JPH03229409A - Pattern detector - Google Patents

Abstract

PURPOSE:To detect patterns having different cross-section structures simultaneously by a method wherein detection signals having directions which are not parallel with each other are subjected to different signal processing operations to detect the patterns. CONSTITUTION:At the time of x-direction scanning, reflected electron signals created by P-N junctions 1-4 are transmitted through a route composed of a first signal processor 7, a first amplifier 11 and a third signal processor 9 by 1st-4th signal route changeover mechanism 14-17. At the time of y-direction scanning, the reflected electron signals are transmitted through a route composed of a second signal processor 8, a second amplifier 12 and a fourth signal processor 10 by the 1st-4th signal route changeover mechanism 14-17. With this process, both the reflected electron signals from marks having different structures are converted into signals which enable calculation of mark positions by a mark position calculator 13. With this constitution, the marks having different structures can be employed as positioning marks in x- and y-directions.

Description

[Detailed description of the invention] Industrial applications The present invention relates to a pattern detection device.

2. Description of the Related Art As an example of a conventional technique, a mark detection device used in an electron beam exposure apparatus is shown in FIG.

FIG. 4 shows a schematic diagram of a conventionally used mark detection method. In FIG. 4, numerals 1 to 4 are pn junctions provided in the exposure chamber and converting reflected electrons into electrical signals. 1, 2 and 3.4 form a pair, 1 and 2 generate signals for scanning in the X direction, and 3 and 4 generate signals for scanning in the X direction.

18 is an alignment mark provided on the substrate to be exposed.

FIG. 5 is a schematic diagram of a conventional pattern detection device. Fifth
In the figure, 19 is a signal processor.

22 is a signal path switching mechanism that switches the x and y signals. An amplifier 21 amplifies the reflected electron signal. 20 is a signal processor. 13 is a mark position calculator.

Next, the functions of each part in FIG. 5 will be explained using FIG. 6.

FIG. 6 shows an example of a convex mark formed by etching on a substrate to be exposed. In FIG. 6, 23 is a substrate to be exposed, 24 is a convex mark, and e is a scanning electron beam. Here, the left side of the figure is assumed to be the -X direction. When the electron beam e is scanned from the -X direction to the +X direction, the reflected electron signals captured at each pn junction are the 7th
Figure (a).

(b). When using the mark shown in Figure 6,
In the signal processor 19 of FIG. 5, a difference signal of pn junction 1 (+xX direction) and pn junction 2 (−X direction) is generated and applied to the signal processor 20 via the amplifier 21. ).In this case, the signal processor 20 generates a pulse signal having an interval of the mark width scanning time shown in FIG. 7(d) from the input signal of FIG. 7(C). Mark position data is obtained by inputting this signal to the mark position calculator 13. When determining the mark position in y, the signal path switching mechanism 22
is switched to yfljl, and the signals from pn junctions 3 and 4 in FIG. 4 are processed in the same way.

FIG. 8 shows an example of dissimilar metal marks provided on a substrate.

Reference numeral 25 denotes a different metal mark, which is usually formed of a heavier element than the substrate 23 to be exposed. Here, the left side of the figure is assumed to be the -X direction. When the electron beam e is scanned from the X direction to the +X direction, the reflected electron signals captured at each pn junction are shown in FIGS. 9(a) and 9(b), respectively. When using the mark shown in FIG. 8, the signal processor 19 shown in FIG.
, a sum signal of pn junction 1 (+xX direction) and pn junction 2 (-x direction) is generated and applied to the signal processor 20 via the amplifier 21. The sum signal is shown in FIG. 9(C). Signal processing In this case, the container 20 is shown in FIG.
), a pulse signal having an interval of the mark width scanning time shown in FIG. 9(d) is generated. By inputting this signal to the mark position calculator 13, mark position data is obtained. In this way, the signal processors 19 and 20 must use different signal processing functions depending on the cross-sectional structure of the mark used.

Problems to be Solved by the Invention In FIG. 5, the gains of the signal processor 19 and the amplifier 21 are set to be the same in both the x and X directions.

However, as mentioned earlier, the backscattered electron signal varies depending on the structure of the mark, and different signal processing methods must be used to obtain mark position data depending on the structure of the mark. Furthermore, since the intensity of the backscattered electron signal increases as the material is a heavier element, the gain of the amplifier 21 must be changed in FIGS. 6 and 8. Therefore, with the configuration of the conventional mark detection device, it is impossible to detect the alignment marks unless the alignment marks in the x and X directions have the same structure. However, in actual overlay exposure, it may be necessary to use alignment marks with different structures as alignment marks in the x and X directions. Figures 3 (a) and (b) show the -
Give an example.

FIGS. 10 and 11 are examples of the case where gate patterns of a gallium arsenide MESFET are overlapped and exposed. In FIGS. 10 and 11, reference numeral 5 indicates an alignment mark formed by an ohmic metal pattern layer to determine the position in the X direction, and 6 indicates an alignment mark formed by a mesa step pattern layer to determine the position in the X direction. , 26 is an ohmic metal pattern, 27 is a mesa step pattern, 28
is the exposure pattern, 29 is the coordinate origin, and f is the scanning direction of the electron beam. In a gallium arsenide MESFET, typically
After forming mesa step patterns 6 and 27 by a contact photolithography method, this pattern 26.2
7 is superimposed using a contact photolithography method to form an ohmic metal pattern 5.26. The overlay error in the contact photolithography method is usually about +-0.5 μm. For this reason, if the alignment mark for the exposure pattern 28 is formed only with the mesa step pattern layer 5 or the ohmic metal pattern layer 6, the exposure pattern will be aligned only with the layer on which the alignment mark is formed. For the other layer +-0,5 μm
This results in an overlay error of more than 100%.

By the way, in a gallium arsenide MESFET, the pattern width of the exposure pattern 28 is made thicker as shown in FIG. 11 in order to prevent pattern breakage at the stepped portion of mesa etching. If this portion overlaps within the mesa step pattern 27, the device characteristics will deteriorate. Also, exposure pattern 2
8 and the ohmic metal pattern 26 is about 2 .mu.m, so the alignment error of the exposure pattern 28 with respect to the ohmic metal pattern 26 must be less than +-0.2 .mu.m. In this way, gallium arsenide MESF
When exposing the ET gate pattern overlappingly, y
The direction must be aligned with the mesa step pattern 27, and the X direction must be aligned with the ohmic metal pattern 26. Therefore, as shown in FIG. 10, the alignment mark 5 in the X direction is formed by an ohmic metal pattern layer, and the alignment mark 6 in the y direction is formed by a mesa step pattern layer, as shown in f in FIG. The position of the coordinate origin 29 is detected by scanning with an electron beam, and an exposure pattern 28 as shown in FIG. 11 is exposed at a predetermined position with respect to the coordinate origin 29.

However, when the ohmic metal pattern layer 5 is scanned with an electron beam, the reflected electron signals are as shown in FIG. 9(a).
(b), and the reflected electron signal when scanning the mesa step pattern layer 6 with an electron beam is shown in FIG. 7(a).
(b). For this reason, it is not possible to obtain the pulse signals shown in FIG. 7(d) and FIG. 9(d) using the same signal processing device. Also, with regard to signal strength, in general, the reflected electron signal obtained from the mark having the structure shown in FIG. 6 is smaller than the signal obtained from the mark having the structure shown in FIG. For this reason, as shown in FIG.
As shown in FIG. 1, it is impossible to form alignment marks in the X+V directions using marks with different cross-sectional structures and to simultaneously detect the respective positions.

The conventional method has the problems described above.

Means for Solving the Problems In order to solve the problems described above, in the present invention, patterns are detected by performing different signal processing on pattern detection signals in two non-parallel directions.

By performing different signal processing on pattern detection signals in two directions that are not parallel to each other, patterns with different cross-sectional structures can be detected simultaneously. This results in
X in the case of exposure equipment.

As the alignment marks in the y direction, alignment marks having different cross-sectional structures can be used.

Embodiment FIG. 1 shows an embodiment of the present invention. FIG. 1 shows a schematic diagram of mark detection. In FIG. 1, numerals 1 to 4 are pn junctions, which convert the strength of reflected electrons into electrical signals.

5 is an alignment mark in the X direction, which is formed of an ohmic metal pattern. Reference numeral 6 denotes a positioning mark in the y direction, which is formed in a convex mesa step pattern. e is an electron beam.

FIG. 2 shows an outline of the signal processing device. 7,8°9.10
are respectively the first. Second. Third. This is a fourth signal processor. 11 and 12 are the first and second amplifiers, respectively.

13 is a mark position calculator, and a third signal processor 9.
The mark position is calculated from the pulse signal output from the fourth signal processor 10. 14.15 and 16.17 are respectively the 1st. Second. Third. This is a fourth signal path switching mechanism.

FIGS. 3(a) to 3(d) schematically show the input/output characteristics of the first to fourth signal processors 7 to 10, respectively. Figure 3 (a
), the first signal processor 7 outputs a sum signal of the input signals. As shown in FIG. 3(b), the second signal processor 8 outputs a difference signal between the input signals. Figure 3 (
As shown in C), the third signal processor 9 outputs a pulse signal having an interval of the mark width scanning time from the sum signal of the output of the first signal processor 7. As shown in FIG. 3(d), the fourth signal processor 10 outputs a pulse signal having an interval of mark width scanning time from the difference signal of the output of the second signal processor 8. 14.15 is pn junction 1
4 reflected electron signals from the 1st. Second signal processor 7, 8
This is a mechanism that switches to 16 is the first and second amplifier 1
1.12 signal from 3rd. Fourth signal processor 9.10
This is a mechanism that switches to

17 is a mechanism for switching the input of the mark position calculator. When detecting the mark shown in FIG. 1, all switching mechanisms A14 to D17 are set to
It is set to switch to lower pick-up during direction scanning.

In the configuration shown in FIG. 1, the backscattered electron signals generated at the pn junctions 1 to 4 are transferred to the first signal processor 7. It passes through the path from the first amplifier 11 to the third signal processor 9. During X-direction scanning, the first to fourth signal path switching mechanisms 14 to 17 switch the second signal processor 8.
Second amplifier 12. It passes through the path of the fourth signal processor 10.

As a result, as shown in FIGS. 3(a) to (d),
The reflected electron signals from marks with different structures are both converted by the mark position calculator 13 into signals from which the mark position can be calculated. Thereby, marks with different structures can be used as alignment marks in the x and y directions.

In addition, if the configurations of the alignment marks in the x and y directions are different, the first to fourth signal path switching mechanisms 14 to 17
．． and 1st. Mark detection becomes possible by setting the gain of the second amplifier so that appropriate signal processing can be performed.

Effects of the Invention Since the present invention performs pattern detection by applying different signal processing to detection signals in two non-parallel directions, it is possible to simultaneously detect patterns having different cross-sectional structures.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram of a mark detection device according to an embodiment of the present invention, FIG. 2 is a schematic diagram of a signal processing device according to an embodiment of the present invention, and FIG. 3 is a diagram for explaining the operation of the above embodiment. A waveform diagram, FIG. 4 is a conventional map diagram, and FIGS. 10 and 11 are schematic diagrams of an example of overlapping exposure for explaining the conventional problems. 1... pn junction (+X direction), 2...
...pn junction (-X direction), 3...
...pn junction (+y direction), 4...
・pn junction (-y direction), 5... alignment mark (ohmic metal), 6... alignment mark (convex mesa step), e... electron beam , 7...signal processor, 8...signal processor, 9...signal processor, 10...
・Signal processor, 11...Amplifier, 12...
...Amplifier, 13...Mark position calculator, 14
......Signal path switching mechanism, 15-...
Signal route switching mechanism, 16... Signal route switching mechanism, 17... Signal route switching mechanism.

Claims (1)

[Claims] 1. A pattern detection device comprising means for performing pattern detection by applying different signal processing to detection signals in two non-parallel directions when detecting a pattern on a substrate.