JP4184983B2 - Alignment method - Google Patents

Description

  The present invention relates to an alignment technique, and more particularly to an alignment technique suitable for application to step-and-repeat proximity exposure.

  As alignment methods employed in step-and-repeat exposure, a die-by-die alignment method and a global alignment method are known. The die-by-die alignment method is a method in which an alignment mark is detected for each step and the wafer and the mask (reticle) are aligned. The global alignment method is a method in which an estimation calculation is performed based on position information obtained from several alignment marks on a wafer, and an area to be exposed on the wafer is moved to an exposure position based on the calculation result.

  The global alignment method described in Patent Document 1 will be described with reference to FIG. FIG. 7 shows a plan view of the wafer. A plurality of unit regions (dies) 101 are defined on the exposed surface of the wafer 100. The unit areas 101 are arranged in a matrix. One unit area 101 is moved to the exposure position, and one unit area 101 arranged at the exposure position is exposed. On the peripheral edge of the wafer 100, the unit region 101 is not disposed, and a plurality of alignment marks 102 are disposed.

  By detecting the alignment mark 102, position information of the wafer 100 is obtained. From this position information, global alignment is performed.

JP-A-11-168053

  Usually, the shape of the unit region is a square or a rectangle, and a plurality of unit regions are arranged in a matrix in the exposed surface of the wafer. Since the wafer is circular, a part of the unit area on the edge of the wafer is missing. A unit area that is partially missing is referred to as an incomplete unit area. When a plurality of chips are included in one unit region, usable chips may be formed in the incomplete unit region. By transferring the pattern also to the incomplete unit region, usable chips can be formed, and the number of chips taken out from one wafer can be increased.

  When using global alignment, it is possible to perform alignment of incomplete unit regions. However, since position detection for each unit region is not performed, a highly accurate and highly stable moving mechanism is required. On the other hand, in die-by-die alignment, the position detection is performed for each unit region to correct the positional deviation, so that the moving mechanism is not required to have high accuracy as required for global alignment. In incomplete unit regions, alignment marks necessary for die-by-die alignment may be lost. In this case, die-by-die alignment cannot be performed.

  The complete unit region must be aligned by die-by-die alignment, and the incomplete unit region must be aligned by global alignment. For this reason, the alignment accuracy of the incomplete unit region is lowered as compared with the complete unit region.

An object of the present invention is to provide a aligned Way Method capable of suppressing the lowering of the alignment accuracy of an incomplete unit area.

According to one aspect of the invention,
(A) A substrate to be processed in which a plurality of complete unit regions whose entire area is included in the processing surface of the substrate to be processed and at least one incomplete unit region that is partially missing from the outer periphery of the substrate to be processed are defined The process of preparing
(B) Based on the result of detecting the position of the substrate to be processed after the complete unit region is arranged by die-by-die alignment by die-by-die alignment for at least a part of the complete unit region. Obtaining unit unit arrangement information;
(C) a step of aligning the substrate to be processed so that the incomplete unit region is arranged at the processing position based on the arrangement information acquired in the step b;
(D) obtaining a positional deviation information of the substrate to be processed by detecting a part of alignment marks formed in the incomplete unit region arranged at the processing position in the step c;
(E) a step of correcting the position based on the positional deviation information obtained in the step d;
Have
The positions of the complete unit area and the incomplete unit area in the substrate to be processed are defined by design coordinates in the substrate to be processed determined at the time of design,
The arrangement information acquired in the step b includes a first correspondence between design coordinates of a unit area and position information of a substrate to be processed when the unit area is arranged at a processing position.
In the step c, the position information of the substrate to be processed is obtained by applying the first correspondence to the design coordinates of the incomplete unit region, and the position of the substrate to be processed is aligned based on the obtained position information.
Step b is
A step of acquiring the amount of expansion / contraction of the complete unit area arranged at the processing position by detecting the alignment mark;
Obtaining a second correspondence between the design coordinates of the unit area and the amount of expansion and contraction;
Including
Step c is
Applying the second correspondence to the design coordinates of the incomplete unit region to estimate the amount of expansion / contraction of the incomplete unit region;
A process of aligning the substrate to be processed in consideration of the estimated expansion and contraction amount;
Alignment method comprising is provided.

  By observing a part of the alignment mark in the incomplete unit region and correcting the position based on the observation result, the position accuracy can be improved as compared with the case where the alignment mark is not observed.

  Taking the electron beam proximity exposure as an example, the alignment method according to the embodiment will be described.

  FIG. 1A is a plan view of a semiconductor wafer 11 that is an object of the alignment method according to the embodiment. A plurality of unit regions 50 arranged in a matrix are defined on the exposed surface of the wafer 11. An XYZ orthogonal coordinate system is defined in which the row direction is the X axis, the column direction is the Y axis, and the normal direction of the exposed surface is the Z axis. The shape of the unit region 50a included in the exposed surface of the wafer 11 is, for example, a square or a rectangle. Such a unit region 50a is referred to as a “complete unit region”. The unit region 50b that extends around the outer peripheral line of the wafer 11 has a shape obtained by cutting out a part of a square or a rectangle. Such a unit region 50b is referred to as an “incomplete unit region”. The unit area 50 is a unit in which exposure is performed, and exposure is performed by moving one unit area 50 to the exposure position of the exposure apparatus. The unit area 50 is usually called a “die”.

  At the time of designing, the position of each unit region 50 in the wafer 11, that is, the coordinates in the XY plane is determined. The coordinates determined at the time of design are called “design coordinates” of the unit area 50. The design coordinates are coordinates in the XY coordinate system defined on the surface of the wafer 11.

  In each unit region 50, a plurality of chip regions 51 are defined. One integrated circuit element is formed in one chip region 51. FIG. 1A shows a case where four chip regions 51 are defined in one complete unit region 50a. The incomplete unit region 50b includes 0 to 3 chip regions 51 depending on the size and shape thereof.

  FIG. 1B shows a plan view of a state in which one unit region 50a and a transfer region 12a of a mask arranged close to the wafer are overlaid, and a schematic view of an observation optical system for alignment. Two X alignment marks WX1, WX2 and two Y alignment marks WY1, WY2 are arranged in the complete unit region 50a. The X alignment marks WX1 and WX2 are usually arranged at different positions with respect to the X-axis direction and the Y-axis direction. The Y alignment marks WY1 and WY2 are also usually arranged at different positions with respect to the X-axis direction and the Y-axis direction.

  Alignment marks MX1, MX2, MY1, and MY2 corresponding to the alignment marks WX1, WX2, WY1, and WY2 on the wafer are formed in the mask transfer region 12a. The positions of these alignment marks are the X-axis direction, the Y-axis direction, the rotation direction in the XY plane, and the mask relative to the mask, which are calculated from the positional deviation between the alignment mark on the wafer and the alignment mark on the mask. It is preferable to arrange the wafer so that the error of the relative expansion / contraction amount of the wafer becomes small.

  The observation optical system 20X1 observes the alignment marks WX1 and MX1. Similarly, the observation optical system 20X2 observes the alignment marks WX2 and MX2, the observation optical system 20Y1 observes the alignment marks WY1 and MY1, and the observation optical system 20Y2 observes the alignment marks WY2 and MY2. The optical axes of the observation optical systems 20X1 and 20X2 are inclined in the Y-axis direction from the normal direction of the exposed surface. The optical axes of the observation optical systems 20Y1 and 20Y2 are inclined in the X-axis direction from the normal direction of the exposed surface.

  FIG. 2 shows a schematic view of an alignment apparatus according to the embodiment. The alignment apparatus according to the embodiment includes a wafer / mask holding unit 10, an observation optical system 20, and a control device 30.

  The wafer / mask holding unit 10 includes a wafer holding table 15, a mask holding table 16, and moving mechanisms 17 and 18. At the time of alignment, the wafer 11 is held on the upper surface of the wafer holding table 15 and the mask 12 is held on the lower surface of the mask holding table 16. The wafer 11 and the mask 12 are arranged substantially in parallel so that a certain gap (proximity gap) is formed between the exposed surface of the wafer 11 and the surface of the mask 12 on the wafer side. When performing electron beam proximity exposure, a stencil type mask 12 is used. A transfer pattern and an alignment mark are formed by an opening provided in the mask membrane.

The moving mechanism 17 is fixed to the reference base 1 and can move the wafer holder 15 so that the relative position of the wafer 11 and the mask 12 in the exposed surface changes. The moving mechanism 18 can move the wafer holder 15 so that the distance between the exposed surface of the wafer 11 and the mask surface of the mask 12 changes. When taking the X axis from the left to the right in the figure, the Y axis from the front surface to the back surface in the direction perpendicular to the paper surface, and the Z axis in the normal direction of the exposed surface, the moving mechanism 17 X-axis, Y-axis direction, to adjust the relative position around the rotational direction (theta _Z direction) of the Z-axis, the moving mechanism 18, the Z-axis direction, rotation about the X-axis and Y-axis (tilt) direction ( The relative position in the θ _X and θ _Y directions) is adjusted.

  The moving mechanism 17 includes a coarse movement stage and a fine movement stage arranged on the coarse movement stage. The position of the coarse movement stage is detected by a linear encoder, and the position accuracy is about 1 μm. The position of the fine movement stage relative to the coarse movement stage is detected by, for example, a capacitance sensor, and the positional accuracy is several nm. The position of the fine movement stage with respect to the reference base 1 is detected by a laser interferometer.

  An electron beam 42 is emitted from the exposure beam source 43. The electron beam 42 is irradiated onto the wafer 11 through the mask 12.

  The observation optical system 20 includes an image detection device 21, a lens 22, a beam splitter 23, and an optical fiber 24. In FIG. 2, the observation optical system 20Y2 shown in FIG. 1B is shown as a representative. The other observation optical systems 20X1, 20X2, and 20Y1 have the same configuration as the observation optical system 20Y2. The optical axis 25 of the observation optical system 20 is arranged so as to be parallel to the XZ plane and oblique to the exposed surface.

  Illumination light radiated from the optical fiber 24 is reflected by the beam splitter 23 to form a light bundle along the optical axis 25, and enters the exposed surface through the lens 22 at an angle.

  When the alignment marks provided on the wafer 11 and the mask 12 have scattering points such as edges or vertices, illumination light is scattered at these scattering points. Of the scattered light, light incident on the lens 22 is converged by the lens 22, and a part of the light passes through the beam splitter 23 and forms an image on the light receiving surface 29 of the image detection device 21. The imaging magnification on the light receiving surface 29 is, for example, 60 to 100 times.

  Light receiving pixels are arranged in a matrix on the light receiving surface of the image detection device 21. Each pixel generates an image signal corresponding to the pixel in accordance with the intensity of light emitted to the pixel. This image signal is input to the control device 30. Note that image signals from the other observation optical systems 20X1, 20X2, and 20Y1 are also input to the control device 30.

  The control device 30 performs image processing to obtain relative position information regarding the Y-axis direction between the alignment mark of the mask 12 and the alignment mark of the wafer 11.

  FIG. 3A is a plan view showing the relative positional relationship between the alignment mark on the wafer and the alignment mark on the mask. Alignment marks 13A and 13B on each wafer are configured by arranging three rectangular patterns in the Y-axis direction and 14 in the X-axis direction in a matrix. The alignment marks 13A and 13B constitute one alignment mark WY2 shown in FIG. One alignment mark 14 on the mask is configured by arranging three similar rectangular patterns in the Y-axis direction and five in the X-axis direction in a matrix. The alignment mark 14 corresponds to the alignment mark MY2 shown in FIG. In the state where the alignment is completed, the alignment mark 14 on the mask is arranged approximately at the center between the alignment marks 13A and 13B on the wafer in the Y-axis direction.

FIG. 3B is a cross-sectional view taken along one-dot chain line B3-B3 in FIG. The alignment marks 13A and 13B on the wafer are formed by patterning, for example, a SiN film or a polysilicon film formed on the exposed surface. A resist film 11 </ b> R is formed on the exposed surface of the wafer 11. The alignment mark 14 on the mask is configured by an opening formed in a mask membrane made of, for example, SiC. FIG. 4 is a sketch of an image on the light receiving surface 29 by scattered light from the edge. The horizontal direction (v-axis direction) in FIG. 4 corresponds to the Y-axis direction in FIG. 3A, and the vertical direction (u-axis direction) corresponds to the X-axis direction in FIG. Images 40A and 40B due to scattered light from alignment marks 13A and 13B on the wafer appear apart in the v-axis direction, and image 41 due to scattered light from alignment mark 14 on the mask appears therebetween. The images 40A and 40B and the image 41 appear at different positions with respect to the u-axis direction. By detecting the positions of the images 40A, 40B and 41 in the v-axis direction, positional information in the Y-axis direction between the alignment marks 13A and 13B on the wafer and the alignment mark 14 on the mask shown in FIG. Can be obtained.

The four sets of alignment marks shown in FIG. 1 (B), X-axis direction, it is possible to obtain positional information in the Y-axis direction, and theta _Z direction, and the relative distortion information of the wafer relative to the mask. Hereinafter, such information will be described. The coordinates of the alignment mark WX1 are (X1x, X1y), the coordinates of the alignment mark WX2 are (X2x, X2y), the coordinates of the alignment mark WY1 are (Y1x, Y1y), and the coordinates of the alignment mark WY2 are (Y2x, Y2y). The amount of deviation in the X-axis direction between the alignment marks WX1 and MX1 is dx1, the amount of deviation in the X-axis direction between the alignment marks WX2 and MX2 is dx2, the amount of deviation in the Y-axis direction between the alignment marks WY1 and MY1 is dy1, and the alignment mark The amount of deviation in the Y-axis direction between WY2 and MY2 is dy2. By detecting dx1, dx2, dy1, and dy2, the amount of deviation dx in the X-axis direction, the amount of deviation dy in the Y-axis direction, the amount of deviation dθ in the θz direction, and the relative expansion / contraction amount dm can be obtained.

With reference to FIG.5 and FIG.6, the alignment method by an Example is demonstrated. The coordinates obtained by measuring the position of the wafer holding table 15 on which the wafer is placed with a laser interferometer are defined as laser coordinates (lx, ly, lθ). Here, lx and ly are each an X and Y coordinates, Erushita shows theta _Z coordinates. The design coordinates of each unit area 50 defined on the wafer is (wx, wy). Further, the coordinates of the wafer holding table 15 obtained by the linear encoder of the coarse movement stage of the moving mechanism 17 and the capacitance sensor of the fine movement stage are set as encoder coordinates (ex, ey).

  In step MS1, the complete unit region 50a is processed, and a conversion matrix for obtaining converted coordinates from the design coordinates of the unit region 50 is acquired. Hereinafter, with reference to FIG. 6, the detailed process of step MS1 will be described.

  In step S <b> 1, the wafer to be exposed is placed on the wafer holder 15. Wafer position detection and leveling are performed using an observation optical system. As a result, the design coordinates (wx, wy) of the unit area defined on the wafer are associated with the encoder coordinates (ex, ey). That is, if design coordinates of a certain unit area are designated, the moving mechanism 17 can be driven to move the unit area to the exposure position. After the wafer 11 is placed on the wafer holder 15 and leveling is performed, the relative positions of the wafer 11, the wafer holder 15, and the fine movement stage of the moving mechanism 17 do not change. For this reason, detecting the position of the fine movement stage with the laser interferometer is substantially the same as detecting the position of the wafer 11 or the position of the wafer holder 15.

  In step S2, based on the design coordinates (wx, wy) of the unit area, the unit area 50 to be exposed is moved to just below the mask 12 (exposure position). More specifically, the encoder coordinates are obtained from the design coordinates (wx, wy), and the wafer holding table 15 is moved to the encoder coordinates (ex, ey). In this state, a laser interferometer that detects the position of the wafer holding table 15 is initialized. This movement of the wafer holder 15 is referred to as “alignment based on design coordinates”.

  Proceeding to step S3, the four observation optical systems 20X1, 20X2, 20Y1, and 20Y2 shown in FIG. 1B are used to observe the relative positions of the corresponding wafer and mask alignment marks. When the positional deviation exceeds the allowable value, high-precision alignment (die-by-die alignment) is performed by controlling the fine movement stage of the moving mechanism 17. In step S4, the unit region 50 arranged at the exposure position is exposed through the mask.

  Proceed to step S5. Hereinafter, the process of step S5 will be described. Laser coordinates (lx, ly, lθ) of the wafer holding table 15 after alignment based on the design coordinates in step S2, encoder coordinates (ex, ey), and positional deviation amount (dx, dy, dθ) and the design coordinates (wx, wy) of the unit region 50 arranged at the exposure position are stored in the control device 30. Even during exposure, the position of the alignment mark is detected at regular intervals, for example, every 30 ms. In this case, as the misregistration amount (dx, dy, dθ) that is actually stored, the misregistration amount immediately before exposure, the average value of misregistration amounts obtained within a certain period immediately before exposure, Any of a positional deviation amount, an average value of the positional deviation amounts obtained during exposure, an average value of the positional deviation amounts obtained immediately before and during exposure, and the like can be employed.

  Here, the subscripts x, y, and θ mean coordinates relating to the X-axis direction, the Y-axis direction, and the θz direction, respectively. A conversion matrix M for associating these coordinates is defined as follows. Since the actual position of the wafer holder is obtained by applying the conversion matrix M to the design coordinates, the conversion matrix M indicates arrangement information indicating how the unit area is arranged with respect to the wafer holder 15. Can be considered.

上記換算行列Ｍの第１行目の要素を求めるためには、

In order to obtain the element of the first row of the conversion matrix M,

を、最小二乗法を用いて解けばよい。その一例として、下記の正規方程式を解く方法が挙げられる。

Can be solved using the method of least squares. One example is a method of solving the following normal equation.

ここで、ｎは、得られた座標データの数であり、１つの単位領域のダイバイダイアライメントが完了すると、１つの座標データが得られる。

Here, n is the number of obtained coordinate data, and one coordinate data is obtained when the die-by-die alignment of one unit region is completed.

  Proceeding to step S6, it is determined whether the conversion matrix M can be obtained from the number of obtained coordinate data. If the conversion matrix M cannot be obtained, the process returns to step S2, and alignment is performed based on the design coordinates of the unexposed unit area to be exposed next. When the conversion matrix M can be obtained from the number of obtained coordinate data, the process proceeds to step S7, and the conversion matrix M is acquired.

  An example of this determination method will be described. If the number of obtained coordinate data is smaller than the number of unknowns to be obtained (three in the formula (2)), the normal equation cannot be solved, and the determination result is “uncalculated”. When the number of obtained coordinate data is greater than or equal to the number of unknowns, the simultaneous equations of Equation (3) are solved using, for example, a singular value decomposition method. If an error occurs during the calculation, the determination result is “uncalculated”. If no error has occurred, a conversion matrix is obtained. In step S7, this conversion matrix is acquired (stored).

  In step S5, the laser coordinates (lx, ly, lθ) and the encoder coordinates (ex, ey) of the wafer holding table 15 are stored, so that the laser coordinates (lx, ly, lθ) are encoded by the same method. A conversion matrix for converting to coordinates (ex, ey) is also acquired.

  Further, in the die-by-die alignment step of step S3, the relative expansion / contraction amount described in FIG. 1B is measured. This relative expansion / contraction amount is defined as dm. For example, it is assumed that the following relationship exists between the relative expansion / contraction amount dm and the design coordinates of the unit area.

  Similar to the above formula (3), the coefficients A, B, and C can be obtained by setting up and solving a normal equation. When the coefficients A, B, and C are determined, the relative expansion / contraction amount dm at the position can be predicted from the design coordinates (wx, wy) of the unit area. The coefficients A, B, and C are referred to as relative expansion / contraction amount estimation coefficients. In step S7, the relative expansion / contraction amount estimation coefficient is also acquired.

  In step S8, the conversion matrix M is applied to the design coordinates (wx, wy) of the unit area to be exposed next, and the conversion coordinates (rx, ry, rθ) to which the wafer holder 15 is to be moved are obtained. Specifically, the converted coordinates are calculated by the following formula.

  In step S9, the converted coordinates (rx, ry, rθ) are converted into encoder coordinates (ex, ey), and the moving mechanism 17 is driven to move the wafer holding table 15 to the encoder coordinates.

  The actual position of the unit area deviates from the position indicated by the design coordinates due to various factors. For example, factors such as misalignment during transfer of a pattern that has already been transferred and thermal expansion of the wafer can be cited. In this embodiment, in step S9, the wafer holder 15 is moved based on the converted coordinates calculated by applying the conversion matrix to the design coordinates instead of the design coordinates of the unit area.

  Even if the position of the wafer holding table 15 is detected by a laser interferometer and the wafer holding table 15 is moved so that the laser coordinates (lx, ly, lθ) coincide with the converted coordinates (rx, ry, rθ). Good. Compared with the case where the wafer holding table 15 is moved using only the encoder coordinates (ex, ey), alignment with higher accuracy can be performed.

  In both cases where encoder coordinates are used and laser coordinates are used, alignment is performed based on the converted coordinates. This alignment is called “alignment based on converted coordinates”.

  In step S10, die-by-die alignment is performed. In step S9, since the alignment based on the converted coordinates is performed, a higher position accuracy can be expected as compared with the case where the alignment is performed based on the design coordinates. For this reason, the moving amount of the stage at the time of die-by-die alignment can be shortened. This makes it possible to shorten the alignment time.

  In step S11, exposure is performed. In step S12, it is determined whether or not exposure of all complete unit areas has been completed. If there is an unexposed complete unit area, the process returns to step S8, and converted coordinates are calculated from the design coordinates of the complete unit area to be exposed next. Subsequently, steps S9 to S12 are executed.

  If exposure of all complete unit areas is completed, the process proceeds to step MS2 in FIG.

  In step MS2, conversion coordinates (rx, ry, rθ) are obtained by applying the conversion matrix M to the design coordinates of the incomplete unit region 50b to be exposed next. Proceeding to step MS3, the wafer holding table 15 is moved based on the converted coordinates. Thereby, the incomplete unit area 50b to be exposed is arranged at the exposure position.

  Proceeding to step MS4, a positional deviation is detected from the detectable alignment mark. Hereinafter, an example of a method for detecting misalignment will be described.

  When the alignment mark WY2 shown in FIG. 1B is missing and the remaining three alignment marks WX1, WX2, and WY1 are observable, the alignment mark positional deviations dx1, dx2, and dy1 can be determined. it can. From formula (4), an estimated value dm of the relative expansion / contraction amount of the incomplete unit region 50b is obtained. The positional deviation amounts dx, dy, dθ can be obtained from the deviation amounts dx1, dx2, dy1 of the alignment mark and the estimated value dm of the relative expansion / contraction amount.

  When the alignment marks WX2 and WY1 are missing and the alignment marks WX1 and WY2 are observable, the positional deviation amounts dx1 and dy2 of the alignment marks can be determined. The relative expansion / contraction amount dm of the incomplete unit region 50b is obtained from Expression (4). If it is assumed that the positional deviation amount dθ in the θz direction is 0, the positional deviation amounts dx and dy can be obtained from these pieces of information.

  When only the alignment mark WX1 can be observed, only the alignment mark positional deviation amount dx1 can be determined. Assuming that the positional deviation amount dy in the Y direction and the positional deviation amount dθ in the θz direction are zero, and assuming that the relative expansion / contraction amount dm is an estimated value obtained from Equation (4), the positional deviation amount dx can be obtained.

  Based on the obtained positional deviation amount, the wafer holding table 15 is moved so as to reduce the positional deviation amount, and the position of the wafer is corrected.

  Proceeding to step MS5, exposure is performed. Proceeding to step MS6, it is determined whether or not processing of all incomplete unit areas to be processed has been completed. If there is an unprocessed incomplete unit area, the process returns to step MS2, and the process is repeated for the unprocessed incomplete unit area. When the processing of all incomplete unit areas is completed, the wafer is unloaded from the wafer holding table 15.

  As described above, in the method according to the embodiment, the position of the wafer is corrected by using the positional deviation information obtained from the detectable alignment mark remaining in the incomplete unit area. For this reason, the alignment accuracy can be increased. If the alignment position accuracy based on the converted coordinates in step MS3 is sufficiently high, the position shift detection and position correction step in step MS4 may be omitted.

  Further, as described above, by adopting the method according to the embodiment, the distance for moving the wafer holder 15 in step S12 can be shortened, and the time required for alignment can be shortened. Thereby, the throughput can be improved.

  When die-by-die alignment is performed in step S10, position data including the position coordinates of the wafer holding table 15, position shift information, and design coordinates of the unit area is obtained. The conversion matrix M acquired in step S7 may be corrected in consideration of this position data. By creating the conversion matrix M from a larger amount of position data, the position accuracy of the converted coordinates can be further increased.

  Moreover, in the said Example, the process of collecting the position data for calculating | requiring the conversion matrix M from step S2 to S5 is shared with an actual exposure process. For this reason, there is almost no decrease in throughput due to the collection of position data.

  When performing electron beam exposure (for example, electron beam proximity exposure disclosed in Japanese Patent No. 2951947), if the amount of positional deviation between the mask and the wafer is small, the positional deviation is compensated by deflecting the electron beam. be able to. In the above embodiment, since the alignment accuracy after the alignment based on the converted coordinates in step S11 is high, the displacement may be compensated by deflecting the electron beam instead of the die-by-die alignment in step S12. Further, instead of correcting the position of the wafer in step MS4, the positional deviation may be compensated by deflecting the electron beam.

  You may perform the process from step S2 to S5 for calculating | requiring the conversion matrix M with respect to the several unit area | region which is not mutually adjacent. According to this method, a conversion matrix is obtained based on position data obtained from a relatively wide area in the wafer. For this reason, when the deviation from the design coordinates has a characteristic tendency in a part of the wafer, it is possible to reduce the influence of the characteristic tendency on the conversion matrix.

  In the above embodiment, the correspondence relationship between the design coordinates of the unit area and the coordinates of the wafer holder 15 is approximated by a linear expression, but it may be approximated by a quadratic expression. When approximated by a quadratic expression, the following relational expression is used instead of the above-described mathematical expression (1).

    Although the present invention has been described with reference to the embodiments, the present invention is not limited thereto. It will be apparent to those skilled in the art that various modifications, improvements, combinations, and the like can be made.

  In the above embodiment, the electron beam proximity exposure has been described as an example. In addition, the above alignment method can be applied to X-ray exposure, ultraviolet exposure, and semiconductor wafer inspection.

It is a top view of the wafer used as the object by the alignment method by the example of the present invention. 1 is a plan view of one unit region of a wafer and a schematic view of an observation optical system. It is the schematic of the alignment apparatus by an Example. It is a top view of the alignment mark formed in the wafer and the mask. It is sectional drawing in the position in which the alignment mark of the wafer and the mask was formed. It is the figure which sketched the image of the alignment mark. It is a flowchart of the alignment method by an Example. It is a flowchart which shows the detail of step MS1 of the alignment method by an Example. It is a top view of the wafer used as the object of the conventional alignment method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Reference base 10 Wafer / mask holding part 11 Semiconductor wafer 11R Resist film 12 Mask 13A, 13B, 14 Alignment mark 15 Wafer holding stand 16 Mask holding stand 17, 18 Moving mechanism 20 Optical system 20X1, 20X2, 20Y1, 20Y2 Observation optical system 21A, 21B Image detection device 22, 28 Lens 23, 26A Half mirror 24 Optical fiber 25 Optical axis 26B Mirror 29 Light receiving surface 30 Control devices 40A, 40B, 41 Image 42 Energy beam 43 Beam source for exposure 50 Unit area (die)
50a Complete unit area 50b Incomplete unit area 51 Chip area 60 Main controller MX1, MX2, MY1, MY2 Alignment mark

Claims (1)

(A) A substrate to be processed in which a plurality of complete unit regions whose entire area is included in the processing surface of the substrate to be processed and at least one incomplete unit region that is partially missing from the outer periphery of the substrate to be processed are defined The process of preparing
(B) Based on the result of detecting the position of the substrate to be processed after the complete unit region is arranged by die-by-die alignment by die-by-die alignment for at least a part of the complete unit region. Obtaining unit unit arrangement information;
(C) a step of aligning the substrate to be processed so that the incomplete unit region is arranged at the processing position based on the arrangement information acquired in the step b;
(D) obtaining a positional deviation information of the substrate to be processed by detecting a part of alignment marks formed in the incomplete unit region arranged at the processing position in the step c;
(E) a step of correcting the position based on the positional deviation information obtained in the step d,
The positions of the complete unit area and the incomplete unit area in the substrate to be processed are defined by design coordinates in the substrate to be processed determined at the time of design,
The arrangement information acquired in the step b includes a first correspondence between design coordinates of a unit area and position information of a substrate to be processed when the unit area is arranged at a processing position.
In the step c, to obtain position information of the substrate by applying the first correspondence to the design coordinates of an incomplete unit area, line physician alignment of the substrate on the basis of the acquired position information ,
Step b is
A step of acquiring the amount of expansion / contraction of the complete unit area arranged at the processing position by detecting the alignment mark;
Obtaining a second correspondence between the design coordinates of the unit area and the amount of expansion and contraction;
Including
Step c is
Applying the second correspondence to the design coordinates of the incomplete unit region to estimate the amount of expansion / contraction of the incomplete unit region;
A process of aligning the substrate to be processed in consideration of the estimated expansion and contraction amount;
Including an alignment method.

Images