JP2015025656A - Device for measuring shape formed on substrate - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve problems that evaluation items of a bump may be further increased in the future, but that it is not considered in conventional technology.SOLUTION: The invention comprises: a top part bright field detection system disposed in a normal direction of a substrate; a first bevel bright field detection system disposed with a first bevel to the substrate; and a second bevel bright field detection system disposed with a second bevel to the substrate. The bump is evaluated using the images acquired from the plural bright field detection systems. According to the invention, the amount of information capable of being acquired increases compared with conventional technology, so that the bump can be evaluated more precisely.

Description

  The present invention relates to an apparatus and method for measuring a substrate. The substrate includes various substrates such as a silicon wafer. The evaluation and measurement include, for example, obtaining various shapes and defects formed in the bump formation process, obtaining bump shapes, and evaluating (classifying) what the defects are.

  Semiconductors used in various electronic products are manufactured by performing various processes on wafers. This semiconductor manufacturing process is roughly divided into a pre-process and a post-process. First, the previous process will be described. The pre-process includes wafer oxidation, photoresist application to the wafer, pattern formation by exposure, etching, oxidation / diffusion / CVD / ion implantation, planarization, electrode formation, and inspection. In the pre-process, necessary semiconductor elements are formed on the wafer. Next, a post process is demonstrated. The post-process includes a process of cutting out the wafer processed in the pre-process for each chip, a process of fixing the cut-out chip to the package substrate, and a molding process.

  Here, the step of fixing the chip to the package substrate may use flip chip bonding in which protruding electrodes called bumps are formed on the chip and the chip package substrate is connected via the bumps. In flip chip bonding, the shape of a bump, the presence or absence of defects on the bump, etc. are important because they affect the electrical characteristics between the chip and the package substrate, that is, the electrical characteristics of the semiconductor. Therefore, it is necessary to evaluate the shape formed on the substrate, for example, to form a bump and measure the formed bump (bump measurement).

  As prior art, optical evaluation apparatuses of Patent Documents 1 to 3 have been proposed.

US Pat. No. 7,286,218 US Pat. No. 7,064,821 US Pat. No. 7,126,699

  Bump evaluation items are expected to vary in the future. In the prior art, no consideration has been given to this point.

  The present invention is characterized in that bumps are evaluated by a plurality of imaging optical systems (for example, bright field optical systems).

  According to the present invention, it is possible to evaluate a bump with higher accuracy than in the past.

Configuration of the bump measurement system of this embodiment. The figure explaining image processing. The figure explaining the detail of the evaluation part 128. FIG. The figure explaining from what kind of information the evaluation item and its evaluation item in a 1st process are obtained. The figure explaining the evaluation item in a 1st process, and what kind of information the evaluation item is obtained from (continuation). The figure explaining from what kind of information the evaluation item and its evaluation item in a 2nd process are obtained. The figure explaining from what kind of information the evaluation item and its evaluation item in a 2nd process are obtained (continuation). The figure explaining from what kind of information the evaluation item and its evaluation item in a 3rd process are obtained. The figure explaining the evaluation item in a 3rd process, and what kind of information the evaluation item is obtained from (continuation). The figure explaining from what kind of information the evaluation item and its evaluation item in a 4th process are obtained. The figure explaining the evaluation item in a 4th process, and what kind of information the evaluation item is obtained from (continuation). The figure explaining side wall measurement and the filling defect measurement of a paste material. The figure explaining reading of an image. The figure explaining the positional relationship of the visual field of the one-dimensional line sensor 106, the 2nd one-dimensional line sensor 107, the two-dimensional sensor 118, and the two-dimensional sensor 123. FIG. 6 is a diagram illustrating a second embodiment.

  Embodiments will be described below with reference to the drawings.

  The configuration of the bump measurement system of this embodiment will be described with reference to FIG. The bump measurement system of the present embodiment is arranged with a stage 1300 for scanning with the substrate fixed, an upper bright field detection system 1000 arranged in the normal direction of the substrate, and a first oblique angle with respect to the substrate. A first oblique bright field detection system 1100 and a second oblique bright field detection system 1200 disposed at a second oblique angle with respect to the substrate are provided.

  First, the configuration of the upper bright field detection system 1000 will be described. Light from the illumination system 101 passes through the lens 10 and the half mirror 102 and is illuminated on the substrate through the objective lens 103. Light from the substrate (for example, regular reflection light) is collected by the objective lens 103. The condensed light is imaged by the imaging lens 105 via the variable aperture 104. Note that a beam splitter 106 is disposed on the imaging side of the imaging lens 105, and the light that has passed through the imaging lens 105 is branched into two. The branched light is detected by the first one-dimensional line sensor 106 (SCU) and the second one-dimensional line sensor 107 (SCD) arranged at a focal position different from the one-dimensional line sensor 106 (focus). The position may be the same). Note that a movable total reflection mirror 108, a lens 109, and an observation camera 110 for observing the obtained image are arranged on the imaging side of the imaging lens 105. The illumination system 101 is a lamp light source 1400 for switching between illumination method 1: bright field epi-illumination / dark field illumination using a lamp (bulb) and illumination method 2: parallel illumination using a multipoint light source using LEDs. The multi-point LED light source 1500 and the switching mechanism 1600.

  Next, the first oblique bright field detection system 1100 arranged at a first oblique angle with respect to the substrate will be described. The light from the light source 111 is branched by the half mirror 112. The branched light is further reflected by the total reflection mirror 113 and enters the first oblique bright field detection system 1100 via the lens 20. The incident light is reflected by the half mirror 114 and illuminated by the objective lens 115 on the substrate at the first oblique angle. Light from the substrate (for example, regular reflection light) is collected by the objective lens 115. The condensed light is imaged by the imaging lens 116. An optical shutter 117 made of a transparent ferroelectric ceramic material (hereinafter referred to as PLZT) is disposed on the imaging side of the imaging lens 116. Here, an outline of an optical shutter using PLZT will be described. The bump measurement system of this embodiment has a power source for driving an optical shutter 117 and an optical shutter 122 described later. When a voltage is applied from this power supply, the birefringence of PLZT changes. That is, PLZT serves as a shutter by applying a voltage. Further, the function of the PLZT as a shutter depends on the frequency to which the voltage is applied, and a high-speed response is possible. Note that a liquid crystal shutter or a mechanical shutter may be employed as an alternative to PLZT.

  The light passing through the optical shutter 117 is imaged by a two-dimensional sensor 118 (SB) (for example, a time delay integration type sensor which is a kind of charge storage type sensor). Here, in imaging using the two-dimensional sensor 118, it is possible not to use all the pixels of the sensor but to selectively use a set of some arbitrary pixels according to the design standard information. It is. This arbitrary set of pixels can be expressed as an effective area.

  A second oblique bright field detection system 1200 arranged at a second oblique angle with respect to the substrate will be described. The light from the light source 111 is branched by the half mirror 112. The other branched light enters the second oblique bright field detection system 1200 via the lens 30. The incident light is reflected by the half mirror 119 and illuminated on the substrate by the objective lens 120 at a second oblique angle. Light from the substrate (for example, regular reflection light) is collected by the objective lens 120. The condensed light is imaged by the imaging lens 121. An optical shutter 122 composed of PLZT is disposed on the imaging side of the imaging lens 121. The light passing through the optical shutter 122 is imaged by a two-dimensional sensor 123 (SF) (for example, a time delay integration type sensor). Here, in imaging using the two-dimensional sensor 123, it is possible not to use all the pixels of the sensor but to selectively use a set of some arbitrary pixels according to the design standard information. It is. This arbitrary set of pixels can be expressed as an effective area.

  Here, the variable apertures 104, 40 and 50 can change the depth of focus according to the nature (size, etc.) of the defect to be detected. Further, the variable apertures 104, 30, and 50 can be changed in depth of focus so as to be focused in accordance with the unevenness of the substrate read from design reference information described later.

The first oblique bright field detection system 1100 is disposed on the opposite side of the traveling direction with respect to the scanning direction of the X axis of the stage 1300, and the second oblique bright field detection system 1200 is composed of the X axis of the stage 1300. It is arranged on the traveling direction side with respect to the scanning direction. In other words, when viewed from above the substrate, the upper bright field detection system 1000, the first oblique bright field detection system 1100, and the second oblique bright field detection system 1200 scan the X axis of the stage 1300. It can also be expressed as being arranged along the X-axis scanning direction of the stage 1300 arranged in the direction.
Note that the upper bright field detection system 1000, the first oblique bright field detection system 1100, and the second oblique bright field detection system 1200 may be arranged on the Y axis scanning direction of the stage 1300.

  Further, at least one of the first oblique bright field detection system 1100 and the second oblique bright field detection system 1200 can change the inclination of the entire system based on design criteria information described later. is there. That is, the first oblique angle and the second oblique angle can be arbitrarily changed. More specifically, at least one of the first oblique bright field detection system 1100 and the second oblique bright field detection system 1200 is in accordance with design criteria information (for example, the angle of the bump surface to be detected). ) The first oblique angle and the second oblique angle can be arbitrarily changed. Furthermore, it becomes possible to observe the inner side wall of the hole described later.

  Further, at least one of the two-dimensional sensor 118 and the two-dimensional sensor 123 can move back and forth with respect to the detection optical axis based on design reference information described later. That is, the focal point can be arbitrarily changed. More specifically, at least one of the two-dimensional sensor 118 and the two-dimensional sensor 123 moves back and forth with respect to the detection optical axis according to at least one of the height and the distance of the bump to be detected. Is.

  Further, the two-dimensional sensor 118 accumulates charges in synchronization with the opening / closing of the optical shutter 117. This means that the start of charge accumulation of the two-dimensional sensor 118 is controlled by the opening operation of the optical shutter 117, and the end of charge accumulation is controlled by the closing operation of the optical shutter 117. The charge accumulation time of the charge accumulation of the two-dimensional sensor 118 corresponds to the time from when the optical shutter 117 is opened until it is closed. Further, the two-dimensional sensor 123 accumulates charges in synchronization with the opening / closing of the optical shutter 122. This means that the start of charge accumulation of the two-dimensional sensor 123 is controlled by the opening operation of the optical shutter 122, and the end of charge accumulation is controlled by the closing operation of the optical shutter 122. The charge accumulation time of the charge accumulation of the two-dimensional sensor 123 means that it corresponds to the time from when the optical shutter 122 is opened until it is closed. Furthermore, the images obtained by the two-dimensional sensor 118 and the two-dimensional sensor 123 can be independently read out.

  Next, image processing will be described.

As described above, two images (first image and second image) are obtained from the first oblique bright field detection system 1100. One image is obtained from the first oblique bright field detection system 1100.
The second oblique bright field detection system 1200 can obtain one image. That is, in the bump measurement system of the present embodiment, four images are obtained from one inspection area. The bump measurement system according to the present embodiment obtains information regarding bumps using at least one of the four images.

  More specific description will be given with reference to FIG.

First, how each detection system acquires an image will be described. First, imaging of the first one-dimensional line sensor 106 (SCU) and the second one-dimensional line sensor 107 will be described.
Imaging of the first one-dimensional line sensor 106 (SCU) and the second one-dimensional line sensor 107 is performed by setting the position of the first bump (1st bump) of the first die (1st die) to start imaging from the chip origin. The first image and the second image are obtained by recognizing based on the design value and imaging the first bump.

  The first image obtained by the first one-dimensional line sensor 106 (SCU) is sent by the hole position recognition unit 50 for recognizing the hole position. The first image is also used to generate a trigger for driving at least one of the optical shutter 117 and the optical shutter 122. The generated trigger is sent to the stage synchronous imaging control unit 124. By this trigger, at least one of the optical shutter 117 and the optical shutter 122 is driven, and the two-dimensional sensor 118 and the two-dimensional sensor are synchronized with the opening / closing of at least one of the optical shutter 117 and the optical shutter 122. At least one of 123 will change the accumulation time.

  The next bump (2nd bump) is imaged by adjusting the stage movement amount to the design pitch from the imaging position of the 1st bump, not from the chip origin. The same applies to the Nth bump. Further, for the 1st bump of the next die (2nd die), an interval from the last bump position of the 1st die to the 1st bump of the 2nd die is obtained, and the amount of movement of the stage is adjusted to this interval.

  The image obtained by the observation camera 110 is sent to the alignment processing unit 125 that obtains information for positioning the substrate. Information for positioning the substrate is obtained from an image obtained by the observation camera 110. Information for positioning the substrate is once transmitted to the host CPU 126 and then transmitted to the mechanical control unit 127. Based on the information for positioning the substrate, the stage is driven up and down. Thereby, so-called autofocus is possible.

  Note that the amount of image processing depends on the density of bumps and shapes to be evaluated and the manner in which they are arranged. Therefore, the stage 1300 of this embodiment changes its speed according to the density of the bumps and the shape formed on the substrate surface and the way of arrangement. For example, if the density is low, the speed of the stage 1300 becomes faster, and if the density is high, the speed of the stage 1300 becomes slower.

  A first image obtained by the first one-dimensional line sensor 106 (SCU), a second image obtained by the second one-dimensional line sensor 107 (SCD), and a first oblique bright field detection system 1100 3 and the fourth image obtained from the second oblique bright field detection system 1200 are sent to the evaluation unit 128. Coordinate values from the stage are also sent to the evaluation unit 128 via the stage synchronous imaging control unit 124. It is possible to know which region is being inspected by associating the four images with the coordinate values. It is also possible to select which region should be inspected.

  The four images transmitted to the inspection area selection unit 127 are sent to the evaluation unit 128. Here, bump evaluation is performed using at least one of the four images. Here, for the evaluation of the bump, defect determination for evaluating the presence or absence of defects in the bump (or paste material to be bump), defect dimension measurement, defect classification, height measurement of the paste material to be bump, paste material This includes depth formation of holes to be filled, hole wall angle measurement, and volume estimation of paste material. The evaluation result is transmitted to the host CPU 126 and displayed on the operator as a map by the monitor 129.

  The evaluation result is simultaneously transmitted to the management system 131 via the I / F unit 130 and used for yield management. A substrate transfer device 132 is also connected to the host CPU 126, and the substrate transfer is performed by the substrate transfer device 132 based on information from the host CPU.

  Next, details of the evaluation unit 128 will be described. FIG. 3 is a diagram for explaining the details of the evaluation unit 128. A first image obtained by the first one-dimensional line sensor 106 (SCU), a second image obtained by the second one-dimensional line sensor 107 (SCD), and a first oblique bright field detection system 1100 The third image obtained from the second image and the fourth image obtained from the second oblique bright field detection system 1200 are sent to the buffer unit 300. In the buffer unit 300, information on the distance between the detectors, the third image, and the fourth image are obtained from the first image, the second image, the third image, and the fourth image, respectively. Using the information about the location in the sensor and the position information, an image of the bump (including an image around the bump if necessary) is cut out. That is, four types of bump images are obtained for one bump.

  The image of the cut bump is sent to the feature amount extraction unit 310. In the feature quantity extraction unit 310, for example, feature quantities such as information about luminance in the image are extracted.

  At the same time, the cut image of the bump is sent to the 3D image composition unit 320. The 3D image composition unit 320 creates a 3D (three-dimensional) image of the bump using information on the first and second oblique angles, the depth of focus of each detection optical system, and the focal plane spacing. The

  The feature amount and the 3D image are sent to the processing unit 330 in the evaluation unit 128. The processing unit 330 evaluates the bumps using the feature amount, the 3D image, the design reference information (eg, normal hole outline, bump arrangement, hole depth, bump size, etc. described later), and threshold values. Here, for bump evaluation, defect determination for evaluating the presence or absence of defects (Particle, Residue, Chipping, Dent, Projection, Cracking, etc., which will be described later) of bumps (or paste material to be bumps), dimension measurement of defects, defects Classification, measurement of the height of the paste material to be bumped (SOD Height described later), measurement of the dimension of the hole (hole) to be filled with the paste material (Geometry described later), angle measurement of the wall of the hole, volume estimation of the paste material Is included. Furthermore, these bump evaluations are all performed in association with the coordinates. Here, the design standard information is not a predetermined value, and may be arbitrarily changed by an operator.

  The evaluation result obtained by the bump evaluation is sent to the result file generation unit 340. The result file generation unit 340 organizes useful information such as the defect position, size, classification result, 3D image, and statistical data by the operator, and transmits the arranged information to the host CPU 126.

  Next, the bump evaluation will be described in more detail. The bump measurement system of this embodiment can change items to be evaluated for each bump forming process to be evaluated. The bump formation process is roughly divided into four steps. The first step is a step of forming a copper base to be the base of the bump. In the second step, a wall is formed around the copper base formed in the first step, and a hole (hole) for filling a paste material to be a bump protrusion is formed. The third step is a step of filling the hole with a paste material. The fourth step is a step of removing the wall formed in the second step. In the bump measurement system of the present embodiment, for example, bumps can be appropriately evaluated for each of these four steps. Hereinafter, although details will be described, the bump evaluation is performed by comparing various information described later with design standard information (for example, normal hole outline, bump arrangement, hole depth, bump size, etc.).

  First, the bump evaluation in the first step will be described. 4 and 5 are diagrams for explaining the evaluation items in the first step and what information the evaluation items are obtained from.

  Particle 400 (foreign matter) is obtained from a dark image in the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second image obtained by the second one-dimensional line sensor 107 (SCD). To be judged. This dark image is caused by scattering of illumination light.

  Residue 410 (residue) is determined from the contrast of the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second image obtained by the second one-dimensional line sensor 107 (SCD). Is done. The contrast includes, for example, a contrast due to scattering when illumination light is illuminated on the Residue 410, a contrast due to a difference in reflectance between the Residue 410 and other regions, and a contrast due to interference when the Residue 410 is a transmissive thin film. If the Residue 410 is thick enough, the end of the Residue 410 should be a dark image. This dark image information is also used.

  Surface Roughness 420 (surface roughness) is the contrast between the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second image obtained by the second one-dimensional line sensor 107 (SCD). Obtained from.

  Chipping 430 (cut end), Dent (dent), and Projection (projection) are obtained by the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second one-dimensional line sensor 107 (SCD). Contrast in the second image and its position, and a third image obtained from the first oblique bright field detection system 1100 and a fourth image obtained from the second oblique bright field detection system 1200. It is judged from the image.

  Chipping 430 (cut edge) in the normal direction of the substrate is the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second image obtained by the second one-dimensional line sensor 107 (SCD). As shown in the image 500 in FIG. 5, the end portion is displayed darkly.

  Further, the chipping 430 (cut end) on the substrate diagonal is brightly displayed on the obtained image 510 if the chipping 430 faces the first oblique bright field detection system 1100 (second diagonal). The same applies to the bright field detection system 1200). Further, when the Chipping 430 is not opposed to the first oblique bright field detection system 1100 (if it is in the direction of approximately 90 ° with respect to the projection on the substrate of the first oblique bright field detection system), it was obtained. The Chipping 430 in the image 520 is displayed darkly (the same applies to the second oblique bright field detection system 1200). Further, even when the Chipping 430 is in the opposite direction to the first oblique bright field detection system 1100, contour extraction (a copper ellipse image 60 actually obtained and a reference ellipse 61 which is a kind of design reference information) are obtained. The chipping 430 can be determined. The determination using the contour extraction is performed when the above-described Chipping 430 faces the first oblique bright field detection system 1100 or when the Chipping 430 does not face the first oblique bright field detection system 1100. Can also be applied.

  As shown in the images 540 to 560 in FIG. 5, the Dent 440 (dent) has a thick outline, and the entire Dent 440 (dent) is displayed darkly. As shown in the images 570 to 590 in FIG. 5, the projection 450 (projection) has a thin outline and a bright central portion. Therefore, by using this difference, it is possible to discriminate between Dent 440 (dent) and Projection 450 (projection).

  Cracking 460 (crack) is the illumination light in the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second image obtained by the second one-dimensional line sensor 107 (SCD). Judged from the dark image due to scattering.

  Cu Die Height 470 (the height of the copper platform) is the third image obtained from the first oblique bright field detection system 1100 and the second oblique illumination, as shown in FIG. 4 image 471 and image 472. This is determined from the shape of the fourth image obtained from the visual field detection system 1200. Further, Geometry (φx, φy) 480 (the shape of a copper base), Alignment 481 (angle from a certain direction), and pitch 482 are obtained by the first one-dimensional line sensor 106 (SCU) as shown in an image 483 in FIG. The first image obtained can be obtained from at least one of the second images obtained by the second one-dimensional line sensor 107 (SCD).

  Next, bump evaluation in the second step will be described. 6 and 7 are diagrams for explaining the evaluation items in the second step and what information the evaluation items can be obtained from.

  Residue 601 (residue) resulting from the formation of the hole is a first image obtained by the first one-dimensional line sensor 106 (SCU), and a second image obtained by the second one-dimensional line sensor 107 (SCD). It is judged from the dark image by scattering of the illumination light in the image.

  Side Wall Crack 602 and Blunt Prominence 603 (smooth swell with spread) are obtained from the third image obtained from the first oblique bright field detection system 1100 and from the second oblique bright field detection system 1200. Obtained fourth image can be obtained. This will be specifically described. Side Wall Crack 602 and Blunt Prominence 603 are obtained from the third image obtained from the first oblique bright field detection system 1100 and the fourth image obtained from the second oblique bright field detection system 1200. This is determined by the difference between the image of the hole and the outline of the normal hole which is one type of design reference data, the height of the hole wall, and the contrast distribution of the abnormal part.

  First, the case where Side Wall Crack 602 is determined will be described. An image 701 in FIG. 7 is an image obtained from the first oblique bright field detection system 1100, and an image 702 is an image obtained from the second oblique bright field detection system 1200. Here, it is determined that Side Wall Crack 602 in the following case. (1) The outline of the obtained hole deviates from the outline of the normal hole. (2) In the image 701, there is a dark part outside the outline of the normal hole. (3) The size of the dark part is larger than a certain reference. (4) In the image 702, there is a bright part outside the outline of the normal hole. (5) The height of the hole wall is lower than a certain standard.

  On the other hand, in the following cases, it is determined that it is Blunt Prominence 603. (1) The outline of the obtained hole deviates from the outline of the normal hole. (2) In the image 703, there is a bright portion outside the outline of the normal hole. (3) The size of the dark part is smaller than a certain standard. (4) In the image 704, there is a bright part outside the outline of the normal hole. (5) The height of the hole wall is higher than a certain standard.

  Geometry (φx) 604, Geometry (φy) 605 (shape of the wall of the formed hole), Alignment 606 (angle from a certain direction), Offset (Δx) 607 (deviation in the x direction), Offset (Δy) 608 ( Pitch 4609 is a first image obtained by the first one-dimensional line sensor 106 (SCU) and a second one-dimensional line sensor 107 (SCD) as shown in an image 610 in FIG. It can be obtained from at least one (preferably two) of the obtained second images.

  Next, the bump evaluation in the third step will be described. FIG. 8 and FIG. 9 are diagrams for explaining the evaluation items in the third step and what information the evaluation items can be obtained from.

  First, Paste Residue 801 (residue resulting from application of paste material) is obtained by the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second one-dimensional line sensor 107 (SCD). It is determined from a dark image due to scattering of illumination light in the obtained second image.

  Miss Filling 802 (a state in which a predetermined paste filling amount is not satisfied) and Over Filling 803 (overfilling of paste material) are a first image and a second one obtained by the first one-dimensional line sensor 106 (SCU). A second image obtained by the dimensional line sensor 107 (SCD), a third image obtained from the first oblique bright field detection system 1100, and obtained from the second oblique bright field detection system 1200. Judgment is made from the fourth image.

  This will be described more specifically. A region 901 in FIG. 9 is a diagram for explaining discrimination between Miss Filling 802 and Over Filling 803 using the detection result of the upper bright field detection system 1000. First, the Miss Filling 802 portion is displayed dark (reference numerals 902 and 903), and the Over Filling 803 portion is displayed brightly (reference numerals 904 and 905). Furthermore, a sharp outline is different between an image obtained at Focus Position Upper (focus position is upward) and an image obtained at Focus Position Lower (focal position is downward). This is because, for example, in an image obtained with Focus Position Upper (focus position is upward), the outline of the image near the top of Over Filling 803 is clearly obtained, and an image obtained with Focus Position Lower (focal position is downward). The figure shows that the outline of the image near the bottom surface of Over Filling 803 can be clearly obtained. For Miss Filling 802, for example, an image obtained with Focus Position Lower (focal position is lower) indicates that the contour of the image can be obtained more clearly. Thus, by focusing on at least one (preferably both) of contrast and contour, it is possible to distinguish between Miss Filling 802 and Over Filling 803.

  The first oblique bright field detection system 1100 and the second oblique bright field detection system 1200 can also distinguish between Miss Filling 802 and Over Filling 803. A region 906 in FIG. 9 is a diagram for explaining the discrimination between Miss Filling 802 and Over Filling 803 using the first oblique bright field detection system 1100 and the second oblique bright field detection system 1200. Here, it is determined as Over Filling 803 in the following cases. (1) In the image 907, there is a bright part outside the outline of the normal hole. (2) In the image 908, there is a bright part inside the outline of the normal hole. On the other hand, in the following case, it is determined to be Miss Filling 802. (1) In the image 909, there is a dark part inside the outline of the normal hole. (2) In the image 910, there is a dark part inside the outline of the normal hole. The above-described discrimination between Miss Filling 802 and Over Filling 803 is performed using the detection result of the upper bright field detection system 1000, the first oblique bright field detection system 1100, and the second oblique bright field detection system 1200. Since the accuracy of discrimination is higher when the discrimination using the detection result is combined, it is better to combine them.

  Next, bump evaluation in the fourth step will be described. 10 and 11 are diagrams for explaining the evaluation items in the fourth step and what kind of information the evaluation items can be obtained from.

Particle 1001 and DFR Residue 1002 (residue resulting from wall removal) are obtained by the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second one-dimensional line sensor 107 (SCD). Further, it is determined from a dark image due to scattering of illumination light in the second image.
Classification is possible by using feature quantities.

  Dent 1003 (dent) and Projection 1004 (protrusion) are a first image obtained by the first one-dimensional line sensor 106 (SCU) and a second image obtained by the second one-dimensional line sensor 107 (SCD). In the meantime, it is determined from the third image obtained from the first oblique bright field detection system 1100 and the fourth image obtained from the second oblique bright field detection system 1200.

  More specific description will be given. A region 1005 in FIG. 11 is a diagram for explaining discrimination between Dent 1003 (dent) and Projection 1004 (projection) using the detection result of the upper bright field detection system 1000. First, the portion of Dent 1003 (dent) is displayed dark (1006, 1007), and the portion of Projection 1004 is displayed brightly (reference numerals 1008, 1009). Further, an image obtained at Focus Position Upper (focus position is upward) and an image obtained at Focus Position Lower (focus position is downward) have different outlines. Furthermore, in Projection 1004, the size of the bright portion that appears in the image differs depending on the location of the focal position. More specifically, the bright part of the image obtained at Focus Position Lower (focal position is lower) is larger than the bright part of the image obtained at Focus Position Upper (focal position is upward). In this way, it is possible to distinguish between Dent 1003 (dent) and Projection 1004 (projection) by paying attention to the detection result of the upper bright field detection system 1000, particularly the contrast and contour.

  The first oblique bright field detection system 1100 and the second oblique bright field detection system 1200 can also discriminate between Dent 1003 (dent) and Projection 1004 (projection). An area 1010 in FIG. 11 is used to discriminate between Dent 1003 (dent) and Projection 1004 (projection) by using at least one of the first oblique bright field detection system 1100 and the second oblique bright field detection system 1200. It is a figure explaining. The images 1011 and 1012 are at least one of images obtained by the first oblique bright field detection system 1100 and the second oblique bright field detection system 1200. If it is Dent 1003 (dent), a unique contrast 1120 is displayed inside the reference outline 1110. In the case of Projection 1004 (projection), a unique contrast 1130 is displayed outside the reference outline 1110. Note that the above-described discrimination between Dent 1003 (dent) and Projection 1004 (projection) is performed using the detection result of the upper bright field detection system 1000, the first oblique bright field detection system 1100, and the second oblique bright field. Since the accuracy of discrimination is higher when the discrimination using the detection result of the detection system 1200 is combined, it is better to combine the discrimination.

  SOD Height 1013 (the height from the copper platform to the paste apex) and Cu Die Height 1014 (the height of the copper platform) were obtained from the first oblique bright field detection system 1100 as shown in region 1015 of FIG. It can be obtained from at least one of the third image and the fourth image obtained from the second oblique bright field detection system 1200. Note that the SOD Height 1013 and the Cu Die Height 1014 in the area 1015 of FIG. 10 reflect the first and second oblique angles. Therefore, in the present embodiment, the true SOD Height 1013 and Cu Die Height 1014 are obtained in consideration of the first and second oblique angles in the SOD Height 1013 and Cu Die Height 1014 in the region 1015 of FIG. Furthermore, as shown in FIG. 10 region 1021, the quality of the formed bump protrusion 1022 can be determined by comparing with the normal bump outline 1023 to be used as a reference.

  Geometry (φx) 1016, Geometry (φy) 1017 (shape of the wall of the formed hole), Offset (Δx) 1019 (deviation in the x direction), Offset (Δy) 1020 (deviation in the y direction), FIG. As shown in an image 1030, from at least one of the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second image obtained by the second one-dimensional line sensor 107 (SCD). Can be obtained.

  Next, side wall measurement and paste material filling failure measurement in the second step and the third step will be described. FIG. 12 is a view for explaining side wall measurement and paste material filling failure measurement. As described above, when imaging using the two-dimensional sensors 118 and 123, it is possible to selectively use a set of some arbitrary pixels instead of using all the pixels of the sensor. This arbitrary set of pixels can be expressed as an effective area. For the side wall measurement and paste material filling failure measurement, the effective areas 1207 and 1208 and their depth of focus are the height 1203 from the top of the copper platform 1201 to the top of the wall 1202 read from the design reference data, It is determined in consideration of the aspect ratio between the hole width 1204 and height 1203 and the hole width 1204. The same applies to the first oblique angle and the second oblique angle.

  The inclination of the wall is determined by the first image obtained by the first one-dimensional line sensor 106 (SCU) and the hole obtained by the second image obtained by the second one-dimensional line sensor 107 (SCD). It can be obtained by the difference between the shape (may be the ellipticity of the hole) and the third image obtained from the effective area of the first oblique bright field detection system 1100. Further, the shape of the hole (hole) obtained by the first image obtained by the first one-dimensional line sensor 106 (SCU) and the second image obtained by the second one-dimensional line sensor 107 (SCD). And the fourth image obtained from the effective area of the second oblique bright field detection system 1200 can also be obtained.

  Further, the poor filling of the paste is caused by the third image obtained from the effective area of the first oblique bright field detection system 1100 and the second oblique bright field, such as areas 1205 and 1206 in FIG. This is reflected in the fourth image obtained from the effective area of the detection system 1200. That is, the filling of the paste from the third image obtained from the effective area of the first oblique bright field detection system 1100 and the fourth image obtained from the effective area of the second oblique bright field detection system 1200. Defects can be estimated.

  Next, defect detection at the upper and lower parts of the hole will be described. In the present embodiment, the first one-dimensional line sensor 106 (SCU) and the second one-dimensional line sensor 107 (SCD) are connected to different focal positions so that defects at the upper and lower portions of the holes having different focal positions can be detected. To place. For example, in the present embodiment, the first one-dimensional line sensor 106 (SCU) is used for detecting defects at the upper part of the hole, and the second one-dimensional line sensor 107 (SCD) is used for detecting defects at the lower part of the hole. Furthermore, in this embodiment, in order to make the illumination light reach the bottom of the hall, for example, a bright field epi-illumination / dark field illumination method using a lamp (bulb) can be adopted (in this case, the illumination light quantity of bright field and dark field) The ratio can be changed from 100: 0 to 0: 100).

  Next, image reading will be described. FIG. 13 is a diagram for explaining reading of an image in this embodiment. The trigger signals 1301 and 1302 for controlling the exposure of the two-dimensional sensor 118 and the two-dimensional sensor 123 are for reading the signals of the first one-dimensional line sensor 106 (SCU) and the second one-dimensional line sensor 107 (SCD). The trigger signals 1303 and 1304 are synchronized with each other. Here, the trigger signal 1301 is a trigger signal corresponding to a t1st bump separated from the alignment mark in the 1st die 1309 (or shot) by Δxforg. The trigger signal 1302 is a trigger signal corresponding to a 2nd bump that is separated from the 1st bump by a distance Px. The opening times of the optical shutters 117 and 122 are synchronized with the trigger signal 1303. Further, the closing time of the optical shutters 117 and 122 can be set to an arbitrary time so as to cope with a measurement object having a low reflectance (region 1305). As a result, the time during which the optical shutters 117 and 122 are open can be controlled. This means that the charge accumulation times of the two-dimensional sensor 118 and the two-dimensional sensor 123 are variable. Furthermore, as described above, the two-dimensional sensor 118 and the two-dimensional sensor 123 may be of a time delay integration type. In that case, the two-dimensional sensor 118 and the two-dimensional sensor 123 shift the charge in synchronization with the scanning of the stage 1300. The shift timing in this case is changed according to the first oblique angle and the second oblique angle with reference to the one-dimensional line sensor 106 (SCU) and the second one-dimensional line sensor 107 (SCD). The number of additions is determined based on the two-dimensional sensor 118 and the two-dimensional sensor 123 (region 1306 in FIG. 13). It is also possible to invalidate the read data at the time of addition scanning (FIG. 13 area 1307). After the exposure of the two-dimensional sensor 118 and the two-dimensional sensor 123 is completed, the shift trigger is changed and data is read (area 1308).

  Next, the positional relationship of the visual fields of the one-dimensional line sensor 106, the one-dimensional line sensor 107, the two-dimensional sensor 118, and the two-dimensional sensor 123 will be described. FIG. 14 is a diagram for explaining the positional relationship of the visual fields of the one-dimensional line sensor 106, the second one-dimensional line sensor 107, the two-dimensional sensor 118, and the two-dimensional sensor 123. In this embodiment, the visual field 1401 of the one-dimensional line sensor 106, the visual field 1402 of the second one-dimensional line sensor 107, the visual field 1403 of the two-dimensional sensor 118 (which can also be expressed as the above-described effective region), and the two-dimensional sensor 123 fields of view 1404 are in different positions. Note that the X axis and the Y axis in FIG. 14 are the scanning directions of the stage 1300, respectively. That is, the stage 1300 moves the substrate in the X-axis and Y-axis directions in FIG. In this embodiment, the stage 1300 moves the substrate in the direction of the arrow 1405. That is, the bump 1406 is temporally imaged in the order of the two-dimensional sensor 118, the one-dimensional line sensor 106, the one-dimensional line sensor 107, and the two-dimensional sensor 123. Since the one-dimensional line sensor 106 and the one-dimensional line sensor 107 are on the same coordinate on the X axis, the bumps are imaged simultaneously in time. In other words, in this embodiment, a plurality of images different in time are obtained for the bump 1406. That is, images for the same bump are stored at different addresses in the memory. Here, if the distance between each field of view is known, it is possible to know how much time the plurality of images obtained are captured by each sensor from the relationship between the distance between each field of view and the scanning speed. it can. Therefore, in this embodiment, a plurality of images that are imaged with respect to the same bump and have a time difference are read from the memory in consideration of this time difference.

  This will be described more specifically. For example, at least one of the one-dimensional line sensor 106 and the one-dimensional line sensor 107 captures an image delayed by a time 1407 with respect to the time captured by the two-dimensional sensor 118. The two-dimensional sensor 123 captures an image with a delay of time 1408 with respect to the time when the two-dimensional sensor 118 captures an image. Therefore, in this embodiment, the captured image is read from the memory in consideration of the time differences 1407 and 1408.

  Further, considering the Y-axis direction, each visual field may have a positional shift 1409, 1410, or 1411 in terms of distance, but the size of each visual field is larger than the expected positional shift, and is desirable. Is sufficiently large, the influence of the positional deviations 1409, 1410, and 1411 can be substantially ignored.

  Next, Example 2 will be described. In the second embodiment, portions different from the first embodiment will be mainly described. In the second embodiment, for example, parallel illumination light from a multipoint light source using LEDs is used.

  The present embodiment will be described with reference to FIG. In the present embodiment, the first multi-point light source 1506 and the second multi-point light source 1506 are arranged on the outer peripheral side around the objective lens 120 of the second oblique bright field detection system 1200. A multi-point light source 1507 is included. Here, the first multipoint light source 1506 includes a plurality of LEDs 1506. The second multipoint light source 1507 has a plurality of LEDs 1505. When the bump 1501 or the copper base 1502 is imaged, the first multi-point light source 1506 and the second multi-point light source 1507 emit light. In this case, the end of the dotted line 1507 is more emphasized and imaged. This is effective for obtaining the above-described bump shape.

  Note that the above-described multi-point light source includes the upper bright field detection system 1000 arranged in the normal direction of the substrate and the first oblique bright field detection system 1100 arranged at a first oblique angle with respect to the substrate. It may be adopted. Alternatively, at least one of the first multi-point light source 1506 and the second multi-point light source 1507 may be selectively made to emit light. Further, as indicated by a dotted line 1503, a part of the multipoint light source may be caused to emit light. By doing in this way, it becomes possible to emphasize and image the edge of an arbitrary place.

  According to the present embodiment, since the amount of information obtained is greater than in the prior art, more accurate bump evaluation can be performed. In addition, this invention is not limited to an Example. For example, any two of the three detection optical systems described above can be employed, or the number of detection optical systems may be further increased. Further, the present invention may be used for evaluations other than bumps.

1000 Upper bright field detection system 1100 First oblique bright field detection system 1200 Second oblique bright field detection system 1300 Stage

Claims (11)

In an apparatus for measuring the shape formed on a substrate,
An upper bright-field detection optical system disposed in the normal direction of the substrate;
A first oblique bright field detection optical system disposed at a first oblique angle with respect to the substrate;
An apparatus for measuring a shape formed on a substrate, comprising: a second oblique bright field detection optical system disposed at a second oblique angle with respect to the substrate. In the apparatus for measuring the shape formed on the substrate according to claim 1,
At least one of the first oblique bright field detection optical system and the second oblique bright field detection optical system starts imaging based on the imaging of the upper bright field detection optical system. A device for measuring the shape formed on the substrate. In the apparatus for measuring the shape formed on the substrate according to claim 2,
At least one of the first oblique bright field detection optical system and the second oblique bright field detection optical system has a shutter, and the shutter is triggered by imaging of the upper bright field detection optical system. An apparatus for measuring a shape formed on a substrate, characterized by opening and closing. In the apparatus for measuring the shape formed on the substrate according to claim 3,
The shutter is made of a transparent ferroelectric ceramic material, and is an apparatus for measuring a shape formed on a substrate. In the apparatus for measuring the shape formed on the substrate according to claim 4,
At least one of the first oblique bright field detection optical system and the second oblique bright field detection optical system includes a charge storage type sensor,
An apparatus for measuring a shape formed on a substrate, wherein a charge accumulation time of the sensor corresponds to an opening / closing time of the shutter. In the apparatus for measuring the shape formed on the substrate according to claim 1,
The apparatus for measuring a shape formed on a substrate, wherein the first oblique bright field detection optical system changes the first oblique angle according to design standard information. In the apparatus for measuring the shape formed on the substrate according to claim 1,
The apparatus for measuring a shape formed on a substrate, wherein the second oblique bright field detection optical system changes the second oblique angle according to design standard information. In the apparatus for measuring the shape formed on the substrate according to claim 1,
The upper bright field detection optical system includes two line sensors having different focal positions, and an apparatus for measuring a shape formed on a substrate. In the apparatus for measuring the shape formed on the substrate according to claim 1,
The first oblique bright field detection optical system includes a two-dimensional sensor having a plurality of pixels, and selects a pixel to be used for imaging from the plurality of pixels according to design standard information. A device for measuring the shape formed on the substrate. In the apparatus for measuring the shape formed on the substrate according to claim 1,
The second oblique bright field detection optical system includes a two-dimensional sensor having a plurality of pixels, and selects a pixel to be used for imaging from the plurality of pixels according to design standard information. A device for measuring the shape formed on the substrate. In the apparatus for measuring the shape formed on the substrate according to claim 1,
The upper bright field detection optical system, the first oblique bright field detection optical system, and the second oblique bright field detection optical system are arranged along a scanning direction of the substrate. A device for measuring the shape formed on the substrate.