JP4520260B2 - Wafer defect detection method and apparatus - Google Patents

Description

  The present invention relates to a wafer defect detection method and apparatus for detecting defects generated at the peripheral edge of a disk-shaped semiconductor wafer in which notches or orientation flats are formed.

  As a first conventional technique, Patent Document 1 discloses a wafer position detection device that detects an orientation flat (hereinafter referred to as an orientation flat). This position detection device causes the wafer to be stepped so that the wafer is angularly displaced about its axis by a predetermined angle, and the shape of the peripheral edge of the wafer is detected by a line sensor.

The position detection device calculates the position where the orientation flat is formed based on the angular displacement amount of the wafer and the image sensor output. Specifically, the angular position of the portion θm _{+ 4} where the distance from the rotation center of the wafer to the peripheral portion is the shortest is detected, and the angular position of the orientation flat is detected based on the angular position.

  As a second conventional technique, Patent Document 2 discloses a substrate processing apparatus having a function of detecting an orientation flat. This substrate processing apparatus detects the angular positions of both side portions P and Q in the circumferential direction of the orientation flat, and detects the central angular position of the orientation flat based on the angular positions of both side portions P and Q.

Japanese Patent Publication No. 7-13998 JP-A-10-270404

  In the first prior art, since it is assumed that the depth of the defect notch formed at the peripheral edge of the wafer is smaller than the depth of the orientation flat where the orientation flat is formed, the defect notch is deeper than the orientation flat. When the depth of the part is the same, the orientation flat part and the defect notch part cannot be distinguished. Further, when the depth of the defect notch portion is larger than the depth of the orientation flat portion, the defect notch portion may be erroneously detected as the orientation flat portion. Therefore, it is impossible to accurately detect the defective notch portion of the wafer.

  The second prior art also does not disclose a configuration for detecting a defect notch portion of a wafer, and cannot detect a defect notch portion. As in the first prior art, the orientation flat portion and the defect notch portion cannot be distinguished from each other, and the defect notch portion and the orientation flat portion may be erroneously detected.

  In addition, the case where the orientation flat part was formed in the wafer was demonstrated as a prior art. However, the problem described above occurs even when the notch is formed in the wafer. The notch portion where the notch is formed is smaller in depth and width than the orientation flat portion. When the above-described prior art is applied to detect the notch portion, the possibility of erroneous detection of the defect notch portion and the notch portion increases.

  Accordingly, it is an object of the present invention to provide a wafer defect detection method and apparatus capable of accurately detecting a defect notch portion generated in the peripheral portion of a semiconductor wafer.

The present invention is a wafer defect detection method for detecting a defect notch portion generated in a peripheral portion of a disc-shaped semiconductor wafer in which a predetermined notch portion is formed in a peripheral portion,
When the peripheral straight line extending from the axis set in advance to the wafer to the peripheral edge of the wafer is angularly displaced around the axis, the distance change of the peripheral straight line is measured.
Based on the measurement result, a notch to which attention is paid is a width representing an angular displacement amount at which the peripheral straight line passes through the notch portion, and a depth related to the peripheral straight line at which the distance of the peripheral straight line becomes the smallest in the notch portion. Calculate for each part,
When at least one of the width and depth of the notch portion to be noticed is outside the specified range set in the notch portion, the notch portion to be noticed is detected as a defect notch portion ,
When the width and depth of the notch part of interest are within the specified range set for the notch part, calculate a curve formula representing the distance change of the peripheral straight line in the notch part of interest,
If the curve equation calculated for the notched portion of interest falls within a predetermined similar range to the prescribed curve equation representing the change in the distance of the peripheral straight line at the prescribed notch portion, the notched portion of interest is defined as the prescribed notch portion. A defect detection method for a wafer, wherein the defect is detected as a chipped portion .

  Further, the present invention is characterized in that, based on the measurement result, an angular displacement position around the axis of the wafer with respect to an arbitrary reference position set for the wafer is obtained for the defect notch portion.

The present invention is, for provisions notch portion, and obtains the angular displacement position about the axis of the wafer relative to an arbitrary reference position set in the wafer.

  Further, the present invention provides a notch portion that is the closest to the specified curve formula when the curve formula calculated for a plurality of notched portions of interest is within a predetermined similar range to the specified curve formula. Is detected as a specified notch.

  Further, the present invention is characterized in that, in the change in the distance of the peripheral straight line, the width and depth of the notch portion are obtained based on a calculation result obtained by removing a frequency component equal to a period in which the peripheral straight line goes around the axis.

  Further, the present invention is characterized in that the notch portion to be noticed is ignored when the width and depth of the notch portion to be noticed are within a predetermined allowable range.

  Further, the present invention is characterized in that the prescribed notch portion is a portion where one of a notch and an orientation flat is formed.

Further, the present invention is a defect detection apparatus for a wafer for detecting a defect notch portion generated in a peripheral portion of a disc-shaped semiconductor wafer in which a predetermined notch portion is formed in a peripheral portion,
The base,
A holding unit for holding the wafer;
Rotation driving means for rotating the holding unit around a central axis set on the base;
Angle measuring means for measuring the amount of angular displacement of the holding part;
A peripheral position measurement unit that measures a position in the wafer radial direction of the peripheral portion of the wafer, facing a part in the peripheral direction of the peripheral portion of the wafer held by the holding unit;
Based on the measurement results of the angle measurement means and the peripheral position measurement means, and having a defect detection means for detecting a defect notch portion,
Defect detection means
When the wafer is angularly displaced and the position change in the wafer radial direction of the wafer peripheral portion becomes larger than a predetermined threshold value, a notch portion is formed in the peripheral portion of the wafer facing the peripheral position measuring means. Detect
The depth associated with the width representing the angular displacement of the wafer while the notch passes through the peripheral position measuring means, and the depth related to the distance of the peripheral straight line where the distance from the central axis to the wafer peripheral edge becomes the smallest at the notch. For each notch to be noticed,
When at least one of the width and depth of the notch portion to be noticed is outside the specified range set in the notch portion, the notch portion to be noticed is detected as a defect notch portion ,
When the width and depth of the notch part of interest are within the specified range set for the notch part, calculate a curve formula representing the distance change of the peripheral straight line in the notch part of interest,
If the curve equation calculated for the notched portion of interest falls within a predetermined similar range to the prescribed curve equation representing the change in the distance of the peripheral straight line at the prescribed notch portion, the notched portion of interest is defined as the prescribed notch portion. A defect detection apparatus for a wafer, wherein the defect detection device detects the defect portion .

Further, the present invention is characterized in that the defect detecting means obtains an angular displacement position around the axis of the wafer with respect to an arbitrary reference position set on the wafer, for the defect notch portion , based on the measurement result .

  According to the present invention, the width and depth of the notch are determined. When at least one of the width and depth of the specified notch is outside the specified range set for the specified notch, the notch is detected as a defect notch. As a result, it is possible to accurately detect the defect notch portion without erroneously detecting the defect notch portion generated on the wafer as the specified notch portion.

Even when both the prescribed notch portion and the defect notch portion are formed on the wafer, the prescribed notch portion and the defect notch portion can be detected separately. As a result, the wafer on which the defect notch is formed can be removed.
Further, a curve equation representing a change in the distance of the peripheral straight line in the notch portion is calculated, and based on the curve equation, it is determined whether the notch portion is a specified notch portion or a defect notch portion. As a result, even if a defect notch that occurs within the specified range of the width and depth of the specified notch is generated, it can be detected as a defect notch, and the specified notch is erroneously detected. Can be prevented.

Further, according to the present invention, the occurrence position of the defect notch portion can be grasped by obtaining the angular displacement position of the defect notch portion with respect to an arbitrary reference position set on the wafer. This enables the score facilities patterning process in a portion except-out defects cutouts among the wafers.

Further, according to the present invention, the occurrence position of the specified notch portion can be grasped by obtaining the angular displacement position of the specified notch portion with respect to an arbitrary reference position set on the wafer .

  Further, according to the present invention, when a calculated curve formula is within a predetermined similar range with respect to the specified curve formula, a curve formula closest to the specified curve formula is obtained for a plurality of notched portions to be noticed. A notch is detected as a specified notch. As a result, it is possible to accurately distinguish and detect one specified notch portion and other defect notch portions from a plurality of notch portions.

  Further, according to the present invention, when the axis set in advance on the wafer is eccentric from the center position of the wafer, the distance change of the peripheral straight line includes a frequency component having a period equal to one period in which the wafer is angularly displaced. . By removing this frequency component, it is possible to prevent the eccentricity of the wafer from affecting the distance change of the peripheral straight line. As a result, the width and depth of the notch portion can be accurately obtained, and therefore the specified notch portion and the defect notch portion can be detected with higher accuracy.

  In addition, even if the wafer is in an eccentric state, the specified notch portion can be detected with high accuracy, so that it is possible to easily measure the peripheral distance change. Further, based on this frequency component, it is also possible to obtain an eccentric amount and an eccentric direction in which the center position of the wafer is eccentric with respect to an axis set in advance on the wafer.

  Further, according to the present invention, when the depth and width of the notch of interest are within the allowable range, the notch of interest is ignored. As a result, it is possible to accurately detect the notched defect notch. For example, it is possible to prevent erroneous detection as a defect even if the depth and width of the notch part are small and have no problem in wafer quality. In addition, by registering the depth and width of the unevenness generated in the wafer manufacturing process as an allowable range, it is possible to prevent the uneven part generated in the manufacturing process from being detected as a defect.

  Further, according to the present invention, the specified notch portion is a portion where either one of the notch and the orientation flat is formed. Thereby, it is possible to distinguish and detect the defect notch portion and the notch portion or the orientation flat portion. In addition, it is possible to determine whether the wafer is a notch wafer in which a notch is formed or an orientation flat wafer in which an orientation flat is formed.

  Furthermore, both the notch part and the orientation flat part may be set as the prescribed notch part as the prescribed notch part. In this case, the detection device can be used regardless of whether the notch part or orientation flat part is formed on the wafer by setting the determination reference values of the notch part and orientation flat part independently. It becomes possible and the convenience can be improved.

  According to the present invention, the wafer is rotated together with the holding unit by the rotation driving means while the wafer is held by the holding unit. As the wafer rotates, the peripheral edge of the wafer facing the angle measuring means changes. When there is a notch in the peripheral edge of the wafer, the measurement result measured by the peripheral position measuring means changes when the wafer is angularly displaced.

The defect detection means obtains the width and depth of the notch portion based on the measurement results by the angle measurement means and the peripheral position measurement means. Then, if the width and depth of the notch portion of interest are within the prescribed range of the width and depth set in the prescribed notch portion determined in advance, it is detected as the prescribed notch portion, and if it is outside the prescribed range, the defect is detected. Detect as a notch. As described above, by determining whether or not the notched portion of interest is a defect based on the width and depth of the notched portion, the defect can be detected with high accuracy.
In addition, the defect detection means calculates a curve equation representing the distance change of the peripheral straight line in the notch portion, and based on the curve equation, whether the notch portion is the specified notch portion or the defect notch portion. Judging. As a result, even if a defect notch that occurs within the specified range of the width and depth of the specified notch is generated, it can be detected as a defect notch, and the specified notch is erroneously detected. Can be prevented.

Further, according to the present invention, the occurrence position of the defect notch portion can be grasped by obtaining the angular displacement position of the defect notch portion with respect to an arbitrary reference position set on the wafer. As a result, a pattern forming process can be performed on a portion of the wafer excluding the defect notch portion .

  FIG. 1 is a block diagram showing a configuration of a wafer shape detection apparatus 20 according to an embodiment of the present invention. The wafer shape detection device 20 serves both as a so-called aligner device that adjusts the angular displacement position of the wafer 21 and a wafer defect detection device that detects defects in the wafer 21.

  The crystal orientation of the semiconductor wafer 21 is set in advance. The semiconductor wafer 21 has a specified notch formed on the peripheral edge thereof for discriminating the crystal direction. The semiconductor wafer 21 may be subjected to various semiconductor processes in an aligned state so that its crystal direction is aligned with a predetermined direction of the manufacturing apparatus. In this case, the wafer 21 is arranged so that the specified notch portion of the wafer 21 faces a predetermined position of the manufacturing apparatus.

FIG. 2 is an enlarged view showing a part of the wafer 21 for explaining the notch 22. The semiconductor wafer 21 is formed in a disk shape. In the first embodiment, the defined notch portion is realized by a notch portion 80 in which the notch 22 is formed. The notch 22 is a recess formed in the peripheral edge 23 of the semiconductor wafer 21. When the wafer 21 is viewed in a cross section perpendicular to the thickness direction, the notch portion 80 forms a substantially U-shaped cut end that tapers toward the center position of the wafer 21 in the peripheral portion 23. The detailed dimension of the notch 22 is SEMI (Semiconductor
Standardized in advance by the Equipment and Materials Institute).

  The wafer shape detection device 20 detects the notch portion 80. The wafer shape detection device 20 angularly displaces the wafer 21 so that the notch portion 80 faces the reference position set in the wafer shape detection device 20. Thereby, the position of the wafer 21 is adjusted so that the crystal direction of the wafer 21 extends in a predetermined direction. In this state, the wafer 21 is transferred from the wafer shape detection device 20 by a transfer device such as a robot hand, and is aligned with a predetermined other semiconductor processing device in consideration of the crystal direction.

  Further, the wafer shape detection device 20 detects a defect notch portion 82 that is formed in the peripheral portion of the semiconductor wafer 21 and in which a recess 83 is formed as a defect. When the defect notch portion 82 is detected, the operator determines that the defect notch portion 82 is formed on the wafer 21 by notifying the generation of the defect notch portion 82. In addition, the wafer shape detection device 20 notifies a transfer device such as a robot hand that a defect notch portion 82 has occurred in the wafer 21. In this case, the transfer apparatus removes the wafer 21 in which the defect notch portion 82 is generated, and prevents the wafer 21 in which the defect is generated from being processed in the subsequent semiconductor processing apparatus.

  The wafer shape detection apparatus 20 includes a base 30, a holding unit 31, a motor 32, an encoder 33, a line sensor 34, a sensor controller 35, and a control unit 36. The base 30 supports the holding unit 31, the motor 32, the encoder 33, and the line sensor 34.

  The holding unit 31 holds the semiconductor wafer 21 by suction. The holding portion 31 is supported by the base portion 30 so as to be rotatable around a central axis 29 preset in the base portion 30. Specifically, the holding portion 31 is formed with a contact surface with which the wafer 21 contacts. A suction hole for sucking air is formed in the contact surface. With the semiconductor wafer 21 in contact with the contact surface, the semiconductor wafer 21 can be sucked and held by the holding portion 31 by sucking air from the suction holes. The semiconductor wafer 21 is brought into contact with the holding portion 31 so that the center position 28 is arranged on the central axis 29 of the base portion 30 as much as possible.

  The motor 32 continuously rotates the holding unit 31 around a predetermined central axis 29 with respect to the base 30. The motor 32 transmits the rotation output to the holding unit 31 via the power transmission mechanism 37. Accordingly, when the output shaft of the motor 32 rotates, the holding unit 31 and the wafer 21 held by the holding unit 31 continuously rotate around the central axis 29. The motor 32 is realized by, for example, a direct current motor, and serves as a rotational drive unit that rotationally drives the holding unit 31.

  The encoder 33 measures the amount of angular displacement in which the output shaft is angularly displaced from the state before rotation. That is, the encoder 33 serves as an angle measuring unit that measures the angular displacement amount of the holding unit 31. The encoder 33 may be such an incremental type, but may be an absolute type.

  The line sensor 34 is fixed to the base 30. The line sensor 34 is disposed at a position facing a part of the peripheral edge 23 of the wafer 22 held by the holding unit 31. The line sensor 34 measures the position in the radial direction r of the wafer 21 with respect to the peripheral edge of the wafer. Specifically, the line sensor 34 includes a light projecting unit 34a and a light receiving unit 34b.

  The light projecting unit 34 a projects belt-shaped parallel light that extends in the wafer radial direction r and travels parallel to the central axis 29. The light receiving unit 34b is disposed to face the light projecting unit 34a with the held wafer 21 interposed therebetween. For example, the light projecting unit 34a is realized by a light emitting diode element or a semiconductor laser element. The light receiving unit 34b is realized by a CCD (Charge Coupled Device) image sensor. Specifically, the light receiving unit 34b is configured such that a large number of fine photodiodes are arranged in the radial direction r of the wafer and can receive the parallel light projected from the light projecting unit 34a.

  A part of the light projected from the light projecting unit 34 a is shielded by the wafer 21. The remaining part passes radially outward from the wafer 21 and enters the light receiving portion 34b. When the position of the peripheral edge of the wafer 21 approaches the radial direction r with respect to the central axis 29, the light shielded by the wafer 21 decreases, and the amount of light incident on the light receiving unit 34b increases. Further, when the position of the peripheral edge of the wafer 21 is moved away from the central axis 29 in the radial direction r, the light shielded by the wafer 21 increases and the amount of light incident on the light receiving unit 34b decreases. The amount of electricity output from the light receiving unit 34b varies depending on the amount of light incident from the light projecting unit 34a.

  The sensor controller 35 gives a light projection command to the light projecting unit 34a so as to project light toward the light receiving unit 34b. Further, the sensor controller 35 detects the position in the radial direction r of the peripheral edge at the portion facing the line sensor 34 based on the amount of electricity output from the light receiving unit 34b.

  The controller 36 is given a position in the radial direction r of the peripheral edge facing the line sensor 34 from the sensor controller 35. Further, the control unit 36 gives a rotation command to the motor 32, and receives an angular displacement amount of the semiconductor wafer 21 from the encoder 33.

  The control unit 36 determines the angular position of the notch portion 80 and the defect notch portion 82 based on information given from the sensor controller 35 and the encoder 33 in a state where the wafer 21 is continuously rotated by the motor 32. To detect. When detecting the defect notch portion 82, the control unit 36 outputs that the defect notch portion 82 is generated on the wafer 21. For example, the control unit 36 notifies the presence or absence of the defect notch portion 82.

  Further, in order to direct the crystal direction of the wafer 21 in a predetermined direction, the control unit 36 gives a control command to the motor 32 so as to arrange the notch portion 80 at a predetermined angular position. For example, after the notch portion 80 passes through the line sensor 34, the motor 32 is controlled so that the wafer 21 stops rotating at the angular displacement position that is angularly displaced by a predetermined arbitrary amount of angular displacement.

  The control unit 36 includes an arithmetic circuit that performs an operation, and a storage circuit that stores a predetermined program and set numerical values. The control unit 36 detects the notch portion 80 and the defect notch portion 82 by executing a program stored in the storage circuit by the arithmetic circuit. For example, the arithmetic circuit is realized by a CPU. For example, the memory circuit is realized by a RAM and a ROM.

  In a state where the wafer 21 is held by the holding unit 31, the wafer 21 is continuously rotated together with the holding unit 31 by one or more rounds by the motor 32. At this time, the encoder 33 detects the angular displacement amount of the wafer 21 with respect to the state before the rotation as an encoder value. The encoder value increases as the angular displacement amount of the wafer 21 increases.

  When the wafer 21 is continuously rotated, the peripheral edge 23 facing the line sensor 34 changes. The line sensor 34 detects the position in the radial direction r of the peripheral edge at the portion facing the line sensor 34 as a sensor value.

  In addition, when the cross-sectional shape in a plane perpendicular to the thickness direction of the wafer 21 is formed in a perfect circle, and the center position 28 of the wafer 21 and the center axis 29 of the base portion 30 coincide with each other, the sensor value is constant. become. On the other hand, when a notch is formed in the wafer 21 and when the wafer 21 is eccentric, the sensor value also changes with the angular displacement of the wafer 21.

  In the present embodiment, the smaller the sensor value, the more the peripheral edge of the portion facing the line sensor 34 is located inward in the radial direction r. In the present embodiment, the greater the encoder value, the greater the angular displacement of the wafer 21 from the state before rotation. In the present invention, the radius 85 of the wafer 21 corresponds to the radius 85 of the wafer 21 when the wafer 21 is formed in a perfect circle without a notch. The center position 28 of the wafer 21 corresponds to the center position of the wafer 21 when the cross section of the wafer 21 is formed in a perfect circle.

  FIG. 3 is a front view showing the notch wafer 21 in which the three notches 25, 26 and 27 are generated. FIG. 4 is a graph showing an encoder value e when the notch wafer 21 shown in FIG. 3 is angularly displaced, and a sensor value c corresponding to the encoder value e. The notch wafer 21 shown in FIG. 3 includes a first cutout portion 127 in which a first cutout 27 is formed, a second cutout portion 125 in which a second cutout 25 is formed, and a third cutout 126. And a third notch portion 126 is formed.

  FIG. 4 shows a measurement result in a state where the center position 28 of the wafer 21 and the center axis 29 of the base portion 30 coincide with each other. The sensor value c corresponds to a distance of a peripheral distance 40 facing the line sensor 34 among the peripheral straight lines 40 extending from the central axis 29 to the peripheral edge of the wafer 21. The encoder value e corresponds to the amount of angular displacement from the state before rotation until the arbitrary wafer peripheral edge 23 faces the line sensor 34.

  In the case of the notch wafer 21 shown in FIG. 3, when the wafer 21 is displaced by about 90 ° from the state before the angular displacement, the first cutout portion 127 faces the line sensor 34. When the wafer 21 is angularly displaced by about 175 ° from the state before the angular displacement, the second notch portion 125 faces the line sensor 34. When the wafer 21 is angularly displaced by about 260 ° from the state before the angular displacement, the third notch portion 126 faces the line sensor 34.

  When the wafer 21 is angularly displaced and the first cutout portion 127 comes close to the line sensor 34, first, a peripheral straight line 40 (hereinafter referred to as a peripheral straight line on the line) facing the line sensor 34 at the first angular displacement position 42. The distance from the radius 85 of the wafer 21 is smaller than the radius 85 of the wafer 21. Further, when the wafer 21 is angularly displaced, the distance of the peripheral straight line 40 on the line is gradually shortened.

  When the wafer 21 is further angularly displaced after reaching the second angular displacement position 44 from the first angular displacement position 42, the distance between the peripheral straight lines on the line gradually increases. When the third angular displacement position 43 is reached, the distance from the peripheral straight line on the line returns from a state where it is smaller than the radius 85 of the wafer 21 to a state where it is substantially equal to the radius 85 of the wafer 21. In the present embodiment, the substantially equal state includes an equal state.

  When the wafer 21 is angularly displaced and the second notch portion 125 and the third notch portion 126 are close to the line sensor 34, the same is true as when the first notch portion 127 is close. Specifically, the distance between the peripheral straight lines on the line gradually decreases after reaching the first angular displacement position 42. Then, after reaching the second angular displacement position 44, it gradually becomes longer, and when it reaches the third angular displacement position 43, it returns to a state almost equal to the radius 85 of the wafer 21. Each angular displacement position 42, 43, 44 described above is obtained based on the amount of angular displacement with respect to the angular displacement position before rotation.

  The controller 36 detects the first angular displacement position 42, the second angular displacement position 44, and the third angular displacement position 43 based on the sensor value c and the encoder value e when the wafer is rotated. When detecting the respective angular displacement positions 42 to 44, the control unit 36 determines that a notch is formed between the first angular displacement position 42 and the third angular displacement position 43.

  Further, the control unit 36 sets the angular displacement amount between the first angular displacement position 42 and the third angular displacement position 43 as the widths W1, W2, and W3 of the notch portions. The distance obtained by subtracting the peripheral straight line distance 40 at the second angular displacement position 44 from the radius 85 of the wafer 21 is defined as the depths H1, H2, and H3 of the notches. Further, an angular displacement position between the first angular displacement position 42 and the third angular displacement position 43 in the notch portion is set as an angular displacement position 45 in the notch portion.

  The prescribed ranges of the width and depth allowed for the notch portion 80 are arbitrary values set in advance. Such prescribed ranges of the width and depth of the notch 22 are stored in advance in the storage circuit.

  In the case of the notch wafer 21 shown in FIG. 3, the first notch portion 127 has a width W1 larger than the upper limit value of the specified range, and a depth H1 larger than the upper limit value of the specified range. Therefore, the control unit 36 does not detect the first notch portion 127 as the notch portion 80 but detects it as the defect notch portion 82.

  Further, the width W2 of the second notch portion 125 is larger than the lower limit value of the specified range and smaller than the upper limit value. The depth H2 is larger than the lower limit value of the specified range and smaller than the upper limit value. Therefore, the control unit 36 detects the second notch portion 125 as the notch portion 80.

  The third notch 126 has a width W3 smaller than the lower limit value of the specified range, and a depth H3 smaller than the lower limit value of the specified range. Therefore, the control unit 36 does not detect the third notch portion 126 as the notch portion 80 but detects it as the defect notch portion 82. Thus, the control part 36 judges as the defect notch part 82, when at least any one of the width W and the depth H of the notch part 25-27 to which it pays attention does not enter in the regulation range.

  FIG. 5 is a flowchart showing the first notch detection operation of the control unit 36. When detecting the defect notch portion 82 and the notch portion 80 from the sensor value c and the encoder value e, the control unit 36 needs to process data so that the detection operation can be executed.

  In step a0, the control unit 36 is given a predetermined range of the width W and the depth H of the notch portion 80 in advance by a user or the like. The control unit 36 stores the given specified range in the storage circuit. In the state where the preparatory operation is completed in this way, the control unit 36 continuously rotates the wafer 21 held by the holding unit 31, and proceeds to step a1.

  In step a <b> 1, the control unit 36 acquires the sensor value c from the sensor controller 35 and acquires the encoder value e from the encoder 33. The control unit 36 stores the encoder value e and the sensor value c acquired at the same time in association with each other in the storage circuit.

  FIG. 6 is a graph in which the sensor value c acquired in step a1 is arranged in the order of the encoder value e. FIG. 6 shows a graph with the sensor value c on the vertical axis and the encoder value e on the horizontal axis. Further, the graph shown in FIG. 6 is a detection result when the center axis 29 of the base 30 and the center position 28 of the wafer 21 are eccentric. The control unit 36 uses the sensor value c associated with each encoder value e as detection data, and proceeds to step a2.

  In step a2, the control unit 36 extracts an eccentric frequency component change corresponding to a period when the wafer 21 goes around from one change in each detection data c. Specifically, the control unit 36 first extracts an arbitrary number of detection data from the detection data c. Then, with respect to the increase of the encoder value e, an arithmetic expression of an eccentric sine wave waveform representing a change in the eccentric frequency component of each extracted detection data c is obtained.

  In the present embodiment, detection data is extracted in order by detecting data c every time the encoder value changes by 12 °. The set range in which the difference between the i-th extracted detection data ci and the i-1th extracted detection data c (i-1) among the detected data c extracted is determined in advance. Is exceeded, the i-th extracted detection data ci is removed, and an arithmetic expression of the eccentric sine wave waveform is obtained using the remaining detection data.

As the arithmetic expression, three of the extracted detection data are used, and the detection data c is substituted into Y of the arithmetic expression in the following expression (1), and the encoder value e corresponding to the detection data c is substituted into X. Then, the values of the constants α, β, γ are obtained. For the three detection data, data having encoder values e separated by 120 ° are used as much as possible.
Y = α · Sin (X + β) + γ (1)

  Furthermore, the constants A, B, and C when three of the extracted detection data are variously changed are obtained, and the average values of the obtained constants α, β, and γ are used as the coefficients of the equation (1). Determine the arithmetic expression. That is, when 30 pieces of detection data are extracted, 10 times, and when only 24 pieces of detection data can be obtained, the average value of constants A, B, and C is obtained for 8 times. Is an arithmetic expression of an eccentric sine wave waveform. Note that the method for obtaining the arithmetic expression of the eccentric sine wave waveform is an example of the present embodiment, and the arithmetic expression may be obtained using another derivation method.

  FIG. 7 is a graph showing the sine wave data d. The control unit 36 assigns each encoder value e to the calculation formula of the eccentric sine wave waveform, and calculates the sine wave data d associated with each encoder value e. When the sine wave data d is calculated, the process proceeds to step a3.

  For example, when the arithmetic expression of the sine wave waveform is expressed by the above-described expression (1), the sine wave data di associated with the encoder value ei of interest is expressed by α · Sin (ei + β) + γ.

  In step a3, the control unit 36 subtracts (c−d) the sine wave data d shown in step a2 from the detection data c shown in step a1 for each of the associated encoder values e, thereby removing the influence of eccentricity. Data f is calculated. FIG. 8 is a graph showing the difference data f.

  In step a4, differential data g obtained by differentiating difference data f is calculated. Specifically, the differential data g is, for each encoder value e of interest in the difference data f, the difference data fi corresponding to the encoder value ei of interest and the encoder value e (i + 1) next to the encoder value ei of interest. This data indicates the difference from the corresponding difference data f (i + 1), and is associated with each encoder value e. When the differential data g is calculated, the process proceeds to step a5.

  For example, if the difference data corresponding to the encoder value ei of interest is fi and the difference data corresponding to the encoder value e (i + 1) next to the encoder value ei of interest is f (i + 1), the difference data is associated with the encoder value ei of interest. The differential data g is f (i + 1) −fi.

  FIG. 9 is a graph showing a defect cutout by enlarging a part of the difference data f. In step a5, the control unit 36 searches for a peripheral portion where an arc equal to the radius 85 of the wafer 21 is formed based on the difference data f, and obtains a starting point from which data detection is started. Difference data f during the displacement of the wafer 21 by 360 ° from this starting point is used.

  Specifically, the encoder value change interval in which the encoder value e increases over the predetermined second threshold value P2 in the state where the differential data g is small and the difference data f continues the predetermined first threshold value P1. Then, the smallest encoder value among the change intervals of the encoder value e is detected as the starting encoder value ea. The distance of the peripheral straight line 40 on the line at the angular displacement position of the starting encoder value ea is substantially equal to the radius 85 of the wafer 21. When the starting encoder value ea is detected, the process proceeds to step a6.

  In step a6, when the encoder value e increases from the starting encoder value ea, if the difference data f is equal to or less than the predetermined third threshold value P3, the control unit 36 sets the encoder value eb at that time to the above-described encoder value eb. A single angular displacement position 42 is determined, and it is determined that a notch is formed from the first angular displacement position 42. When the first angular displacement position 42 is detected, the process proceeds to step a7.

  In step a7, when the encoder value e increases from the first angular displacement position 42, the control unit 36 determines that the difference data f is less than or equal to a predetermined fourth threshold value P4, the differential data g is small, and the difference data f In the state where the encoder value e increases over the predetermined sixth threshold value P6, and the starting point of the change interval of the encoder value e is detected. Is detected as a return encoder value ec. The distance of the peripheral straight line 40 on the line at the angular displacement position of the return encoder value ec is substantially equal to the radius 85 of the wafer 21. When the return encoder value ec is detected, the process proceeds to step a8.

  In step a8, when the difference data f becomes equal to or smaller than a predetermined seventh threshold value P7 when the encoder value e decreases from the return encoder value ec, the control unit 38 determines that the encoder value at that time is Let ed be the third angular displacement position 43 described above, and proceed to step a9.

  In step a9, the width of the notch is determined based on the encoder value e and the difference data value f between the first angular displacement position 42 obtained in step a6 and the third angular displacement position 43 obtained in step a8. W, depth H, and notch position 45 are obtained, and the process proceeds to step a10.

  In step a10, the control unit 36 determines whether or not the next first angular displacement position 42 exists when the encoder value is increased from the return encoder value ec based on the difference data f. If it is determined that the single-angle displacement position 42 exists, the process returns to step a7. For the next notch, the notch width W, depth H, and notch position 45 are determined. In Step a10, even if the 360 ° encoder value e is increased from the starting encoder value ea, if the next first angular displacement position 42 does not exist, the process proceeds to Step a11.

  In step a 11, one notch portion having a width and a depth within a specified range of the notch portion 80 is detected from the detected one or more notch portions, and the notch portion is determined as the notch portion 80. . Other notch portions are detected as defect notch portions 82.

  When there are a plurality of notch portions having a width and a depth within the prescribed range of the notch portion 80, the notch portion closest to the optimum width and depth preset in the notch portion 80 is the notch portion. You may judge as 80.

  In this way, when the control unit 36 detects one notch portion 80 from one or a plurality of notch portions, it detects the remaining notch portion as the defect notch portion 82. Then, the process proceeds to step a12, and the notch detection operation is terminated.

  When the wafer shape detection device 20 is an aligner, the notch portion 80 is angularly displaced to a predetermined angular position. Then, the hand robot that transfers the wafer is brought into a state in which the wafer 21 can be transferred. In addition, when the wafer shape detection device 20 detects the defect notch 82, the wafer shape detection device 20 notifies that the wafer 21 has a defect.

  As described above, the wafer shape detection device 20 detects the width and depth of the notch portion by executing the first defect detection method described above. When at least one of the width and the depth of the notch portion is outside the specified range set in the notch portion 80, the notch portion is detected as a defect notch portion. In other words, the notch portion is detected as the notch portion 80 only when both the width and depth of the notch portion are within the specified range. Even if a defect such as a scratch is formed on the wafer 21, the notch portion 80 can be detected accurately.

  Thereby, the defect notch portion 82 can be detected with high accuracy without erroneously detecting the defect notch portion 82 and the notch portion 80 formed on the wafer. By accurately detecting the defect notch portion 82, the wafer 21 on which the defect notch portion 82 is formed can be removed during the semiconductor manufacturing process. By removing the wafer 82 in which the defect has occurred thereby at an early stage, it is possible to prevent the wafer having the defect from being subjected to semiconductor processing.

  In the present embodiment, wafer shape detection apparatus 20 detects defect notch portion 82 simultaneously with detection of notch portion 80. As a result, the process for detecting the defect notch 82 need not be a separate process, and the time spent for detection can be shortened. In other words, it is possible to detect a defect in the peripheral portion of the wafer without affecting the tact time of the apparatus.

  In the present embodiment, the width and depth of the notch are detected based on the difference data f obtained by subtracting the sine wave data d from the detection data c. Therefore, even if the center position 28 of the wafer 21 is eccentric with respect to the rotation axis 29, the influence of the eccentricity can be eliminated and the width and depth of the notched portion can be detected. As a result, the width and depth of the notch portion can be accurately obtained, and therefore the notch portion 80 and the defect notch portion 82 can be detected with higher accuracy.

  Further, by obtaining an arithmetic expression of the eccentric sine wave waveform, the degree of eccentricity of the wafer 21 can be obtained from the amplitude and phase. For example, by transmitting the degree of eccentricity to the robot hand, the robot hand can grasp the degree of eccentricity of the wafer, and the wafer 21 held by the wafer shape detection device 20 can be accurately held. Further, as described above, even when the wafer 21 is held in the holding portion 31 in an eccentric state, the notch portion 80 and the defect notch portion 82 can be accurately detected, and the robot hand position teaching operation can be performed. Such trouble can be reduced.

  Further, according to the present embodiment, the defect of the wafer 21 is detected together with the adjustment of the arrangement position of the wafer 21 by changing the calculation program of the so-called aligner device that arranges the arrangement position of the wafer 21 based on the notch portion 80. Can do. Therefore, it is not necessary to add a new physical configuration, and a wafer defect detection apparatus can be easily realized.

  In the present embodiment, based on the differential data g, the peripheral edge 23 where the notch is not formed is detected. As a result, it is possible to accurately detect the peripheral portion 23 where the notch portion is not formed, and to detect the notch portion 80 and the defect notch portion 82 more accurately than based on the difference data f. Further, detection of the notch portion 80 and the defect notch portion 82 is started from the starting encoder value ea where no notch is formed in the peripheral portion. This prevents the detection of the cutout part from the middle of the cutout part, and the cutout part can be detected more accurately.

  10 and 11 are graphs for explaining the defect detection method for a second wafer of the present invention. FIG. 10 is a graph showing a notch portion 80 by enlarging a part of the difference data f. FIG. 11 is a graph showing a defect notch 82 by enlarging a part of the difference data f.

  In the notch portion 80, on the graph, the difference data f that changes as the encoder value e increases from the first angular displacement position 42 to the second angular displacement position 44 is the encoder value e and the vertical axis is the difference data f. Plot to. In this case, the plotted difference data f is arranged along the prescribed curve 50 drawn on the graph. The specified curve 50 is represented by a predetermined specified curve formula.

In the present embodiment, the prescribed curve 50 is represented by a quadratic curve formula of A · e ² + B · e + C. A is a coefficient of the square of the encoder value e, and is a shape specifying coefficient A representing the shape of the prescribed curve 50. B and C differ depending on the arrangement position of the wafer 21. In the above equation, e represents an encoder value.

  On the other hand, in the defect notch portion 82, the plotted difference data f is arranged along a non-specified curve 51 having a shape different from that of the specified curve 50. Further, in the defect notch portion 82, the plotted difference data f varies greatly with respect to the prescribed curve 50. Moreover, the approximate curve equation in which the plotted difference data f is arranged can be easily derived by using an approximation method such as a least square method.

  If the shape specifying coefficient A of the notch portion of interest is outside the specified range of the shape specifying coefficient A representing the notch portion 80, the notch portion of interest can be detected as the defect notch portion 82.

  The shape specifying coefficient A is a determination value for determining whether or not the curve equation representing the notch portion is similar to the prescribed curve equation representing the notch portion. That is, by extracting a coefficient indicating the shape of the notched portion from the approximate curve formula, and determining the similarity of the coefficient with respect to the coefficient indicating the shape of the specified notch portion of the specified curve formula, the curve formula is similar. Determine whether or not. For example, the degree of similarity is determined by a difference between a coefficient indicating the shape of the specified notch portion and a coefficient indicating the shape of the notch portion of interest.

  FIG. 12 is a graph showing a part of the graph of FIG. 10 in an enlarged manner. Each plot point in the notch portion 80 falls within a predetermined range in which an error average value from the predetermined curve 50 is predetermined. The error average value is a value obtained by adding all the differences between the difference data f of each plot point 60 and the difference data f of the specified curve 50 at the encoder value e set at each plot point 60.

  When the difference data f at an arbitrary plot point 60 is xi and the difference data f at the encoder value e set at the arbitrary plot point 60 is yi, the error average value V is expressed by the following equation.

Here, xi represents i-th difference data among n plot points from the first angular displacement position 43 to the second angular displacement position 44. Yi indicates difference data of the prescribed curve 50 at the same encoder value e as the i-th plot point.

  When the error average value V at the notched portion to be noticed is outside the predetermined range, the notched portion to be noticed can be detected as the defect notched portion 82.

  The error average value V is also a determination value for determining whether or not the curve equation representing the notch portion is similar to the prescribed curve equation representing the notch portion. It shows that the shape of the notch part to which the attention is paid is similar to the shape of a notch part, so that the average error value V with respect to a regulation curve is small.

  FIG. 13 is a flowchart showing the second notch detection operation of the control unit 36. In step b0, the control unit 36 is given a specified range of the shape specifying coefficient A at the notch portion 80. The control unit 36 stores the given specified range in the storage circuit. Then, the control unit 36 rotates the wafer 21 in a state where the wafer 21 is sucked and held by the holding unit 31, and proceeds to step b1.

  In step b1, the controller 36 performs an operation corresponding to step a1 to step a10 shown in FIG. However, instead of the controller 36 calculating the depth and width of the notch portion in step a9, an approximate curve equation of the difference data value f between the first angular displacement position 43 and the second angular displacement position 44 is obtained. Is calculated.

  When the approximate curve equation of each notch is obtained in step b1, the process proceeds to step b2. In step b2, one or more notch portions whose shape specifying coefficient A is within a specified range are detected from the detected approximate curve equations of one or more notch portions, and the process proceeds to step b3.

  In step b3, an error average value V with respect to the prescribed curve is obtained for one or more notch portions detected in step b2, and the process proceeds to step b4. In step b4, a notch whose average error value with respect to the prescribed curve is within a prescribed range for the one or more notches detected in step b2 is detected as a notch portion 80.

  In step b4, if there are a plurality of cutout portions where the error average value V is within the specified range, the error average value is the largest of the plurality of cutout portions where the error average value is within the specified range. A small one is detected as a notch portion 80. In this way, when the control unit 36 detects one notch portion 80 from one or a plurality of cutout portions, the control portion 36 detects the remaining cutout portion as a defect cutout portion. Then, the process proceeds to step b5, and the notch detection operation is terminated.

  As described above, when the wafer shape detection apparatus 20 executes the second defect detection method described above, the wafer shape detection apparatus 20 calculates a curve equation at the notch portion, and based on the curve equation, the cut shape is calculated. It is determined whether the notch portion is the notch portion 80 or the defect notch portion 82.

  Specifically, the notch portion is determined based on the shape of the defect notch portion by detecting the shape specifying coefficient A and the error average value V at the notch portion. As a result, even if it is a defect notch part included in the prescribed range of the width and depth of the notch part 80, the defect notch part 82 and the notch part 80 can be distinguished and detected, thereby preventing erroneous detection. Can do. Moreover, the same effect as that of the first defect detection method can be obtained for the step of performing the same procedure as that of the first defect detection method.

  FIG. 14 is a flowchart showing the third defect detection operation of the control unit 36. In the third defect detection operation, the defect notch portion and the notch portion of the wafer are detected using the depth, width, shape specifying coefficient A of the notch portion, and the average error value with the specified curve. Note that the notch width, depth, the shape specifying coefficient A, and the error average value V with respect to the specified curve can be obtained by the same procedure as described above, and thus the description thereof is omitted.

  In step c0, the control unit 36 stores an allowable range of the width and depth of the notch portion that should be ignored as a defect. Further, the control unit 36 stores the width and depth of the notch portion 80 to be detected and the specified range of the shape specifying coefficient A. In a state where the preparatory operation is completed in this way, the control unit 36 rotates the wafer 21 held by the holding unit 31, and proceeds to step c1.

  In step c1, the control unit 36 acquires the width and depth of the notch portion as in the first defect detection operation, and proceeds to step c2 when acquired. In step c2, the control unit 36 determines whether or not the acquired width and depth of the cutout portion are within an allowable range that should be ignored as a defect. If it is determined that the width and depth of the acquired notch are not within the allowable range, the process proceeds to step c3. In step c3, it is determined whether or not the width and depth of the notch portion acquired in step c2 are within a specified range defined as the notch portion 80. If the acquired width and depth of the notch are within the specified range, the process proceeds to step c4.

  In step c4, it is determined whether or not the acquired shape of the notch portion corresponds to the shape of the notch portion 80. Specifically, it is determined whether or not the shape specifying coefficient A and the error average value V of the notch portion described above are included in the specified range. If it is determined that they are within the specified range, the process proceeds to step c5. In step c5, the notch portion determined to be within the specified range of the notch portion 80 is counted as the notch portion 80, information such as the notch position is stored, and the process proceeds to step c7.

  In step c2, if the width and depth of the acquired notch portion are within an allowable range that should be ignored as a defect, the control unit 36 proceeds to step c7 without counting the notch portion. In step c3 and step c4, when the acquired notch part is out of the specified range of the notch part 80, the process proceeds to step c6. In step c6, the control unit 36 counts the notch portion determined to be outside the specified range of the notch portion 80 as the defect notch portion 82, stores information such as the notch position, and proceeds to step c7.

  In step c7, it is determined whether data for one round of the wafer has been retrieved. If the data for one round of the wafer is not retrieved, the process returns to step c1 to detect whether there is another notch portion. If it is determined in step c7 that data for one round of the wafer has been retrieved, the process proceeds to step c9, and the defect detection operation is terminated.

  In step c1, the control unit 36 proceeds to step c8 when it cannot acquire the data of the width and depth of the notch portion, similarly to the first defect detection operation. In step c8, it is determined whether data for one round of the wafer has been retrieved. If the data for one round of the wafer is not retrieved, the process returns to step c1 to detect whether there is another notch portion. If it is determined in step c8 that data for one round of the wafer has been detected, the process proceeds to step c9 and the defect detection operation is terminated.

  In the third defect detection method as described above, the effects of both the first defect detection method and the second defect detection method can be obtained. As a result, even if a defect notch portion 82 occurs that falls within the prescribed range of the width and depth of the notch portion 80, it can be detected as the defect notch portion 82. Thereby, erroneous detection of the notch portion 80 can be prevented.

  Further, in step c4, even when the shape specifying coefficient A of a plurality of notched portions to be noticed is included in the specified range, the notched portion that becomes the curve equation closest to the specified curve equation, that is, the error average value A Is detected as a notch portion 80. As a result, the notch can be detected with higher accuracy.

  In step c2, when the depth and width of the notch of interest are within the allowable range, the notch of interest is ignored. This makes it possible to accurately detect the defect notch portion 82 of interest. For example, it is possible to prevent erroneous detection as a defect even if the depth and width of the notch part are small and have no problem in wafer quality. By registering the depth and width of the unevenness generated in the wafer manufacturing process as an allowable range, it is possible not to detect the uneven part generated in the manufacturing process as a defect.

  In addition, the shape specifying coefficient A and the error average value V are obtained only for the notch portions whose width and depth are within the specified ranges. Accordingly, the calculation time can be shortened as compared with the case where the shape specifying coefficient A and the error average value V are obtained for all the notched portions.

  In the above-described embodiment, the case where the notch 22 is used as the predefined notch is described. However, for example, an orientation flat (hereinafter referred to as an orientation flat) other than the notch 22 is used as the defined notch. However, the same effect can be obtained.

  FIG. 15 is a front view showing the semiconductor wafer 21 on which the orientation flat 24 is formed. In another embodiment, the defined notch portion is realized by an orientation flat portion 73 in which the orientation flat 24 is formed. The orientation flat 24 is a recess formed in the peripheral edge 23 of the semiconductor wafer 21. The orientation flat portion 73 is formed with a plane 70 perpendicular to a radius line 71 extending outward from the center position 28 of the wafer 21 when the wafer 21 is viewed in a cross section perpendicular to the thickness direction. The detailed dimensions of the orientation flat 24 are standardized in advance by SEMI or the like.

  FIG. 16 is a graph showing a part of the difference data f in the orientation flat 24. Even in the orientation flat 24, the width W, the depth H, and the approximate curve 72 are defined in the same manner as the notch 22. Therefore, it can be determined that the notched portion is the orientation flat 24 when the width W, the depth H, the shape specifying coefficient A of the approximate curve equation, and the error average value V are within the specified range set in the orientation flat 24.

  FIG. 17 is an enlarged graph showing a portion where the difference data f in the orientation flat 24 is minimized. When the detection accuracy of the line sensor 34 is low, the encoder value e having the smallest difference data f covers a predetermined range 73 as shown by the plot points shown in FIG. Therefore, it may be difficult to determine one angular position of the orientation flat portion 73 determined based on the encoder value e with the smallest difference data f.

  On the other hand, by obtaining the approximate curve 72 determined based on the difference data f, the encoder value e that minimizes the difference data f can be easily obtained. The position, in other words, the notch angle displacement position 45 can be determined. As a result, the angular position of the orientation flat portion 73 can be detected with high accuracy. Thereby, the wafer shape detection apparatus 20 can improve the performance as an aligner. The same applies to the case of the notch portion 80 described above.

  Further, even when a chip is formed in the orientation flat portion 73, a defect notch portion formed in the orientation flat portion 73 can be detected by taking the difference between the approximate curve 72 and the difference data f. In this case, the notch portion formed in the orientation flat portion 73 becomes the difference data f having a large error from the orientation curve 72 of the orientation flat. Therefore, it is possible to determine a defect notch portion formed in the orientation flat portion 73 from the difference data f having a large error. Except for the difference data f having a large error, the approximate curve equation of the notched portion may be obtained again and compared with the approximate curve 72 of the orientation flat portion.

  In addition, as the orientation flat 24, two orientation flats of a main orientation flat and a sub orientation flat may be formed. In such a case, the control unit 36 stores the specified ranges of the main orientation flat and the secondary orientation flat so that the main orientation flat portion, the secondary orientation flat portion, and the defect notch portion can be distinguished and detected. it can.

  The specified range stored by the control unit 36 may store both the specified range set for the notch 22 and the specified range set for the orientation flat 24. This makes it possible to use the present detection device regardless of whether the prescribed notch portion formed in the wafer 21 is the notch portion 80 or the orientation flat portion 73, and the convenience can be improved. For example, the size of the wafer 21 is determined, and it is determined which of the specified ranges of the notch portion 80 and the orientation flat portion 73 is used to distinguish the specified notch portion, and the specified range is determined based on the determined specified range. You may detect a notch part. As described above, the specified range defined as the specified notch portion and the allowable range ignored as the notch can be arbitrarily input numerical values, and it is preferable that a plurality can be set.

  The specified range may be specified by measuring an actual notch portion or orientation flat portion. Further, it may be determined based on a predetermined standard. The wafer shape detection apparatus 20 is preferably provided with setting means that allows the user to input and change the specified range stored in the control unit 36. For example, the setting means is realized by a keyboard or the like. By changing the specified range for each semiconductor wafer to be measured by the setting means, it is possible to further prevent erroneous detection of the notch portion.

  Further, in the present embodiment, based on the width and depth of the notch portion, the notch portion may be detected based on the area of the region notched by the notch portion. The area of the notched region can be obtained, for example, by adding up the width and depth of the notched portion.

  In the present embodiment, the wafer 21 is continuously rotated with respect to the line sensor 34. On the other hand, the line sensor 34 may be continuously rotated with respect to the wafer 21. Also, an operator may perform part or all of the operation of the wafer shape detection device 20.

  In this case, the operator measures the distance change of the peripheral straight line 40 when the peripheral straight line extending from the axis set in advance to the wafer 21 to the peripheral edge of the wafer 21 is angularly displaced around the axis, and the measurement result is Based on the width representing the amount of angular displacement that the peripheral straight line passes through the notch part and the depth related to the peripheral straight line where the distance of the peripheral straight line is the smallest in the notch part. When at least one of the width and the depth is outside the specified range set in the specified notch portion, the notch portion of interest is detected as the defect notch portion.

It is a block diagram which shows the structure of the wafer shape detection apparatus 20 which is one Embodiment of this invention. It is a figure which expands and shows a part of wafer 21 for demonstrating the notch 22. FIG. It is a front view which shows the notch wafer 21 in which the three notches 25, 26, and 27 generate | occur | produce. It is a graph which shows the encoder value e at the time of carrying out the angular displacement of the notch wafer 21 shown in FIG. 3, and the sensor value c corresponding to the encoder value e. 4 is a flowchart showing a first notch detection operation of the control unit 36. It is the graph which put sensor value c and encoder value e acquired at Step a1 in order of encoder value e. It is a graph which shows the sine wave data d. It is a graph which shows difference data f. It is a graph which expands a part of difference data f and shows a defect notch. 5 is a graph showing a notch portion 80 by enlarging a part of difference data f. It is a graph which expands a part of difference data f, and shows the defect notch part 82. FIG. It is a graph which expands and shows a part of graph of FIG. 7 is a flowchart showing a second defect notch detection operation of the control unit 36. 7 is a flowchart showing a third defect detection operation of the control unit 36. It is a front view which shows the semiconductor wafer 21 in which the orientation flat 24 is formed. 4 is a graph showing a part of difference data f in the orientation flat 24. 4 is an enlarged graph showing a portion where differential data f in the orientation flat 24 is minimized.

Explanation of symbols

DESCRIPTION OF SYMBOLS 20 Wafer shape detection apparatus 21 Semiconductor wafer 22 Notch 23 Peripheral part 24 Orientation flats 25, 26, 27 Notch 28 Center axis 29 Center position 30 Base 31 Holding part 32 Motor 33 Encoder 34 Line sensor 35 Sensor controller 36 Control part 40 Peripheral straight line 41 Reference position 50 Specified approximate curve W Width H Depth

Claims (9)

In a disk-shaped semiconductor wafer in which a predetermined notch portion defined in advance in the peripheral portion is formed, a wafer defect detection method for detecting a defect notch portion generated in the peripheral portion,
When the peripheral straight line extending from the axis set in advance to the wafer to the peripheral edge of the wafer is angularly displaced around the axis, the distance change of the peripheral straight line is measured.
Based on the measurement result, a notch to which attention is paid is a width representing an angular displacement amount at which the peripheral straight line passes through the notch portion, and a depth related to the peripheral straight line at which the distance of the peripheral straight line becomes the smallest in the notch portion. Calculate for each part,
When at least one of the width and depth of the notch portion to be noticed is outside the specified range set in the notch portion, the notch portion to be noticed is detected as a defect notch portion ,
When the width and depth of the notch part of interest are within the specified range set for the notch part, calculate a curve formula representing the distance change of the peripheral straight line in the notch part of interest,
If the curve equation calculated for the notched portion of interest falls within a predetermined similar range to the prescribed curve equation representing the change in the distance of the peripheral straight line at the prescribed notch portion, the notched portion of interest is defined as the prescribed notch portion. A method for detecting a defect of a wafer, wherein the defect is detected as a chipped portion .   2. The wafer defect detection method according to claim 1, wherein an angular displacement position around the axis of the wafer with respect to an arbitrary reference position set for the wafer is obtained for the defect notch portion based on the measurement result. For provisions notch portion, a defect detection method according to claim 1 or 2, wherein the wafer and obtaining the angular displacement position about the axis of the wafer relative to an arbitrary reference position set in the wafer. When the curve formula calculated for a plurality of notched portions to be noticed is within a predetermined similarity range with respect to the specified curve formula, the cutout portion that is closest to the specified curve formula is defined as the specified notch portion. The defect detection method for a wafer according to any one of claims 1 to 3, wherein: In the distance change of the peripheral linear, either peripheral straight line on the basis of the calculation result of removing the frequency component equal to the period of around an axis, according to claim 1-4, characterized in that to determine the width and depth of the notch portion The wafer defect detection method according to claim 1. The width of the notch portion of interest and the depth, if within the allowable range specified in advance, the wafer according to any one of claims 1-5, characterized in that ignore the notch portion of the attention Defect detection method. The defining notch portion, a defect detection method for a wafer according to any one of claims 1-6, characterized in that either part one is formed of a notch and an orientation flat. In a disk-shaped semiconductor wafer in which a predetermined notch portion that is predetermined in the peripheral edge portion is formed, a wafer defect detection device that detects a defect notch portion generated in the peripheral edge portion,
The base,
A holding unit for holding the wafer;
Rotation driving means for rotating the holding unit around a central axis set on the base;
Angle measuring means for measuring the amount of angular displacement of the holding part;
A peripheral position measurement unit that measures a position in the wafer radial direction of the peripheral portion of the wafer, facing a part in the peripheral direction of the peripheral portion of the wafer held by the holding unit;
Based on the measurement results of the angle measurement means and the peripheral position measurement means, and having a defect detection means for detecting a defect notch portion,
Defect detection means
When the wafer is angularly displaced and the position change in the wafer radial direction of the wafer peripheral portion becomes larger than a predetermined threshold value, a notch portion is formed in the peripheral portion of the wafer facing the peripheral position measuring means. Detect
The depth associated with the width representing the angular displacement of the wafer while the notch passes through the peripheral position measuring means, and the depth related to the distance of the peripheral straight line where the distance from the central axis to the wafer peripheral edge becomes the smallest at the notch. For each notch to be noticed,
When at least one of the width and depth of the notch portion to be noticed is outside the specified range set in the notch portion, the notch portion to be noticed is detected as a defect notch portion ,
When the width and depth of the notch part of interest are within the specified range set for the notch part, calculate a curve formula representing the distance change of the peripheral straight line in the notch part of interest,
If the curve equation calculated for the notched portion of interest falls within a predetermined similar range to the prescribed curve equation representing the change in the distance of the peripheral straight line at the prescribed notch portion, the notched portion of interest is defined as the prescribed notch portion. A defect detection apparatus for a wafer, wherein the defect detection device detects the defect portion . The defect detecting means, on the basis of the measurement result, the portion-out defects notches, of the wafer of claim 8, wherein the determination of the angular displacement position about the axis of the wafer relative to an arbitrary reference position set in the wafer Defect detection device.

Images