44...sample hold,
45.46...Resistor, 47...Buffer amplifier,
48... Decrementer, 49... Adder, 50... Divider. Representative Patent Attorney Katsuo Ogawa 1, 20 1st ward 20th 30th horse 4th group 8th mouth 11 old tradition 12 Kunimi I3 Edict

Claims (1)

[Claims] 1. A first optical system that projects and images a light pattern that periodically combines bright and dark areas onto an object to be measured via a half mirror and a lens; A second image sensor that images the reflected light pattern through the lens and half mirror on an array image sensor made up of a plurality of photoelectric elements.
and a contrast detection circuit for obtaining brightness information corresponding to a light pattern from an output signal photoelectrically converted by an array image sensor and detecting a focal position of the object to be measured based on the brightness difference. The focal position detection device includes two sets of the arrayed image sensor and the contrast detection circuit, and each image sensor detects the reflected light pattern from the object to be measured at the focal position of the first optical system. The image sensor is installed so that it is shifted forward and backward on the optical axis with respect to the image formation plane when detected by the second optical system, and both image sensors detect contrast at the same location on the object to be measured. A focus position detection device that sets a field of view and compares the contrast detected in the field of view to determine the direction of defocus. 2. At least one of the image sensors is provided with a magnification adjusting lens for equalizing the size of the detection pattern imaged on the two image sensors. Focus position detection device. 3. When detecting the contrast of a periodic light-dark pattern as a difference in light and dark levels, the average value of both adjacent bright (dark) parts that sandwich the dark (bright) part is used as the bright (dark) level for the dark (bright) part. The focus position detection device according to claim 1 or 2, wherein the focus position detection device is calculated using the following.