Patent Grant
7489395
The present invention relates to inspection of patterns and/or foreign matters, for detecting or examining defects, such as short-circuit and/or open-circuit or the like, on the patterns as a target of inspection, and relates, in particular, to a method and an apparatus, for inspecting the defects and/or foreign matters of the patterns formed on, such as, a semiconductor wafer, a liquid crystal display, and a photo-mask, etc., for example. Hereinafter, it is assumed that the defects include the foreign matters, in the meaning thereof.
Conventionally, in such kind of an inspection apparatus, as is described in Japanese Patent Laying-Open No. Hei 7-318326 (1995) <hereinafter, conventional art 1>, an image is detected on the patterns to be examined or inspected by means of an image pick-up element, such as a line sensor, etc., while moving the patterns to be inspected, so as to compare an image signal detected with one which is delayed by a predetermined time in gradation thereof, thereby acknowledging inconsistency or anti-coincident to be a defect.
Also, other conventional art, relating to the defect inspection of patterns to be inspected, is known from Japanese Patent Laying-Open No. Hei 8-320294 (1996) <hereinafter, conventional art 2>. In this conventional art 2 is described a technology, for examining the patterns to be inspected, for example, a semiconductor wafer, in which areas high pattern density, such as the memory mat portions, etc., and low pattern density, such as peripheral circuit portions, etc., are mixed with each other, wherein gradation conversion is conducted on digital image or video signal, which is obtained through A/D conversion of the image signal detected, so that a predetermined relationship can be established between the high density areas and the low density areas on the patterns to be inspected, in particular in the brightness or the contrast thereof, rather than the frequency distribution thereof, and this image signal converted in gradation is compared with other gradation conversion image signal for comparison, under the condition of being fitted to each other in position, thereby inspecting minute or microscopic defects thereon with high accuracy.
Furthermore, the conventional art for inspecting the patterns on a photomask is already known, for example, Japanese Patent Laying-Open No. Hei 10-78668 (1998) <hereinafter, conventional art 3>. In this conventional art, it is described that the UV light rays are irradiated upon the mask, equally, in which coherence is lowered by rotating a diffuser panel inserted within an optical path, with using a UV laser beam as a light source, such as, an excimer laser or the like, so as to calculate out a characteristic amount or quantity from the obtained image data of the mask, thereby deciding or examining good or bad, i.e., the quality thereof. And, a projection exposure apparatus with using an excimer laser therein is already known, for example, in Japanese Patent Laying-Open No. Sho 59-226317 (1964) or Japanese Patent Laying-Open No. Sho 62-231924 (1987), etc.
For LSI manufacturing in recent years, the circuit patterns formed on a wafer come to be minimized, such as, 0.25 μm or less than that in the pattern width thereof, responding to the needs of high integration; thus it nearly reaches up to the limit on the resolving power of an image-forming optical system. For this reason, development is advanced on application of high NA technology or an optical super resolving power technology.
However, the NA technology also reaches up to the limit thereof, physically. Accordingly, a substantial or intrinsic approach is trying to shift the light wavelengths to be applied for the detection into the regions of the UV light and DUV light, i.e., shortening the light wavelength.
Further, due to the necessity of conducting the inspection with high speed, it is impossible to adopt the method of scanning the laser beam focused or converged thinly on a sample. On the contrary, when trying to illuminate with irradiating the laser beam spreading out all over a visual field, however the illumination of the laser beam generates speckles, or an overshoot or an undershoot, being called by a “wringing”, at an edge portion of the circuit pattern, thereby bringing about a problem that the image cannot be obtained of good quality.
An object, according to the present invention, for dissolving such the drawbacks of the conventional arts mentioned above, is to provided a method and an apparatus, for inspecting or examining the microscopic circuit patterns with high resolving power at high speed, thereby achieving inspection of the defects.
Also, other object, according to the present invention, is to provide a manufacturing method of semiconductor devices, with applying the method and the apparatus for inspecting the pattern defects mentioned above, thereby enabling the manufacture of the super microscopic semiconductor devices.
According to the present invention, in a pattern defects inspecting apparatus using a UV light source or a UV laser beam source as a light source, being provided with a light-amount adjustment optical system for adjusting a light-amount of the UV light or the UV laser beam and a coherence reduction optical system for suppressing generation of speckles of the UV light or the UV laser beam, in an optical path thereof, wherein the UV light lowered in coherency is irradiated upon the surface of a target, thereby detecting an image of the target to be inspected. Herein, the UV light includes a DUV light therein.
Namely, according to the present invention, a pattern defects inspecting apparatus is constructed with providing an illumination optical system, comprising: a light source for emitting a UV light, a laser light, or a UV laser light, etc.; a light-amount adjustment optical system for adjusting an light-amount of the UV light, the laser light, or the UV laser light; a slit-like light flux optical system for forming the UV light, the laser light, or the UV laser light into a slit-like light flux for fitting to a light receiving portion, such as a TDI image sensor, or the like; and a coherence reduction optical system for lowering coherency of the UV light, the laser light, or the UV laser light, emitting from the slit-like light flux optical system, and further being provided with a detection optical system, having an image sensor for detecting a detected image signal by picking up an image of reflection light from the sample irradiated by the illumination optical system, and an image processing portion for detecting the defects on the patterns formed on the sample upon the basis of information relating to the image signal of the sample detected by the detection optical system.
Also, according to the present invention, in the pattern defects inspecting apparatus mentioned above, it is characterized that as an image sensor of the detection optical system is applied an image sensor of time delay integrated (TDI) type having a sensitivity to the UV light.
And also, it is characterized that the image sensor of the time delay integrated (TDI) type is an anti-blooming TDI sensor, a surface irradiation type TDI sensor, in which an organic thin-film coating is treated on a cover glass thereof, or a reverse surface irradiation type TDI sensor.
Also, according to the present invention, there is provided a method for inspecting defects on pattern formed on a sample, comprising the following steps of: irradiating the UV laser light lowered in coherency thereof upon surface of a sample; obtaining an image signal by picking up an image of the surface of the sample irradiated by the UV laser light; detecting defects on the sample, being equal or less than 10 nm in size, by processing the image signal; and outputting information relating to position on the sample with respect to the detected defects being equal or less than 100 nm in size.
Further, according to the present invention, there is provided a method for inspecting defects, comprising the following steps of: irradiating the UV laser light lowered in coherency thereof upon a wafer having a diameter of 200 mm; detecting an image of the wafer by picking up an image of the wafer irradiated with the UV laser light; and detecting defects on patterns formed on the wafer, being equal or less than 100 nm, by processing the detected image of the wafer at throughput of equal or greater than three (3) pieces per an hour.
And also, according to the present invention, in the method for inspecting pattern defects, on the sample are formed with patterns having repetitiveness.
Those and other objects, features and advantages of the present invention will be apparent from the following more particular description of preferred embodiments of the invention, as illustrated in the accompanying drawings.
a) and 2(b) are views for showing illumination conditions on a pupil and a visual field of an objective lens for detection, under the illumination by means of a discharge tube;
a) to 3(d) are views for showing illumination conditions on a pupil and a visual field of an objective lens for detection, a pattern on the visual field and a signal detected therefrom, but under the illumination by means of laser radiation;
a) and 4(b) are views for showing illumination conditions on a pupil and a visual field of an objective lens for detection, under the illumination by means of laser radiation, being widened or spread on the pupil thereof;
a) to 5(d) are views for showing illumination conditions on a pupil and a visual field of an objective lens for detection, under the illumination by means of laser radiation according to the present invention;
a) and 8(b) are views for showing a relationship in an attachment of a laser beam source, according to the present invention;
a) and 9(b) are views for showing a adjustment manner in the attachment of the laser beam source, according to the present invention;
a) to 12(d) are views for explanation about an idea of exchanging the illumination area or region, according to the present invention;
a) and 14(b) are views for explanation on an aperture diaphragm or stop, according to the present invention;
a) to 16(c) are also views for explanation about the idea for lowering spatial coherence of the illumination by laser radiation, according to the present invention;
a) to 18 (c) are views for explanation about a second idea for lowering spatial coherence of the illumination by laser radiation, according to the present invention;
Hereinafter, embodiments, i.e., a method and an apparatus for inspecting defects on patterns to be tested, according to the present invention will be fully explained by referring to the attached drawings.
A light flux L1 from the laser beam source 3 is reflected upon a mirror 4, which is rotated around an optical axis at a desired position, and is limited or restricted in an amount, down to that necessary for inspection, by means of a ND filter (an adjustment optical system for light amount) 5 for restricting the amount of lights. The UV laser light is expanded to a certain size of light flux by means of a beam expander 6.
An optical system for exchanging optical illumination light path (an optical system for exchanging illumination area) 7 is used for setting the area of illumination upon the sample 1, exchanging between when inspecting and when adjusting an inclination. Thus, the optical system for exchanging optical illumination light path (the optical system for exchanging illumination area) 7 exchanges the illumination area, so that an illumination is made on a rectangle area or region (i.e., a slit-like area) 71 covering an image sensor 51 when inspecting, while it is made all over a detection area of a detector 18 when adjusting the inclination. However, when inspecting, in the optical system 7 for exchanging optical illumination light path, an aperture diaphragm 8 is set at a conjugate point with a pupil 14a of an objective lens 14, thereby functioning to restrict the NA incident upon the pupil 14a. Through a lens 9, the laser beam is guided to an optical system 10 for lowering coherence.
The coherence lowering optical system 10 is provided for lowering the coherence of the laser beam emitted from the illumination light source 3. This coherence lowering optical system 10 may be a one, which can lower the time or the spatial coherence. As such the coherence lowering optical system 10, for example, it may be built up with a mechanism for scanning the laser beam from the illumination light source 3 on the pupil of the objective lens 14. An image of the laser beam from the coherence lowering optical system 10 is formed on the pupil 14a of the objective lens 14 through a lens 11.
A beam splitter 12 may be constructed with a polarization beam splitter in a case, and is constructed, so that the UV illumination light from the UV illumination light source 3 reflects thereupon, to be given upon the sample 1 through the objective lens 14, thereby giving a bright-field illumination thereupon, for example. When the beam splitter 14 is constructed with the polarization beam splitter, it has a function of reflecting the UV laser beam thereupon if the polarization direction of the UV laser beam is in parallel with the reflection surface, while passing it therethrough if the polarization direction is perpendicular to the reflection surface. Accordingly, since the UV laser beam is inherently a polarized light, it is possible to bring about total reflection of the UV laser beam, by means of the polarization beam splitter 14.
A group 13 of polarization elements has a function of adjusting a ratio pf polarizing of the illumination light arbitrarily, through controlling the polarization directions of the UV laser illumination light and the reflection light, so that the reflection light will not reach to an image sensor 20 with accompanying a blur or unevenness in brightness, due to the shape of pattern and/or the difference in density, and it may be constructed with a ½ wavelength plate and a ¼ wavelength plate, for example.
And, a beam splitter 16 and a lens 17, as well as, the detector 20, which are provided in the optical path defined between the beam splitter 12 and the image sensor 20, are so constructed, that an image with a wide visual field can be obtained by means thereof. Thus, when adjusting the inclination of the sample, the wide field image of the sample 1 is formed on the detector 18 by means of the lens 17, for enabling observation (picking-up of pictures) of the wide field image thereof.
Also, by means of an image-forming lens 19 is formed an image of the reflection light on the image sensor 20, which is detected by the objective lens 14. Insertion of a mirror 21, which can be inserted, on the optical path between the image-forming lens 19 and the image sensor 20 enables an image-forming of the sample 1 on the detector 22, being same to that formed on the image sensor 20.
In this manner, by setting a rotation angle of the ½ wavelength plate and the ¼ wavelength plate, respectively, with rotating them around the optical axes thereof, upon the basis of the spatial image formed on the pupil surface, which is detected from the detector 18, a CPU (not shown in the figure) controls the polarization condition of the reflection light generated from the circuit patterns formed on a semiconductor wafer 1, i.e., the diffraction light in the reflection light; therefore, it is possible to detect the image, for example, with reducing the diffraction light of 0th order but hardly reducing diffraction lights of other higher orders, by means of the image sensor 20. As a result, the image of the patterns can be improved progressively in the contrast, and thereby stable detection sensitivity can be obtained. However, for the purpose of observation (picking-up of the pictures) of the spatial image on the pupil surface 14a of the objective lens 14 by means of the detector 18, it is necessary to replace the lens 17 with a lens or the like, which can form the spatial image on the pupil 14a on the detector 18.
As the objective lens 14, because of adoption of the UV light or the DUV light therein, it is possible to lessen an ill influence due to chromatic aberration by using a reflection objective lens. Also, when irradiating the UV light upon the sample, as the objective lens 14 is used one having the NA greater than 0.75.
The image sensor 20 has pixels, each being from 0.05 μm to 0.3 μm in the size, converted onto the sample, and providing an output of an image signal of density (or gradation) depending upon the brightness (gradation) of the reflection light from the semiconductor wafer 1, as one example of the patterns to be inspected. However, those optical systems are expanded on an optical frame or base not shown in the figure, and are constructed with the laser and the illumination optical system, the detection optical system, and the sensors, in one body. The optical base is provided on an upper portion of the stage 2, but in the manner not shown in the figure. For this reason, it is impossible to obtain a stable detection thereof, in particular with respect to change in temperature, and disturbance due to vibration, etc.
With such the structure mentioned above, the UV light (for example, the UV laser beam) L1 emitted from the illumination light source 3, being reflected upon the mirror 4 and penetrating through the ND filter 5 for limiting the light amount, is expanded by the beam expander 6, to be incident upon the objective lens 14 through the coherence reducing optical system 10, the lens 11, the beam splitter 12 and the group 13 of polarization elements, thereby being irradiated upon the sample (i.e., the semiconductor wafer) 1. Namely, after being converged in the vicinity of the pupil of the objective lens 14 by means of the lens 11, the UV light L1 is given upon the sample 1 as the Koehler illumination. The reflection light from the sample 1 is detected by the image sensor 20, after passing through the objective lens 13, the group 13 of polarization elements, the beam splitter 12, and the image-forming lens 19, sequentially from the above of the sample 1 in the vertical direction.
When being inspected, the sample 1 is always detected on the surface to be tested, i.e., the position in the Z direction thereof, in a focus detector system 15, through a method not shown in the figure, while moving the semiconductor wafer 1, as an example of the pattern to be tested, at a constant speed with scanning the stage 2, thereby controlling the stage 2 in the Z direction so that the distance between the objective lens is constant. The image sensor 20 detects brightness information (i.e., gradation image signal) of the pattern to be tested, which is formed on the semiconductor wafer 1.
A signal processor circuit 24 is constructed with an A/D converter 200, a gradation conversion portion 201, a video filter 215, a delay memory 202, an position alignment portion 203, a local gradation conversion portion 204, a comparing portion 205, a CPU 212, an image input portion 206, a scatter diagram producing portion 207, a memory means 208, a display means 209, an output means 210, and an input means 211, etc.
The A/D conversion portion 200 converts gradation image signal, which can be obtained from the image sensor 20, into digital video signal, thereby providing an output of video signal of the sample. For example, ten (10) bits data is adopted. The gradation converter 201 performs such the gradation conversion described in Japanese Patent Laying-Open Hei 8-320294 (1996) upon the 10 bits digital video signal, which is outputted from the A/D converter 200. Thus, the gradation converter 201 treats it with the logarithm transformation, the exponential transformation, or the multinomial transformation, etc., thereby compensating the image, and it is so constructed that it can provide an output of eight (8) bits digital signal, for example.
The video filter 215 is a filter provided for removing noises being characteristic to the image detected under the UV light from the image, which is gradation-converted and compensated, with high efficiency. The delay memory 202 is a memory portion for memorizing a reference image signal therein, and it memorizes the output image signal from the video filter 215, for a one cell or a plural number of cells, or for an one tip or a plural number of tips, being repetitively formed on the semiconductor wafer, so as to delay. Herein, the cell constitutes a unit of repetition within the tip. However, the video filter 215 may be provided after passing the delay memory 202. The position alignment portion 203 is a portion for detecting an amount of positional shift or deviation between the image signal (i.e., the detected image signal obtained from the sample) 213 treated with the gradation conversion, which is outputted from the gradation converter 201, and the delayed image signal (i.e., the reference image signal) 214 obtained from the delay memory 202, through the normal function, thereby aligning the position by a unit of the cell. The local gradation conversion portion 204 is a portion for conducting the gradation-conversion upon both or one of the image signals, being different in the characteristic quantity (i.e., the brightness (the difference in tone), the contrast, the differential value, the standard deviation, the texture, etc.) of a normal portion, locally (i.e., for each predetermined area or region), so that the characteristic quantity of the said normal portion coincides with each other locally (i.e., for each predetermined area or region). Thus, both or one of the image signals are compensated through the local gradation-conversion, so that the characteristic quantity of the detected image signal and that of the reference image signal are coincident with each other, with the normal portion. With doing this, if a difference is generated between the detected image signal and the reference image signal, in the brightness, the contrast or the like, it is possible to bring them to be coincident with, through the local gradation-conversion, and therefore, without generating erroneous information, it is possible to detect the defects with high sensitivity, by means of a decision threshold value upon basis of the difference in the characteristic quantity thereof.
The comparison portion 205 is a portion for comparing the detected image signals by themselves, which are treated with the gradation-conversion in the local gradation-conversion portion 204 locally (i.e., for each predetermined area or region), thereby detecting the defects or the defect candidates by means of the decision threshold value upon basis of the difference in the characteristic quantity. Thus, the comparison portion 205 compares the reference image signal, which is outputted from the delay memory 202 and is delayed by an amount corresponding to the cell pitch, etc., with the detected image signal.
By inputting the coordinate, such as alignment data or the like on the semiconductor wafer 1, through the input means 211, being constructed with, such as a keyboard, or a disc, etc., the CPU 212 produces defect inspection data upon the basis of the coordinate, such as the alignment data, etc., on the semiconductor wafer 1, and stores it into the memory means 208. This defect inspection data can be displayed on the display means 209, such as a display, if necessary, and also can be outputted to the output means 210. Further, the details of the comparison portion 205 may be such, as shown in Japanese Patent Laying-Open No. Sho 61-212708 (1986), for example, and may be constructed with a position alignment circuit for an image, a differential image detector circuit for the image aligned in position, an anti-coincidence detector circuit for digitizing the difference image, and a characteristic extractor circuit for calculating out an area, a length (a projection length), or the coordinate, etc., from the digitized output. In brief, from the comparison process portion 205 is outputted information in relating with the position and the sizes of the defects on the pattern to the CPU 212, therefore the defect inspection data is produced in the CPU 212.
The image input portion 206 is a portion for inputting in synchronism or in non-synchronism, for the purpose of producing scattered diagrams for both images, which are aligned in the position by a unit of the cell in the position alignment portion 203. The scattered diagram producing portion 207 is provided for producing the scattered diagrams, between the characteristic quantity of the detected image by a unit of pixel and the characteristic quantity of the reference image by a unit of pixel, for respective categories (i.e., the brightness (the difference in tone), the contrast, the differential value, the standard deviation, the texture), for each the predetermined area (locally), with respect to the both images inputted in the image input portion 206, thereby for displaying them on the display means 209, for example.
An example of the video filter 215 will be explained.
F(i,j)=B·⅛+D·⅛+F·⅛+H·⅛+E·⅛ (Eq. 1)
The size and the coefficient of the filter can be changed flexibly, by using a lookup table thereof. As another example, there is known a median filter. With this, taking a central value among the brightness values in the vicinity where the setting is made, it is possible to remove an ill influence of a singular point. Alternatively, in other example is used the Gaussian function. With this, smoothing of image is conducted through the convolution; i.e., first convoluting an average 0 upon the image f(x,y), and then two (2)-dimensional Gaussian function (Eq. 2) upon (Eq. 3).
G(x,y)=(½πσ2)·exp(−(x2+y2)/2σ2) (Eq. 2)
F(x,y)=G(x,y)f(x,y)=∫∫G(x+u,y+v)·f(x,y)dudv (Eq. 3)
where, means the convolution.
Also, in other example, it is possible to remove the noises generating irregularly, by using the Fourier transform.
Next, restoration is made (282) on the image, which is deteriorated through the noise removal (281). As an example of this, the restoration is made by means of the Wiener filter. In this, the image is given, so that an averaged square error between the image f(x,y) after being inputted and the image f′(x,y) after the restoration comes to the minimal.
Further, it is checked whether there is a large difference or not, in a view between the detected image and the reference image to be compared with. Indexes for evaluation include the characteristic quantity (i.e., the brightness (tone value), the contrast (the brightness value), the dispersion of brightness (the standard deviation), the frequencies of noise components, the texture). Then, upon the detected image and the reference image, which are aligned in the position by a unit of pixel, the characteristic quantities are calculated out (283) for the respective categories; i.e., the brightness (tone value), the contrast, the dispersion of brightness (the standard deviation), the frequencies of noise components, the texture, etc., for each the predetermined area corresponding thereto (for each location) In the normal portion, if there is local difference between the detected image and the reference image in any one of the characteristic quantities, then first of all, sequential image compensation (i.e., sequential gradation-conversion) is conducted, locally, so that those characteristic quantities come close to each other. Thus, for each category, the characteristic quantities (such as, the tone value, or the brightness value, etc.) of the detected image and the reference image are compared with, for each area corresponding thereto, thereby extracting the difference therebetween. For example, for each category, and for each area, the scattered diagram is produced, by depicting the brightness value, as the characteristic quantity of the detected image by a unit of pixel of, on the horizontal axis, while the brightness value of the reference image, as the characteristic quantity by a unit of pixel, on the vertical axis. Then, compensation coefficients (a,b) are obtained upon the basis of an intersection point b between an inclination a of a line drawn in the area mentioned above, by fitting the straight line thereto (i.e., a linear approximation), for example, on the scattered diagram that is produced for each area. And, for example, compensating one of the images, i.e., the detected image f(x,y) (i.e., fitting of the image) (285) through the compensation coefficient mentioned above, for each area, enables to obtain the detected image signal f′(x,y), being coincident with the reference image in the characteristic quantity thereof, for each area, if it is the normal portion. Namely, as a result of the compensation, the detected image and the reference image are scattered under the condition that they are restricted onto a straight line of 45 degree in the scattered diagram, and thereby setting a decision threshold value for deciding to be the defects or the defect candidates, with respect to the difference (the difference image) between the detected image and the reference image, by a straight line having a certain width to that of 45 degree. In this manner, the decision threshold value can be obtained from the scattered diagram.
f′(x,y)=a·f(x,y)+b
In this fitting of image, the Wiener filter mentioned above may be applied between the detected image and the reference image. Also, in video processing, if the image is at a level where the fitting of the local characteristic quantity is impossible, the sensitivity is lowered down, by increasing the decision threshold value in the comparison process portion 205, for extracting the defects and the defect candidates with respect to the difference (the difference image) between the detected image and the reference image, thereby suppressing generation of the erroneous information (286).
Further, the method for calculating the defects in the video processing portion 24 can be realized, in the details thereof, by that shown in Japanese Patent Laying-Open No. 2000-194323, for example.
Next, explanation will be given on the illumination light source 3. Although shortening is necessary, in particular, in the wavelength thereof, for achieving the high resolution, it is thought appropriate to use the laser for the light source, as a means for obtaining an illumination of high luminance within wavelength region of the UV, being most effective in the effect thereof. As was mentioned in the above, adopting the UV laser light as the light source brings about a large merit. According to the present invention, there is shown a method of using the illumination by means of the UV laser light.
a) and 2(b) show the condition of illumination on the pupil of the objective lens and a visual field, when being illuminated under an ordinary white light. AS in
Next, in
Under the illumination of an ordinary white light, an illumination is performed to have a certain size on the pupil AS, thereby illuminating to the visual field FS from the directions having an angle range corresponding to the NA (i.e., the number of aperture) of the objective lens 14. Under the coherent light (i.e., having coherency), such as, the laser light, σ (in proportion to the size of a light source on the pupil) comes to be zero (0). This is because an image of the light source is a point with such the coherent light, therefore the image on the pupil becomes a point. Of course, as shown in
Accordingly, there is a necessity of a means for lowering the coherence of the UV laser light. For the purpose of lowering the coherence, it is sufficient to reduce either one of the time coherence or the spatial coherence thereof.
Then, according to the present invention, as a function of alight scanning mechanism (or a light scanning system) for scanning on the pupil, which constructs a light modulator as the coherence reducing optical system 10, it is proposed that illumination 40 is made upon the visual field shown in
As shown in
a) and 8(b) show the structure of the laser beam source. The laser beam source 3 is fixed on a plate 102. A plate 101 is fixed with aligning to an optical base 100. Position alignment is made, for example, by fixing with using a pin 103 fixed on the optical base 100 as a guide. Herein, it is assumed that the pin 103 is adjusted with respect to an optical axis of the optical system. The plate 102 is fixed on the plate 101. The laser beam source 3 happens to be replaced with new one, due to the lifetime of a laser oscillator. When the laser beam source 3 is replaced, there is a probability that the stoppage time of the apparatus comes to be long if adjusting the optical axis of the optical system. For this reason, preferably, the laser beam source 3 is completed in adjustment of the optical axis thereof before it is mounted for replacement, so as to obtain a desired performance at the lowest.
a) and 9(b) show an example of a jig for adjustment of optical axis. An optical axis adjustment base 104 is fixed by the pin 103 on the position being same to that of the optical base 100. Targets 105 and 106 are same in the height, and are fixed in parallel relationship with the position of the pin 103, while each being opened with a pinhole 107 for passing the laser beam therethrough. The distance between the targets 105 and 106 has a sufficient length for adjusting the laser beam source 3. In this structure, the plate 101 is fixed onto the optical adjustment base 104. The laser beam source 3 is provisionally fixed on the plate 102 in advance. Then, the plate 102 is mounted on the plate 101, thereby emitting the laser beam therefrom. The light flux L is adjusted in the right-hand and the left-hand side directions (i.e., in the horizontal direction) by means of the plate 102, so that it can pass through the pinholes 107 of the targets 105 and 106, while by the laser beam source 3 itself in the gate direction (i.e., in the vertical direction). After the adjustment, the plate 102 and the laser beam source 3 are fixed, and further the plate 102 is fixed onto the plate 101. In this manner, the light flux of the laser beam source 3 is adjusted upon the basis of the pin 103. Thereafter, replacing with the laser beam source 3 fixed on the plate 101 onto the optical base 100 brings about the coincidence of the optical axis thereof.
Next, explanation will be given on the ND filter (a light amount adjustment optical system) for restricting (adjusting) a light amount, by referring to
Next, explanation will be given on the optical path exchanger system (i.e., illumination area exchanger means) 7.
First, explanation will be given when inspecting. As was mentioned previously, because of use of the linear sensor, it is effective to perform a rectangular illumination 71 (i.e., illumination with using a slit-like light flux), irradiating all over the surface in the pixel direction while focusing or condensing onto the sizes of the sensor opening in the scanning direction, within the range of illumination when inspecting. For this reason, for forming a slit-like spot light (i.e., the light flux being focused or converged in the width direction of the linear sensor, such as the TDI sensor, etc.), the homogenizer 74 is inserted after emission of the light from the beam expander 6, as shown in
By the way, the sample 1 is shifted onto the stage 2 by a method not shown in the figure. The sample 1 must be positioned, so that the patterns formed on the sample 1 are correctly aligned or fitted to the inspection direction of the apparatus, i.e., the scanning direction, for reducing the detection errors. For that purpose, it is necessary to compensate an inclination of the sample 1. In the present embodiment, a method will be explained, in which the inclination compensation is carried out by using a detector 18, which is provided in the optical path. First, by using tips at both ends of the sample 1, alignment marks provided within the tips, but not shown in the figure, are detected, so as to calculate out the inclination of the sample 1. Since the inclination of the sample 1 is unknown, it is necessary to enlarge the detection range of the detector 18, by making the optic magnification smaller than that for use in the inspection. As was mentioned previously, since the illumination area for use in the inspection is rectangular, the target falls down into a condition where no illumination is made in the direction perpendicular to the scanning direction when it is detected by the detector of wide visual field. For this reason, it is necessary to make illumination all over the detection area of the detector.
Explanation will be given on the illumination when compensating the inclination. In
Next, explanation will be given on the aperture diaphragm system.
Next, explanation will be given on the coherency reduction optical system 10. In the present embodiment, the reduction of coherency is achieved by a method of providing a diffuser panel within the laser optical path.
Also, similar effect can be obtained by provision of a rotational phase plate in the place of the diffuser plate 92 mentioned above.
Next, the structure of the optical path exchanger mechanism will be shown.
However, the image-formation lens 19 is used, by changing it to have the magnification corresponding to the pixel size. Thus, when the pixel size differs, not the objective lens 14, but the image-formation lens 19 is exchange. Further, the image-forming positions of the image-formation lenses 19a, 19b and 19c, having the respective magnifications, are set at the same position, irrespective of the difference in magnification thereof. Therefore, there is no necessity to change the position of the sensor 20 if the magnification is different, thereby enabling a stable detection. Also, the magnification is determined through the focal distance of the objective lens 14 and those of the image-formation lenses 19a, 19b and 19c, and the pixel size is determined by the size of aperture of the image sensor 20. However, due to errors occurring in manufacturing the objective lens 14 and the image-formation lenses 19a, 19b and 19c, as well as in install thereof, there is a probability that the magnification differs for each lot. Therefore, there will occur a problem that the detection sensitivity differs for each lot for the optical system, thus, for each apparatus. For this reason, each of those image-formation lenses 19a, 19b and 19c has a mechanism for making the focal distance thereof being variable.
Next, explanation will be given on an example of the TDI sensor, being able to detect the UV light, in particular the DUV light in this manner.
For further improving the sensitivity to the DUV light, it is preferable to adopt an image sensor of reverse-surface irradiation type.
Next, explanation will be given on the installation or attachment of the image sensor 20.
Further, in addition to the high resolution by means of the UV light, an explanation will be given on a method for improving the pattern contrast, through controlling the polarizing elements group 13, as was mentioned previously. For the purpose of increasing the pattern contrast, by paying attention to the fact that the polarization condition of the UV light can be controlled freely upon the basis of the control of the polarizing elements group 13, the direction of polarization of the illumination light, i.e., the elliptical ratio is controlled, thereby enabling the detection of a part of polarized light elements of the detected light by means of the image sensor 20. As characteristics of the illumination by the UV light, there are facts that it has a single wavelength, and that it is a linearly polarized or plane polarized light. For this reason, it is possible to control the polarization condition, with high efficiency, by means of the polarizing elements group 13, including a ½ wavelength plate and a ¼ wavelength plate provided within the optical path. The control can be achieved by rotating the ½ wavelength plate and/or the ¼ wavelength plate. Since the pattern contrast changes greatly depending upon the polarization condition of illumination, performances of the optical system can be improve, by bringing the polarization condition to be controllable (i.e., positioning by rotation of the wavelength plate(s)). In more details, the linearly polarized light can be controlled in the direction by means of the ½ wavelength plate, while the elliptical ratio by means of the ¼ wavelength plate. With this, it is possible to obtain an improvement of the detection sensitivity. With combination of those, it is also possible to achieve the horizontal Nicol and the vertical Nical. Of course, it is also possible to achieve the circularly polarized light. However, those do not depend upon the illumination wavelength itself. Also, if such an idea mentioned above can be established, the structure may be in any kind for achieving it. Of course, a spatial filter may be disposed at the position being conjugate with the pupil of the objective lens 14, thereby attenuating the 0th dimensional light (or, it may be possible to guide the scattered light from the foreign matter into the image sensor, through blocking the diffracted light from the pattern by means of the spatial filter). However, controlling the polarized light enables more effective extraction of the diffracted light therefrom. According to experiments made by the present inventors, it is clear that the contrast can be improved by about 20%-300%. Also, an analyzer (not shown in the figure) may be positioned between the sample 1 and the image sensor 20, to be controlled on rotation thereof, thereby detecting the image of the sample from the reflection light while controlling the polarization condition in the detection optical system.
As was explained in the above, using the DUV light, such as that having wavelength of 266 nm, 248 nm or 192 nm, enables the detection of defects on the device, being under less than 0.07 μm rule. Also, it can be applied to the inspection of Cu damascene, as a target to be inspected, Further, the inspection can be made even onto a portion where no pattern is formed on the target to be inspected, since no erroneous information is produced when comparing between the detected image and the reference image because of no generation of speckles.
However, with the UV light less than 365 nm in wavelength, to be used as the illumination light, since it has high optical energy, contaminants, such as of organic material or the like, are decomposes or react upon due to the energy thereof when the UV light is irradiated upon the optical elements, thereby attaching or adhering upon the surface thereof. For this reason, with provision of a means for discharging air compellingly upon the surface of the optical system by a manner not shown in the figure, and/or a means for puffing air compulsively thereupon, it is possible to prevent the optical parts from deterioration thereof.
Moreover, in the present embodiment, the explanation was given only on the structure of the light-field optical system, however it is also possible to obtain a similar effect if applying the structure of a co-focal microscope into the detection optical system.
According to the present invention mentioned in the above, it is possible to obtain a UV or DUV illumination of high brightness, thereby enabling a pick-up of image with high resolution in short time-period, and as a result, it is possible to obtain an inspection apparatus with high speed and high sensitivity. The defects on patterns detected are outputted with information of the positions and sizes thereof. In particular, the target (the sample) 1 to be inspected mentioned above includes the damascene, of such as Cu, etc., being formed by burying conductive metal, such as Cu, etc., into openings and/or grooves for via (or contac) -holes and/or wirings formed on the insulator film, such as of SiO2, and thereafter removing excess portion deposited or accumulated through polishing by means of CMP, etc. Accordingly, the inspecting method and the apparatus thereof, according to the present invention, can be applied to the damascene of, such as Cu, etc. Also, the inspecting method and the apparatus thereof, according to the present invention, in particular using the DUV light (the light having wavelength of 266 nm, 248 nm, or 193 nm), can be applied very effectively, to the device under the designing rule of 0.07 μm, since it can detects super microscopic or minute defects less than 0.07 μm in size.
Also, according to the present invention, in the method for inspecting the defects on patterns formed on the sample, the UV laser beam is irradiated upon the sample surface, which is lowered in the coherency thereof, and then the image or video signal is processed, being obtained via image pick-up upon the surface of the sample, upon which this UV laser beam is irradiated, therefore it is possible to detect the defects less than 100 nm on the sample, and thereby to obtain information relating to the positions on the sample about the detected defects less than 100 nm.
And also, according to the present invention, the defects less than 100 nm can be detected, on patterns formed on the wafer having a diameter of 200 mm, with throughput of three (3) pieces per one (1) hour.
And further, according to the present invention, it is possible to obtain an effect that an image can be obtained with much higher sensitivity and higher speed, being equal or superior to that obtained under the ordinary discharge tube illumination, by means of the laser beam source, under illumination of short-wavelength necessary for high resolution, and further being able to achieve the and being advantageous for an actual practice thereof, thereby enabling the detection of microscopic defects with high sensitivity.
The invention may be embodied in other specific forms without departing from the spirit or essential characteristics or features thereof. The present embodiment is therefore to be considered in all aspects as illustrative and not restrictive, the scope of the present invention being indicated by the appended claims rather than by the foregoing description and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2001-277681 | Sep 2001 | JP | national |
This application is a continuation application of U.S. application Ser. No. 11/192,021, filed Jul. 29, 2005, now U.S. Pat. No. 7,110,105, which is a continuation application of U.S. application Ser. No. 10/218,463, filed on Aug. 15, 2002, now U.S. Pat. No. 6,927,847, the contents of which are incorporated by reference, and is related to U.S. application Ser. No. 10/650,756, filed Aug. 29, 2003, now U.S. Pat. No. 6,900,888, and U.S. application Ser. No. 11/131,379, filed May 18, 2005.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4634240 | Suzuki | Jan 1987 | A |
| 5774222 | Maeda et al. | Jun 1998 | A |
| 6031607 | Miyazaki | Feb 2000 | A |
| 6084716 | Sanada | Jul 2000 | A |
| 6091488 | Bishop | Jul 2000 | A |
| 6288780 | Fairley et al. | Sep 2001 | B1 |
| 6369888 | Karpol et al. | Apr 2002 | B1 |
| 6621571 | Maeda et al. | Sep 2003 | B1 |
| 6900888 | Yoshida | May 2005 | B2 |
| 6927847 | Yoshida | Aug 2005 | B2 |
| Number | Date | Country |
|---|---|---|
| 59-226317 | Dec 1984 | JP |
| 62-231924 | Oct 1987 | JP |
| 07-318326 | Dec 1995 | JP |
| 08-320294 | Dec 1996 | JP |
| 10-078668 | Mar 1998 | JP |
| Number | Date | Country | |
|---|---|---|---|
| 20070002318 A1 | Jan 2007 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 11192021 | Jul 2005 | US |
| Child | 11518195 | US | |
| Parent | 10218463 | Aug 2002 | US |
| Child | 11192021 | US |
