JP6445943B2 - Measuring method of semiconductor device - Google Patents

Description

  The present embodiment relates to a method for measuring a semiconductor device.

  In a semiconductor device, a bonding wire may be used.

JP 2007-129121 A JP 2006-292647 A

  Automatic and accurate measurement of wire height.

The measurement method of the semiconductor device according to the present embodiment corrects the tilt of the substrate, corrects the tilt of the semiconductor chip disposed on the substrate, and is disposed on the electrode on the substrate and the substrate. A local coordinate system along the path of the bonding wire disposed between the electrode pads on the semiconductor chip and the bonding wire is set for the plurality of bonding wires, and the heights of the bonding wires correspond to the bonding wires. multiple measurements along the front kilometers Karu coordinate system.

The typical block diagram which shows the structure of the measuring apparatus which concerns on embodiment. The flowchart which showed the flow of the whole measurement which concerns on embodiment. The figure which showed an example of the coordinate data file of wire bonding. The figure which showed typically the structure of the spreadsheet for measurement program creation which produces a measurement program. The flowchart which shows the flow of the measuring method which concerns on embodiment. The top view explaining the measuring method concerning an embodiment. The top view explaining the measuring method concerning an embodiment. The top view explaining the measuring method concerning an embodiment. The top view explaining the measuring method concerning an embodiment. The top view explaining the measuring method concerning an embodiment. The top view explaining the measuring method concerning an embodiment. The top view explaining the measuring method concerning an embodiment. The top view explaining the measuring method concerning an embodiment. Sectional drawing explaining the measuring method which concerns on embodiment. The top view explaining the measuring method concerning an embodiment. (A) The top view explaining the measuring method which concerns on embodiment, (B) Typical sectional drawing explaining the measuring method which concerns on embodiment.

  Hereinafter, embodiments will be described with reference to the drawings. In the following description, about the same function and component, the same code | symbol is attached | subjected. In addition, the dimension and magnitude | size in each figure are reference. For ease of viewing, each figure is appropriately changed and adjusted in dimensions and length.

(Description of Embodiment)
FIG. 1 is a schematic block diagram illustrating a measuring apparatus 10 according to the present embodiment.

  The measuring apparatus 10 includes a head unit 20, a stage 50, a controller 70, and a display unit 80.

  The head unit 20 includes a first drive unit 30 and a camera 40. The first drive unit 30 can move the head unit 20 in the Z-axis direction (vertical direction). That is, the first drive unit 30 can move the head unit 20 in the direction of the stage 50, and can make the distance between the head unit 20 and the stage 50 closer or further away.

  The camera 40 can photograph the semiconductor chip 100 and the substrate 110 on the stage 50, for example. The camera 40 has a lens (not shown), for example. The camera 40 shoots a focused image by adjusting the height of the lens based on the input of the controller 70, for example. The camera 40 outputs the captured image data to the controller 70.

  The stage 50 has a second drive unit 60. The stage 50 can hold the substrate 110 and the semiconductor chip 100 thereon. The second drive unit 60 moves the stage 50 in the XY directions. The second drive unit 60 may move the stage 50 in the Z-axis direction (vertical direction).

  The controller 70 controls each part.

  For example, the controller 70 can output a signal to the first drive unit 30 or the second drive unit 60 to move the head unit 20 or the stage 50. For example, the controller 70 outputs a signal to the camera 40 to adjust the focus of the camera 40. The camera 40 outputs the lens position at that time to the controller 70. The controller 70 calculates the distance to the subject from the information on the lens position. The controller 70 can measure the height between the two points from the difference in distance to each of the two points. The controller 70 may directly calculate the height between the two points from the lens position where the two points are in focus.

  In addition, the controller 70 may adjust the focus by moving the head unit 20 or the stage 50. In this case, the controller 70 can calculate the height from the moving distance of the head unit 20 or the stage 50.

  For example, the controller 70 receives image data from the camera 40 and outputs the image data to the display unit 80. The controller 70 can perform image processing on the image data received from the camera 40. The controller 70 can output the image data subjected to the image processing to the display unit 80. For example, the controller 70 performs processing on the image data in the range specified by the caliper (measurement mark on the screen of the display unit 80) in the acquired image data of the entire screen, thereby performing predetermined processing in the captured image data. The length and position can be measured.

  The display unit 80 has a screen (not shown). The display unit 80 displays the image data input from the controller 70. The display unit 80 is, for example, a display or a monitor.

  The semiconductor chip 100 is disposed on the substrate 110. The semiconductor chip 100 is connected to the substrate 110 by a bonding wire described later. The bonding wire is, for example, a gold wire. The substrate 110 is an arbitrary substrate such as a glass epoxy substrate or a lead frame.

  In the following description, so-called “positive bonding” in which bonding is performed from the semiconductor chip 100 toward the substrate 110 will be described as an example. So-called “reverse bonding” in which bonding from the substrate 110 toward the semiconductor chip 100 may be used.

  In the following description, an example in which two semiconductor chips 100 are mounted on the substrate 110 will be described. A multi-stage chip in which the semiconductor chips 100 are stacked stepwise may be used, and the number of the semiconductor chips 100 is arbitrary.

  FIG. 2 is a flowchart showing the overall flow of measurement.

  The measurer prepares a coordinate data file at the time of bonding between the semiconductor chip 100 and the substrate 110 (step S100). This coordinate data file is originally a file for instructing coordinates when bonding to the wire bonder, and is used here. Since it is not a newly created file for bonding wire measurement, there is little burden on the measurer.

  FIG. 3 is an example of a coordinate data file.

  The first part 200 in FIG. 3 describes setting items such as a product name, a package name, the number of alignment points, and the number of bonding wires. In the second portion 210 of FIG. 3, coordinate information for aligning the substrate 110 is described. For example, an example in which there are two measurement points and the alignment coordinates are (X, Y) = (8000, 8500), (−8000, -8500) is shown. The third part 220 and the fourth part 230 in FIG. 3 describe coordinate information when the semiconductor chip 100 is aligned.

  In the fifth portion 240 of FIG. 3, bonding coordinate information is described. That is, for example, the first bonding wire has a ball bonding coordinate (−3300, 3200) on the semiconductor chip 100 side and a stitch bonding coordinate (−4500, 4500) on the substrate 110 side.

  In this case, the unit of the coordinate value is a micrometer. Since the coordinate unit instructed to the measuring apparatus 10 may be in millimeters, the unit is converted from micrometer to millimeter if necessary in the “bonding coordinate readout” of step S110 in FIG.

  Returning to FIG. 2, the measurement flow will be described.

Data of wire bonding coordinates corresponding to the fifth portion 240 of FIG. 3 is read from the coordinate data file (step S110). The measurer prepares design data such as information on the alignment mark coordinates of the substrate 110, the size of the semiconductor chip, and the offset amount of the semiconductor chip (step S120).
Then, based on the wire bonding coordinate data extracted in step S110 and the design data such as the alignment mark coordinate information of the substrate 110 prepared in step S120, the size of the semiconductor chip 100, and the offset amount of the semiconductor chip 100. A measurement program is created (step S130).

  The measurement program is usually timed by the measurer to execute the actual measurement procedure one step at a time, and takes time.

  In order to reduce the measurement program creation time, for example, the following is performed.

  The measurer prepares a spreadsheet for creating a measurement program (with a macro). The data of wire bonding coordinates extracted from the coordinate data file is input to a list of bonding coordinate values provided in advance in the spreadsheet for measuring program creation (with macro). From the design data, design data such as information on the alignment mark coordinates of the substrate 110, the size of the semiconductor chip 100, and the offset amount of the semiconductor chip 100 is input to a measurement program creation spreadsheet (with macro). Thereafter, the macro in the spreadsheet sheet is executed to automatically create a measurement program.

  FIG. 4 is a diagram showing the configuration of a spreadsheet for creating a measurement program (with a macro).

  The procedure for measuring the bonding wire with the measuring apparatus 10 is constant regardless of the type of semiconductor device, and the moving position of the head and stage, the display position of the caliper, etc. are determined according to the bonding wire position that changes depending on the type. Instructed coordinate values, measurement parameters, etc. change. By rewriting the coordinate values, measurement parameters, and the like written in each program line in the measurement program, a measurement program for a different type of semiconductor device can be obtained. FIG. 4 shows a procedure for rewriting the coordinate values, measurement parameters, and the like necessary for each program line in the spreadsheet sheet according to the type of semiconductor device.

  A first area 300 in FIG. 4 is a list of bonding coordinate values. For example, the user inputs a bonding coordinate value in a predetermined area. In the second area 310, one line of the measurement program is shown as an example. One line of the measurement program is divided into separate cells for each element, and includes cells in which measurement commands and the like are written in advance and cells referring to the first area 300. Here, the cell referring to the first region 300 of the second region 310 is based on design data such as alignment mark coordinate information (not shown), the size of the semiconductor chip 100, and the offset amount of the semiconductor chip 100. The value of the area 300 may be converted and referenced.

  For example, the bonding coordinate value instructed to the wire bonder usually has the origin at the center of the semiconductor device. That is, the bond coordinate value extracted from the wire bonder coordinate data file and written in the first region 300 in FIG. 4 also has the center of the semiconductor device as the origin.

  On the other hand, the bonding coordinate value instructed to the measuring apparatus 10 has the origin of the alignment mark (B), the corner of the semiconductor chip 100, etc. for the convenience of creating the measurement program. That is, coordinate conversion is required between the bonding coordinate value written in the first area 300 and the bonding coordinate value written in the second area 310. If a calculation formula for coordinate conversion is inserted between the first region 300 and the second region 310, the bonding coordinate value with the origin of the alignment mark (B), the corner of the semiconductor chip 100, etc., necessary for the measurement program is obtained. Can be converted.

  Not only the bonding coordinate values but also the measurement parameters entered by the measurer can be actually converted into the required form in the measurement program line.

  The third area 320 in FIG. 4 is obtained by combining the cell values of the second area 310 as a character string. In the third area 320 of FIG. 4, an example of the measurement command for one line is shown, but actually, a plurality of program lines necessary for measuring one bonding wire × the number of bonding wires to be measured. The data of the third area 320 is generated.

  The data in the third area 320 is collectively written out as a text file, thereby creating a measurement program. This is shown in the fourth region 330.

  Returning to FIG. 2 again, the measurement flow will be described.

  By the measurement program created up to step S130, the measuring apparatus 10 measures various data of the bonding wire (step S140). The measuring device 10 outputs measurement data (step S150).

  Details of step S140 in FIG. 2 will be described with reference to FIG.

  FIG. 5 is a flowchart illustrating a specific measurement procedure of the measurement apparatus 10. This measurement procedure is automatically processed based on the above-described measurement program. Alternatively, the measurement procedure is automatically processed except for a response when an error occurs.

  In step S200, the alignment mark on the substrate 110 is measured and the tilt is corrected. Specific processing will be described with reference to FIG.

  As shown in FIG. 6A, semiconductor chips 100a and 100b are arranged on a substrate 110. The semiconductor chip 100a is, for example, a memory controller, and the semiconductor chip 100b is, for example, a NAND memory. The semiconductor chip 100a is electrically connected to the substrate 110 via the bonding wires 150.

  On the substrate 110, alignment marks 400a and 400b are arranged. An example of the alignment mark 400 is shown in FIGS. 6B and 6C. FIG. 6B shows an example of a donut shape, and FIG. 6C shows a circular example. The alignment mark 400 may be captured by the camera from above and the boundary can be seen as shown in FIGS. 6B and 6C. For example, the alignment mark 400 may be a conductive layer formed on the substrate 110 or a hole penetrating the substrate 110. Further, the alignment mark 400 may be aligned using an arbitrary pattern on the substrate 110 instead of a dedicated mark.

  Note that the alignment mark 400 may have a cross shape, an L shape, or the like instead of the donuts and circles shown in FIGS. 6B and 6C. However, in order to measure the center of an alignment mark such as a cross or an L character, a large number of program lines are required. If it is a donut shape or a circle, it can be measured by a program of several lines using a circular caliper 410 described later. Unless there is a pattern similar to the alignment mark in the vicinity and there is a risk of measurement errors, the alignment mark is preferably donut-shaped or circular.

  Here, the circular caliper 410 is a measurement symmetry with respect to a program module that detects an edge of a circular object to be photographed existing in an image and estimates the center coordinates of an approximate circle of the edge using a least square method or the like. It is a circular cursor image for indicating the position on the image.

  First, the positions of the alignment marks 400a and 400b are measured using a circular caliper 410 shown in FIG. The controller 70 can measure the diameter and center position of the alignment mark, for example, by performing image processing on the circular caliper 410 shown in FIG. 6D and the image data input from the camera 40.

  Specifically, as shown in FIG. 6E, the circular caliper 410 is disposed so as to overlap the alignment mark 400. The controller 70 measures the brightness of the image in the circular caliper 410 radially from the center of the circular caliper 410. The edge of the alignment mark 400 is detected by measuring the brightness of the image. The controller 70 calculates the diameter and center position of the alignment mark 400 from the detected positions of the plurality of edges. If the alignment mark 400 is not a perfect circle, the controller 70 calculates the diameter and center position of the proximity circle of the alignment mark 400 by calculation.

  The controller 70 sets the local coordinate system 500 for the substrate 110 with the straight line connecting the calculated center positions of the alignment marks 400a and 400b as the X axis.

  FIG. 6 shows an example in which the local coordinate system 500 is arranged on the alignment mark 400a, for example. In FIG. 6, since the alignment marks 400 a and 400 b are arranged in parallel to the horizontal direction of the substrate 110, the X axis of the local coordinate system 500 set here is also parallel to the horizontal direction of the substrate 110. Since the measurement program after this designates an operation position with respect to the local coordinate system 500, even if the board | substrate 110 is inclined and set with respect to the stage of the measuring apparatus 10, it is operated by the assumed position.

  FIG. 7 is a diagram for explaining an example of the arrangement of another alignment mark 400. Alignment marks 400a and 400c are arranged obliquely with respect to substrate 110.

  The X axis of the local coordinate system 500 is set in a direction rotated clockwise by an angle formed by a straight line connecting the alignment marks 400a and 400c with the lateral direction of the substrate 110.

  Note that a value calculated from the design values of the alignment marks 400a and 400c is used as the angle formed with the lateral direction of the substrate 110. In this way, even if the alignment marks 400a and 400c are arranged obliquely with respect to the substrate 110, the set inclination of the substrate 110 with respect to the stage can be corrected.

  Subsequently, in step S210 of FIG. 5, the mounting position and inclination of the semiconductor chip 100a with respect to the substrate 110 are corrected. Specific processing contents will be described.

  The controller 70 moves the stage and the head unit to a position where an image including the edge of the semiconductor chip can be taken. The movement destination is indicated in the measurement program line as a coordinate value for the local coordinate system 500 set in FIG. 6 or FIG.

  The coordinate value of the movement destination described in the measurement program line is calculated from information such as the chip size of the semiconductor chip 100a and the offset amount from the substrate in the measurement program creation spreadsheet (with macro). The edge of the semiconductor chip is calculated or measured from the image data photographed by the camera 40.

  For example, when the semiconductor chip 100a has a substantially rectangular shape, the controller 70 measures two locations on each side. That is, the controller 70 measures, for example, eight edge positions. By measuring two locations on each side, the edge position of each side of the semiconductor chip 100a can be calculated. Further, the controller 70 sets a local temporary coordinate system 510 for the semiconductor chip 100a.

  The edge measurement position of the semiconductor chip 100a with respect to the local coordinate system 500 is included in advance when the measurement program is created (step S130 in FIG. 2).

  More specific processing contents will be described with reference to FIG. FIG. 8 is an enlarged view of the region A1 in FIG. 6, and is a diagram for explaining a state in which edge positions with respect to two sides of the semiconductor chip 100a are detected one by one.

  In FIG. 8, a metal pad 605 is disposed on the semiconductor chip 100a. A metal ball 600 is disposed on the metal pad 605. The metal ball 600 is a part of the bonding wire 150 (not shown). Further, the wire portion of the bonding wire is not shown in the drawings of FIGS.

  The controller 70 uses the camera 40 to capture an image including the edge (end) of the semiconductor chip 100a as shown in FIG. The controller 70 calculates or measures the edge position of the semiconductor chip 100a from the image data captured by the camera 40 using the rectangular calipers 420a and 420b. Here, the rectangular caliper 420 refers to a measurement symmetrical position of a program module that can detect a linear edge of an object to be imaged existing in an image and can measure the length between two points as necessary. It is a rectangular cursor image for instructing on the image.

  Similarly to the circular caliper 410 described above, the controller 70 uses the rectangular caliper 420a to detect an edge as follows. That is, the controller 70 measures the brightness of the image in the rectangular caliper 420 from one side of the rectangular caliper 420 toward the other side. The controller 70 detects the edge of the semiconductor chip 100a by measuring the brightness of the image.

  Further, a local temporary coordinate system 510 with the corner of the semiconductor chip 100a as the origin is set for the semiconductor chip 100a from the measured edge information of the semiconductor chip 100a. The edges of the semiconductor chip 100a may vary with respect to the electrode pads and the like in the semiconductor chip 100a due to dicing line misalignment during dicing. That is, the local temporary coordinate system 510 is not necessarily set accurately for the electrode pads and the like in the semiconductor chip 100a. Therefore, in order to set the local temporary coordinate system 510 with high accuracy, the following measurement is performed.

  FIG. 9 is a diagram for explaining how the local temporary coordinate system 510 is reset to the first local coordinate system 520 with higher accuracy.

  As shown in FIG. 9, the controller 70 searches the image data stored in advance by the search caliper 430 by matching within the range designated for the local temporary coordinate system 510. Here, the search caliper 430 is a search target image (an image obtained by binarizing the search target image or the like) that is stored in advance using a least square method or the like by detecting the edge of the target to be photographed in the image. This is a cursor image for searching for a search symmetrical image for a program module that searches for a search module (including data obtained by processing a target image) or estimates an intersection of approximate lines. In FIG. 9, as an example, the search caliper 430 searches for a corner of the metal pad 605. The local temporary coordinate system 510 is reset as the first local coordinate system 520 with the position shifted by the distance from the corner of the metal pad 605 to the corner of the semiconductor chip 100a searched by the search caliper 430 as the origin.

  The reason why the corner of the semiconductor chip 100a is the origin of the first local coordinate system 520 is that the coordinate value of each metal ball 600 can be easily calculated in the spreadsheet for measuring program creation (with macro).

  That is, by setting the first local coordinate system 520 with respect to the pattern formed on the semiconductor chip 100a, the semiconductor chip 100a is not affected by variations during dicing, the inclination of the semiconductor chip 100a with respect to the substrate 110, or the like. It is possible to set the first local coordinate system 520 with high accuracy for the arranged metal pads and the like.

  However, since a large number of metal pads 605 are arranged close to the semiconductor chip 100a, if there are a plurality of metal pads 605 within the range photographed by the camera 40, what is the assumed metal pad 605? There is a risk of searching different metal pads 605 incorrectly. In that case, a wiring pattern having no similar pattern can be stored as a search target in the vicinity. In this case, since the distance from the stored wiring pattern to the corner of the semiconductor chip 100a is not accurately known, the origin of the first local coordinate system 520 does not correctly come to the corner of the semiconductor chip 100a.

  When the measurement position of the circle caliper 410 for measuring the metal ball 600 is instructed to the first local coordinate system 520 set in this way, the metal ball 600 and the circle caliper 410 are misaligned. This can be avoided by finely adjusting the position of the circular caliper 410 and then saving the measurement program again.

  If all the metal balls are designated with respect to the first local coordinate system 520, fine adjustment of the position is required for all the circular calipers. Therefore, in step S220 in FIG. A second local coordinate system 530 for the metal ball 600 is set.

  FIG. 10 is a diagram for explaining the setting of the second local coordinate system 530.

  When the measurement program is partially executed on the metal ball 600 up to the point where the circular caliper 410 is displayed, the controller 70, as shown in FIG. 440 is displayed.

  Although the metal ball 600 and the circular caliper 410 may be misaligned, the first local coordinate system is set with respect to the circuit pattern on the semiconductor chip 100a, so even if there is a dicing line misalignment or the like, The positional deviation direction and the positional deviation amount are at the same level. If the measurement program is modified so that the circle caliper 440 comes on the metal ball 600 by finely adjusting the display position of the circle caliper 440 with respect to the first local coordinate system, it will always be continued regardless of dicing line misalignment, etc. The circle caliper 440 is displayed at the correct position.

  If the display position of the circle caliper 440 with respect to all the metal balls 600 is designated with respect to the first local coordinate system 520, the display position of the circle caliper 440 with respect to all the metal balls 600 is finely adjusted. Since it is necessary, fine adjustment work becomes enormous.

  Since the positional relationship between the metal pads on the semiconductor chip 100a is determined by the chip design and hardly varies compared to the bonding position or the chip position, the second local coordinates are set at the center of the first metal ball. If the system 530 is set and the circle caliper display position for the second and subsequent metal balls is indicated by the distance from the first metal ball, the position adjustment of the circle caliper with respect to the second and subsequent metal balls is unnecessary. become. The adjustment work of the measurement program is greatly reduced.

  The setting of the second local coordinate system 530 described in FIG. 10 is desirably performed for each side of the semiconductor chip 100a. In other words, the possibility that the metal ball 600 and the first local coordinate system 520 are shifted in different directions for each side is small, but the second local coordinate system is formed on the first metal ball on the first chip side. When 530 is set, both the X coordinate value and the Y coordinate value of the coordinate values of the metal balls on the second to fourth chip sides change for each metal ball. If the second local coordinate system is set on the first metal ball on each chip side, either the X coordinate value or the Y coordinate value is always zero, so the coordinate value must be specified. It becomes easy. Since the distance from the second local coordinate system is relatively short, an error in the circular caliper display hardly occurs.

  Therefore, it is desirable to set the second local coordinate system 530 for the metal ball 600 for each side.

  The setting of the second local coordinate system 530 is automatically set by the measurement program based on the measurement result (metal ball position) of the first metal ball on each chip side.

  After partially executing the measurement program to the point where the circle caliper that measures the first metal ball on each chip side is displayed, finely adjust the position of the circle caliper on the metal ball. Save again. This operation only needs to be performed once before entering the actual measurement. Even if the bonding wire measurement of the semiconductor device (same type) is performed many times, the circular caliper is correctly displayed on the metal ball.

  Subsequently, in step S230 of FIG. 5, reference points for measuring the ball crush diameter and the wire loop height are set.

  As shown in FIG. 11, the height is measured with respect to a point on the metal pad 605 where the metal ball 600 is not arranged, and is set as a reference point. In FIG. 11, as a specific example, a laser is used to focus at the point of the cursor 460, and the height is measured.

  Further, as shown in FIG. 12, the metal ball 600 and the metal pad 605 may overlap each other. In this case, the height of the surface of the semiconductor chip 100a is measured around the metal pad 605. The surface of the semiconductor chip 100a is covered with a polyimide film (not shown) for circuit pattern protection. The polyimide film is translucent. Here, there are cases where the measured value is experimentally higher than the actual height by about ½ of the thickness of the translucent film. In other words, an approximate correct height can be obtained by subtracting the height of half of the thickness of the translucent film from the measured height value.

  Here, the reason why the measured value is about 1/2 higher than the thickness of the polyimide film is obtained, for example, as follows.

  When focusing using laser light, the laser light is reflected on the surface of the polyimide film, reflected in the middle of the polyimide film, or reflected on the upper surface of the circuit pattern below the polyimide film You could think so. Here, the upper surface of the circuit pattern is substantially at the same height as the surface of the metal pad 605. In addition, the above three cases are assumed to occur with almost the same probability. Therefore, on average, the measured value is about 1/2 of the thickness of the polyimide film and higher than the actual metal pad 605.

  Here, it is assumed that the polyimide film is several um, which is several times thicker than the circuit pattern, and the thickness of the circuit pattern itself can be ignored. When the circuit pattern is thick, such as a three-dimensional memory, the measured value is not necessarily about half that of the polyimide film.

  Subsequently, in step S240 of FIG. 5, the ball collapse diameter and the ball position are measured for the metal ball 600.

  A method for measuring the ball crush diameter and the ball position will be described with reference to FIGS. FIG. 14 is a schematic cross-sectional view of a metal ball 600 to be measured. First, as shown in FIG. 14, the length in the Z direction of the metal ball 600, that is, h <b> 1 is referred to as a collapsed thickness h <b> 1 of the metal ball 600. The length that is ½ of the collapsed ball thickness, that is, h2 is referred to as a ball collapsed diameter measurement height h2 of the metal ball 600. The diameter of the metal ball 600 at the ball collapse diameter height h2 is referred to as a ball collapse diameter w1.

  The ball collapse thickness h1 may be measured with laser light, or may be set as a predetermined value in the measurement program in advance. Then, the ball collapse diameter measurement height h2 is calculated from the ball collapse thickness h1.

  As shown in FIG. 13, after the focus of the camera 40 is adjusted to the ball collapse diameter height h2, the length of the ball collapse diameter w1 is measured using a circular caliper 470.

  Further, in FIG. 13, the ball crush diameter w1 is measured, and the position of the metal ball 600 is measured.

  Subsequently, in step S250 of FIG. 5, the third local coordinate system 540 is set for the wire portion 610 of the bonding wire 150.

  FIGS. 15A and 15B are enlarged views of regions A2 and A3 of FIG. 6, respectively. That is, FIGS. 15A and 15B are enlarged views of the vicinity of the metal ball 600 and the bonding position 630 with respect to the wire portion 610, respectively. The bonding position 630 is a bonding position where the wire portion 610 is stitch bonded to the electrode 620 on the substrate 110, for example. The bonding position 630 can be obtained by using, for example, a value extracted from a wire bonding coordinate data file or a value calculated using the value.

  Then, a third local coordinate system 540 is set along a straight line connecting the metal ball 600 connected to the wire portion 610 to be measured and the bonding position 630. The third local coordinate system 540 is disposed at a position where the edge of the semiconductor chip 100a and the wire portion 610 overlap.

  Subsequently, in step S <b> 260 of FIG. 5, the height of the wire portion 610 is measured along the third local coordinate system 540 with respect to the wire portion 610.

  As shown in FIG. 16A, at the origin of the third local coordinate system 540, the controller 70 measures the bonding wire height using the height measurement caliper 480a. Here, the measurement caliper 480a detects the edge of the measurement target existing in the image, and the measurement target image stored in advance using the least square method or the like (here, measurement having two substantially parallel straight lines) The object image is a cursor image for searching for a measurement symmetry image for a program module that measures the in-focus height. That is, the controller 70 measures the height of the wire portion 610 at the edge portion of the semiconductor chip 100a. The height of the wire portion 610 is also read as the loop edge height.

  As shown in FIG. 16A, the controller 70 follows the third local coordinate system 540 in the direction of the metal ball and the direction of the bonding position 630 around the point where the bonding wire is assumed to have the highest value. Measure several points and set the maximum value among them as the loop height.

  FIG. 16 shows an example in which the height of the bonding wire is measured at four points as an example. Of course, the height of the bonding wire may be measured more than four points or less.

  The point where the bonding wire takes the maximum value can be roughly estimated from the forward bonding / reverse bonding, the bonding loop mode of the wire bonder, and the like.

  The controller 70 measures the height of the object, that is, the wire portion 610 as follows, for example. That is, the controller 70 moves the lens of the camera 40 up and down so that the object, that is, the wire portion 610 is focused inside the height calipers 480a to 480d. The controller 70 calculates the difference between the height of the metal pad obtained in FIG. 11 and the like and the bonding wire height obtained by the height caliper.

  Here, when the height of the wire portion 610 is measured, it is desirable that the contrast of the image data of the upper end of the bonding wire photographed by the camera 40 is large. That is, when the contrast is low, it is difficult for the camera 40 to focus at the upper end of the bonding wire, which may cause an error in the height measurement of the wire portion 610.

  Next, a configuration capable of further enhancing the contrast will be described. FIG. 16B is a cross-sectional view schematically showing a part of the camera 40 of the head unit 20. A ring light 700 is disposed around the lens 720 included in the camera 40.

  When the height of the wire portion 610 is measured, only part of the ring light 700 is lit. For example, the ring light 700 is a ring-shaped light source, and half of the light is lit. Then, since light is applied to the wire portion 610 from one direction, the brightness of the wire portion 610 is emphasized. That is, the controller 70 can easily focus the camera 40 on the wire portion 610.

  Further, depending on the orientation and material of the wire portion 610, the light and darkness of the wire portion 610 may be emphasized more in the light bulb illumination 710 disposed above the lens 720 than in the ring light 700. In this case, the bonding wire may be illuminated using the light bulb illumination 710 without using the ring light 700.

  When the thickness of the wire portion 610 is sufficient, the height of the wire portion 610 may be directly measured using a laser displacement meter (laser rangefinder) or the like.

  Subsequently, as shown in step S260 of FIG. 5, the controller 70 outputs the above measurement data to a text file or the like.

  Furthermore, as shown in the flow of FIG. 5, the height of the wire portion 610 as much as necessary is measured by performing the processing of step S230 to step S270 by the number of wire portions 610 that need to be measured. The result is output.

  Although the measurement procedure for one semiconductor chip 100a has been described as an example, the semiconductor chips 100a may be stacked in multiple stages. That is, when a plurality of semiconductor chips 100 are arranged on the substrate 110, the height of the wire portions 610 of the plurality of semiconductor chips 100 is increased by repeating the above procedure for each semiconductor chip 100. It is possible to measure.

  According to the present embodiment, the third local coordinate system 540 is set on the wire portion 610 based on the metal ball 600 and the bonding position 630. Then, the height of the wire portion 610 is measured based on the third local coordinate system 540.

  The wire portion 610 may be arranged with various inclinations. Since the third local coordinate system 540 is arranged based on the metal ball 600 and the bonding position 630, the height measuring caliper 480 shown in FIG. 16 can be easily arranged.

  In addition, according to the present embodiment, by using the measurement program creation macro, it is possible to easily create a measurement program that takes a lot of time and effort. Furthermore, the height of the wire portion 610 can be automatically measured by causing the measurement apparatus 10 to read the measurement program.

  That is, after setting the substrate 110 to the measuring apparatus 10, it is possible to automatically measure a large number of samples with little manual intervention. Furthermore, it is possible to detect defective products more efficiently by saving the measurement work and increasing the number of measurement samples.

  Although the embodiment of the present invention has been described, this embodiment is presented as an example and is not intended to limit the scope of the invention. The novel embodiment can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. This embodiment and its modifications are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 10 ... Measuring apparatus 20 ... Head part 30 ... First drive part 40 ... Camera 50 ... Stage 60 ... Second drive part 70 ... Controller 80 ... Display part 100 ... Semiconductor chip 110 ... Substrate 120 ... Step 150 ... Bonding wire 400 ... Alignment Mark 410 ... Circle caliper 420 ... Rectangular caliper 430 ... Search caliper 440 ... Circle caliper 460 ... Cursor 470 ... Circle caliper 480 ... Measurement caliper 500 ... Local coordinate system 510 ... Local temporary coordinate system 520 ... First local coordinate system 530 ... Second Local coordinate system 540 ... Third local coordinate system 600 ... Metal ball 605 ... Metal pad 610 ... Wire portion 620 ... Electrode 630 ... Bonding position 700 ... Ring light 710 ... Light bulb illumination 720 ... Lens

Claims (5)

Correct the tilt of the board,
Correcting the tilt of the semiconductor chip disposed on the substrate,
A local coordinate system is set for the plurality of bonding wires along the path of the bonding wire disposed between the electrode on the substrate and the electrode pad on the semiconductor chip disposed on the substrate;
The height of the bonding wire, multiple measurements along each kilometer Karu coordinate system before corresponding to the bonding wire,
A method for measuring a semiconductor device. The reference point of the local coordinate system is set at the intersection of the bonding wire and the edge of the semiconductor chip when viewed from above.
The method for measuring a semiconductor device according to claim 1. Measure the collapse diameter of the ball connected to the bonding wire, and the position of the ball,
The method for measuring a semiconductor device according to claim 1, wherein the crushing diameter of the ball is measured by measuring a diameter at a predetermined height. After measuring the height on the upper surface of the semitransparent film disposed on the semiconductor chip, the reference point when measuring the height of the bonding wire by reducing the film thickness of approximately half of the semitransparent film To calculate,
The method for measuring a semiconductor device according to claim 1 . Measurement method of a semiconductor device according to the height of the bonding wire was measured the plurality of times, in any one of claims 1 to 4 for calculating the highest height of the bonding wire.

Images