JPH04302131A - Projection exposure system - Google Patents

Abstract

PURPOSE:To continue the servo-alignment of a Z stage while collecting the data on the shifting range and the velosity characteristics of the Z stage previ ously set up to increase the focal length as the offset of an AF sensor in the focus servo state based upon the detection signal from a surface position detec tor (AF sensor). CONSTITUTION:A halving 12 is controlled to be tilted having the previously specified velosity characteristics during the exposure step with the closed loop (feed back) of an autofocus (AF) mechanism in the servo state (hereinafter to be referred to as 'focus servo state'). When the halving 12 is thus tilted in the focus servo state, the focus offset amount is follow-fluctuated corresponding to the tilting degree. Resultantly, a Z stage 20 is follow-shifted in the optic axial direction corresponding to the tilting degree of the halving 12. That is, the AF mechanism can constantly and continuously tracts the surface of a wafer W as a focal point.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a projection exposure apparatus that enables an exposure method that increases the apparent depth of focus of a projection optical system when projecting and exposing a minute pattern onto a photosensitive substrate.

0002

[Conventional technology]

Conventionally, as this type of projection exposure apparatus, one disclosed in, for example, Japanese Patent Application Laid-Open No. 63-64037 is known. This publication describes that the resist layer formed on the surface of the semiconductor wafer and the image plane of the projection optical system that projects the reticle pattern are moved relative to each other in the optical axis direction and multiple exposure is performed, thereby developing the resist layer. It is disclosed that the image has an image quality as if it were transferred through a projection optical system with a large depth of focus. In this case, in the apparatus disclosed in the above publication, after performing the first exposure, the Z stage that holds the wafer is moved by a certain amount in the optical axis direction, and then the second exposure is performed using the same reticle pattern. ing. Furthermore, it is shown that the step position of the Z stage in the optical axis direction may be two or more points.

[0003] Similarly, as a method to obtain the effect of enlarging the apparent depth of focus, Japanese Patent Application Laid-Open No. 58-1744
As disclosed in Japanese Patent No. 6, a method of vibrating a wafer in the optical axis direction during one shot exposure operation on the wafer is also known. In this case, although the waveform of the vibration is not specified in the above publication, considering that it is a general sine wave vibration, by setting the vibration amplitude and frequency to appropriate values, it is possible to The effect of expanding the depth of focus can be obtained in exactly the same way as in the publication.

0004

However, among the above two conventional techniques, in the method of performing exposure at each of a plurality of points separated from each other in the optical axis direction, it is difficult to adjust the shutter speed to complete one shot of exposure on the wafer. This increases the number of opening/closing operations, which inevitably causes a problem of lower throughput. Furthermore, in the method of exposing the wafer while vibrating the wafer, the shutter only needs to be opened and closed once, but instead it is necessary to set the number of vibration cycles during the opening time of the shutter to be exactly an integral multiple. Furthermore, in the above two conventional technologies, if the vibration method is used, it is difficult to accurately set the center of vibration at which position along the optical axis direction, and if it is a method that exposes at each discrete position in the optical axis direction, the discrete There is also no clear disclosure as to how the position of the surface is guaranteed relative to the wafer surface. For this reason, the conventional technology is extremely unsatisfactory in terms of control accuracy in the optical axis direction (Z direction) and throughput.

The present invention has been made based on these requirements, and aims to provide a projection exposure apparatus that not only improves throughput but also significantly improves position control accuracy, reproducibility, etc. in the optical axis direction. purpose.

[0006]

[Means for Solving the Problems] In order to solve the above-mentioned problems, in the present invention, the pattern of the mask (R) is formed on the photosensitive substrate (
A projection optical system (PL) that forms and projects an image onto a photosensitive substrate (
A Z stage (20) that can hold the surface of the photosensitive substrate (W) and move in the optical axis direction of the projection optical system (PL), a motor (19) that drives the Z stage, and a P
surface position detection means (1 to 11, 14 to 17) that detects the amount of relative positional deviation in the optical axis direction with respect to the predetermined image forming plane of L) and outputs a detection signal according to the amount of positional deviation; control means (18) for servo-controlling a motor (19) so that the detection signal from the surface position detection means coincides with a predetermined value; In a projection exposure apparatus that exposes a pattern onto a photosensitive substrate (W) by setting a predetermined relationship between
, 14 to 17), the value of the detection signal from the photosensitive substrate (W)
correction means (12, 13) for correcting the surface position detection means so as to include a predetermined offset amount (MS) with respect to the actual positional deviation amount between the surface of the pattern and the predetermined image forming plane; storage means (304) for storing data for changing the offset amount (MS) over a predetermined range with predetermined speed characteristics between the start and end of the exposure operation; At the start, storage means (3
04) to the correction means (12, 13), and during the exposure operation, the Z stage (19) is servo-controlled by the control means (18). It was decided to move it in the axial direction.

[0007]

[Operation] The present invention controls the offset of the focus sensor that detects the surface position of the substrate, and while the Z stage is always driven to the focus zero point, the Z stage is moved by sequentially changing the focus offset during the exposure operation. The technical element is to continuously expose the substrate at different positions to improve accuracy and expand the Z movement range.

[0008]

[Embodiment] Figure 1 shows an outline of a projection exposure apparatus used in the present invention. Emits illumination light with uniform illuminance. A 1/5 or 1/10 reduction projection lens PL projects and exposes the pattern of the reticle R onto the wafer W under telecentric conditions on both sides. A photosensitive layer (photoresist) (not shown) is applied to the wafer W, and the wafer W is attracted onto a Z stage 20 including a wafer holder. This Z stage is provided on the XY stage 21 so as to be able to move slightly in the optical axis AX direction (Z direction) of the projection lens PL. Z stage 2
0 is normally used to align the surface of the wafer W with the correct focus plane (the conjugate plane of the projection lens PL with the reticle R), and the XY stage 21 aligns the projected image of the reticle R and the wafer W with the optical axis. Two-dimensional movement is performed for positioning within a plane perpendicular to AX. This Z stage 20 moves at a resolution of, for example, 0.1 μm or less so that the surface of the wafer W can be positioned with higher accuracy within the effective focal range that decreases as the resolution (numerical aperture N.A.) of the projection lens PL improves. controlled.

It is also known that the optical performance of the projection lens PL, such as the focus position and magnification (distortion), varies depending on the atmospheric pressure and temperature in the usage environment, the exposure power during wafer processing, the amount of irradiation light, etc. . Therefore, in the apparatus of this embodiment, the information S11 regarding these environmental conditions and exposure conditions is
a fluctuation amount monitor 100 that calculates the fluctuation in optical performance of the projection lens PL by inputting the following information, and a pressure regulator 1 that adjusts the air pressure in the projection lens PL according to the calculated fluctuation amount.
02 is provided to correct variations in optical performance.

FIG. 2 shows an oblique incident light type automatic focusing mechanism provided in the projection exposure apparatus of FIG. 1, and FIG. 3 is a block diagram summarizing the hardware and software mechanisms of the control system in the apparatus of FIG. It is a diagram. First, an automatic focusing (hereinafter referred to as AF) mechanism will be described based on FIG. 2. The illumination light IL, which is insensitive to the resist layer on the wafer W and has a wide wavelength range, illuminates the slit plate 1 uniformly. The light passing through the slit of the slit plate 1 passes through the lens system 2,
An imaging light beam obliquely enters the wafer W via the mirror 3, the aperture 4, the projection lens 5, and the mirror 6. As a result, a one-dimensional slit image is formed at a position on the wafer W through which the optical axis AX of the projection lens PL passes, that is, at the center of the projection area of the pattern image of the reticle R. The reflected light reflected by the wafer W passes through the mirror 7, the objective lens 8, the relay lens 9, the vibrating mirror 10, and the parallel flat glass 12, and forms an image on the slit plate 14 for light reception. That is, when the wafer W comes to a predetermined position (focusing position) in the optical axis direction, the three points of the light transmitting slit plate 1, the wafer W, and the light receiving slit plate 14 are set to be conjugate with each other. has been done. Further, the above-described system from the light traveling slit plate 1 to the light receiving slit plate 14 is arranged without slight movement with respect to the projection lens PL. Here, assuming that the wafer W matches the best image forming plane of the projection lens PL (is in focus), the slit image on the light receiving slit plate 14 is formed on the light receiving slit plate by the action of the vibrating mirror 10. It will vibrate around 14 slits. When the wafer W deviates from the focused position, the center of vibration of the slit image on the slit plate 14 is displaced from the slit of the slit plate 14. A photomultiplier (PMT) 15 photoelectrically detects the amount of light passing through the light receiving slit plate 14, and the photoelectric signal is output to a synchronous detection circuit (hereinafter referred to as PSD) 17. Further, the oscillator (OSC.) 16 outputs an AC drive signal of frequency f to the actuator (M-DRV) 11 that drives the vibrating mirror 10, and outputs a reference signal of frequency f to the PSD 17. PSD17 is OSC.
The reference signal from PMT 16 and the photoelectric signal from PMT 15 are input, and the photoelectric signal is synchronously detected with respect to the reference signal.

This synchronous detection is the same as that of a conventionally well-known photoelectron microscope, and when the center of vibration of the slit image vibrating on the slit plate 14 coincides with the center of the slit of the slit plate 14, The photoelectric signal of PMT15 is OSC. The frequency (2f) is exactly twice the frequency f of the oscillation signal of PSD 16, and the detection output of PSD 17 becomes zero. When it deviates from this state, the PSD 17 moves in the direction of the deviation.
That is, a detection output having a different polarity is generated depending on which direction the position of the wafer W is displaced with respect to the in-focus position.

Therefore, the detection output of the PSD 17 is a continuous voltage change that is zero in the focused state, becomes positive when the wafer W approaches the projection lens PL, and becomes negative when the wafer W moves away from the projection lens PL. shows. The detection output is usually also called an S-curve signal, and there is a range near the zero point in which the relationship between voltage change and positional change in the Z direction of the wafer W is almost linear, and this range is used for automatic focusing. Used for servo control.

Now, the Z stage 20 is the XY stage 2.
A drive circuit (Z-DRV) for the motor 19 provided on the
It is moved by servo control by 18. Z-D
The RV 18 inputs the detected output from the PSD 17 as deviation information, and drives the motor 19 so that the detected output becomes zero. As described above, the vibrating mirror 10, M-DRV11, slit plate 14, PMT15, OSC. 16, PSD17
, Z-DRV 18, motor 19, and Z stage 20 constitute a closed-loop control system for automatic focusing.

By the way, the parallel flat glass (hereinafter referred to as harbing) 12 is tilted so as to shift the reflected light flux from the wafer W in the vibration direction of the slit image on the slit plate 14, and includes a motor and an encoder. Drive part(
D-DRV) 13. The harving 12 works as an offset mechanism that shifts the surface to be focused by a predetermined amount in the direction of the optical axis AX, and how much focus offset to give is determined by the main control unit (MCU).
It is controlled by the signal DS from 30. Furthermore, MCU3
0 inputs the encoder signal ES that monitors the amount of inclination of the harving 12 provided in the H-DRV 13, compares whether the amount of focus offset monitored by this signal ES has reached the target value, and A driving signal DS is output to the H-DRV 13 according to the difference.

Optical focus offset setting using the harving 12 allows offset to be applied over a much wider range than the method of adding an electrical offset to the detection output of the PSD 17. Normally, Herbing 1
2 is used for various focus offsets, and one is a process offset arbitrarily determined by the operator. Other equipment-specific offsets are shown in Figure 1.
Information S12 calculated by the fluctuation amount monitor 100 shown in
Of these, information f2 (t) regarding focus fluctuations and information regarding atmospheric pressure changes may be input.

Now, in the embodiment of the present invention, the projection lens P
As a method of increasing the apparent depth of focus of L, the Z stage 20 is moved along the optical axis at a predetermined speed characteristic during one exposure operation (while the shutter is open), as in the previous Japanese Patent Application Laid-Open No. 58-17446. Continuously move in the direction. However, the difference from the above publication is that the Z stage 20 moves only in one direction from bottom to top or from top to bottom in one exposure operation. This is to maximize throughput.

Furthermore, in the embodiment of the present invention, the above-mentioned A
In a state in which the closed loop control (feedback) of the F mechanism is activated (hereinafter referred to as a focus servo state), the harving 12 is controlled to be tilted at a predetermined speed characteristic during the exposure operation. When the harving 12 is tilted in the focus servo state in this way, the focus offset amount changes accordingly, and as a result, the Z stage 20 moves in the optical axis direction in accordance with the amount of inclination of the harving 12. In other words, since the AF mechanism in FIG. 2 always captures the surface of the wafer W (or the internal surface of the resist layer) as the focal point, the amount of offset with respect to the best imaging plane of the projection lens PL is determined from the encoder information ES. It can be determined accurately and extremely precise control is possible. Such control of the harving 12 is performed by information f1 (t) added to the MCU 30 in FIG. Therefore, the functions of each unit inside the MCU 30 will be explained below with reference to FIG.

In FIG. 3, a basic data storage section (memory) 33 is externally connected to the MCU 30, and standardized basic data as shown in FIG. 4 is stored here. The horizontal axis of FIG. 4 represents the memory address of the storage unit 33, and the vertical axis represents the drive characteristics of the harving 12 (or the position change characteristics of the Z stage 20 in the optical axis direction) determined in advance through experiments etc.
represents the level corresponding to In this embodiment, it is assumed that basic data is stored in addresses AD1 to AD4 of the storage unit 33, and as shown in FIG. It is set to have a curved waveform. That is, the rate of change is made small in both the vicinity of the address AD1 and the vicinity of the address AD4, and the rate of change is made large in the vicinity of the central address ADc. and the start point (or end point)
When the numerical value of the address AD1 is K2, and the numerical value of the end point (or starting point) address AD4 is K1, the setting is made so that (K1 + K2)/2=K0. Then, the data in FIG. 4 is the information f1 mentioned earlier.
(t) Then, the drive data creation section 3 in the MCU 30
04 or is sent to the correction data creation unit 302. MCU
A timer 300 in the storage section 30 changes the address of the storage section 33 to AD1 during one exposure operation time determined in advance.
It has a clock counter that sequentially accesses from AD4 to AD4. Further, the clock count of the timer 300 starts at the same time as the exposure starts (when the shutter starts opening). The correction data creation unit 302 performs a correction calculation on the level of the basic data shown in FIG. 4 with respect to one movement width of the Z stage 20 set by the operator. The calculation result is stored in the correction data storage section 306 in a format similar to that shown in FIG. In addition, timer 300
is the value counted during one exposure time at the address AD.
Since it is necessary to uniquely correspond to the number of 1 to AD4, a programmable device (hardware counter,
or a software counter).

As another data correction method, only the correction amount (correction multiplier, etc.) for the basic data may be calculated by the correction data creation section 302 and stored in the correction data storage section 306. In this case, it is necessary to correct the shape of the curve of the basic data (Figure 4) to make it thinner, so address AD1 ~
A correction multiplier corresponding to each basic data level of AD4 is created in the address space in the storage unit 306.

[0020] The drive data creation unit 304 generates information (offset amount) corresponding to the change in focus among the fluctuations in the optical characteristics described above.
f2 (t) and the data stored in the correction data storage unit 306 are read, addition and subtraction are performed, and the results are sequentially output to the digital-to-analog converter (DAC) 308. This output operation is performed in synchronization with the clock counter of the timer 300, and the DAC 308 outputs the voltage MS corresponding to the temporal position change of the Z stage 20 in the optical axis direction to the servo amplifier 310 as a target value. This servo amplifier 3
The output of 10 becomes the aforementioned signal DS and is supplied to the motor in the H-DRV 13. Further, the signal ES from the encoder in the H-DRV 13 is counted up (or counted down) by an up/down counter 312, and the value is converted into an analog voltage in real time by a DAC 314, and is applied to the feedback input of the servo amplifier 310 as a deviation signal. Ru. With this, Harbing 12
Servo amplifier 310, encoder, counter 312,
The amount of inclination is sequentially changed by a closed loop control system using a DAC 314 or the like. Note that when only the correction multiplier is stored in the correction data storage section 306, the drive data creation section 304 simultaneously reads the basic data f1 (t) and the correction multiplier from the storage section 306 in synchronization with the timer 300. After multiplication, the multiplied value is given information about focus fluctuation f2 (t)
The result of adding (or subtracting) is output to the DAC 308.

Next, the operation of this embodiment will be explained with reference to FIGS. 5, 6, and 7.
As will be explained with reference to FIG. 7, the final required staying probability in the optical axis direction in this embodiment is as shown in FIG. The staying probability means how long it takes Δt to pass through a minute distance ΔZ in the optical axis direction when the Z stage 20 moves in the optical axis direction, and is equivalent to the mathematical differential value dt/dz. . Therefore, the staying probability when the Z stage 20 moves at a constant speed is a constant value. In FIG. 7, the horizontal axis represents the magnitude of the stay probability, and the vertical axis represents the position in the optical axis direction (Z). In this figure 7, Z
The midpoint +ΔZf in the direction corresponds to the midpoint of the movement range of the Z stage 20, that is, the level K0 (address ADc) in FIG. The movement range of the Z stage 20 is the position + in FIG.
It is between ΔZ1 and position -ΔZ2, and the midpoint +Δ
The local maximum values of the stay probability exist at two points separated from Zf by ±ΔD in the Z direction. It is desirable that the maximum values at these two locations are almost the same, and if the distance ±ΔD is set to the same level as or greater than the depth of focus (±D0f) of the projection lens PL itself, the appearance will be the same as that of the multifocal exposure method. Experiments have confirmed that the depth above can be expanded.

Now, in order to obtain the distribution shown in FIG. 7, it is necessary to obtain in advance the opening/closing characteristics of the shutter during the exposure operation and the total exposure time per shot. These data can be easily obtained by using the exposure amount monitor (light amount integrator) provided in the projection exposure apparatus shown in FIG. You can ask for it.

The illuminance characteristic IP shown in FIG.
The shutter begins to open at time 1, fully opens at time T2, starts closing at time T3, and closes at time T4.
The shutter closes and the exposure for one shot ends. The characteristics at the next time T5 to T8 mean similar exposure to another shot on the wafer W.

[0024] Considering the characteristics of such a shutter,
The reading of the basic data f1 (t) or corrected equivalent data is synchronized as shown in FIG. 5(A). That is, if time (T2 - T1) and time (T4 - T3) are equal in FIG. 5(C), data f1 (
The timing at which the address ADc at the center level K0 in t) is accessed is set to the midpoint between time T1 (or T2) and time T4 (or T3). Therefore,
Time T1 is made to match the access timing of address AD1 in FIG. 4, and time T4 is made to match the access timing of address AD4. Furthermore, time T2 is made to match the access timing of address AD2 in FIG. 4 (the point at which the rate of change of data begins to increase with time),
Time T3 is made to coincide with the access timing of address AD3 in FIG. 4 (the point at which the rate of change of data begins to decrease with time).

FIG. 5A corresponds to a change in the position of the wafer surface in the Z direction obtained only by the basic data f1 (t), and 0 point means the best focus position. However, the actual best focus position is determined by the offset amount +ΔZf according to the information f2 (t), as shown in FIG. 5(B).
This is the position where . Although the offset amount +ΔZf is shown here as being constant, it may change from moment to moment (however, it may change much more slowly than the change in the data f1 (t)). The drive data creation unit 304 generates the data in FIG.
Add the characteristics in (B) to create data as shown in Figure 6. As a result, the Z stage 20 moves vertically by a symmetrical amount with the focus offset amount +ΔZf as the midpoint. In addition, in FIG. 6, the data f1 (t) and f2 (t) are directly added to form the signal MS, but the data f1 (t)
) is corrected as necessary as mentioned above. Further, time Tc in FIG. 6 corresponds to the access timing of address ADc in FIG.

When one shot of exposure is thus completed between time T1 and time T4, the XY stage 21 steps for exposure of the next shot. Time T4
to T5 is the stepping time. Now, for the next shot, the Z stage 20 can be moved in the opposite direction from the previous shot by accessing the address of each storage unit 33, 306 in the reverse direction from AD4 to AD1. Can be done.

If each shot is exposed as described above, the shutter only needs to be opened and closed once for each shot, and the Z stage 20 only needs to be moved once in one direction, which is faster than the conventional exposure method. Throughput is also significantly increased. Moreover, since the AF mechanism constantly monitors the surface of the moving wafer, the position of the wafer W in the optical axis direction can be controlled extremely accurately. In addition, Fig. 5 (A
), or the characteristics shown in FIG. 6 are uniquely related to the speed characteristics of the Z-direction movement of the Z stage 20,
The speed characteristic at 0 has a trapezoidal shape similar to the illuminance change in FIG. 5(c).

As described above, in the embodiment of the present invention, Harbing 1
2 is servo-controlled in response to the signal MS, and the existing AF mechanism is also servo-controlled using the zero method, so even if the amount of movement of the Z stage 20 is relatively large,
The system has high stability and can perform highly accurate position control. As mentioned above, in addition to the method of adding an offset using the slope of the harving 12, there is also a method of electrically adding an offset to the detection output (S curve signal) of the PSD 17, but this causes problems as described below. Figure 8 shows P
In the waveform example of the S-curve signal of SD17, the horizontal axis represents the deviation amount ΔF of the wafer W from the in-focus point (zero point), and the vertical axis represents the signal voltage V. The S-curve characteristic PV1 shown by the solid line in Figure 8 is obtained in an ideal state with no offset, and the signal voltage is approximately linear within a range of ±ΔF1 that is approximately symmetrical in the optical axis direction across the zero crossing point. ing. This ±ΔF1
The range becomes the range where servo is possible. Generally, when using an S-curve signal obtained by synchronous detection in a servo system, the accuracy of tracking to the zero cross point will increase as the sensitivity of the servo system improves if the voltage change within the range of ±ΔF1 becomes steeper than a certain level. Become. However, making the voltage change in the linear region of ±ΔF1 steeper means that the servo possible range 2
This leads to narrowing ΔF1. This servo-enabled range 2ΔF1 is determined to such an extent that the sensitivity does not decrease while assuming the range of flatness and local unevenness on the wafer W.

Now, for such an S curve signal, ΔV0f
When a voltage offset of is added, the entire S-curve characteristic shifts in the voltage axis direction, as shown by the broken line in FIG. The zero point of the voltage of this broken line S curve characteristic PV2 is -Δ
Shift by ΔOf in the F direction. That is, this ΔOf
corresponds to a focus offset, and the focus servo system positions the wafer surface at a position minus ΔOf from the original zero point. However, as is clear from Figure 8,
The linear region in the -ΔF direction from the voltage zero point of the S-curve characteristic PV2 indicated by the broken line almost disappears. This means that the focus servo system no longer works stably. Therefore, the range to which electrical offset can be applied is limited to an extremely narrow range.

On the other hand, in the optical offset shown in this embodiment, the S curve characteristic PV1 of the solid line is changed to ±Δ
Since it will be moved parallel to the F direction (optical axis direction),
The stability of the focus servo system is not compromised. Therefore, by using the harbing 12, it is possible to apply a stable focus offset over a fairly wide range as long as it does not affect the slit image to be photoelectrically detected.

In the above embodiment, the harving 12 is provided on the light receiving system side of the AF mechanism, but it is provided between the slit plate 1 and the lens system 2 on the projection system side, or between the projection lens 5 and the mirror 6. Exactly the same effect can be obtained even if it is provided between the two. Further, as a method of providing an optical offset, the inclination of the mirror 3 in the light projecting system or the direction of reflection at the vibration center of the vibrating mirror 10 may be slightly changed using a piezo element.

The present invention can also be applied to a static type focus detection circuit other than the synchronous detection type. For example, the vibrating mirror 10 is a fixed mirror, and a slit plate 14 for receiving light is used.
A two-split optical detector is placed at the position of , and the slit image of the slit plate 1 is projected with a width wider than the dividing line (dead zone) of the optical detector, so that the output signals of both elements of the two-split detector are A waveform similar to the S-curve signal can also be obtained by using a differential signal.

FIG. 9 shows a case where the stay probability has a maximum value at three points in the optical axis direction, and step 9 (A) is a characteristic f3 (t
), and FIG. 9(B) shows the distribution of the stay probability obtained thereby. In this case, the three maximum points are 2 in the optical axis direction.
It is preferable to leave an interval of ΔD (2D0f). The position of the central maximum point in the Z direction is set to the best focus plane of +ΔZf. Again, the characteristic f3 (t)
The movement of the Z stage 20 according to the following is one exposure operation time (
(from time T1 to time T4), it is necessary to reduce the speed of the Z stage 20 by one degree at the intermediate time Tc. Regarding this characteristic f3 (t), an optimal curve is also obtained through simulation, experiment, etc., and is stored in the storage unit 33 as basic data.

By the way, since the basic data shown in FIG. 4 is created based on standard wafer exposure conditions, the data may vary depending on differences in shutter opening/closing operating characteristics between devices, extreme exposure time settings, etc. 4 data curve needs to be corrected. Therefore, some modification examples will be shown with reference to FIGS. 10 and 11, but here it is assumed that two points on the stay probability distribution have maximum values. FIG. 10(A) shows the illuminance change on the wafer when the exposure time is extremely short, and FIG. 11(A)
) shows the change in illuminance when the exposure time is long. First of all, short exposure time means shutter full-open time (T3
-T2) is almost non-existent or zero. In this case, the passing speed of the best focus plane ΔZf of the Z stage 20 is set to approximately the maximum speed. Figure 10
(B) shows the relationship between the position of the Z stage 20 in the Z direction and the time t, and shows the relationship between the position of the Z stage 20 in the Z direction and the time t. Set the determined straight line L1. Next, at time T1, a straight line L2 passes approximately the position ΔZ1 and has a slope smaller than the straight line L1, and at time T4, a straight line L2 passes approximately the position ΔZ2 and has a slope smaller than the straight line L1.
3 (actually, the slopes of L2 and L3 are set to be equal). At this time, the shutter opening operation time (T2
-T1) and the closing operation time (T4 -T3) are equal, and each operation time is the fully open time (T3 -T2).
If the length is longer than , then the intersection of straight lines L1 and L2,
and the intersection of straight lines L1 and L3 may both exist outside the fully open time (during the operating time). Z stage 2
The speed of Z stage 20 (Harving 12) is controlled along these straight lines L1, L2, and L3, but in reality, the Z stage 20 (harving 12) starts moving at time Ts, which is earlier than time T1, and completely stops (speed reaches zero). ) is to correct the polygonal line portion into a smooth curve so as to stop at time Te, which is after time T4, and store it as basic data. In the case of FIG. 10(B), the slopes of the straight lines L2 and L3 may be zero, but it is better to give them a certain slope to improve Z.
This is preferable in terms of drive control of the stage 20. On the other hand, Fig. 11 (
When the exposure time is long as in A), the movement characteristics of the Z stage 20 are a straight line L1' passing through the best focus position ΔZf at time Tc and a time T1 as shown in FIG.
It is approximated by a straight line L2' passing through the position ΔZ1 at time T4, and a straight line L3' passing through the position ΔZ2 at time T4. When the full-open time of the shutter is long, the intersection of straight lines L1' and L2' and the intersection of straight lines L1' and L3' are located near the center of the full-open time. Further, the slope of the straight line L1' may be set to match the maximum speed of the Z stage 20, but since the fully open time is long, it may be set to a speed lower than the maximum speed.

From the characteristics shown in FIGS. 10(B) and 11(B), there are some differences in the speed characteristics when the Z stage 20 starts moving and when it stops. It is preferable to store this in the storage unit 33 as reference data. Furthermore, in each of the above embodiments, data is stored in advance in the storage unit 33, but the optimum characteristics can be determined by numerical calculation using the polygonal line approximation as explained in FIG. 10(B) and FIG. 11(B). may be calculated. Furthermore, the projection exposure apparatus described above controls the exposure amount using a shutter, but in an exposure apparatus using a pulsed laser as a light source (such as an excimer stepper), exposure control is performed by the number of pulses from the light source and fine adjustment of the light amount for each pulse. It is carried out. Even in this case, the present invention can be applied in exactly the same way and the same effects can be obtained.

[0036]

As described above, according to the present invention, the positioning of the Z stage can be accurately and continuously servoed in real time. Furthermore, the Z stage movable range can be extended, and nonlinear movement of the Z stage due to changes in resist thickness or wafer surface reflectance is prevented, making position control in the optical axis direction extremely accurate and increasing the depth of focus. The effect is always stable.

[Brief explanation of drawings]

FIG. 1 is an overall configuration diagram of a projection exposure apparatus suitable for an embodiment of the present invention.

[Fig. 2] A configuration diagram for realizing a focus tracking method according to an embodiment of the present invention,

[Figure 3] Block diagram of main control system functions,

[Figure 4] Data strings stored in the basic data storage unit,

[Figure 5] Chart diagram showing the relationship between focus offset and exposure operation,

[Figure 6] A chart showing the actual change in focus offset over time.

[Figure 7] A graph showing an example of the stay probability obtained by moving the Z stage,

[Figure 8] Graph showing what happens when an electrical offset is applied to the synchronous detection output,

[Figure 9] A graph showing the situation when the maximum value of the stay probability is given to three points in the optical axis direction.

[Figure 10] Graph showing how to determine the speed characteristics of the Z stage,

FIG. 11 is a graph showing how to determine the speed characteristics of the Z stage.

[Explanation of symbols]

PL Projection lens R Reticle W Wafer IL Autofocus system light beam 1 Light transmission slit 10 Oscillating mirror 11 Mirror drive amplifier 12 Harving 13 for focus offset
Harving driver 14 Light receiving slit 15 Photomultiple 17 Synchronous detection circuit 18 for focus position detection
Z stage driver 19 Z stage movement motor 20 Z stage 21 XY stage 30 Main control system 33 Basic data storage unit

Claims (2)

[Claims] 1. A projection optical system that images and projects a pattern of a mask onto a photosensitive substrate; a Z stage that holds the photosensitive substrate and is movable in the optical axis direction of the projection optical system; and a Z stage that drives the Z stage. a motor, and a surface position detection means for detecting a relative positional deviation amount in the optical axis direction between the surface of the photosensitive substrate and a predetermined imaging plane of the projection optical system and outputting a detection signal according to the positional deviation amount. and control means for servo-controlling the motor so that the detection signal from the surface position detection means coincides with a predetermined value, the position of the surface of the photosensitive substrate and the predetermined imaging plane in the optical axis direction In a projection exposure apparatus that exposes the pattern onto the photosensitive substrate by setting the patterns in a predetermined relationship, the value of the detection signal from the surface position detection means is based on the actual relationship between the surface of the photosensitive substrate and the predetermined image forming plane. a correction means for correcting the surface position detection means so as to include a predetermined offset amount with respect to the amount of positional deviation; storage means for storing data for changing the speed at a predetermined speed characteristic over a predetermined range; and means for applying data in the storage means to the correction means upon the start of the exposure operation; A projection exposure apparatus characterized in that during an exposure operation, the Z stage is moved in the optical axis direction under servo control by the control means. 2. The surface position detection means includes a light projecting system that obliquely irradiates a light beam onto the surface of the photosensitive substrate, and a photoelectric sensor that receives reflected light of the light beam from the surface and detects a change in the light receiving position. the correction means is arranged in the optical path from the light projecting system to the light receiving system, and the correction means is arranged in the optical path from the light projecting system to the light receiving system to adjust the light flux or the reflected light with respect to the light receiving position detection direction of the light receiving system. a movable optical element for offsetting;
2. Projection according to claim 1, comprising: a second motor that drives the movable optical element; and second control means that servo-controls the second motor based on data in the storage means. Exposure equipment.