JP2711582B2 - Positioning method - Google Patents

Description

Description: FIELD OF THE INVENTION The present invention relates to a method for detecting the position of an alignment mark when sequentially aligning (aligning) a plurality of regions on a substrate at a desired position, and particularly relates to semiconductor manufacturing. In a step-and-repeat type exposure apparatus, a position related to a shot area on a semiconductor wafer is measured, and a position detection method of an alignment mark when each shot is aligned.

2. Description of the Related Art Conventionally, in a step-and-repeat type exposure apparatus for manufacturing a semiconductor, that is, a stepper, a method of aligning a shot area on a semiconductor wafer is disclosed in, for example, Japanese Patent Application Laid-Open No. No. -232321.

This alignment method can measure all the shot areas on the wafer with high accuracy by simply measuring the position related to the shot area on the wafer, and is extremely excellent. is there.

[Problems to be Solved by the Invention] However, semiconductor wafers are produced from various processes. At this time, alignment marks in each process are variously changed due to a change in a material of a wafer base, a mark step, and a covering of a photosensitive material (resist). To change. In the conventional alignment method, the processing parameters are fixed. Therefore, the parameters are not always optimal for the change of the alignment mark as described above.

The present invention has been made in view of such circumstances, and an object of the present invention is, for example, when the present invention is applied to a step-and-repeat type exposure apparatus for semiconductor manufacturing, the position of a measurement mark on a semiconductor wafer. It is to further improve the measurement accuracy at the time of measurement.

Means and Action for Solving the Problems In order to achieve the above object, the present invention provides a window of a predetermined size for a signal obtained by imaging an alignment mark by an imaging unit, In the position detection method of calculating a matching degree using a template at each position in the measurement direction of the signal of the signal and detecting the position of the alignment mark based on the calculated plurality of matching degrees, the effective range of the template is changed. Detecting the position of the alignment mark a plurality of times using each of the changed templates and calculating a variation in the detection result; and detecting the position of the alignment mark based on the variation for each template. Deciding a template to be performed.

The changing may include changing a width of an effective range of the template.

The changing may include changing a center of an effective range of the template.

[Example] Hereinafter, the present invention will be described in detail based on an illustrated example.

FIG. 1 shows an embodiment of a step and repeat type semiconductor exposure apparatus according to the present invention. In this figure, RT is a reticle on which a pattern PT for manufacturing a semiconductor element is formed, WF is a semiconductor wafer having a number of shots SH, and LN is a reduced projection of the pattern PT on the reticle RT onto one shot SH of the wafer WF. CU is a control unit for controlling the entire stepper, and CS is a console for inputting necessary information such as alignment data and exposure data to the control unit CU.

The control unit CU has a plurality of computers, a memory, an image processing device, an XY stage control device, and the like. In addition, in order to detect the positional shift amount and the feature amount of the imaged mark from the image output of the image pickup device CM, as shown in FIG. 2, an A / Q that quantizes each pixel signal from the image pickup device CM is used.
A D conversion device 21, an integration device 22 that integrates the quantized pixel signal from the A / D conversion device 21 in a predetermined direction, and a position detection device 23 that detects a position shift amount of a mark from the signal integrated by the integration device 22. have.

 This configuration will be described later in detail.

The reticle RT follows a command from the control unit CU,
It is held by suction on a reticle stage RS that moves in the X, Y, and θ directions. Reticle RT projects reticle RT to projection lens LN
Reticle alignment marks RAMR / RAML and reticle used when aligning to a predetermined positional relationship with respect to
It has reticle marks RMR and RML used when detecting the positional relationship between the RT and the shot SH on the wafer WF.

The reticle set marks RSMR and RSML are formed on a member fixed to the lens barrel of the projection lens LN so as to have a predetermined positional relationship with respect to the projection lens LN. The alignment of the reticle RT with respect to the projection lens LN is performed by using a set of the mark RAMR and the mark RSMR and a set of the mark RAML and the mark RSML in the imaging device CM.
The control unit CU moves the reticle stage RS so that the positional deviation amount detected from the image output at this time is within a predetermined allowable value.

The wafer WF is held by suction on the wafer stage WS.
The wafer stage WS moves the wafer WF in the Z and θ directions with respect to the XY stage XYS.

MX and MY are motors for moving the XY stage XIS in the X and Y directions.

MRX / MRY is a mirror fixed to the XY stage XYS.

 IFX / IFY is a laser interferometer.

XY stage XY for moving wafer WF in X and Y directions
The S is constantly monitored on the XY coordinates by mirrors MX and MY called laser interferometers IFX and IFY, and the motor MX
Move to the position instructed by MY from the control unit CU.

Control unit CU keeps laser interferometer IFX / IFY after moving
The XY stage XYS is held at the specified position based on the output of the XY stage.

A large number of patterns (shots SH) have already been formed on the wafer WF in the previous exposure step, approximately in the X and Y directions, and the wafer alignment marks WAML / WAMR have been formed.
Are formed. Each shot SH is provided with a wafer mark WML / WMR. Returning to FIG. 1, the OS is an off-axis scope for imaging the wafer alignment marks WAML / WAMR in order to detect the positions on the XY coordinates of the wafer alignment marks WAML / WAMR on the wafer WF.
The off-axis scope OS is firmly fixed so as to maintain a predetermined positional relationship with the projection lens LN.

IL is an illumination device for illuminating the reticle RT with printing light when printing the pattern PT of the reticle RT onto the shot SH of the wafer WF via the projection lens.

SHT is a shutter for controlling an exposure amount at the time of printing, and these also operate according to a command from the control unit CU.

LS is a laser light source that emits laser light having a wavelength substantially the same as the printing wavelength.In order to detect the amount of positional shift between the pattern PT of the reticle RT and the shot SH of the wafer WF through the projection lens, the image pickup device CM uses the reticle. This is used to illuminate each mark when each mark is superimposed and imaged with a set of the mark RML and the wafer mark WMR.

The laser light from the laser light source LS is diffused and smoothed by the diffusion plate DP, and then used as illumination light for each mark.

LSH can be used when laser light is not required
Is a shutter for blocking the laser light from the laser light source LS so as not to reach the wafer WF when the laser light is stepped.

The detection of the amount of displacement by such a configuration is as follows. In the following description, with respect to the front direction indicated by the arrow in FIG. 1, the right hand direction is referred to as a right direction, and the left hand direction is referred to as a left direction.

The laser light emitted from the laser light source LS is diffused by the diffusion plate DP and then scanned by the polygon mirror PM. After that, the scanning is converted into constant-speed scanning by the fθ lens Fθ, passes through the beam splitter BS, and is split right and left by the roof prism DAP. The laser beam split in the left direction is irradiated onto the region including the reticle mark RMR from above the reticle RT by the right objective mirror AMR.

The laser beam transmitted through the reticle RT passes through the reduction projection lens LN
From the area including the wafer mark WMR on the right side of the shot. The reflected light from the area including the wafer mark WMR follows the optical path opposite to that described above, passes through the area including the projection lens LN and the reticle mark RNR, and reaches the roof prism DAP. Similarly, the laser beam split rightward by the roof prism DAP is also irradiated from the left objective mirror AML to the area including the reticle mark RML, and then passes through the same optical path and is reflected from the area including the wafer mark WML to the roof prism DAP. Return. After the right and left laser beams are aligned by the roof prism DAP, the laser beams pass through the beam splitter BS, are expanded by the elector EL, and are formed as images shown in FIG. 3 on the imaging surface of the imaging device CM. Wafer mark WML / W imaged on imaging device
The image of the MR is adjusted by an
It has been expanded by EL and others. The imaging device CM is, for example,
It is a photoelectric conversion device such as an ITV camera or a two-dimensional image sensor, and converts images of the formed reticle marks RSL / RSR and wafer marks WML / WMR into two-dimensional electric signals.

FIG. 3 shows the reticle mark RSL / R formed on the image pickup device CM.
FIG. 4 is an explanatory diagram of an area including an SR and a wafer mark WSL / WSR.
In this figure, the RML / R
Each of MR and wafer mark WML / WMR is specified in more detail. In this figure, the reticle mark RML is RML _x
RML _y is indicated, and reticle mark RMR is indicated as RMR _x · RMR _y . In Fig. 3, the left half is the mark on the left side of shot SH.
An image of WML _x / WML _y and mark RML _x / RML _y on the right side of reticle RT is shown, and the right half is mark WMR _x / WMR _y on the right side of shot SH.
And images of marks RMR _x and RMR _y on the left side of the reticle RT.

The image of the reticle mark _{RML x · RML y · RMR x} · RMR y appear black in Figure 3, the reticle RT to the illumination from the back, and imaging the transmitted light imaging device CM is by the light reflected from the wafer WF Because it is.

The image converted into a two-dimensional electric signal by the imaging device is digitized by the A / D conversion device 21 shown in FIG.
For example, binarized and xy corresponding to the position of each pixel on the imaging surface
It is stored in an image memory having an address. The contents of the image stored in the image memory correspond to those obtained by adding an X address (coordinate) in the horizontal direction and a Y address (coordinate) in the vertical direction in FIG.

The deviation amount measurement is performed independently for each of the four sets of mark images in FIG.

That is, the amount of displacement D1 _x in the left visual field X direction via the objective mirror AML is calculated from the difference between the positions of the reticle mark RML _x and the wafer mark WML _x within the screen, using the reticle mark RML _y and the wafer mark WML _y.
Similarly, the shift amount D1 _y in the left visual field Y direction is obtained. Similarly, a shift amount D in the right visual field X direction from the reticle mark RMR _x and the wafer mark WMR _x via the objective mirror AMR.
The _rx, seeking each deviation amount D _ry Similarly from the reticle mark RMR _y and wafer marks WMR _y in the right-field Y direction. The measurement of each shift amount is the same except for the difference in the measured values at the XY coordinates, and therefore, the measurement in the X direction of the left visual field will be described below as an example.

FIG. 4 (a) is a set of marks RML _{x at the} upper left of FIG.
-Represents WML _x . In the superposition of the pattern in the shot SH and the pattern PT of the reticle RT, the marks of each set described above are designed such that when the patterns are accurately superimposed, the relative displacement amount becomes zero. . That is, the fourth
In FIG. (A), PRL imaging plane position of the left mark components of the reticle mark RML _x, the imaging plane located right mark component P
If RR and the position of the wafer mark WML _{x in} the imaging plane are PWM,
The displacement amount D1 _x is as follows: D1 _x = PWM− (PRL + PRR) / 2.

Next, a method of calculating the positions PRL, PRR, and PWM will be described. W _k (k = 1 to n) in FIG. 4A represents a two-dimensional window set on the imaging surface.

Second accumulation shown in FIG. 22 is a respective window W _k of the 2-dimensional, perpendicular to the direction (in this case the X direction) for detecting a deviation amount (in this case the Y direction) to the A / D converter Device 2
The pixel values from 1 are integrated, and a one-dimensional integrated waveform S _k (x)
Get. If the pixel data value P (X, Y) in the image memory range in the Y direction of the window W _k and the _{Y k1 ≦ Y ≦ Y k2 S} k (x) is It is expressed as Actually, n windows W _k are set as shown in FIG. 4A, and a projection integrated waveform as shown in FIG. 4B is obtained for each window W _k .
Since in the image captured, the edge signal portion of the reticle mark RML _x and wafer marks WML _x is the contrast changes abruptly than in other areas, integrated waveform S
_{As for k} (x), the contrast in the direction perpendicular to the integration direction (X direction) is enhanced, and the S / N is enhanced, so that a steep peak or dip is observed in the external signal portion.

Position detection device 23 shown in FIG. 2, the integrated waveform S _k above
From (x), mark positions PRL, PRR, and PWM are detected. The position detecting device 23, the same process for each of the integrated waveform _{S 1 (x) ~S n (} x) of FIG. 4 (b) is being performed. In the following description, an arbitrary integrated waveform S (x) is taken as an example. Mark position detection is based on the reticle mark position.
It is divided into PRL and PRR detection processing and wafer mark position PWM position detection. The process of detecting each mark position is divided into a process of obtaining a coarse position and a process of obtaining a precise position.

In the processing for obtaining the rough position, both the wafer mark position detection and the reticle mark position detection use the template matching method. First, the position PWM detection of wafer marks WML _x. The ideal waveform obtained by integration is shown in FIG.
And S (x) as shown in FIG. 5 and P (x) as a template shown in FIG. 5 (b), a matching evaluation equation shown by the following equation provides a matching degree E (x _k ) at an arbitrary point x _k . .

The parameters c and w in the above equation indicate the effective range of the template, where c is the center of the effective range and w is the width of the effective range. Matching degree E for any point x _k
The value of (x _k ) has a peak at a rough position of the wafer mark WML as shown in FIG.

The x _k coordinate value at which the match degree E (x _k ) becomes a peak is x _p ,
The peak value at that time is defined as a peak matching degree E (x _p ).
In practice, the integrated waveform does not always become S (x) due to the semiconductor manufacturing process or the resist film thickness. Therefore, in reality, a multi-template method is used in which similar processing is performed using several types of templates, the degree of matching of each is calculated, and the maximum one is adopted.
Representative templates other than FIG. 5 (b) are shown in FIGS. 5 (d) to 5 (f). Precise mark position detection is attached to the adopted matching degree function E (x _k), are determined by centroid computation of several points around the position x _p. Or E
(X _k ) may be approximated by a curve and determined from the peak value of the approximate curve.

Position PRL of each mark component of the reticle mark RML _x, a rough detection processing same template matching process PRR. After determining the center position of the two reticle marks in the template matching process, the left and right reticle mark positions are determined from the detected center positions of the two reticle marks.
After roughly calculating the position of each reticle mark, the center of gravity of the integrated waveform S (x) is calculated around the calculated coarse position.

In this manner, the position detection device 23 sets each window W _k
Reticle mark positions PRL _k , PR shown in FIG.
After calculating R _k and the wafer mark position PWM _k , each window W
the average value of the positional deviation amount D _Lxk reticle marks RML _x and the wafer mark WML _x per _k, as indicated by the following equation is obtained as a position deviation amount D _lk reticle marks RML _x and the wafer mark WML _x. In addition, the reticle mark
The positional deviation amount D _ly from the positional displacement amount D _lyk for each window W _k of RML _y and the wafer mark WML _y, reticle marks RMR
The positional deviation amount D _ly from the positional displacement amount D _lyk for each window W _{k x} and the wafer mark WML _y, the position shifted from the position displacement amount D _Ryk for each window W _k of the reticle mark RMR _x and wafer marks WMR _{y Determine the} amount _Dry .

Next, a method of optimizing the parameters of the mark measurement will be described.

The measurement parameters include, for example, an integrated width (Y _k2 −Y _k1 ) in each window, a template shape, and a template effective range (c, w). Has been adjusted.

In each semiconductor manufacturing process, one of the main factors related to the changing alignment is the line width variation of the alignment mark. This depends on whether the mark part is etched or the part other than the mark is etched at the time of transferring the alignment mark, and changes depending on the degree of etching or the degree of etching, or the presence or absence of a substance covered on the mark. Things. For example, FIG.
Is a mark cross section when the mark portion 61 is etched and Al62 is deposited by sputtering.In this case, clearly, an edge appears inside the line width transferred from the reticle, and the edge position of the mark is narrowed. You can see that there is. On the other hand, FIG. 6 (b) shows a case where a portion other than the mark is etched and Al62 is deposited. In this case, the edge of the mark appears at a position wider than the line width transferred by the reticle. I understand. Such a change in line width may be large depending on a process in which the mark is formed. In that case, the position where the edge signal of the mark appears varies, and the edge signal portion deviates from the fixed template effective width, thereby causing a reduction in S / N. In other words, it may be said that the optimal parameter in terms of S / N is to capture the edge signal portion without excess or deficiency.

As an evaluation scale for setting an optimal parameter from the viewpoint of the S / N, it is possible to employ measurement variation when measuring different portions of a mark that is strongly affected by stationary random noise. . That is, in general, there is a characteristic that the dispersion or variance of these measurement values tends to increase as the influence of stationary random noise increases.

FIG. 7 is a diagram for explaining a parameter optimization portion. The imaging device CM, the A / D converter 21, and the integrating device 2 in FIG.
2. The position detecting device 23 is the same as the mark detecting portion. FIG. 71 shows a parameter set generator, which comprehensively generates a set of parameters in which the effective range of the template is changed within a meaningful range and passes it to the position detector. Seventh
FIG. 72 shows a feature amount extraction device, in which each window W
_{The variance} of the positional deviation amounts D _lxk , D _lyk , D _rxk , and D _ryk for each _k Holds the variance value of the set of parameters and the amount of positional deviation between the windows. Next, after the parameter generator 71 generates a new parameter set and performs position detection, the feature extractor holds the new parameter set and the variance. The above loop is repeated until the parameter generator finishes generating new parameters, and finally the set with the smallest variance among these sets is output as the optimal parameter.

As can be easily understood, the parameter generator 71 generates a parameter by a so-called hill-climbing search method and obtains a minimum value by feeding back a variance value that is an evaluation value without exhaustively generating a parameter. Can obtain substantially the same effect.

In addition, the rotation angle θ between the shot SH and the reticle RT is large,
If the wafer mark WML _x as shown in FIG. 8 (a) is significantly inclined relative to the reticle mark RML _x is the deviation amount is dependent on the position of the window W _k as shown in Figure 8 (b) The variance becomes large even when the error of the measurement value of each window is small. So if in this way the wafer mark is rotating may use the least square error when the linear approximation the deviation amount for each position of the window W _k in place of the dispersion.

FIG. 9A shows an example of an integrated waveform of a wafer mark in a wafer process. FIG. 9B is a contour graph showing a change in variance of measured values between windows when the center c of the template effective range is set on the horizontal axis and the width w of the template effective range is set on the vertical axis. FIG. 9 (b)
As shown at 91, there is a position where the variance shows the minimum, and this indicates that the set of parameters at that position is optimal for S / N. FIG. 10A shows an example of an integrated waveform of another pattern having a different line width. FIG. 10 (b) is a contour graph showing the change of the variance similarly to FIG. 9 (b). In this case, as shown at 101, the position showing the minimum value of the variance due to the change of the line width is shown. This is different from the case of FIG. 9 (b), and shows that the optimal set of parameters has changed.

Next, the procedure of the positioning according to the present embodiment will be described. First, the reticle RT is sent and set on the reticle stage RS by the transfer hand mechanism (not shown), and the wafer WF is set.
Is transported onto the wafer stage WS by a transfer hand mechanism (not shown), and is fixed on the wafer stage WS by vacuum suction. Wafer has wafer alignment marks WAML and WAMR
It is moved to the XY stage, imaged via the off-axis scope OS, and roughly aligned. Next, the alignment mark of the shot SH to be observed on the wafer WF is sent to the position below the projection lens according to the measurement values of the XY stage XYS and the interferometer IFX, IFY, and held at a position observable by the imaging device CM. Observe the alignment mark, optimize the measurement parameters, and determine the optimal parameter set
In the storage device in the server. Thus, one shot
It is also possible to optimize parameters with SH, but in order to further improve the accuracy of optimization, optimize the parameters with multiple shots in the wafer, output the average value,
It may be a final set of parameters. The purpose of the averaging is to remove a change in the characteristic amount depending on the mark position in the wafer, and in that sense, it is desirable to select a plurality of shots at positions evenly distributed over the entire area in the wafer. The optimal set of parameters is the left and right marks WMR and WML of the shot.
It may be determined every time, or the average of the left and right optimum parameters may be obtained. The same wafer is then aligned and step-and-repeat exposed according to the aligned position using this optimally determined set of parameters. Further, when aligning a wafer under the same process conditions as the measured wafer, the alignment is performed by referring to the optimal parameters stored in the storage device in the CU.

[Second Embodiment] In the first embodiment, the relative variation from the design position of the measurement value of a different part of the mark is used as a feature amount reflecting the accuracy of measurement of the alignment mark. In the present embodiment, as a feature amount reflecting the accuracy of the alignment mark measurement, the mark position is held by an XY stage XYS that performs servo feedback at the observation position, and measurement is performed a plurality of times.
The variation of the measured value is used as a feature value. Since the XY stage is positioned with high precision, the main cause of the dispersion of the measured values is the influence of stationary random noise. Therefore, it is possible to employ the variation of the measured value as a feature amount reflecting the accuracy of the measurement. As in the first embodiment, the minimum value of the variation of the measurement multiple times when the measurement parameter is changed is determined, and the parameter set indicating the minimum value is determined as the optimal parameter set. Other parts are the same as in the first embodiment.

[Effects of the Invention] As described above, according to the present invention, when an image of an alignment mark is taken and its position is detected using a template, the position of the alignment mark is detected a plurality of times using different templates, and the detection result is obtained. Since the variation is calculated and the template for detecting the position of the alignment mark is determined based on the variation of each template, the alignment mark position can be detected with high accuracy by using a template having an optimum effective range. it can.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of a step-and-repeat type semiconductor exposure apparatus according to the present invention, FIG. 2 is an explanatory view of a mark detection portion of the present embodiment, and FIG. 4 (a) and 4 (b) are illustrations of an ideal mark image and mark signals, and FIGS. 5 (a) to 5 (f) are illustrations of marks. FIGS. 6 (a) and 6 (b) are explanatory diagrams of a change in mark line width, and FIG. 7 is an explanatory diagram of a mark measurement parameter optimizing portion. 8 (a) and 8 (b) are explanatory diagrams and graphs of the state of change of the measured position shift amount when the mark is rotating, and FIGS. 9 (a), (b) and 10th. FIGS. 9A and 9B show the relationship between mark line width and measurement variation between windows. FIG. XYS: XY stage, WS: Wafer stage, WF: Wafer, LN:
Reduction projection lens, RT: reticle, SHT: shutter, IL: exposure light source, CU: control unit, CS: console, LS:
Laser light, CM: camera.

 ──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-61-44429 (JP, A) JP-A-62-54436 (JP, A) JP-A-3-91917 (JP, A)

Claims (3)

(57) [Claims] 1. A window having a predetermined size is provided for a signal obtained by imaging an alignment mark by an imaging means,
In a position detection method for calculating a matching degree using a template at each position in a measurement direction of a signal in the window and detecting a position of the alignment mark based on the calculated plurality of matching degrees, an effective range of the template Using the changed templates, detecting the position of the alignment mark a plurality of times, calculating a variation in the detection result, and, based on the variation for each template, Deciding a template for detecting a position. 2. The position detecting method according to claim 1, wherein said changing step changes a width of an effective range of said template. 3. The method according to claim 1, wherein said changing step changes a center of an effective range of said template.