JPH04249340A - Alignment method of semiconductor wafer - Google Patents

Abstract

PURPOSE:To perform an alignment operation with high accuracy by detecting the accurate position of an orientation flat. CONSTITUTION:The position of an edge at a semiconductor wafer 1 is measured and the position of an orientation flat is detected. It is examined whether the semiconductor wafer 1 is to be aligned first or not. When it is to be aligned first, the orientation-flat part is made parallel to the traveling direction of a Y-stage 5 for orientation-flat detection use. After that, an error in the mounting angle of a line sensor 7 is measured and the semiconductor wafer 1 is aligned. When the semiconductor wafer 1 is not to be aligned first, the orientation-flat part is made parallel to the line sensor 7. The semiconductor wafer 1 is turned by a portion of said error in the mounting angle; after that, said semiconductor wafer 1 is aligned.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for aligning semiconductor wafers used in semiconductor manufacturing equipment.

0002

2. Description of the Related Art As a method for aligning a semiconductor wafer used in a highly integrated LSI manufacturing apparatus, the position of the outer edge of the semiconductor wafer is detected in a non-contact manner, and the result is used to determine the orientation flat portion (hereinafter referred to as " A method has been proposed in which a semiconductor wafer is aligned in a predetermined direction and position based on the position of the orientation flat.

The basic configuration for carrying out the above-mentioned semiconductor wafer alignment method is as shown in FIG.
While the orientation flat detection chuck 2 holding the semiconductor wafer 1 by suction is rotated by the orientation flat detection θ stage 3, the light emitting diode 6, which is a non-contact outer edge position detection means, is rotated.
The edge position for one circumference of the outer edge of the semiconductor wafer 1 is measured using a signal processing unit 9 using a line sensor 7 consisting of a photoelectric conversion element, etc., and the orientation flat position is determined from the edge position measured by an arithmetic control unit 13. . The line sensor 7 is installed parallel to the moving direction of the X stage 4 for orientation flat detection. Then, based on the obtained orientation flat position, corrected drive amounts for each of the X, Y, and θ axes are calculated to move the semiconductor wafer 1 to a predetermined position, and according to the results, the orientation flat detection The semiconductor wafer 1 is positioned by driving the X stage 4, the Y stage 5 for orientation flat detection, and the θ stage 3 for orientation flat detection.

0004

However, the above-mentioned conventional techniques do not take into consideration the case where the line sensor deviates from a predetermined position during assembly. Current line sensor assembly technology allows positioning with an accuracy of several μm in the X and Y directions by using a high-precision jig such as an optical reticle, but in the θ direction It is very difficult to achieve an accuracy of 10-3 rad or more.

The error that occurs when the edge position of the semiconductor wafer 1 is measured using a line sensor assembled in an inclined state will be explained with reference to FIGS. 12 and 13.

FIG. 12 shows an x-
This shows the semiconductor wafer 1 in the y-coordinate system, and the line sensor 7 is assembled in an inclined state by Δθx with respect to the x-axis. The assembly in the X and Y directions is performed at the center of the line sensor 7, and the accuracy thereof is assumed to be sufficiently high. FIG. 13 shows the relationship between the line sensor 7 in FIG. 12 and the outer edge of the wafer passing over the line sensor 7.
e) shows a process in which the portion of the orientation flat portion 102 passes over the line sensor 7 when the semiconductor wafer 1 is rotated in the rotation direction (direction of arrow θ in the figure) in FIG.

In FIG. 13(a), the vicinity of the boundary PA between the orientation flat portion 102 and the circular arc portion 101 is measured. At this time, originally the outer edge on the x-axis should be measured, but what is actually measured is the outer edge on the line sensor 7, so the distance from the center of the chuck is measured longer by the error Δl. . FIG. 13(b) shows a state further rotated in the θ direction from this state. Since the orientation flat portion 102 becomes closer to perpendicular to the x-axis, the error Δl decreases. FIG. 13C shows a state in which the orientation flat portion 102 is perpendicular to the x-axis after further rotation in the θ direction. In this state, the measurement error is purely due to the inclination of the line sensor 7, and if the length along the line sensor 7 from the center P of the line sensor 7 to the measurement point is x, then in this state The error Δl is expressed as Δl=x*(1-cos Δθx), and as mentioned above, the orientation flat part 10 is
This is an order of magnitude that can be ignored compared to the error when 2 is tilted. In the state of FIG. 13(d), which is further rotated in the θ direction, the state is opposite to FIG. 13(b), and the outer edge on the line sensor 7 is shorter by the error Δl than the outer edge on the x-axis to be measured. It will be measured. In FIG. 13(e), FIG.
The vicinity of the boundary PB with the circular arc portion 101 is measured in the reverse state of (a), and the error Δl increases compared to (d) of FIG.

Although not shown, when the arcuate portion 101 is on the line sensor 7, it is approximately equal to (c) in the image 13;
There is almost no effect. Therefore, (a) to (e) in FIG.
The edge position of the orientation flat portion 101 measured in the state shown in FIG. 12 becomes an S-shaped measurement value as shown by the broken line in FIG. 12, and as a result, the orientation flat detection θ accuracy deteriorates. However, the S-shaped error curve shown in FIG. 12 is an example, and the orientation flat portion 101
A different curve will be obtained depending on the position on the line sensor 7 at which the measurement is performed. Since this is related to the eccentricity between the center of the chuck and the center of the semiconductor wafer 1, a different amount of orientation flat detection θ error occurs for each semiconductor wafer.

The present invention has been made in view of the problems of the conventional techniques described above, and provides a semiconductor wafer alignment method that enables highly accurate alignment by detecting the accurate position of the orientation flat. is intended to provide.

0010

[Means for Solving the Problems] The present invention uses non-contact outer edge position detection means for a light emitting means and a line sensor serving as a light receiving means for the light emitting means while rotating a semiconductor wafer for semiconductor manufacturing by a rotating means. The edge position of the semiconductor wafer is measured for one circumference of the outer edge of the semiconductor wafer, the orientation flat part of the semiconductor wafer is detected from the change in the measured edge position, and the orientation flat position indicating the orientation of the orientation flat part is determined. moving the orientation flat part in a specified direction with reference to the direction of movement in two directions: a first movement direction parallel to the orientation of the line sensor and a second movement direction perpendicular to the first movement direction. In the method for aligning semiconductor wafers in which the semiconductor wafer is moved to a designated position using means, after determining the orientation flat position, the angle between the orientation flat portion and the first moving direction is determined based on the orientation flat position. determining the difference, and driving the rotating means according to the determined angular difference to make the orientation flat section and the first movement direction approximately parallel, and the center of the orientation flat section approximately coincides with the center of the line sensor. While scanning the semiconductor wafer in the second moving direction, an angular difference between the line sensor and the orientation flat portion is determined, and the rotating means is rotated according to the determined angular difference. Drive to make the orientation flat part and the line sensor parallel, and measure in advance.
Driving the rotation means according to an assembly angle error of the line sensor with respect to the first movement direction to make the orientation flat part parallel to the first movement direction, and making the orientation flat part parallel to the first movement direction and the designated direction. The rotation means is driven according to the angular difference between the orientation flat portion and the orientation flat portion in a specified direction.

Furthermore, when aligning the first semiconductor wafer, after determining the orientation flat position of the semiconductor wafer, the angular difference between the orientation flat portion and the first movement direction is determined based on the orientation flat position. , driving a rotating means according to the determined angular difference to make the orientation flat portion and the second moving direction substantially parallel;
While scanning the semiconductor wafer in the second movement direction, measure the edge positions of arbitrary two points of the orientation flat part, and from the measured two edge positions, move the orientation flat part in the second movement direction. The rotation means is driven according to the determined inclination to make the orientation flat part parallel to the second movement direction, and then the rotation means is further driven to make the orientation flat part parallel to the second movement direction. While scanning the semiconductor wafer in the second moving direction, An angular difference between the line sensor and the orientation flat part is determined, the determined angular difference is stored as an assembly angle error of the line sensor, and the orientation flat part is determined by the angular difference between the second movement direction and the designated direction. The rotating means may be driven in accordance with the specified direction, and when determining the angular difference between the line sensor and the orientation flat portion, both ends of the line sensor may be rotated while scanning the orientation flat portion. In some cases, the amount of movement of both pixels in the second movement direction from completely transparent to completely blocked is measured, and the angular difference is determined based on the amount of movement.

0012

[Operation] The semiconductor wafer positioning method of the present invention is as follows:
After determining the orientation flat position of the semiconductor wafer, the orientation flat portion is made substantially parallel to the first moving direction of the moving means based on the orientation flat position. In this state,
The orientation flat portion of the semiconductor wafer passes over the line sensor by scanning the semiconductor wafer in the second moving direction of the moving means with the center of the orientation flat portion substantially coinciding with the center of the line sensor. By observing the output signal of the line sensor at that time, it is possible to determine the angular difference between the line sensor and the orientation flat portion. According to this angular difference, the semiconductor wafer is rotated to make the orientation flat part parallel to the line sensor, and then the semiconductor wafer is rotated according to the previously measured assembly angle error of the line sensor with respect to the first movement direction. By rotating the wafer, the orientation flat portion is made parallel to the first movement direction. Since the angular difference between the designated direction of the orientation flat section and the first movement direction is predetermined, the orientation flat section is moved in the first direction as described above.
By rotating the semiconductor wafer in accordance with the angular difference from the specified direction, the orientation flat portion can be oriented in the specified direction.

[0013]

Embodiments Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a diagram showing an orientation flat detection portion of a semiconductor exposure apparatus for implementing the semiconductor wafer alignment method of the present invention.

In this embodiment, the orientation flat detection chuck 2 that holds the semiconductor wafer 1 moves the semiconductor wafer within its plane with the center of the orientation flat detection chuck 2 (hereinafter referred to as "chuck center") as an axis. The orientation flat detection θ stage 3 rotates the semiconductor wafer 1 in the θ direction in the figure.
It is placed on an X stage 4 for orientation flat detection and a Y stage 5 for orientation flat detection, which are moving means for moving in the axis and y axis directions. Also, at the same y-coordinate position as the center of the chuck, there is a light emitting diode (
Hereinafter, it will be referred to as "LED". ) 6 is fixed, and a line sensor 7, which is a light receiving element for the LED 6, is fixed at the same y-coordinate position as the center of the chuck and at a prescribed x-coordinate position, sandwiching the semiconductor wafer 1. ing.

The light beam emitted from the LED 6 is made into parallel light by a lens (not shown) and illuminates the edge portion of the semiconductor wafer 1, and is partially blocked by the semiconductor wafer 1 to form an image on the line sensor 7. .

An output signal corresponding to the amount of light received by the line sensor 7 is taken in by a signal processing section 9. The signal processing section 9 A/D converts the output signal from each pixel of the line sensor 7, and further detects pixels corresponding to edges. Then, the arithmetic processing unit 13 calculates an edge position indicating the edge distance from the chuck center to the outer edge of the semiconductor wafer 1 from the pixel number detected by the signal processing unit 9 and the x-coordinate position of the line sensor 7 with respect to the chuck center. , the orientation flat portion of the semiconductor wafer 1 (hereinafter referred to as “
It is called "Ori-Fura". ) is detected. Further, the signal processing section 9 can take in output signals of desired pixels of the line sensor 7.

The θ stage 3 for orientation flat detection, the X stage 4 for orientation flat detection, and the Y stage 5 for orientation flat detection are operated via a θ-axis drive unit 10, an X-axis drive unit 11, and a Y-axis drive unit 12, respectively. It is driven by the control section 13.

In order to move the semiconductor wafer 1 to a desired position based on the position of the orientation flat detected as described above, the calculation control unit 13 controls the orientation flat detection θ stage 3 and the orientation flat detection X stage 4. and the drive amount of the Y stage 5 for orientation flat detection are determined and driven. Further, the arithmetic control section 13 drives the LED 6 via the LED drive section 8 to turn it on and off at the start and end of orientation flat position detection.

The operation of this embodiment will now be explained with reference to the flowchart shown in FIG.

In this embodiment, the position for transferring the semiconductor wafer 1 to a load hand (not shown) that transports the semiconductor wafer 1 to an exposure wafer stage (not shown) is designated as a designated position, and the semiconductor wafer 1 is transferred to the designated position. Perform alignment.

First, in order to detect the orientation flat position of the semiconductor wafer 1, the edge position of the semiconductor wafer 1 is measured (S201).

In this edge position measurement start state, the semiconductor wafer 1 is attracted to the orientation flat detection chuck 2 with the orientation flat facing in an arbitrary direction, and at this time, the center of the semiconductor wafer 1 and the center of the chuck are aligned. During this period, an eccentricity within a specified value occurs due to an error that occurs when the semiconductor wafer 1 is transported. Further, the orientation flat detection X stage 4 and the orientation flat detection Y stage 5 are in positions for performing measurement.

In this state, the arithmetic control unit 13 lights up the LED 6 via the LED drive unit 8 and drives the θ stage 3 for orientation flat detection via the θ axis drive unit 10 to move the chuck 2 for orientation flat detection. is rotated, and while the orientation flat detection θ stage 3 is rotating at a constant speed, the signal processing unit 9
The edge position of one circumference of the outer edge of the semiconductor wafer 1 is measured at predetermined angular intervals. After the measurement is completed, the LED 6 is turned off via the LED drive unit 8.

The edge position measurement method is to pass the charge accumulated by the amount of light received by the line sensor 7 from the LED 6 within a certain period of time at each measurement point to the signal processing unit 9, and obtain the edge position as an output. . From this edge position output and the mounting coordinate position of the line sensor 7 from the center of the chuck, the distance from the center of the chuck to the outer edge can be determined.

After the measurement of the edge position for one circumference of the outer edge of the semiconductor wafer 1 is completed, the angle rotated by the orientation flat detection θ stage 3 until it stops is stored as the overrun rotation amount β, and is fed back at the time of correction drive.

FIG. 3 shows the locus of change in the edge distance l from the chuck center to the outer edge of the semiconductor wafer 1 with respect to the rotation angle θ of the orientation flat detection θ stage 3.
(a) shows a semiconductor wafer with one orientation flat, (b)
shows a semiconductor wafer on which two orientation flats, first and second, are formed.

Note that when there is no eccentricity between the center of the chuck and the center of the semiconductor wafer 1, the gentle curve corresponding to the arc of the semiconductor wafer 1 in FIGS. 3(a) and 3(b) becomes a straight line.

Next, the orientation flat portion is determined from the edge position measured as described above (S202).

FIG. 4 shows a method for determining the orientation flat portion.
This will be explained according to the flowchart shown in .

Regarding the trajectory of the edge distance l, the data string l is scanned to search for the minimum value (S401), and the rotation angle at which the minimum value is obtained is stored. The rotation angle that takes this minimum value is 2 of θ1a and θ2a in the locus shown in FIG. 3(a).
In the trajectory shown in FIG. 3(b), θ1b, θ2b
, θ3b.

Next, it is checked whether all of the minimum values detected in step S401 satisfy the orientation flat condition. Various methods can be used to set the orientation flat condition, but for example, in an enlarged view near the minimum value as shown in FIG.

At this time, the specified number of measurement points must be definitely points on the orientation flat, and can be determined from the wafer diameter, the length of the orientation flat, and the maximum possibility of eccentricity.

Specifically, if the radius of the semiconductor wafer is r, the length of the orientation flat is s, the maximum value of eccentricity is d, and the number of measurement points at the edge position for one circumference of the outer edge is n, the specified number of measurement points is (1
) is expressed as an integer m obtained by the formula.

[0035]

[Equation 1] In FIG. 5, the specified number of measurement points is 1 including the minimum value.
It is 0 points.

(2) The difference between the value lm of the measurement point separated by the specified number of measurement points from the minimum value and the value lmin of the minimum value is greater than the specified value Δl on each of the left and right sides. The method for determining Δl at this time is to make it larger than the amount of change of m points from the minimum value on the arc at the time of maximum eccentricity, and smaller than the amount of change of m points from the minimum value on the orientation flat at the time of maximum eccentricity. The amount of change Δlm for m points on the orientation flat at the maximum eccentricity is expressed by equation (2), so
It is sufficient to set Δl to a value slightly smaller than Δlm.

[0037]

[Equation 2] Since the minimum value caused by the eccentricity does not satisfy the orientation flat condition ■, in step S402, (a
θ1a shown in ) and θ2b shown in FIG. 3(b) are excluded. Furthermore, the minimum value that occurs when there is a chip or the like in the semiconductor wafer 1 is eliminated by the orientation flat condition (2). Therefore, only the minimum value of the portion corresponding to the orientation flat passes through step S402.

Next, in step S402, the number of local minimum values that satisfy the orientation flat condition is checked (S40
3). As shown in FIG. 3A, when there is only one minimum value that satisfies the orientation flat condition, that portion corresponds to the orientation flat. Therefore, in this case, the rotation angle θ1a is set as the orientation flat position and the process of determining the orientation flat portion is ended. If there are three or more, it is determined that the semiconductor wafer 1 has cracks, chips, etc. that cannot be distinguished from orientation flats, and the process ends with an error. Also,
As shown in FIG. 3(b), in the case of two θ1b and θ3b, it is checked based on the ID information of the semiconductor wafer 1 whether the semiconductor wafer 1 currently undergoing orientation flat detection has a second orientation flat ( S404). If a semiconductor wafer without a second orientation flat is being measured, step S403
It ends with an error for the same reason as the error branch from . When measuring a semiconductor wafer 1 having a second orientation flat, since the angle between the first orientation flat and the second orientation flat is known in advance, the second orientation flat of the semiconductor wafer as seen from the first orientation flat is A first orientation flat is determined from the angle (S405).

In this embodiment, in the x-y coordinate system with the chuck center as the origin shown in FIG.
The angle of the second orientation flat 62 seen from 1 is 270°,
If the orientation flat detection θ stage 3 is rotated in the positive direction of the θ axis during measurement, a second orientation flat can be measured on the line sensor 7 at an angle of 90° after the first orientation flat has been measured. Therefore, in FIG. 3(b), the portion θ1b is determined to be the first orientation flat.

If the eccentricity between the center of the chuck and the center of the semiconductor wafer 1 is large, the locus of change in the edge distance l may move beyond the measurement range of the line sensor 7 when the edge position is measured. It will be done. In that case,
Although accurate wafer outer edge position information cannot be obtained, in the present invention, the wafer edge position information for one circumference of the outer edge of the semiconductor wafer 1 is used only for determining the approximate position of the orientation flat, so it is possible to recognize the orientation flat even in such a case. If this is possible, even if the eccentricity is large, there will be no effect on the orientation flat detection accuracy.

Furthermore, for the same reason, the orientation flat detection accuracy can be maintained even if the orientation flat detection speed is improved by reducing the number of edge position measurement points or increasing the rotational speed.

Next, it is determined whether the semiconductor wafer 1 whose orientation flat portion has been determined as described above is the first semiconductor wafer 1 to be processed by the orientation flat detection section of this embodiment (S203).

When the semiconductor wafer 1 is the first one, in order to measure the assembly angle error of the line sensor 7 attached to the orientation flat detection section, the orientation flat section and the running direction of the orientation flat detection Y stage 5 and The arithmetic control unit 13 adjusts them so that they are parallel (S204).

This adjustment method will be explained with reference to the flowchart shown in FIG.

In step S202, the overrun rotation amount β during measurement is added to the rotation angle θi which takes the minimum value determined to indicate the orientation flat position, so that the orientation flat faces the direction of the line sensor 7 ( 3) θ using Eq.
Calculate the corrected drive amount Δθi1 and drive the θ drive unit 10 to rotate the θ stage 3 for orientation flat detection (S701)
. Δθi1=θi −β (3) Next, the Y stage 5 for orientation flat detection is scanned, the edge positions at both ends of the orientation flat are measured by the line sensor 7 through the signal processing unit 9, and their X and Y coordinates are determined. (S702
). At this time, the semiconductor wafer 1 has an eccentricity with respect to the orientation flat detection chuck 2, and the amount of eccentricity is unknown, so the actual position is not at both ends of the orientation flat, but at the inner edge position that takes into account the maximum eccentricity. Measure.

FIG. 8 shows the semiconductor wafer 1 with the orientation flat facing the line sensor 7 with the maximum eccentricity within the specified values. In this state, the orientation flat is approximately parallel to the running direction of the orientation flat detection Y stage 5, but this is not accurate. In FIG. 8, s is the length of the orientation flat, and d is the maximum value of eccentricity. The scanning range in the Y-axis direction from the center of the chuck should be determined so as to be reliably on the orientation flat, and the Y coordinates YA and YB of both ends A and B of the scanning range are respectively expressed by equation (4).

[0047]

[Equation 3] Next, based on the edge position measured in step S702, the θ correction drive amount Δθ1 for making the running directions of the orientation flat and the orientation flat detection Y stage 5 parallel to each other is calculated (
S703). The X coordinate of the edge position at one end A is
If the X coordinate of the edge position at A and the other end B is XB,
The θ correction drive amount Δθ1 is expressed by equation (5).

[0048]

[Equation 4] Then, drive the θ-axis drive unit 10 to detect θ for orientation flat detection.
The stage 3 is rotated by the θ-corrected drive amount Δθ1 determined in step S703 (S704).

Next, the θ correction drive amount Δθ1 is checked (S705). This corrected drive amount Δθ1 is
If it is within the predetermined tolerance, the next step S706
The orientation flat detection Y stage 5 is returned to the origin, and this parallelization processing is completed. On the other hand, if the θ-corrected drive amount Δθ1 is outside the tolerance, the process returns to step 702 and repeats the process.

Next, the assembly angle error of the line sensor 7 is measured (S205).

FIG. 9(a) shows the semiconductor wafer 1 at the end of the above-mentioned parallel alignment, that is, in a state where the orientation flat is parallel to the running direction of the Y stage 5 for orientation flat detection. In this state, first, measure the distance l from the center of the chuck to the orientation flat.
OF is determined using the line sensor 7.

Then, the θ-axis drive unit 10 is driven to rotate the orientation flat detection θ stage 3 by 90° in the negative direction of the θ axis, and the Y-axis drive unit 12 is driven to rotate the orientation flat detection Y stage 5 on the Y axis. in the positive direction by an amount slightly larger than the distance lOF from the center of the chuck to the orientation flat. moreover,
In this state, the X-axis drive unit 11 is driven to move the orientation flat detection X stage 4 in the positive direction of the X-axis by the distance from the center of the chuck to the center of the line sensor 7. At this time, the orientation flat and the running direction of the orientation flat detection X stage 4 are parallel to each other. The positions of the semiconductor wafer 1 and the line sensor 7 in this state are shown in FIG. 9(b).

From this position, while driving the Y-axis drive section 12 to scan the Y stage 5 for orientation flat detection in the negative direction of the Y-axis, the output signals of the pixels at both ends of the line sensor 7 are detected through the signal processing section 9. do.

The amount of Y driving until the pixels at both ends of the line sensor 7 change from being completely transparent to completely blocking light, that is, the distance δy in the y direction shown in FIG. 9(c), is set as δy1. Due to this distance δy1, the assembly angle error δθ of the line sensor 7
1 is expressed by equation (6), where L is the distance between pixels at both ends of the line sensor 7, and lP is the width of the pixels of the line sensor 7.

[0055]

[Equation 5] This assembly angle error δθ1 is used in the orientation flat detection sequence for the second and subsequent semiconductor wafers, so
Remember it.

After that, step S206 and step S2
In step 07, the semiconductor wafer 1 is positioned at the delivery position to the aforementioned load hand (not shown).

First, in step S206, while reading out the output signal of a pixel located on the X axis of the line sensor 7 (in this embodiment, the central pixel of the line sensor 7) through the signal processing section 9, the pixel is semi-shaded. The Y-axis drive unit 12 is driven to move the Y stage 5 for orientation flat detection so that After the movement, the semiconductor wafer 1 is in a state in which the orientation flat portion is along the x-axis, as shown in FIG. Figure 10
, the portion indicated by the dashed line is the designated orientation flat position relative to the delivery position. Then, X shown in FIG.
The Y-axis drive unit 12 is driven by the designed distance YOF from the axis to the orientation flat delivery position to move the orientation flat detection Y stage 5 to y.
Move the axis in the negative direction.

In step S207, while detecting the edge position of the semiconductor wafer 1 through the signal processing section 9, the X-axis drive section 11 is driven to move the X stage 4 for orientation flat detection, and the semiconductor wafer is driven from one direction. Position the end of 1 at the specified position. Then, the LED drive unit 8 is driven to turn off the LED 6, and the orientation flat detection sequence is completed.

On the other hand, if it is determined in step S203 that it is not the first semiconductor wafer, then it is the second or subsequent semiconductor wafer, and the above-mentioned assembly angle error of the line sensor 7 has already been determined. Therefore, the semiconductor wafer is aligned.

In the case of the second and subsequent semiconductor wafers, first, in step 208, the orientation flat is made parallel to the line sensor 7.

The alignment of the orientation flat and the line sensor 7 will be explained with reference to the flowchart shown in FIG. 11.

First, the orientation flat is set to be approximately parallel to the line sensor 7 (7 ) is used to calculate the θ correction drive amount Δθi2, and the θ-axis drive unit 10 is driven to rotate the orientation flat detection θ stage 3 (S1101).

Δθi2=θi −β−90° (
7) After the rotation, the semiconductor wafer 1 is in a state in which the orientation flat faces the negative direction of the Y-axis and is substantially parallel to the line sensor 7. Then, the Y-axis drive unit 12 is driven to move the orientation flat detection Y stage 5 in the positive direction of the Y-axis by an amount slightly greater than the edge distance l corresponding to the rotation angle θi, and the X-axis drive unit 11 to move the orientation flat detection X stage 4 in the positive direction of the X axis by the distance from the center of the chuck to the center of the line sensor 7 (S1102).

At this time, the semiconductor wafer 1 is close to the state shown in FIG. 9(b), but more precisely, the orientation flat is not parallel to the running direction of the orientation flat detection X stage 4, and the orientation flat is not parallel to the running direction of the orientation flat detection X stage 4. The accuracy of the direction is only the accuracy of the measurement interval angle of the edge position in step S201. Further, the orientation flat detection X stage 4 and the orientation flat detection Y stage 5 serve as scan start positions for detecting the θ correction drive amount for making the orientation flat and the line sensor 7 parallel.

Next, from that position, the Y-axis drive unit 12
While scanning the orientation flat detection Y stage 5 in the negative direction of the Y axis, the output signals of the pixels at both ends of the line sensor 7 are detected through the signal processing unit 9 (S1103).
. Then, the Y driving amount until the pixels at both ends of the line sensor 7 go from completely transparent to completely blocked, that is, (c) in FIG.
The distance δy in the y direction shown in is set as δy2.

Using this Y drive amount δy2, the θ correction drive amount δ is used to make the orientation flat and the line sensor 7 parallel.
Calculate θ2. The θ correction drive amount δθ2 is expressed by equation (8), where L is the distance between pixels at both ends of the line sensor 7, and lP is the width of the pixels of the line sensor 7.

[0067]

[Equation 6] Then, the θ-axis drive unit 10 is driven to rotate the orientation flat detection θ stage 3 by the determined θ correction drive amount δθ2 (S1105).

Next, a tolerance check of this θ correction drive amount δθ2 is performed (S1106). Correction drive amount δ
If θ2 is outside the tolerance, return the orientation flat detection Y stage 5 to the scan start position (S1107),
The follow-up process for the θ correction drive amount δθ2 is performed again from step S1303. The θ correction drive amount δθ2
If it is within the tolerance, the parallel alignment process between the orientation flat and the line sensor 7 is completed.

When the parallel alignment process between the orientation flat and the line sensor 7 is completed, in step S209, the step S
In order to make the orientation flat aligned with the line sensor 7 in step 208 perpendicular to the running direction of the orientation flat detection Y stage 5, an angle correction is performed for the assembly angle error δθ1 of the line sensor 7. The assembly angle error δθ1 of the line sensor 7 is
Step S of the orientation flat detection sequence for the first wafer
The measurement is completed in 205 and stored in memory. Therefore, here, the θ-axis drive unit 10 is driven to rotate the orientation flat detection θ stage 3 in the θ-axis negative direction by the assembly angle error δθ1 of the line sensor 7, and then the processing is performed in steps S206 and 207. and perform positioning in the Y and X directions.

In the explanation of this embodiment, as a method for determining the orientation flat portion, the minimum value of the wafer outer edge position data string l is searched to match the orientation flat condition, and when two orientation flat portions are present, Although a method of determining the first orientation flat using the relative angle information was used, of course, other orientation flat determination methods may be used.

Furthermore, when aligning the orientation flat and the running direction of the Y stage 5 for orientation flat detection, the θ correction amount was calculated by setting a measurement point that takes into account the safety of the eccentricity, but the Y stage for orientation flat detection The edge position of the semiconductor wafer is measured in synchronization with the scanning in step 5, and the point where the rate of change in the edge position changes is recognized as a circular arc, and both ends of the orientation flat are used as measurement points for calculating the θ correction amount. You can also do that. Using this method takes time, but the θ correction drive amount Δθ
Since the denominator of calculation formula (5) of 1 becomes larger, the θ accuracy is further improved.

In this embodiment, the assembly angle error of the line sensor 7 was measured only on the first sheet of the normal orientation flat detection sequence, but the assembly angle error may be measured and stored during device assembly or maintenance. However, there is also a method in which the assembly angle error is not measured in the normal orientation flat detection sequence.

[0073]

[Effects of the Invention] Since the present invention is constructed as described above, it produces the following effects. (1) Orientation of the semiconductor wafer The flat part of the semiconductor wafer is made parallel to the line sensor, and from that state, the semiconductor wafer is rotated by the assembly angle error of the line sensor to correct the orientation of the semiconductor wafer, so at this point The angular error in the orientation flat position due to the assembly angle error of the line sensor is removed, and the orientation flat position is accurately recognized.
The alignment accuracy of the semiconductor wafer is improved. (2) After the orientation flat position of the semiconductor wafer is detected, the angle detection error of the orientation flat position is corrected, so regarding the detection of the orientation flat position based on the edge position of the semiconductor wafer,
It is sufficient to be able to recognize the approximate position of the orientation flat, and it is possible to reduce the number of measurement points for the edge position or increase the rotational speed when measuring the edge position, thereby improving the throughput of orientation flat detection.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an orientation flat detection unit for carrying out the semiconductor wafer positioning method of the present invention.

FIG. 2 is a flowchart showing an embodiment of the semiconductor wafer positioning method of the present invention.

FIG. 3 is a diagram showing changes in the distance from the chuck center to the outer edge of the semiconductor wafer with respect to the rotation angle, (a) shows the case of a semiconductor wafer with only one orientation flat part, and (b)
) shows the case of a semiconductor wafer in which two orientation flat parts, first and second, are formed.

FIG. 4 is a flowchart showing the operation of orientation flat position determination.

FIG. 5 is a diagram showing changes in the distance from the center of the chuck to the outer edge of the orientation flat portion with respect to the rotation angle.

FIG. 6 is a plan view showing a semiconductor wafer in an xy coordinate system with the origin at the center of the chuck.

FIG. 7 is a flowchart showing an operation for making the orientation flat part parallel to the running direction of the orientation flat detection Y stage.

FIG. 8 is a diagram showing a semiconductor wafer facing a line sensor direction in an xy coordinate system with the origin at the center of the chuck.

FIG. 9 is a plan view showing a semiconductor wafer and a line sensor in an x-y coordinate system with the origin at the center of the chuck; (a) shows a state in which the orientation flat is parallel to the y-axis, and (b) ) shows a state in which the orientation flat is parallel to the x-axis, and (c) shows the orientation of the line sensor.

FIG. 10 is a diagram showing a semiconductor wafer oriented in a specified direction in an xy coordinate system with the origin at the center of the chuck.

FIG. 11 is a flowchart showing an operation for making the semiconductor wafer parallel to the line sensor.

FIG. 12 is a diagram for explaining a conventional method for positioning a semiconductor wafer.

FIG. 13 is a plan view showing the outer edge of the semiconductor wafer passing over the line sensor, and (a) to (e) sequentially show states in which the orientation flat part of the semiconductor wafer passes over the line sensor. .

[Explanation of symbols]

1 Semiconductor wafer 2 Orientation flat detection chuck 3 θ stage for orientation flat detection 4 X stage for orientation flat detection 5 Y stage for orientation flat detection 6 Light emitting diode 7 Line sensor 8 LED drive unit 9 Signal processing section 10　θ-axis drive section 11 X-axis drive section 12 Y-axis drive section 13 Arithmetic control section 61 First orientation flat 62 Second orientation flat

Claims (3)

[Claims] Claim 1: While a semiconductor wafer for semiconductor manufacturing is rotated by a rotating means, a non-contact outer edge position detection means of a light emitting means and a line sensor serving as a light receiving means for the light emitting means is used to detect one circumference of the outer edge of the semiconductor wafer. , detect the orientation flat portion of the semiconductor wafer from a change in the measured edge position, determine the orientation flat position indicating the direction of the orientation flat portion, and use the determined orientation flat position as a reference to determine the orientation flat portion of the semiconductor wafer. the semiconductor wafer in a designated direction using means for moving it in two directions: a first movement direction parallel to the direction of the line sensor and a second movement direction perpendicular to the first movement direction. In a method for aligning a semiconductor wafer that is moved to a designated position, the orientation flat position is determined, and then an angular difference between the orientation flat part and the first moving direction is determined based on the orientation flat position, and the determined angular difference is performed. driving the rotation means according to the orientation flat portion and the first
The semiconductor wafer is moved so that the center of the orientation flat section substantially coincides with the center of the line sensor, and while scanning the semiconductor wafer in the second moving direction, determining the angular difference between the line sensor and the orientation flat portion, and driving the rotating means according to the determined angular difference;
The orientation flat part and the line sensor are made parallel to each other, and the rotation means is driven according to a previously measured assembly angle error of the line sensor with respect to the first moving direction, so that the orientation flat part is moved to the first direction. of the semiconductor wafer, wherein the rotating means is driven in accordance with the angular difference between the first moving direction and the designated direction to orient the orientation flat portion in the designated direction. Alignment method. 2. When aligning the first semiconductor wafer, the orientation flat position of the semiconductor wafer is determined, and then the angular difference between the orientation flat portion and the second moving direction is determined based on the orientation flat position. , drive the rotating means according to the obtained angular difference to make the orientation flat section and the 22nd movement direction substantially parallel, and while scanning the semiconductor wafer in the second movement direction, rotate the orientation flat section. Measure edge positions at arbitrary two points, determine the inclination of the orientation flat portion with respect to the second moving direction from the measured edge positions, and drive the rotation means according to the determined inclination to change the orientation. The flat part is
Then, the rotating means is further driven to move the orientation flat part parallel to the second direction of movement.
The semiconductor wafer is moved perpendicularly to the moving direction of the line sensor so that the center of the orientation flat portion substantially coincides with the center of the line sensor, and while scanning the semiconductor wafer in the second moving direction, the line The angular difference between the sensor and the orientation flat part is determined, the determined angular difference is stored as an assembly angle error of the line sensor, and the orientation flat part is adjusted to the angular difference between the second movement direction and the designated direction. 2. The method of positioning a semiconductor wafer according to claim 1, wherein said rotating means is driven to orient in said designated direction. 3. When determining the angular difference between the line sensor and the orientation flat part, while scanning the orientation flat part, the pixels at both ends of the line sensor are set to a 3. The semiconductor wafer positioning method according to claim 2, further comprising measuring the amount of movement in the moving direction of the semiconductor wafer, and determining the angular difference based on the amount of movement.