JP2779569B2 - Output control device of laser device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is mainly used as a light source of a step-and-step type reduction projection exposure apparatus (hereinafter referred to as "stepper"), and an output control apparatus of a laser apparatus which oscillates a laser by being excited by discharge. About.

[0002]

2. Description of the Related Art In a stepper, strict exposure control is required to maintain the resolution of a circuit pattern at a certain level or more. On the other hand, an excimer laser used as a light source for this stepper has a variation in pulse energy for each pulse because of a so-called pulse discharge excitation gas laser, and it is necessary to reduce this variation in order to improve the accuracy of exposure control. is there. In addition, since the light is an intermittent light, an exposure amount control different from the conventional shutter control when a mercury lamp which is a continuous light is used as a light source is required.

[0003] Therefore, as described in the literature (Miyaji et al., “Excimer Laser Lithography”, International Laser / Application '91, Seminar L-5, P36-51), exposure is performed by continuously oscillating a plurality of pulses. Do,
There is a technique that attempts to improve the accuracy of exposure control by so-called multiple pulse exposure.

According to this method, the variation in the energy of the oscillating pulse of the excimer laser can be approximated by a normal distribution. 1 / (n) 1/2 is used.
That is, the variation in energy of one pulse is ΔP /
Assuming that P and required exposure amount control accuracy are A, the necessary number N of exposure pulses is given by the following relationship.

N ≧ ｛(ΔP / P) / A｝ 2 For example, the energy variation ΔP / P of one pulse is 1
5% (3σ), and the required exposure amount control accuracy A is 1.5% (3σ).
If σ), then N ≧ 100, and the desired accuracy can be achieved with 100 or more continuous pulse oscillations.

The stepper alternately repeats exposure and stage movement. For this reason, the operating state of the excimer laser as a light source is necessarily in a so-called burst mode. Note that the burst mode refers to repeating the operation of stopping the pulse oscillation for a predetermined time after continuously oscillating the laser beam for a predetermined number of times. That is, a short continuous pulse oscillation period and a short oscillation suspension period are alternately repeated.

As described above, since the excimer laser is a pulse discharge pumped gas laser, it is difficult to continuously oscillate at a constant pulse energy. As a cause of this, the density of the laser gas is disturbed in the discharge space due to the discharge, and the next discharge becomes non-uniform or unstable. This is because a local temperature rise occurs, and the next discharge is further deteriorated to make the discharge uneven and unstable. In particular, the tendency is remarkable at the beginning of the continuous pulse oscillation period, and in the first pulse after the lapse of the oscillation pause period, a stable discharge is obtained and a relatively high pulse energy is obtained, but thereafter, the discharge deteriorates. A so-called spiking phenomenon in which the pulse energy gradually decreases appears. This phenomenon is shown in FIG.

As described above, in the excimer laser apparatus in the burst mode operation, the above-described variation in energy for each pulse lowers the accuracy of the exposure amount control, and the spiking phenomenon further significantly increases the variation, thereby increasing the accuracy of the exposure amount control. Is further reduced.

In recent years, the sensitivity of the photosensitive agent applied to the wafer has been improved, and exposure with a small number of continuous pulses has become possible, and the number of pulses tends to decrease.

However, when the number of pulses is reduced, the variation in pulse energy is correspondingly increased, and it becomes difficult to maintain the accuracy of the exposure amount control only by the above-described multiple pulse exposure control. For this reason, it is desired to improve the variation of the pulse energy, particularly to eliminate the influence of the spiking phenomenon in the burst mode.

Therefore, utilizing the property that the energy of the pulse oscillated increases as the discharge voltage increases, the discharge voltage of the first pulse of the continuous pulse oscillation in the burst mode is reduced, and thereafter the discharge voltage of the pulse is reduced. The present inventors have proposed an invention relating to control for preventing the initial energy rise due to the spiking phenomenon by changing the discharge voltage for each pulse, that is, gradually increasing the discharge voltage, that is, the spiking occurrence prevention control. At the same time as a patent application (Japanese Patent Application No. 4-11056).
Already adopted and implemented. According to the present invention, since the influence of the spiking phenomenon is removed, the accuracy of the exposure control can be improved even with a small continuous pulse oscillation.

Here, the present inventors have found that (a) the occurrence pattern of the spiking phenomenon becomes more prominent as the oscillation pause time Tpp in the burst mode increases, and (b) a pulse of a predetermined size in accordance with the deterioration of the laser gas. The discharge voltage required to obtain energy changes
The above-mentioned patent application discloses a technique in which the generation pattern of the spiking phenomenon is changed by the so-called power lock voltage Vpl, and a technique for removing the spiking phenomenon with high accuracy based on these is disclosed. In this patent application, as shown in the following equation (1),
A function Vi (Tpp, Vpl) having Tpp and Vpl as variables is set, and the discharge voltage V (i) corresponding to the pulse in the oscillation order i is set.
Is obtained as a function value of the function Vi (Tpp, Vpl).

V (i) = Vi (Tpp, Vpl) (1) where V (i) is the discharge voltage of the i-th pulse of the continuous pulse oscillation. Vi: is the discharge voltage of the i-th pulse of the continuous pulse oscillation. Function formula to be determined Tpp: oscillation pause time Vpl: power lock voltage Then, a continuous pulse (i = 1, 2, 3,...) In which the values of the variables Tpp and Vpl are the same is set as a set of discharge voltage data strings,
That is, there is no single voltage pattern table, and the table is stored in the memory in table units. Then, simultaneously with the start of the continuous pulse oscillation, from the top of the table, that is, i = 1, 2, 3
.. Are controlled by transmitting voltage data to the laser power supply in the order of.

[0014]

The above-mentioned voltage pattern table is prepared so as to be adapted to a predetermined repetition frequency, generally, the maximum repetition frequency f of the laser device. This is because high repetition oscillation (up to 1 kHz) has become possible in the excimer laser device, so that if a voltage pattern table suitable for any repetition frequency is prepared, the consumption of the storage area of the memory for storing the table becomes enormous. This is because the cost of the control device is significantly increased as the storage capacity of the memory is increased, and this is to avoid this.

Therefore, in the related art, there is a problem that the spiking occurrence prevention control effectively works only when the operation is performed at the maximum repetition frequency f or the repetition frequency very close to the maximum repetition frequency f. That is, for example, assuming that the maximum repetition frequency is f (= 1 kHz), the discharge voltages V (1), V (2), V (3)... It is remembered. Therefore, when the operation is performed at the maximum repetition frequency f, the discharge voltages V (1), V (2), V (3)...
When these voltage data are transmitted to the laser power supply, the trigger signal reception interval and the time interval T match, so that the control is performed accurately. However, when the repetition frequency changes and deviates from the maximum value f, the discharge voltages V (1), V (2), V (3)... Even if these voltage data are transmitted to the power supply, the actual trigger signal reception interval does not match the time interval T corresponding to the maximum repetition frequency f, so that the control is not performed accurately.

The present invention has been made in view of such circumstances, and even if the repetition frequency is changed, a spike generation prevention control suitable for each repetition frequency can be performed at low cost. Its purpose is to provide.

[0017]

Accordingly, the main invention of the present invention is applied to a laser device which repeatedly performs an operation of pulsating laser light for a predetermined number of times and then stopping the pulse oscillation for a predetermined time. In the output control device of the laser device, which controls the charging voltage so that the energy of the pulse becomes a predetermined value based on the voltage data, before starting the operation of the laser device, the energy of each pulse of the continuous pulse is made the same. The charging voltage data to be stored is stored in advance for each magnitude of the pause time immediately before the continuous pulse oscillation according to the elapsed time from the start of the continuous pulse oscillation.

[0018]

[Effects] That is, the present inventors conducted various experiments, and the following became clear.

That is, the power-lock voltage was fixed, and the occurrence patterns of spiking phenomena at various repetition frequencies were measured (the oscillation pause time immediately before continuous pulse oscillation was the same). Then, it was found that the same curve was drawn at any repetition frequency, with the pulse energy E on the vertical axis and the elapsed time τ from the start of continuous pulse oscillation on the horizontal axis in FIG. This means that there is a correlation between the magnitude of the pulse energy E and the elapsed time τ from the start of the continuous pulse oscillation regardless of the magnitude of the repetition frequency. 4 does not change, indicating that the pattern of occurrence of the spiking phenomenon depends only on the elapsed time from the start of continuous pulse oscillation without depending on the repetition frequency.

Therefore, before starting operation of the laser device,
Charge voltage data that makes the magnitude of the energy of each pulse of the continuous pulse the same is stored for each magnitude of the pause time immediately before the continuous pulse oscillation in accordance with the elapsed time from the start of the continuous pulse oscillation. This eliminates the need to store charging voltage data corresponding to various repetition frequencies, and can reduce the memory capacity. If the charging voltage is controlled so as to obtain the stored charging voltage corresponding to the measured elapsed time, the control can be performed accurately even if the repetition frequency changes. Also, since the charging voltage data is stored for each magnitude of the pause time immediately before the continuous pulse oscillation, no matter how large the magnitude of the pause time changes during the operation of the laser device, the corresponding charging voltage data is stored. The charging voltage is accurately controlled with the data.

[0021]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an output control device for a laser device according to the present invention will be described below with reference to the drawings.

As shown in FIG. 1, the apparatus of the embodiment is an excimer laser apparatus 1 for outputting an excimer laser beam L.
And the output laser light L using the excimer laser device 1 as a light source.
And a stepper 9 for performing reduced projection exposure.

The oscillator 2 of the laser device 1 comprises a laser chamber, an optical resonator and the like.
It is filled with a laser gas such as F2. And
Electric discharge is performed between electrodes provided in the laser chamber to excite a laser gas to perform laser oscillation. The oscillated laser light is resonated in the optical resonator, and is output as an effective oscillated laser light L from a front mirror (not shown). The discharge is performed at a predetermined interval with a predetermined pulse width, and the laser light L is output intermittently.

The laser light thus oscillated from the oscillator 2 is partially sampled by the beam splitter 3 and is incident on the output monitor 5 via the lens 4. The output monitor 5 detects the energy per pulse of the output laser light L, that is, the pulse energy E.

The pulse energy E detected by the output monitor 5 is applied to an output control unit 6, and the output control unit 6 generates a desired pulse energy Ed based on the input pulse energy E as described later. Voltage data is applied to the laser power supply unit 8 so as to be obtained. in this case,
Power lock control is performed.

The power lock control is a control method for preventing the pulse energy E from decreasing even when the laser gas is deteriorated and the same discharge voltage is applied, by increasing the discharge voltage in accordance with the deterioration. One. In general,
The energy of multiple oscillated pulses is integrated and averaged,
This is feedback control for determining the discharge voltage after the next pulse by comparing with the desired energy Ed.
The determined discharge voltage (discharge voltage for obtaining the desired energy Ed) is called a power lock voltage Vpl. In addition,
"POWERLOCK" is a registered trademark of Questek, USA. On the other hand, the spiking prevention control
This is prediction control in which the pulse energy of one pulse to be oscillated is predicted and the discharge voltage is determined before oscillation.

The laser power supply unit 8 applies the potential difference V between the electrodes according to the applied voltage data, and performs the discharge. Here, the voltage for discharging is the laser power supply unit 8.
Is temporarily charged by a charging circuit disposed therein, and is discharged by the operation of a switch element such as a thyratron.

The output controller 6 is connected to a stepper controller 10 in the stepper 9 via a signal line, and receives a trigger signal Tr sent from the stepper controller 10.
The output control unit 6 has a built-in timer, and the timer measures the time between reception times of the transmitted trigger signal Tr. The detection result E of the output monitor 5 is also added to the control unit 7, and data is exchanged with the output control unit 6.

In the embodiment, the discharge voltage V corresponding to each pulse of the continuous pulse is determined in consideration of the following items.

(A) Pause period of pulse oscillation Tpp When the excimer laser is operated in the burst mode, a spiking phenomenon appears in which the pulse energy E increases (the energy gradually decreases thereafter) immediately after the start of the continuous pulse oscillation ( (See S in FIG. 5). The experiment by the present inventors has revealed that the spiking phenomenon becomes more remarkable as the oscillation pause time Tpp in the burst mode becomes longer. Further, the pulse energy E has a property of increasing as the discharge voltage V for exciting the laser gas increases.

Then, the pause time Tpp of the pulse oscillation is measured, and each pulse is measured based on the measured pause time Tpp so that the energy E of each pulse of the next continuous pulse becomes the same desired magnitude Ed. The magnitude of the corresponding discharge voltage V is changed. That is, as the discharge voltage V is lowered for the first pulse of the continuous pulse oscillation, and thereafter the discharge voltage V is gradually increased, the discharge voltage is changed for each pulse to prevent the initial energy increase due to the spiking phenomenon. Moreover, the degree of change of the discharge voltage V is made different according to the pause time Tpp. As a result, the energy level of each pulse always becomes the same value Ed.

(B) Power lock voltage Vpl In addition, the operation time of the laser oscillation becomes longer, and the pulse energy E decreases as the laser gas deteriorates. Power lock control for increasing the lock voltage Vpl is performed. However, experiments by the present inventors have revealed that when the power lock control is performed, the generation pattern of the spiking phenomenon changes. That is, it has been clarified that the pulse energy E changes according to the magnitude of the power lock voltage Vpl, and that the number of pulses affected by the spiking phenomenon changes. Therefore, it is necessary to change the discharge voltage V so that the pulse energy E becomes a desired magnitude Ed according to the power lock voltage Vpl.

(C) Real-time processing When the repetition frequency of the laser oscillation becomes high, the output interval of the trigger signal Tr is measured, and the output control unit 6 outputs the discharge voltage V corresponding to each pulse at each reception.
Is calculated, and it is difficult to perform a process of sequentially outputting the calculated values to the laser power supply unit 8. Therefore, it is required to perform quick processing and realize real-time processing.

Therefore, in the embodiment, the above (a) to (c)
In consideration of the above, the oscillation pause time Tpp immediately before continuous pulse oscillation and the discharge voltage V corresponding to the power lock voltage Vpl are set in order to make the energy E of each pulse the same desired value Ed, and continuous pulse oscillation starts. It is obtained for each count number i of the time since the completion, and this is stored in a predetermined memory in advance as voltage data. Then, required voltage data is sequentially read out from the memory and output to the laser power supply unit 8 so that the processing is quickly performed.

That is, FIG. 2 is a flowchart showing the processing executed by the output control section 6. As shown in FIG. 2 (a), in the main routine simultaneously with the start-up, as shown in the following equation (2): A function Vi (Tpp, Tpp, Tp, Vpl, Tk, which is the discharge voltage corresponding to the pulse with the elapsed time i, and Tk (the elapsed time from the start of the continuous pulse oscillation).
Vpl, Tk), and these are stored in a memory as a voltage pattern table. In this case, the variable Tpp,
A voltage data sequence having the same value of Vpl (i = 1,
Are arranged in the order of 2, 3,...), And are stored in the memory in units of sets.

V (i) = V (Tpp, Vpl, Tk) (2) where V (i): the i-th voltage data in the voltage pattern table V: a functional expression for determining the voltage data Tpp: oscillation pause time Vpl: Power lock voltage Tk: Elapsed time from start of continuous pulse oscillation Here, voltage data V (1), V (2), V (3) ...
Is 1 when the maximum repetition frequency of the laser device is f.
It is stored for each time interval Tst equal to or less than the cycle of / f. (Step 101) The output control section 6 performs a predetermined interrupt process at the timing of receiving the trigger signal Tr and the timing of two types of timer interrupts. That is, when the above-described voltage pattern table creation and storage processing is completed, an interrupt acceptance state is set (step 102). At the same time as the output control unit 6 enters the interrupt accepting state, the process shifts to the timer interrupt routine A of FIG. In the timer interrupt routine A, the counter is sequentially incremented and the count number n is incremented by +1 until the next time the trigger signal Tr is sent after the previous trigger signal Tr is sent (step 104).

When the next trigger signal Tr is received, the process proceeds to the trigger interrupt routine shown in FIG. Here, first, the count-up in the timer interrupt routine A is stopped, the process of converting the count number n at that time into time is performed, and the converted time is set as the oscillation suspension time Tpp. For example, the timer interrupt interval is 10 ms
ec, and if the count number n is 15, the pause time Tpp is set to 150 msec by multiplying the two (step 106).

Subsequently, the count n of the timer is reset to zero (step 107). Next, it is determined whether or not the oscillation suspension time Tpp obtained in step 106 is equal to or longer than a predetermined upper limit value Tul. When the upper limit value Tul is longer than this, the effect of changing the pulse energy E of the spiking phenomenon is constant, and is determined by experiments or the like as being no longer dependent on the oscillation pause time, and stored in a predetermined memory. (Step 10
8). Further, the oscillation suspension time T obtained in step 106
It is determined whether pp is greater than or equal to a predetermined lower limit value Tbs. That is, if the time interval between pulse oscillations is sufficiently small, the influence of density disturbance or the like due to the immediately preceding pulse oscillation remains strongly in the discharge space, and the spiking phenomenon does not occur.
Therefore, for a shorter time, the lower limit time Tbs at which the spiking phenomenon does not occur is determined by experiments or the like.
It is stored in a predetermined memory (step 11
0).

Here, if it is determined that the oscillation suspension time Tpp is equal to or greater than the upper limit value Tul (determination Y in step 108)
ES), Tpp = Tul (step 109), and the count i of a voltage data timer counter (not shown) is set to 1 (step 111). At this point, the output control unit 6 recognizes that the next continuous pulse oscillation has started, and measures the elapsed time from the start of the continuous pulse oscillation in the timer interrupt routine B of FIG. The timer starts counting the voltage data. The timer counter for counting the elapsed time differs from the increment interval of the timer interrupt routine A, and the count number i is sequentially increased by +1 at the same time interval Tst set when the voltage pattern table of the formula (2) is created. It is incremented (step 105).

When the output control unit 6 recognizes that the continuous pulse oscillation has started, the current power lock voltage Vpl,
Oscillation pause time immediately before continuous pulse oscillation Tpp = Tul, i = 1
Is read from the memory, and the discharge voltage V (i) corresponding to
This is output to the laser power supply unit 8 and discharge is performed. As a result, the first pulse of the continuous pulse oscillation is made one in which the influence of the spiking phenomenon is removed, and a desired pulse energy Ed is obtained. Moreover, when Tpp is equal to or more than the upper limit value Tul, it is not necessary to store data of a plurality of discharge voltages V (i) corresponding to Tpp, and the discharge voltage V (i) is uniquely determined according to the constant value Tul. Since only the data needs to be stored in the memory, the storage capacity can be reduced and the cost can be reduced (step 113).

When the pause time Tpp is equal to or longer than the lower limit value Tbs,
If it is smaller than the upper limit value Tul (YES in step 110), the count i of the elapsed time timer counter is set to 1 in order to eliminate the effect of the spiking phenomenon again from the first pulse of the continuous pulse oscillation (step 110). Step 111). Then, the oscillation pause time Tpp immediately before the continuous pulse oscillation obtained in step 106, the current power lock voltage Vpl, and the discharge voltage V (i) corresponding to i = 1 are read out from the memory and output to the laser power supply unit 8. And discharge is performed. As a result, the first pulse of the continuous pulse oscillation has the effect of the spiking phenomenon removed,
A desired pulse energy Ed is obtained (step 11)
3).

If the pause time Tpp is smaller than the lower limit value Tbs (NO in step 110), it is determined that continuous pulse oscillation is continuing, and the time elapsed from the start of continuous pulse oscillation is detected. The count i of the timer counter for measuring the elapsed time is read (step 11).
2), i of the voltage pattern table used immediately before
The third voltage data V (i) is read from the memory, output to the laser power supply unit 8, and discharge is performed. As a result, the i-th pulse of the continuous pulse oscillation is made the one in which the influence of the spiking phenomenon is removed, and the desired pulse energy Ed is obtained (step 114).

As described above, in the embodiment, the voltage calculation formula (2) is a function formula using Tpp, Vpl, and Tk as variables. As described above with reference to FIG. 4, if the variables Tpp and Vpl are determined, The curve indicating the spike phenomenon occurrence pattern is uniquely determined. In other words, if the variables Tpp and Vpl are determined, the change pattern of the discharge voltage will also be determined. That is, as shown in FIG. 3 where the vertical axis indicates the discharge voltage V and the elapsed time τ from the start of the continuous pulse oscillation is the horizontal axis, the voltage data of the spiking prevention control is the same regardless of the repetition frequency. It will be on the curve in the figure. For this reason, a formula for calculating the voltage pattern curve of FIG. 3 from the elapsed time Tk is prepared for each set of (Tpp, Vpl), and the voltage data is calculated using the following function using only the elapsed time Tk as a variable. An implementation using equation (3) is also possible.

V (i) = Vtv (Tk) (3) where V (i): the i-th voltage data in the voltage pattern table Vtv: The voltage data prepared for each combination of Tpp and Vpl is determined. In the embodiment, the voltage data is stored at a pitch Tst such that Tst ≦ 1 / f, but the change in the repetition frequency of the continuous pulse oscillation is the maximum repetition frequency. In the case where f is near f, Tst = 1 / f may be set. Therefore, in this case, the pitch is the same as the pitch of the voltage data in the conventional technique, the conventional voltage pattern table can be used as it is, and the improvement can be performed at low cost.
Moreover, in this case, it is possible to accurately perform the spiking prevention control when the repetition frequency, which is impossible with the related art, is an integer fraction of the maximum repetition frequency f. In the embodiment, the discharge voltage V is obtained by a function in consideration of the oscillation pause time Tpp and the power lock voltage Vpl. However, spiking such as the elapsed time after newly filling the laser gas into the laser chamber is performed. The discharge voltage V may be obtained by a function using various parameters that are considered to affect the phenomenon occurrence pattern as variables.

[0045]

As described above, according to the present invention,
Since it is not necessary to store spiking occurrence prevention control data for each repetition frequency, cost reduction can be achieved, such as a small memory capacity. Spiking occurrence prevention control suitable for the repetition frequency will be performed accurately according to the elapsed time from the start of continuous pulse oscillation,
The reliability of the device can be further improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of an output control device of a laser device according to the present invention.

FIG. 2 is a flowchart illustrating a processing procedure according to the embodiment;

FIG. 3 is a graph showing a relationship between an elapsed time from the start of continuous pulse oscillation and a discharge voltage.

FIG. 4 is a graph showing a relationship between an elapsed time from the start of continuous pulse oscillation and pulse energy.

FIG. 5 is a graph used to show a burst mode and a spiking phenomenon, and is a graph showing how the pulse energy changes with time.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Laser apparatus 6 Output control part 10 Stepper control part

Claims (3)

(57) [Claims] The present invention is applied to a laser device that repeatedly performs an operation of suspending pulse oscillation for a predetermined time after continuously oscillating a laser beam for a predetermined number of times, and based on charging voltage data, the energy of the pulse has a predetermined magnitude. In the output control device of the laser device for controlling the charging voltage so that the charging voltage data for making the magnitude of the energy of each pulse of the continuous pulse equal to each other before the operation of the laser device is started, the continuous pulse oscillation is started. An output control device for a laser device, which stores in advance for each magnitude of a pause time immediately before continuous pulse oscillation in accordance with an elapsed time from the laser beam. 2. The output control of a laser device according to claim 1, wherein, when a maximum value of the repetition frequency of the continuous pulse oscillation is f, charging voltage data is stored at time intervals Tst such that Tst ≦ 1 / f. apparatus. 3. During operation of the laser device, an elapsed time from the start of continuous pulse oscillation is counted, and stored charging voltage data corresponding to the counted elapsed time is read out. 3. The output control device for a laser device according to claim 1, wherein the charging voltage is controlled so as to obtain the charging voltage.