JPH10214765A - Alignment method and apparatus - Google Patents

Abstract

(57) [Problem] To perform high-accuracy alignment by reducing the influence of disturbance of a charged particle beam detection signal at an edge of an alignment mark on a photosensitive substrate. A pitch P1 of a projected image of an alignment mark 34X on a mask onto a wafer and a wafer mark 39 are provided.
The pitch P of X is slightly different from the pitch P of the alignment mark 34 on the wafer mark 39X by irradiating an electron beam.
An X image is projected to form moire fringes, and reflected electrons from the moire fringes are detected. In this state, the amount of displacement between the two marks is detected from the reflected electron signal obtained when the two marks are relatively scanned.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an alignment method for aligning a mask with a photosensitive substrate in a charged particle beam transfer apparatus used in a lithography process for manufacturing, for example, a semiconductor integrated circuit. The present invention is particularly effective when applied to a case where the amount of displacement is detected using a positioning mark in which a plurality of patterns are arranged in the measurement direction.

[0002]

2. Description of the Related Art In recent years, studies have been made on a charged particle beam transfer apparatus capable of improving both the resolution of a transfer pattern and the throughput (productivity). In such a transfer device, when a pattern to be transferred is a pattern obtained by repeating a predetermined plurality of basic patterns, those basic patterns are formed on an aperture plate, and the transferred patterns are formed on a photosensitive substrate. A transfer device such as a cell projection system for transferring a desired pattern by sequentially transferring images of the basic pattern in a predetermined arrangement has been developed. In recent years, the pattern to be transferred to the photosensitive substrate is divided into a plurality of small areas smaller than the size corresponding to one die on a mask, and the divided patterns are transferred to the photosensitive substrate. A transfer device of the type has also been proposed.

In general, a semiconductor element is formed by stacking multilayer circuit patterns on a substrate in a predetermined positional relationship. Therefore, in such charged particle beam transfer apparatuses, the semiconductor element is formed on a photosensitive substrate in the previous steps. Circuit pattern
An alignment device is provided for performing alignment with a mask pattern to be transferred. As a method for detecting the amount of displacement of the conventional alignment device, (a)
An SEM (Scanning Electron Microscope) system and a (b) beamlet system are known.

In the SEM method (a), a photosensitive substrate (hereinafter, “wafer” is used) by a charged particle beam (hereinafter, “electron beam”) transmitted through a small opening of a square or a circle on a mask. By scanning a wafer mark as an upper alignment mark, a reflected electron signal from the wafer mark is obtained. Focusing on the waveform of the reflected electron signal at the edge of the wafer mark, a threshold (threshold) method for setting a slice level, a correlation method for obtaining an autocorrelation or a cross-correlation with a template, or the like is used. Displacement from the target position (displacement between mask and wafer)
Was detected.

As the wafer mark, as shown in FIG. 6A, for example, a mark 51 formed of a concave portion formed on the wafer is used, and the reflected electron signal I obtained in this case is the scanning position of the electron beam. For X, the level changes from I1 to I2 as shown in FIG. The disturbance of the reflected electron signal I at the edge of the mark 51 between the levels I1 and I2, and the signal level I2 inside the mark 51
Depends on the position of the backscattered electron detector, the expected angle of the irradiation area of the electron beam, and the energy of the electron beam.

FIG. 7A shows a wafer mark.
As shown in the figure, tantalum (Ta) deposited on a wafer
Also, a mark 52 made of heavy metal such as tungsten or tungsten (W) is used. In this case, the number of reflected electrons in a certain direction is
Since the principle of Rutherford scattering is proportional to the product of the square of the atomic number and the atomic number density, the intensity of the reflected electron signal from the heavy metal mark increases. However, even in this case, in the obtained reflected electron signal I, as shown in FIG. 7B, there is a disturbance of the signal of the width ΔM corresponding to the edge portion of the mark 52,
The signal level inside the mark 52 changes depending on the energy of the electron beam and the like. Normally, the position shift amount of the wafer mark is detected using a portion where the change of the reflected electron signal I is sharp (signal at the edge of the mark 51 or 52).

On the other hand, in the beamlet method (b), for example, RCFarrow, et.al. "Mark detection for alignme
nt and registration in a high-throughput projectio
n electron lithography tool ", J.Vac.Sci.Technol., B1
0 (6), Nov / Dec, p2780, as disclosed in (1992),
The duty ratio of the alignment mark on the mask is 1: 1.
And the same pitch as the pitch when it is projected on the wafer,
Again, a line and space pattern with a 1: 1 duty ratio was used as a wafer mark. Then, heavy metal is attached to the surface of the wafer mark, and a part of the reflected electrons from the entire wafer mark by the electron beam passing through the alignment mark on the mask is detected. That is, the wafer mark used in this beamlet method is shown in FIG.
(A) is a multi-mark in which heavy metal marks 52 are arranged at a predetermined pitch.

[0008]

In the prior art as described above, in the SEM method of (a), the position of a wafer mark is determined based on information of a portion where the waveform of a reflected electron signal from a single mark is sharp. The shift amount was detected. However, there is an inconvenience that the inclination is apt to change slightly due to the influence of noise or the like, and it is difficult to obtain good measurement reproducibility in one measurement. In order to improve the reproducibility of measurement, conventionally, measures such as averaging the measurement results by performing a plurality of measurements have been taken, but this reduces the throughput.

Further, in the beamlet method (b), reflected electrons from the entire mark portion of the wafer mark are detected, but the reflected electron at the edge portion of the mark 52 as shown in FIG. The disturbance of the reflected electron signal I similarly exists when a plurality of marks are scanned with an electron beam having the same pitch. Therefore, the waveform of the reflected electron signal is disturbed at the edge thereof, which causes a deterioration in measurement reproducibility.

In view of the above, the present invention reduces the influence of the disturbance of the detection signal of the charged particle beam at the edge of the alignment mark on the substrate, and enables alignment with high accuracy. It is an object of the present invention to provide a method and an alignment apparatus capable of performing the alignment method.

[0011]

According to the alignment method of the present invention, a pattern on a mask (1) is transferred onto a photosensitive substrate (5) via a projection system (14, 15) under a charged particle beam. In this case, in the alignment method for aligning the mask and the photosensitive substrate, a plurality of charged particle beam slit-shaped transmission portions (38A) are arranged on the mask (1) at a predetermined pitch in the measurement direction. A first alignment mark (34X) is formed on the photosensitive substrate (5), and a linear alignment mark is formed in a region on the photosensitive substrate (5) having the same width as the projected image of the first alignment mark in the measurement direction. A second alignment mark (39X) is formed by arranging the charged particle beam reflecting portions (40A) in the measurement direction at a slightly different pitch from the projected image of the first alignment mark. , Predetermined charged particle beam The reflected charged particle beam generated from the second alignment mark while relatively moving the projection image of the first alignment mark and the second alignment mark in the measurement direction under the Alternatively, the secondary charged particle beam is detected, and the position of the mask and the photosensitive substrate are aligned based on the detection result.

According to the present invention, the pitch between the projected image of the first alignment mark (34X) on the mask and the second alignment mark (39X) on the photosensitive substrate is slightly different. ing. A slight difference in pitch means that, as an example, when the width of the second alignment mark (39X) in the measurement direction is MD, the reflection portion (40A) of the second alignment mark (39X) within the width MD ) Differs from the number of images of the transmission portion (38A) in the projected image of the first alignment mark (34X) by one to several. When the numbers of both marks are different by one, the minimum required number of marks can be formed by the least number of marks. In the present invention, when an image of the first alignment mark (34X) is projected on the second alignment mark (39X), so-called moiré fringes are formed, and reflected charged particle beams from the entire moiré fringe are formed. Alternatively, a secondary charged particle beam is detected. Therefore, when the two marks are relatively scanned, almost the same amount of charged particles from the edges of the marks are added to the detection signal regardless of the amount of positional deviation between the two marks. Waveform disturbance is eliminated, and good measurement reproducibility can always be expected. As a result, positioning can be performed with high accuracy.

In this case, the first alignment mark (3
4X) The ratio (duty ratio) of the width in the measurement direction of the slit-shaped transmission portion (38A) to the width of the non-transmission portion (38B) between the transmission portions in the measurement direction, or the second alignment mark. At least one of the ratio between the width in the measurement direction of the linear reflection portion (40A) of (39X) and the width in the measurement direction of the low reflection portion (40B) between the reflection portions is substantially 1: 1. Further, the length of the projection image of the first alignment mark (34X) in the non-measurement direction is the second alignment mark (39X).
Is desirably different from the length in the non-measurement direction.

When the duty ratio of at least one of the alignment marks is approximately 1: 1 as described above, the area where the detection signal does not change in the amount of displacement between the two marks, that is, the dead zone is reduced. Also, for example, as shown in FIG.
The length H2 of the non-measurement method of the second alignment mark (39X) on the photosensitive substrate is equal to the first alignment mark (3X).
In the case where the projection image of 4X) is shorter than the length H1 in the non-measurement direction, the entire mark on the photosensitive substrate has an advantage that it can contribute to the detection signal, but the charged particle beam in the protruding portion becomes a DC component. Conversely, the length H2 of the non-measurement method of the second alignment mark (39X) on the photosensitive substrate is equal to the length H1 of the projection image of the first alignment mark (34X) in the non-measurement direction.
In the wider case, such a DC component disappears,
Unused portions will remain in the marks on the photosensitive substrate. In any case, even if a certain degree of displacement occurs, the two marks can be completely overlapped, and the level fluctuation of the obtained detection signal is reduced.

Further, the alignment apparatus according to the present invention comprises:
A transfer device is provided for transferring a pattern on a mask (1) onto a photosensitive substrate (5) via a projection system (14, 15) under a charged particle beam, and aligns the mask with the photosensitive substrate. An alignment apparatus for performing a first charged particle beam on a mask (1) at a predetermined pitch on a mask (1);
Mark (34X) and projection system (14, 1)
5) irradiating the second alignment mark (39X) formed on the photosensitive substrate (5) at a slightly different pitch from the projected image of the first alignment mark (34X). System (10, 11A to 11C, 12A, 12B, 3
1), a charged particle beam detector (29) for detecting a reflected charged particle beam or a secondary charged particle beam generated from the second alignment mark, and a projected image of the first alignment mark Relative movement device (13A) for relatively moving the second alignment mark and the second alignment mark in the measurement direction, and projecting the first alignment mark and the second alignment via the relative movement device Calculating means (19) for calculating the amount of displacement between the mask (1) and the photosensitive substrate (5) based on a detection signal obtained from the charged particle beam detector (29) when the mark is relatively moved. It is provided.

According to the alignment apparatus of the present invention, the alignment method of the present invention can be performed using the detection signal from the charged particle beam detector (29).

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. In this embodiment, the present invention is applied to the case where alignment is performed by an electron beam transfer device. FIG. 1 shows a schematic configuration of a split transfer type electron beam transfer apparatus used in this embodiment. In FIG. 1, a Z axis is taken in parallel with an optical axis AX of an electron optical system, and in a plane perpendicular to the Z axis. The X axis is perpendicular to the plane of FIG. 1 and the Y axis is parallel to the plane of FIG.
A description will be given taking the axis. First, in the deflection illumination system of the present example, the electron beam EB emitted from the electron gun 10 is focused once by the first condenser lens 11A, and then focused again by the second condenser lens 11B. An aperture plate 31 having a variable aperture is arranged near the second condenser lens 11B, and the electron beam E passing through the aperture of the aperture plate 31 is provided.
B is deflected mainly in the Y-direction by the first field-selection deflector 12A, is converted into a parallel beam by the third condenser lens 11C, and is then turned back by the second field-selection deflector 12B so that one of the masks 1 It is guided to the irradiation area 33 on the small area. The amount of deflection in the field-of-view selecting deflectors 12A and 12B is determined by the main controller 19 which controls the overall operation of the apparatus.
It is set via the deflection amount setting device 25. In FIG. 1, the locus of the electron beam EB indicated by a solid line indicates the conjugate relationship of the crossover image, and the locus indicated by the dotted line indicates the conjugate relationship of the mask pattern image.

The electron beam EB that has passed through the mask 1 is deflected by a predetermined amount by a deflector 13A composed of a two-stage electromagnetic deflector, and once forms a crossover image CO by a projection lens 14, and then the objective lens 15 Is focused on the wafer 5 on which the electron beam resist is applied, and an image obtained by reducing the pattern in one small area of the mask 1 at a predetermined position on the wafer 5 at a predetermined reduction rate β (for example, １ ／) is formed. Transcribed. The projection lens 14 and the objective lens 15 of this example constitute an imaging system of a symmetric magnetic doublet (SMD) system, and an aperture stop 32 is arranged at a position where a crossover image CO is formed. I have. A deflector 13B composed of a two-stage electromagnetic deflector is also arranged between the vicinity of the objective lens 15 and the wafer 5, and deflectors 13A and 13B
Is controlled by the main controller 19 to determine the deflection amount setter 26.
Set via. In the split transfer method, each small area on the mask 1 is arranged with a strut in between, while each small area on the corresponding wafer 5 is arranged closely, so that the deflectors 13A, 13B is used to shift the electron beam laterally by the amount of the strut and to correct a synchronization error between the mask 1 and the wafer 5. Further, in this example, when the image of the alignment mark on the mask 1 is scanned on the wafer mark on the wafer 5 in the measurement direction, the deflector 13A or 13B is used.

A backscattered electron detector 29 for detecting backscattered electrons and the like from the wafer 5 is disposed near the bottom surface of the objective lens 15, and a backscattered electron signal RE from the backscattered electron detector 29 is transmitted to the main controller 19. Supplied. The mask 1 is mounted on the mask stage 16 in parallel with the XY plane. The mask stage 16 is continuously moved in the X direction by the driving device 17 and is stepwise moved in the Y direction. The position of the mask stage 16 in the XY plane is detected by the laser interferometer 18 and output to the main controller 19.

On the other hand, the wafer 5 is held on a wafer stage 21 on a sample stage 20 in parallel with the XY plane. The wafer stage 21 can be continuously moved by the driving device 22 in the direction opposite to the direction of continuous movement of the mask stage 16 along the X axis, and can be step-moved in the Y direction. The direction opposite to the X direction is because the mask pattern image is inverted by the lenses 14 and 15. Wafer stage 21
Are detected by the laser interferometer 23 and output to the main controller 19. Further, the wafer stage 2
1 also incorporates a mechanism for correcting the rotation angle of the wafer 5.

At the time of exposure, the main controller 19 is operated by the input device 24.
Exposure data input from the laser interferometers 18 and 23
Stage 16 and wafer stage 21 detected by
Field deflectors 12A and 12A based on the
B and the amount of deflection of the electron beam EB by the deflectors 13A and 13B are calculated, and information (for example, position and moving speed) required to control the operations of the mask stage 16 and the wafer stage 21 is calculated. The calculation result of the deflection amount is output to deflection amount setting devices 25 and 26, and these setting devices set the deflection amount of the electron beam by the field-of-view selection deflectors 12A and 12B and the deflectors 13A and 13B, respectively.

Calculation results regarding the operations of the mask stage 16 and the wafer stage 21 are output to drivers 27 and 28, respectively. The drivers 27 and 28 operate the driving device 1 so that the stages 16 and 21 operate according to the calculation results.
7 and 22 are controlled. As the input device 24, a device for reading magnetic recording information created by an exposure data creating device, a device for reading exposure data recorded on the mask 1 and the wafer 5 when these are loaded, and the like may be appropriately selected. .

Next, the basic operation of transferring the mask pattern by the division transfer method using the transfer apparatus of the present embodiment will be described with reference to FIG. FIG. 2 is a perspective view showing the correspondence between the mask 1 and the wafer 5 of the present example. In FIG. 2, the mask 1 is rectangular with a predetermined pitch in the X direction and the Y direction by struts 3 as boundary regions. Are divided into a number of small areas 2A, 2B, 2C,..., And the electron beam EB is applied to the irradiation area 33 in the small area to be transferred (small area 2A in FIG. 2).
Is irradiated. In each of the small areas, a transmission part of the electron beam corresponding to the pattern to be transferred is provided, and the strut 3 is an area that scatters the electron beam. The mask 1 of this example has a thickness of 1
It is a so-called scattering (hole) stencil type in which a hole formed in a scattering substrate made of silicon (Si) of about 0 μm is used as a transmission part of an electron beam. However, a so-called scattering mask in which an electron beam transmitting portion is formed of a thin film of silicon nitride (SiN) or the like and a tungsten scattering portion is appropriately provided on the surface thereof may be used as a mask.

The electron beam EB that has passed through the small area 2A on the mask 1 passes through the projection lens 14 and the objective lens 15 shown in FIG. A reduced image of the pattern is formed. At the time of exposure,
The irradiation of the electron beam EB is repeated in units of the small regions 2A, 2B, 2C,..., And the reduced images of the patterns in each small region are sequentially transferred to different small transfer regions 7A, 7B, 7C,. You. At this time, the deflectors 13A and 13B in FIG.
Is driven to shift the electron beam EB laterally by the width of the strut 3 that separates each small area, so that the small transfer areas 7A, 7B, 7C,...

At this time, under the control of the main controller 19, through the mask stage 16 and the wafer stage 21,
As shown in FIG. 2, the mask 1 is moved at a predetermined speed VM in the −X direction.
The wafer 5 is continuously moved at the speed VW corresponding to the + X direction in synchronization with the continuous movement of the wafer 5. As the mask 1 and the wafer 5 move in the X direction, the optical axis A
The operation of sequentially transferring the patterns in the plurality of small areas in a row reaching the position crossing X on the wafer 5 is repeated,
When the patterns in all the small areas on the mask 1 are transferred onto the wafer, the transfer of the pattern to the transfer area 6A for one die on the wafer 5 is completed. Thereafter, the pattern of the mask 1 is similarly transferred to another adjacent transfer area 6B for one die on the wafer 5.

In order to transfer such a mask pattern to the second and subsequent layers on the wafer 5, alignment (alignment) between the projected image of the pattern of the mask 1 and each transfer area on the wafer 5 must be performed in advance. It must be done with high precision.
Therefore, a plurality of alignment marks are formed on the mask 1. FIG. 3 shows an example of an arrangement of alignment marks on the mask 1. In FIG. 3, small areas 2A, 2B,... Are formed inside the mask 1 at a predetermined pitch in the X direction and the Y direction. 4 around small area of
The same X-axis alignment marks 34X to 37X at the corners,
Also, the same Y-axis alignment marks 34Y to 37Y are formed. The X-axis alignment mark 34X is
As shown in FIG. 4A, a rectangular opening pattern 38A having a width L1 in the X direction and a width H1 in the Y direction is pitched P1 in the X direction.
Are marks arranged in a predetermined number. Specifically, the shape of the alignment mark 34X of the present example is
At the stage of the projected image projected on the wafer 5 via the objective lens 15, the width L1 in the X direction of the opening pattern 38A is 4.5 μm, the width H1 in the Y direction is 55 μm, and the pitch P1 is 10 μm. The number of 38A is 10. Therefore, the width of the alignment mark 34X in the measurement direction (X direction) is about 94.5 μm at the stage of the projected image, and the ratio (duty ratio) of the width in the measurement direction between the opening pattern 38A and the scattering portion 38B therebetween. Is 4.5: 5.
5, or approximately 1: 1. At this time, the irradiation area of the electron beam irradiated onto the alignment mark 34X from the deflection illumination system of FIG. 1 is an area having a width in the X direction of 110 μm and a width in the Y direction of about 70 μm (at the stage of the projected image). In this case, the entire alignment mark 34X completely fits within the irradiation area.

The Y-axis alignment mark 3 shown in FIG.
4Y is a mark obtained by rotating the X-axis alignment mark 34X by 90 °. Each transfer area on the wafer 5 shown in FIG.
4Y to 37Y are the projection lens 14 and the objective lens 1 of FIG.
A plurality of wafer marks are formed in a positional relationship projected through the reference numeral 5. X-axis alignment mark 34
The X-axis wafer mark 39X corresponding to X is shown in FIG.
Shown in

In FIG. 4B, the X-axis wafer mark 39X is formed by forming a rectangular convex pattern 40A of heavy metal (for example, tantalum or tungsten) having a width L2 in the X direction and a width H2 in the Y direction at a pitch in the X direction. P is a mark arranged in a predetermined number. Specifically, the shape of the wafer mark 39X of this example is such that the width L2 of the rectangular convex pattern 40A in the X direction is 4.
5 μm, Y direction width H2 is 50 μm, pitch P is 9 μm
And the number of patterns 40A is eleven. Accordingly, the width of the wafer mark 39X in the measurement direction (X direction) is about 94.5 μm, and the ratio (duty ratio) of the width in the measurement direction between the convex pattern 40A and the low reflection portion 40B therebetween is 4. 5: 4.5, ie exactly 1: 1. That is, the projected image of the alignment mark 34X and the wafer mark 39X are line and space patterns having slightly different pitches, and a moire fringe is formed by overlapping the two. The width H2 of the wafer mark 39X in the non-measurement direction is shorter than the width H1 of the projection image of the alignment mark 34X in the non-measurement direction. Further, the Y-axis wafer mark is a mark obtained by rotating the X-axis wafer mark 39X of FIG. 4B by 90 °.

When detecting the amount of displacement in the X direction between the alignment mark 34X and the wafer mark 39X during alignment, the mask stage 1 shown in FIG.
6, the alignment mark 34X is moved to the irradiation area of the electron beam EB by the deflecting illumination system, and then the wafer stage 21 is driven so that the wafer mark 39X on the wafer 5 is designed to be a projected image of the alignment mark 34X. To the target position (target position). In this example, as shown in FIG.
Width H of projection image of alignment mark 34X in non-measurement direction
1 is wider than the width H2 of the wafer mark 39X.
, The wafer mark 39X is the alignment mark 34
Moire fringes are formed within the width of the projected image of X, and reflected electrons are generated from bright portions of the moire fringes (portions where the electron beam irradiation area and the convex portion (reflection portion) of the wafer mark 39X overlap). Then, the reflected electrons are detected by the reflected electron detector 29. Therefore, the main controller 19 uses the deflector 13A (or the deflector 13B) of FIG.
The projected image of the alignment mark 34X is scanned in the measurement direction (X direction) on 9X, and the reflected electron signal RE from the reflected electron detector 29 is monitored. The mask stage 1
6, or the wafer stage 21 may be driven to perform relative scanning between them.

Specifically, in FIG. 4, the difference between the pitch P1 of the projected image of the alignment mark 34X on the mask and the pitch P of the wafer mark 39X is ΔP (= | P1-P
|), The number of the patterns 40A of the wafer mark 39X is (2N + 1), and the amount of displacement of the center of the projected image of the alignment mark 34X with respect to the center of the wafer mark 39X in the X direction is ΔX, and a predetermined proportional coefficient k is used. The reflected electron signal RE is as follows.

RE = k ｛N ² · P / (2N−1) −2 | ΔX |｝ (1) However, the range of the displacement amount ΔX in which the expression (1) is satisfied is as follows. −ΔP / 2 ≦ ΔX ≦ ΔP / 2 (2) When the actual parameters in FIG. 4 are substituted into the equation (1), the pitch P is 9 μm, and the integer N that defines the number is 5.
The reflected electron signal RE is expressed as (25-2 | ΔX |) as follows.
Becomes a signal proportional to. That is, within the range of the expression (2), the reflected electron signal RE is a signal that decreases linearly with respect to the absolute value of the positional deviation amount ΔX.

RE = k (25−2 | ΔX |) (3) FIG. 5 shows that not only the backscattered electron signal RE represented by the equation (3) but also that | ΔX | is larger than the range of the equation (2). The reflected electron signal RE in the case is also shown. The horizontal axis in FIG. 5 is the displacement amount ΔX (range of ≧ 0), and the vertical axis is the displacement amount ΔX.
This is the level of the reflected electron signal RE corresponding to X. However,
In FIG. 5, a signal level in a range where the displacement amount ΔX exceeds ΔP / 2 indicates a signal level calculated using specific parameters in FIG.

Also in this example, as described with reference to FIG. 7 for the conventional example, at the edge of each of the convex marks of the wafer mark 39X, the reflected electron signal is generated within the range of the width ΔM in the measurement direction. Disturbance occurs. In this example, the width ΔM
The difference ΔP between the mark pitches appearing in the expression (2) is set to be larger than that. As a result, the disturbance of the signal at the edge of each of the convex marks of the wafer mark 39X is always included in the reflected electron signal RE, and as shown in FIG.
Within the range of the expression (2), the reflected electron signal RE has no waveform disturbance. Therefore, as an example, in the main control device 19,
The slice level RE1 slightly lower than the highest level of the reflected electron signal RE is set, and the reflected electron signal R
It detects two positional deviation amounts ΔX1 and ΔX2 when E crosses the slice level RE1. In practice, the positional deviation amount .DELTA.X1, .DELTA.X2 is the amount of deflection of the projected images of the alignment marks 34X from the target position of the design, their positional deviation amount .DELTA.X1, average target of [Delta] X _AV wafer mark 39X of .DELTA.X2 This can be regarded as a positional deviation amount from the position.

In the main controller 19, the reflected electron signal RE shown in FIG. 5 has a straight line in a range of ΔX ≧ 0 (a straight line having a negative gradient) and a straight line in a range of ΔX <0 (a straight line having a positive gradient). )
May be obtained by least-squares approximation or the like, and the amount of displacement at the intersection of these straight lines may be regarded as the amount of displacement of the wafer mark 39X. At this time, if a large amount of data is used, it goes without saying that the SN ratio is improved and the measurement reproducibility is improved.

If there is a possibility that the amount of displacement ΔX that can be driven in exceeds the range of the expression (2), the waveform of the reflected electron signal RE of FIG. 5 is stored in advance as a template using a standard wafer or the like. In advance, when actually detecting the amount of positional deviation, the correlation between the actually obtained reflected electron signal RE and the template is calculated, and the amount of positional deviation when the correlation becomes the highest is calculated as the positional deviation of the wafer mark 39X. It may be an amount.

Thereafter, similarly for the other wafer marks, the corresponding alignment marks 35X to 37X, 34
By relatively scanning the projected images of Y to 37Y, the amount of displacement in the X direction or the Y direction can be detected. Then, the main controller 19 calculates the amount of displacement of the transfer area of the exposure target on the wafer 5 in the X direction, the Y direction, and the rotation direction with respect to the pattern image of the mask 1 from these displacement amounts. The position of the wafer 5 is corrected by driving the wafer stage 21 shown in FIG. 1 so as to correct the amount of displacement. Thereafter, the mask pattern is transferred onto the wafer 5 by the division transfer method while maintaining the positional relationship set as described above, whereby high overlay accuracy is obtained.

As described above, according to this embodiment, the projected image of the alignment mark and the wafer mark are superimposed to form a moire fringe, and the amount of displacement between the two marks is determined based on the reflected electron signal from the moire fringe. Has been detected. Therefore,
The backscattered electron signal at the edge portion of each pattern constituting the wafer mark is generated in the same degree for any amount of displacement, and the reflected electrons change almost linearly with respect to the amount of displacement at least in a range satisfying the expression (2). A signal is obtained. Therefore,
The position shift amount can be detected with high accuracy, and as a result, position alignment can be performed with high accuracy.

In the above-described embodiment, a mark in which heavy metal marks are arranged at a predetermined pitch is used as a wafer mark, but a mark in which an uneven pattern is arranged at a predetermined pitch on a substrate is used as a wafer mark. May be. Further, in order to detect the amount of displacement between the alignment mark and the wafer mark, reflected electrons from the wafer mark are detected. However, secondary electrons from the wafer mark may be detected. Further, the above-described embodiment
Although the present invention is applied to a transfer device of a split transfer system, the present invention can be similarly applied to a case where alignment is performed by a transfer device of a cell projection system or the like.

Further, it goes without saying that the present invention can be applied not only to a transfer device using an electron beam but also to a case where alignment is performed by a transfer device (exposure device) using another charged particle beam such as an ion beam. . As described above, the present invention is not limited to the above-described embodiment, and can take various configurations without departing from the gist of the present invention.

[0040]

According to the alignment method of the present invention, the pitch of the projected image of the first alignment mark on the mask is slightly different from the pitch of the second alignment mark on the photosensitive substrate. However, since the charged particle beam from the moiré fringes formed when the two are superimposed is detected, the same amount of charged particle beam from the mark edge always remains, regardless of the amount of misalignment between the mask and the photosensitive substrate. In addition to the detection signal, the detection signal itself has the advantage of eliminating waveform disturbance. Further, since charged particle beams from other than the edge portion are also detected at the same time, the number of charged particle beams to be detected increases, and the SN of the detection signal increases.
The ratio is also improved, and good measurement reproducibility can be expected. Therefore,
Positioning can be performed with high accuracy.

When at least one of the duty ratio of the first alignment mark and the duty ratio of the second alignment mark is substantially 1: 1, the dead zone is reduced. Further, when the length of the projection image of the first alignment mark in the non-measurement direction is different from the length of the second alignment mark in the non-measurement direction, the two marks are only roughly positioned. Since these marks can be completely overlapped in the non-measurement direction, there is an advantage that the level of the obtained detection signal is always stable. Further, according to the alignment apparatus of the present invention, the alignment method of the present invention can be performed.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram showing an example of an electron beam transfer apparatus according to an embodiment of the present invention.

FIG. 2 is a perspective view for explaining an operation when transferring is performed by a division transfer method.

FIG. 3 is a plan view showing an example of the arrangement of alignment marks on a mask according to the embodiment.

FIG. 4 is an enlarged plan view showing a configuration of an alignment mark and a wafer mark on a mask.

FIG. 5 is a diagram showing an example of a backscattered electron signal detected in the embodiment.

FIG. 6 is a diagram showing a conventional wafer mark formed of a convex mark and a reflected electron signal obtained from the mark.

FIG. 7 shows a wafer mark made of a conventional heavy metal mark,
FIG. 4 is a diagram showing reflected electron signals obtained from the marks.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Mask 5 Wafer 13A, 13B Deflector 14 Projection lens 15 Objective lens 16 Mask stage 19 Main controller 21 Wafer stage 25, 26 Deflection amount setting unit 29 Backscattered electron detector 34X-37X, 34Y-37Y Alignment mark 39X on mask Alignment mark on wafer

Claims (3)

[Claims] An alignment method for aligning the mask and the photosensitive substrate when transferring a pattern on the mask onto a photosensitive substrate via a projection system under a charged particle beam; A first alignment mark formed by arranging a plurality of charged particle beam slit-shaped transmission portions at a predetermined pitch in a measurement direction is formed on the first alignment mark, and the first alignment mark of the first alignment mark on the photosensitive substrate is formed. In a region having the same width as the projected image in the measurement direction, the reflecting portions of the linear charged particle beams are arranged at a slightly different pitch from the projected image of the first alignment mark in the measurement direction. A second alignment mark is formed, and the projected image of the first alignment mark and the second alignment mark are relatively moved in a measurement direction under a predetermined charged particle beam. While the second An alignment method comprising: detecting a reflected charged particle beam or a secondary charged particle beam generated from an alignment mark; and performing positioning between the mask and the photosensitive substrate based on the detection result. 2. The alignment method according to claim 1, wherein a width of the first alignment mark in a measurement direction of the slit-shaped transmission portion and a measurement direction of a non-transmission portion between the transmission portions are measured. At least one of a ratio of a width or a ratio of a width of the second alignment mark in the measurement direction of the linear reflection portion to a width of the low reflection portion between the reflection portions in the measurement direction is substantially at least one. Wherein the length of the projection image of the first alignment mark in the non-measurement direction is different from the length of the second alignment mark in the non-measurement direction. Method. 3. An alignment apparatus provided in a transfer apparatus for transferring a pattern on a mask under a charged particle beam onto a photosensitive substrate via a projection system, for aligning the mask with the photosensitive substrate. In the above, a predetermined charged particle beam through the first alignment mark and the projection system at a predetermined pitch on the mask, at a pitch slightly different from the projected image of the first alignment mark An irradiation system for irradiating a second alignment mark formed on the photosensitive substrate, and a charged particle beam detection for detecting a reflected charged particle beam or a secondary charged particle beam generated from the second alignment mark A relative movement device for relatively moving the projected image of the first alignment mark and the second alignment mark in the measurement direction; and the first alignment mark via the relative movement device. Calculating means for calculating a displacement amount between the mask and the photosensitive substrate based on a detection signal obtained from the charged particle beam detector when the projected image of the mask and the second alignment mark are relatively moved. And an alignment device.