JPH04283606A - Gap measuring apparatus - Google Patents

Abstract

PURPOSE:To make it possible to perform continuous gap computation by reading a signal which is stored in another signal memory means when another signal is being stored in any of a plurality of signal memory means, and measuring the gap between one surface and the other surface. CONSTITUTION:A first MPU 3 has a first digital filter 31 which removes noises by the average of movement of a CCD signal Si in digital value and outputs a movement-average signal Ti, a first norm computing part 32 which reads the signal Ti stored in a first RAM 51 and computes the norm Nk and a first gap computing part 33 which computes the gap based on the norm Nk. A second MPU has a second digital filter 41, a second norm computing part 42 and a second gap computing part 42 which have the similar functions as in the MPU 3. In this constitution, the input of the signal Si is alternately repeated. Thus, the computing time is shortened extremely and the continuous gap computation can be performed.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention relates to the gap (
This invention relates to a device for measuring gap).

0002

[Prior Art] The gap measurement principle includes, for example, "Jo
Urnal of Vacuum Science a
nd Technology B1 (4) Oct-
Dec. 1983. P1196 D. C. ,Flan
ders and T. M. There is a method reported in ``Lyszcazrz''. As shown in FIG. 12, this is achieved by irradiating a mask 100 and a wafer 101 with laser light emitted from a laser 10, causing the light reflected on the mask 100 surface and the light reflected on the wafer 101 surface to interfere with each other. This method measures the gap g by detecting the interval p between interference fringes.

[0003] The gap can be measured by the method described in Japanese Unexamined Patent Publication No. 2-61508 (March 1, 1990), which uses this principle and is described in the X-ray mask wafer gap measurement method. This is shown in FIG. First, interference fringes are detected by CCD1, and A/
After being A/D converted by the D converter 2, it is taken into the MPU 3 via the RAM 20. Inside this MPU 3, a digital filter 31 is used from the input CCD signal Si.
Find the moving average signal Ti.

0004

[Math 1]

The noise is removed by this moving average and the moving average signal Ti is stored in the RAM 5. next,

000
6]

[Math 2]

The norm Nk expressed by
2 to emphasize the period of the interference fringe, and this norm Nk
The neighborhood of the minimum where k is the minimum among the minimums of is quadratic approximated by a parabola of N=ak2 +bk+c (3), and the period of interference fringes (interval of fringes) P is expressed as P=-(b/2a) ( Find it using 4). Period P of interference fringes, wavelength λ of laser light, CC from mask 100
When the gap g is calculated by the gap calculation unit 33 from the distance l to D1,

[0008]

[Math 3]

It can be found as follows. Figures 14 (A), (B), (
C) shows the respective signal waveforms, (A) shows the input CCD signal Si, (B) shows the moving average signal Ti from which noise has been removed by a digital filter, and (C) shows the norm NK. . Next, the operation of the conventional device based on the above configuration will be explained based on FIGS. 15 and 16. FIG. 15 is a flowchart of each operation of A/D (analog/digital) conversion and moving average, and FIG. 16 is a flowchart of operation of norm calculation.

In FIG. 15, the entire apparatus is initialized, and the cell number i of the image sensor of the CCD 1 is set to 0 (step 100). The CCD signal Si input from the CCD 1 is A/D converted by the A/D converter 2 and stored in the RAM.
20, and the cell number i is incremented (step 101). The CCD signals Si for the total number (total number of bits) n of each image sensor of the CCD 1 are all A/D
It is converted and stored in the RAM 20. Here, all C
It is determined whether or not the CD signal Si has been captured based on i≧n (step 102).

In step 102, all the CCDs
When the CCD signal Si from 1 is taken in, the moving average calculation executed by the digital filter 31 is initialized by cell number i←0, and the moving sum which is the sum of the data output from the cell to be subjected to the moving average. This is done by setting the value W←0 (swipe step 103). In this moving average calculation, find the first term (initial term) and increment the cell number i (
Step 104), it is determined whether or not the number is subject to the moving average if i≧2r+1 (Step 105). In this step 105, r refers to the number of cells before and after the cell of interest.

When the preprocessing of the moving average is completed in steps 104 and 105, the moving average calculation is executed for the next term after the initial term, and the moving average signal Ti is stored in the RAM 5.
(step 106). The operation in step 106 is performed by determining whether the total number of cells n of the CCD 1 has been reached (
Step 107), if it is determined that the moving average signal Ti has not been reached, repeat the moving average calculation operation of step 6 and store the moving average signal Ti in the RAM 5, and if it is determined that the moving average signal Ti has been reached, perform the moving average calculation operation. end.

Furthermore, in the norm calculation operation in the norm calculation unit 32, the norm calculation is initialized by k←1 (step 200). Here, k indicates the number of times (1 to L) to perform norm calculation, and is used to obtain norms N1 to NL. Further, when performing an arbitrary norm calculation, the previous norm calculation state is initialized as i←0, NK←0 (step 201). This step 200,
201 completes all initialization of norm calculation.

The norm calculation unit 32 calculates NK ←Nk + (T
Calculate the norm based on k+i −Ti )2 and increment the cell number i to read the next moving average signal Ti from the RAM 5 (step 202).
. By this increment, it is determined whether the cell number i has reached the norm cell number m that is the target of norm calculation (step 203), and if it is determined that it has not reached the norm cell number m, the norm calculation in step 202 is repeated and the norm cell number m has been reached again. If it is determined that it is, the process moves to the next step 204. In this step 204, the norm number k is incremented to calculate the next norm Nk (step 204
). It is determined whether or not the number of norm calculations k has reached the number L+1 by this increment (step 205). If it is determined that it has not reached the number L+1, the process returns to step 201 and repeats steps 202 to 205. Further, when the norm number k reaches the number L+1, the norm calculation operation ends, and the gap calculation section 33 calculates the gap g based on each of the norms N1 to NL.

FIG. 17 shows a block diagram of another conventional mask wafer gap measuring device. In the same figure, CCD1
is a Z stage 1 that adjusts the distance between the mask 100 and the wafer 101 together with the position sensor 11 and the actuator 12;
Configure 0. The CCD 1 and the position sensor 11 convert analog CCD signals Si and sensor data Z into A/D converters.
It is converted into a digital value by converters 21 and 22 and M
Output to PU7 side. Further, the actuator 12 is M
The digital value control signal output from PU7 is D/A
The converter 23 converts it into an analog value and inputs it.

The MPU 7 calculates a moving average signal Ti using a digital filter 71 based on the CCD signal Si of the input digital value, and calculates a moving average signal Ti using the digital filter 71.
The norm calculation section 73 calculates the norm Nk from the above, and the cap calculation section 74 calculates the gap g. Further, the target value calculation unit 75 calculates the difference ω between the sensor data Z from the position sensor 11 and the gap g, and adds this difference ω to the gap target value gr that was previously provided to obtain the Z-axis target value. Zr
Calculate. This Z-axis target value Zr is input to the actuator 12 as a control signal via the feedback control unit 76 and the D/A converter 23, and the actuator 12 performs drive control to match the detected gap g with the target gap gr. be done.

[0017]

[Problems to be Solved by the Invention] In the above method, it is necessary to calculate a large number of norms Nk in order to find the minimum of the norm NK, but as shown in equation (2), the norm Nk is the value of the moving average signal Ti. Since it is the sum of the squares of differences, a large amount of calculation time is required. Also, the CCD signal S
The analog signal from CCD 1 that is output every clock is A/D converted to the input of i, but due to the clock of CCD 1 or the conversion time limit of A/D converter 2 (or 21), MPU 3 (or 7) You need to wait a while to input data. Therefore, it takes time from inputting the CCD signal Si to calculating the gap g, resulting in a reduction in the throughput of the exposure apparatus. Furthermore, it takes even more time to increase the accuracy of the gaps by continuously calculating the gaps g and taking the average of them. For these reasons, there has been a problem that the throughput of the exposure apparatus is reduced.

The present invention has been made to solve the above problems, and an object of the present invention is to propose a gap measuring device that can perform continuous gap calculations by following input signals.

[0019]

[Means for solving the problem] The invention according to claim 1 includes:
Projecting light onto one surface (100) and the other surface (101),
The norm (Nk) is calculated for the signal (Si) that is sequentially input regarding the interference fringes formed by the light reflected on each surface (100, 101), and the norm (Nk)
A gap measuring device that calculates the period (P) of the interference fringes from the minimum value of and measures the gap between the one surface (100) and the other surface (101) based on the period (P) of the interference fringes. is provided with a plurality of signal storage means (51, 52...) for sequentially storing the signal (Si) for each predetermined amount of data, and stores the signal (Si) in any one of the plurality of signal storage means (51 or 52...). ) while storing other signal storage means (
The signal (Sj) already stored in 52... or 51)
is read out to measure the gap between the one surface (100) and the other surface (101).

[0020] The invention according to claim 4 has one aspect (100).
and another surface (101), and each surface (100
, 101), the norm (Nk
), calculate the period (P) of the interference fringe from the minimum value of the norm (Nk), and calculate the period (P) of the interference fringe.
In the gap measuring device that measures the gap between the one surface (100) and the other surface (101) based on the above, each partial signal (S2 or S3, ~Sm) is input, based on the already input signal Si'(i'=1, ~, m-1) and some of the input signals (S2 or S3, ~Sm). This is used to calculate part of the norm calculation.

The invention of claim 5 projects light onto one surface (100) and the other surface (101), and each surface (100, 1
The norm (Nk) for the signal (Si) that is sequentially input regarding the interference fringes formed by the light reflected at 01)
The period (P) of the interference fringe is calculated from the minimum value of the norm (Nk), and the one surface (100) and the other surface (101) are determined based on the period (P) of the interference fringe. In a gap measuring device that measures a gap between
3, ~Sm) is input, it moves from the already input signal Si'(i'=1, ~, m-1) and some of the input signals (S2 or S3, ~Sm). An average value is obtained, and the norm is calculated based on the moving average value.

[0022] The invention according to claim 6 has one aspect (100).
and another surface (101), and each surface (100) is irradiated with light.
, 101), the norm (Nk
), the vicinity of the minimum value of the norm (Nk) is approximated by a curve to calculate the period (P) of the interference fringe, and the period (P) of the interference fringe is calculated based on the period (P) of the interference fringe. ) and another surface (101), the gap measuring device includes fast Fourier transform means () for calculating an approximate period (P') of interference fringes by fast Fourier transform of the input signal (Si). 62), the norm (Nk) is calculated only for a specific range based on the approximate period (P') of the interference fringes, and the gap between the one surface (100) and the other surface (101) is calculated. It is used to measure.

[0023] The invention according to claim 7 has one aspect (100).
and another surface (101), and each surface (100
, 101), the norm (Nk
), the vicinity of the minimum value of the norm (Nk) is approximated by a curve to calculate the period (P) of the interference fringe, and the period (P) of the interference fringe is calculated based on the period (P) of the interference fringe. ) and another surface (101), when the first surface (100) or the other surface (101) is newly loaded, the norm (Nk) is calculated over a wide range. The period (P) of the interference fringe is calculated by approximating the vicinity of the minimum value of the norm (Nk) with a curve, and the one surface (100) or the other surface (101) is not newly loaded. In this case, the period of the interference fringes can be determined from the relationship between the sensor data (Z) of the position sensor that detects the position of the stage that adjusts the distance between one surface (100) and the other surface (101), and the already measured gap g. (P') and calculates the norm Nk within the range necessary for approximating the curve based on the estimated period (P') of the interference fringes.

[0024]

[Operation] In the invention described in claim 1, the signal (Si) is sequentially stored in a plurality of signal storage means (51, 52, . . . ), and any one of the signal storage means (51 or 52, . . . ) (Si) while storing the signal (Si) already stored in the other signal storage means (52,... or 51).
;j<i) and measure the gap, when calculating the gap continuously based on the sequentially input signal (Si), the output of the gap calculation value can be approximately equal to the input of the signal (Si). Follow and output.

In the invention according to claim 4, the signal (S
i ), while inputting the already input signal (Si
′;i′<i) and the portion of the input signal (Si), the norm calculation can be performed at the same time as each portion of the signal Si is input. This makes it possible to shorten the time from inputting the signal Si to calculating the gap as much as possible.

In the invention according to claim 5, the signal (S
i ), while inputting the already input signal (Si
';i'<i) and the input signal (Si), and by calculating the norm based on the moving average value, the time required to calculate the gap is shortened. In particular, when an analog value signal (Si) is input, the moving average is performed during the waiting time for A/D converting the signal (Si), so that the time required to calculate the moving average can be shortened.

In the invention as set forth in claim 6, the approximate period of the interference fringe is obtained by the fast Fourier transform means, and the norm is calculated for a specific range based on this approximate period of the interference fringe. The number of pieces can be reduced, and the time required for gap calculation can be shortened. In the invention according to claim 7, by changing the range of norm calculation depending on whether or not there is new loading on one side or the other side,
The number of norms to be calculated can be reduced, reducing the time required for gap calculation.

[0028]

Embodiments (1) An embodiment of the first invention FIG. 1 shows a block diagram of the present embodiment. In the figure, the gap measuring device according to the present embodiment includes a CCD 1 that receives each laser beam reflected from a mask 100 and a wafer 101, detects interference fringes of each laser beam, and outputs a CCD signal Si; An A/D converter 2 that converts an analog value of the CCD signal Si into a digital value, and first and second converters that alternately calculate the gap g between the mask 100 and the wafer 101 based on the CCD signal Si of the digital value. MPU
3 and 4, and first and second RAMs 51 and 52 that store a moving average signal Ti for calculating the gap g in the first and second MPUs 3 and 4, respectively.

[0029] The first MPU 3 inputs a digital value C
a first digital filter 31 that removes noise from the CD signal Si by a moving average and outputs a moving average signal Ti; and a moving average signal T stored in the first RAM.
Norm calculation unit 3 that reads i and calculates norm Nk
2, and a first gap calculation unit 33 that calculates the gap g based on the norm Nk. Further, the second MPU 3, like the first MPU 3,
Second digital filter 41, second norm calculation section 4
2. The configuration includes a second gap calculation section 43.

Next, the operation of the apparatus of this embodiment based on the above configuration will be explained based on FIGS. 2(A) and 3. In FIG. 2, ■ (or ■') indicates the input operation of the CCD signal Si;
■ (or ■') indicates the operation of calculating and storing the moving average signal Ti, ■ (or ■') indicates the operation of calculating the norm Nk, and ■ indicates the operation of calculating the gap g. Said CCD1 and A/D
Each of the converters 2 operates in the same manner as the conventional device, and sequentially converts the digital CCD signal Si to the first and second MPs.
Input to U3 and 4.

The first MPU 3 inputs a predetermined number of CCD signals Si (the predetermined number of image pickup elements of the CCD 1), and
') In parallel with this input operation (■'), the CCD signal Si
The moving average value of is calculated by the first digital filter 31 and the moving average signal Ti is stored in the first RAM 51 (
■′). Based on this moving average signal Ti, the first norm calculation section 32 starts calculating the norm Nk (■).

While the first norm calculation unit 32 is performing the norm calculation (■), the second MPU 4
The CCD signal Si sequentially output from the D converter 2 is input (■'), and the moving average signal Ti of the input CCD signal Si is calculated by the second digital filter 42, and the moving average signal Ti is converted into the second digital filter 42. Store it in the RAM 52 (■'). Based on this moving average signal Ti,
The norm calculation unit 42 starts calculating the norm Nk (
■).

Furthermore, while the second norm calculation section 42 is performing norm calculation (■), the first MPU
3 completes the norm calculation (■) and gap calculation (■) and inputs the next CCD signal Si. By alternately repeating the input of this CCD signal Si in the first and second MPUs 3 and 4, the calculation time required for each of the first and second norm calculation sections 32 and 42 is reduced as much as possible, and the calculation time is continuously reduced. It becomes possible to calculate the gap g. As described above, in this embodiment, the time TA (
(see FIG. 2(A)) is shorter than the time TB of the prior art shown in FIG. 2(B), and the gap calculation time is shortened by the corresponding time.

As shown in FIG. 3, the CCD signal Si
The input operation of , the calculation operation of the moving average signal Ti, the calculation operation of the norm Nk, and the calculation operation of the gap g will be explained next. The cell number i of the CCD signal Si is set to "0" and the moving addition value W is set to "0" to initialize the entire device (step 10). First, this CCD signal Si is input to the first MPU 3 via the A/D converter 2 to increment the cell number i and update the moving addition value W (
Step 11), it is determined whether the cell number i has reached the moving average number 2r+1 (step 12). Step 1 until the moving average number 2r+1 is reached in this step 12.
1 is repeated, and when it is determined that the number has been reached, the number of CCD signals Si corresponding to 2r+1 is input, the moving average signal Ti is calculated by Ti -r-1←W/(2r+1), and the moving sum value W is calculated. is updated, and the cell number i is incremented (step 13). Further, the calculation of the moving average signal Ti is repeated until the number n required for norm calculation is reached (step 14).

(2) Another embodiment of the first invention FIG. 4 shows a block diagram of another embodiment. In the same figure, the gap measuring device according to another embodiment includes a CCD 1, an A/D converter 2, a digital filter 31, a norm calculating section 32, and a gap calculating section 33, similar to the conventional device shown in FIG.
In addition to this configuration, first and second filters sequentially and alternately store the moving average signal Ti output from the digital filter 31 and output the stored moving average signal Ti to the norm calculation section 32. The MPU 3 includes the digital filter 31, the norm calculation section 32, and the gap calculation section 33, and has two tasks T1 and T2.

The relationship between the tasks T1 and T2 of the MPU 3 is the same as the relationship between the first MPU 3 and the second MPU 4 in the embodiment shown in FIG. (3) An embodiment of the second invention The gap measuring device according to this embodiment includes a CCD 1, an A/D converter 2, a digital filter 31, a norm calculation unit 32, and a gap calculation unit, similar to the conventional device shown in FIG. Part 3
3 and RAM 5, and the norm calculation operation of the norm calculation unit 32 is different.

The norm calculation unit 32 receives CCD signals Si (i
Moving average signal Ti (i=1 to m) corresponding to the moving average signal Ti (i=1 to m
) of each portion of the moving average signal T2 (or T3 , ~, T
m) is input, the already input moving average signal T1 (or T2, ~, Tm-1) and a part of the input moving average signal T2 (or T3, ~, Tm)
The configuration is such that part of the norm calculation is sequentially calculated based on the following.

Next, the operation of this embodiment based on the above configuration will be explained based on FIGS. 5 and 6 and with reference to FIG. 13 of the conventional device. FIG. 5 is an operation flowchart of this embodiment, and FIG. 6 is an explanatory diagram of the calculation development of norms N1 to NL, in which l norms Nk with norm numbers k=1 to L are obtained by taking the sum of m signals. be. In each of the above figures, initialization is performed for each calculation by the number of times L of norm calculations (step 20), and the CCD signal S1 is inputted via the A/D converter 2 and stored in the RAM 5 as a moving average signal T1 by the digital filter 31. , the norm calculation unit 32 reads out the moving average signal T1 and calculates the norm N.
1. Calculate the first term of . Furthermore, CCD signal S2
is converted into a moving average signal T2 by the digital filter 31.
When stored in the RAM 5 as
2 is read out by the norm calculation unit 32 and the second
term and the first term of norm N2, respectively. Also. C
When the CD signal S3 is stored in the RAM 5 as a moving average signal T3 by the digital filter 31, the norm calculation unit 32 reads out the moving average signal T3 and calculates the third term of the norm N1, the second term of the norm N2, and the norm N3.
Calculate the first term of each. In this way, the norm NL
When the CCD signal SL is input, calculations are performed from the Lth term of norm N1 to the first term of norm NL, and when the CCD signal Sm is input, calculations are performed from the Lth term of norm N1 to the first term of norm N1. Calculate from the mth term to the m-L+1th term of the norm Nn, and when the CCD signal Sm+1 is input, calculate from the mth term of the norm N2 to the m-L+2th term of the norm Nn, CCD signal S
When m+L-1 signals are input, the m-th term of norm NL is calculated. From the above, the flowchart of norm calculation is as shown in FIG.

The above operation will be specifically explained by substituting specific numerical values. Here, if the number of norm calculations L=3 and the total number of signals m=5, the result will be as shown in FIG. 7, and the flowchart of FIG. 5 will be operated m+L-1=7 times. Also, C
The moving average signal Ti corresponding to the CD signal Si will be described. In FIGS. 5 and 7, N1, N2, N3
Initialize the device by assigning O to (step 20)
, the cell number i of the input CCD signal Si starts from 1 (step 21). The moving average signal T1 corresponding to this CCD signal S1 is read out from the RAM 5 (step 22), and i=1 is compared with L=3. Since i is smaller than L (step 23), the final number k for norm calculation is
e (that is, the value of k at which addition of the norm Nk ends) is set to 1 (=i) (step 24). Next, i=1 is m
=5 (step 26), the initial number ks for norm calculation (that is, the value of k at which the addition of the norm Nk starts) is set to 1 (step 27).

The above norm calculation is performed for the norm Nk of number k from the initial number ks to the final number ke.
←Nk + (Ti - Ti-k )2 is calculated (step 29). That is, in step 29, the norm N1
←N1 + (T1 - To)2. moreover,
The cell number i is incremented to i=2 (step 30), and it is determined whether this cell number i=2 and m+L=8 match (step 31). The process returns to step 22 and repeats the above operation.

The operations after step 22 are as follows: reading the moving average signal T2 from the RAM 5 (step 22);
=2 and L=3, it is determined that cell number i=2 is smaller (step 23), and in step 24, the final number ke
is assumed to be 2. Further, in step 26, it is determined that cell number i=2 is smaller than m=5, and in step 27, the initial number ks is set to 1. In step 29, from the initial number ks to the final number ke, the norm Nk
When calculating, N1 ←N1 + (T2 - T1 )2
and N2 ←N2 + (T2 - T0 )2,
In step 30, increment the cell number i to i=3
shall be. Since this cell number i=3 is not equal to m+L=8 (step 31), the process returns to step 22 and repeats the above operation.

Similar to the above operation, when the moving average signals T4, T5 and T6 are read out in step 22, it is determined in step 26 that cell number i=6 is greater than m=5, and the process moves to step 28. This step 2
In step 8, the initial number ks is set to i-m+1=2, and in step 29, the norm N2 ←N2 + (T6 - T4 )
2 and norm N3 ←N3 + (T6 - T3 )2
seek. Furthermore, when the moving average signal T7 is read in step 22, the initial number of times ks is read out in step 28.
Set i-m+1=7-5+1=3, and in step 29 find the norm N3 ←N3 + (T7 - T4 )2,
At step 30, i←8 is determined, and the operation ends at step 31.

(4) An embodiment of the third invention FIGS. 8 and 9 show a gap measuring device according to an embodiment of the third invention. FIG. 8 shows a block diagram of this embodiment, and FIG. 9 shows an operation flowchart of this embodiment. In the same figure, CC
D1, A/D converter 2, digital filter 61 (
31) and a gap calculation unit 65 (corresponding to 33), and in addition to this configuration, a fast Fourier transform (FF
T; Fast Fourier Transforma
tion ) unit 62 and a norm range calculation unit 63.

First, a CCD signal Si is input from the CCD 1 and converted into a digital CCD signal Si by the A/D converter 2 (step 40). This CCD signal Si is input to the digital filter 61, and the above formula (
A moving average signal Ti is calculated based on 1) to remove noise components, and this moving average signal Ti is subjected to FFT section 62.
and is output to the norm calculation unit 64 (step 41). This FFT unit 62 extracts the frequency component of the moving average signal Ti (step 42), and searches for the frequency fmax having the maximum value component of this frequency component (step 43).
). Since an approximate value is sufficient for this frequency fmax search, there is no need to search for all the data of the moving average signal Ti, and FF
T can be performed, and the calculation time of the FFT unit 62 can be further shortened.

Furthermore, the norm range calculation unit 63 multiplies the reciprocal of the frequency fmax by an appropriate constant ds or constant de to calculate an initial number ks corresponding to the norm calculation start position and a final number ke corresponding to the norm calculation end position. (step 44). Initial number ks that falls within this norm range
and the final number ke, the norm calculation unit 64 calculates the moving average signal Ti output from the digital filter 61.
The norm Nk is calculated for (step 45). The gap calculation unit 65 calculates the calculated norm Nk by 2
The next approximation is obtained (step 46), and the period P and gap g of the interference fringes are calculated (step 47).

From the frequency fmax found in the FFT section 62 in this way, the norm calculation number k required for quadratic approximation is
This means that the cap g can be found by calculating the norm Nk only within the range of the norm calculation number k. Therefore, the FFT unit 62 calculates n×log for the number n of data of the input CCD signal Si.
Since the period can be calculated by multiplication and addition/subtraction corresponding to 2 n, calculation can be performed in a shorter time than calculating all the norms Nk.

(5) An embodiment of the fourth invention FIGS. 10 and 11 show a gap measuring device according to an embodiment of the fourth invention. FIG. 10 shows a block diagram of this embodiment, and FIG. 11 shows a flowchart of this embodiment. In the figure, the gap measuring device according to this embodiment includes a Z stage 10, A/D converters 21, 22, and a D
This configuration includes a /A converter 23 and an MPU 7, and in addition to this configuration, the MPU 7 includes a norm range calculation section 72 that calculates a range for norm calculation.

New mask 100 or wafer 101
When the mask 100 and the wafer 10 are loaded into the exposure apparatus, the mask 100 and the wafer 10
1 gap g and position sensor 1 on Z stage 10
The relationship with sensor data Z of No. 1 is arbitrary. That is, this is because the thickness of the mask 100 or wafer 101 differs each time, and there are variations in the mounting of the mask 100 or wafer 101. However, once the mask or wafer is loaded, there is a certain relationship between the gap g and the sensor data Z of the position sensor 11. In addition, since the norm Nk is quadratic approximated by equation (2), if only the norm Nk of the neighborhood to be quadraticly approximated is calculated, the norm Nk to be calculated is
Since the total number of is reduced, the total computation time is also reduced. Therefore, only when a new mask 100 or wafer 101 is loaded into the exposure apparatus, the minimum position of the norm Nk is searched for using the conventional method, and the gap g is calculated by quadratic approximation in the vicinity of the minimum position. From the first time onwards, the relationship between the gap g and the sensor data Z of the position sensor 11 at that time is memorized, the range of the number k of norm calculations required for quadratic approximation is calculated from the result, and the range of the number k of norm calculations is calculated from the result. only norm Nk
Calculate and find the gap g.

Further, the operation will be specifically explained in detail with reference to FIG. When a new mask 100 or wafer 101 is loaded into the exposure apparatus, the mask 100 or wafer 10
Since the relationship between the gap g between 1 and the sensor data of the position sensor 11 is unknown, the norm Nk must be calculated using a wide range of norm calculation number k from 1 to kmax. Therefore, the initial number ks for norm calculation =
1 and the final number ke = kmax (step 50). After this setting, the CCD signal Si is sent from the CCD 1 to the digital filter 71 via the A/D converter 21.
(step 51), and this digital filter 7
1, the moving average signal Ti is calculated based on the above equation (1), and noise is removed by the moving average (step 52).
.

Based on the moving average signal Ti, a norm Nk is calculated in the range from the initial number ks to the final number ke (step 53). It is determined whether this norm calculation is the first time or not (step 54). If it is determined that it is the first calculation, the minimum value where k is the minimum among the minimum values of the norm Nk is searched for, and the norm number k at this time is determined. ko (step 55). The range of quadratic approximation is calculated by multiplying this minimum number ko by constants es and ee (step 56). This range has an initial number ks = es ko and a final number k
It will be specified by e = ee ko .

The norm Nk is quadratic approximated in the above range (step 57), and the period P (=b/2a) of the interference fringes and the cap g (obtained by the above formula (5)) are calculated and sent to the target value calculation section 75. Output (step 58). Here, sensor data Z from the position sensor 11 of the front Z stage 10 is
It is input to the norm range calculation section 72, target value calculation section 75, and feedback control section 76 via the D converter 22 (step 59). The target value calculation unit 75 calculates the difference ω between the sensor data Z and the gap g (step 60).
, the Z stage target value Zr (=gr +ω) is calculated by adding this difference ω and a pre-existing gap target value gr (step 61). This Z stage target value Zr
is input to the feedback control unit 76, and after being feedback-controlled based on the sensor data Z, the actuator 12 is driven and controlled via the D/A converter 23 (step 62).

Furthermore, the norm range calculation unit 72 calculates the sensor data Z detected by the position sensor 11 after the drive control.
' is input (step 63), and the period P' of interference fringes that will occur at this time is calculated (step 64). Further, the norm range calculation unit 72 calculates the period P′ of this interference fringe.
A norm range for calculating the norm Nk is calculated by multiplying the initial number ks and final number ke, which are ranges to be quadratic-approximated, by constants ds and de (step 65). This ks = (ds P'), ke (=de
Based on the norm calculation range specified by P'), the process returns to step 51 and the norm calculation is repeated.

[0053]

Effects of the Invention As explained above, in the invention according to claim 1, a plurality of signal storage means (51, 52, . . . )
While one of the signal storage means (51 or 52,...) is storing the signal (Si), the other signal storage means (52,... or 51) has already been stored with the signal (Si). By reading out the stored signal (Si) and measuring the gap, when calculating the gap continuously based on the sequentially input signal (Si), the output of the gap calculation value can be used as the output of the signal (Si). It has the effect of outputting almost the same as the input.

In the invention according to claim 4, the signal (S
i ), while inputting the already input signal (Si
′;i′<i) and the portion of the input signal (Si), the norm calculation can be performed at the same time as each portion of the signal Si is input. This has the effect of reducing the time from inputting the signal Si to calculating the gap as much as possible. In the invention according to claim 5, while inputting the signal (Si), a moving average value is calculated from the already inputted signal (Si';i'<i) and the inputted signal (Si). By calculating the norm based on the moving average value,
Reduce the time it takes to calculate the gap. In particular, when an analog value signal (Si) is input, moving average is performed during the waiting time for A/D conversion of the signal (Si).
This has the effect of reducing the time required to calculate the moving average.

In the invention as set forth in claim 6, the approximate period of the interference fringe is determined by the fast Fourier transform means, and the norm is calculated for a specific range based on this approximate period of the interference fringe. The number can be reduced, which has the effect of shortening the time required for gap calculation. In the invention according to claim 7, by changing the range of norm calculation depending on the presence or absence of new loading on one surface or the other surface, the number of norms to be calculated can be reduced, and the time required for gap calculation can be reduced. has.

[Brief explanation of the drawing]

FIG. 1 is a block configuration diagram for explaining a device according to an embodiment of the first invention.

FIG. 2 is an operation timing chart for explaining the embodiment shown in FIG. 1;

FIG. 3 is an operational flowchart of moving average signal output for explaining the embodiment shown in FIG. 1;

FIG. 4 is a block configuration diagram for explaining another example device of the first invention.

FIG. 5 is an operation flowchart of an embodiment of the apparatus of the second invention.

6 is a diagram illustrating the development of norm calculation for explaining the norm calculation of the embodiment device shown in FIG. 5; FIG.

[Fig. 7] Norm calculation N1, N2, in the embodiment described in Fig. 5;
FIG. 3 is a diagram illustrating a specific calculation development of N3.

FIG. 8 is a block configuration diagram for explaining an embodiment of the third invention.

FIG. 9 is an operation flowchart of the embodiment shown in FIG. 8;

FIG. 10 is a block configuration diagram for explaining an embodiment of the fourth invention.

FIG. 11 is an operation flowchart of the embodiment shown in FIG. 10;

FIG. 12 is a schematic explanatory diagram of a general gap measurement principle.

FIG. 13 is a block diagram of a conventional gap measuring device.

FIG. 14 is a signal waveform diagram of a CCD signal, a moving average signal, and a norm.

FIG. 15 is an operation flowchart for calculating a moving average signal in a conventional device.

FIG. 16 is an operational flowchart of norm calculation in a conventional device.

FIG. 17 is a block diagram of another conventional device.

[Explanation of symbols]

1...CCD 2...A/D converter 3, 4...MPU 10...Z stage 11...position sensor 12...actuator 21, 22...A/D converter 23...D/A converter 31, 41, 61, 71...digital filter 32 ,4
2, 64, 73...norm calculation section 33, 43, 65, 74
... Gap calculation section 51, 52 ... RAM 62 ... FFT section 63, 72 ... Norm range calculation section 73 ... Norm calculation section 75 ... Target value calculation section 76 ... Feedback control section

Claims (7)

[Claims] Claim 1: One side (100) and the other side (101)
For the interference fringes formed by the light reflected from each surface (100, 101), the norm (Nk) is calculated for the signal (Si) that is sequentially input, and the minimum of the norm (Nk) is calculated. From the value, the period of the interference fringe (P)
In a gap measuring device that calculates the gap between the one surface (100) and the other surface (101) based on the period (P) of the interference fringes, the signal (Si) is converted into a predetermined amount of data. A plurality of signal storage means (51, 52
), any one of the plurality of signal storage means (51
or 52...), the signal (Sj) already stored in the other signal storage means (52... or 51) is read out and the signal (Sj) is stored on the one surface (100) and the other signal storage means (52... or 51). A gap measuring device characterized in that it measures a gap between a surface (101) and a surface (101). 2. The gap measuring device according to claim 1, wherein the signal storage means (51, 52...) is provided corresponding to the plurality of signal storage means (51, 52...) and stored in each signal storage means (51, 52...). (Si) and the first surface (100)
A gap measuring device characterized by comprising a plurality of calculation processing means (3, 4, . . . ) for calculating the gap between the surface and another surface (101). 3. In the gap measuring device according to claim 1, during the waiting time while inputting the signal (Si) to any one of the plurality of signal storage means (51, 52, . . . ), Arithmetic processing of reading the signal (Si) from other signal storage means (51, 52...) into which input of the signal (Si) has been completed and calculating the gap between the one surface (100) and the other surface (101). A gap measuring device comprising means (3). Claim 4: One side (100) and the other side (101)
For the interference fringes formed by the light reflected from each surface (100, 101), the norm (Nk) is calculated for the signal (Si) that is sequentially input, and the minimum of the norm (Nk) is calculated. From the value, the period of the interference fringe (P)
In a gap measuring device that calculates the gap between the one surface (100) and the other surface (101) based on the period (P) of the interference fringes, the input signal Si (i=
1 to m) each part of the signal (S2 or S3, ~Sm
) is input, the already input signal Si ′(i
'=1, ~, m-1) and some input signals (S2
or S3, ~Sm). Claim 5: One side (100) and the other side (101)
For the interference fringes formed by the light reflected from each surface (100, 101), the norm (Nk) is calculated for the signal (Si) that is sequentially input, and the minimum of the norm (Nk) is calculated. From the value, the period of the interference fringe (P)
In a gap measuring device that calculates the gap between the one surface (100) and the other surface (101) based on the period (P) of the interference fringes, the input signal Si (i=
1 to m) each part of the signal (S2 or S3, ~Sm
) is input, the already input signal Si ′(i
'=1, ~, m-1) and some input signals (S2
or S3, ~Sm), and calculates the norm based on the moving average value. Claim 6: One side (100) and the other side (101)
For the interference fringes formed by the light reflected from each surface (100, 101), the norm (Nk) is calculated for the signal (Si) that is sequentially input, and the minimum of the norm (Nk) is calculated. The period (P) of the interference fringe is calculated by approximating the vicinity of the value with a curve, and the period (P) of the interference fringe is calculated.
In the gap measuring device that measures the gap between the one surface (100) and the other surface (101) based on ) for calculating fast Fourier transform means (6
2), the norm (Nk) is calculated only for a specific range based on the approximate period (P') of the interference fringes, and the gap between the one surface (100) and the other surface (101) is calculated. A gap measuring device characterized by measuring. Claim 7: One side (100) and the other side (101)
For the interference fringes formed by the light reflected from each surface (100, 101), the norm (Nk) is calculated for the signal (Si) that is sequentially input, and the minimum of the norm (Nk) is calculated. The period (P) of the interference fringe is calculated by approximating the vicinity of the value with a curve, and the period (P) of the interference fringe is calculated.
In the gap measuring device between the one surface (100) and the other surface (101), when the one surface (100) or the other surface (101) is newly loaded, the norm ( Nk ) over a wide range to find the norm (Nk
The period (P) of the interference fringes is calculated by approximating the vicinity of the minimum value of k ) with a curve, and if the one surface (100) or the other surface (101) is not newly loaded, Sensor data (Z) of a position sensor that detects the position of the stage that adjusts the distance between the surface (100) and another surface (101)
The period (P') of the interference fringe is estimated from the relationship between the gap (g) and the already measured gap (g), and the period (P') of the estimated interference fringe is calculated.
) based on the norm (N
A gap measuring device characterized in that it calculates k).