JP2011138967A - Generating method, program, light source for exposure, and exposure apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To generate a light source image of an exposure apparatus by a simple method. <P>SOLUTION: A generating method of generating the light source image to be formed in an illumination light incident from a light source so that a desired illumination distribution is formed in an incident light entering an illumination optical system of the exposure apparatus having the illumination optical system guiding the illumination light to a mask includes: a calculating step of calculating a plurality of optimized light source images corresponding to an exposure pattern by the light source mask optimization technique; an acquisition step of acquiring a plurality of basis images perpendicular to one another from the plurality of optimized light source images by the principal component analysis; a composition step of composing a basis image subset by excluding a part of the plurality of basis images; an initial irradiation step of emitting an initial light source image onto the illumination optical system; a measuring step of measuring an incident image entering the illumination optical system; and a generation step of generating a corrected image by correcting the initial light source image based on a difference of coefficients in a case where the initial light source image and the incident image are developed into the basis image subsets. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a generation method, a program, an exposure light source, and an exposure apparatus.

  There is an exposure apparatus that can form an arbitrary intensity distribution in illumination light (see Patent Document 1). There is a light source mask optimization method (SMO: Source and Mask Optimization) in which a mask pattern (reticle pattern) and a light source image are optimized using such an exposure apparatus and a fine pattern is exposed with high accuracy.

JP 2002-353105 A

  However, even if the optimized light source image calculated by the light source mask optimization method is used, the calculated irradiation pattern is not always obtained due to the characteristics of the optical element. At that time, if the illumination light was to be brought close to the optimized light source image while maintaining the two-dimensional image information, a large amount of calculation processing and time were required.

  In order to solve the above-mentioned problems, as a first aspect of the present invention, in an exposure apparatus having an illumination optical system that guides illumination light to a mask, an illuminance distribution required for incident light incident on the illumination optical system is formed. A generation method for generating a light source image to be formed on illumination light incident from a light source, a calculation step for calculating a plurality of optimized light source images according to an exposure pattern by a light source mask optimization method, and a plurality of components by principal component analysis An acquisition step of acquiring a plurality of orthogonal base images from the optimized light source image, a composition step of excluding a part of the plurality of base images and composing a base image subset, and irradiating the illumination optical system with the initial light source image The initial irradiation step, the measurement step for measuring the incident image incident on the illumination optical system, and the initial light source image and the incident image are each expanded into a base image subset. Generating method comprising a generating step of generating a corrected image obtained by correcting the initial source image is provided based on the difference in coefficient when.

  As a second aspect of the present invention, there is provided a program that causes a processing device to execute the generation method. Furthermore, the light source for exposure which performs the said production | generation method is provided as 3rd aspect of this invention. Furthermore, as a fourth aspect of the present invention, there is provided an exposure apparatus that executes the above generation method.

  The above summary of the present invention does not enumerate all necessary features of the present invention. A sub-combination of these feature groups can also be an invention.

1 is a schematic diagram showing the overall structure of an exposure apparatus 100. FIG. 3 is a cross-sectional view schematically showing an individual structure of a reflecting mirror 222. FIG. 3 is a schematic perspective view of a spatial light modulator 220. FIG. FIG. 6 is a diagram illustrating the operation of the spatial light modulator 220. It is a flowchart showing the generation procedure of the source image I _2. It is a figure which shows typically the optimized light source image group 620 and the base image group 630. FIG. FIG. 10 is a diagram illustrating a calculation method of a base image group 630. FIG. 10 is a diagram illustrating a calculation method of a base image group 630. FIG. 10 is a diagram illustrating a calculation method of a base image group 630. FIG. 10 is a diagram illustrating a calculation method of a base image group 630. It is a flowchart which shows the manufacturing process of a microdevice. It is a flowchart which shows the content of a board | substrate process step.

  Hereinafter, the present invention will be described through embodiments of the invention. The following embodiments do not limit the invention according to the claims. Not all combinations of features described in the embodiments are essential for the solution of the invention.

  FIG. 1 is a schematic diagram showing the overall structure of the exposure apparatus 100. The exposure apparatus 100 includes a light source unit 200, an illumination optical system 300, and a projection optical system 400.

The light source unit 200 includes a light source 110, a control unit 210, a spatial light modulator 220, a prism 230, an imaging optical system 240, a beam splitter 250, and a measurement unit 260. The light source 110 generates illumination light L. The illumination light L generated by the light source 110 has an illuminance distribution according to the characteristics of the light emitting mechanism of the light source 110. For this reason, the illumination light L has the original image I ₁ in a cross section orthogonal to the optical path of the illumination light L.

  The illumination light L emitted from the light source 110 is incident on the prism 230. The prism 230 guides the illumination light L to the spatial light modulator 220 and then emits the light again to the outside.

  The spatial light modulator 220 modulates the illumination light L incident under the control of the control unit 210. The structure and operation of the spatial light modulator 220 will be described later with reference to other drawings.

The illumination light L emitted from the prism 230 via the spatial light modulator 220 is incident on the illumination optical system 300 at the subsequent stage via the imaging optical system 240. The imaging optical system 240 forms an illumination light image I ₃ on the incident surface 312 of the illumination optical system 300.

  The beam splitter 250 is disposed on the optical path of the illumination light L between the imaging optical system 240 and the illumination optical system. The beam splitter 250 separates a part of the illumination light L before entering the illumination optical system 300 and guides it to the measurement unit 260.

The measurement unit 260 measures the image of the illumination light L at a position optically conjugate with the incident surface 312 of the illumination optical system 300. Thereby, the measurement unit 260 measures the same image as the illumination light image I ₃ incident on the illumination optical system 300. Therefore, the control unit 210 can feedback control the spatial light modulator 220 with reference to the illumination light image I ₃ measured by the measurement unit 260.

  The illumination optical system 300 includes a fly-eye lens 310, a condenser optical system 320, a field stop 330, and an imaging optical system 340. A mask stage 420 holding an exposure mask 410 is disposed at the exit end of the illumination optical system 300.

The fly-eye lens 310 includes a large number of lens elements arranged densely in parallel, and forms a secondary light source including illumination light images I ₃ as many as the number of lens elements on the rear focal plane. The condenser optical system 320 condenses the illumination light L emitted from the fly-eye lens 310 and illuminates the field stop 330 in a superimposed manner.

The illumination light L that has passed through the field stop 330 forms an irradiation light image I ₄ that is an image of the opening of the field stop 330 on the pattern surface of the exposure mask 410 by the imaging optical system 340. Thus, the illumination optical system 300, the pattern surface of the exposure mask 410 disposed at its exit end, Koehler illuminated by illumination light image I _4.

Note that the illuminance distribution formed at the entrance end of the fly-eye lens 310 that is also the entrance surface 312 of the illumination optical system 300 is higher than the overall illuminance distribution of the entire secondary light source formed at the exit end of the fly-eye lens 310. Show correlation. Therefore, the illumination light image I ₃ that the light source unit 200 enters the illumination optical system 300 is also reflected in the illumination light image I ₄ that is the illuminance distribution of the illumination light L that the illumination optical system 300 irradiates the exposure mask 410.

  The projection optical system 400 is disposed immediately after the mask stage 420 and includes an aperture stop 430. The aperture stop 430 is disposed at a position optically conjugate with the emission end of the fly-eye lens 310 of the illumination optical system 300. A substrate stage 520 that holds a substrate 510 coated with a photosensitive material is disposed at the exit end of the projection optical system 400.

The exposure mask 410 held on the mask stage 420 has a mask pattern composed of a region that reflects or transmits the illumination light L irradiated by the illumination optical system 300 and a region that absorbs it. Therefore, by irradiating the illumination light image I ₄ to the exposure mask 410, the projection light image I ₅ produced by the interaction of the illuminance distribution of the illumination light image I ₄ itself as a mask pattern of an exposure mask 410. The projected light image I ₅ is projected onto the photosensitive material of the substrate 510 to form a resist layer having the required pattern on the surface of the substrate 510.

  In FIG. 1, the optical path of the illumination light L is drawn in a straight line, but the exposure apparatus 100 is miniaturized by bending the optical path of the illumination light L. Although FIG. 1 depicts the illumination light L so that it passes through the exposure mask 410, a reflective exposure mask 410 may be used.

  FIG. 2 is a diagram illustrating the structure of the spatial light modulator 220, and shows a part of the spatial light modulator 220 in an enlarged manner. The spatial light modulator 220 includes a reflecting mirror 222, a substrate 224, a flexure 226, and electrodes 223 and 225. The reflecting mirror 222 is suspended from the lower surface of the substrate 224 via the flexure 226. The flexure 226 is formed of a material that can be easily deformed. Therefore, the reflecting mirror 222 is supported so as to be swingable with respect to the substrate 224.

  One electrode 223 is disposed on the back surface of the reflecting mirror 222 so as to face the substrate 224. The other electrode 225 is disposed on the surface of the substrate 224 so as to face the back surface of the reflecting mirror 222. The electrode 225 disposed on the substrate 224 is divided into a plurality of parts, and a voltage can be applied individually. With such a structure, a voltage can be applied to any of the electrodes 225 on the substrate 224 to cause an electrostatic force to act between the electrodes 223 of the reflecting mirror 222 and to give the required tilt to the reflecting mirror 222. .

  FIG. 3 is a schematic perspective view of the spatial light modulator 220. As shown in the drawing, the spatial light modulator 220 is formed by two-dimensionally arranging a structure including the reflecting mirror 222 and the electrodes 223 and 225 as described above on a single substrate 224. The plurality of reflecting mirrors 222 individually swing according to the control of the control unit 210.

  For the purpose of clarifying the drawing, FIG. 3 shows a spatial light modulator 220 including 16 reflecting mirrors 222. However, the spatial light modulator 220 mounted on the exposure apparatus 100 includes a large number of reflecting mirrors 222 according to the accuracy of the pattern to be formed.

  FIG. 4 is a partially enlarged view of the light source unit 200 and shows the operation of the light source unit 200 including the spatial light modulator 220. The prism 230 has a pair of reflecting surfaces 232 and 234. The illumination light L incident on the prism 230 is irradiated toward the spatial light modulator 220 by the one reflecting surface 232.

As already described, the spatial light modulator 220 has a plurality of reflecting mirrors 222 that can be individually swung. Therefore, the control unit 210 controls the spatial light modulator 220 can be formed of any light source image I ₂ corresponding to the request.

The light source image I ₂ emitted from the spatial light modulator 220 is emitted from the prism 230 by the other reflecting surface 234 of the prism 230. The light source image I ₂ emitted from the prism 230 forms an illumination light image I ₃ on the incident surface 312 of the illumination optical system 300 by the imaging optical system 240.

Next, a method for generating the light source image I ₂ generated by the spatial light modulator 220 in the exposure apparatus 100 as described above will be described.

Figure 5 is a flowchart illustrating a procedure for generating source image I _2. As shown in the figure, the method for generating the light source image I ₂ includes a preparation stage P 201 that is performed for each mask pattern, and an operation stage P 202 that is repeated throughout the operation period of the exposure apparatus 100.

  FIG. 6 is a diagram illustrating an example of an image that is a target of the procedure illustrated in FIG. 5. FIG. 6 shows a pattern 610 intended to be formed on the substrate 510, an optimized image group 620 including a plurality of optimized light source images 622, 624, and 626 calculated for the pattern 610, and an optimized image group. A base image group 630 acquired from 620 is shown. Hereinafter, the preparation stage P201 will be described with reference to FIG.

First, the pattern 610 for the purpose of forming on a substrate 510, a suitable light source image I ₁ in the case of exposing the pattern 610 is calculated by a light source mask optimization method (step S201). The optimized light source image group 620 can be expressed by, for example, the following formula [Equation 1].

  At this stage, the optimized light source image group 620 including more optimized light source images is calculated by changing the optimization conditions in various ways. Note that the optimization conditions that can be changed include, for example, selection of which basic pattern to be targeted among the masks to be exposed. In addition, the algorithm of the light source mask optimization method tends to seek a solution that reduces the area of the light source, so the maximum area of the illumination light source determined by the energy density is allowed. Can be exemplified as one of the optimization condition options.

  Next, a base image group 630 including a plurality of base images orthogonal to each other is obtained by principal component analysis of the plurality of optimized light source images included in the above-described optimized light source image group 620 (step S202). At this stage, the number of base images acquired here is large, including those with small coefficients. Theoretically, Z base images can be obtained for Z pixels of the optimized light source image.

7 to 10 are diagrams illustrating an example of a procedure for acquiring the base image group 630 in step S202. First, as shown in FIG. 7, respectively _{[I k (x, y)} ] of a plurality of optimization the light source images calculated in step S201 one-dimensionally arranged in accordance with a certain rule. Thereby, the correlation between the plurality of optimized light source images can be statistically analyzed as a high-dimensional vector [I _k ] with the number of pixels forming the image as the number of dimensions.

Then, the high-dimensional vector [I _k] of one (here, vector [I _1]) regarded as a column vector and multiplying the horizontal vector obtained by transposing the vector [I _1]. Thus, as shown in FIG. 8, the two-dimensional matrix M ₁ to a value multiplied by the respective component and component is formed.

Formation of such a two-dimensional matrix M ₁ is performed for each of the plurality of optimized source image. Thereby, a two-dimensional matrix _Mk corresponding to the number of optimized light source image groups included in the optimized light source image group 620 is formed.

Next, as shown by the following formula [Equation 2], a plurality of two-dimensional matrices M _k obtained by the above processing are added.

The matrix thus obtained represents the degree of correlation of each high-dimensional vector [I ₁ ]. Hereinafter, eigenvalue analysis is performed using this covariance matrix.

The matrix obtained by the above processing is decomposed as a symmetric matrix into a product form formed by a matrix U and a diagonal matrix formed by k eigenvalues as shown in the following equation [Equation 3]. it can.

Therefore, if the matrix U is regarded as n vertical vectors arranged in the horizontal direction, n eigenvectors corresponding to the respective eigenvalues, that is, the matrix M, as shown in the following equation [Formula 4]. A vector is obtained that is a simple product of the eigenvalues for the operation.

  Next, as shown in FIG. 9, the eigenvectors obtained above can be returned to the two-dimensional image information according to the same rules as the operations shown in FIG. Thus, n base images can be acquired as shown in FIG. The base image acquisition method is not limited to the above procedure. For example, the calculation can be speeded up by other methods such as singular value decomposition.

  The base image group 630 obtained as described above has a property that the base image having the largest eigenvalue (or expansion coefficient) has the most weight. Therefore, the base image group 630 can be efficiently operated by rearranging the base images included in the base image group 630 in order from the largest eigenvalue.

  That is, as will be described later, when an image measured by the measurement unit 260 is evaluated, a feature of many images is represented by a base image having a large eigenvalue. Therefore, the processing amount can be reduced by processing by removing the base image having a small eigenvalue. Therefore, referring to FIG. 5 again, a base image is selected from the base image group 630 having a larger eigenvalue, and a base image subset 640 (FIG. 6) having a small number of base images is composed (step S203).

  The base image subset 640 excludes, for example, eigenimages belonging to eigenvalues of less than 1% of the maximum eigenvalues among the base images included in the base image group 630 having the eigenvalue λ group expressed by the formula [Equation 3]. The composition can be obtained. This is because the approximate accuracy with which an arbitrary image can be developed and replaced with a base image is about 1%. However, it goes without saying that the subset may be composed by setting other threshold values as appropriate.

Hereinafter, in the operational phase of the base image subset 640 obtained (P202), a description will be given of a procedure for generating a source image _{I 1.}

First, expansion coefficients c ₀₁ , c ₀₂ , c ₀₃ ... _Are given as initial values to the light source unit 200 of the exposure apparatus 100. As the initial value, for example, an expansion coefficient when one optimized image is selected from the optimized light source image group 620 and expanded in the base image subset 640 can be exemplified. In this case, one of the optimized images formed with the initial value may be one of the target images of the illumination light image.

Using the initial values c ₀₁ , c ₀₂ , c ₀₃ ..., The control unit 210 generates an initial light source image I ₂ as a linear sum of the base image subset 640 represented by the following equation [Equation 5].

Source image _{I 2} of the initial is generated by the spatial light modulator 220 is irradiated to the illumination optical system 300 (step S204). If the initial source image I ₂ is irradiated to the illumination optical system 300, the measuring unit 260 measures receiving part of the illumination light L is split by the beam splitter 250 (step S205). The image measured by the measurement unit 260 is equal to the irradiation light image I ₃ for the illumination optical system 300.

Next, the image measured by the measurement unit 260 is developed into a linear sum of the base images included in the base image subset 640 as shown in the following equation [Formula 6]. Thereby, the expansion coefficient C (C = c ₁₁ , c ₁₂ , c ₁₃ ...) Unique to the irradiation light image I ₃ is calculated (step S206).

Therefore, based on the difference between the expansion coefficients c ₀₁ , c ₀₂ , c ₀₃ ... As the initial values and the expansion coefficients c ₁₁ , c ₁₂ , c _13. I ₃ can be evaluated (step S207).

Or the measured value and the initial value are equal, or, if sufficiently small difference (step S207: YES), the irradiation light image I ₃ is seen that well reflects the light source image I ₂ as an initial value. More specifically, for example, when the following formula [Equation 7] is satisfied, it can be said that the difference between the measured value and the initial value is sufficiently small (RSS means the square sum of squares).

Therefore, the expansion coefficient as the initial value is not corrected. However, when the measured value and the initial value are significantly different (step S207: NO), the initial value is corrected to calculate correction values c ₂₁ , c ₂₂ , c ₂₃ ... (Step S208).

As a method for calculating the correction values c ₂₁ , c ₂₂ , c ₂₃ ..., For example, the correction values c ₂₁ , such that the difference between the initial value and the measurement value becomes smaller in the entire linear sum of the base image subset 640. It calculates a _{c 22,} c ₂₃ ···, based on the correction value _{c 21, c 22, c 23} ···, a method of swinging the reflector 222 of the spatial light modulator 220 can be exemplified. Thus, a very simple process, the irradiation light image I _3, can be brought close to the light source image I _2. Such calculation of the correction values ₂₁ , c ₂₂ , c ₂₃ ... Is more strictly performed by repeating the calculation using the calculated correction value as the next initial value, thereby changing the illumination image I ₃ to the light source image I _2. You can get closer.

  Further, the eigenvalue (expansion coefficient) of the optimized light source image calculated in step S201 as the optimized light source image may be calculated in advance and held in the control unit 210, and compared with the expansion coefficient in the measurement value. Thereby, an optimized light source image with a close distribution of expansion coefficients can be selected and used as the next initial value.

For example, for the purpose of generating an initial value, one optimized light source image 622 is selected to measure the illumination light image I _3, and expansion coefficients c ₁₁ , c ₁₂ , c _13. If it is close to another optimized light source image 624, the optimized light source image 624 can be selected for the next initial value generation. Thereby, the processing amount and time required until the correction value converges can be shortened.

Thus, by executing the operational phase P202 again on that reflects the difference between the measured value, the illumination light image I ₃ can be brought closer to the desired pattern. In either case, the causal relationship between the images is easy to understand and the amount of information processing is small compared to the process of performing feedback for each pixel of the light source image I ₂ or the illumination light image I ₃ . Therefore, the processing can be speeded up, which contributes to the improvement of the throughput of the exposure apparatus 100.

  Of course, the correction method of the initial value is not limited to the above method. Further, the operation stage P202 may be repeatedly performed at some point, such as when the lot of the substrate 510 to be exposed is updated. Thereby, there is no variation among lots, and a stable exposure process can be executed over a long period of time.

Furthermore, in the above-described embodiment, the exposure apparatus 100 including the control unit 210 that executes the above-described method has been described. However, as the target of the exposure apparatus 100 having the hardware resources that can change the light source image I _2, may be provided a program for executing the above-described generation method. Further, the light source unit 200 including the control unit 210 in which such a control program is mounted may be provided in a form that can be replaced with the light source unit 200 of the existing exposure apparatus 100.

  Furthermore, in the above example, the illuminance distribution of the illumination light L has been described. However, for the spatial frequency distribution (power spectrum distribution) of the illumination light L, a base image is prepared and measured based on the expansion coefficient. The value can be evaluated. That is, when the power spectrum analysis method related to the correlation function analysis method is used through Fourier transform, the above method can be applied to the evaluation of the power spectrum image.

  FIG. 11 is a flowchart showing a manufacturing process of a microdevice to which the above generation method can be applied. Note that a micro device includes a semiconductor device such as an integrated circuit or a large-scale integrated circuit, a liquid crystal display panel, an image sensor such as a CCD or C-MOS, a thin film magnetic head, a micromachine, or the like.

  First, the function and performance of the target microdevice are determined, and a structure that realizes the function and performance is designed (for example, circuit design of a semiconductor device). Moreover, the design of the pattern formed in a semiconductor substrate etc. is also included (step S101: design step).

  Next, a mask for forming the circuit pattern designed in the design step is manufactured (step S102: mask manufacturing step). Further, a substrate is manufactured using a material such as silicon, compound semiconductor, glass, ceramics (step S103: substrate manufacturing step).

  Subsequently, the circuit board is manufactured by processing the substrate manufactured in the substrate manufacturing step using the mask manufactured in the mask manufacturing step (step S104: substrate processing step). Furthermore, a device is assembled using the substrate processed in the substrate processing step (step S105: device assembly step). The device assembly step includes processes such as dicing, bonding, and packaging (chip encapsulation) as required.

  The device assembled as described above is subjected to inspections such as an operation confirmation test and a durability test (step S106: inspection step). The microdevice thus inspected is shipped as a product.

  FIG. 12 is a flowchart showing the contents of the substrate processing step. That is, FIG. 12 is a diagram showing details of step S104 shown in FIG. In the substrate processing step, a circuit is formed on the substrate by lithography, for example.

  Steps S111 to S114 described below form a pretreatment process of the substrate processing step. The processing in each step is not necessarily essential, and is appropriately selected according to the specifications of the microdevice to be manufactured.

  In the pretreatment process, first, the surface of the substrate is oxidized to form an oxide film (step S111: oxidation step). Next, an insulating film is formed on the surface of the substrate (step S112: CVD step). Hereinafter, for example, electrodes are formed by vapor deposition (step S113: electrode formation step), and ions are also implanted into the substrate (step S114: ion implantation step).

  Next, a post-processing step in the substrate processing step is executed. In the post-processing step, first, a photosensitive material is applied to the substrate (step S115: resist formation step). Next, the photosensitive material is exposed by the exposure apparatus 100 with an exposure pattern corresponding to the target pattern (step S116: exposure step). Next, the exposed photosensitive material is developed to form a mask layer having a pattern on the surface of the substrate (step S117: development step).

  Next, the region exposed from the mask layer is removed by etching (step S118: etching step). Further, the already used mask layer is removed (step S119: resist removal step). In this way, a circuit is formed on the surface of the substrate, but the pre-processing step and the post-processing step are repeatedly executed, and a multi-layered circuit is formed on the substrate.

  As mentioned above, although this invention was demonstrated using embodiment, the technical scope of this invention is not limited to the range as described in the said embodiment. It will be apparent to those skilled in the art that various modifications or improvements can be added to the above-described embodiment. It is apparent from the scope of the claims that the embodiments added with such changes or improvements can be included in the technical scope of the present invention.

  The execution order of the procedures, steps, stages, etc. in the operation and method of the apparatus shown in the claims, the specification, and the drawings is clearly indicated as “before”, “prior”, etc. Alternatively, it can be realized in any order except when the output of the previous stage is used in the subsequent stage. Even if “first”, “next”, and the like are used for convenience in the claims, the description, and the drawings, it does not necessarily mean that execution in this order is essential.

DESCRIPTION OF SYMBOLS 100 Exposure apparatus, 110 Light source, 200 Light source part, 210 Control part, 220 Spatial light modulator, 222 Reflection mirror, 223, 225 Electrode, 224 Substrate, 226 Flexure, 230 Prism, 232, 234 Reflection surface, 240, 340 Imaging Optical system, 250 beam splitter, 260 measuring unit, 300 illumination optical system, 312 entrance plane, 310 fly-eye lens, 320 condenser optical system, 330 field stop, 400 projection optical system, 410 exposure mask, 420 mask stage, 430 aperture stop , 510 substrate, 520 substrate stage, 610 pattern, 620 optimized light source image group, 622, 624, 626 optimized illumination image, 630 base image group, 640 base image subset

Claims (10)

In an exposure apparatus having an illumination optical system that guides illumination light to a mask, a light source image to be formed on the illumination light incident from the light source is generated so that the illuminance distribution required for the incident light incident on the illumination optical system is formed A generation method to
A calculation step of calculating a plurality of optimized light source images according to the exposure pattern by a light source mask optimization method;
Acquiring a plurality of base images orthogonal to each other from the plurality of optimized light source images by principal component analysis;
A composition step of excluding a portion of the plurality of base images and composing a base image subset;
An initial irradiation step of irradiating the illumination optical system with an initial light source image;
A measurement step of measuring an incident image incident on the illumination optical system;
A generation method including: generating a corrected image obtained by correcting the initial light source image based on a difference in coefficients when each of the initial light source image and the incident image is expanded into the base image subset.   The generation method according to claim 1, wherein in the composition step, a base image having a small eigenvalue is excluded.   3. The generation method according to claim 1, wherein, in the initial irradiation step, the initial light source image is generated as a linear sum of the base image subset given a coefficient as an initial value.   4. The correction image according to claim 1, wherein in the generation step, the corrected image is calculated so that a difference between eigenvalues of the initial image and the incident image is reduced in a whole linear sum of the base image subset. The generation method described. The composition step further includes a storage step of storing, as a reference value, a linear sum coefficient developed in the base image subset for each of the plurality of optimized light source images.
4. The generation method according to claim 1, wherein in the generation step, one of the plurality of optimized light source images having a coefficient close to a linear sum of the measurement images is selected as the correction image.   The generation method according to claim 1, wherein the calculation step, the acquisition step, and the composition step are performed for each mask loaded in the exposure apparatus.   The generation method according to any one of claims 1 to 6, wherein the initial irradiation step, the measurement step, and the generation step are executed each time an exposure target lot is updated.   A program that causes a processing apparatus to execute the generation method according to claim 1.   An exposure light source that executes the generation method according to claim 1.   An exposure apparatus that executes the generation method according to claim 1.