JPS6352767B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an apparatus for aligning two objects, and more particularly to an apparatus for aligning a mask or a wafer before printing a semiconductor integrated circuit pattern on a mask or reticle onto a wafer. suitable for

The semiconductor manufacturing process includes a step of sequentially transferring several patterns onto a wafer to form a semiconductor integrated circuit. In this case, in order to accurately position another pattern on the wafer onto which the pattern from the previous process has already been transferred, it is necessary to align the mask with the pattern and the wafer with high precision. This alignment is usually achieved automatically by photoelectrically detecting alignment marks written on the mask and the wafer, respectively, and using the detected signals.

On the other hand, recently, devices have been known that are equipped with a mask change mechanism that stores a mask or automatically attaches a mask in a carrier to a mask set position. This mechanism requires a pre-positioning function (referred to as mask alignment) to accurately set the mask in a predetermined position on the printing stage.

An object of the present invention is to provide an apparatus that achieves accurate mask alignment in a short time.

Embodiments of the present invention will be described below with reference to the drawings, with FIG. 1 showing the external appearance. 1 is a mask provided with an integrated circuit pattern, and is provided with other mask alignment marks and mask/wafer alignment marks. 2 is the mask stage,
The mask 1 is held and moved in the plane and in the rotational direction. 3 is a reduction projection lens, 4 is a wafer provided with a photosensitive layer, and is provided with a mask/wafer alignment mark and a wafer alignment mark. 5 is a wafer stage. The wafer stage 5 holds the wafer 4 and moves it in the plane and in the rotational direction, and also moves between the wafer printing position (inside the projection field) and the television wafer alignment position.
6 is an objective lens of a detection device for television wafer alignment, 7 is an image pickup tube or a solid-state image sensor, and 8 is a television receiver for video observation. A binocular unit 9 is useful for observing the surface of the wafer 4 through the projection lens 3. Reference numeral 10 denotes an upper unit housing an illumination optical system for converging the mask illumination light emitted from the light source 10a and a detection device for mask-wafer alignment. A signal processing circuit 12 calculates an alignment error and drives the mask stage 2 to remove this error.
The wafer stage 5 holds the wafer transported by a wafer transport means (not shown) at a predetermined position, and first moves to a position where the alignment mark on the wafer is within the field of view of the television wafer alignment objective lens 6. The positional accuracy at this time is due to mechanical alignment accuracy, and the field of view of the objective lens 6 is approximately 1 mm to 2 mm in diameter.
That's about it. The alignment mark within this field of view is detected by the image pickup tube 7, and the coordinates of the alignment mark from there are detected with reference to a reference mark for TV/wafer alignment (described later) provided in the optical system for TV/wafer alignment. The position is detected. On the other hand, since the detection position for auto alignment of the projection optical system and the position of the reference mark for TV/wafer alignment described above are set in advance, the auto alignment position can be determined from the coordinate position of these two points and the TV/wafer alignment mark. The feeding amount of the wafer stage 5 is determined.

The position detection accuracy of TV/wafer alignment is less than ±5μ, and even when taking into account the error caused by the movement of the wafer stage from the TV/wafer alignment position to the mask/wafer alignment position, it is about ±10μ. Therefore, alignment can be performed within a range of approximately ±10 μ, which is less than 1/100 of the field of view of conventional alignment, and alignment can be performed faster than conventional alignment. Note that the TV/wafer alignment will be explained in detail later.

FIG. 2 shows an embodiment for achieving mask alignment and mask-wafer alignment. In the figure, a mask 1, a reduction projection lens 3, a wafer 4, and a wafer stage 5 are as shown in FIG. The projection lens 3 is schematically depicted for convenience.
Reference numerals 11 and 11' are mask alignment marks, which are carved in an immovable location such as a lens barrel or a part of the device.

On the other hand, at the positions indicated by 20 and 20' on the mask 1, there are line or slit-shaped alignment lines arranged at an angle of 45 degrees with respect to the scanning line 60, as indicated by numbers 75 and 76 in FIG. A mark is provided. Further, at the positions indicated by 21 and 21' on the wafer 4, there are lines or slits arranged at an angle of 45 degrees with respect to the scanning line 60 as indicated by numbers 71, 72, 73, and 74 in FIG. Alignment marks are provided. Normally, positioning is performed using signals from both the left and right observation systems.

Note that the alignment marks on the mask 1 and the alignment marks on the wafer are arranged so that when a system other than the same-magnification projection system is used, the dimensions of both alignment marks do not change even if the projection or back projection is performed. The dimensions are set to be different, and in this case, settings are made such that when the dimension of the alignment mark on the mask is removed from the dimension of the alignment mark on the wafer, the reduction magnification is obtained.

Returning to FIG. 2, 22 is a laser light source, and 23 is an optical deflector such as an acousto-optic device. Optical deflector 23
switches the light emission direction upward, horizontally, or downward in response to a switching signal from the outside. 24 and 25 are convergent cylindrical lenses, which are arranged so that their generating lines are perpendicular to each other, and have the function of converting the cross-sectional shape of the laser beam into a linear shape. 26 and 27 are trapezoidal prisms, which have the function of refracting the light deflected upward and downward by the optical deflector 23 in opposite directions. 28
is a rotating polygon mirror (polygon) that rotates around the rotation axis 29.

Depending on the state of the optical deflector 23, the laser beam 30 emitted from the laser light source 22 becomes a slit-shaped beam 30a passing through the cylindrical lens 24 and the prism 26, a slit-shaped beam 30b passing through the cylindrical lens 25 and the prism 27, or The spot-shaped light ray 30c traveling straight takes one of the optical paths, but in each case it converges on a point 31 on the surface of the rotating polygon mirror 28. 32, 33, 34 are intermediate lenses, 35 is an optical path splitting mirror, and 36 is a visual observation system 3.
A half mirror guides light to 7 and 38, 37 is a field lens, and 38 is an eyepiece lens, which transmits an image of the wafer surface. 39 is a visual observation system illumination system 40;
41 is a half mirror, 40 is a condenser lens, and 41 is an illumination lamp. 42 is a half mirror that guides light to photoelectric detection systems 43, 44, 45, and 46, 43 is an optical path bending mirror, 44 is a lens, 45 is a spatial filter, 46 is a condenser lens, and 47 is a photodetector. be. 48,
49, 50, and 51 are total reflection mirrors, 52 is a prism, and 53 is an f-θ objective lens. 54 is the position of an alignment pattern provided on the mask 1 for mask alignment.

As can be seen from Figure 2, the signal detection system consists of completely symmetrical left and right systems.If the operator side is on the front side of the paper, the system shown with a dowel is the signal detection system on the right, and the system without a dowel is on the left. I'll call you.

Intermediate lenses 32, 33, 34 are rotating polygon mirrors 28
The origin of vibration from the aperture position 55 of the objective lens 53
It is formed in the pupil 56 inside. Therefore, the laser beam is scanned over the mask and wafer by the rotation of the rotating polygon mirror 28.

Further, in the objective lens system, an objective lens 53,
The aperture 55, mirror 51, and prism 52 are movable in the XY directions by a moving means (not shown),
The observation and measurement positions of the mask 1 and wafer 4 can be changed arbitrarily. For example, when moving in the X direction, when the mirror 51 moves in the direction indicated by arrow A in the figure, the objective lens 53 and the aperture 55 simultaneously move in the A direction, and furthermore, in order to keep the optical path length constant, the prism 52 also moves in the A direction. The mirror 51 is moved by 1/2 the amount of movement.

On the other hand, when moving in the Y direction, the entire optical system for observation and position detection moves in the Y direction (direction perpendicular to the plane of the paper).

Optical path 30 passing through cylindrical lens 24
The linear scanning beam 61 of a forms an angle θ=45° with respect to the scanning axis 60, and almost marks 71, 72, 7
parallel to 5. When scanning in this state, the photodetectors 47, 47' have S ₇₁ , as shown in FIG. 3B.
Signals S ₇₅ and S ₇₂ are obtained. S ₇₁ , S ₇₅ , and S ₇₂ are signals corresponding to the alignment marks 71, 75, and 72, respectively. Further, even if there is minute dust on the scanning surface, unlike the case of a spot beam, it is averaged out and is not practically detected as an output.

On the other hand, the scanning beam 62 of the optical path 30b passing through the cylindrical lens 25 is inclined at θ=-45° with respect to the scanning axis 60 and is parallel to the marks 73, 74, and 76, so the detection signal is S ₇₃ in FIG. 3b. , S ₇₆ ,
It becomes _S74 . Therefore, the detection signals S ₇₁ , S ₇₅ , S ₇₂ , S ₇₃ ,
By measuring the interval between S ₇₆ and S ₇₄ , the amount of deviation between the mask and the wafer can be detected, and when they match, the intervals between the detection signals become equal.

The present applicant has proposed auto-alignment in Japanese Patent Laid-Open Nos. 53-90872 and 53-91754. Furthermore, in this example, the light beam scanned by the rotating polygon mirror is divided into left and right detection systems in the middle of the optical path, but each detection system may be provided with a light source and a scanner, in which case the entire detection system X direction and Y
It is also possible to move in the direction.

The alignment of the alignment mark 20 on the mask 1 and the alignment mark 21 on the wafer surface, that is, the mask-wafer alignment, has been explained above using FIGS. 2 and 3. Alignment mark 5 on one side of the mask with the optical system and alignment mark of the same shape.
4, the mask reference mark 11 supported by means not shown in the lens barrel 3 can be aligned, that is, mask alignment can be performed. In this case, as described above, the objective lens 53, pupil 55, mirror 51, and prism 52 are moved in the direction A in FIG. 2 to the position of the mask reference mark 11 for alignment.

The patterns for mask alignment are numbered 71, 72, 7 in FIG. 3 for the mask reference mark 11.
The alignment marks 54 of the mask 1 are provided with patterns indicated by numerals 75 and 74 in FIG. Therefore, in mask alignment, the mask and the mask reference mark are aligned in exactly the same way as in the mask-wafer alignment, and the mask is set at a predetermined position with respect to the lens 3.

In the case of mask alignment, since the reduction projection lens 3 is not used, the alignment mark 54 and the mask reference mark 11 on the mask surface may be patterns with the same magnification.

By the way, stepper-type printing equipment uses a reduction projection optical system, so the mask stage (Figure 1, 2) that holds the mask is usually moved for positioning during wafer alignment. . On the other hand, also in mask alignment, the mask stage moves to perform positioning. Therefore, in any case, the two-line patterns 75 and 76 are provided on the moving mask side, but this is only one embodiment of the present invention, and the four-line patterns 71 and 76 are provided on the mask side. 72,7
3, 74 may be provided, or the number of patterns on the mask may be different between mask alignment and wafer alignment, and there is no limitation on the selection of patterns.

Next, the mask alignment method will be explained in more detail. During mask alignment, the mask is transported from the mask change mechanism to the mask stage and held there, but the positioning at this time is done mechanically, so the precision is low and the number of
An error of about 100 μm will occur.

Therefore, when the field of view of the objective lens is narrowed, the mask alignment mark is not necessarily located within the field of view, that is, within the scanning range of the laser beam. The alignment mark on the mask may be out of view due to an error when setting the mask. In that case, the signal in Figure 3B
Since only S ₇₁ , S ₇₂ , S ₇₃ , and S ₇₄ are detected, and the signals of S ₇₅ and S ₇₆ are not detected, auto-alignment is not realized.

In such cases, a process of searching for marks on the mask is required, but in conventional methods, the stage holding the wafer or mask moves and searches for the marks by repeating what is called a groping drive. came.

In order to make the following explanation easier to understand, an example of a conventional groping drive will be explained with reference to FIG. 4A.

The figure depicts the locus of the objective lens when searching for an alignment mark on a mask or wafer surface. However, in reality, instead of moving the objective lens in the X and Y directions, it is preferable to move the stage holding the mask or wafer, but for convenience of explanation, the objective lens is moved here.

It is assumed that the mask has been transported to position A (FIG. 4), which is approximately below the objective lens 53, by a well-known transport means (position A is referred to as the initial setting position). However, if there is an error in the mask settings and the length L and width T (see Figure 3) of the beam scanning range are not large enough to include the setting error, the mark M will not be detected at position A. Therefore, the objective lens moves to the boundary of the groping driving range, starts groping driving from point B, and searches for the mark while moving from B→C→D→E, and so on. In this example, J of the third round trip
-Detect the mark M in the K process.

The following difficulties can be pointed out in the conventional example described above.

(1) Even though the probability of a mark existing is high near point A, the operator moves to point B, a boundary area where the probability of its existence is low, and starts searching from point B, so it takes a long time to detect the mark. .

(2) Since the stage is generally driven at low speed and with high precision, when searching for an alignment mark for alignment of a mask, etc., it takes time to move the stage and the detection time increases.

In the embodiment described later, the following configuration is adopted, so that the mark detection time can be shortened.

(1) Search from the periphery of the initial setting position of the objective lens toward the boundary area.

(2) More preferably, the stage is fixed during the groping drive, and the objective lens 53 and the aperture 55 are arranged in the X direction.
Then, the prism 52 constituting the optical trombone is moved, and the entire scanning optical system 22 to 55 is moved in the Y direction. Generally, the stage is required to have the most precise positioning, so the transport mechanism must have a structure that meets this requirement. Therefore, it is not suitable for moving the stage at high speed.

FIG. 4B shows the locus of the groping drive according to the embodiment of the present invention. Here, the objective lens system moves from the initial setting position A on the mask in a spiral pattern from A→P→Q→R→S→..., searching for marks, and in this example, from U to V.
Mark M is detected on the route to . And since the probability of a mark existing is generally high near point A,
This has the advantage that the time required for detection is short.
It is assumed that the objective lens system is always set at the same position at the starting point of operation.

FIG. 5 shows a flow diagram showing movement of the objective lens system. When the groping drive starts at 501, the X-direction movement counter and the Y-direction movement counter are first cleared at 502. These two counters are counters that determine the amount of movement of the objective lens in the X and Y directions. (See Figure 4B for the X and Y directions.
). Next, in step 503, a limit check is performed to determine whether the limit in the X direction of the groping drive has been reached, and if it is within the limit, the content N of the X direction movement counter is incremented in step 504.

If the limit is reached after going through this loop several times as described below, skip 504 and move from 503 to 505. At 505, movement is performed in the positive direction of X, and the amount of movement at this time is, for example, T×N. (T is the laser scanning width as shown in FIG. 3) Here, in the first loop, the movement from A to P shown in FIG. 4B is performed.

Once a predetermined amount of movement has been performed in the X direction, then
Perform a limit check in the (-)Y direction at 506,
Similar to the X direction, if it is within the limit, the content M of the Y direction movement counter is incremented at 507, and if the limit has been reached, the process at 507 is jumped and the process is returned to 508.
I do. 508 moves in the (-)Y direction by an amount of L×M. (L is the laser scan length shown in FIG. 3).

Similarly, processes 509, 510, and 511 are performed in the (-)X direction, and then processes 512, 513, and 514 are performed in the Y direction.
Next, it is checked in 515 whether or not the limit of the groping drive has been reached in both the X and Y directions. Repeat this loop until you reach the limit.

Here, the amount of movement in the X direction and Y direction can be explained using the symbols in Fig. 4B.
is three times...and twice as much as or,
It is three times... Therefore, the amount increases once in each of the X and Y directions, and this amount is counted by the X direction movement counter and the Y direction movement counter. The amount of the curve is equal to the laser scan width T and the laser scan length L, respectively.

The situation in which one of the directions in Figure 4 B reaches the limit is, for example, taking the (-) Y direction as an example.
In the state shown as W'→Z', the amount of movement is the same as the previous movement W→Z. This corresponds to passing the process 507 in the flow diagram of FIG. 5 and performing the process flow 506→507.

Note that the example explained here is a method of searching in which the two alignment patterns 75 and 76 on the mask are always detected. For example, after first detecting one of the above-mentioned patterns, If a book detection method is used, the pitch of the groping drive may be rough. That is, and may be longer than T and L, respectively, and therefore the amount of movement of the groping drive is not particularly limited to the amount described above.

Next, the interlocking and independent driving of the right objective lens system and the left objective lens system, which are other advantages of the present invention, will be described.

The explanation of the operation in Figures 3 and 4 was given for one objective lens system, but in reality, as shown in Figure 2, the alignment mark groping drive is performed using both left and right objective lens systems. . In this case, there are two modes: a mode in which the objective lens systems on both sides are driven in conjunction with each other, and a mode in which they are driven independently, and a desirable one of the two modes can be selected.

For example, in the linked drive mode, both the left and right objective lens systems perform the drive trajectory shown in FIG. 4B. In the case of this interlocking drive, the time to detect the alignment mark is, on average, longer than in the case of independent drive, which will be described later.
Although it is a little longer, the operation of the objective lens system and its control are linked, so the apparatus itself is simple.

On the other hand, in the case of independent drive, for example, one objective lens system is driven clockwise as shown in Figure 4B, and the other objective lens system is driven counterclockwise as shown in Figure 4C. Although the operation is slightly more complicated than interlocking drive, alignment and
The average time to detect marks becomes shorter.

The present invention resides in that the interlocking drive mode and the independent drive mode can be selected as necessary. An example of combining these methods is, for example, first performing interlocking driving near the initial position of the objective lens system, and performing independent driving toward the periphery.

Another example is a method in which interlocking driving is performed until one pattern of alignment marks is detected, and then independent driving is performed after detection, so that both objective lens systems detect four patterns. .
This method is particularly effective when the driving pitch of the aforementioned groping drive is rough or when there is a deviation in the θ direction of the mask.

Next, we will discuss the selection and use of the scan beam shape. As explained in FIG. 2, it is also possible to use the optical path 30c in which the laser beam 30 does not pass through the cylindrical lenses 24 and 25 and travels straight as it is. In this case, the illumination area by the beam becomes a so-called spot shape and does not have a tilt characteristic like a slit beam, so the patterns 71, 72,
Signals can be obtained for any of 73, 74, 75, and 76. In particular, this method is effective in mask alignment with sufficient light intensity and a good S/N ratio, but it can also be used in mask-wafer alignment without being limited thereto. Alternatively, for example, in mask alignment, a spot beam is used until a total of 6 beams are detected, such as 4 beams on the mask or 4 beams on the mask and 2 reference marks, and then a slit beam is used during scanning. You may also use a method of switching. By doing so, complicated control is not required, and highly accurate alignment can be achieved using the sheet-shaped beam. In other words, the advantage of this method is that it provides a structure that generates both sheet-like beams and spot-like beams, and that they can be selected depending on the object.

Next, the positioning of the mask after the objective lens is groped will be described using FIG. 4B. When the mask alignment mark on the mask is detected by the groping drive as described above, the objective lens position M can be easily determined from, for example, the amount of movement of the reference mark from the position A. Therefore, the objective lens system (52 to 5
3), and mask stage 2 is also changed from M to A.
move in the direction. When the movement is complete, both the mask reference mark and the mask alignment mark can be observed within the field of view of the objective lens, resulting in the state shown in FIG. 3A. In this state, six alignment signals can be detected, so the positional relationship between the mask reference mark and the mask alignment mark can be detected and aligned by measuring the intervals between the six signals.

Next, the detection device for television wafer alignment will be explained using FIG. 6.

The reduction projection lens 3, wafer 4, objective lens 6, and image pickup tube 7 in the figure are the same as in FIG. On the other hand,
91 is a light source for illumination, and uses a halogen lamp, for example. 92 is a condenser lens. 93A and 9
3B is a bright field diaphragm and a dark field diaphragm that can be attached and detached interchangeably, and in the figure, a bright field diaphragm 93A is installed in the optical path. Condenser lens 92 images light source 11 onto a bright field aperture. 94 is a relay lens for illumination, and 95 is a cemented prism. The cemented prism 95 has a function of making the optical axis of the illumination system and the optical axis of the light receiving system coaxial, and includes an inner reflective surface 95a and a semi-transparent reflective surface 95b. . Here, a light source 91, a condenser lens 9
2. Bright or dark field apertures 93A and 93B, illumination relay lens 94, cemented prism 95, objective lens 6
constitutes an illumination system, and the chief ray emitted from the objective lens 6 enters the wafer 6 perpendicularly.

Next, 96 is a relay lens, and 97 is a mirror that bends the optical path. Reference numeral 98 denotes a glass plate having a reference mark for TV/wafer alignment, and the reference mark has the function of providing the origin of coordinates, so to speak. Therefore, the wafer alignment mark is detected as an X coordinate value and a Y coordinate value. 99 is an imaging lens, which includes the above-mentioned cemented lens 95 and relay lens 9.
6, a mirror 97, a glass plate 98, an imaging lens 99, and an imaging tube 7 constitute a light receiving system, and an objective lens 6
The optical path passing through is reflected by the inner reflective surface 95a of the cemented prism, is reflected by the semi-transparent surface 95b, is reflected again by the inner reflective surface 15a, and is directed toward the relay lens 96. The wafer alignment mark image on the wafer 4 is formed on a glass plate 98 having a reference mark, and then is imaged on the imaging surface of the image pickup tube 7 together with the reference mark image.

Next, the action will be explained. The light flux from the illumination light source 91 is converged by the condenser lens 12, illuminates the aperture of the bright field diaphragm 93A or the dark field diaphragm 93B, further passes through the illumination relay lens 94, and is transmitted through the semi-transparent surface 95b of the cemented prism. Reflective surface 95a
, and illuminates the wafer 4 through the objective lens 6.

The light beam reflected on the surface of the wafer 4 is subjected to an imaging action by the objective lens 6, enters the cemented prism 15, is reflected by the reflective surface 95a, and is then reflected by the semi-transparent surface 95.
b. It is reflected by the reflective surface 15a and emitted, relayed by the relay lens 96, reflected by the mirror 97, and imaged on the glass plate 98, and then imaged on the imaging tube 7 by the imaging lens 99. At that time, as described above, the reference mark on the glass plate 98 is imaged with the bright field diaphragm 93A inserted, and the origin of the coordinates is determined using that image, and then the wafer alignment mark image is clearly seen by switching to the dark field state. The position of the wafer alignment mark image is detected by imaging it. Then, in accordance with the position of the wafer alignment mark detected by electrical processing, the wafer stage 5 moves and stops so that the wafer 4 occupies a prescribed position 4' in the projection field of the projection lens 3.

Note that the wafer 4 may be deformed so that it is once aligned at the standard position and then moved into the projection field.

The present invention described above can shorten the search time because the stage that supports the mask is fixed when detecting the alignment mark on the mask, and the alignment mark and the reference mark can be brought close together before fine alignment. The operation will be short-lived. In particular, when this alignment is achieved using a mask and wafer alignment detection device, alignment can be achieved with extremely high precision.

[Brief explanation of the drawing]

FIG. 1 is a perspective view showing an apparatus according to an embodiment of the present invention. FIG. 2 is a side view of the optical system according to the embodiment. FIG. 3A is a plan view of the alignment mark, and FIG. 3B is a diagram of an example of an output signal when scanning the alignment mark. FIG. 4A is a plan view for explaining the operation of the conventional example.
FIGS. 4B and 4C are plan views for explaining the operation of the embodiment. FIG. 5 is a flow diagram showing the operation of the groping drive.
FIG. 6 is a perspective view of the wafer alignment system. In the figure, 1 is a mask, 2 is a mask stage, and 3
is a reduction projection lens, 4 is a wafer, 5 is a wafer stage, 6 is an objective lens, 7 is an image pickup tube, 8
is a television receiver, 11 is a fixed mask alignment mark, 22 is a laser light source, 23 is a light deflector, 24 and 25 are cylindrical lenses, 26 and 27 are prisms, 28 is a rotating polygon mirror, 53 is an objective lens, 52 is prism, 71, 72, 73, 7
4 is an element constituting a wafer-side alignment mark, 75 and 76 are elements constituting a mask-side alignment mark, A is an initial setting position, and M is an alignment mark.

Claims (1)

[Claims] 1 The positional relationship between the reference mark provided on the printing device and the alignment mark provided on the mask is detected by a position detection optical system, and the mask stage that supports the mask is driven based on the detected positional relationship. , a method of setting a mask in a predetermined positional relationship in a printing device, in which the position detection optical system is moved on the mask surface by a moving means to detect an alignment mark provided on the mask; a first step of determining the position of the alignment mark with respect to the reference mark based on the position of the optical system; and based on the alignment mark position, driving the mask stage to move the mask and moving the position detection optical system; a second step of bringing the mark and the alignment mark within the detection range of the position detection optical system; and detecting the positional relationship between the reference mark and the alignment mark within the detection range of the position detection optical system to form a mask. and a third step of setting the mask in a predetermined positional relationship on the printing device.