JP2009064024A - Device for detecting physical quantity of observation object and detection method using the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device which detects, in a short period of time, a quantity of phase change of an observation object from a differential interference image obtained by a differential interference microscope. <P>SOLUTION: The device includes: a light source; the differential interference microscope having an illuminating optical system which splits light from the light source into two polarization components and guides them to the observation object and an image-forming system which recomposes the two polarization components into which the light has been split by the illuminating optical system, and forms an image of the observation object; a means for varying a quantity of retardation of each of the two polarization components; a means for detecting each quantity of retardation; and a means for picking up the image of the observation object, and the device uses an imaging means to pick up two differential interference images of the observation object, which are equal in the quantity of retardation and different in sign, wherein an operation means is provided. The operation means acquires subtraction image information and addition image information by performing a subtraction operation and addition operation for corresponding pixels in the two differential interference images, and detects a quantity of phase on the face of the observation object by using expression (1):δΦ(x,y)/δr=k×ä(1-cosθ)×D(x,y)}/ä2sinθ×S(x,y)}. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an apparatus for detecting physical quantities such as the shape of an observation object, the coordinates of each observation point of the observation object, steps, and phase changes from the image information of the observation object obtained by a differential interference microscope, and detection using the apparatus. Regarding the method.

  The differential interference microscope is widely used for observing a fine structure such as a living body or an IC pattern because it can visualize the phase change and step information of an observation object by polarization interference. In particular, inspection of minute protrusions (bumps) for preventing adhesion of magnetic heads formed on the surface of magnetic disks in recent years, measurement of defects and retardation (phase difference) in phase shift reticles used for semiconductor pattern exposure, semiconductors Attempts have been made to use a differential interference microscope in a wafer positioning device or the like.

For example, the following Patent Document 1 and Patent Document 2 disclose a method of applying the differential interference microscope to a defect detection and phase measurement of a phase shift reticle, assuming that the differential interference microscope is a shearing interferometer and a Mach-Zehnder interferometer. Patent Document 3 below discloses a method for positioning a semiconductor wafer by detecting an edge portion of a positioning mark on the semiconductor wafer using a differential interference microscope.

However, these methods apply a conventional interference measurement technique to a differential interference microscope, and do not consider the influence of light diffraction on the observation object surface. Further, no consideration is given to the influence on the light intensity change due to the change in the light reflectance or transmittance of the observation object.
Note that the effects of diffraction and intensity change of light on the observation object surface are obtained by the differential interference microscope by the present inventor in the following Patent Document 4 by clarifying the imaging characteristics of the differential interference microscope. A method to extract the phase information of the observed object from the captured image is presented.

JP-A-5-149719 JP 7-248261 A JP 7-239212 A JP-A-9-15504

The differential interference microscope converts the phase change on the observation object surface into a light and shade distribution in the image. On the contrary, it is considered that the phase change on the observation object plane can be detected by analyzing the distribution of shading in the differential interference image. Also, since the edge of the step of the observed object is accompanied by an abrupt phase change, the intensity value of the image also changes abruptly, so it is possible to observe by extracting the part where the intensity value changes abruptly from the differential interference image The above-mentioned patent document 3 shows that the position of the step of the object can be detected.
Further, Japanese Patent Application Laid-Open No. 5-256695 discloses a method of detecting a foreign matter mixed in an observation object by using a differential interference image of a normal specimen as a reference image and comparing the reference image with the observation object image. Yes.

By the way, when obtaining the phase distribution of the observation object from the density distribution in the differential interference image, there are factors other than the phase change of the observation object, such as the light transmittance and reflectance change in the observation object and the illumination light intensity change. If they are mixed, the phase change of the observation object cannot be accurately detected.
These problems can be solved to some extent by acquiring information of an object to be observed in advance as a reference image and performing image processing of the observed object based on the data of the reference image. However, this method is not preferable in view of the fact that it takes a long time to perform the comparison process with the reference image, and that it is desired to shorten the inspection time in the inspection of semiconductors and the like.

Further, in detection of an edge portion such as a step of the observation object, the density distribution of the differential interference image is inverted between a portion that changes from a convex portion to a concave portion and a portion that changes from a concave portion to a convex portion. Therefore, in order to detect the edge portion, both the maximum value and the minimum value of the gray value must be detected in the differential interference image.
However, in this case, the detection characteristics greatly change due to the conversion of the detection characteristics of the element that captures the differential interference image and the intensity characteristics of the illumination light. Furthermore, when detecting a gradient in a specific region on the observation object, a specific value is detected from the differential interference image, and there is a problem that it is particularly susceptible to disturbance light.

By the way, when measuring a change in phase amount such as a step of the observation object, if the step is relatively small, the phase information of the observation object is extracted by combining the differential interference microscope with the fringe scanning used in the interference measurement. be able to. When performing fringe scanning, four images having different polarization component retardation amounts must be captured and calculated, and the task of shortening the processing time cannot be achieved.
In addition, since the gray value of the image changes greatly between the case where the polarization component retardation amount is 0 and π, an image sensor with a wide dynamic range is required to detect an accurate phase amount. The device becomes complicated.
In addition, in this patent document 4 , the inventor paid attention to the imaging characteristics of the differential interference microscope and showed a method for separating phase information and intensity information from the differential interference image.

  The present invention has been made in view of the above-described problems of the prior art, and the object of the present invention is to provide specific information such as the gradient of an observation object, a minute plane portion, an edge portion, a step, and a phase change amount. An apparatus for detecting a physical quantity in a shorter time than a conventional method by separating phase information and intensity information from a differential interference image of an observation object obtained by a differential interference microscope, and a detection method using the apparatus are provided. There is.

In order to achieve the above object, a detection apparatus according to the present invention includes an illumination optical system for guiding light from the light source, which includes a light source and a member for separating light from the light source into two polarization components, to an observation object. A differential interference microscope comprising: a system; and an imaging optical system for forming an image of an observation object provided with a member for reconstructing two polarization components separated in the illumination optical system; The retardation amount formed by the means for changing the retardation amount is provided with means for changing the retardation amount of the polarization component, means for detecting the retardation amount, and means for capturing an image of the observation object. In the detection apparatus for capturing two differential interference images of observation objects having the same sign but different from each other, difference calculation and summation are performed for each corresponding pixel in the images of the two differential interference images. By calculating the difference image information D (x, y) and the sum image information S (x, y), the retardation amount of the polarization component detected by the means for detecting the retardation amount is θ, and each of the images When the phase amount on the observation object plane corresponding to the information is Φ (x, y), the phase on the observation object plane corresponding to the separation direction r of the two polarization components is calculated using any one of the following equations. Computation means for detecting a phase amount Φ (x, y) on the observation object plane by detecting the differential value ∂Φ (x, y) / ∂r of the quantity and integrating in the r direction is provided. A detection device characterized by that.
∂Φ (x, y) / ∂r = k · {(1-cos θ) · D (x, y)}
/ {2sinθ · S (x, y)}
∂Φ (x, y) / ∂r = k · tan ⁻¹ [{(1-cos θ) · D (x, y)}
/ {2 sin θ · S (x, y)}]
However, regarding k, when the wavelength of the light emitted from the light source of the detection device is λ, k = λ / 2π when observing the observation object, and k = λ / when reflecting the observation object. 4π.

In the detection apparatus of the present invention, an illumination optical system for guiding light from the light source, which includes a light source and a member for separating light from the light source into two polarization components, to the observation object, and the illumination A differential interference microscope having an imaging optical system for forming an image of an observation object having a member for reconstructing two polarization components separated in the optical system, and retardation of the two polarization components Observations comprising means for changing the amount, means for detecting the amount of retardation, and means for capturing an image of the observation object, and the retardation amounts formed by the means for changing the amount of retardation are equal and different in sign A method used in a detection apparatus for imaging two differential interference images of an object by the imaging means, wherein a difference calculation and a sum are performed for each corresponding pixel in the images of the two differential interference images. By calculating the difference image information D (x, y) and the sum image information S (x, y), the retardation amount of the polarization component detected by the means for detecting the retardation amount is θ, and each of the images When the phase amount on the observation object plane corresponding to the information is Φ (x, y), the phase on the observation object plane corresponding to the separation direction r of the two polarization components is calculated using any one of the following equations. A method of detecting a phase amount Φ (x, y) on an observation object plane by detecting a differential value ∂Φ (x, y) / ∂r of a quantity and performing integration processing in the r direction.
∂Φ (x, y) / ∂r = k · {(1-cos θ) · D (x, y)}
/ {2sinθ · S (x, y)}
∂Φ (x, y) / ∂r = k · tan ⁻¹ [{(1-cos θ) · D (x, y)}
/ {2 sin θ · S (x, y)}]
However, regarding k, when the wavelength of the light emitted from the light source of the detection device is λ, k = λ / 2π when observing the observation object, and k = λ / when reflecting the observation object. 4π.

According to the present invention, it is possible to accurately detect a minute planar portion of an observation object. Further, by detecting a minute flat portion, it is possible to detect the shape and defect of the minute protrusion for preventing the magnetic head from contacting the magnetic disk.
Next, by extracting the image information and forming its absolute value or square value, the edge portion can be detected without being affected by disturbances such as scattered light by the edge portion. It is possible to accurately detect the edge portion of the formed positioning uneven sample.
Since the phase distribution of the observation object can be measured from the two images having the polarization component retardation amount of ± θ, the time for measuring the phase distribution of the observation object can be shortened. In addition, compared with the conventional fringe scanning method, it is less affected by the dynamic range of the element that captures an image, and the apparatus can be simplified.

The detection apparatus according to the present invention includes a light source and an illumination optical system for guiding light from the light source, which includes a member for separating light from the light source into two polarization components, to the observation object, and the illumination optical system. A differential interference microscope having an imaging optical system for forming an image of an observation object having a member for reconstructing two separated polarization components, and changing the amount of retardation of the two polarization components It comprises means, and means for capturing an image of the observation object, and means for performing an operation of an image captured by the imaging means.

  In the detection apparatus of the present invention, the amount of retardation of the two polarization components separated in the illumination optical system is detected, and two differential interference images of observation objects having the same polarization component retardation amount and different signs are formed. Then, a difference calculation is performed for each corresponding pixel in these two differential interference images to obtain a difference image, and a gradient of the observation object can be detected by extracting image information in a predetermined range from the difference image ( First method).

  In the detection apparatus of the present invention, the retardation amount of two polarization components separated in the illumination optical system is detected, and two differential interference images of observation objects having the same polarization component retardation amount and different signs are formed. Then, a difference operation and a sum operation are performed for each corresponding pixel in these two differential interference images to obtain a difference image and a sum image, and a ratio of image information in the difference image and the sum image is calculated and obtained. The gradient of the observation object can also be detected by extracting image information in a predetermined range from the result (second method).

  Further, if image information in a predetermined range centering on 0 value is extracted from the image information obtained as a result of the calculation according to the first or second method, a portion having no phase change or a plane portion on the observation object is obtained. It can be detected (third method).

  Further, by comparing the area or shape of the portion or plane portion of the observation object detected by the third method without phase change with the area or shape of the plane portion obtained in advance from another reference sample, observation is performed. A difference between the object and the specimen can be detected (fourth method).

  Moreover, if the detection apparatus of this invention is used, the edge part on an observation object is also detectable. That is, in the detection apparatus of the present invention, the amount of retardation of two polarization components separated in the illumination optical system is detected, and two differential interference images of observation objects having the same polarization component retardation amount and different signs are formed. Then, a difference calculation is performed for each corresponding pixel in the two differential interference images to obtain a difference image, an absolute value of image information in the difference image is obtained, a predetermined threshold value is set, and an image area exceeding the threshold value is determined. By extracting, the edge part on the observation object can be detected (fifth method).

  In the apparatus of the present invention, the amount of retardation of two polarization components separated in the illumination optical system is detected, and two differential interference images of observation objects having the same polarization component retardation amount and different signs are formed. A difference operation and a sum operation are performed for each corresponding pixel in the two differential interference images to obtain a difference image and a sum image, and a ratio of image information is calculated for each pixel in the difference image and the sum image. The edge portion on the observation object can also be detected by obtaining the absolute value of the image information of the observation object from the image, setting a predetermined threshold value, and extracting the image information exceeding the threshold value (sixth method).

  In the detection apparatus of the present invention, the retardation amount of two polarization components separated in the illumination optical system is detected, and two differential interference images of observation objects having the same polarization component retardation amount and different signs are formed. Then, a difference calculation is performed for each corresponding pixel in these two differential interference images to obtain a difference image, and this difference image is divided by a value between the maximum value and the minimum value in the difference image, which is 2 By obtaining a power, setting a predetermined threshold value, and extracting an image region exceeding the threshold value, it becomes possible to detect an edge portion on the observation object (seventh method).

  In the detection apparatus of the present invention, the retardation amount of two polarization components separated in the illumination optical system is detected, and two differential interference images of observation objects having the same polarization component retardation amount and different signs are formed. Then, a difference image and a sum image are obtained by performing a difference operation and a sum operation for each corresponding pixel in these two differential interference images, and a ratio of image information is calculated for each pixel in the difference image and the sum image, The image information obtained as a result is divided by a value between the maximum value and the minimum value in the image information, and the value obtained as a result is squared to set a predetermined threshold value. By extracting the region, it is possible to detect the edge portion on the observation object (eighth method).

  Further, the position of the edge portion on the observation object can be detected also by obtaining the coordinates of the image area exceeding the threshold according to the fifth to eighth methods (the ninth method).

  In addition, if the reference coordinates are set in particular in the image area exceeding the threshold according to the fifth to eighth methods, the position of the edge portion on the observation object can be detected with higher accuracy (tenth method).

  By obtaining the coordinates of the image area exceeding the threshold according to the fifth to eighth methods, the interval between the edge portions on the observation object can also be detected (eleventh method).

  In addition, if the reference coordinates are set in particular in the image area exceeding the threshold according to the fifth to eighth methods, the interval between the edge portions on the observation object can be detected with higher accuracy (the twelfth method). .

Furthermore, in the detection apparatus of the present invention, first, the retardation amount of two polarization components separated in the illumination optical system is detected, and two differential interference images having the same polarization component retardation amount and different signs are formed. Thus, difference image information and sum image information are obtained by performing difference calculation and sum calculation for each corresponding pixel in these two differential interference images. Next, the retardation amount of the detected polarization component is θ, the difference image information is D (x, y), the sum image information is S (x, y), on the observation object plane corresponding to each image information. Is the phase on the observation object plane corresponding to the separation direction r of the two polarization components using either of the following conditional expressions (1) and (2): If the differential value ∂Φ (x, y) / ∂r of the quantity is detected and integration processing in the r direction is performed, the phase quantity Φ (x, y) on the observation object can be detected (the thirteenth method). ).
∂Φ (x, y) / ∂r = k · {(1-cos θ) · D (x, y)}
/ {2sinθ · S (x, y)} (1)
∂Φ (x, y) / ∂r = k · tan ⁻¹ [{(1-cos θ) · D (x, y)}
/ {2sinθ · S (x, y)}] (2)
However, regarding k, when the wavelength of the light emitted from the light source of the detection device is λ, k = λ / 2π when observing the observation object, and k = λ / when reflecting the observation object. 4π.

In the detection apparatus of the present invention, first, the retardation amount of two polarization components separated in the illumination optical system is detected, and two differential interference images having the same polarization component retardation amount and different signs are formed. Thus, difference image information and sum image information are obtained by performing difference calculation and sum calculation for each corresponding pixel in these two differential interference images. Next, the retardation amount of the detected polarization component is θ, the difference image information is D (x, y), the sum image information is S (x, y), on the observation object plane corresponding to each image information. When the phase information of Φ (x, y) is deconvoluted by using the optical response characteristics of the differential interference microscope, the difference image information D (x, y) is d (x, y). By using either of the conditional expressions (3) and (4), the phase amount Φ (x, y) on the observation object plane can be detected (fourteenth method).
Φ (x, y) = k · {(1-cos θ) · d (x, y)}
/ {2sinθ · S (x, y)} (3)
Φ (x, y) = k · tan ⁻¹ [{(1-cos θ) · d (x, y)}
/ {2 sin θ · S (x, y)}] (4)
However, regarding k, when the wavelength of light emitted from the light source of the detection device is λ, k = λ / 2π for transmission observation of the observation object, and k = λ / for reflection observation of the observation object. 4π.

In the detection apparatus of the present invention, first, the amount of retardation of the two polarization components separated in the illumination optical system is detected to form two differential interference images having the same polarization component retardation amount and different signs, In these two differential interference images, difference image information and sum image information are obtained by performing difference calculation and sum calculation for each pixel at a corresponding position. Next, the retardation amount of the detected polarization component is θ, the difference image information is D (x, y), the sum image information is S (x, y), on the observation object plane corresponding to each image information. Φ (x, y), the difference image information D (x, y) is deconvoluted using the optical response characteristics of the differential interference microscope, d (x, y), and the sum image information When the image information enveloping the local minimum value in S (x, y) is L (x, y), the phase on the observation object is obtained by using either of the following conditional expressions (5) and (6). The quantity Φ (x, y) can be detected (fifteenth method).
Φ (x, y) = k · {(1-cos θ) · d (x, y)}
/ {2 sin θ · L (x, y)} (5)
Φ (x, y) = k · tan ⁻¹ [{(1-cos θ) · d (x, y)}
/ {2 sin θ · L (x, y)}] (6)
However, regarding k, when the wavelength of light emitted from the light source of the detection device is λ, k = λ / 2π for transmission observation of the observation object, and k = λ / for reflection observation of the observation object. 4π.

In the detection apparatus of the present invention, first, the amount of retardation of two polarization components separated in the illumination optical system is detected, two differential interference images having the same polarization component retardation amount and different signs, and the polarization component A differential interference image having a retardation amount of 0 is formed, and difference and sum operations are performed for each corresponding pixel in two differential interference images having the same polarization component retardation amount and different signs. Get image information. Next, the detected polarization component retardation amount is θ, the polarization component retardation amount is 0, O (x, y) image information, the difference image information is D (x, y), and the sum image. When the information is S (x, y) and the phase amount on the observation object plane corresponding to each image information is Φ (x, y), one of the following conditional expressions (7) and (8) is used. By detecting the differential value ∂Φ (x, y) / ∂r of the phase amount of the observation object corresponding to the separation direction r of the two polarization components and performing integration processing in the r direction, the phase amount on the observation object Φ (x, y) can be detected (sixteenth method).
∂Φ (x, y) / ∂r = k · {(1-cos θ) · d (x, y)}
/ {2sinθ · B (x, y)} (7)
∂Φ (x, y) / ∂r = k · tan ⁻¹ [{(1-cos θ) · d (x, y)}
/ {2 sin θ · B (x, y)}] (8)
However, regarding k, when the wavelength of light emitted from the light source of the detection device is λ, k = λ / 2π for transmission observation of the observation object, and k = λ / for reflection observation of the observation object. 4π. B (x, y) = S (x, y) −2 · O (x, y).

In the detection apparatus of the present invention, first, the amount of retardation of two polarization components separated in the illumination optical system is detected, two differential interference images having the same polarization component retardation amount and different signs, and the polarization component A differential interference image having a retardation amount of 0 is formed, and difference and sum operations are performed for each corresponding pixel in two differential interference images having the same polarization component retardation amount and different signs. Get image information. Next, the detected polarization component retardation amount is θ, the polarization component retardation amount is 0, O (x, y) image information, the difference image information is D (x, y), and the sum image. The information is S (x, y), the phase amount on the observation object plane corresponding to each image information is Φ (x, y), and the difference image information D (x, y) is the optical response characteristic of the differential interference microscope. When the image information deconvoluted using d is set to d (x, y), the phase amount Φ (x, y on the observation object plane is obtained by using one of the following conditional expressions (9) and (10). ) Can be detected (17th method).
Φ (x, y) = k · {(1-cos θ) · d (x, y)}
/ {2sinθ · B (x, y)} (9)
Φ (x, y) = k · tan ⁻¹ [{(1-cos θ) · d (x, y)}
/ {2 sin θ · B (x, y)}] (10)
However, regarding k, when the wavelength of light emitted from the light source of the detection device is λ, k = λ / 2π for transmission observation of the observation object, and k = λ / for reflection observation of the observation object. 4π. B (x, y) = S (x, y) −2 · O (x, y).

In the detection apparatus of the present invention, first, the amount of retardation of two polarization components separated in the illumination optical system is detected, two differential interference images having the same polarization component retardation amount and different signs, and the polarization component A differential interference image having a retardation amount of 0 is formed, and difference and sum operations are performed for each corresponding pixel in two differential interference images having the same polarization component retardation amount and different signs. Get image information. Next, the detected polarization component retardation amount is θ, the polarization component retardation amount is 0, O (x, y) image information, the difference image information is D (x, y), and the sum image. The information is S (x, y), the phase amount on the observation object plane corresponding to each image information is Φ (x, y), and the difference image information D (x, y) is the optical response characteristic of the differential interference microscope. Deconvolved image information using d (x, y), {S (x, y) -2 · O (x, y)} enveloping image information b (x, y) Then, by using any of the following conditional expressions (11) and (12), the phase amount Φ (x, y) on the observation object plane can be detected (eighteenth method).
Φ (x, y) = k · {(1-cos θ) · d (x, y)}
/ {2 sin θ · b (x, y)} (11)
Φ (x, y) / ∂r = k · tan ⁻¹ [{(1-cos θ) · d (x, y)}
/ {2 sin θ · b (x, y)}] (12)
However, regarding k, when the wavelength of light emitted from the light source of the detection device is λ, k = λ / 2π for transmission observation of the observation object, and k = λ / for reflection observation of the observation object. 4π.

The inventor has described in detail the imaging characteristics of the differential interference microscope from the method of deriving the imaging characteristics to the results in the above-mentioned Patent Document 4 .
The characteristic of the differential interference microscope is that the image intensity distribution in the differential interference microscope is I (x, y), the retardation amount of the two polarization components in the differential interference microscope is θ, and the light transmission (reflection) rate of the observation object is T. If (x, y), the phase information of the differential interference microscope is P (x, y), and the intensity information of the image is A (x, y), it can be simply expressed by the following equation (13).
I (x, y, θ) = T (x, y) {(1−cos θ) · A (x, y) / 2
+ Sinθ · P (x, y)} (13)
Further, when −θ in the equation (13),
I (x, y, −θ) = T (x, y) {(1−cos θ) · A (x, y) / 2
−sin θ · P (x, y)} (14)
It becomes.

When calculating (13)-(14), (13) + (14) from the equations (13), (14),
I (x, y, θ) −I (x, y, −θ)
= 2T (x, y) · sin θ · P (x, y) (15)
I (x, y, θ) + I (x, y, −θ)
= T (x, y). (1-cos.theta.). A (x, y) (16)
It becomes.
Therefore, it is possible to separate the phase information and the image intensity information from the differential interference image by forming the difference image information and the sum image information.

Furthermore, by extracting the phase information from the differential interference image, it is possible to obtain image information corresponding to the phase change of the observation object. In particular, since the phase information of the differential interference image correlates with the differential value of the phase distribution of the observation object, the gradient corresponding to the predetermined value is detected in the observation object by extracting the predetermined value from the phase information of the image. can do.
Therefore, as shown in the first method described above, a difference image is acquired from two differential interference images having the same polarization component retardation amount and different signs, and image information in a predetermined range is extracted from the difference image. Thus, the gradient of the observation object can be easily detected. An apparatus for detecting the gradient of the observation object can be configured by systematizing this method.

By the way, when the transmittance (reflectance) T (x, y) of the observation object is relatively close to 1, the gradient of the observation object can be detected sufficiently accurately only by acquiring the difference image information. In a general case, the influence of T (x, y) can be eliminated by calculating the ratio of Equation (15) and Equation (16).
Therefore, as shown in the second method, by obtaining the difference image and the sum image and obtaining the ratio thereof, the influence of the light transmittance and reflectance of the observation object is removed, and then the gradient of the observation object is obtained. Can be detected. Moreover, an apparatus for detecting the gradient of the observation object can be configured by systematizing this method.

In addition, as described in the third method, when an image region within a predetermined range is extracted from the image information obtained by the first or second method around a portion where the value is 0, the phase gradient is substantially reduced. A region that becomes 0 can be extracted. This is equivalent to detecting a portion having no phase change or a plane portion from the observation object. This is because when the observation object has a curved surface or a spherical surface, the vicinity of the surface top or a portion where no phase change occurs can be regarded as a flat surface. Further, when wear, deformation, or alteration occurs on the curved surface or spherical surface of the observation object, the top portion of the curved surface or spherical surface also changes.
Therefore, as shown in the third method, it is possible to detect wear, deformation, or alteration of the observation object by detecting a portion or plane in which the observation object has no phase change. Further, by systematizing this method, it is possible to configure a device that detects wear, deformation, and alteration of an observation object.

Further, as shown in the fourth method, by determining the area of the region where the phase does not change or the region of the plane portion and the envelope shape, and comparing with the area and shape of the sample as a reference in advance, Deformation and alteration can be measured quantitatively. Furthermore, it is possible to detect defects by performing correlation detection.
If this method is automated, a detection apparatus that implements the fourth method can be configured.

Specific examples of observation objects include minute protrusions called bumps for preventing adhesion of magnetic heads formed on a magnetic disk, spherical conductors formed on IC chips and IC lead frames, and semiconductors. There is a spherical conductor formed on a substrate.
Therefore, by applying the above-described fourth method, when the observation object is homogeneous, a deteriorated portion is detected, so that a relative defect can be detected. For example, it is possible to detect alteration of liquid crystal cells and electrodes used in displays and the like.

Further, the use of a differential interference microscope to detect the edge portion of the step is disclosed in Patent Document 3 described above . However, this does not show any imaging characteristics of the differential interference microscope, and only the edge portion can be detected because the phase change occurring at the step portion changes the contrast of the interference fringes. When the edge portion of the step is actually observed with a microscope, it can be confirmed that light is scattered at the edge portion. Due to this effect, when observing the edge of the step with a differential interference microscope, it is confirmed that the density of the image is asymmetric between the portion where the convex portion of the step changes from the concave portion to the concave portion and vice versa. Has been.

The cause of this can be considered that the scattered light component of the edge portion is added to the differential interference image. Therefore, assuming that the scattered light at the edge portion has no polarization characteristic, and assuming that the scattered light component at the edge portion is N (x, y), the above equation (13) is
I (x, y, θ) = T (x, y) {(1−cos θ) · A (x, y) / 2
+ Sin θ · P (x, y) + N (x, y)} (17)
It can be expressed. Even if the equation (17) is used, the difference image information can be expressed by the equation (15) similarly to the equation (13).
For Japanese image information,
I (x, y, θ) + I (x, y, −θ) = T (x, y) {(1−cos θ)
A (x, y) + 2N (x, y)} (18)
However, since the scattered light component is considered to be weak compared with the intensity component of the image, there is no problem even if the equation (16) is approximately used instead of the equation (18).

  Therefore, by extracting the difference image information, it is possible to extract phase information correlated with the phase change of the edge portion of the observation object. Specifically, the edge portion of the step can be detected by obtaining both the maximum value and the minimum value of the image information from the difference image and extracting the phase information. The phase information to be extracted is formed as shown in the fifth method because the change in density of the image is symmetrical between the portion where the convex portion of the step changes from the concave portion to the concave portion and vice versa. By forming image information that takes the absolute value of the image information of the difference image that has been obtained, and extracting a portion that exceeds a predetermined threshold from this image information, the edge portion of the step can be detected.

Further, as shown in the sixth method, the light transmittance of the observation object is calculated by calculating the ratio of the image information of the difference image and the sum image for each pixel from the acquired difference image and the sum image. And adverse effects of reflectance can be eliminated. Furthermore, the edge portion of the step can be accurately detected by obtaining the image information in which the absolute value of the result obtained by the above calculation is taken and the portion having a predetermined threshold value or more is obtained.
Further, when each pixel data of the formed difference image is squared to form a square image of the difference image information, the change is further increased in a portion where the density of the image changes abruptly, such as an edge portion of a step. Therefore, by doing in this way, the precision of the 6th method which sets a predetermined threshold and detects a part more than that value can be improved more.

The image captured by the imaging means is normally handled as 256-gradation integer data from 0 to 255 for each pixel. When a difference image is formed, integer data of 512 gradations from −255 to 255 is obtained. Even if this integer data is squared as it is, edge detection can be performed satisfactorily. Furthermore, as a method for improving the accuracy of edge detection, a value between the maximum value and the minimum value of the formed difference image is set, and the image data of the difference image formed by this value is divided to obtain the image data of the difference image. To a real number. Thereby, the image data of the difference image is composed of one or less area and one or more area. Further, the accuracy of edge detection can be improved by setting a value of 1 or more as the threshold value.
Therefore, the edge portion of the step can be detected with higher accuracy by using the seventh method described above.

  Further, as shown in the eighth method, the ratio of the image information of the difference image and the sum image is obtained, the adverse effect of the light transmittance and reflectance of the observation object is removed, and then the formed image information By forming an image divided by using a value between the maximum value and the minimum value, it is possible to eliminate the influence of noise or the like of the image sensor and to detect the edge portion of the step more accurately. .

The ninth method is an application example of the fifth to eighth methods. In this method, first, a coordinate system is set on the captured image, and the detected edge information is made to correspond to the coordinates on the image. Then, the observation object can be positioned by relatively moving the observation object until the coordinates of the edge portion overlap with predetermined coordinates.
Further, when the detected edge portion has a specific size and occupies a certain area, the center or the center of gravity in two or one direction in the coordinate system of the area is obtained, and the center value or the center of gravity value is obtained. The observation object can be positioned by using as a representative point. Further, as shown in the ninth and tenth methods, the observation object can be positioned by detecting the edge portion on which the difference image is formed.

Similar to positioning, two specific areas are set from the detected edge information, the distance is calculated from the coordinates of the two areas, and the distance on the observation object plane is calculated using parameters such as the optical magnification of the imaging optical system. By converting, the distance between two predetermined points on the observation object plane can be measured.
Therefore, as shown in the eleventh and twelfth methods, the edge interval or length of the observation object can be obtained by forming the difference image and detecting the edge portion.

Although the fifth to twelfth methods have been described above, it is easy to use these methods as detection devices by automating these methods.
Further, since the image when the retardation amount of the polarization component is 0 is considered to represent the confusion light component N (x, y) at the edge portion in the equation (17), the retardation of the polarization component is calculated from the sum image. By forming an image obtained by subtracting an image obtained by doubling the image information with 0 as the amount of foundation, and replacing this image with a sum image, the ratio of the difference image is obtained, so that edge detection with higher accuracy can be performed.

Incidentally, a method of measuring the phase distribution of an observation object by combining a differential interference microscope with fringe scanning used in interference measurement is disclosed in the above-mentioned Patent Document 1 and the like. Usually, when a fringe scanning method is combined with a differential interference microscope, four differential interference images with polarization component retardation amounts of 0, π / 2, π, and 3π / 2 are captured, and these four differential interference images are captured. Using each pixel data of
tan ⁻¹ [{I (π / 2) −I (3π / 2)} / [I (0) −I (π)]]
.... (19)
The phase information of the observation object is obtained by performing the above calculation.

The fringe scanning method in the interferometry is based on the premise that no light diffraction or scattering occurs on the surface of the observation object. However, since the differential interference microscope converts the light diffracted on the object surface into a differential interference image, an image having a form in which both phase information and intensity information are combined is obtained as shown in equation (13). It is done.
When an image having a polarization component retardation θ of ± π / 2 is captured and a difference image and a sum image are formed, the difference image is converted into “I (π / 2) −I (3π / 2)” in Expression (19). Corresponding information is obtained.
Therefore, in the differential interference microscope, information similar to the fringe scanning method can be obtained by using an image of θ = ± π / 2. The same effect can be obtained for any θ.

Therefore, as shown in the thirteenth method, while detecting the retardation amount of the polarization component with the differential interference microscope, the difference image and the sum image are obtained from the two differential interference images having the same polarization component retardation amount and different signs. Then, a differential value of the phase amount in the shear direction of the observation object can be detected by calculating an image ratio for each pixel and obtaining a value of tan ⁻¹ . Further, by obtaining the integral value of the differential value in the shear direction, the phase amount at each point of the observation object can be quantitatively measured more accurately.

In addition, since the density of the differential interference image varies depending on the amount of the object to be observed and the shear amount of the differential interference microscope, high-accuracy measurement can be performed by capturing an image with the retardation amount that provides the best density and detecting the phase.
When the detected phase amount is small, the differential value of the phase amount can be obtained approximately without obtaining the value of tan ⁻¹ , and the phase amount can also be obtained by performing an integration process.
In this case, since the phase distribution of the observation object can be obtained from the two differential interference images, the measurement time can be shortened as compared with the conventional fringe scanning method.

Also, in the differential interference microscope, the image density distribution changes greatly between when the polarization component retardation amount is 0 and when it is π. Therefore, when the fringe scanning method is used, an image sensor having a very wide dynamic range. Is required.
In the present invention, since an image having a polarization component retardation amount of π is not particularly required, measurement is possible without using an image sensor having a very wide dynamic range.

  When measuring the phase distribution of an observation object using a differential interference microscope, the pupil diameter of the differential interference microscope is finite, so that all the light diffracted by the observation object cannot be acquired as image information. The imaging characteristics of the differential interference microscope are expressed by the above equation (13). The phase information P (x, y) and the image intensity information A (x, y) are convolved with specific optical response functions as shown in the following equations (20) and (21). It has become. Therefore, in order to accurately obtain the phase information of the observation object, it is necessary to consider the optical response characteristics of the differential interference microscope.

.... (20)
.... (21)
However,
Here, Q (ξ, ζ) is the pupil function of the illumination optical system, R (ξ, ζ) is the pupil function of the imaging optical system, Δ _x and Δ _y are the x component, y component of the shear amount, f _x and f _y indicate spatial frequencies in the x and y directions.

  However, in the present invention, as described in the above-described fourteenth method, after the difference image is formed, the difference image is deconstructed using the optical response characteristic corresponding to the phase information of the differential interference microscope shown in Equation (20). The phase amount of the observation object can be accurately obtained without performing an integration process by performing a volume process and obtaining a difference from the sum image.

  However, when the phase amount of the observation object becomes larger than the weak phase region, the influences of the second and third terms of Expression (21) begin to increase. In order to accurately obtain the phase amount of the observation object, it is necessary to detect zero-order light that is not diffracted and scattered by the observation object. However, since the intensity can be detected only in a state where it is actually mixed with the light diffracted and scattered, the zero-order light component must be separated from the intensity information.

  As a method of separation, the 0th-order light component can be separated by Fourier transforming the density information of the image and extracting only low frequencies. In addition, it is considered that the 0th-order light component is a minimum value in the density information of the image. Therefore, as shown in the fifteenth method, by forming the sum image, the density information of the image is extracted to obtain the minimum value of the sum image, and the image information L (x, y) enveloping the minimum value and By obtaining a ratio with the image information d (x, y) obtained by deconvolution of the difference image, it is possible to improve the detection accuracy of the phase amount of the observation object.

  Although the thirteenth to fifteenth methods have been described above, it is easy to use these methods as detection devices by automating these methods. As a result, it is possible to obtain effects such as shortening the detection time and being hardly affected by the dynamic range of the image sensor that detects the differential interference image.

When the phase change of the observation object is relatively large and an edge portion such as a step exists, light scattering occurs at the edge portion. This confusion at the edge can cause a decrease in the accuracy of phase detection. In order to eliminate this influence, as described in the sixteenth method, a total of 3 images, one each when the polarization component retardation amount is 0 and one when the polarization component retardation amount is ± θ, are 3 in total. Image information O (x, difference image information D (x, y), sum image information S (x, y), and retardation amount 0, which are obtained from an image of which the polarization component retardation amount is ± θ. , Y), a value of B (x, y) = S (x, y) −2 · O (x, y) is obtained, and difference image information D (x, y) and B (x, y) The scattered light at the edge portion can be removed by obtaining the ratio.

  Further, as shown in the seventeenth method, by performing the deconvolution process similarly to the fourteenth method, phase detection can be performed in consideration of the influence of the response characteristics of the differential interference microscope. Further, as described in the eighteenth method, the image information enveloping the minimum value of the image information is formed as in the fifteenth method, and the ratio between the difference image and the deconvolution image is obtained, so that Phase detection in consideration of the influence of scattered light.

Hereinafter, the present invention will be described in detail by showing reference examples and examples of the present invention.

First Reference Example This reference example shows a method for detecting minute protrusions (bumps) for preventing adhesion of a magnetic head formed on a magnetic disk. In this reference example , a detection apparatus incorporating an epi-illumination type differential interference microscope for observing a metal object is used. The outline is shown in FIG.

As shown in FIG. 1, the detection device used in this reference example captures an image obtained by a differential interference microscope 4 including a light source 1, an illumination optical system 2, and an imaging optical system 3, and the differential interference microscope 4. And a microcomputer 6 that calculates an image picked up by the CCD camera 5.
The illumination optical system 2 includes a lens 7, a diaphragm 8, a lens 9, a polarizer 10, a quarter wavelength plate 11, a half mirror 12, a Nomarski prism 13, and an objective lens 14 arranged in this order from the light source 1 side. ing. The imaging optical system 3 includes an objective lens 14, a Nomarski prism 13, a half mirror 12, an analyzer 16, and a lens 17 arranged in this order from the observation object 15 side. The half mirror 12, Nomarski prism 13, and objective lens 14 are common to the illumination optical system 2 and the observation optical system 3.

  In the detection apparatus shown in FIG. 1, the light emitted from the light source 1 is polarized through the polarizer 10, then passes through the quarter-wave plate 11, and is reflected downward by the half mirror 12. The reflected light passes through the objective lens 14 through the Nomarski prism 13 arranged so that the branch point between the ordinary light and the extraordinary light is localized at the pupil position of the objective lens 14. The light is separated by a predetermined shear amount. Then, the ordinary light and the abnormal light reflected by the observation object 15 are transmitted again through the objective lens 14, reconstructed by the Nomarski prism 13, further transmitted through the half mirror 12, and then transmitted through the analyzer 16. The ordinary light and the abnormal light interfere with each other, and a differential interference image of the observation object 15 is formed on the imaging surface of the CCD camera 5 via the lens 17.

Here, the polarizer 10 is rotatable about the optical axis. Further, a pulse motor 18 is connected to the polarizer 10 and can be controlled by the microcomputer 6 so that the rotation angle of the polarizer 10 can be freely set. Accordingly, the retardation amount of the polarization component can be set by the polarizer 10 by controlling the rotation of the pulse motor 18 by the microcomputer 6.
The quarter-wave plate 11 is fixed so that its fast axis or slow axis coincides with the polarization direction of the analyzer 16.

In this reference example , first, a reference mirror is placed in place of the observation object 15 for observation, and the microcomputer 6 is operated to capture the light intensity distribution while rotating the polarizer 10, and the rotation angle of the polarizer 10. And the change in the retardation amount of the polarization component.
Next, a magnetic disk is placed as the observation object 15 instead of the mirror, and the polarizer 10 is rotated so that the retardation amount of the polarization component is set to a predetermined value θ, and an image on the magnetic disk is captured. In this case, the retardation amount θ of the polarization component is set to an optimum value by observing a typical magnetic disk or the like.
Then, the polarizer 10 is rotated so that the amount of retardation of the polarization line becomes −θ, and an image is captured. A difference image is formed from the captured two images to extract phase information. Further, a portion near 0 value is extracted from the difference image, and an image obtained by binarizing the portion near 0 value and the other portion is formed.

FIGS. 2A and 2B show differential interference images of the bumps when the retardation amounts of the polarization components are θ and −θ, respectively. FIG. 4C shows a difference image formed for extracting phase information (here, the maximum and minimum values are indicated by 256 gradations for the convenience of display). Further, FIG. 4D shows an image when a value near 0 value is extracted from the difference image, and a portion near 0 value and the other portion are binarized.
The cross section of the pump formed on the magnetic disk has a shape as shown in FIG. Therefore, the gradient is 0 at the top of the bump. However, when the bump is worn or deformed by the magnetic head, the flat surface portion of the top surface is deformed.

In the present reference example , a circle R that minimally approximates the data of the surface top portion is obtained from the binarized image of FIG. 2D, and the radius or diameter is parameterized as shown in FIG. The degree of wear can be detected.
Further, since the magnetic disk has a substantially uniform light reflectance, phase information can be detected simply by forming a difference image. However, a sum image of the images shown in FIGS. 2 (a) and 2 (b) is further formed for an observation object whose light reflectance is not uniform, and each of the corresponding pixels in the sum image and the difference image is formed. By forming image information that takes a ratio for each data, it is possible to eliminate the influence of changes in the reflectance of light.

In this reference example , the method for detecting the bumps of the magnetic disk has been described. However, the detection method shown in this reference example is not limited to bump detection. For example, it can be used as a method for inspecting a spherical solder of a BGA (ball grit array) formed on an IC chip or a lead frame. In this case, when the size of the spherical solder changes, the area of the top portion of the surface also changes. Therefore, the size of the spherical solder can be determined by parameterizing the radius and area of the circle to be detected by detecting the plane portion. Can be detected. In addition, shape changes and defects can be detected by detecting the top of the surface.

In this reference example , a detection apparatus incorporating an epi-illumination type differential interference microscope is used. However, if a transmission type differential interference microscope is used, a homogeneous portion of a transmission specimen can be detected. In this case, it is also possible to detect a non-homogeneous part by detecting a homogeneous part.

Second Reference Example This reference example shows a method of detecting an edge of a semiconductor wafer alignment uneven specimen (box mark). As in the first reference example , a detection apparatus incorporating the epi-illumination type differential interference microscope shown in FIG. 1 is used.

  First, in the detection apparatus shown in FIG. 1, the microcomputer 6 is operated to rotate the polarizer 10 to change the amount of retardation of the polarization component, and capture an image with the polarization component retardation amounts θ and −θ. . Next, a difference image is formed from the two images, an absolute value of the difference image is obtained, and a predetermined threshold value is set. To do.

  FIGS. 5 (a) and 5 (b) show differential interference images of a concavo-convex sample with the polarization component retardation amounts θ and −θ, respectively. FIG. 2C shows a difference image (here, the maximum and minimum values are indicated by 256 gradations for the convenience of display) formed to extract phase information. FIG. 4D shows an image in which a threshold value is set by taking the absolute value of the difference image, and binarized based on this threshold value by a portion above the threshold value and a portion below the threshold value.

According to the detection method of the present reference example , the edge portion of the concavo-convex sample can be detected from the image shown in FIG. Then, an xy coordinate is set on the image, and the concavo-convex sample can be positioned by placing the concavo-convex sample on a stage that can be moved two-dimensionally corresponding to the xy coordinate on the image (see FIG. 6).
Furthermore, in the image shown in FIG. 6, the length of the concavo-convex specimen can be measured by measuring the interval between the edge portions.

  Here, as a method for improving the accuracy of position detection of the edge portion, a reference position of an area equal to or larger than the threshold obtained by the above-described method is obtained in the xy coordinates on the image, and the reference position is represented as a representative of the edge portion. There are a method of making a point and a method of detecting the maximum value of the gray value of the image in an area equal to or greater than a threshold value and setting the coordinates of the edge portion based on this.

In this reference example , the absolute value of the difference image is shown. However, a predetermined value may be set between the maximum value and the minimum value of the difference image, and the difference image may be divided by this value. Thereby, the difference image can be separated into an area of ± 1 or more and an area of ± 1 or less. Then, by squaring the value of each pixel data of the difference image, the area of ± 1 or more has an image gray value of 1 or more, and the area of ± 1 or less has an image gray value of 1 or less. Therefore, the edge signal can be made steeper and the edge detection accuracy can be improved.

  Furthermore, since the reflectance of the concavo-convex specimen is almost uniform as in the magnetic disk, phase information can be extracted only by forming a difference image. However, for an uneven sample with uneven light reflectance, a sum image of the two images shown in FIGS. 5A and 5B is formed, and each pixel data in the difference image and the sum image is formed for each pixel data. By forming the image information with the ratio of the above, it is possible to remove the influence of the change in the reflectance of light. This method can be applied to IC pattern positioning and length measurement.

Further, the detection method of this reference example can be applied to positioning of transparent electrodes of liquid crystals, measurement of intervals between transparent electrodes, and the like by using a transmission type differential interference microscope.

Embodiment This embodiment shows a detection method using a phase grating for measuring the phase amount of a phase object. In this embodiment, a detection apparatus incorporating a transmission type differential interference microscope is used. The outline is shown in FIG.

As shown in FIG. 7, the detection apparatus used in the present embodiment captures an image obtained by the light source 21, the differential interference microscope 24 including the illumination optical system 22 and the imaging optical system 23, and the differential interference microscope 24. And a microcomputer 26 for calculating an image picked up by the CCD camera 25.
The illumination optical system 22 includes a lens 27, a stop 28, a lens 29, a polarizer 30, a ¼ wavelength plate 31, a Nomarski prism 32, and a condenser lens 33. The observation optical system 23 includes an objective lens 34, a Nomarski prism 35, an analyzer 36, and a lens 37.

  In the detection apparatus shown in FIG. 7, the light emitted from the light source 21 is polarized by the polarizer 30, then passes through the ¼ wavelength plate 31, and a branch point between ordinary light and abnormal light at the pupil position of the condenser lens 33. The normal light and the abnormal light are separated by a predetermined shear amount on the observation object 15 through the condenser lens 33 by the Nomarski prism 32 arranged so as to be localized. Ordinary light and extraordinary light transmitted through the observation object 15 are transmitted through the objective lens 34 and then reconstructed by a Nomarski prism 35 arranged so that the combined point of the ordinary light and extraordinary light is localized at the pupil position of the objective lens 34. Further, when passing through the analyzer 36, the ordinary light and the extraordinary light interfere with each other, and a differential interference image of the observation object 15 is formed on the imaging surface of the CCD camera 25 via the lens 37.

In the present embodiment, an interference filter is arranged in the light source 21 so that the emitted illumination light is quasi-monochromatic light having a wavelength of 550 nm.
Further, the polarizer 30 is rotatable around the optical axis. Further, a pulse motor 38 is connected to the polarizer 30 and is controlled by the microcomputer 26 so that the rotation angle of the polarizer 30 can be freely set. Therefore, by controlling the rotation of the pulse motor 38 by the microcomputer 26, the retardation amount of the polarization component can be set by the polarizer 30.
The quarter-wave plate 31 is fixed so that its fast axis or slow axis coincides with the polarization direction of the analyzer 36.

First, in the present embodiment, a uniform phase object is observed, and the light / dark distribution of an image is captured while rotating the polarizer 30, and the change in the rotation angle of the polarizer 30 and the retardation amount of the polarization component is obtained.
Next, the image of the observation object 15 is captured by rotating the polarizer 30 so that the retardation amount of the polarization component becomes θ. Note that the rotation angle of the polarizer 30 at this time is detected by the microcomputer 26, and at the same time, the retardation amount of the polarization component at this time is also stored by the microcomputer 26. Similarly, the image of the observation object 15 is captured by rotating the polarizer 30 so that the retardation amount of the polarization component becomes −θ. Further, the rotation angle of the polarizer 30 at this time is detected by the microcomputer 26, and at the same time, the retardation amount of the polarization component at this time is also stored by the microcomputer 26. Then, a difference image and a sum image are formed from two images in which the amount of retardation of the captured polarization component is θ and −θ.
Further, the formed difference image is deconvoluted by using the optical response characteristic of the differential interference microscope shown in FIG. 8 in the microcomputer 26 to newly form a phase information image considering the optical response function. Thereafter, each formed image information is divided by the sum image to obtain a value of tan ⁻¹ , and is converted into a phase distribution by multiplying the following value obtained from the detected retardation amount θ of the polarization component.

k · (1-cos θ) / 2 sin θ where k = λ / 2π (λ = 550 nm)
.... (22)

FIGS. 9A and 9B show images of phase objects in which the polarization component retardation amounts are θ and −θ, respectively. FIGS. 9C and 9D show a difference image and a sum image formed by the images shown in FIGS. FIG. 9E shows a reproduction diagram (cross-sectional view of the grating) of the phase distribution at this time.
For reference, FIG. 10 shows a cross-sectional view of a grating reproducing a phase distribution obtained by performing a deconvolution process using a conventional fringe scanning method.

  As described above, according to the present embodiment, it is understood that the phase distribution of the observation object equivalent to the measurement by the conventional fringe scanning method is obtained from the two images having the polarization component retardation amount of ± θ. In particular, when the conventional fringe scanning method is used, four images having different polarization component retardation are required, whereas in this embodiment, only two images having different polarization component retardation amounts are used. Since the phase distribution can be obtained in the same manner as the method, the measurement time can be shortened.

By the way, in the detection method of the present embodiment, when it is known in advance that the phase amount of the object to be observed is relatively small, the approximation of tan φ = φ is established, and thus the processing for obtaining the value of tan ⁻¹ is omitted. Measurement time can be further reduced. The same result can be obtained even if the image after the ratio between the difference image and the sum image is obtained is deconvolved.

  The optical response characteristic of the differential interference microscope varies depending on the diameter of the aperture stop of the illumination optical system. Depending on the observation object, the phase distribution may be biased to a specific spatial frequency band. When performing phase detection of such an object, the response characteristic can be maintained at a value close to 1 by setting the diameter of the aperture stop to an appropriate size, so that phase distribution can be achieved without performing deconvolution processing. Can be obtained accurately.

In this embodiment, the retardation amount of the polarization component is detected by detecting the rotation angle of the polarizer 30, but a means for detecting the retardation amount of the polarization component is added to the detection device, and this detection means. The retardation amount of the polarization component may be obtained using a signal from
In addition, when changing the retardation amount of the polarization component, not only a method of rotating the polarizer 30 but also a liquid crystal element is inserted between the polarizer 30 and the quarter wavelength plate 31, and the applied voltage of the liquid crystal element is changed. The same effect can be obtained by using a method of changing, a method of arranging a half-wave plate in a detachable manner between the polarizer 30 and the quarter-wave plate 31.

  Moreover, although the example which permeate | transmits is shown in a present Example, the same phase distribution can be calculated | required even if it uses the epi-illumination type differential interference microscope shown in FIG. However, when the epi-illumination type is used, when the observation object is a reflective object such as a metal, the detected phase distribution is doubled. Therefore, in such a case, it is necessary to obtain the phase distribution with k = λ / 4π in the above-described equation (22).

  In this embodiment, a specific method for obtaining the phase distribution of the observation object from two images having a polarization component retardation amount of ± θ is shown. Similarly, the polarization component retardation amount is ± θ 2. One image and an image with a polarization component having a retardation amount of 0 are captured, and B (x, y) is subtracted from the image shown in FIG. 9D by twice the image information with a polarization component having a retardation amount of 0. (Where B (x, y) = S (x, y) −2 · O (x, y)), and the ratio between the difference image and B (x, y) shown in FIG. 9C is formed. By doing so, it is possible to detect the phase distribution in consideration of the influence of light scattering at the edge portion of the observation object.

As described above, in the present invention, as described with reference to the first and second reference examples and the examples , the image pickup device (CCD camera) is used as one unit and the image is picked up while changing the retardation amount of the polarization component. Have said. However, the intention of the present invention is not limited to the use of a single image pickup means, and the same effect can be obtained even when two image pickup means are used.

For example, as shown in FIGS. 11 and 12, a PBS (polarization beam splitter) is arranged in the observation optical system in place of the analyzer and separated into two orthogonal two polarization components, and each polarization component is separated by a CCD camera. You may make it receive light. In this way, differential interference images having the same polarization component retardation amount and different signs are formed on the imaging surfaces of the two CCD cameras.
Therefore, by calculating the images from the two CCD cameras, physical quantities such as the gradient of the observation object, the plane portion, the birefringence portion, the distortion, the edge portion of the step, and the phase distribution can be detected.

It is a figure which shows the structure of the apparatus for detecting the physical quantity of the observation object by this invention. (A), (b) is a figure which shows the differential interference image of the observation object when the retardation amount of the polarization component acquired by the method by the 1st reference example using the apparatus shown in FIG. 1 is respectively θ and −θ. It is. (C) is a figure which shows the difference image formed with the image shown to (a), (b). (D) is a figure which shows the image when the value near 0 value is extracted from the difference image shown in (c), and the part near 0 value and the other part are binarized. It is sectional drawing of the bump currently formed on the magnetic disc. It is a figure which shows the image which approximated the information of the surface top part of the said bump minimum from the image shown in FIG.2 (d) for detecting the degree of wear of the bump shown in FIG. (A), (b) is the figure which shows the differential interference image of the observation object when the retardation amount of the polarization component acquired by the method by the 2nd reference example using the apparatus shown in FIG. 1 is respectively θ and −θ. It is. (C) is a figure which shows the difference image formed with the image shown to (a), (b). (D) is a figure which shows the image binarized by taking the absolute value of the difference image shown in (c) and setting a threshold, and based on this threshold, a portion above the threshold and a portion below the threshold. It is a figure for demonstrating the procedure which detects the edge part of an observation object by the method of a 2nd reference example , and positions an observation object. It is a figure which shows the structure of the apparatus for detecting the physical quantity of the observation object by this invention. It is a graph which shows the optical response characteristic of a differential interference microscope. (A), (b) is a figure which shows the differential interference image of the observation object when the retardation amount of the polarization component acquired by the method by the Example using the apparatus shown in FIG. 7 is respectively θ and −θ. . (C), (d) is a figure which shows the difference image and sum image which were formed by the image shown to (a), (b), respectively. (E) is a phase distribution reproduction diagram. It is a phase distribution reproduction figure detected by the conventional fringe operation method. It is a figure which shows the structure of the apparatus for detecting the physical quantity of the observation object by this invention. It is a figure which shows the structure of the apparatus for detecting the physical quantity of the observation object by this invention.

Explanation of symbols

1,21 Light source 2,22 Illumination optical system 3,23 Imaging optical system 4,24 Differential interference microscope 5,25 CCD camera 6,26 Microcomputer 7, 9, 17, 27, 29, 37 Lens 8, 28 Aperture 10 , 30 Polarizer 11, 31 1/4 Wave plate 12 Half mirror 13, 32, 35 Nomarski prism 14, 34 Objective lens 15 Observation object 16, 36 Analyzer 18, 38 Pulse motor 33 Condenser lens

Claims (2)

A light source, an illumination optical system for guiding light from the light source to an observation object provided with a member for separating light from the light source into two polarization components, and two separated in the illumination optical system A differential interference microscope having an imaging optical system for forming an image of an observation object having a member for reconstructing a polarization component, means for changing the amount of retardation of the two polarization components, and the retarder Means for detecting the amount of retardation and means for capturing an image of the observation object, and the two differential interference images of the observation object having the same retardation amount and different signs formed by the means for changing the retardation amount In a detection apparatus that captures an image using an imaging unit,
Difference image information D (x, y) and sum image information S (x, y) are obtained by performing difference calculation and sum calculation for each corresponding pixel in the images of the two differential interference images, and the amount of retardation When the retardation amount of the polarization component detected by the detecting means is θ and the phase amount on the observation object plane corresponding to each image information is Φ (x, y), one of the following equations is used. By detecting the differential value ∂Φ (x, y) / ∂r of the phase amount on the observation object plane corresponding to the separation direction r of the two polarization components and performing integration processing in the r direction, the observation object plane is obtained. A detection apparatus comprising an arithmetic means for detecting the upper phase amount Φ (x, y).
∂Φ (x, y) / ∂r = k · {(1-cos θ) · D (x, y)}
/ {2sinθ · S (x, y)}
∂Φ (x, y) / ∂r = k · tan ⁻¹ [{(1-cos θ) · D (x, y)}
/ {2 sin θ · S (x, y)}]
However, regarding k, when the wavelength of the light emitted from the light source of the detection device is λ, k = λ / 2π when observing the observation object, and k = λ / when reflecting the observation object. 4π. A light source, an illumination optical system for guiding light from the light source to an observation object provided with a member for separating light from the light source into two polarization components, and two separated in the illumination optical system A differential interference microscope having an imaging optical system for forming an image of an observation object having a member for reconstructing a polarization component, means for changing the amount of retardation of the two polarization components, and the retarder Means for detecting the amount of retardation and means for capturing an image of the observation object, and the two differential interference images of the observation object having the same retardation amount and different signs formed by the means for changing the retardation amount A method used in a detection device that captures an image with an imaging means,
Difference image information D (x, y) and sum image information S (x, y) are obtained by performing difference calculation and sum calculation for each corresponding pixel in the images of the two differential interference images, and the amount of retardation When the retardation amount of the polarization component detected by the detecting means is θ and the phase amount on the observation object plane corresponding to each image information is Φ (x, y), one of the following equations is used. By detecting the differential value ∂Φ (x, y) / ∂r of the phase amount on the observation object plane corresponding to the separation direction r of the two polarization components and performing integration processing in the r direction, the observation object plane is obtained. A method of detecting the upper phase quantity Φ (x, y).
∂Φ (x, y) / ∂r = k · {(1-cos θ) · D (x, y)}
/ {2sinθ · S (x, y)}
∂Φ (x, y) / ∂r = k · tan ⁻¹ [{(1-cos θ) · D (x, y)}
/ {2 sin θ · S (x, y)}]
However, regarding k, when the wavelength of the light emitted from the light source of the detection device is λ, k = λ / 2π when observing the observation object, and k = λ / when reflecting the observation object. 4π.