JPH06229935A - Inspection method for defect - Google Patents

Abstract

PURPOSE:To inspect precisely a defect of a pattern surface and a glass surface of a reticle having an arbitrary thickness. CONSTITUTION:A distance in the direction of the thickness of a reticle R between the position of convergence of a laser beam Lp converging toward a pattern surface PP of the reticle R and the position of convergence of a laser beam Lg converging toward a glass surface PG of the reticle R is made a reference value. First a dimensional difference between the reference value and the thickness of the reticle R is determined as a first process. As a second process, subsequently, a process wherein the position of the reticle R is adjusted so that the pattern surface PP be within the depth of the focus of the beam Lp and the glass surface PG be within the depth of the focus of the beam Lg before inspection of a defect of the pattern surface PP and the glass surface PG is executed, or a process wherein the pattern surface PP is made to coincide with the position of convergence of the beam Lp before the inspection of the defect and then the glass surface PG is made to coincide with the position of convergence of the beam Lg before the inspection of the defect, is selected in accordance with the dimensional difference.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for inspecting minute defects on a substrate, and more particularly to a defect for inspecting defects on both sides of a substrate such as a photomask, reticle or semiconductor wafer used in the manufacturing process of integrated circuits. It relates to the inspection method.

[0002]

2. Description of the Related Art In a lithography process, which is one of manufacturing processes of integrated circuits, a circuit pattern of a mask (reticle) is transferred onto a substrate such as a semiconductor wafer. At this time,
If a defect such as a foreign substance such as dust adheres to the surface of the reticle, the image of the foreign substance is transferred onto the wafer like the circuit pattern. As a result, an image of a foreign substance appears as a defect in the circuit pattern transferred onto the wafer, which causes a reduction in manufacturing yield.

Here, the reticle has a surface on which a circuit pattern is provided (pattern surface) and a surface on which a glass substrate such as quartz is exposed (glass surface). In particular, foreign matter attached to the pattern surface poses a problem. Becomes This is because the pattern surface of the reticle and the surface of the wafer are arranged so as to have an optically conjugate relationship, so that the image of the foreign matter attached to the pattern surface is image-projected onto the wafer. On the other hand, since the glass surface is defocused from the wafer conjugate surface (pattern surface) by the thickness of the reticle, the image of the foreign matter adhering to the glass surface is out of focus on the wafer and is not transferred onto the wafer. However, when the foreign matter is large, it causes uneven irradiation of exposure light on the wafer. From the above, it is necessary to inspect defects on the surface of the reticle on both the pattern surface and the glass surface before transferring the circuit pattern onto the wafer.

As a conventional defect inspection method, in particular, a method for inspecting a pattern surface and a glass surface of a reticle, a laser beam is irradiated on each surface of the pattern surface and the glass surface of the reticle to scatter light generated from each surface. There is known a method of detecting a defect by receiving light with a photoelectric detector arranged on each surface side. When the laser beam for irradiating the pattern surface is the first beam and the laser beam for irradiating the glass surface is the second beam, one line on the pattern surface (the X direction of the XY coordinate system of the reticle is defined by the first beam. (Predetermined scanning range), the optical path is switched, and one line (predetermined scanning range in the X direction) on the glass surface is once scanned with the second beam. Then, by repeating the scanning of the first beam and the scanning of the second beam in synchronization with the driving of the mounting table on which the reticle is mounted in the Y direction, the first beam is spread over the entire pattern surface of the reticle. ,
The second beam is scanned across the glass surface of the reticle. Then, the scattered light generated from each surface is received by each photoelectric detector provided on the pattern surface side and the glass surface side.

Each of the photoelectric detectors converts the scattered light generated from the pattern surface and the glass surface into a photoelectric signal having a level (voltage value) corresponding to the light quantity (magnitude) thereof. Then, each photoelectric signal is compared with a predetermined voltage value determined on the glass surface side and the pattern surface side (hereinafter, this value is referred to as a slice level). When the photoelectric signal detected on the pattern surface side and the glass surface side has a value larger than the slice level, it is detected as a defect, and the size of the defect is determined according to the level of the photoelectric signal.

By the above method, the defects on both the pattern surface and the glass surface of the reticle can be detected by moving the reticle once in the Y direction.

[0007]

However, in recent years, the shape of the reticle (mask) has become diversified with the miniaturization of the LSI pattern in the lithography process and the improvement of the manufacturing yield. Conventionally, a reticle having a shape with a side of 5 inches and a thickness of 0.09 inches has been the mainstream, but at present, even a square with a side of 5 inches has a thickness of 0.
An 18-inch reticle is also used. Further, a reticle having a side of 5 inches as well as a 6 inch has come to be used, and a reticle having a thickness of 0.12 inch or 0.25 inch is used according to the application.

However, in the conventional defect inspection method for inspecting both surfaces of the reticle, only the reticle having a predetermined thickness is inspected. That is, the first beam is focused on the pattern surface of the reticle having a predetermined thickness, and the second beam is focused on the glass surface. Therefore, it was not possible to position the focusing positions of the first beam and the second beam on the pattern surface and the glass surface of the reticle having an arbitrary thickness different from the predetermined thickness. . Further, since the foreign matter on the glass surface is defocused from the wafer conjugate plane (pattern surface) by the thickness of the reticle as described above, the size of the defect to be detected varies depending on the thickness of the reticle. That is, the thinner the reticle, the smaller the size of the defect to be detected, and the higher the detection sensitivity must be. However, the conventional defect inspection method cannot switch the detection sensitivity of defects on the glass surface according to the thickness of the reticle. Further, the thickness of the reticle usually has an error of about ± 0.1 mm, and due to this thickness error, the beam diameter (spot diameter) of the second beam on the glass surface, that is, the light amount per unit area (luminance).
Change causes a difference in scattered light generated from the defect, that is, an error in the detection signal of the defect.

The beam converging position referred to here is the position where the laser beam for irradiating the surface of the reticle is most narrowed by the condensing lens, that is, the position where the beam has the smallest spot diameter on the surface of the reticle. is there. Hereinafter, this position will be referred to as a beam focusing position. The present invention has been made in view of the above problems, and in a method for inspecting defects on the first surface and the second surface of the substrate, the first surface and the second surface of the substrate having an arbitrary thickness, For example, it is an object of the present invention to provide a defect inspection method capable of accurately inspecting a pattern surface and a glass surface of a reticle.

[0010]

In order to make the present invention easier to understand, the present invention will be described with reference to FIG. 1 which is an embodiment. In order to solve such a problem, the present invention provides a pattern surface PP of a plate-shaped substrate (reticle) R having a predetermined thickness.
The first laser beam Lp focused toward the laser beam and the pattern surface PP are relatively scanned to inspect for defects on the pattern surface PP, and the glass surface PG facing the pattern surface PP of the reticle R and the glass surface PG. Toward the laser beam Lg and the laser beam Lg condensed at a position away from the condensing position of the laser beam Lp by a predetermined reference value in the thickness direction of the reticle R are inspected for defects on the glass surface PG. In the defect inspection method, the first step of obtaining the thickness information corresponding to the thickness of the reticle R, the pattern surface PP within the depth of focus of the beam Lp, and the glass surface PG with the beam Lg.
The position of the reticle R is adjusted to be within the depth of focus of the first reticle R, and then the first inspection for inspecting the defects on the pattern surface PP and the glass surface PG is made to substantially coincide with the converging position of the pattern surface PP and the beam Lp. After inspecting the pattern surface PP for defects,
And a second step of selecting a second inspection for inspecting a defect of the glass surface PG by making the converging position of the glass surface PG and the beam Lg substantially coincide with each other according to the thickness information obtained in the first step. Is characterized by.

[0011]

In the present invention, according to the thickness of the substrate, the first substrate
Each of the surfaces and the second surface (for example, the pattern surface and the glass surface of the reticle) are accurately positioned within the focal point or the focal point of the first light beam and the second light beam for irradiating the respective surfaces. It is possible to perform surface defect inspection. Therefore, regardless of the thickness of the substrate used now or in the future, it is always possible to accurately inspect defects on both surfaces (first surface and second surface) of the substrate regardless of the thickness of the substrate. Is possible.

[0012]

1 is a perspective view showing a schematic structure of a defect inspection apparatus according to an embodiment of the present invention. A description will be given below with reference to FIG. The reticle R is mounted on the mounting table 13 with only its peripheral portion supported. The mounting table 13 can be moved in the Y direction in the XYZ coordinate system in the figure by the motor 14 and the feed screw 15, and can also be moved in the Z direction in the XYZ coordinate system in the figure by the motor 16 and the feed screw 17. Has become. And the mounting table 1
The movement amounts of the No. 3 in the Y direction and the Z direction are measured by length measuring devices (linear encoders) 18 and 19.

A laser beam L emitted from a light source LS enters a light splitter (beam splitter) 1. The laser beam Lg reflected by the beam splitter 1 is reflected by the mirror M1.
And is directed in a direction substantially parallel to the laser beam Lp transmitted through the beam splitter 1. Rotating shutter 2
Is rotated at a predetermined angular velocity by the motor 3, and the beam L
The light blocking and passing of g and Lp are repeated. As shown in FIG. 1, the rotary shutter 2 has two sets of fan-shaped light-shielding portions having a central angle of 90 °, and each light-shielding portion is provided rotationally symmetrically with respect to the rotation axis. When the light beam Lg is blocked, the beam Lp is passed, and when the light beam Lp is blocked, the beam Lg is passed. Therefore, the shutter 2 does not pass the respective beams at the same time. In FIG. 1, the beam Lp passes through the shutter 2 and
The state where g is shielded from light is shown.

Further, the rotary shutter 2 can not only rotate but also stand still at a predetermined position. For example, as shown in FIG. 1, it is possible to make the light-shielding portion of the shutter 2 stand still in the optical path of the beam Lg, and to transmit only the beam Lp. Naturally, the opposite (the light-shielding portion of the shutter 2 is kept stationary in the optical path of the beam Lp, and the
It is also possible to allow only the light to pass through, and such rotation or standstill of the shutter 2 is controlled by the motor 3 through the main control system MC.
It is controlled by S.

The laser beam L passing through the rotary shutter 2
g is the size (area) after being reflected by the mirrors M2 and M3
The beam diameter of the predetermined size is adjusted by the light shielding plate 4 provided with the two circular openings 4a and 4b different from each other. The light shielding plate 4 can appropriately position the circular openings 4a and 4b in the optical path of the laser beam Lg by the motor 5. The beam Lg that has passed through the circular aperture (aperture 4a in FIG. 1) is reflected by a scanning mirror (for example, a galvanometer mirror or a polygon mirror) 7, is condensed in a spot shape by a condenser lens 11, and is focused on the glass surface PG of the reticle R. Is projected.
Then, the laser beam Lg is scanned within a predetermined range in the X direction on the glass surface PG of the reticle R by the galvanometer mirror 7 and the motor 8 (hereinafter, this range is referred to as a scanning range Sg). Further, the main control system MCS controls the motor 8 to cause the laser beam Lg to scan the reticle R one-dimensionally, and at the same time to control the motor 14 to move the mounting table 13 on which the reticle R is mounted in the Y direction. Laser beam L
It is possible to scan g over the entire glass surface PG of the reticle R.

The laser beam L passing through the rotary shutter 2
After being reflected by the mirrors M4 and M5, p is adjusted to have a predetermined beam diameter by a light shielding plate 6 provided with a circular opening 6a having a predetermined size (area). Then, the beam Lp that has passed through the circular aperture 6a is reflected by a scanning mirror (for example, a galvano mirror or a polygon mirror) 9, is condensed in a spot shape by a condenser lens 12, and is a pattern surface PP of the reticle R.
Projected on. Then, the laser beam Lp scans a predetermined range in the X direction on the pattern surface PP of the reticle R by the galvano mirror 9 and the motor 10 (hereinafter, this range is referred to as a scanning range Sp). Further, the main control system MCS controls the motor 9 to scan the reticle R one-dimensionally with the laser beam Lp, and controls the motor 14 to move the mounting table 13 on which the reticle R is mounted in the Y direction. The laser beam Lp is applied to the pattern surface P of the reticle R.
The entire surface of P can be scanned.

Here, in the present embodiment, the diameter of the circular opening 4b is 1/2 of the diameter of the circular opening 4a. Generally, for a laser beam with the same wavelength, the diameter of the beam entering the condenser lens and the numerical aperture of the convergent beam after passing through the condenser lens
It is almost proportional to NA. A beam with a large NA has a smaller spot diameter than a beam with a small NA. Therefore, the diameter of the beam Lga which passes through the circular opening 4a and is irradiated onto the glass surface PG of the reticle R is equal to the diameter of the beam Lg which passes through the circular opening 4b and is irradiated onto the glass surface PG.
It becomes 1/2 of the diameter of b. Further, since the light amount of the beam Lga is larger than that of the beam Lgb, the light amount (luminance) of the beam Lga per unit area is significantly larger than the light amount (luminance) of the beam Lgb per unit area. In this type of device, defects can be inspected with high sensitivity by increasing the amount of light (luminance) per unit area of the beam. From this, the main control system MCS arranges the circular opening 4a in the optical path of the beam Lg when inspecting a defect on the glass surface PG with high sensitivity, corresponding to the thickness of the reticle R to be inspected. For inspecting defects on the glass surface PG with low sensitivity, the circular opening 4b is arranged in the optical path of the beam Lg.

The circular aperture 4a and the circular aperture 6a have the same size (area), and the brightness of the beam Lp passing through the circular aperture 6a and irradiated on the pattern surface PP of the reticle R is the above-mentioned beam Lga. Is exactly the same as. Therefore, the defect inspection on the pattern surface PP of the reticle R is always highly sensitive regardless of the thickness of the reticle R. Further, a photoelectric detector 23 is arranged above the reticle R (on the glass surface PG side), and scattered light generated from the glass surface PG is condensed by the condenser lens 21 and enters the photoelectric detector 23. Similarly, the photoelectric detector 24 is arranged below the reticle R (on the pattern surface side), and the scattered light generated from the pattern surface PP is condensed by the condenser lens 22 to be collected by the photoelectric detector 24.
Incident on. Further, the photoelectric detectors 23 and 24 are respectively arranged at positions where the specular reflection light of the laser beams Lg and Lp from the reticle R and the diffracted light from the circuit pattern provided on the pattern surface PP are avoided.

In this embodiment, these photoelectric detectors 2
It is possible to switch the sensitivities of the photoelectric signals obtained from 3 and 24. For example, when photomultipliers are used as the photoelectric detectors 23 and 24, when the scattered light generated from the foreign matter is converted into photoelectric signals by the photoelectric detectors 23 and 24, the main control system MCS operates as the photoelectric detectors 23 and 24. By changing the applied voltage applied to each, the level (voltage value) of the photoelectric signal can be changed to an appropriate magnitude.
Further, by changing the slice level (detection threshold value) of the photoelectric detectors 23 and 24, the level of the photoelectric signal to be detected, that is, the size of the foreign matter can be changed.

As described above, the beam diameter is switched by the light shielding plate 4 in FIG. 1 according to a large change in the thickness of the reticle R to be inspected, and a smaller difference in thickness and fine adjustment of sensitivity are performed. As a result, the defect on the reticle R can be detected more accurately by changing the slice level and the applied voltage. Further, the light source 20 obliquely irradiates a laser beam or the like from above the reticle R (on the glass surface PG side) at a predetermined angle with respect to the glass surface PG, and the beam specularly reflected by the glass surface PG is a position detector (for example, a position detector). The light is received by a sense detector or a four-division silicon photodiode) 21. The light source 20 and the position detector 21 are calibrated such that when a reticle having a predetermined thickness (reference thickness) is irradiated with a laser beam, specularly reflected light from the reticle enters the center of the position detector 21. ing. When the thickness of the reticle changes, the height of the glass surface of the reticle in the Z-axis direction changes, so that the reflected light that enters the position detector 21 also shifts in the Z-axis direction and shifts from the center of the position detector 21. Incident on the right position. By measuring the shift amount with the position detector 21, the error amount between the reference thickness and the thickness of the reticle to be inspected, that is, the thickness of the reticle can be measured. Hereinafter, the light source 20 and the position detector 21 will be collectively referred to as a "thickness measuring device". When measuring the thickness of the reticle R, the main control system MCS places the mounting table 13 on which the reticle R is mounted.
Is positioned at a predetermined position in the Z-axis direction (hereinafter, this position is referred to as a reference position). At this reference position, the focus position of the laser beam Lp and the pattern surface PP of the reticle always match, and when the thickness of the reticle is the reference thickness, the focus position of the laser beam Lg and the glass surface PG also match. To do.

The position information from the position detector 21 is sent to the main control system MCS, and the thickness amount corresponding to the reference thickness and the amount of error from the reference thickness is stored in advance in the main control system MCS.
Further, the distance between the focus position of the laser beam Lg and the focus position of the laser beam Lp in the thickness direction of the reticle R (scanning range S
The distance between g and the scanning range Sp in the thickness direction of the reticle R) is the reference thickness. Further, the defect inspection apparatus of this embodiment has a reticle thickness of 0.25 inch (about 6.3 mm).
The standard thickness is used. That is, in FIG. 1, when the thickness of the reticle R is 0.25 inch, the pattern surface PP of the reticle R and the focus position of the laser beam Lp coincide, and the focus position of the glass surface PG and the focus position of the laser beam Lg. Match.

The main control system MCS in FIG. 1 controls the drive of the motors 3, 5, 8, 10, 14, 16.
Controls the entire device. FIG. 2 shows an example of a flow chart of the defect inspection method of this embodiment, and FIGS. 3, 4 and 5 show laser beams Lp and L on the pattern surface PP and the glass surface PG of reticles having different thicknesses in this apparatus.
It is a top view which shows the state which is irradiating each g. Less than,
A method of inspecting the defect state of the pattern surface PP and the glass surface PG of the reticle R using this apparatus will be described with reference to FIGS. 2, 3, 4, and 5. In each of FIGS. 2 to 5, members having the same functions as those in FIG. 1 are designated by the same reference numerals.

Now, in step 101, the main control system MCS places the reticle R to be inspected on the mounting table 13. Next, in step 102, the thickness of the reticle R is determined. The reticle R has a certain thickness as described in the section of the problem to be solved by the above invention, and this thickness is used by the operator through the terminal to control the main control system M.
Enter into CS. And the thickness of the reticle R is 0.25.
If it is inch, proceed to the next step 103. If it is other than 0.25 inch, proceed to step 104.

Next, step 103 will be described. As described above, since the reticle thickness has an error of about ± 0.1 mm, even if it is determined in step 102 that the reticle R thickness is 0.25 inch, which is the reference, the actual size and the reference There is an error with the thickness T (= 0.25 inch). Therefore, the actual thickness t of the reticle is measured by the thickness measuring device (20, 21) shown in FIG. 1, and the error Δt (= | T−t |) from the reference thickness T is measured. Here, the laser beam Lg for irradiating the glass surface PG of the reticle R
Let DOFG be the depth of focus of the laser beam Lp and DOFP be the depth of focus of the laser beam Lp that illuminates the pattern surface PP.
If OFG, go to step 106, DOFG <Δt ≦ DO
When FG + DOFP, go to step 107, DOFG + D
When OFP <Δt, the process proceeds to step 108.

Now, the depth of focus of the beam referred to here will be described. The laser beam irradiating the reticle has the smallest irradiation area at the focusing position, that is, the spot diameter of the beam on the reticle, and the irradiation area increases as the distance from the focusing position increases. As the irradiation area increases, the amount of light per unit area (luminance) naturally decreases in proportion. In this embodiment, the variation amount of the brightness of the beam from the focusing position (variation amount of the irradiation area of the beam from the focusing position) is 2
The distance in the Z-axis direction between the position of the beam (center point) and the focus position (center point) at 0% is the depth of focus. If the laser beam irradiates the surface to be inspected of the reticle within this range, it is not necessary to correct the sensitivity of the photoelectric detector that receives the scattered light generated from the defect. This range for each beam will be referred to as depth of focus DOFG, DOFP
Is written.

Next, step 106 will be described with reference to FIGS. 3 (a) and 3 (b). FIG. 3A shows the case where the thickness error Δt = 0 in step 103, and FIG. 3B shows the case where the thickness error is 0 <Δt ≦ DOFG in step 103. When Δt = 0, the pattern surface PP of the reticle R coincides with the condensing position 31 of the laser beam Lp, and the glass surface PG coincides with the condensing position 30 of the laser beam Lg. Also, 0 <Δ
When t ≦ DOFG, the glass surface PG of the reticle R is within the focal depth DOFG of the laser beam Lg, and the pattern surface PP coincides with the converging position 31 of the laser beam Lp. Therefore, both can perform the defect inspection of the glass surface PG and the pattern surface PP of the reticle R without moving the mounting table 13 from the reference position. That is, the movement amount ΔZ = 0 in the Z-axis direction from the reference position.

Here, the reticle R having a thickness of 0.25 inch is a so-called thick reticle, and the detection sensitivity of defects on the glass surface PG may be relatively low. Therefore, when performing the defect inspection, the main control system MCS drives the motor 5 to arrange the circular opening 4b in the light shielding plate 4 in the optical path of the laser beam Lg. Then, the main control system MCS controls the driving of the motors 3, 8 and 10 to control the laser beam Lg on the glass surface PG (scanning range S
The scanning of g) and the scanning of the laser beam Lp on the pattern surface PP (scanning range Sp) are alternately repeated to control the motor 16 to move the mounting table 13 in the Y direction.

Next, step 107 will be described with reference to FIGS. 4 (a) and 4 (b). Step 103
, The dimensional difference Δt (= | T−t |) between the thickness t of the reticle R and the reference thickness T is larger than DOFG.
When the value is equal to or less than the sum of G and DOFP (DOFG + DOFP), the glass surface PG of the reticle R is located outside the depth of focus DOFG of the laser beam Lg, as shown in FIG. 4A. However, since the dimensional difference Δt between the reticle R and the reference thickness T is smaller than the sum DOFG + DOFP of the focal depths of the laser beams Lg and Lp, the mounting table 13 is moved from the reference position in the Z-axis direction. Each of the glass surface PG and the pattern surface PP of the reticle R can be arranged within the depth of focus of each of the laser beams Lg and Lp.

FIG. 4B is a plan view showing a state after the reticle R is moved by ΔZ in the Z-axis direction from the reference position. Here, the depth of focus DOFG of the laser beam Lg and the depth of focus DOFP of the laser beam Lp in the present embodiment.
Satisfies the relationship of DOFG: DOFP = 2: 1. As shown in FIG. 4B, the converging position 30 of the glass surface PG and the laser beam Lg after the reticle R is moved by ΔZ.
, And the distance (defocus amount) between the pattern surface PP and the converging position 31 of the laser beam Lp is Z2 (= ΔZ), the main control system MCS sets each defocus amount Z1. , Z2 so that the value of the ratio of the depths of focus of the respective laser beams becomes the same, that is, Z1:
The mounting table 13 is positioned so that Z2 = 2: 1 (ΔZ = Δt · DOFP / (DOFP + DOFG)).
In this way, by dividing the error amount Δt of the thickness of the reticle R into the defocus amount of the laser beam Lg on the glass surface PG and the defocus amount of the laser beam Lp on the pattern surface PP, the above-described step 106 is performed. With the same defect inspection method (method of alternately inspecting the glass surface PG and the pattern surface PP), it is possible to accurately inspect the defect state of each surface PG, PP. However, since this reticle R is a reticle having a thickness of approximately 0.25 inch and has a slight error in thickness, the main control system MCS performs a circular opening in the optical path of the laser beam Lg when performing a defect inspection. Place 4b.

Now, step 108 will be described with reference to FIG. When the value of the difference Δt between the thickness t of the reticle R and the reference thickness T is larger than the sum of the depths of focus (DOFG + DOFP) of the laser beams in step 103, as shown in FIG. The glass surface PG is located outside the depth of focus of the laser beams Lg, and even if the mounting table 13 is moved in the Z-axis direction, the glass surface PG and the pattern surface P of the reticle R are simultaneously exposed to the laser beams Lg and Lp, respectively. It cannot be placed within the depth of focus. Therefore, the main control system MCS first performs a defect inspection on the pattern surface PP in a state where the focus position 31 of the pattern surface PP laser beam Lp is matched. At this time, the main control system MC
S positions the light blocking portion of the rotary shutter 2 in FIG. 1 in the optical path of the laser beam Lg so that only the laser beam Lp passes through. When the defect inspection of the pattern surface PP is completed, the main control system MCS sets the mounting table 13 to ΔZ =
The glass surface PG is moved by Δt to position the glass surface PG at the focusing position 30 of the laser beam Lg, and the defect inspection of the glass surface PG is performed. At this time, the main control system MCS positions the light shielding portion of the rotary shutter 2 in FIG. 1 in the optical path of the laser beam Lp so that only the laser beam Lg passes.

In this way, even if the mounting table 13 is moved in the Z-axis direction, if both the glass surface PG and the pattern surface PP of the reticle R cannot be arranged within the focal depths of the laser beams Lg and Lp, the respective surfaces are moved. Inspect for defects separately. However,
Since the reticle R has a thickness of about 0.25 inch and has a slight thickness error, the main control system MCS arranges the circular opening 4b in the optical path of the laser beam Lg when performing the defect inspection. To do. That is, the detection sensitivity of the glass surface PG may be low.

Here, if Δt becomes very large due to an error in inputting the thickness information of the reticle R in step 102, the operator may be notified of the error information by an indicator or the like (not shown). Next, steps 104 and 105 will be described. Step 102
In step 104, the main control system MCS measures the actual thickness t of the reticle R with the thickness measuring device (20, 21) in FIG. And the difference Δ with the reference thickness T
Measure t. When the measurement of the thickness of the reticle R is completed, the routine proceeds to step 105, where the glass surface PG and the pattern surface PP of the reticle R are inspected for defects. Here, when a reticle having a thickness of 0.09 inch is arranged in the apparatus of this embodiment,
The state is the same as that of FIG. 5 used in the description of step 108. That is, even if the mounting table 13 is moved in the Z-axis direction, both the glass surface PG and the pattern surface PP of the reticle R cannot be arranged within the focal depths of the laser beams Lg and Lp. Therefore, in step 105, similarly to step 108, the glass surface PG and the pattern surface PP of the reticle R are separately inspected for defects. The defect inspection method is the same as step 108 above.
However, since the thickness of the reticle R is 0.09 inch, the detection sensitivity of defects on the glass surface PG is 0.2.
It should be higher than with a 5 inch reticle. Therefore, the main control system MCS is the glass surface P of the reticle R.
When the defect inspection of G is performed, the motor 5 is driven to arrange the circular opening 4a in the light shielding plate 4 in the optical path of the laser beam Lg.

By the above method, an arbitrary thickness t
Glass surface PG and pattern surface PP of reticle R having
It is possible to accurately inspect the above defects. Also,
In step 102 of this embodiment, the reticle thickness information is input by the operator. However, by providing a barcode with thickness information on the reticle and a barcode reader that reads the barcode on this inspection device, it is possible to automatically determine the thickness of the reticle without operator intervention. Is. Further, the case where the thickness of the reticle is thinner than 0.25 inch has been described in steps 104 and 105 of the present embodiment.
If it is thicker than 0.25 inch, the glass surface may be inspected with low sensitivity. Further, in the present embodiment, in step 105, if the reticle thickness is other than 0.25 inch, all are detected with high sensitivity. For example, the beam of the laser beam Lg is bordered by the thickness of 0.17 inch. Select the diameter, that is, the circular openings 4a and 4b in FIG.
It is also possible to select whether to detect G with high sensitivity or low sensitivity.

Further, step 102 and step 103
And step 105 may be simultaneously performed by the thickness measuring device (20, 21). That is, after the reticle is placed in step 101, the thickness of the actual reticle is measured by the thickness measuring device (20, 21), and an error Δt from the reference thickness T is obtained.
Corresponding to the value of
You may make it proceed to 8). At this time, if the thickness error Δt is DOFG + DOFP <Δt, it is necessary to determine whether to proceed to step 108 or step 105. Therefore, for example, if the thickness of the reticle is thin at the boundary of 0.17 inch, the glass surface is inspected with high sensitivity, and if it is thick, it is inspected with low sensitivity in advance in the main control system MCS. . Therefore, when Δt <0.08, the main control system MCS executes step 108 to place the circular aperture 4b in the optical path of the beam Lg. Then, when Δt ≧ 0.08, the main control system MCS executes step 105 to arrange the circular aperture 4a in the optical path of the beam Lg.

Further, not only the detection sensitivity of the photoelectric detectors 23 and 24 is switched between circular apertures in accordance with the thickness of the reticle, but also the applied voltage applied to each photoelectric detector is changed as described above. By changing the slice level, it is possible to obtain detection sensitivity corresponding to the thickness of the reticle. Of course, in any sensitivity switching method,
For example, the detection sensitivity of the glass surface is calibrated with polystyrene particles of φ2 μm at high sensitivity and polystyrene true spherical beads of φ15 μm at low sensitivity as reference particles.

Further, although the reticle is used as the object to be inspected in the defect inspection method of this embodiment, the reticle with the pellicle is effective for the defect inspection of the reticle pattern surface and the glass surface as well as the pellicle surface. Is the way. That is, when the pellicle is provided on one side or both sides of the reticle, the entire thickness of the reticle with a pellicle may be considered to be the reticle thickness described in this embodiment. The pellicle surface is separated from the surface of the reticle by, for example, 4 mm, 5 mm, and 6.3 mm, and the error is about ± 0.2 mm. Therefore, the defect inspection can be performed by a method substantially similar to the flowchart of FIG.

The thickness measuring device (20, 2) of this embodiment is also used.
In place of 1), for example, an optical oblique incidence type surface position detection device is provided, and the reference position of the surface position detection device and the laser beam L
The light collecting position of g is arranged in the same plane,
If the position of the reticle R is moved in the Z direction so that the glass surface PG coincides with the reference position, the glass surface PG can be accurately positioned at the focus position of the laser beam Lg without measuring the thickness of the reticle R. You can

[0038]

As described above, according to the present invention, it is possible to accurately inspect for defects on both surfaces of a plate-shaped substrate having an arbitrary thickness. Therefore, no matter what thickness of substrate is used in the future, it is possible to always perform accurate defect inspection regardless of the thickness of the substrate.

[Brief description of drawings]

FIG. 1 is a perspective view showing a schematic configuration of a defect inspection apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a flowchart in a defect inspection method of the present embodiment.

FIG. 3 shows that the thickness error Δt is Δt ≦ Δ in step 103.
FIG. 6 is a plan view showing a state where both surfaces of a reticle that is a DOFG are irradiated with laser beams.

FIG. 4 shows that the thickness error Δt is DOF in step 103.
FIG. 9 is a plan view showing a state where both surfaces of a reticle satisfying G <Δt ≦ DOFG + DOFP are irradiated with a laser beam.

FIG. 5 shows that the thickness error Δt is DOF in step 103.
FIG. 6 is a plan view showing a state where both sides of a reticle satisfying G + DOFP <Δt are irradiated with laser beams.

[Explanation of symbols]

 L ... Laser beam R ... Reticle MCS ... Main control system 2 ... Rotating shutter 4 ... Shading plate 7, 9 ... Galvano mirror 20 ... Laser light source 21 ... Position sense Detector

Claims (3)

[Claims] 1. A defect of the first surface is inspected by relatively scanning a first surface of a plate-like substrate having a predetermined thickness and a first light beam focused toward the first surface. A second surface facing the first surface of the substrate, a second surface facing the first surface, and a position separated from the converging position of the first light beam by a predetermined reference value in the thickness direction of the substrate. A defect inspecting method for inspecting defects on the second surface by relatively scanning with a focused second light beam, the first step of obtaining thickness information corresponding to the thickness of the substrate; Is within the depth of focus of the first light beam, and the second surface is within the depth of focus of the second light beam.
After inspecting for defects on the first surface by making the first inspection for inspecting defects on the surface and the second surface and the converging position of the first surface and the first light beam substantially coincide with each other, the first inspection is performed. And a second step of selecting a second inspection for inspecting a defect on the second surface by substantially matching the condensing position of the second light beam with the second surface, and a second step of selecting according to the thickness information. Characteristic defect inspection method. 2. The defect inspection method according to claim 1, wherein the thickness information is a dimensional difference between the thickness of the substrate and the reference value. 3. In the first inspection, an interval between the first surface and a converging position of the first light beam, and an interval between the second surface and a converging position of the second light beam. The ratio of
3. The defect inspection method according to claim 1, wherein the ratio of the depth of focus of the first light beam to the depth of focus of the second light beam is substantially the same.