JPH0525171B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a light source control device, and more specifically to an exposure light source for a device that repeatedly exposes a circuit pattern or the like on a semiconductor wafer or a photomask in the manufacturing process of high-density semiconductor integrated circuits. The present invention relates to a control device for controlling the light variation operation of a light source suitable for.

[Background of the invention]

In recent years, in the production of semiconductor integrated circuits, especially LSIs, there has been a demand for increasingly finer circuit patterns, and extremely high-precision exposure control is also required for exposure equipment that prints patterns with line widths of 1 μm or less on wafers. It is rare. As one of the exposure methods that performs such highly accurate exposure control, a so-called flash exposure method has been proposed in which the intensity of the light emitted by the light source is varied in intensity in synchronization with the opening and closing operations of the shutter. This method uses a mercury discharge lamp as a light source, which has a quick response to changes in emission intensity in response to step changes in input power, and during the period when the shutter is closed, the emission intensity of the mercury discharge lamp is lower than the level required for exposure. The input power (for example, the rated power of a mercury discharge lamp) is such that the emission intensity of the mercury discharge lamp is at the level required for exposure during the exposure period when the shutter is opened. (2 to 3 times the electric power) and alternately performs this in accordance with the step-and-repeat operation.

In such an exposure method, the ratio of input power between the non-exposure period when the shutter is closed and the input power during the exposure period when the shutter is open is constant throughout the light changing operation. For this reason, when the exposure time and repetition period are changed, the average input power changes, resulting in disadvantages such as a reduction in luminous efficiency and a reduction in the life of the mercury discharge lamp. That is, to explain this with reference to Fig. 1, first of all, in Fig. 1 a, the horizontal axis is time;
The vertical axis is the input power to the mercury discharge lamp, the exposure period t _e1 when the shutter is open is the input power P _U that is powered up, and the non-exposure period t _S1 when the shutter is closed is the rated input power value P _N. When a mercury discharge lamp is turned on and operated to change its light periodically, the average input power is
P _A1 is now shown by a broken line, and the duty ratio Dr ₁ at that time is expressed as Dr ₁ =te ₁ /(te ₁ +t _S1 ). Here, the non-exposure period t _S1 is a preparation period for the next exposure, such as movement and positioning of the wafer stage, and the length of the exposure period te ₁ is determined by the length of the exposure period te 1, which is a preparatory period for the next exposure such as movement and positioning of the wafer stage. Type, thickness, and type of photoresist applied
It is determined as appropriate depending on the reflectance of the base.

Now, let us consider a case where the input power for the exposure period and the non-exposure period is unchanged, but the duty ratio for both periods is changed, as shown in FIG. 1b, compared to the case shown in FIG. 1a. In FIG. 1b, the exposure period te ₂ is te ₂ >te ₁ , the non-exposure period t _SS2 is t _S2 <ts ₁ , and the duty ratio Dr ₂ is Dr ₂ =te ₂ /(te ₂ +ts _S2 )> It is named Dr ₁ .

In this way, when the duty ratio increases as shown in Figure 1b, even if the input powers Pu and P _N do not change, the average input power will be lower than P _A1 in Figure 1a, as shown in Figure 1b.
P _A2 of is larger. As a result, not only does luminous efficiency decrease due to an increase in the temperature of the electrodes of the mercury discharge lamp, but also an unexpected accident may occur due to an increase in the pressure of the gas sealed inside the mercury discharge lamp and an increase in the temperature of the tube wall.

[Purpose of the invention]

An object of the present invention is to provide a light source control device that maintains the average input power to the light source at a constant value even if the duty ratio between the exposure period and the non-exposure period is changed.

[Summary of the invention]

In order to achieve the above-mentioned object, the light source control device of the present invention changes the light emission intensity of the light source into a first light emission intensity and a lower second light emission intensity by periodically changing the input power to the light source. A light source device that performs a light emission operation that repeatedly emits light between is equipped with a power control device that controls the input power to the light source during the first period in which the light emission intensity is the first emission intensity or the second period in which the light emission intensity is the second emission intensity. This solves the problem.

[Embodiments of the invention]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the present invention is applied to control of light variation operation of a light source of a reduction projection exposure apparatus will be described in detail below with reference to the drawings.

FIG. 2 is an optical path diagram showing a schematic configuration of a reduction projection type exposure apparatus, and FIG. 3 is an enlarged plan view of the shutter. First, this exposure apparatus will be explained with reference to FIGS. 2 and 3.

In FIG. 2, light from a mercury discharge lamp 1 is focused on a second focal point f of an elliptical mirror 2. In FIG. A rotary shutter 3 is arranged at the second focal point f, and is rotated by a constant angle by a pulse motor 4 to transmit and transmit the light beam.
Perform a cutoff. As shown in Fig. 3, the shutter 3 has a blade portion 3a and a cutout portion 3b formed at equal angular intervals, and the cutout portion 3b allows the light beam to pass through, while the blade portion 3a blocks the light beam. do.

In the example shown in Figure 2, the shutter 3 is in the optical path.
The blade portion 3a is provided at an angle of 45 degrees, and the light source side surface of the blade portion 3a is a mirror surface, and this mirror surface reflects the light flux from the light source in a different direction, so that it can be used for other lighting in the device, e.g. It is used for illuminating a wafer and mask positioning optical system (alignment microscope).

The light that has passed through the shutter 3 enters the first condenser lens 5 and becomes a parallel beam of light, and then passes through the optical member 6, which includes a fly-eye lens, an interference filter that selects the exposure wavelength (λ = 436 mm), etc. As a result, a parallel light beam l with a uniform intensity distribution is obtained. This light flux l further passes through the second condenser lens 7,
A reticle R on which a predetermined pattern is drawn is uniformly illuminated. The pattern light image of the reticle R is reduced and projected onto the wafer W by the reduction projection lens 8 . A light receiving element 9 disposed above the reticle R so as to receive light arriving outside the pattern area of the reticle R out of the light illuminating the reticle R
Outputs a photoelectric signal according to the intensity of the light illuminating the
It is used to measure the amount of exposure to detect the timing of opening and closing of the shutter 3.

The wafer W is placed on a two-dimensional moving stage 10, and by moving the stage 10, a reduced image of the pattern on the reticle R is sequentially exposed in a step-and-repeat manner so as to fill a predetermined area on the surface of the wafer W. It is something to do.

The entire control system of the exposure apparatus as described above is shown in FIG. In FIG. 4, the photoelectric signal _S1 from the light receiving element 9 is amplified by the amplifier 20, and then
It is input to the photometry circuit 21. The exposure time corresponding to the appropriate amount of exposure to the wafer W (the exposure period te ₁ described above)
A digital setting switch 22 is provided to set the shutter opening time (or shutter opening time corresponding to te ₂ ), and a digital switch 23 is provided to set the amount of stepping of the stage 10. The light source control circuit 24 is an input power control device that controls the intensity change operation of the light emission intensity of the mercury discharge lamp (light source) 1, and the shutter drive circuit 25 receives signals from the photometry circuit 21 and the microcomputer (CPU) 30. A drive pulse is outputted to rotate the pulse motor 4 by an angle θ. In order to measure its two-dimensional position, the stage 10 is equipped with a high-precision lengthening device, such as a laser interferometer, which is equipped with an internal counter that counts the movement of the interference fringes of the laser beam.
A stage driver 27 is provided to move the stage 10 two-dimensionally according to commands from the CPU 30.

The photometry circuit 21, digital switches 22 and 23, light source control circuit 24, shutter drive circuit 25, length measurement circuit 26, and stage driver 27 are connected via an interface bus 29.
Exchanges various information with the CPU 30,
30 programmatically controls the sequence of the entire device.

FIG. 5 is a circuit block diagram showing a specific example of the light source control circuit 24 shown in the previous figure. Current from a commercial AC power supply 40 is rectified by a rectifier 41, smoothed by a capacitor 42, and converted into a DC voltage. The DC voltage smoothed by the capacitor 42 is applied from the collector-emitter circuit of the control transistor 43 to one electrode of the mercury discharge lamp 1 via the bleeder resistor 44, and the other electrode is grounded.

The voltage drop occurring across the bleeder resistor 44 is proportional to the magnitude of the current flowing through the discharge lamp 1, and this voltage drop is input to one input terminal of the multiplier 47 via the amplifier 45. The other input terminal of this multiplier 47 is connected to the mercury discharge lamp 1 via an amplifier 46.
As a result, the multiplier 47 produces an output corresponding to the product of the output voltage of the amplifier 45 and the output voltage of the amplifier 46, that is, the input power value of the mercury discharge lamp 1. The output of the multiplier 47 is input to one input terminal of an error amplifier 48, and the digital data _D0 from the CPU 30 is input to the other input terminal of the error amplifier 48.
49, it is converted to an analog value and then input,
An error amplifier 48 detects the deviation between the two input signals. The output of the error amplifier 48 corresponding to this deviation is applied to the base of the control transistor 43,
The collector current of the control transistor 43 is controlled by this deviation output. The current control loop of the transistor 43 including the error amplifier 48 operates so that the deviation between the outputs of the multiplier 47 and the DAC 49 is always zero, so the input power of the mercury discharge lamp 1 is adjusted according to the data _D0 from the CPU 30. is variable.

In this way, for the power supply of the mercury discharge lamp 1,
A feedback control system is set up so that the power set by data _D0 is input to the discharge lamp 1.

On the other hand, the photometry circuit 21 has the function of an exposure control circuit, a specific example of which is shown in FIG. In FIG. 6, the photoelectric signal _S1 from the light receiving element 9 is input to a voltage-frequency converter (VFC) 60, and a pulse signal having a frequency corresponding to the voltage value of the photoelectric signal _S1 is input.
It is converted into S ₂ and input to the counter 61. The counter 61 responds to a clear signal CL from the CPU 30, clears its counting contents, and then sequentially counts the number of pulses of the pulse signal _S2 . counter 61
The count data _D1 is compared with target value data D2 given from the CPU ₃₀ by a comparator (CP) 62, and when both data DA and _D2 match, the comparator 62 outputs a stop signal SP. . in this case,
The target value data _D2 is a value corresponding to the appropriate exposure amount.

The shutter drive circuit 25 receives the clear signal.
A drive pulse for opening the shutter 3 is output to the motor 4 in response to CL, and a drive pulse for closing the shutter 3 is output to the motor 4 in response to the stop signal SP.

Note that the CPU 30 does not control the DAC until it outputs the clear signal CL, that is, until it starts the exposure operation.
49 is given data D ₀ corresponding to the rated input power P _N of the lamp 1, and therefore the lamp 1 is lit at the rated input power, for example 500W.

Next, the light source control operation of the exposure apparatus having the above configuration will be explained with reference to the flowcharts of FIGS. 7 and 8.

FIG. 7 shows the main program flow for light source control, the first step 100 being for initialization. In step 100, the exposure time te of the step-and-repeat exposure operation, the step (movement) time _tS , and the reduced power value Pd of the light source power down are determined. The details are as follows.

<Step 200> As the first step of initialization, the degree of increase in the input power of the lamp 1, that is, the power up rate C, which indicates how much the input power is increased relative to the rated input power, is set. For example, the power-up rate C takes a value of 0C1, and in order to shorten the exposure time te, specify C=1, that is, 100% power-up with respect to the rated input power, and reduce the input power to lamp 1 at power-up. Increase it to twice the rating.

<Step 201> Power up rate C set in step 200
and the rated input power value P _N of lamp 1, calculate the input power value Pu at power-up. That is, in this step, the calculation Pu=P _N (1+C) is performed.

<Step 202> Power-up electric power Pu obtained in step 201
is smaller than the maximum allowable instantaneous input power Pum of the lamp 1. If Pu>Pum, proceed to step 203; if PuPum, set the value of Pu as is and proceed to the next step 204.

<Step 203> This is a case where it is determined in step 202 that Pu>Pum. In this case, the value of Pum is set as Pu instead of the value of Pu calculated in step 201, and the process proceeds to the next step 204.

<Step 204> The CPU 30 reads data corresponding to the appropriate exposure amount set in the digital switch 22.

<Step 205> Based on the data read in step 204, the CPU
30 provides data D ₂ to a comparator 62 .

<Step 206> The CPU 30 outputs a clear signal CL to cause the shutter drive circuit 25 to open the shutter 3.

<Step 207> At the same time as the clear signal CL is generated in step 206, the content of data _D0 is changed from the rated input power value P _N to the power-up power value Pu, and the light intensity of the lamp 1 is increased via the control circuit 24.

<Step 208> At the same time as the clear signal CL is generated in step 206
Timer counter in CPU 30 (not shown)
and measure the time that shutter 3 is actually open.

<Step 209> In this step, the stop signal SP is monitored, and when the stop signal SP is output from the counter 61, the process proceeds to the next step 210.

<Step 210> Here, the stop signal SP from the counter 61
At the same time, the timer counter in the CPU 30 stops clocking. At this time, the shutter 3 is closed because the shutter drive circuit 25 responds to the stop signal SP.

<Step 211> Simultaneously with the arrival of the stop signal SP, the CPU 30 returns the contents of the data _D0 from Pu to _PN again.

By the operation described above, the wafer W to be exposed is
Shutter 3 necessary to give proper exposure to
The opening time, that is, the exposure time te is measured.

Note that the above operation is necessary not only when starting up the system, but also in order to know the accurate exposure time te when the light intensity decreases due to deterioration of the lamp 1. For example, if lamp 1 is new, the rated input power is 500W, and the input power at power-up is 1000W.
Assuming this, the light intensity is approximately 2I when the power is increased compared to the light intensity I at the rated time. However, due to deterioration of the lamp, the light intensity at the rated input power decreases to I′ (where I′<
When it comes to I), the light intensity when the power is turned up is approximately
2I', and the exposure time required to give the same amount of exposure to the wafer W is longer than te. Also, due to the deterioration of lamp 1, the power is lower than 500W when the power is turned on.
Even if the input power is increased to 1000W, the light intensity will still be
The exposure time te cannot be calculated accurately from the input power of the lamp 1, as the value does not necessarily change from I' to 2I'. For this reason, it is necessary to perform the operations of steps 200 to 211 as appropriate to obtain an accurate exposure time te in advance before the exposure operation of the wafer W.

Now, in step 211, the input power to the lamp 1 is returned to the rated value PN, and then the process proceeds to the next step 212.

<Step 212> In this step, the CPU 30 continues the value set in the digital switch 23, that is, the amount of stage movement (stepping amount) DL from one exposure field to the next exposure field on the wafer W in step-and-repeat operation. .

<Step 213> From this step onwards, the stepping time t _S and the power down power Pd are determined based on the movement amount DL, and will be explained with reference to FIG. 9 as follows.

FIG. 9 shows the relationship between the moving distance of the stage 10 and time, and the stage 10 is shown as being subjected to constant acceleration control.

In FIG. 9, when the stage 10 moves from position P ₀ to position P ₁ which is a distance DL ₁ away,
From P ₀ , the speed increases at a constant slope according to the acceleration characteristic LA (acceleration a), and from a position DL ₁ /2 away from P ₀ , the speed increases at a constant slope according to the deceleration characteristic LB (acceleration - a). The speed decreases and at position P ₁ the speed reaches O ₁ or stops. Therefore, the stepping time t _S in this case is determined.

It is clear from these characteristics LA and LB that the time it takes for stage 10 to move by DL ₁ /2 from P ₀ is exactly half the time it takes to move from P ₀ to P ₁ , so the initial velocity of stage 10 Under the condition that is zero, DL ₁ /2 = at ² /2 (t: time) Therefore, t = √ ₁ , and in the end, the stepping time t _S from position P ₀ to P ₁ is t _S = 2√ ₁ ...(1).

On the other hand, the acceleration characteristic LA does not increase infinitely, but is limited by the maximum speed Vm. For this reason, DL
When becomes as long as DL ₂ , the stage 10 is controlled in the order of acceleration characteristic LA → constant speed characteristic LC → deceleration characteristic LD.

Therefore, when the stage 10 moves from the position P ₀ to the position P ₂ which is DL ₂ apart, the distance dm to be traveled between the acceleration characteristic LA and the deceleration characteristic LD and the time tm required for it are determined in advance on the stage. Since it is known as a characteristic value specific to 10, the constant speed characteristic LC
The travel time tc inside is determined by tc=(DL ₂ −2dm)/Vm. As a result, the total stepping time _tS can be calculated as _tS =2tm+(DL-2dm)/Vm (2).

In this way, in this step 213, the movement amount DL is
It is determined whether or not DL>2dm, and if DL>2dm is true, proceed to step 214, and if DL2dm, proceed to step 215.

<Step 214> If DL>2dm, calculate _tS using equation (2).

<Step 215> In the case of DL2dm, t _S is determined using equation (1).

<Step 216> Based on t _S obtained in step 214 or 215 and the exposure time te measured up to step 210, a reduced power value Pd during power down is calculated. That is, if the number of repetitions of exposure on one wafer W is m times, the power down is performed during stepping between each exposure timing, so the number of times is (m-1). This power down power value Pd
The calculation of , with reference to FIG. 10, is as follows.

In FIG. 10, the time T _k from the start of the first exposure to the end of the last exposure for one wafer W
_The normal _power P _A input to the _{lamp 1 at} For example, from the above formula, the power down power Pd is Pd={P _N・T _k −m・te・Pu}/{(m−1)・
t _S } Here, T _k =m・te+(m−1)t _S , so Pd=Pk−{m・te(Pu−P _N )/(m−1)t _S }
...(3) becomes.

For example, when P _N = 500W and Pu = 1000W, te
=0.3sec, _tS =0.8sec, and if m=50, Pd is calculated as Pd=500-50×0.3×(1000-500)/49×0.8≈309W.

Note that if m is large, equation (3) may be approximated as Pd=P _N -te(Pu-P _N )/t _S (3a).

<Step 217> In this step, it is determined whether the power-down power Pd calculated in the previous step is larger than the instantaneous minimum input power Pdm. If the input power is lower than Pdm, the lamp 1 can remain lit, but a lamp starting operation is required to restart the lamp.
Therefore, in this step 217, Pd and Pdm are compared,
If PdPdm, the program returns to the original program shown in FIG. 7 and proceeds to step 101, and if Pd<Pdm, the program proceeds to step 101 after passing through the next step 218.

<Step 218> This is a case where it is determined in step 217 that Pd<Pdm, and in this case, the value of Pdm is set to be used as Pd. That is, the lower limit of the power-down power value is limited here.

When the subroutine shown in FIG. 8, ie, the initialization step 100, is completed, the process returns to the flow shown in FIG. 7 and starts the exposure operation on the wafer W. In FIG. 7, the first step after initialization is step 101 for alignment of the wafer W.

<Step 101> The wafer W is transferred to the stage 10, and the wafer W is accurately positioned (aligned) at the exposure position. In the operation after this alignment, the light source control operation is performed in synchronization with the operation of the stage 10 for step-and-repeat exposure using a lamp input power pattern as shown in FIG. Ru.

<Step 102> The projection position of the pattern image of the reticle R is adjusted to a predetermined position on the wafer W by stepping movement of the stage 10.

<Step 103> The CPU 30 outputs a clear signal CL to open the shutter 3.

<Step 104> At the same time as the clear signal CL is output in step 103
The data _D0 given from the CPU 30 to the DAC 49 is made to correspond to the power-up power value Pu, and the input power of the lamp 1, which is lit at the rated power P _N , is increased to the power-up power Pu.

<Step 105> The arrival of the stop signal SP is monitored, and the comparator
When the stop signal SP from the CP is input to the CPU 30, the process advances to step 106. Incidentally, the opening time of the shutter 3 at this time is the exposure time te already set in steps 204 and 205 described above.

<Step 106> Check the exposure number on the wafer W, and if it is less than a predetermined number m, proceed to step 107, and if it reaches m, proceed to step 108.

<Step 107> This is a case where it is determined in step 106 that the number of exposures has not reached m, and in this case, the input power of the lamp 1 is changed by changing the contents of the data _D0 by the CPU 30 to correspond to the power down power value Pd. to Pd and step again after stepping time t _S.
Repeat exposure returning to step 102.

<Step 108> When it is detected in step 106 that the number of exposures has reached m, the input power of the lamp 1 is returned to the rated value P _N and the exposure operation for one wafer is completed.

This completes the light source control operation during exposure for one wafer, but if the next wafer is to be exposed subsequently, the same steps can be repeated from step 101. In FIG. 10, time _Tk is the time required to expose one wafer, and Tc is the time required to carry out the wafer after exposure and carry in and align the next wafer. Further, if the type of wafer changes, such as a difference in photoresist or thickness, or if the exposure time changes, initialization is executed in step 100.

In the embodiment described above, steps 217 and 218 in FIG. 8 limit the lower limit input power during power down, but depending on the values of exposure time te, stepping time t _S and power up power Pu. In some cases, Pd<Pdm. For example, P _N =500W
When Pu=1000W and te= _tS , if m=50 and formula (3) is followed, Pd≈0.

Generally, the instantaneous minimum input power Pdm of a mercury discharge lamp is about 1/2 to 1/3 of the rated power P _N. Therefore, in the above case, the value of Pdm is set as Pd.
If an exposure operation is performed in this state, the average input power P _A
will be larger than the rated power P _N.

Therefore, in such a case, in step 217, Pd<
When it is determined that Pdm, the power-up rate C may be corrected to be smaller in step 200 and recalculation may be performed to determine the optimum value of the power-up power Pu. If that doesn't work, it will issue an alarm or flash (variable light operation).
It is better to switch to normal exposure, which does not allow exposure.

In addition, the power Pu at power-up was inputted at the power-up rate C to give an input value proportional to the power, but the light intensity at the time of exposure was calculated using the rated power.
An input method such as how many times the light intensity at P _N may be used may be used. In this case, compare the light intensity when the rated power is P _N and the light intensity when the power is turned up, change the data D _{0 in various ways so that the ratio becomes the input setting value, and input the data D 0} to the set light intensity. Set the data D ₀ at the time of power-up to be the value of the power Pu at power-up. For this purpose, it is convenient to provide, for example, between the lamp 1 and the shutter 3 a light receiving element that always receives the light from the lamp 1 regardless of the operation of the shutter 3.

Furthermore, in this embodiment, the average input power P _A is set to be equal to the rated input power P _N , but the average input power P _A may be larger than the rated input power value P _N.
The amount is governed by the characteristics of the lamp 1 and is determined by how much larger the maximum power that can be continuously input can be allowed to be than the rated input power P _N. In the case of mercury discharge lamps, the tolerances are relatively small and also vary depending on the conditions of use, especially the cooling conditions. Preferably, the mercury discharge lamp is forcedly air-cooled, and the amount of air blown by the air-cooling fan during exposure operation is increased.

For example, P _N = 500W, Pu = 1000W, te = 0.3sec,
If m is large under the condition of t _S = 0.5 sec, Pd may be approximately 200 W, and in this case Pdm = 250 W.
Then, since Pd is set to 250W, the average input power P _A is 531W. Therefore, a control system may be configured to increase the amount of _air blown by the air cooling fan in accordance with the increase in average input power PA from rated input power P _N.

In addition, in the above example, conversely, Pd=250W (=
As mentioned above, it is also possible to set the power-up power Pu and the exposure time te so that Pdm). For example, by giving P _N = 500 W and t _S = 0.5 sec,
Using approximation formula (3a) of formula (3), when te=0.3sec, Pu=
Assuming that the relationship of 1000W is approximately inversely proportional, and assuming that the exposure amount K=Pu・te=300, we get: 250=500−te(Pu−500)/0.5 Therefore, te=0.35sec, Find Pu=857W.

Further, in the above-described embodiment, the shutter 3 was opened once and the exposure time te was preliminarily measured before starting the wafer exposure operation, but this may not necessarily be necessary. For example, during the first exposure shot on a wafer.
te is measured, and based on that value, the calculation shown in Figure 8 is completed before the next exposure, and the power is turned down.
It is also possible to calculate Pd in advance and power down after the second exposure shot. In this case, between the first shot and the second shot, power down is either not performed, or even if it is, the power is down to, for example, Pdm, but the greater the number of exposures m, the shorter the time T. The difference between the average input power P _A and the rated input power P _N within _K becomes small, and there is not much difference in practical use.

FIG. 11 is a block diagram of a photometric circuit system according to a second embodiment of the present invention, in which an analog-to-digital converter (ADC) 65 is used as the photometric circuit 21, and a VFC6
0 and the light amount integral by the counter 61, the shutter 3 is time-controlled.

In FIG. 11, the ADC 65 converts the level of the photoelectric signal _S1 into digital data _D3 . In this case, the ADC 65 responds to the signal S ₃ from the CPU 30 and converts the level of the photoelectric signal S ₁ at that time. An oscillator (OSC) 66 supplies a constant cycle clock signal CLK to the CPU 30, and the CPU 30 supplies the shutter drive circuit 25 with an open signal (clear signal) CL for the shutter 3 and a stop signal SP for closing the shutter 3.

The operation of the example shown in FIG. 11 will be explained with reference to the flowchart shown in FIG. Note that FIG. 12 shows the exposure operation for one wafer.

The first step 150 is a stepping step similar to step 102 in FIG.

By the way, the shutter 3 is operated from the time when the open signal CL is generated until it reaches full open, or when the stop signal is
Each operation time is required from when SP occurs until fully closed. The operating characteristics of the shutter 3 as a result of changes in light intensity with respect to the wafer will be briefly explained with reference to FIG. 13.
When the release signal CL is output at time _t1 , the shutter becomes fully open at time _t2 , and the operating time is ( _t2 - _t1 ). Also, stop signal SP
is output at time _t3 , it becomes fully closed at time _t4 , and its operating time is ( _t4 - _t3 ). Generally, this operating time is equal for the opening and closing operations of the shutter 3. In the embodiment related to FIG. 6 above, the time point
Although it is assumed that the lamp 1 is powered up from _t1 to _t4 , in reality, as in this example, the power down is started at the same time as shutter 3 is closed, that is, at time _t3 (that is, at time t3). ₁ to _t3 ). Therefore _, the light intensity characteristics on the wafer are controlled as shown in _FIG .
1000W is input, and from time _t3 , 250W, half of the rated power of 500W, is input with power down, and in this way, the exposure amount when the shutter is closed ( _t4 −
t ₃ )·I ₂ /2 can be controlled to be extremely small.

Therefore, in this embodiment, the exposure amount from time t ₁ to t ₃ is set as the appropriate exposure amount, and the amount (t ₂ - t ₁ )I ₁ /
2 + (t ₃ − t ₂ )I ₁ to calculate the full open time (t ₃ − t ₂ ),
Assume that shutter 3 is closed. in this case,
Since (t ₂ − _{t 1} ) is a constant value unique to the shutter itself, let it be Ta, and let (t ₃ − t ₂ ) be Tx and the appropriate exposure amount be E, then (Ta/2+Tx)I=KE (However, K is a constant) Therefore, Tx=KE/I-Ta/2. Here, the light intensity I can be determined from the data _D2 from the ADC65, so
Tx is obtained as Tx=KE/D ₃ −Ta/2 (4).

At step 151, a value corresponding to the operating time of the shutter 3, that is, Ta, is set in the internal counter TM1 (not shown) of the CPU 30. step
Power up at step 152 and close shutter 3 at step 153.

In step 154, the presence or absence of the clock signal CLK is determined. If CLK is present, the counter TM1 is set in step 155.
If the value of TM1 is 0, then the value of TM1 in step 156 is changed by one clock, that is, 1
, and return to step 154 again. When this is repeated and TM1 becomes 0 (time t ₂ in FIG. 14), the CPU 30 outputs the signal S ₃ in step 157, and the ADC 6
5 is performed, and the converted data _D3 is read in step 158.

In step 159, Tx is determined based on the above equation (4), and the value of Tx is set in another internal counter TM2 (not shown) of the CPU 30. It goes without saying that E is given in advance from the digital switch 22. Further, Ta/2 is a characteristic value of the shutter itself, and is usually 10 to several milliseconds.

In step 160, the presence or absence of the clock signal CLK is monitored, and if CLK is present, the counter is activated in step 161.
It is determined whether TM2 is 0 or not, and if TM2≠0, the value of counter TM2 is decremented (-
1) and repeat from step 160.

When TM2 becomes 0 in step 161 (time T ₃ in FIG. 14), the CPU 30 outputs a stop signal SP in step 163 and starts the closing operation of the shutter 3, and simultaneously starts power down in step 164. By the way, in the case of mercury discharge lamps, the response time for power up and power down is generally extremely short, about 1 msec.

In step 165, it is determined whether m exposures have been performed on the wafer, and if it is less than m, step 150 is performed.
The process is repeated from 1 to 3, and when the process reaches m times, the lamp input power is returned to the rated value P _N in step 166, and the process ends.

According to this embodiment, the amount of exposure due to the delay in the closing operation of the shutter 3 is extremely small, so even if stepping of the stage 10 is started immediately after the closing of the shutter 3 starts, the amount of unnecessary exposure given to the wafer is is small and does not cause pattern image flow. Therefore, there is an advantage that the stepping operation can be started earlier, and the overall time for exposure processing m times for one wafer can be shortened.

As another modification of the present invention, after determining the exposure time te at the rated input power of the lamp 1, the exposure time te required for setting the exposure time te during the power-up exposure operation to a predetermined value is determined by preliminary exposure or the like. The exposure of the wafer may be started after calculating the power-up power Pu and calculating the power-down power Pd based on Pu, te, and _tS .

Furthermore, when controlling the exposure amount according to the present invention, the light intensity corresponding to the supplied power is obtained from the lamp 1, so the exposure amount for each exposure can be controlled by, for example,
It is not necessarily necessary to measure and control using an integrating means such as a counter, and it is sufficient to control the exposure amount by simply measuring the light intensity at a steady state once before starting repeated exposure.

〔Effect of the invention〕

As described above, according to the present invention, since the average power supplied to the lamp is constant, the gas pressure inside the lamp and the electrode temperature can be kept almost constant, and the intermittent power increase due to repeated exposure can be avoided. However, there is no change in light intensity due to a change in luminous efficiency between the beginning and the end. Furthermore, by controlling in this manner, it is possible to prevent excessive input to the lamp, to prevent shortening of lamp life, and to prevent lamp breakage accidents.

[Brief explanation of the drawing]

1A and 1B are diagrams showing the temporal change pattern of the input power to the light source, FIG. An enlarged plan view of the shutter, FIG. 4 is a block diagram showing the overall control system of the exposure apparatus, FIG. 5 is a block diagram showing a specific example of the light source control circuit, and FIG. 6 is a first embodiment of the photometry circuit. 7 is a flowchart showing an example of the light source control operation, FIG. 8 is a flowchart showing the subroutine of the initialization step in FIG. 7, and FIG. 9 is a flowchart showing the speed of stepping movement of the stage. 10 is a diagram showing the change pattern of the input power to the lamp during exposure operation, FIG. 11 is a block diagram showing the second embodiment of the photometry circuit, and FIG.
The figure is a flowchart for explaining the operation, No. 13.
This figure is a dynamic characteristic diagram of the shutter, and FIG. 14 is a diagram showing exposure characteristics on a wafer. 1: light source lamp, 3: shutter, R: reticle, 9: light receiving element, 10: stage, W: wafer, 21: photometry circuit, 22, 23: digital switch, 24: light source control circuit, 25: shutter drive circuit, 26: Length measuring device, 27: Stage driver, 30: CPU, 60: VFC, 61: Counter, 62: Comparator, 65: Digital-analog converter, 66: Oscillator.

Claims (1)

[Scope of Claims] 1. In a light source control device that causes a light source to periodically and repeatedly emit light with a first light emission intensity and a second light emission intensity that is lower than the first light emission intensity, Controlling the input power to the light source during the first period when the light source has a first emission intensity or during the second period when the light source has a second emission intensity so as to maintain the average input power at a predetermined value. 1. A light source control device comprising a power control device. 2. The light source control device according to claim 1, wherein the average input power to the light source during the light emission operation is maintained at the rated input power value of the light source. 3. Control the input power to the light source during the second period based on the duty ratio of the light emission operation between the first period and the second period and the input power value to the light source during the first period. A light source control device according to claim 1 or 2.