JP2007304054A - Surface inspection method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a surface inspection method which enables certain and favorable execution of inspection, and to provide a surface inspection device suitable therefor. <P>SOLUTION: The surface inspection method is constituted so that the pattern on a substrate to be inspected is illuminated by polarized illumination light and a reflection signal showing the polarized state of the regular reflected light formed therein is acquired to inspect the substrate to be inspected on the basis of the reflection signal. The reflection signal of a prepared test substrate is repeatedly acquired (S2) while changing over the wavelength of the illumination light (S1, S3 and S4) prior to the inspection of the substrate to be inspected and the inspection wavelength of the substrate to be inspected is determined on the basis of a plurality of the acquired reflection signals (S5). <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a surface inspection method and a surface inspection apparatus for inspecting the surface of a semiconductor wafer or a liquid crystal substrate.

One surface inspection device such as a semiconductor wafer or a liquid crystal substrate illuminates the substrate surface, and the image of the substrate is diffracted by the repetitive pattern (line and space pattern such as a wiring pattern) provided there. The one that is formed and automatically detects defects on the substrate from the image has been put into practical use (see Patent Document 1).
In addition, as an apparatus to which this is applied, an apparatus in which illumination light is previously polarized and an image of a substrate is formed by a specific polarization component of specularly reflected light generated in a repetitive pattern has been proposed (Patent Document 2 and the like). reference). This apparatus utilizes the phenomenon that the polarization state of specularly reflected light changes due to the structural birefringence of the repetitive pattern, and the amount of change depends on the shape of the repetitive pattern. Such an apparatus has an advantage that inspection can be performed even if the wavelength of illumination light is longer than the pitch of the repetitive pattern.
Japanese Patent Laid-Open No. 10-232122 International Publication No. 2005 / 040776A1 Pamphlet

However, it has been found that the detection accuracy and detection sensitivity of defects depend on the wavelength of illumination light and the type of inspection object, and satisfactory inspection is not necessarily performed.
Therefore, an object of the present invention is to provide a surface inspection method capable of surely performing a good inspection and a surface inspection apparatus suitable for the surface inspection method.

  The surface inspection method of the present invention illuminates a pattern on a test substrate with polarized illumination light, acquires a reflection signal indicating the polarization state of specular reflection light generated there, and based on the reflection signal, the test substrate A surface inspection method for inspecting a plurality of reflections obtained by repeatedly acquiring the reflection signal of a test substrate prepared in advance while switching the wavelength of the illumination light prior to inspection of the substrate to be tested. A wavelength for inspection of the test substrate is determined based on the signal.

  The surface inspection apparatus of the present invention illuminates a pattern on a test substrate with polarized illumination light, and generates a reflected signal indicating the polarization state of specularly reflected light generated there from the measurement optical system. A controller that acquires the reflected signal of the test substrate and inspects the test substrate based on the reflected signal, and the measurement optical system is capable of switching the wavelength of the illumination light, Prior to the inspection of the test substrate, the control unit performs test measurement by repeatedly acquiring the reflected signal of the test substrate prepared in advance by the measurement optical system while switching the wavelength of the illumination light. To do.

The control unit may determine a wavelength for inspecting the substrate to be inspected based on a plurality of reflection signals acquired by the test measurement.
Further, the control unit may present information of a plurality of reflected signals acquired by the test measurement to an inspector, and cause the inspector to determine a wavelength for inspecting the substrate to be inspected.
Further, the control unit may store information on the determined wavelength for inspection.

The controller sets the wavelength of illumination light of the measurement optical system to the determined wavelength for inspection when inspecting the substrate to be inspected.
Further, on the test substrate, a plurality of test patterns having the same shape as the pattern of the substrate to be inspected and having different defect amounts are formed, and the wavelength for inspection has the plurality of test patterns. It is desirable that the wavelength be determined so as to produce the largest difference between the reflected signals.

  The test substrate and the test substrate are resist layers that have undergone exposure and development of the pattern, and the plurality of test patterns are a plurality of test patterns having different defocus amounts during the exposure. It is desirable to be.

  ADVANTAGE OF THE INVENTION According to this invention, the surface inspection method which can perform a favorable test | inspection reliably, and a surface inspection apparatus suitable for it are implement | achieved.

[First Embodiment]
A first embodiment of the present invention will be described. The present embodiment is an embodiment of a surface inspection apparatus and a surface inspection method using the same.
First, the configuration of the surface inspection apparatus will be described.
As shown in FIG. 1, the surface inspection apparatus 10 includes a stage 11 that supports a wafer 20 to be tested, an alignment system 12, an illumination system 13, a light receiving system 14, a control arithmetic device 15, an image display device 16, an input device 17, and the like. Is provided.

  The test wafer 20 is a semiconductor wafer whose uppermost layer is a resist film that has been exposed and developed, and a pattern (described later) is formed on the surface. The test wafer 20 is carried from a wafer cassette (not shown) or a developing device by a transfer system (not shown) and fixedly held on the stage 11. The test wafer 20 can be rotated by the stage 11 around the normal line 1A at the center of the upper surface as a rotation axis. The rotation position of the wafer 20 under rotation is detected by the alignment system 12 from the shape of notches and orientation flats provided on the outer edge. The alignment system 12 and the stage 11 set the rotation position of the wafer 20 to be tested to a position (described later) suitable for surface inspection.

The illumination system 13 is a decentered optical system in which a light source 31, a wavelength selection unit 32, a light guide fiber 33, a polarizing filter 34, and a concave reflecting mirror 35 are arranged in this order. The light source 31 is a discharge light source such as a mercury lamp or a metal halide lamp, and the wavelength selection unit 32 is a mechanism for switching the wavelength of light traveling from the light source 31 to the light guide fiber 33.
The wavelength selection unit 32 includes a collimator lens 25 that converts light from the light source 31 into a parallel light beam, and a plurality of wavelength selection filters F1, F2, F3, F4, F5, and F6 having different transmission wavelength ranges with respect to the parallel light beam. A turret 12b that is selectively inserted, a condensing lens 26 that condenses the parallel light flux that has passed through the turret 12b onto the incident end of the light guide fiber 33, and a motor 23c that rotates the turret 12b are provided. When the motor 23c is driven, the wavelength of light incident on the incident end of the light guide fiber 33 is switched. The emission line spectrum of the light source 31 and the transmission wavelength ranges of the wavelength selective filter filters F1, F2, F3, F4, and F5 will be described later.

  Light incident on the incident end of the light guide fiber 33 propagates through the light guide fiber 33 and exits from the exit end. The light is converted into linearly polarized light by a polarizing filter 34 disposed in the vicinity of the exit end, and then the substantially entire surface of the test wafer 20 is illuminated from an oblique direction via a concave reflecting mirror 35. The exit optical axis O1 of the concave reflecting mirror 35 passes through the center of the stage 11 and is inclined by a predetermined angle θ with respect to the normal line 1A of the stage 11.

  The concave reflecting mirror 35 is a reflecting mirror whose inner surface is a reflecting surface, and its front focal point substantially coincides with the exit end of the light guide fiber 33 and its rear focal point substantially coincides with the surface of the wafer 20 to be examined. . By this concave reflecting mirror 35, the entire surface of the wafer 20 to be tested is illuminated with telecentric linearly polarized light L1. The principal ray (traveling direction) of each linearly polarized light L1 directed to each point on the test wafer 20 is substantially parallel to the exit optical axis O1 of the concave reflecting mirror 35. The direction of the transmission axis of the polarizing filter 34 described above is set in advance so that the polarization direction of the linearly polarized light L1 is p-polarized with respect to the surface of the wafer 20 to be tested.

  The light receiving system 14 is a decentered optical system in which a concave reflecting mirror 36, a polarizing filter 38, an imaging lens 37, and an image sensor 39 are sequentially arranged. The concave reflecting mirror 36 is a reflecting mirror similar to the concave reflecting mirror 35 of the illumination system 13, and its incident optical axis O2 is on the same plane as the normal 1A and the outgoing optical axis O1 of the concave reflecting mirror 35, and The angle formed by the incident optical axis O2 and the normal line 1A is the same as the angle formed by the outgoing optical axis O1 and the normal line 1A. Therefore, the regular reflection light L2 from the test wafer 20 travels along the incident optical axis O2 of the concave reflecting mirror 36. The concave reflecting mirror 36 reflects the specularly reflected light L <b> 2 and guides it toward the polarizing filter 38, and forms a reflected image of the test wafer 20 on the imaging surface of the imaging device 39 in cooperation with the imaging lens 37. .

  Here, the direction of the transmission axis of the polarizing filter 38 is set so as to be orthogonal to the transmission axis of the polarizing filter 34 in the illumination system 13 (in a crossed Nicols state). Therefore, the linearly polarized light L4 that has passed through the polarizing filter 38 corresponds to a polarization component that is orthogonal to the polarization direction of the linearly polarized light L1 in the regular reflected light L2. The intensity of the linearly polarized light L4 determines the intensity of the reflected image described above.

  This reflected image is captured by the image sensor 39. The image sensor 39 is, for example, a CCD image sensor or the like, photoelectrically converts the reflected image to generate an electric signal, and sends the electric signal to the control arithmetic unit 15 including a circuit, a computer, and the like. When the control arithmetic unit 15 recognizes the reflection image of the test wafer 20 based on the electrical signal, the control arithmetic unit 15 determines the degree of pattern defects on the test wafer 20 based on the brightness of each region on the reflection image as a shot region (described later). ) Judge every time. The control arithmetic device 15 has a control function for controlling each part of the surface inspection device 10 in addition to such an arithmetic function.

Next, the rotation position of the test wafer 20 suitable for surface inspection will be described.
As shown in FIG. 2, a plurality of shot areas 21 are arranged in the XY direction on the surface of the test wafer 20, and a common pattern is formed in each shot area 21. FIG. 2 shows an example in which only one type of repetitive pattern 22 is formed in each shot area 21. The repeated pattern 22 is a resist pattern (for example, a wiring pattern) in which a plurality of line portions are arranged at a constant pitch along the short direction (X direction). In this specification, the pitch direction (X direction in FIG. 2) is referred to as “repetition direction”.

  The rotation position suitable for the surface inspection is such that the repetition direction (X direction in FIG. 2) and the incident surface 3A of the linearly polarized light L1 that illuminates the test wafer 20 are other than 0 degrees on the surface of the test wafer 20. The rotational position forms a predetermined angle (preferably 45 degrees). The incident surface 3A includes a traveling direction (principal ray) of the linearly polarized light L1 and is a plane parallel to the normal line 1A (the incident surface 3A shown in FIG. 2 is the linearly polarized light L1 incident on the center of the test wafer 20). Incident surface). Incidentally, since the linearly polarized light L1 is p-polarized light, the incident surface 3A of the linearly polarized light L1 coincides with the vibration surface of the linearly polarized light L1.

  When the linearly polarized light L1 is incident on the repetitive pattern 22, it is affected by the structural birefringence of the repetitive pattern 22. At this time, the specularly reflected light L2 generated by the repeated pattern 22 is not linearly polarized light but elliptically polarized light. Therefore, the intensity of the linearly polarized light L4 after passing through the polarizing filter 38 of the light receiving system 14, that is, the intensity of the reflected image represents the degree of ovalization.

  If there was no defocus defect (shape error due to excessive or insufficient defocus amount during exposure) in the repetitive pattern 22, the intensity of the reflected image (reflected image intensity) was the highest and a defocus defect occurred. Sometimes, the reflected image intensity decreases according to the degree of the defocus defect (defocus defect amount). Such a relationship between the defocus amount and the reflected image intensity ideally draws a saddle-shaped curve as shown by a thick solid line in FIG. Hereinafter, a curve indicating the relationship between the defocus amount and the reflected image intensity is referred to as a “sensitivity curve”.

  Therefore, the control arithmetic unit 15 of the surface inspection apparatus 10 refers to the brightness value (average brightness value in the area) of each shot area image (shot area image) shown in the reflected image, and the brightness value is less than the threshold value. It may be determined that a low shot area is “defect” and a shot area whose luminance value is higher than a threshold value is “no defect”. In particular, as shown in FIG. 2, when the incident surface 3A and the repeating direction form an angle of 45 degrees, the slope of the sensitivity curve (FIG. 3) is the steepest and the sensitivity of the surface inspection is the highest.

  However, the sensitivity curve is improved in this way only when the wavelength (illumination wavelength) of the linearly polarized light L1 is optimally set for the design shape of the repeated pattern 22. For example, if the illumination wavelength is too short, extra diffracted light is generated in the repetitive pattern 22, and therefore periodic noise is superimposed on the sensitivity curve as shown by a fine dotted line in FIG. On the other hand, if the illumination wavelength is too long, the degree of ovalization due to structural birefringence is reduced, so that the slope of the sensitivity curve becomes gradual as shown by the rough dotted line in FIG.

  Further, the design shape of the repetitive pattern 22 varies depending on the type of chip (product) that can be cut out from the shot region 21 and the processes that have been performed up to the surface inspection. Even if the product type is the same, when the product includes a plurality of types of repetitive patterns (such as a logic circuit), it varies depending on the position in the shot area. For this reason, the illumination wavelength needs to be set for each repeated pattern to be subjected to surface inspection. Therefore, in the surface inspection apparatus 10, the inspection wavelength is set (conditions are set) prior to the surface inspection. At that time, the wavelength selection unit 32 of the surface inspection apparatus 10 is effectively used.

Next, the wavelength selection unit 32 will be described in detail.
Here, a case where a mercury lamp is used as the light source 31 of the wavelength selection unit 32 will be described. In this case, as shown in FIG. 4, the emission line spectrum of the light source 31 has a wavelength λ ₁ = 248 nm, a wavelength λ ₂ = 313 nm, a wavelength λ ₃ = 365 nm, a wavelength λ ₄ = 405 nm, a wavelength λ ₅ = 436 nm, and a wavelength λ ₆ = Each wavelength component is 546 nm.

Accordingly, the transmission wavelength region of the wavelength selection filter F1 is set to the wavelength λ ₁ , the transmission wavelength region of the wavelength selection filter F2 is set to the wavelength λ _2, and the transmission wavelength region of the wavelength selection filter F3 is the wavelength λ ₃ , the transmission wavelength band of the wavelength selection filter F4 is set to the wavelength λ ₄ , the transmission wavelength band of the wavelength selection filter F5 is set to the wavelength λ _5, and the transmission wavelength band of the wavelength selection filter F6 is the wavelength λ ₆ is set.

Therefore, when the motor 23c of the wavelength selection unit 32 is driven by the control arithmetic unit 15, the illumination wavelengths of the surface inspection apparatus 10 are six types of wavelengths λ ₁ , λ ₂ , λ ₃ , λ ₄ , λ ₅ , λ _6. Switch between.
The rotational position of the turret 12b of the wavelength selection unit 32 is detected by an absolute rotary encoder (not shown) or the like, and the detection signal is given to the control arithmetic device 15. Therefore, the control arithmetic unit 15 can recognize the illumination wavelength being set from the detection signal.

Next, a method for setting (conditioning) the inspection wavelength will be described.
Here, a case will be described in which a surface inspection regarding a defocus defect is assumed and an illumination wavelength optimal for the surface inspection is set as the inspection wavelength. For simplicity, it is assumed that only one type of repetitive pattern 22 is formed in each shot region 21 of the wafer 20 to be tested, as shown in FIG.

First, the inspector prepares the test wafer 20 _T as shown in FIG. The test wafer 20 _T is a semiconductor wafer that has undergone the same process as the test wafer 20, and a plurality of shot regions 21 _T are two-dimensionally arranged on the surface, like the test wafer 20. In each shot area 21 _T , the same test pattern 22 _T as the repetitive pattern 22 of the wafer to be tested 20 is formed. However, each shot area 21 _T test wafer 20 _T is different exposure conditions (here, a combination of the defocus amount and dose.) To each other and is formed with.

In the example shown in FIG. 5, the defocus amount differs by one step in order from the left between the shot regions arranged in the left-right direction, and the defocus amount is zero in the shot region located in the center. In the example shown in FIG. 5, the dose amount (exposure amount at the time of exposure) differs sequentially from the bottom between shot regions arranged in the vertical direction, and the shot region located in the center of the shot regions has a dose amount. Is optimal (ie, the dose defect amount is zero). Accordingly, the shot regions 21 _T-2 , 21 _T−1 , 21 _T0 , 21 _{T + 1} , 21 _{T + 2} (shot region group G) aligned in the horizontal direction at the center in the vertical direction all have zero dose defects. Yes, and the defocus amount is different from each other by one step.

Here, the defocus amounts of the shot areas 21 _T-2 , 21 _T-1 , 21 _T0 , 21 _{T + 1} , and 21 _{T + 2} are set to −2, −1, ± 0, +1, and +2, respectively. However, this numerical value does not represent the absolute value of the defocus amount, but represents the stage of the defocus amount. “−2” is the maximum value on the minus side of the defocus amount that can occur during exposure, and “+2” is the maximum value on the plus side of the defocus amount that can occur during exposure.

In this embodiment, since surface inspection related to defocus defects is assumed, this shot region group G (shot regions 21 _T-2 , 21 _T-1 , 21 _T0 , 21 _{T + 1} , 21 _{T + 2} ) is used. (The other shot areas are used in another inspection.) If the reflection image of the test wafer 20 _T is acquired by the surface inspection apparatus 10, the shot regions 21 _T-2 , 21 _T−1 , 21 _T0 , 21 _{T + 1} , 21 _{T + 2} are included in the reflection image. The brightness of the reflected image becomes stepwise, and the approximate shape of the sensitivity curve (FIG. 3) described above can be estimated.

Note that the test wafer information such as the coordinates, defocus amount, and dose amount of each shot region 21 _T of the test wafer 20 _{T described} above is known in advance. Incidentally, in order to make the defocus amount known, measurement using a microscope is effective.
Examiner, the more test wafers 20T, while set to the test wafer 20 similarly to the surface inspection device 10 (FIG. 1), a test wafer information about the test wafer 20 _T, the control operation through the input device 17 Input to the device 15. The test wafer information may not include information on shot areas that are not used in the present embodiment.

Further, the inspector gives an instruction to start the condition setting to the control arithmetic device 15 of the surface inspection apparatus 10 via the input device 17. In accordance with the instruction, the control arithmetic device 15 operates as follows.
FIG. 6 is an operation flowchart of the control arithmetic device 15 relating to the setting of the inspection wavelength. As shown in FIG. 6, the control calculation unit 15 first sets the illumination wavelength of the surface inspection apparatus 10 to lambda ₁ (step S1), the drive surface inspection apparatus 10 in this state, the test wafer 20 _T reflecting image Is acquired (step S2). Then, the control arithmetic unit 15 switches the illumination wavelength of the surface inspection apparatus 10 to lambda ₂ (step S4), and similarly obtains the reflection image of the test wafer 20 _T (Step S2). Further, the control arithmetic unit 15 acquires a reflection image of the test wafer 20 _T even in each state in which the illumination wavelengths are set to λ ₃ , λ ₄ , λ ₅ , and λ ₆ (step S 4 → S 2 → S 3 loop). When the acquisition of the reflection image for all wavelengths is completed (step S3 yes), the control arithmetic device 15 determines the inspection wavelength (step S5), and registers the determined inspection wavelength in the inspection recipe (step S6). Among these, the process of step S1-S4 is test measurement.

Next, determination of the inspection wavelength (S5) and registration of the inspection wavelength (S6) will be described in detail.
Determination of inspection wavelength (S5):
FIG. 7 is a conceptual diagram of the reflection images I ₁ , I ₂ ,..., I ₆ acquired by the test measurement. The reflection image I _i is a reflection image acquired under the wavelength λ _i . Reflection image I _1, I _2, ..., the I ₆ is photographed statue of shot areas 21 _T described above (shot area image) 21 _T '. Reference numeral G ′ represents an image of the above-described shot region group G (shot region group image). At this step S5, such a reflection image I _1, I _2, ..., individually evaluated value from each of the _{I 6 E 1, E 2,} ..., and calculates the E _6. The evaluation value E _i is an evaluation value related to the reflected image I _i and is an evaluation value of a sensitivity curve obtained under the wavelength λ _i .

In the calculation of the evaluation value E ₁ , first, as shown in FIG. 8, the control arithmetic unit 15 selects the shot region image 21 _T0 ′ corresponding to zero defocus amount among the shot region group images G ′ on the reflected image I ₁ . Reference is made to the luminance value a ₀ and the luminance values a ₋₂ and a _{+2 of the} two shot area images 21 _T−2 ′ and 21 _{T + 2} ′ corresponding to the maximum defocus amount (± 2). The luminance value a _j is an average luminance value of the shot area image 21 _Tj .

Next, the control arithmetic device 15 calculates the evaluation value E ₁ by the following equation (1) based on the referenced luminance values a ₋₂ , a ₀ , a ₊₂ , for example.
E ₁ = (a ₀ −a ₋₂ ) + (a ₀ −a ₊₂ ) (1)
The first term (a ₀ −a ₋₂ ) on the right side of Equation (1) simply represents the amount of inclination on the negative side of the sensitivity curve (FIG. 3) obtained under the wavelength λ ₁ . The term (a ₀ −a ₊₂ ) simply represents the amount of inclination on the plus side of the sensitivity curve (FIG. 3) obtained under the wavelength λ ₁ . Accordingly, if the slope of the sensitivity curve (FIG. 3) obtained under the wavelength λ ₁ is steep, the first term (a ₀ −a ₋₂ ) and the second term (a ₀ −a ₊₂ ) are Both take large values. On the other hand, if the slope of the sensitivity curve (FIG. 3) obtained under the wavelength λ ₁ is gentle, both the first term (a ₀ −a ₋₂ ) and the second term (a ₀ −a ₊₂ ) are small. Take the value. Therefore, if this evaluation value E ₁ is large, it can be determined that a good sensitivity curve can be obtained by the wavelength λ ₁ .

The calculation method described above, other wavelengths _{λ 2, λ 3, λ 4} , ..., the evaluation value of _{λ 6 E 2, E 3,} E 4, ..., is calculated in the same manner for E _6. Then, the control device 15 selects the largest value among the calculated evaluation values E ₁ , E ₂ , E ₃ , E ₄ ,..., E ₆ and determines the corresponding wavelength as the inspection wavelength. To do. Therefore, the inspection wavelength is determined to be a wavelength that provides the best sensitivity curve.

Inspection wavelength registration (S6)
The control arithmetic unit 15 registers the inspection wavelength information in the inspection recipe of the wafer 20 to be inspected. The inspection recipe is stored in the storage unit or the like of the control arithmetic unit 15, and for example, as shown in FIG. 9, the item of the wafer ID of the wafer 20 to be tested, the item of the type of the wafer to be tested 20, There are items of directionality of the repeated pattern 22 of the wafer 20 and items of illumination light quantity (inspection light quantity) suitable for surface inspection of the wafer 20 to be examined. In this embodiment, in addition to these items, a dotted line The item of the inspection wavelength is provided as enclosed in. When the registration of the inspection wavelength is completed, the control arithmetic device 15 displays an image of the registered inspection recipe on the image display device 16. For example, when the determined inspection wavelength is λ ₄ , a numerical image such as “405 nm” is displayed on the display screen. This completes the determination (conditioning) of the inspection wavelength.

Thereafter, when the inspector sets the wafer 20 to be inspected in place of the test wafer 20 _T to the surface inspection apparatus 10 and inputs an instruction to start inspection to the control arithmetic apparatus 15 via the input unit 17, the control arithmetic apparatus 15 Then, the surface inspection of the test wafer 20 is performed according to the contents of the inspection recipe described above. For example, when the inspection wavelength registered in the inspection recipe is λ ₄ (405 nm), the control arithmetic unit 15 sets the illumination wavelength of the surface inspection apparatus 10 to λ ₄ (405 nm) and then the wafer 20 to be inspected. Perform surface inspection.

Above, the control arithmetic unit 15 of the present embodiment, the reflected image of a prepared test wafer 20 _T, the wavelength lambda ₁ to the illumination wavelength, lambda _2, lambda _3, ..., repeatedly acquires while switching between the lambda ₆ test measurements (FIG. 6 step S1 to S4), the plurality of reflection obtained with test measurement image _{I 1, I 2, I 3} , ..., the I _6, the evaluation value E ₁ indicating the quality of sensitivity curves, When E ₂ , E ₃ ,..., E ₆ are extracted, the optimum wavelength that provides a good sensitivity curve is determined as the inspection wavelength based on them (step S5 in FIG. 6). Therefore, according to the present embodiment, the surface inspection of the test wafer 20 can be performed reliably and with high sensitivity regardless of the design shape of the repeated pattern 22.

In addition, since the control arithmetic unit 15 of the present embodiment registers the determined inspection wavelength together with the wafer ID in the inspection recipe (FIG. 9), the inspector can easily manage inspection information on a large number of wafers.
Further, the control arithmetic unit 15 of the present embodiment calculates the evaluation value E _{i by using} a plurality of shot region images 21 _T-2 ′, 21 _T0 ′21 _T with a dose defect amount of zero and different defocus amounts. _{Since +2} ′ (FIG. 8) is used, it is possible to reliably set an inspection wavelength suitable for defocus defect inspection.

Further, the control arithmetic device 15 of the present embodiment limits the shot area image used for calculating the evaluation value E _i to only three shot area images 21 _T−2 ′, 21 _T0 ′, and 21 _{T + 2} ′. The amount of calculation related to the calculation can be suppressed. Therefore, the inspection wavelength is set at high speed.
The control arithmetic unit 15 of the present embodiment limits the shot area image used for calculating the evaluation value E _i to only three shot area images 21 _T−2 ′, 21 _T0 ′, and 21 _{T + 2} ′. You may restrict _| limit only to two shot area | region images ( _21T-2 ', _21T0 ', or _{21T + 2} ', _21T0 '). Thus, if the number of shot area images to be used is further limited, the amount of calculation can be further suppressed. In this case, the following formula (2) or formula (3) is used instead of formula (1).

E ₁ = a ₀ −a ₋₂ (2)
E ₁ = a ₀ −a ₊₂ (3)
Further, in the surface inspection apparatus 10 of the present embodiment, six types of wavelength selection filters F1, F2,..., F6 are used according to the emission line spectrum of the mercury lamp, but the number is limited to six types. No. When the inspection wavelength is set in detail, it is desirable that the number of wavelength selection filters is large. When the inspection wavelength is set at high speed, it is desirable that the number of wavelength selection filters is small.

In the surface inspection apparatus 10 of the present embodiment, the mercury lamp is applied to the light source 31, but other discharge light sources such as a metal halide lamp and a halogen lamp may be applied. Each transmission wavelength region of the wavelength selection filters F1, F2,..., F6 is appropriately set according to the type of the applied radiation source.
Further, in the surface inspection apparatus 10 of the present embodiment, the switching of the plurality of wavelength selection filters is performed by the turret 12b (rotation mechanism), but may be performed by another mechanism (slide mechanism or the like). Furthermore, the wavelength may be switched by a method other than the switching of the wavelength selection filter.

Moreover, in the surface inspection apparatus 10 of this embodiment, the state of the elliptically polarized light of the regular reflection light L2 generated from the repetitive pattern 22 is slightly rotated around the traveling direction. For this reason, it is preferable to finely adjust the direction of the transmission axis of the polarizing filter 38 of the light receiving system 14 in consideration of the rotation angle.
In the surface inspection apparatus 10 of the present embodiment, the linearly polarized light L1 is p-polarized light, but the present invention is not limited to this. It may be s-polarized light instead of p-polarized light. Incidentally, the p-polarized light is advantageous for acquiring defect information related to the edge shape of the repetitive pattern 22, and the s-polarized light efficiently captures defect information on the surface of the wafer 20 to be tested and improves the SN ratio. Is advantageous.

  Further, not only p-polarized light and s-polarized light but also linearly polarized light whose vibration surface has an arbitrary inclination with respect to the incident surface may be used. In this case, the repeating direction (X direction) of the repeating pattern 22 is set to an angle other than 45 degrees with respect to the incident surface of the linearly polarized light L1, and the direction of the vibrating surface of the linearly polarized light L1 on the surface of the wafer 20 to be tested is repeated. It is preferable to set the angle formed by the repeating direction (X direction) of the pattern 22 to 45 degrees.

  In the surface inspection apparatus 10 of the present embodiment, the angle (inclination angle) formed by the direction of the vibration plane of the linearly polarized light L1 on the surface of the wafer 20 to be tested and the repeating direction (X direction) of the repeating pattern 22 is 45 degrees. Although set, the present invention is not limited to this. When the tilt angle is set to 45 ° ± 15 °, the luminance is reduced by about 70%, and when it is 45 ° ± 20 °, the luminance is reduced by about 55%. That is, if the inclination angle is in the range of 30 degrees to 60 degrees, a luminance value of 70% or more when set to 45 degrees can be secured, and thus the same surface inspection as described above can be sufficiently performed. Further, the inclination amount (for example, 50%) of the sensitivity curve is considered to be constant regardless of the inclination angle. Therefore, it is considered that the inclination amount of the sensitivity curve becomes smaller as the inclination angle deviates from 45 degrees. However, if the inclination angle is in the range of 30 degrees to 60 degrees, sufficient inspection can be performed. In actual operation, it is more preferable that the tilt angle is in the range of 35 to 55 degrees.

  In the surface inspection apparatus 10 of the present embodiment, the concave reflecting mirrors 35 and 36 are used to illuminate the test wafer 20 from an oblique direction and receive specularly reflected light emitted from the test wafer 20 in the oblique direction. However, even if a refracting lens having the same function is used to illuminate the wafer 20 to be tested from the front (that is, θ = 0 degrees) and receive the specularly reflected light emitted from the wafer 20 to the front. Good. In that case, many parts of the illumination system and the light receiving system may be shared, and a polarization beam splitter may be used to separate the illumination light and the regular reflection light. This polarizing beam splitter serves as both a polarizing filter for the illumination system and a polarizing filter for the light receiving system.

  In the surface inspection of the present embodiment, the test wafer 20 is the test substrate, but the present invention is not limited to this. The present invention can also be applied when performing a defect inspection of a liquid crystal substrate (substrate to be tested) in the manufacturing process of the liquid crystal display element. Further, a general-purpose computer may be used instead of the dedicated control arithmetic device 15 for the surface inspection apparatus 10. In addition, a part of the operation of the control arithmetic device 15 may be manually performed by an inspector.

  In the surface inspection apparatus 10 of the present embodiment, a two-dimensional sensor such as a CCD is used as the image sensor 39, but a one-dimensional sensor may be used. In this case, the one-dimensional sensor that is the image sensor and the stage on which the test wafer (or liquid crystal substrate) that is the test substrate is moved relative to each other, and the one-dimensional sensor moves the entire surface of the test wafer (or liquid crystal substrate). Scanning may be performed to capture an image of the entire surface of the test wafer (or liquid crystal substrate).

[Second Embodiment]
A second embodiment of the present invention will be described. This embodiment is an embodiment of a surface inspection apparatus. Here, only differences from the first embodiment will be described. The difference is, the control arithmetic unit 15 evaluation value E _1, E ₂ by, ..., in the method of calculating the E _6. Here, a case where eleven types of shot regions 21 _T are arranged in the shot region G of the test wafer 10 _T will be described.

FIG. 10 is a conceptual diagram of the reflection images I ₁ , I ₂ ,..., I ₆ acquired by the test measurement. In FIG. 10, the symbol G ′ is an image of the shot region group G (shot region group image). ', The 11 kinds of shot area image 21 _T-5' shot area unit image _{G, ..., 21 T0 ',} ..., 21 T + 5' are arranged. The shot area images 21 _T-5 ′,..., 21 _T0 ′,..., 21 _{T + 5} ′ are all images of 11 shot areas having zero dose defects and different defocus amounts by one step. It is. Here, the shot area image _{21 T-5 ', ...,} 21 T0', ..., the defocus amount of 21 _{T + 5} ', respectively -5, -4, -3, -2, -1, ± 0, +1, +2, +3, +4, +5. However, this numerical value does not represent the absolute value of the defocus amount, but represents the stage of the defocus amount. “−5” is the maximum value on the minus side of the defocus amount that can occur during exposure, and “+5” is the maximum value on the plus side of the defocus amount that can occur during exposure.

Now, in the calculation of the evaluation value E ₁ of the wavelength lambda _1, the control arithmetic unit 15, first, as shown in FIG. 10, the shot area image _{21 T-5 ', ...,} 21 T0', ..., 21 T + 5 ' each of the luminance value a _-5 a, ..., a _0, ..., referring to a _+5. The luminance value a _j is an average luminance value of the shot area image 21 _Tj ′.
Based on the referenced luminance values a ₋₅ ,..., A ₀ ,..., A ₊₅ , the control arithmetic device 15 calculates the evaluation value E ₁ by the following equation (4), for example.

E ₁ = (a ₀ −a ₋₁ ) + (a ₋₁ −a ₋₂ ) +... + (A ₋₄ −a ₋₅ )
+ (A ₀ −a ₊₁ ) + (a ₊₁ −a ₊₂ ) +... + (A ₊₄ −a ₊₅ ) (4)
Each term in the first half of the right side of Equation (4) represents the amount of inclination of each position on the minus side of the sensitivity curve (FIG. 3) obtained under the wavelength λ ₁ , and each term in the latter half represents the wavelength λ ₁ . The inclination amount of each position on the plus side of the sensitivity curve (FIG. 3) obtained below is shown.

Thus, if the slope of the sensitivity curve (FIG. 3) obtained under the wavelength λ ₁ is steep, each of the first half term and each second half term should take a large value. On the other hand, if the slope of the sensitivity curve (FIG. 3) is gentle, each term in the first half and each term in the second half should take a small value. If periodic noise is superimposed on the sensitivity curve (FIG. 3), some terms take negative values. Therefore, if this evaluation value E ₁ is large, it can be determined that a good sensitivity curve can be obtained by the wavelength λ ₁ .

The calculation method described above, the control arithmetic unit 15, the other wavelengths _{λ 2, λ 3, λ 4} , ..., the evaluation value of _{λ 6 E 2, E 3,} E 4, ..., is calculated in the same manner for E ₆ . In the present embodiment, these evaluation values are used to determine the inspection wavelength.
As described above, the evaluation value E _i of this embodiment reflects not the approximate amount of inclination of the sensitivity curve (FIG. 3) but the amount of inclination of each position of the sensitivity curve (FIG. 3). In addition, the quality of the sensitivity curve can be evaluated more reliably. Therefore, according to the present embodiment, the determination accuracy of the inspection wavelength is further increased.

[Third Embodiment]
A third embodiment of the present invention will be described. This embodiment is an embodiment of a surface inspection apparatus. Here, only differences from the first embodiment will be described. The difference is in the operation of the control arithmetic unit 15 after the end of the test measurement, and the inspection wavelength is determined by the inspector. Here, a case where the same test wafer as that of the second embodiment is used will be described.

FIG. 11 is an operation flowchart of the control arithmetic device 15 relating to the setting of the inspection wavelength of the present embodiment. As shown in FIG. 11, after completion of the test measurement, the processing device 15, the reflection image I _1, I ₂ obtained in test measurements, ..., each wavelength lambda ₁ from I _6, lambda _2, ..., the sensitivity of the lambda ₆ Curves C ₁ , C ₂ ,..., C ₆ are extracted, and their sensitivity curves C ₁ , C ₂ ,..., C ₆ are displayed on the image display device 16 (step S5A). Further, the control arithmetic device 15 causes the inspector to select one of the wavelengths λ ₁ ,..., Λ ₆ on the image display device 16 (step S5Byes). Then, the wavelength selected by the inspector is registered as an inspection wavelength in the inspection recipe as in the first embodiment (step S6).

Hereinafter, the processing after step S5A will be described in detail.
Figure 12 is a diagram illustrating processing (step S5A) for extracting a sensitivity curve C ₁ from the reflection image I _1. In this step, the control arithmetic unit 15 determines the brightness of each of the shot area images 21 _T-5 ′,..., 21 _T0 ′,..., 21 _{T + 5} ′ in the reflected image I ₁ shown in FIG. Reference values a _-5 ,..., A _{0 ,.} The luminance value a _j is an average luminance value of the shot area image 21 _Tj ′. Then, as shown in FIG. 12B, the control arithmetic unit 15 displays an image of a curve (a broken line) representing the luminance values a ₋₅ ,..., A _{0 ,.} Created and used as an image of the sensitivity curve C ₁ of wavelength λ ₁ . In FIG. 12B, the horizontal axis indicates the defocus amount, and the vertical axis indicates the luminance value.

Further, similarly to the create this image, the control arithmetic unit 15, the reflection image I _2, ..., wavelength lambda ₂ on the basis of I _6, ..., the sensitivity curve C ₂ in lambda _6, ..., the image of the C ₆ create.
Thereafter, the control arithmetic device 15 displays the created sensitivity curves C ₁ ,..., C ₆ side by side on the image display device 16 as shown in FIG. At this time, on the image display device 16, the sensitivity curve C _1, ..., with the image of C _6, and character image 61, such as a "wavelength selection screen", and it is also arranged a cursor 62. These images prompt the examiner to select any sensitivity curve on the image display device 16.

Here, as shown in FIG. 13, the sensitivity curve varies depending on the wavelength. In FIG. 13, the sensitivity curves C ₄ , C ₅ , C ₆ of the relatively long wavelengths λ ₄ , λ ₅ , λ ₆ are too gentle. Further, shorter wavelengths lambda _1, the sensitivity curve C ₁ of lambda _2, C ₂ is wavy. The sensitivity curve C ₃ having a wavelength λ ₃ between the two is steep and has a good shape.
Therefore, the examiner can immediately find the best sensitivity curve (sensitivity curve C _{3 in} FIG. 13) from the sensitivity curves of all wavelengths on the image display device 16. When the operator operates the input device 17 and selects the sensitivity curve C ₃ on the image display device 16, the control arithmetic device 15 switches the display screen to a confirmation screen (not shown) and the corresponding wavelength λ ₃ (here, And 365 nm), the inspector confirms whether or not the wavelength λ ₃ (365 nm) is the inspection wavelength, and is registered in the inspection recipe as in the first embodiment.

Above, in this embodiment, the inspector makes a decision inspection wavelength, the wavelength lambda ₁ that time, ..., sensitivity curve C ₁ in lambda _6, ..., the image of C ₆ is displayed on the image display device 16 Thus, the examiner can make the determination easily and immediately.
[Others]
The control arithmetic unit 15 of the first embodiment and the second embodiment, in Formula (1) as a calculation expression evaluation value E _i, (2), (3), was used either (4), Other evaluation formulas may be used as long as an evaluation value indicating the quality of the sensitivity curve (FIG. 3) can be obtained. Further, for example, the control arithmetic unit 15 may calculate a sensitivity curve by approximating the luminance values of a plurality of shot area images as a function, and may perform evaluation based on how the inclination of the sensitivity curve changes.

  In the above-described embodiment, the case where only one type of repetitive pattern is formed in the shot area 21 has been described. For example, as shown in FIG. 14, different partial areas A and B of the shot area 21 are different from each other. In addition, when different types of patterns are formed, the optimal inspection wavelength for the partial region A and the optimal inspection wavelength for the partial region B may be different. This is performed for each of the partial areas A and B. In that case, the control arithmetic unit 15 performs each process performed for each shot area in the above-described embodiment for each partial area. However, the test measurement described above need not be performed for each partial region, and may be performed for each wafer.

  The distribution information of the partial areas A and B in the shot area 21 is input to the control arithmetic device 15 in advance by the inspector together with other information. Incidentally, the inspection recipe created in this case is as shown in FIG. 15, for example. In FIG. 15, the portion surrounded by the dotted line is an item relating to the inspection wavelength, and it can be seen that the inspection wavelength is registered for each partial region.

It is a figure showing the whole surface inspection device 10 composition of a 1st embodiment. It is a figure explaining the inclination state of the incident surface (3A) of the linearly polarized light L1 and the repeating direction (X direction) of the repeating pattern 22. FIG. It is a figure (sensitivity curve) which shows the relationship between a defocus amount and a reflected image intensity | strength. It is a figure which shows an example of the bright line spectrum of the light source. It illustrates a test wafer 20 _T. It is an operation | movement flowchart of the control arithmetic unit 15 regarding the setting of the test | inspection wavelength of 1st Embodiment. It is a conceptual diagram of reflection images I ₁ , I ₂ ,..., I ₆ acquired by test measurement. The method of calculating the evaluation value E ₁ is a diagram illustrating a. It is a figure which shows an example of an inspection recipe. FIG. 7 is a conceptual diagram of reflection images I ₁ , I ₂ ,..., I ₆ acquired by test measurement of the second embodiment. It is an operation | movement flowchart of the control arithmetic unit 15 regarding the setting of the test | inspection wavelength of 3rd Embodiment. The process of extracting from the reflection image I ₁ the sensitivity curve C ₁ wavelength lambda ₁ is a diagram illustrating a (step S5A). It is a figure which shows a wavelength selection screen. FIG. 3 is a diagram for explaining different partial areas A and B in a shot area 21; It is a figure which shows another example of an inspection recipe.

Explanation of symbols

10: surface inspection device, 11: stage, 12: alignment system, 13: illumination system, 14: light receiving system, 15: control arithmetic device, 20: wafer to be tested, 21: shot area, 22: repetitive pattern, 31: light source , F1 to F6: wavelength selection filter, 33: light guide fiber, 34, 38: polarization filter, 35, 36: concave reflecting mirror, 37: imaging lens, 39: imaging device

Claims (8)

A surface inspection method for illuminating a pattern on a test substrate with polarized illumination light, obtaining a reflection signal indicating the polarization state of specular reflection light generated there, and inspecting the test substrate based on the reflection signal There,
Prior to the inspection of the test substrate,
Repetitively acquiring the reflected signal of a test substrate prepared in advance while switching the wavelength of the illumination light, and determining a wavelength for inspecting the test substrate based on the plurality of acquired reflected signals. Surface inspection method. A measurement optical system that illuminates the pattern on the test substrate with polarized illumination light and generates a reflected signal indicating the polarization state of the specularly reflected light generated there, and
A controller that acquires the reflected signal of the test substrate from the measurement optical system, and inspects the test substrate based on the reflected signal;
The measuring optical system is
It is possible to switch the wavelength of the illumination light,
The controller is
Prior to the inspection of the substrate to be inspected, the surface inspection apparatus is characterized in that the measurement optical system repeatedly acquires the reflected signal of the test substrate prepared in advance while switching the wavelength of the illumination light. In the surface inspection apparatus according to claim 2,
The controller is
A surface inspection apparatus for determining a wavelength for inspecting the substrate to be inspected based on a plurality of reflection signals acquired by the test measurement. In the surface inspection apparatus according to claim 2,
The controller is
A surface inspection apparatus characterized in that information on a plurality of reflected signals acquired by the test measurement is presented to an inspector, and the inspector determines a wavelength for inspecting the substrate to be inspected. In the surface inspection apparatus according to claim 3 or claim 4,
The controller is
Information on the determined wavelength for inspection is stored. In the surface inspection apparatus according to any one of claims 3 to 5,
The controller is
In the inspection of the substrate to be inspected, the wavelength of illumination light of the measurement optical system is set to the determined wavelength for inspection. In the surface inspection apparatus according to any one of claims 3 to 6,
On the test board,
A plurality of test patterns having the same shape as the pattern of the test substrate and having different defect amounts are formed,
The inspection wavelength is
The surface inspection apparatus, wherein the wavelength is determined such that the largest difference occurs between the reflected signals of each of the plurality of test patterns. The surface inspection apparatus according to claim 7,
On the test substrate and the test substrate,
It is a resist layer that has undergone exposure and development of the pattern,
The plurality of test patterns are:
A surface inspection apparatus comprising a plurality of test patterns having different defocus amounts at the time of exposure.

Images