JPH0564874B2 - - Google Patents

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a laser processing device, and is particularly used in the manufacturing process of semiconductor devices, and is applied when cutting and annealing circuit patterns of semiconductor integrated circuit devices. It is suitable.

[Summary of the invention]

The present invention provides, in a laser processing device having a laser as a processing light source, a permissible excitation time that allows the laser to oscillate stably, and laser excitation pulse energy is generated at a selected time point during the permissible excitation time. Accordingly, processing laser light can be generated without wasting time.

[Conventional technology]

2. Description of the Related Art As semiconductor integrated circuit devices become increasingly highly integrated, patterns produced on wafers or photomasks tend to become finer.
Therefore, even if a minute defect occurs in a pattern during the manufacturing process of a semiconductor integrated circuit device, this will cause the entire manufactured integrated circuit device to be defective.
As a result, there is a problem of poor yield.

One way to solve this problem, for example in the case of semiconductor memory, is to configure a spare circuit part as a redundant circuit in addition to the main circuit part that makes up the memory circuit that should be manufactured on each chip. A method has been proposed in which if a defect occurs in a part of the main circuit of a chip, by replacing it with a redundant circuit, the chip can be treated as a passing product, thus improving the yield of passing products. (Japanese Patent Application Laid-open No. 165338/1983).

The present applicant previously filed an application for an invention relating to such a processing device (Japanese Patent Application No. 44900/1983). This earlier application includes a sixth method for processing patterns using laser light.
A laser processing device as shown in the figure has been proposed. In FIG. 6, 1 is a laser constituting a processing light source, and the laser beam LA1 is transmitted to a total reflection mirror 2,
The laser beam is irradiated onto the aperture member 4 through the aperture member 3 and passes through the aperture 4A of the aperture member 4, and is imaged onto the wafer 6 by the projection lens 5. The circuit pattern formed on the wafer 6 is aligned with this laser light (this is called alignment), and the circuit pattern is thus processed.

The alignment of the wafer 6 with respect to the imaging position of the laser beam is performed by moving the stage 11 on which the wafer 6 is placed in directions orthogonal to each other, that is, in the X direction and the Y direction. This stage 11
The position coordinates of are separately detected by a length measuring means (not shown), which is a light wave interferometer using a laser beam, for example.

Integrated circuit patterns (referred to as chips) are formed on the wafer 6 so as to be arranged in a grid pattern, and alignment marks are placed in each chip in advance. This alignment mark is used for X alignment microscope 12 and Y alignment microscope 1.
3. Detected by the θ alignment microscope 14. In this device, the laser beam LA2 sent out from the laser 15 constituting the alignment light source is supplied to the X alignment microscope 12 via a half mirror 16 and a total reflection mirror 17, and is also supplied to the Y alignment microscope 12 via half mirrors 16 and 18. Microscope 1
3, and further half mirrors 16, 18,
The light is supplied to the θ alignment microscope 14 through a total reflection mirror 19, and illuminates the alignment mark. A photoelectric detector is provided in each of the microscopes 12, 13, and 14, and the position of the alignment mark can be detected by detecting diffracted light or scattered light arriving from the alignment mark. being done.

By the way, since the position of each alignment mark on the wafer 6 and the positional relationship of the processed parts in each chip are known, and the positional relationship of the projection lens 5 and each alignment microscope 12, 13, 14 is also known, the projection Based on the current position of the processing laser beam imaged by the lens 5, the distance from this imaging position to the defect position (therefore, the processing position) that has occurred in each chip (the amount of movement of the stage 11)
can be determined by calculation, and thus the processing laser beam can be positioned at the processing position with sufficient accuracy.

Note that the position of the defect occurring within the wafer 6 is inspected and confirmed by a separate inspection device, and based on the inspection result, calculation of the amount of movement to the processing position or positioning is executed.

[Problem that the invention seeks to solve]

In a laser processing device with such a configuration,
Conventionally, the waiting time from when the stage 11 stops at a predetermined processing position until the processing laser beam is irradiated onto the wafer 6 is relatively long, and this has been a factor that hinders an increase in throughput.

Lasers used as processing light sources (e.g.
(YAG laser) generally has a curve K in Fig. 7.
As shown by curve 1, the output variation (this is called output stability) reaches a minimum value at a specific frequency f ₀ , and at the same time, as shown by curve K2, the output reaches a maximum near the same frequency f ₀ . It has similar characteristics. Therefore, the laser 1 generally used as a processing light source is controlled to oscillate with an excitation frequency of f ₀ or a frequency near it, which provides good stability during oscillation and maximum output. Therefore, as shown in FIG. 8B, the laser 1 has a constant period T, regardless of the positioning operation of the stage 11.
A laser excitation pulse EP is given.

On the other hand, when positioning the stage 11 at a predetermined processing position, according to _the control pattern shown in _FIG . By doing so, the predetermined moving speed is increased, and then the stage 11 is moved to the processing target position at a constant speed. When the machining target position is finally approached, deceleration is started at time _t3 , and when the target value is reached at time _t4 , the stage 11 is stopped.

With the processing location of the wafer 6 positioned at the laser beam position in this way, a giant pulse is extracted from the laser 1 by the first laser excitation pulse EP generated after time _t4 , and this is used as the processing laser beam. Machining is performed by irradiating the work area with the beam.

In this way, in conventional equipment, stage 1
1, the processing laser beam is obtained by the first laser excitation pulse EP after completing the positioning of the processing location, so after the time t ₄ when the stage 11 stops at the target processing position, the first During the time ΔT until the laser excitation pulse EP is generated, it is necessary to wait without performing any processing operation on the wafer 6.

This average waiting time ΔT _w has a value of ΔT _w =T/2 (1), and especially when the excitation frequency f ₀ of the laser 1 is relatively low, for example, several tens [pps] (pulse
per second) In the following cases, the processing capacity of the laser processing device may be limited in practice based on this waiting time.

The present invention has been made in consideration of the above points, and
The present invention attempts to propose a laser processing apparatus that can shorten as much as possible the waiting time until the laser excitation pulse EP is generated after the stage 11 reaches the processing target position.

[Means for solving problems]

In order to solve this problem, the present invention includes a laser 1 as a processing light source, and an excitation pulse generator that generates an excitation pulse that generates laser excitation pulse energy at a period that allows the laser 1 to stably oscillate. means 26; positioning means 30, 31, 32 for positioning by changing the relative position of the workpiece 6 and the laser 1 in order to irradiate the workpiece 6 with the laser light emitted by the laser 1; After the positioning by 32 is completed, in order to generate laser excitation pulse energy immediately, the time required to complete the positioning is predicted, and the generation timing of the laser excitation pulse is determined based on the prediction result to ensure stable oscillation of the laser 1. A biasing means 26 for biasing within a range that does not cause damage is provided.

[Effect]

When the wafer 6 to be processed is positioned at the laser irradiation position, the generating means 26 can selectively generate laser excitation pulse energy for generating processing laser light during the excitation allowable time. . Therefore, compared to the conventional case where the processing laser beam can be generated only at the fixed generation point of the laser excitation pulse energy, the processing laser beam can be generated without wasting time. In this way, the throughput per unit time can be significantly improved.

〔Example〕

An embodiment of the present invention will be described in detail below with reference to the drawings.

First, the present invention is constructed so that the generation of a processing laser beam is synchronized with the timing when positioning of the stage 11 is completed based on the principle shown in FIG. That is, the laser excitation pulse EP for the laser 1 (FIG. 1A) is as described above with respect to FIG. 8B.
By generating at a predetermined reference period T, the laser 1 is made to operate stably. In addition, in the present invention, one (or more) of the EP of the laser excitation pulse Even if the timing of the pulse is shifted, as long as the period of the other laser excitation pulses maintains the predetermined reference period T, the laser 1 can be stably oscillated. It is detected whether or not the interval between the time point and the time _point to when the first laser excitation pulse EP is to be generated is within a predetermined excitation allowable time, and if the interval is within a predetermined excitation allowable time, the laser excitation pulse is
The timing is controlled so that the generation of EP coincides with the completion of positioning of the stage 11.

There are the following two conditions when controlling the timing of the laser excitation pulse EP.

In the first case, as shown in FIG. 1 B1, the positioning completion time t _o-1 of the stage 11 occurs before the time t _o of the reference period T. In this case, the laser 1 is emitted at the reference period T. It has an excitation allowable time of time ΔT ₁ for the corresponding time t _o . In other words, within the allowable excitation time ΔT ₁ , the laser 1 can stably oscillate even if the excitation pulse EP is generated singly at any timing. Therefore stage 11
The positioning completion point t _o-1 is this excitation allowable time ΔT ₁
When the laser excitation pulse EP is generated within ΔT _x1 from the reference time _to , as shown in FIG.
occurs at the previous time t _o-1 .

In the second case, as shown in Figure 1 C1,
In the case where the positioning completion time t _o+1 of the stage 11 occurs after the time t _o corresponding to the reference period T, in this case, the laser 1 is activated at the time corresponding to the reference period T.
It has an excitation allowable time of ΔT ₂ for t _o . Therefore, when the positioning completion time t _o+1 of the stage 11 occurs within this excitation allowable time ΔT ₂ , the positioning completion time t o+1 in FIG.
As shown in FIG. 2, the laser excitation pulse EP is generated at a time t _o+1 delayed by ΔT _x2 from the reference time t _o .

In this way, when the positioning of the stage 11 is completed at time t _{o -1} , processing on the wafer 6 can be performed at a time ΔT _x1 earlier than in the case of the reference period T.

On the other hand, when the stage 11 reaches the positioning completion state at time _to+1 , the laser excitation pulse EP is not generated at time _to , but is generated at a time delayed by ΔT _x2 . In contrast, according to the conventional method, processing is performed using the laser excitation pulse EP of the next reference period T delayed by T-T _x2 after the completion of positioning, whereas according to the present invention,
The wafer 6 can be processed at the same time as the positioning is completed.

The laser processing device 25 shown in FIG.
6, and the CPU 26 single-handedly controls the timing of the laser excitation pulse EP of the laser 1 as a processing light source based on the above-mentioned principle.

That is, the oscillation pulses of the oscillation circuit 27 are counted by the counter 28, and the count contents are supplied to the CPU 26 as time data DT. The contents of the time data DT are repeated every reference period T (Fig. 1), and at the timing when the count contents match the reference period T, the counter 28 is activated.
An interrupt signal IR (FIG. 5C) is supplied to the CPU 26. In this way, the CPU 26 is supplied with data representing the timing of the laser excitation pulse EP of the reference period T and the time position between each pulse.
The CPU 26 interrupts the processing from the main routine shown in FIG. 3 to the subroutine shown in FIG. 4, and supplies the laser excitation pulse EP shown in FIG. 5E to the laser 1.

In addition to this configuration, position data PD representing the current position of the stage 11 is transmitted from the coordinate measuring device 29.
It is supplied to the CPU 26. The CPU 26 calculates the relationship between the target position at which the stage 11 should be positioned and the current stage position until the coordinate value corresponding to the current position of the stage 11 reaches the target coordinate value corresponding to the target position. Drive control signal via digital-to-analog converter 30
Output RV. This drive control signal RV is input to a drive circuit 31 for a motor 32 as a drive source for the stage 11, and the motor 32 is controlled according to the control signal RV.

On the other hand, the CPU 26 sends a processing output timing signal PO (FIG. 5D) to the Q switch 35 provided in the output system of the laser 1. Here, the Q switch 35 realizes a shutter function for the output of the laser 1, and when the processing laser beam is irradiated from the projection lens 5 to the wafer 6, the Q switch 35
When receiving the processing output timing signal PO from
Laser for processing giant pulse from laser 1
It operates as if it were drawn out as LA1.

The laser processing device 25 in FIG. 2 is similar to the one shown in FIGS.
Excitation control for the laser 1 is executed according to the processing procedure shown in the figure.

First, in step SP1 of FIG. 3, the CPU 26 uses the position data obtained from the alignment mark on the wafer 6 as to where the wafer 6 is located on the stage 11, the arrangement data of the chips on the wafer 6, and the processing inside the chips. Calculations are performed based on the data regarding the position, and target coordinate values of the stage 11 at which the laser 1 can irradiate the processing position are determined. Subsequently, in step SP2, the CPU 26 inputs the position data PD representing the current coordinate values of the stage 11 from the coordinate measuring device 29, and calculates the distance to the target coordinate values. Based on the calculation result, CPU26
Then, the process moves to the next step SP3, and it is determined whether the stage 11 has reached the target coordinate value.

If a negative result is obtained in step SP3,
The CPU 26 moves to the next step SP4 and motor 3
The drive signal RV for SP2 is sent out, and the process returns to step SP2 described above. Thus, as shown in FIG. 5B, the CPU 26 accelerates the stage 11 (acceleration mode), causes it to travel at a constant speed (constant speed mode), and decelerates the stage 11 when it approaches the target coordinate value (deceleration mode). Motor 3 according to the speed pattern SPT to stop at
2, thereby controlling the position of the stage 11 (therefore, the contents of the position data PD of the coordinate measuring device 29) in a speed pattern as shown in FIG. 5A.
Movement is controlled to match the machining position to the target position according to SPT.

At the same time that the stage 11 starts moving toward the target value in this manner, the interrupt signal 1R is supplied from the counter 28 to the CPU 26 at intervals of time T.
In response to this, the CPU 26 executes step SP21 each time.
The interrupt routine shown in FIG. 4 is entered. Here CPU2
6 first performs stage 11 at step SP22.
Determine how close the coordinates are to the target coordinate values. That is, in step SP22, the distance from the current coordinate value to the target coordinate value of the stage 11 and the mode of the speed pattern SPT are determined, and it is determined whether the stage 11 can finish positioning within the allowable time ΔT ₂ shown in FIG. 1 due to its characteristics. Predict. step
If a negative result is obtained in SP22, it is not possible to complete positioning within the excitation allowable time _ΔT2 , so the CPU 26 moves to step SP23 and laser 1
After supplying excitation pulse EP to
Return to the original main routine at SP24.

On the other hand, if a positive result is obtained in step SP22, the CPU 26 immediately proceeds to step SP24.
, return to the original routine, and without generating the laser excitation pulse EP with the reference period T for laser 1,
The stage 11 enters a state of waiting for the stage 11 to reach the target coordinate value.

Incidentally, in this state, the stage 11 is approaching the target coordinate value, and the allowable time ΔT ₂ at time t _o in FIG.
This means that the stage 11 is predicted to reach the target coordinate value at some point during
Continue the movement control in step 1.

While the CPU 26 is performing the positioning operation of the stage 11 in this manner, the interrupt signal IR (FIG. 5C) is supplied to the CPU 26 from the counter 28 at a predetermined period T interval. , the CPU 26 receives a laser excitation pulse EP with a period T
continues to send out (Fig. 5E), which causes laser 1
is maintained in a state where the oscillation operation is repeated stably.

Eventually, at time t ₁₀ in Figure 5, stage 1
1 reaches the target coordinate value (FIG. 5A) and the stage 11 stops (FIG. 5B), the CPU 26 obtains a positive result in step SP3 and moves to step PS5.

In step SP5, the stage 11 stops after the start of the excitation allowable time ΔT ₁ and the end _{of ΔT 2 in} FIG. (this time is set as ΔT).

If a negative result is obtained here, it means that the laser 1 has not yet entered the allowable time ΔT ₁ or ΔT ₂ for stable operation (Fig. 1 B1), and therefore the CPU 26 receives a positive result. It will be in a waiting state until it is obtained. Eventually, the excitation allowable time ΔT ₁
Upon entering, the CPU 26 obtains a positive result in step SP5 and moves to step SP6, where it supplies the laser excitation pulse EP to the laser 1 and sends the processing output timing signal to the Q switch 35.
Processing is performed by supplying PO (FIG. 5D) and irradiating the wafer 6 with laser light using giant pulses.

Here, there are the following two states in which a positive result can be obtained in step SP5.

In the first case, step SP22 of the interrupt routine
This is when the CPU 26 is in a state of waiting for the stage 11 to stop because a positive result has been obtained in , and at this time the stage 11 stops at the allowable time ΔT ₂ (FIG. 1, C1 and C2). Thus
The CPU 26 does not generate a machining pulse at the reference time _to , but delays it until the stage 11 stops.

The second case is when the stage 11 stops within the allowable time ΔT ₁ (B1 and B2 in Figure 1),
At this time, the CPU 26 generates a machining pulse at a time before the reference time t _o .

As a result, the average waiting time ΔT _w from the completion of positioning until the laser 1 is excited becomes ΔT _w = T/2-1/2 (ΔT ₁ + ΔT ₂ ) ...(2), and in the conventional case ((1 ) can be further shortened.

In addition, in practice, the allowable time ΔT at step SP5
If a positive result is not obtained even though the period of time has elapsed, measures are taken such that the CPU 26 returns to the above-mentioned step SP2.

According to the above configuration, the CPU 26 operates the stage 11 during the allowable time ΔT (=ΔT ₁ +ΔT ₂ ) shown in FIG.
When the laser beam reaches the target position, the laser 1 can be immediately excited to generate a processing laser beam at that point. Therefore, unlike the conventional case, the generation of the laser excitation pulse EP with the reference period T is not necessary. No need to wait.

Note that in the above embodiments, the timing of generation of laser excitation pulse energy is shifted once, but the present invention is not limited to this, and also includes a case where the shift is performed in multiple steps. for example,
The timing at which the positioning of the stage 11 is completed and the timing at which the laser excitation pulse energy is generated based on the reference period T are significantly different from each other, and one deviation may disturb the stability of the laser oscillation operation, which attempts to synchronize both timings. In this case, this deviation is divided into multiple times and 1
By suppressing the amount of deviation of the timing to an amount that does not disturb the stability of the laser, the two timings can be synchronized without any problem.

〔Effect of the invention〕

As described above, according to the present invention, the processing laser beam is sent out immediately after the processing object is positioned at the irradiation position of the processing laser beam.
Processing laser light can be supplied with a significantly shorter waiting time than in the conventional case, and as a result, a laser processing apparatus with a much improved throughput per unit time can be easily obtained.

[Brief explanation of the drawing]

Fig. 1 is a signal waveform diagram showing the operating principle of the laser processing apparatus according to the present invention, Fig. 2 is a block diagram showing an embodiment of the laser processing apparatus according to the invention, and Figs. 3 and 4 are its data processing procedure. 5 is a signal waveform diagram showing the signals of each part in FIG. 2, FIG. 6 is a schematic perspective view showing the configuration around the stage of the laser processing device, and FIG. 7 is the output characteristics of the laser. FIG. 8 is a signal waveform diagram for explaining the operation of a conventional laser processing apparatus. 1...Laser, 6...Wafer, 26...CPU,
30...Digital-to-analog converter, 31...
Drive circuit, 32...motor.

Claims (1)

[Claims] 1. (a) A laser as a processing light source; (b) Excitation pulse generation means for generating an excitation pulse that generates laser excitation pulse energy at a period that allows the laser to operate stably. (c) positioning means that positions the workpiece by changing the relative position of the workpiece and the laser so that the workpiece is irradiated with laser light emitted by the laser; and (d) after the positioning by the positioning means is completed. In order to immediately generate the laser excitation pulse energy, predict the time required to complete the positioning, and vary the generation time of the laser excitation pulse according to the predicted time within a range that does not impair stable oscillation of the laser. A laser processing device characterized by having a biasing means for causing the laser beam to shift.