JPS59101832A - Automatic focuser - 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 Application of the Invention] The present invention relates to an automatic focusing device, and more particularly to an automatic focusing device for an imaging device of a semiconductor pattern inspection device or a minute dimension measuring device.

[Prior art]

As a conventional automatic focusing device, a photomask visual inspection device that requires a high degree of automatic focusing will be taken as an example, and the drawbacks of the conventional technology will be explained. FIG. 1 is a diagram showing an example of a conventional automatic focusing device in a photomask visual inspection device. In the figure, a photomask 1, which is an object to be inspected, is placed on an X-Y stage 2. It is placed on the Z stage 6. An objective lens 4 is disposed above the photomask 1, and the pattern of the photomask 1 is enlarged and projected by the objective lens 4, and an image is formed on a pattern sensor f5 disposed above the objective lens 4. On the other hand, the magnetic tape 6 storing the design data of the circuit pattern of the photomask 10 as an image signal and the output signal of the pattern sensor 5 are electrically connected to the defect determination section 7. A nozzle is formed below the objective lens 4 for pattern detection, and air is supplied to the nozzle. This air is supplied through the gap detection section 8. The gap detection unit 8 detects the gap g by detecting the difference between the back pressure of the nozzle of the objective lens 4 and the air supply pressure), and the detection of the gap g is based on the principle of an air micrometer. It uses A controller 9 that drives the two stages 3 is connected to a gap detection section 8.

The inspection of the photomask 1 is performed on the entire surface of the photomask 1 in the order indicated by the arrows in FIG.

While scanning the inspection friend X-Y stage 2, the pattern sensor f
5 and the output signals of the magnetic tape 6 are compared in a defect determination section 7,
If a difference is detected, it is determined that there is an abnormality. During this inspection,
Since the photomask 1 has a lot of ridges and there is also a difference in the linear motion of the x-y stage 2, it is necessary to keep the objective lens 4 in focus at all times. Moreover, with the recent miniaturization of LSI patterns, it has become necessary to detect defects of 1 μm or less. For this reason, lenses with high resolving power, such as N
It is necessary to use a high-resolution lens with a depth of focus of 0.2 μm. For these reasons, focusing accuracy of about 0.1 μm is required, and high-precision focusing 9
It is necessary to match. Focusing of 0.1 .mu.m is also possible with the conventionally used air cyclometer method. That is, the gap detection section 8 converts the displacement of the gear z g into voltage, and the controller 9 drives the drive source (not shown) of the z stage 3 based on the change in the voltage of the gap detection section 8 to determine the gap gJkg±. It is made to be within the range of 0.2 μm.

However, in the air micrometer method shown in Fig. 1, a gas such as air is blown onto the object to be observed (photomask 1), so if this gas is dirty, there is a risk of blowing foreign matter onto the object to be observed. There is a risk that this foreign material may be judged as a defect, and there is also the drawback that it may damage or stain the observed object.

Furthermore, if the atmosphere around the object to be observed is dirty due to the gas being blown out, there is a drawback that foreign matter is rolled up and attached to the high-resolution objective lens 4, which deteriorates the resolution. Furthermore, since this method performs gap detection via fluid (gas), the focal point detection speed is influenced by the transmission speed of the fluid. That is, in FIG. 1, the gap detection section 8
The piping 10 from to the objective lens 4 is made shorter and thicker,
It is necessary to minimize fluid resistance, increase focus detection speed, and shorten inspection time. However, in the focusing method using this fluid, it takes only several tens to 100 hours.
It has the disadvantage that only a response speed of about z can be obtained. Moreover,
In this method, it is impossible to directly detect the detection part of the objective lens 4, and the field of view of the objective lens 4 is
Since it detects a gap of about 100 degrees, it has the disadvantage that accurate focusing cannot be achieved if the object to be observed has fine ridges or irregularities.

FIG. 3 is a diagram showing a completely non-contact type focusing device using light, which was devised to eliminate the drawbacks of the above-mentioned air micro type automatic focusing device. This method uses illumination light for focusing, which is different from the detection light, projects an image of this illumination light onto the object to be observed, and detects it using a light intensity sensor that detects the intensity of the reflected light. Focusing is performed by detecting the strength and weakness of light intensity due to the movement of an object. To explain in more detail with reference to FIG. 3, an objective lens 4 is disposed above the photomask 1, and the pattern of the photomask 1 is enlarged and projected onto the pattern sensor 5 by the objective lens 4 to form an image. A mirror 11 is disposed in the middle of the optical path L1 that forms an image on the pattern sensor 9'5, and the irradiation light is incident through the mirror 11. An image S of this irradiation light is placed on the optical path L2, and this image S is focused by the objective lens 4 onto the surface of a photomask. Using this image as 81, the optical path L2 is further branched by a mirror 12, and the image S1 on the photomask 1 is projected onto the detection surface of the light intensity sensor 13 and detected. The image on the detection surface of the light intensity sensor 13 is denoted by 82. Here, the light intensity sensor f13 is configured to be able to reciprocate in the vertical direction.

When the enlarged projected image of the pattern of the photomask 1 becomes the focused point at cuff < Turnsen f5, that is, the photomask 1 is at 0 in the figure.
The center position of the reciprocating movement of the light intensity sensor 13 is set at O in the figure so that the output of the light intensity sensor 13 is maximized when
Adjust it so that it is ’ side. Now, when photomask 1 is on surface 0, is the light intensity sensor? The detected image S2 of 13 also has the maximum light intensity on the O' plane, and the relationship between the amount of movement and the output when the light intensity sensor 16 is moved back and forth is as shown in FIG. 4(a). That is, the maximum output is achieved on the 02 plane, and the output becomes smaller on the p' Ifrt Q' plane, which has moved by L'. In order to determine the focal point range, as shown in FIG. 4(b), when the photomask 1 is on the 0 plane, the light intensity sensor 16 has a maximum output on the 01 plane.

At this time, perform threshold value processing at, for example, 90 degrees of the maximum output on the 0' plane, and set the center of the threshold value when an output of 90% or more is obtained as the in-focus point position, and this in-focus point position is There is also a method of using the focal point as long as it is within the range of θ. However, at this time, the pattern sensor 5 must also be in focus.

The photomask 1 moves during the inspection, and the photomask 1 moves as indicated by P in FIG.

If the light intensity sensor 1 moves in a reciprocating manner,
As shown in FIG. 4(C), the output of No. 6 is maximum at the P/ position, becomes small at the 0' position, and becomes even smaller at Q'. The direction in which the photomask 1 is moving can be detected by this change in output. In order to focus here, the output of the light intensity sensor f16 is
) The photomask 1 may be moved so that it falls within the range of θ shown in ).

However, the automatic focusing device shown in FIG. 3 requires a reciprocating movement of the light intensity sensor f13, so a driving device is not required, and the device as a whole becomes complicated and dusty. Further, regarding the movement stroke, for example, when an image is formed by the objective lens 4 with a magnification of 40 times and a focus shift of 1 μm occurs, the light intensity sensor 13 must be moved by about 1.6 degrees. Therefore, even if the defocus detection stroke is ±5 μm, the stroke of the light intensity sensor f13 is 16101). Since this mechanical reciprocating movement is performed with 16 strokes, there is a limit to the movement speed and the detection speed is slow. Further, if there is a change in the reflectance of the object to be observed, the output of the light intensity sensor f13 will also change, and there is a drawback that it is impossible to focus an object to be observed whose reflectance is twice or more.

In order to eliminate the drawbacks of the automatic focusing device shown in FIG. 3 described above, FIG. 5 is characterized in that it is fixed using two light intensity sensors. As shown,
A mirror 1 is placed in the middle of the optical path L1 that forms an image on the pattern sensor 5.
6 is arranged, and the irradiation light is incident through the mirror 16. The image S of this irradiation light is placed on the optical path L3, and the image S is formed on the photomask 1 by the objective lens 4 as an image S1.

This image S1 is an optical path L4 branched by mirrors 17 and 18.
, is projected onto LS as image S2. 1, as shown in the figure, two light intensity sensors 14 are installed in front and behind the image S2.
, 15 are placed.

When the enlarged projected image of the pattern of the photomask 1 is focused on the enlarged projected image F/ζTernsen+5, that is, when the photomask 1 is on the 0 plane, the image S2 is formed on the Q' plane. Also, photomask 1
The light intensity is centered on the focused plane P' of the 02 plane above the 00 plane by λ! A light intensity sensor 14 is arranged on the focused plane Q' plane of the Q plane, which is 2 below from 0. With the above configuration, when the photomask 1 is located on the 0 plane, the light intensity Sensors 14 and 15 show equivalent outputs, and the output difference is 0.
become. Next, as the photomask 1 approaches the surface 0 to surface 2 during inspection, the output of the light intensity sensor f15 increases and the output of the light intensity sensor 14 decreases. At this time, the light intensity sensor 14.
If we take the difference between the outputs of 15, it is clear that the direction of movement of Hoya Mask 1 is caninal. Therefore, if the photomask 1 is moved until the output difference between the light intensity sensors f14 and f15 reaches 〇, the image will come into focus. By repeating this operation repeatedly during an inspection, focusing can be achieved automatically.

Compared to the automatic focusing device shown in FIG. 6, the automatic focusing device shown in FIG. 5 does not require a driving device because the light intensity sensor is not movable.

This has the advantage of increasing detection speed. Further, even if there is a difference in the reflectance of the object to be observed, there is an advantage that accurate focusing can be achieved because the difference in the outputs of the light intensity sensors 14 and 15 is used to perform l/ without multi-focusing.

However, since the automatic focusing device number shown in FIG.
The ζ turn sensor 5 indirectly uses the ζ turn sensor 5 (because it focuses on the turn centimeter 5, for example, if the position of the image S goes out of order during an inspection or during long-term use, even if the focus is in focus in the focus detection section) There is a drawback that the focal point of the photomask 10 detected by this method is out of focus. In addition, because it detects the spots and spots on the photomatrix 1 caused by illumination light, it is difficult to improve automatic focusing accuracy for objects with patterns formed on their surfaces, and the reflectance of objects differs depending on the presence or absence of patterns. Can not.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to eliminate the above-described drawbacks of the prior art and to provide an automatic focusing device for, for example, a semiconductor pattern inspection device, which enables highly accurate focusing.

[Summary of the invention]

The automatic focusing device of the present invention includes a first imaging device using a lens for observing an object to be observed, and a second imaging device for detecting a deviation from the in-focus plane of the first imaging device. It is characterized by

[Embodiments of the invention]

The present invention will now be described in more detail with reference to embodiments shown in the accompanying drawings.

FIG. 6(a) is a diagram showing the principle of the automatic focusing device of the present invention. As shown in the figure, an objective lens 4 is placed above the photomask 1, and the objective lens 4 allows the photomask to be
The pattern is enlarged and projected, and an image is formed on butter:/sen f5 placed above the objective lens 4. A semi-transparent mirror 2J is placed in the middle of the optical path M1 of the optical system that images the pattern sensor fSKMi, and the optical path M1 is branched.
Set it to 2. Furthermore, a semi-transmissive mirror 22 is similarly arranged in the middle of the optical path M2 to branch the optical path M2, and the branched optical path is converted into the optical path M3.
shall be.

FIG. 6(b) is an enlarged view of the field of view of the objective lens 4, and the pattern centimeter 5 for detecting the pattern 26 of the photomask 1.
The detection range 24 is shown.

This is the focused plane of the vane II-n 23 on the 0 plane of the photomask 1 by the objective lens 4. Arrange pattern sensor f5 on the 1st side, and move on to the 0th side! A pattern sensor f20 is placed on the 21st focused plane of Prl1i at the top, and the 0th focused plane is below the 0th plane. The first side has Ki Turnse:/f19.

In the above configuration, as the photomask 1 moves during inspection and approaches the P plane from zero reduction, the focus of the pattern centimeter 5 gradually shifts and comes into focus on the pattern centimeter f20. At pattern sensor f19, the focus is further shifted. This will be explained using an off diagram that is an example of signal processing at this time. Off diagram (a) &L pattern sensor? 19.20 shows the batashi edge signal when pattern 26 of photomask 1 is detected, and off diagrams (b) and (d) show the pattern sensor f.
19.20 shows the differential signal of the pattern edge signal, and off-graphs (0) and (e) show the difference between the above-mentioned differential signals as a differential signal.

Now, when the photomask 1 approaches the P side from the 0 side, a difference occurs in the pattern edge signals at pattern centimeters 19 and 20. The differential signal of the pattern edge signal at this time is as shown in off-line diagram (b). Furthermore, if the difference between the two differential signals is taken, the result will be as shown in the off-line diagram (c), and the area where the output of this difference signal becomes 0 is located at the surface of the photomask 100.

Therefore, in order to focus the enlarged projected image of the pattern 24 of the photomask 1 on the pattern cm 5, it is necessary to
d) The photomask 1 may be moved so that the output is as shown in (e).

FIG. 8 is a diagram showing an embodiment of an automatic focusing device using the above-mentioned automatic focusing method, and is an example applied to the photomask appearance inspection device shown in FIG. 1. The difference from Fig. 1 is that the gap detection section 8 using the focusing button is replaced by the pattern 779''19.20 shown in Fig. 6.
and semi-transparent mirrors 21 and 22, and a signal processing circuit 25 are added, and the output signals of pattern centimeters 19 and 20 are electrically connected so as to be input to the signal processing circuit 25.

The signal processing circuit 25 is connected to a controller 9 that drives the two stages 3. The other configuration and operation of the inspection device are the same as those shown in FIG.

The pattern edge signal of the photomask 1 detected by the pattern sensor f19.20 is input to the signal processing circuit 25.

The signal processing circuit 25 processes the pattern edge signal as described above, and controls the amount and direction of defocus by the controller 9.
Output to. The controller 9 controls the two stages 3 based on the output of the signal processing circuit 25. If this operation is performed in real time during detection, the focus will be automatically adjusted.

FIGS. 9 and 10 are diagrams showing a second embodiment of the present invention. In a method in which a pattern such as a photomask is projected using an imaging device using a lens, and the enlarged projected image is detected using photoelectrons, a part of the field of view of the objective lens is usually detected depending on the shape of the photoelectric element. For this reason, a method may also be considered in which a part of the enlarged projected image, which is not used for detection within the field of view of the objective lens, is used for focusing. FIG. 9(a) is a diagram showing the principle of a focusing method using the above method.

Further, FIG. 9(1)) is an enlarged view of the field of view of the objective lens 4. That is, in FIGS. 9(a) and 9(b), the pattern sensor 26 for detecting defects in the pattern 29 of the photomask 1 detects the detection range 26', and the pattern sensor 27 and 28 used for focusing detects the detection range 27. ', 28' is detected. Pattern 290 of surface 0 on photomask 1 Focus point W
A pattern sensor f26 is arranged on the JO1 plane, a pattern sensor 28 is arranged on the two focused planes P' which are two planes above the zero plane, and a nine focused plane which is further 2 below the zero plane. Pattern sensor f27 is arranged on the Q' side. Even with this type of focusing device, is there a pattern sensor? It goes without saying that if the detection sensitivities of 27 and 28 are adjusted, the in-focus position can be detected in the same manner as the signal processing method shown in the off-line diagram.

FIG. 10 is a diagram showing an embodiment of an automatic focusing device embodying the focusing method shown in FIG. 9. The automatic focusing device shown in FIG. 10 is similar to the pattern sensor of the device shown in FIG. 19. Pattern Senna 27. instead of 20.
28 ft, and a mirror 30.31 is used in place of the translucent mirror 21.22. Note that Pattern Sen+ 5.27.28 is 9 near centimeters. The rest of the configuration is exactly the same as that in FIG. Figures O6 and O9 above
The focusing method shown in the figure has no moving parts for detecting the in-focus point, and also requires no special illumination for focusing, making it possible to downsize the automatic focusing device. Further, in this method, since the projected image of the detection pattern is directly observed and focused, it is possible to perform focusing with high precision and without deviation. Furthermore, according to the method shown in FIG. 9, since focusing is performed using an optical path independent of the pattern detection optical path, there is no need to use an optical element such as a mirror in the pattern detection optical path. There is no loss of light amount to the detection light. Furthermore, by rearranging the position of the pattern sensor that detects the pattern, focusing can be performed before detecting the pattern, and focusing can be performed with high precision and high speed detection.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, since focusing is performed by directly observing the field of view of the objective lens without contacting the object to be observed, high-precision focusing of 0.1 μm or less is possible. It is possible to perform high-speed focus detection on the order of KHz. When this focusing device is applied to the automatic focusing device of LSI inspection equipment, high-speed and high-precision focusing can be performed using a high-resolution objective lens with a depth of focus of ±0.2 μm @, O, S ~ 1 μm Even small defects can now be inspected at high speed. As a result, it has become possible to significantly improve the yield of LSI photomasks and, by extension, the yield of LSI.

It should be noted that the present invention is not limited to LSI photomask inspection (it goes without saying that it can be applied to an automatic focusing device for visual inspection and dimension inspection of fine patterns such as LSIs, LSIs, and F1 bubble memories). .

[Brief explanation of drawings]

Figure 1 is an explanatory diagram showing a conventional automatic focusing device, Figure 2 is an explanatory diagram showing a conventional automatic focusing device;
The figure is an explanatory diagram showing the photomask defect inspection procedure, Fig. 6 is an explanatory diagram showing the principle of the conventional automatic focusing method, and Fig.
Figures (a), (1)), and (0) are diagrams showing the relationship between the amount of movement and the output when the light intensity sensor shown in Figure 6 is moved back and forth, and Figure 5 is for the conventional automatic focusing method. FIG. 6(a) is an explanatory diagram showing the principle of the automatic focusing device of the present invention. FIG. 6(b) is an enlarged view of the field of view of the pattern sensor shown in FIG. 6(a). , OFF diagram (a) is a diagram showing the pattern edge signal of t<turnsen shown in FIG. 6(a), OFF diagram (b) j (d) is a diagram showing the differential signal of the pattern edge signal, OFF diagram (0), (θ)
is a diagram showing a differential signal obtained by taking the difference between the above differential signals, Figure O8 is a diagram showing the first embodiment of the automatic focusing device of the present invention, and Figure O9 is a diagram showing the principle of the automatic focusing device of the present invention. The explanatory diagram shown in FIG. 10 is a diagram showing a second embodiment of the automatic focusing device of the present invention. 1... Photomask, 2... X-Y stage, 3...
...2 stages, 4...objective lens, 5.19.2
0, 26, 27.28...pattern centimeters, 7...
Defect determination section, 8... Gear, detection section, 9... Controller, 1!1, 14.15... Light intensity cm, 2
5...Signal processing circuit. Figure 1 Nishiki 2 Figure 3 Ward Otoko 4 Figure 5 Figure 6 (ku) Otoko 7 Flash (b) (d) 3rd figure'

Claims (1)

[Claims] A first imaging device using a lens for observing an object to be observed, and a second imaging device for observing within the field of view of the lens and detecting a deviation from a focused plane of the first imaging device. An automatic focusing device characterized by: