JP5920610B2 - Pupil intensity distribution setting method, illumination optical system and adjustment method thereof, exposure apparatus, and device manufacturing method - Google Patents

Description

  The present invention relates to a pupil luminance distribution setting method, an illumination optical system and its adjustment method, an exposure apparatus, and a device manufacturing method.

In an exposure apparatus used for manufacturing a device such as a semiconductor element, high resolution has been demanded as the device pattern becomes finer. The resolution R is expressed by the Rayleigh equation, that is, R = k ₁ (λ / NA). Here, λ is the wavelength of illumination light (exposure light), NA is the numerical aperture of the projection optical system, and k ₁ is a process factor that depends on the resolution and / or process controllability of the resist. Therefore, conventionally, resolution has been improved by shortening the wavelength of illumination light and increasing the numerical aperture of projection optical systems (higher NA).

However, the shortening of the exposure wavelength involves difficulties in developing light sources and glass materials, for example. In addition, the increase in NA causes a decrease in the DOF (Depth Of Focus) of the projection optical system, which in turn causes deterioration in imaging performance (imaging characteristics). Therefore, the increase in NA cannot be promoted unnecessarily. For these reasons, it has been difficult to cope with miniaturization of circuit patterns and sizes only by shortening the wavelength of illumination light and increasing the numerical aperture of projection optical systems (higher NA). For this reason, in the exposure apparatus, improved lithography performance (low-k) due to the introduction of deformation illumination technology such as annular illumination and optical life extension technology such as phase shift mask, material development in resist technology and process development technology such as thinning and multilayering effort of ₁ reduction) have been made.

In recent years, an exposure apparatus that employs a local immersion exposure technique has been put into practical use as an apparatus that achieves higher NA. However, even in an immersion exposure apparatus, there is a limit in increasing the NA well shorter wavelength of the illumination light, low k ₁ reduction has become essential. In recent years, low-k ₁ not only makes full use of super-resolution technology such as modified illumination and phase shift technology, but also transfers optical system errors such as aberrations and mask (reticle) patterns onto a wafer (photosensitive substrate). In this case, it has become necessary to consider optical proximity correction (OPC) that compensates for the deterioration of the fidelity of the pattern with a mask pattern. However, low k ₁ reduction, since lead to reduced contrast of the pattern image, also considered a need for this.

In this technical background, SMO (Source and Mask Optimization) optimizes the mask pattern and the light intensity distribution at the illumination pupil simultaneously using an optical model in order to provide an optical imaging solution that can be mass-produced with a low k ₁ value. ) Method and SO (Source Optimization) method for optimizing the light intensity distribution at the illumination pupil (see, for example, Patent Document 1). Hereinafter, the light intensity distribution in the illumination pupil is referred to as “pupil luminance distribution”. The illumination pupil is defined as a position where the illumination surface becomes the Fourier transform plane of the illumination pupil by the action of the optical system between the illumination pupil and the illumination surface (a mask or a wafer in the case of an exposure apparatus). The

US Pat. No. 6,563,566

  However, the conventional SMO and SO methods focus on maximizing the PW (Process Window), and MEEF (Mask Error Enhancement Factor: Mask Error Enhancement), which is important as an exposure evaluation index. Factor) is not taken into account. Incidentally, MEEF is an index indicating the degree to which the mask pattern manufacturing error is amplified on the wafer.

  The present invention has been made in view of the above-described problems, and an object thereof is to provide a setting method capable of setting the pupil luminance distribution so that the process window becomes a predetermined value or more while suppressing the MEEF to be small. . In addition, the present invention appropriately adjusts the pupil luminance distribution by using the setting method of setting the pupil luminance distribution so that the process window becomes a predetermined value or more while keeping the MEEF small. An object of the present invention is to provide an illumination optical system capable of performing the above. Another object of the present invention is to provide an exposure apparatus that can perform good exposure under appropriate illumination conditions using an illumination optical system that appropriately adjusts the pupil luminance distribution.

In order to solve the above-mentioned problem, in the first embodiment, an image of a pattern arranged on the first surface is formed on the illumination pupil of the illumination optical system that supplies illumination light to the imaging optical system that forms the second surface. A method for setting a power pupil luminance distribution,
Calculating a process window using a predetermined pattern arranged on the first surface and a pupil luminance distribution represented by a plurality of luminance elements distributed two-dimensionally along the illumination pupil;
And setting a pupil luminance distribution so that the process window becomes a predetermined value or more with a permissible MEEF (mask error enhancement factor) as a constraint condition. I will provide a.

In the second embodiment, in the adjustment method of the illumination optical system that illuminates the illuminated surface with light from the light source,
Setting a pupil luminance distribution to be formed on the illumination pupil of the illumination optical system using the setting method of the first embodiment;
And adjusting the pupil luminance distribution formed on the illumination pupil with the set pupil luminance distribution as a target.

  In the third embodiment, an illumination optical system characterized by being adjusted by the adjustment method of the second embodiment is provided.

In the fourth embodiment, in the illumination optical system that illuminates the illuminated surface with light from the light source,
A pupil distribution measuring device for measuring a pupil luminance distribution formed on the illumination pupil of the illumination optical system;
A pupil adjustment device for adjusting a pupil luminance distribution formed in the illumination pupil;
And a control unit that controls the pupil adjustment device to adjust the pupil luminance distribution formed in the illumination pupil with the pupil luminance distribution set by using the setting method of the first form as a target. An illumination optical system is provided.

  According to a fifth aspect, there is provided an exposure apparatus comprising the illumination optical system according to the third or fourth aspect for illuminating a predetermined pattern, and exposing the predetermined pattern onto a photosensitive substrate.

In the sixth embodiment, using the exposure apparatus of the fifth embodiment, exposing the predetermined pattern to the photosensitive substrate;
Developing the photosensitive substrate having the predetermined pattern transferred thereon, and forming a mask layer having a shape corresponding to the predetermined pattern on the surface of the photosensitive substrate;
And processing the surface of the photosensitive substrate through the mask layer. A device manufacturing method is provided.

  In the setting method of the pupil luminance distribution according to the embodiment, the pupil luminance distribution can be set so that the process window becomes a predetermined value or more while suppressing MEEF small. In the illumination optical system according to the embodiment, the pupil luminance distribution is appropriately set by using the setting method for setting the pupil luminance distribution so that the process window becomes a predetermined value or more while suppressing the MEEF to be small. Can be adjusted. In the exposure apparatus of the embodiment, it is possible to perform good exposure under appropriate illumination conditions by using an illumination optical system that appropriately adjusts the pupil luminance distribution, and as a result, it is possible to manufacture a good device.

It is a flowchart which shows the algorithm of the setting method of the pupil luminance distribution concerning embodiment. It is a flowchart which shows the process of the optimization of DOF in step S11 and S15 of FIG. It is a flowchart which shows the process of optimization of MEEF performed in step S21 and S22 of FIG. It is a figure which shows the ellipse which defines the magnitude | size of a process window. It is a figure which shows a mode that the illumination pupil was divided | segmented into grid shape. It is a figure which shows a mode that each illumination pixel has a significant blur. It is a figure explaining the position which restrict | limits an intensity value with a constraint condition. It is a figure explaining the evaluation point for optimization of a process window. It is a figure explaining the evaluation point for optimization of MEEF. It is a figure explaining the evaluation point for optimization of an OPE error. It is a figure which shows the pupil luminance distribution obtained by applying the setting method of the pupil luminance distribution concerning this embodiment. It is a figure which shows the OPE error corresponding to each pupil luminance distribution of FIG. It is a figure which shows schematically the structure of the exposure apparatus concerning embodiment. It is a figure which shows roughly the structure and effect | action of a spatial light modulation unit. It is a fragmentary perspective view of the spatial light modulator in a spatial light modulation unit. It is a figure which shows roughly the internal structure of a pupil distribution measuring apparatus. It is a flowchart of the adjustment method of the illumination optical system concerning embodiment. It is a flowchart which shows the manufacturing process of a semiconductor device. It is a flowchart which shows the manufacturing process of liquid crystal devices, such as a liquid crystal display element.

  Hereinafter, embodiments will be described with reference to the accompanying drawings. First, a method for setting a pupil luminance distribution according to the embodiment will be described. In the setting method of the pupil luminance distribution according to the present embodiment, an image of a pattern (mask pattern) arranged on the first surface is formed on the second surface (wafer surface: exposure surface), assuming an exposure apparatus as an example. The pupil luminance distribution to be formed on the illumination pupil of the illumination optical system that supplies illumination light to the imaging optical system (projection optical system) is improved. In the present specification, “improvement” means that the process window in the pupil luminance distribution after improvement is larger than the process window in the pupil luminance distribution in the initial state before improvement. At this time, the pupil luminance distribution in the initial state may be a distribution in which the luminance is uniformly zero in the pupil plane.

  In this embodiment, MEEF (Mask Error Enhancement Factor), OPE (Optical Proximity Effect) error, and PW (Process Window), which are important as exposure evaluation indexes, are improved. We propose an algorithm for deriving the free-form pupil luminance distribution. The OPE error in the exposure apparatus is, for example, a line width error of a resist pattern formed on a wafer (photosensitive substrate). A user-specified value can be given to each of the MEEF and OPE errors. In the present embodiment, as an example, a free-form pupil luminance distribution that is improved so that the process window is maximized is obtained after each MEEF and OPE error is suppressed to a user-specified value or less.

  FIG. 1 is a flowchart illustrating an algorithm of a pupil luminance distribution setting method according to the embodiment. FIG. 2 is a flowchart showing a process of optimizing the best focus position and DOF in steps S11 and S15 of FIG. FIG. 3 is a flowchart showing the MEEF optimization process executed in steps S21 and S22 of FIG. In the optimization of the best focus position and DOF (Depth Of Focus: the allowable range of the defocus amount from the best focus position) in steps S11 and S15 in FIG. 1, steps S21 and S22 in FIG. 2 are executed.

  In steps S21 and S22 of FIG. 2, steps S31 to S35 of FIG. 3 are executed while changing the best focus position and the DOF value. The best focus position and DOF are constants in the linear programming problem solved in step S32 of FIG. 3, that is, in the process of “linear optimization”. In the “linear optimization” step of step S32, the following linear programming problem is solved. That is, the linear programming problem of maximizing the merit function represented by the equation (a) under the constraint conditions represented by the equations (A) to (Q) is solved.

FIG. 4 is a diagram showing an ellipse that defines the size of the process window. In FIG. 4, the vertical axis indicates the light intensity I on the exposure surface of the wafer on which the pattern image is formed, and the horizontal axis indicates the defocus amount with respect to the Gaussian focus position on the exposure surface. Here, the Gaussian focus position is a focus position when the three-dimensional effect of the aberration of the projection optical system, the multilayer structure of the wafer, and the mask is not taken into consideration. Therefore, the defocus amount at the focus position (Gauss focus position) when the aberration of the projection optical system, the multilayer structure of the wafer, and the three-dimensional effect of the mask are not taken into consideration is zero, but the aberration of the projection optical system, The defocus amount ΔF with respect to the Gaussian focus position of the best focus position when considering the multilayer structure and the three-dimensional effect of the mask does not become zero.
In FIG. 4, the position in the defocus direction (horizontal axis direction in FIG. 4) where the width in the vertical direction (vertical axis direction indicating intensity I) of the ellipse representing the process window is maximized is indicated by the aberration of the projection optical system and the multilayer of the wafer. The best focus position when considering the three-dimensional effect of the film structure and mask. The defocus amount at the best focus position with respect to the Gaussian focus position is ΔF. In other words, a straight line defocus direction (horizontal axis direction in FIG. 4) that bisects the ellipse in the vertical direction (vertical axis direction indicating intensity I) through the center of gravity of the ellipse representing the process window in FIG. ) Is the best focus position in consideration of the aberration of the projection optical system, the multilayer structure of the wafer, and the three-dimensional effect of the mask. What is to be maximized in the linear programming problem in step S32 is the process window. That is, the process window defined by the area of the ellipse in FIG. 4 is maximized. The best focus position and DOF are constants when the linear programming problem is solved in step S32. The best focus position and DOF are optimized (improved) in the algorithm of FIG. W ′ and w shown in FIG. 4 are optimization variables, which are two intensity values when the ellipse defining the size of the process window intersects the vertical axis. w ′ and w will be described in detail in the explanation of the constraint conditions (L), (M), and (N).

  Hereinafter, each constraint condition will be described. In the constraint (A), Si is the intensity value of each pixel i of the illumination pupil divided in a grid as shown in FIG. 5, and n is the total number of illumination pixels. How finely the illumination pupil is divided is user-defined (a matter determined by the user), and is a known constant at the time of linear optimization in step S32. SMax is defined as the maximum value of Si. Si (1 ≦ i ≦ n) and SMax are optimization variables in the linear programming problem. By optimizing Si, the free-form pupil luminance distribution is optimized (improved). FIG. 5 shows an example in which the illumination pupil is represented by 33 pixels vertically and 33 pixels horizontally, but in many cases, the expression is expressed by 65 pixels vertically by 65 pixels horizontally or 129 pixels vertically and 129 pixels larger than this. Significantly better results can be obtained with a single pixel representation. Thus, in step S32, a pupil luminance distribution represented by a plurality of luminance elements Si distributed two-dimensionally along the illumination pupil is used.

  When the aberration is small enough to be ignored, if the pupil luminance distribution has asymmetry in the x and y directions, it will lead to image distortion during defocusing. Usually, since the aberration is sufficiently small, Si representing the pupil luminance distribution should be symmetric with respect to the x axis and the y axis. In this case, the optimization variable in the linear programming problem is limited to Si in the first quadrant pixel of the illumination pupil. When actually realizing a pupil luminance distribution expressed by a plurality of grid-divided illumination pixels, each illumination pixel 61 has a significant blur 62 as shown in FIG. The contribution each pixel has to the light intensity distribution on the wafer must be calculated taking into account the blur of each illumination pixel. The performance of the optimal solution obtained when calculating the contribution of each illumination pixel without considering blur is different from the performance when actually realized. The blur is caused by hardware, and an actually realizable pupil luminance distribution can be obtained by considering blur specific to hardware.

In the constraint (B), PFR (Pupil
Fill Ratio) is a user-defined value. By the constraint (B), when the sum ΣSi (i = 1 to n) when all Si is equal to SMax is set to 1, ΣSi (i = 1 to n) has a value at least by the ratio of PFR. To do. For example, a value of about 0.05 or more is entered in PFR. If the PFR is too small, the location used in the projection optical system is localized, and thus there is an effect that the influence of thermal aberration is likely to occur.

  In the constraint conditions (C), (D), (E), and (F), a constraint is made so that a difference from a user-defined reference pupil luminance distribution (for example, a designed pupil luminance distribution) does not increase. The intensity value of the reference pupil luminance distribution at each illumination pixel i is ri. In order to obtain the absolute value di of the difference between the intensity values Si and ri of the optimum pupil luminance distribution, the constraint conditions (C) and (D) are used. nref is the number of pixels having a significant intensity of the reference pupil luminance distribution among n illumination pixels, and nref ≦ n.

  The amount of difference from the reference pupil luminance distribution is constrained by the constraint conditions (E) and (F). If ri is normalized so that the maximum value is 1, the sum of the intensity of the reference pupil luminance distribution is SMaxΣri (i = 1 to nref). In the constraint condition (E), the ratio of the sum Σdi (i = 1 to nref) of the absolute value di of the intensity difference to the SMaxΣri (i = 1 to nref) is constrained by a user-defined constant R. However, the restriction condition (E) is monitored only for nref pixels where ri has a significant value, and restricts the upper limit of the amount to be reduced from the reference pupil luminance distribution. By combining with the constraint condition (F), the upper limit of the amount increased from the reference pupil luminance distribution is also limited. If nref = n, the constraint condition (F) is unnecessary and is deleted. If it is not necessary to constrain the difference from the reference pupil luminance distribution, the constraints (C), (D), (E), and (F) are deleted from the linear programming problem. Thus, in step S32, an improved pupil luminance distribution is obtained from, for example, a user-defined standard pupil luminance distribution.

  In the constraint condition (G), the anchor gauge strength value (threshold strength: threshold) on the wafer is normalized to 1. From the incoherent imaging theory, the light intensity I (χ, Bias) on the wafer is expressed by the following equation (b) by the linear sum of the illumination pixel intensity Si. In equation (b), Ci (χ, Defocus, Bias) is a position χ (vector (x, y)) on the wafer that is displaced by the defocus amount Defocus when each illumination pixel i has a certain unit intensity. The light intensity given through diffraction by the mask pattern thickened by a certain amount Bias is shown. That is, the Bias amount is, for example, an adjustment amount with respect to the design line width of the mask pattern.

  In the constraint conditions (H), (I), (J), and (K), the intensity values of points on the wafer sampled discretely are constrained, and a target structure (Target structure) specified by the user is realized globally. So that FIG. 7 is a diagram for explaining the positions where the intensity values are restricted by the constraint conditions (H), (I), (J), and (K). In FIG. 7, the rectangular portion shown in gray is the target structure (the structure of the pattern image that the user wants to make). The resist structure obtained as a result of exposure is approximately obtained as shown by a thick black line in FIG. 7 by binarizing the light intensity distribution on the wafer with a threshold value (threshold intensity).

  As shown in FIG. 7, Edge_Dark point and Edge_Bright point are taken across the edge of the target structure. Since the threshold value is 1 due to the constraint condition (G), a position of intensity 1 exists between the Edge_Dark point and the Edge_Bright point in FIG. 7 according to the constraint conditions (H) and (I). The edge position of the structure can be located between the Edge_Dark point and the Edge_Bright point. The distance between the Edge_Dark point and the Edge_Bright point is a user-defined constant. When the distance between the Edge_Dark point and the Edge_Bright point is reduced, the resist structure can be brought closer to the target structure.

  The constraint conditions (J) and (K) are constraint conditions for allowing the resist structure to have a normal intensity value even at a position away from the edge. Bright points are taken at positions where the intensity value should be greater than 1 (positions that should be brightened), dark points are taken at positions where the intensity value should be smaller than 1 (positions that should be darkened), and intensity values at each position Are constrained by user-specified values IBright (> 1) and IDark (<1). The intensity of Bright and Dark points is also restricted on the defocus plane. Discrete defocus surfaces below DOF in which the elliptical process window of FIG. 4 is defined, that is, the defocus surface Defocus represented by the following equation (c) (where Δz is a user-defined value, and the wavelength of illumination light λ = In the case of 193 nm and the image-side numerical aperture NA = 1.35 of the projection optical system, it is recommended that Δz be about 5 nm or less, but if Δz is small, many defocus surfaces need to be evaluated. In order to reduce the calculation load, it may be effective to set Δz to a large value such as 20 nm).

  4 is constrained by the constraint conditions (L), (M), and (N). The upper half of the ellipse in FIG. 4 is represented by the following equation (d), and the lower half is represented by the following equation (e). It is constrained by a discrete defocus surface Defocus = kdefΔz + ΔF. From the definition of the process window, in the inner region of the ellipse in FIG. 4, the CD value (line width value of the resist pattern image) on each defocus plane has a user-specified tolerance ± ΔCD (the tolerance of the line width of the resist pattern image: It is an intensity value that falls within a range of (allowable OPE error).

  As shown in FIG. 8, the PW_Bright point is taken on the bright side and the PW_Dark point on the dark side at a position shifted from the CD (Target CD) by ± ΔCD. The intensity values of all the PW_Bright points are above the upper half of the ellipse represented by the equation (d) in each discrete defocus plane, and the intensity values of all the PW_Dark points are represented by the equation (e) in each discrete defocus plane. It is necessary to be lower than the lower half of the ellipse represented by the constraints (L), (M), and (N) to guarantee this.

  The process window optimization method (improvement method) based on the merit function and the constraint conditions (L), (M), and (N) is the same as long as the shape of the figure for evaluating the process window is determined. The shape of the figure to be evaluated may be another shape such as a rectangle or a rhombus. For example, when a figure for evaluating a process window is represented by a rectangle, the merit function represented by the expression (a) and the constraint conditions (L) and (M) are represented by the following expression (a ′). What is necessary is just to replace with the constraint conditions represented by a function, Formula (L '), and Formula (M'). At this time, w ″ is unnecessary, so the constraint condition (N) is deleted.

The MEEF is constrained by the constraint condition (O). D. Ziger, D., Proc.
of SPIE 7122, 71223W (2008), MEEF can be approximately expressed by the following equation (f). However, CD_Biased and CD_Default in equation (f) correspond to the light intensity distribution by the default mask (designed mask) and the light intensity distribution by the biased mask in FIG. 9, respectively. I (χ, Bias), I (χ, 0), and I (χ + Bias, Bias) are intensity values of points indicated by reference numerals 63, 64, and 65 in FIG. 9, respectively. In Expression (f), the bias amount of the mask (converted value in the wafer dimension) and the minute movement amount of the position on the wafer (the difference in position between the points 63 and 64 and the point 65) are assumed to be the same Bias. .

  If the user-specified allowable maximum value MEEFMax is specified in MEEF, the following inequality (g) should hold. Here, as shown in FIG. 9, if the Bias at the position in the CD value evaluation direction is always set so that the light intensity decreases, and the mask bias is set so that the intensity on the wafer increases, the inequality (g) The absolute value in can be removed and the following inequality (h) is obtained. Since the denominator of inequality (h) is positive, multiplying both sides by {I (χ, Bias) −I (χ + Bias, Bias)} yields the following inequality (i). Thus, the MEEF constraint becomes a linear inequality with respect to the strength, and the MEEF can be constrained to a user-specified allowable value MEEFMax or less by the constraint (O).

 The OPE error is constrained by the constraint conditions (P) and (Q). FIG. 10 is a diagram for explaining evaluation points OPE_Bright points and OPE_Dark points for optimizing the OPE error. The algorithm for suppressing the OPE error, that is, the CD error to a value specified by the user or less is as follows. Let ΔCDtarget be the maximum allowable value specified by the user for the CD error from the Target CD. As shown in FIG. 10, the OPE_Bright point and the OPE_Dark point are taken at positions separated by ΔCDtarget / 2 in opposite directions in the CD measurement direction from the end point defining Target CD.

  When the constraint condition (P) that the intensity of the OPE_Bright point is larger than the threshold (threshold intensity) is imposed and the constraint condition (Q) that the intensity of the OPE_Dark point is smaller than the threshold is imposed, the threshold ( = 1) is always between the OPE_Bright point and the OPE_Dark point. In other words, the edge of the structure obtained by binarization using the threshold value is located between the OPE_Bright point and the OPE_Dark point. Since the error between the two edge positions defining the CD value can be made equal to or less than ΔCDtarget / 2, the error between the CD values defined at the two edge positions can be suppressed to less than ΔCDtarget. From the above, the OPE error can be optimized (improved) by the constraint conditions (P) and (Q).

  As described above, in step S32, a process window is formed using a predetermined pattern arranged on the first surface and a pupil luminance distribution represented by a plurality of luminance elements Si distributed two-dimensionally along the illumination pupil. calculate. Then, with the allowable MEEF and the allowable OPE error as constraints, the pupil luminance distribution so that the area of the ellipse representing the process window approaches the maximum value (generally, the process window becomes equal to or larger than a predetermined value). To improve.

  In the linear programming problem in step S32, MEEF is approximately constrained. If there is no position shift of the evaluation pattern and the Bias amount is sufficiently small, the approximation accuracy of MEEF is good, but if not, it is necessary to correct the MEEF by exact calculation. FIG. 3 shows a loop for performing MEEF correction by the exact calculation. In the linear programming problem in step S32, MEEFMax = MEEFTarget is set to an initial value (step S31). Here, MEEFTarget is an allowable maximum value designated by the user.

  The MEEF is strictly calculated using the pupil luminance distribution obtained by solving the linear programming problem in step S32, and the value obtained by the exact calculation is substituted into MEEFcurrent (step S33). Then, it is determined whether or not the absolute value abs (MEEFTarget−MEEFcurrent) of the difference between MEEFTarget and MEEFcurrent is sufficiently small (step S34). If abs (MEEFTarget-MEEFcurrent) is sufficiently small, the process ends. If abs (MEEFTarget-MEEFcurrent) is not sufficiently small, MEEFMax is adjusted (step S35), and the process of solving the linear programming problem in step S32 is repeated. In general, the smaller the MEEFMax, the smaller the MEEFcurrent. Therefore, it is only necessary to make the MEEFMax small enough to realize the MEEFTarget.

  As described above, in steps S32 to S35, the MEEF is calculated based on the pupil luminance distribution improved (set) in step S32, and whether or not the difference between the calculated MEEF and the allowable MEEF is within the allowable range. Determine whether. Then, when the difference between the calculated MEEF and the allowable MEEF is not within the allowable range, the pupil luminance distribution is improved until the difference between the calculated MEEF and the allowable MEEF is within the allowable range. Is repeated to improve MEEF.

  In steps S31 to S35, the best focus position and the DOF are constants. In the “optimization of the best focus position and the DOF” in FIG. 2, the best focus position and the DOF are optimized. In consideration of the aberration of the projection optical system, the multilayer structure of the wafer, and the three-dimensional effect of the mask, the process window can be increased by changing the best focus position. If the required DOF is determined by the convenience of the user or the device, the DOF may be fixed and only the best focus position may be optimized. FIG. 2 shows that the best focus position and the DOF should be optimized in two stages. In steps S21 and S22 of FIG. 2, the algorithms of steps S31 to S35 are repeatedly executed while changing the best focus position and the DOF, respectively. Specifically, in step S21, the best focus position and the fixed pitch (for example, 2 nm, such as 2 nm, a value that is fine enough that two or more local optimum combinations of the best focus position and the DOF cannot exist). While changing the DOF, the best focus position and the DOF value when the merit function value (and hence the value of the process window) is maximized are obtained.

  In step S22, the best focus position and DOF value at which the merit function becomes larger are obtained with higher accuracy in the vicinity of the best focus position and DOF obtained in step S21. Step S22 ends when the value of the merit function has sufficiently converged at the maximum value. It can be said that the best focus position and the DOF obtained by performing the two-dimensional grid search at a sufficiently fine pitch in step S21 and optimizing the best focus position and the DOF are globally optimum values. Thus, in steps S21 and S22, it is repeated to improve the MEEF for a plurality of different best focus positions and DOFs, and the improved best focus when the area of the ellipse representing the process window approaches the maximum value. Determining position and DOF values.

  In order to express the light intensity I on the wafer as a linear sum of the illumination pixel intensity Si as shown in the above formula (b), the evaluation position (x, y) on the wafer must be fixed. Therefore, in the linear programming problem in step S32, the evaluation position is fixed, and the evaluation position is adjusted in the outer loop. FIG. 1 shows the evaluation position adjustment loop. In the best focus position and DOF optimization in step S11, the algorithms in steps S21 and S22 are executed. Next, using the pupil luminance distribution obtained by the optimization in steps S21 and S22, a position shift is calculated within the DOF range centered on the optimized best focus position (step S12).

  In the calculation of the position shift, the difference between the two points defining the CD evaluated in the linear programming problem in step S32 and the two points defining the CD when the CD is actually obtained is examined (step S13). When the deviation is sufficiently small (for example, about 0.01 nm or less), it is not necessary to adjust the evaluation position, so that the loop ends. If the deviation is not sufficiently small, optimization is performed to maximize the value of the process window using the evaluation position as an optimization variable (steps S14 to S16). The evaluation position is adjusted as appropriate, and an evaluation position where the value of the process window is maximized is searched. When the value of the process window converges to the maximum value, the loop ends.

  One method of adjusting the evaluation position is to use the evaluation position shift amount for each defocus plane calculated using the optimal pupil luminance distribution until immediately before as the adjustment amount. Since the linear programming problem can obtain a global optimum solution for a specific evaluation position (which may be different for each defocus plane), this method always improves the pupil luminance distribution which is better than the optimum pupil luminance distribution until immediately before. Can be obtained. Here, it is used that the optimum best focus position and the optimum DOF derived by optimizing the best focus position and the DOF through the two-dimensional grid search as in steps S21 and S22 are also globally optimum. If the user does not want to shift the evaluation position much, the adjustment amount of the evaluation position may be limited.

  As described above, in steps S11 to S16, the shift amount from the current evaluation position of the pattern image formed on the wafer is obtained by using the improved (set) best focus position and the pupil luminance distribution corresponding to the DOF. . It is determined whether or not the obtained shift amount is larger than a predetermined value. Then, when the shift amount is larger than the predetermined value, the evaluation position is adjusted so that the area of the ellipse representing the process window converges toward the maximum value. When the evaluation position is improved, the shift amount calculated according to the plurality of defocus positions on the wafer surface is used as the adjustment amount for each defocus of the evaluation position.

  In application to an actual semiconductor exposure apparatus, it is preferable to be robust against individual differences of apparatuses and short-term fluctuations of apparatus parameters. In order to improve the robustness, individual differences between devices and short-term fluctuations in device parameters may be introduced in the linear programming problem in step S32. Specific differences in equipment and short-term fluctuations in equipment parameters include specific differences in light wavelength, projection optical system aberration and transmittance, MSD (moving standard deviation), image height, etc. This is a change in Ci (χ, Defocus, Bias) in equation (b) for calculating the light intensity I on the wafer.

  Constraints (G) to (M) and (O) to (Q) using Ci (χ, Defocus, Bias) into which an amount that can be considered as an individual difference between devices and a change in device parameters is introduced into the linear program in step S32. By adding to the problem, it is possible to maximize the common process window in consideration of such changes and obtain a pupil luminance distribution that satisfies the constraint condition even under the change, that is, a pupil luminance distribution that is robust to the change. As described above, by using a constraint condition that introduces individual differences of apparatuses including the imaging optical system and the illumination optical system and / or fluctuations of apparatus parameters as a constraint condition regarding the allowable OPE error and the allowable MEEF. Thus, it is possible to obtain a pupil luminance distribution with improved robustness.

Although the expression (b) has been described as “light intensity on the wafer”, it may be expressed by a linear sum of the illumination pixel intensity Si instead of the light intensity on the wafer. While it is obvious that light intensity in water or in resist can be used, the intensity in resist is multiplied by a linear factor as in P. Liu et atl., Proc. Of SPIE 8326, 83260A (2012). PAG (photo-acid)
generator) distribution. In addition, P. Liu et
atl., Proc. of SPIE 8326, 83260A (2012)
The Gaussian convolution approximates the acid diffusion effect, but this may be used.

  The linear programming problem in the above algorithm may be extended to a convex optimization problem (convex optimization problem). Simply squaring the current merit function becomes a convex optimization problem. Further, by extending to the convex optimization problem, functions up to the second order can be introduced into the merit function and constraint conditions as necessary. That is, the pupil luminance distribution may be set using the convex programming method so that the process window becomes a predetermined value or more using the allowable MEEF as a constraint.

  FIG. 11 is a diagram showing a pupil luminance distribution obtained by applying the pupil luminance distribution setting method according to the present embodiment. In the numerical example shown below, the degree of focus on each index is adjusted, and the default annular luminance distribution (standard annular luminance distribution) 66a shown in FIG. ) To (e), improved pupil luminance distributions 66b to 66e were obtained. Table (1) shows the performance of each of the pupil luminance distributions 66a to 66e shown in FIGS. In the numerical example, the pupil luminance distribution was optimized (improved) for 45 one-dimensional line and space patterns with a pitch of 90 nm to 700 nm (converted on the wafer).

  Among the 45 patterns, pattern numbers 1 to 13 are patterns whose target line width is 45 nm converted on the wafer and whose pitch changes in the range of 90 nm to 150 nm converted on the wafer. Pattern numbers 14 to 29 are patterns in which the target line width is 75 nm as converted on the wafer and the pitch changes in the range of 150 nm to 300 nm as converted on the wafer. Pattern numbers 30 to 45 are patterns in which the target line width is 100 nm as converted on the wafer and the pitch changes in the range of 300 nm to 700 nm as converted on the wafer. In order to optimize the periodic line in the x direction, the pattern was irradiated with linearly polarized light in the y direction.

Table (1)
Process window MEEF Maximum OPE error
(% ・ Μm) (nm)
Pupil luminance distribution 66a 0.93 2.8 −
Pupil luminance distribution 66b 1.44 2.6 5.6
Pupil luminance distribution 66c 0.99 2.0 6.0
Pupil luminance distribution 66d 1.17 2.7 0.1
Pupil luminance distribution 66e 1.21 2.5 0.5

  Using the default pupil luminance distribution 66a shown in FIG. 11A, OPC processing (Biasing of line width) was performed on the mask pattern. A pupil luminance distribution 66b shown in FIG. 11B is a pupil luminance distribution obtained by ignoring MEEF and OPE errors and focusing only on process window optimization. The MEEF improvement over the default despite ignoring the MEEF may be due to the intensity contrast being improved by maximizing the process window.

  A pupil luminance distribution 66c shown in FIG. 11C is a pupil luminance distribution obtained by setting the user-specified maximum value 2 to MEEF. A pupil luminance distribution 66d shown in FIG. 11D is a pupil luminance distribution obtained by setting a user-specified maximum value of 0.1 nm to the OPE error. A pupil luminance distribution 66e shown in FIG. 11E is a pupil luminance distribution obtained by setting the MEEF to 2.5 while suppressing the OPE error to 0.5 nm or less, and improving the process window and the MEEF. This is a balanced pupil luminance distribution so that the OPE error does not become too large. FIG. 12 is a diagram showing OPE errors corresponding to the pupil luminance distributions of FIG. In FIG. 12, the horizontal axis indicates pattern numbers 1 to 45, and the vertical axis indicates an OPE error (CD error: nm). In the pupil luminance distributions 66d and 66e constrained by OPE, the OPE error could be suppressed as small as specified. As described above, in the pupil luminance distribution setting method according to the present embodiment, the pupil luminance distribution can be set so that the process window becomes equal to or greater than a predetermined value while keeping MEEF small.

The pupil luminance distribution setting method according to the present embodiment is the stage in which the pupil luminance distribution and the mask pattern shape are optimized on the computer from the target pattern, and the pupil luminance distribution and the mask pattern are actually created and intended. It is used in the stage of correcting the pupil luminance distribution so that an exposure result is obtained. At the stage of optimizing the pupil luminance distribution and the mask pattern shape from the target pattern, SMO (Source and Mask Optimization: For example, T. Muelders et al.
al., Proc. of SPIE 8326, 83260G (2012)), etc., and the pupil luminance distribution is optimized using the aerial image. When optimizing the pupil luminance distribution, only the pupil luminance distribution is optimized (improved) by the setting method according to the present embodiment without changing the mask pattern shape. In the setting method according to the present embodiment, high-speed optimization is possible by utilizing a linear programming problem. As a result, it is possible to obtain a higher optimization result by increasing the number of illumination pixels that are optimization variables, and thus to obtain a pupil luminance distribution that leads to a better exposure result.

  At the stage of correcting the pupil luminance distribution so that the intended exposure result is obtained, the exposure result (target value) based on the mask pattern and pupil luminance distribution derived by calculation, and the exposure result based on the actually created mask pattern and pupil luminance distribution To correct the deviation. In general, when correcting the deviation, a mask or pupil luminance distribution is recreated so that a desired result is obtained. However, by using a technique for actively adjusting the free-form pupil luminance distribution as described in, for example, US Patent Publication No. 2011/0069305, the pupil luminance distribution is cheaper and quicker than remaking the mask. It is possible to adjust the exposure result.

At this time, how to adjust the pupil luminance distribution is derived using the method for setting the pupil luminance distribution according to the present embodiment. The evaluation position is set so as to bridge the difference between the experiment and the calculation. If the line width CDcalc is obtained on the computer, the line width CDexp is obtained in the exposure result, and the desired line width value is CDtarget, the converted CDtarget as shown in the following equation (j) is obtained on the computer. If the pupil luminance distribution as obtained in (1) is derived using the setting method of this embodiment, it is expected that a desired line width CDtarget can be obtained in the exposure result.
Conversion CDtarget = CDtarget × (CDexp / CDcalc) (j)

  FIG. 13 is a drawing schematically showing a configuration of an exposure apparatus according to the embodiment. In FIG. 13, the Z axis is along the normal direction of the transfer surface (exposure surface) of the wafer W, which is a photosensitive substrate, and the X axis is along the direction parallel to the paper surface of FIG. In the transfer surface of the wafer W, the Y axis is set along the direction perpendicular to the paper surface of FIG.

  Referring to FIG. 13, exposure light (illumination light) is supplied from the light source 1 to the exposure apparatus of the present embodiment. As the light source 1, for example, an ArF excimer laser light source that supplies light with a wavelength of 193 nm, a KrF excimer laser light source that supplies light with a wavelength of 248 nm, or the like can be used. The exposure apparatus of this embodiment supports an illumination optical system IL including a spatial light modulation unit 3, a mask stage MS that supports a mask M, a projection optical system PL, and a wafer W along the optical axis AX of the apparatus. Wafer stage WS.

  The light from the light source 1 illuminates the mask M through the illumination optical system IL. The light transmitted through the mask M forms an image of the pattern of the mask M on the wafer W via the projection optical system PL. The illumination optical system IL that illuminates the pattern surface (illuminated surface) of the mask M based on the light from the light source 1 is a multipolar illumination (bipolar illumination, quadrupole illumination, etc.) Deformation illumination such as annular illumination, normal circular illumination, etc. are performed. Further, modified illumination based on a pupil luminance distribution having a complicated outer shape and distribution is performed according to the pattern characteristics of the mask M.

  The illumination optical system IL includes, in order from the light source 1 side along the optical axis AX, a beam transmission unit 2, a spatial light modulation unit 3, a relay optical system 4, and a micro fly's eye lens (or fly eye lens) 5. A condenser optical system 6, an illumination field stop (mask blind) 7, and an imaging optical system 8. The spatial light modulation unit 3 forms a desired light intensity distribution (pupil luminance distribution) in the far field region (Fraunhofer diffraction region) based on the light from the light source 1 via the beam transmitter 2. The internal configuration and operation of the spatial light modulation unit 3 will be described later.

  The beam transmitter 2 guides the incident light beam from the light source 1 to the spatial light modulation unit 3 while converting the incident light beam into a light beam having an appropriate size and shape, and changes the position of the light beam incident on the spatial light modulation unit 3. And a function of actively correcting the angular variation. The relay optical system 4 condenses the light from the spatial light modulation unit 3 and guides it to the micro fly's eye lens 5. The micro fly's eye lens 5 is, for example, an optical element composed of a large number of micro lenses having positive refractive power arranged vertically and horizontally and densely. The micro fly's eye lens 5 is formed by etching a parallel plane plate to form a micro lens group. Has been.

  In a micro fly's eye lens, unlike a fly eye lens composed of lens elements isolated from each other, a large number of micro lenses (micro refractive surfaces) are integrally formed without being isolated from each other. However, the micro fly's eye lens is the same wavefront division type optical integrator as the fly's eye lens in that the lens elements are arranged vertically and horizontally. A rectangular minute refracting surface as a unit wavefront dividing surface in the micro fly's eye lens 5 is a rectangular shape similar to the shape of the illumination field to be formed on the mask M (and the shape of the exposure region to be formed on the wafer W). It is.

  The micro fly's eye lens 5 divides the incident light beam into a wavefront to form a secondary light source (substantial surface light source; pupil luminance distribution) composed of a large number of small light sources at the illumination pupil at or near the rear focal position. . The incident surface of the micro fly's eye lens 5 is disposed at or near the rear focal position of the relay optical system 4. For example, a cylindrical micro fly's eye lens can be used as the micro fly's eye lens 5. The configuration and operation of the cylindrical micro fly's eye lens are disclosed in, for example, US Pat. No. 6,913,373.

  In this embodiment, the secondary light source formed by the micro fly's eye lens 5 is used as a light source, and the mask M arranged on the irradiated surface of the illumination optical system IL is Koehler illuminated. For this reason, the position where the secondary light source is formed is optically conjugate with the position of the aperture stop AS of the projection optical system PL, and the formation surface of the secondary light source can be called the illumination pupil plane of the illumination optical system IL. . Typically, the irradiated surface (the surface on which the mask M is disposed or the surface on which the wafer W is disposed when the illumination optical system including the projection optical system PL is considered) is optical with respect to the illumination pupil plane. A Fourier transform plane.

  The pupil luminance distribution is a light intensity distribution on the illumination pupil plane of the illumination optical system IL or a plane optically conjugate with the illumination pupil plane. When the number of wavefront divisions by the micro fly's eye lens 5 is relatively large, the overall light intensity distribution formed on the incident surface of the micro fly's eye lens 5 and the overall light intensity distribution (pupil luminance distribution) of the entire secondary light source. ) And a high correlation. For this reason, the light intensity distribution on the incident surface of the micro fly's eye lens 5 and the surface optically conjugate with the incident surface can also be referred to as a pupil luminance distribution.

  The condenser optical system 6 condenses the light emitted from the micro fly's eye lens 5 and illuminates the illumination field stop 7 in a superimposed manner. The light that has passed through the illumination field stop 7 forms an illumination region that is an image of the opening of the illumination field stop 7 in at least a part of the pattern formation region of the mask M via the imaging optical system 8. In FIG. 13, the installation of the optical path bending mirror for bending the optical axis (and thus the optical path) is omitted, but the optical path bending mirror can be appropriately disposed in the illumination optical path as necessary. .

  A mask M is placed on the mask stage MS along the XY plane (for example, a horizontal plane), and a wafer W is placed on the wafer stage WS along the XY plane. The projection optical system PL forms an image of the pattern of the mask M on the transfer surface (exposure surface) of the wafer W based on the light from the illumination area formed on the pattern surface of the mask M by the illumination optical system IL. . In this way, batch exposure or scan exposure is performed while the wafer stage WS is two-dimensionally driven and controlled in a plane (XY plane) orthogonal to the optical axis AX of the projection optical system PL, and thus the wafer W is two-dimensionally driven and controlled. As a result, the pattern of the mask M is sequentially exposed in each exposure region of the wafer W.

  With reference to FIGS. 14 and 15, the internal configuration and operation of the spatial light modulation unit 3 will be described. As shown in FIG. 14, the spatial light modulation unit 3 includes a prism 32 and a spatial light modulator 30 disposed close to a side surface 32 a parallel to the YZ plane of the prism 32. The prism 32 is made of an optical material such as fluorite or quartz.

  The spatial light modulator 30 includes, for example, a plurality of mirror elements 30a that are two-dimensionally arranged along the YZ plane, a base 30b that holds the plurality of mirror elements 30a, and a cable (not shown) connected to the base 30b. And a drive unit 30c for individually controlling and driving the postures of the plurality of mirror elements 30a. In the spatial light modulator 30, the attitude of the plurality of mirror elements 30a is changed by the action of the drive unit 30c that operates based on the control signal from the control unit CR, and each mirror element 30a is set in a predetermined direction. The

  The prism 32 is obtained by replacing one side surface of the rectangular parallelepiped (the side surface facing the side surface 32a where the plurality of mirror elements 30a of the spatial light modulator 30 are arranged close to each other) with side surfaces 32b and 32c that are recessed in a V shape. It has an obtained form and is also called a K prism because of its cross-sectional shape along the XZ plane. Side surfaces 32b and 32c of the prism 32 that are recessed in a V shape are defined by two planes P1 and P2 that intersect to form an obtuse angle. The two planes P1 and P2 are both orthogonal to the XZ plane and have a V shape along the XZ plane.

  The inner surfaces of the two side surfaces 32b and 32c that are in contact with a tangent line (straight line extending in the Y direction) P3 between the two planes P1 and P2 function as the reflection surfaces R1 and R2. That is, the reflective surface R1 is located on the plane P1, the reflective surface R2 is located on the plane P2, and the angle formed by the reflective surfaces R1 and R2 is an obtuse angle. As an example, the angle between the reflection surfaces R1 and R2 is 120 degrees, the angle between the incident surface IP of the prism 32 perpendicular to the optical axis AX and the reflection surface R1 is 60 degrees, and the prism 32 perpendicular to the optical axis AX. The angle formed by the exit surface OP and the reflective surface R2 can be 60 degrees.

  In the prism 32, the side surface 32a on which the plurality of mirror elements 30a of the spatial light modulator 30 are arranged close to each other and the optical axis AX are parallel, and the reflection surface R1 is on the light source 1 side (upstream side of the exposure apparatus: FIG. 14 on the left side), the reflecting surface R2 is positioned on the micro fly's eye lens 5 side (downstream side of the exposure apparatus: right side in FIG. 14). More specifically, the reflecting surface R1 is obliquely arranged with respect to the optical axis AX, and the reflecting surface R2 is obliquely inclined with respect to the optical axis AX symmetrically with respect to the reflecting surface R1 with respect to a plane passing through the tangent line P3 and parallel to the XY plane. It is installed.

  The reflecting surface R1 of the prism 32 reflects the light incident through the incident surface IP toward the spatial light modulator 30. The plurality of mirror elements 30a of the spatial light modulator 30 are arranged in the optical path between the reflecting surface R1 and the reflecting surface R2, and reflect the light incident through the reflecting surface R1. The reflecting surface R2 of the prism 32 reflects the light incident through the spatial light modulator 30 and guides it to the relay optical system 4 through the exit surface OP.

  The spatial light modulator 30 applies the spatial modulation according to the incident position to the light incident through the reflecting surface R1 and emits the light. As shown in FIG. 15, the spatial light modulator 30 includes a plurality of minute mirror elements (optical elements) 30a arranged two-dimensionally within a predetermined plane. For ease of explanation and illustration, FIGS. 14 and 15 show a configuration example in which the spatial light modulator 30 includes 4 × 4 = 16 mirror elements 30a. The number of mirror elements 30a is typically large, typically about 4000 to 100,000.

  Referring to FIG. 14, among the light beams incident on the spatial light modulation unit 3 along the direction parallel to the optical axis AX, the light beam L1 is the mirror element SEa of the plurality of mirror elements 30a, and the light beam L2 is the mirror element. The light is incident on a mirror element SEb different from SEa. Similarly, the light beam L3 is incident on a mirror element SEc different from the mirror elements SEa and SEb, and the light beam L4 is incident on a mirror element SEd different from the mirror elements SEa to SEc. The mirror elements SEa to SEd give spatial modulations set according to their positions to the lights L1 to L4.

  The array surface of the plurality of mirror elements 30 a of the spatial light modulator 30 is disposed at or near the front focal position of the relay optical system 4. The light reflected by the plurality of mirror elements 30a of the spatial light modulator 30 and given a predetermined angular distribution forms predetermined light intensity distributions SP1 to SP4 on the rear focal plane 4a of the relay optical system 4. That is, the relay optical system 4 determines the angle that the plurality of mirror elements 30a of the spatial light modulator 30 gives to the emitted light on the surface 4a that is the far field region (Fraunhofer diffraction region) of the spatial light modulator 30. Has been converted.

  Referring again to FIG. 15, the incident surface of the micro fly's eye lens 5 is positioned at the rear focal plane 4 a of the relay optical system 4. Therefore, the pupil luminance distribution formed in the illumination pupil immediately after the micro fly's eye lens 5 is the light intensity distributions SP1 to SP4 formed on the incident surface of the micro fly's eye lens 5 by the spatial light modulator 30 and the relay optical system 4. Corresponding distribution. As shown in FIG. 15, the spatial light modulator 30 includes, for example, a large number of minute mirror elements 30 a that are regularly and two-dimensionally arranged along one plane with a planar reflecting surface as an upper surface. Is a movable multi-mirror.

  Each mirror element 30a is movable, and the tilt of the reflecting surface (that is, the tilt angle and tilt direction of the reflecting surface) is independently controlled by the action of the drive unit 30c that operates according to a command from the control unit CR. Each mirror element 30a can rotate continuously or discretely by a desired rotation angle with two directions (Y direction and Z direction) parallel to the arrangement plane and orthogonal to each other as rotation axes. That is, it is possible to two-dimensionally control the inclination of the reflecting surface of each mirror element 30a.

  When the reflection surface of each mirror element 30a is discretely rotated, the rotation angle is set in a plurality of states (for example,..., -2.5 degrees, -2.0 degrees, ... 0 degrees, +0.5 degrees). ... +2.5 degrees,. Although FIG. 15 shows a mirror element 30a having a square outer shape, the outer shape of the mirror element 30a is not limited to a square. However, from the viewpoint of light utilization efficiency, the shape can be arranged so as to reduce the gap between the mirror elements 30a (the shape that can be packed most closely). Further, from the viewpoint of light utilization efficiency, the interval between two adjacent mirror elements 30a can be minimized.

  In the spatial light modulator 30, the posture of the plurality of mirror elements 30a is changed by the action of the drive unit 30c that operates according to the control signal from the control unit CR, and each mirror element 30a is set in a predetermined direction. The The light reflected by the plurality of mirror elements 30a of the spatial light modulator 30 at a predetermined angle is applied to the illumination pupil immediately after the micro fly's eye lens 5 via the relay optical system 4 in a plurality of poles (bipolar). Light intensity distribution (pupil luminance distribution) such as a ring shape or a circular shape. Further, a pupil luminance distribution having a complicated outer shape and distribution is formed on the illumination pupil immediately after the micro fly's eye lens 5 in accordance with the pattern characteristics of the mask M.

  That is, the relay optical system 4 and the micro fly's eye lens 5 form a predetermined light intensity distribution in the illumination pupil of the illumination optical system IL based on the light that has passed through the spatial light modulator 30 in the spatial light modulation unit 3. A distribution forming optical system is configured. Another illumination pupil position optically conjugate with the rear focal position of the micro fly's eye lens 5 or in the vicinity thereof, that is, the pupil position of the imaging optical system 8 and the pupil position of the projection optical system PL (of the aperture stop AS) At the position), a pupil luminance distribution corresponding to the light intensity distribution in the illumination pupil immediately after the micro fly's eye lens 5 is also formed.

  The exposure apparatus of this embodiment includes a pupil distribution measurement device DT that is attached to, for example, a mask stage MS and measures a pupil luminance distribution formed on the illumination pupil of the illumination optical system IL. As shown in FIG. 16, the pupil distribution measurement device DT includes a pinhole member 10, a condenser lens 11, and a photodetector 12 such as a two-dimensional CCD image sensor. The pinhole member 10 is disposed on the irradiated surface of the illumination optical system IL (that is, the height position at which the pattern surface Pm of the mask M is to be positioned during exposure) during measurement. Further, the pinhole member 10 is disposed at the front focal position of the condenser lens 11, and the detection surface of the photodetector 12 is disposed at the rear focal position of the condenser lens 11.

  Therefore, the detection surface of the photodetector 12 is disposed at a position optically conjugate with the illumination pupil of the illumination optical system IL, that is, at a position optically conjugate with the pupil plane of the imaging optical system 8. In the pupil distribution measurement device DT, the light that has passed through the illumination optical system IL passes through the pinhole of the pinhole member 10, receives the condensing action of the condenser lens 11, and then reaches the detection surface of the photodetector 12. A light intensity distribution corresponding to the light intensity distribution (pupil luminance distribution) on the pupil plane of the imaging optical system 8 is formed on the detection surface of the photodetector 12.

  Thus, the pupil distribution measuring device DT is on a surface optically conjugate with the illumination pupil of the illumination optical system IL (the pupil plane of the imaging optical system 8) based on the light that has passed through the illuminated surface of the illumination optical system IL. Measure the light intensity distribution. Specifically, the pupil distribution measurement device DT measures a pupil luminance distribution (a pupil luminance distribution formed at the exit pupil position of the illumination optical system by the light incident on each point) for each point on the irradiated surface by the illumination optical system. To do.

  The control unit CR that comprehensively controls the operation of the exposure apparatus refers to the measurement result of the pupil distribution measurement device DT, so that a desired pupil luminance distribution is formed in the illumination pupil, so that the spatial light modulator 30 The plurality of mirror elements 30a are controlled. For a more detailed configuration and operation of the pupil distribution measuring device DT, reference can be made to, for example, Japanese Patent Laid-Open No. 2000-19012. Further, as the pupil distribution measuring device DT, for example, a device that detects the pupil luminance distribution through a pinhole disclosed in US Pat. No. 5,925,887 can be used.

  Further, instead of or in addition to the pupil distribution measurement device DT, based on light that has passed through the projection optical system PL (that is, based on light that has passed through the image plane of the projection optical system PL), A pupil distribution measurement device DTw (not shown) that measures the pupil luminance distribution on the pupil plane of the projection optical system PL (the exit pupil plane of the projection optical system PL) may be provided. Specifically, the pupil distribution measurement device DTw includes, for example, an imaging unit having a photoelectric conversion surface arranged at a position optically conjugate with the pupil position of the projection optical system PL, and each point on the image plane of the projection optical system PL. Is measured (a pupil luminance distribution formed at a pupil position of the projection optical system PL by light incident on each point). As for the detailed configuration and operation of these pupil distribution measuring apparatuses, reference can be made to, for example, US Patent Publication No. 2008/0030707. Further, as a pupil distribution measuring device, the disclosure of US Patent Publication No. 2010/0020302 can be referred to.

  In the exposure apparatus, in order to transfer the pattern of the mask M onto the wafer W with high accuracy and faithfulness, it is important to perform exposure under appropriate illumination conditions according to the pattern characteristics of the mask M, for example. In the present embodiment, the pupil luminance distribution formed by the action of the spatial light modulator 30 is calculated using the spatial light modulation unit 3 including the spatial light modulator 30 in which the postures of the plurality of mirror elements 30a individually change. It can be freely and quickly changed, and thus various illumination conditions can be realized. That is, since the illumination optical system IL uses the spatial light modulator 30 having a large number of mirror elements 30a whose postures are individually controlled, the external shape of the pupil luminance distribution formed on the illumination pupil and the change of the distribution are related. The degree of freedom is high.

  However, the pupil luminance distribution actually formed on the illumination pupil is slightly different from the designed pupil luminance distribution due to various causes, or optical characteristics other than the pupil luminance distribution satisfy the conditions for designing the pupil luminance distribution. It is difficult to achieve a desired imaging performance according to the designed pupil luminance distribution. In the exposure apparatus of the present embodiment, the pupil luminance distribution setting method according to the present embodiment should be formed on the illumination pupil of the illumination optical system IL so as to obtain an imaging performance sufficiently close to the desired imaging performance. The pupil luminance distribution is set, the pupil luminance distribution formed on the illumination pupil is adjusted using the set pupil luminance distribution as a target, and the illumination optical system IL is adjusted accordingly.

  FIG. 17 is a flowchart of the adjustment method of the illumination optical system according to the present embodiment. In the adjustment method of the illumination optical system IL according to the present embodiment, the pupil luminance distribution to be formed on the illumination pupil of the illumination optical system is set using the pupil luminance distribution setting method described with reference to FIGS. (Step S01). That is, step S01 includes steps S11 to S16 of FIG. 1, and improves (sets) the pupil luminance distribution so that the process window becomes equal to or greater than a predetermined value while keeping MEEF small.

  Next, in the method for adjusting the illumination optical system IL according to the present embodiment, the pupil luminance distribution formed on the illumination pupil is adjusted using the improved (set) pupil luminance distribution as a target (step S02). Specifically, in step S02, the pupil luminance distribution improved in step S01 is obtained by controlling the postures of the plurality of mirror elements 30a of the spatial light modulator 30 while referring to the measurement result of the pupil distribution measuring device DT. A pupil luminance distribution as close as possible is formed in the illumination pupil, and consequently the pupil luminance distribution formed in the illumination pupil of the illumination optical system IL is adjusted.

  In other words, in step S02, the pupil luminance distribution formed on the illumination pupil of the illumination optical system IL is appropriately adjusted with a desired imaging performance corresponding to the designed pupil luminance distribution as a target. Note that, after step S02, test exposure is performed based on the improved (set) pupil luminance distribution as necessary, and a resist pattern line formed on the wafer W placed on the image plane of the projection optical system PL. It may be confirmed by measuring the width that the OPE error is within an allowable range.

  As described above, in the adjustment method of the illumination optical system IL according to the present embodiment, desired image formation is performed by using the setting method for setting the pupil luminance distribution so that the process window becomes a predetermined value or more while suppressing the MEEF to be small. The pupil luminance distribution can be appropriately adjusted with the performance as a target. As a result, in the exposure apparatus (IL, MS, PL, WS) according to the present embodiment, the illumination optical system IL that appropriately adjusts the pupil luminance distribution with the desired imaging performance as a target is used, and an appropriate illumination condition is obtained. A device with good performance can be manufactured by performing good exposure with the above.

  In the above-described embodiment, the pupil distribution measurement device DT uses the illumination pupil (imaging) of the illumination optical system IL based on the light that has passed through the illuminated surface of the illumination optical system IL (the position of the pattern surface of the mask M). The light intensity distribution on a plane optically conjugate with the pupil plane of the optical system 8 is measured. However, the present invention is not limited to this, and the light intensity distribution on the surface optically conjugate with the illumination pupil can be measured based on the light traveling toward the irradiated surface of the illumination optical system IL. As an example, a part of the illumination light is extracted from the optical path between the micro fly's eye lens 5 and the condenser optical system 6, and the extracted light is optically conjugate with the illumination pupil immediately after the micro fly's eye lens 5. By detecting, the light intensity distribution corresponding to the pupil luminance distribution in the illumination pupil immediately after the micro fly's eye lens 5 is measured.

  Further, in the above-described embodiment, the K prism 32 integrally formed with one optical block is used as the prism member having the optical surface facing the arrangement surface of the plurality of mirror elements 30a of the spatial light modulator 30. Yes. However, the prism member having the same function as that of the K prism 32 can be configured by a pair of prisms without being limited thereto. In addition, a prism member having the same function as the K prism 32 can be configured by one plane-parallel plate and a pair of triangular prisms. Further, an assembly optical member having the same function as that of the K prism 32 can be constituted by one parallel plane plate and a pair of plane mirrors.

  In the present embodiment, as the spatial light modulator 30, for example, a spatial light modulator that continuously changes the directions of a plurality of mirror elements 30a arranged two-dimensionally is used. As such a spatial light modulator, for example, European Patent Publication No. 779530, US Pat. No. 5,867,302, US Pat. No. 6,480,320, US Pat. No. 6,600,591 U.S. Patent No. 6,733,144, U.S. Patent No. 6,900,915, U.S. Patent No. 7,095,546, U.S. Patent No. 7,295,726, U.S. Patent No. 7, No. 424,330, U.S. Pat. No. 7,567,375, U.S. Patent Publication No. 2008/0309901, U.S. Pat. Publication No. 2011/0181852 and U.S. Pat. Publication No. 2011/188017. A spatial light modulator can be used. Note that the orientations of the plurality of mirror elements 5a arranged two-dimensionally may be controlled so as to have a plurality of discrete stages.

  In the above-described embodiment, as the spatial light modulator having a plurality of mirror elements that are two-dimensionally arranged and individually controlled, the directions (angle: inclination) of the plurality of two-dimensionally arranged reflecting surfaces are individually set. The controllable spatial light modulator 30 is used. However, the present invention is not limited to this. For example, a spatial light modulator that can individually control the height (position) of a plurality of two-dimensionally arranged reflecting surfaces can be used. As such a spatial light modulator, for example, the spatial light modulator disclosed in FIG. 1d of US Pat. No. 5,312,513 and US Pat. No. 6,885,493 can be used. In these spatial light modulators, by forming a two-dimensional height distribution, an action similar to that of the diffractive surface can be given to incident light. The spatial light modulator having a plurality of reflection surfaces arranged two-dimensionally as described above is modified in accordance with the disclosure of, for example, US Pat. No. 6,891,655 and US Patent Publication No. 2005/0095749. May be.

  In the above-described embodiment, the pupil adjustment device that adjusts the pupil luminance distribution formed on the illumination pupil of the illumination optical system IL is a reflective type having a plurality of mirror elements 30a arranged in a predetermined plane and individually controllable. A spatial light modulator 30 is used. However, the present invention is not limited to this, and a transmissive spatial light modulator including a plurality of transmissive optical elements arranged in a predetermined plane and individually controlled can also be used.

  In the above-described embodiment, a variable pattern forming apparatus that forms a predetermined pattern based on predetermined electronic data can be used instead of a mask. As the variable pattern forming apparatus, for example, a spatial light modulation element including a plurality of reflection elements driven based on predetermined electronic data can be used. An exposure apparatus using a spatial light modulator is disclosed, for example, in US Patent Publication No. 2007/0296936. In addition to the non-light-emitting reflective spatial light modulator as described above, a transmissive spatial light modulator may be used, or a self-luminous image display element may be used.

  The exposure apparatus of the above-described embodiment is manufactured by assembling various subsystems including the respective constituent elements recited in the claims of the present application so as to maintain predetermined mechanical accuracy, electrical accuracy, and optical accuracy. Is done. In order to ensure these various accuracies, before and after assembly, various optical systems are adjusted to achieve optical accuracy, various mechanical systems are adjusted to achieve mechanical accuracy, and various electrical systems are Adjustments are made to achieve electrical accuracy. The assembly process from the various subsystems to the exposure apparatus includes mechanical connection, electrical circuit wiring connection, pneumatic circuit piping connection and the like between the various subsystems. Needless to say, there is an assembly process for each subsystem before the assembly process from the various subsystems to the exposure apparatus. When the assembly process of the various subsystems to the exposure apparatus is completed, comprehensive adjustment is performed to ensure various accuracies as the entire exposure apparatus. The exposure apparatus may be manufactured in a clean room where the temperature, cleanliness, etc. are controlled.

  Next, a device manufacturing method using the exposure apparatus according to the above-described embodiment will be described. FIG. 18 is a flowchart showing manufacturing steps of a semiconductor device. As shown in FIG. 18, in the semiconductor device manufacturing process, a metal film is vapor-deposited on a wafer W to be a semiconductor device substrate (step S40), and a photoresist, which is a photosensitive material, is applied on the vapor-deposited metal film. (Step S42). Subsequently, using the exposure apparatus of the above-described embodiment, the pattern formed on the mask (reticle) M is transferred to each shot area on the wafer W (step S44: exposure process), and the transfer of the wafer W after the transfer is completed. Development, that is, development of the photoresist with the transferred pattern is performed (step S46: development step).

  Thereafter, using the resist pattern generated on the surface of the wafer W in step S46 as a mask, processing such as etching is performed on the surface of the wafer W (step S48: processing step). Here, the resist pattern is a photoresist layer in which unevenness having a shape corresponding to the pattern transferred by the exposure apparatus of the above-described embodiment is generated, and the recess penetrates the photoresist layer. is there. In step S48, the surface of the wafer W is processed through this resist pattern. The processing performed in step S48 includes, for example, at least one of etching of the surface of the wafer W or film formation of a metal film or the like. In step S44, the exposure apparatus of the above-described embodiment performs pattern transfer using the wafer W coated with the photoresist as a photosensitive substrate.

  FIG. 19 is a flowchart showing manufacturing steps of a liquid crystal device such as a liquid crystal display element. As shown in FIG. 19, in the liquid crystal device manufacturing process, a pattern forming process (step S50), a color filter forming process (step S52), a cell assembling process (step S54), and a module assembling process (step S56) are sequentially performed. In the pattern forming process of step S50, a predetermined pattern such as a circuit pattern and an electrode pattern is formed on the glass substrate coated with a photoresist as the plate P using the exposure apparatus of the above-described embodiment. In this pattern formation process, an exposure process for transferring the pattern to the photoresist layer using the exposure apparatus of the above-described embodiment and development of the plate P to which the pattern is transferred, that is, development of the photoresist layer on the glass substrate are performed. And a developing step for generating a photoresist layer having a shape corresponding to the pattern, and a processing step for processing the surface of the glass substrate through the developed photoresist layer.

  In the color filter forming process in step S52, a large number of sets of three dots corresponding to R (Red), G (Green), and B (Blue) are arranged in a matrix or three R, G, and B A color filter is formed by arranging a plurality of stripe filter sets in the horizontal scanning direction. In the cell assembly process in step S54, a liquid crystal panel (liquid crystal cell) is assembled using the glass substrate on which the predetermined pattern is formed in step S50 and the color filter formed in step S52. Specifically, for example, a liquid crystal panel is formed by injecting liquid crystal between a glass substrate and a color filter. In the module assembling process in step S56, various components such as an electric circuit and a backlight for performing the display operation of the liquid crystal panel are attached to the liquid crystal panel assembled in step S54.

  In addition, the present invention is not limited to application to an exposure apparatus for manufacturing a semiconductor device, for example, an exposure apparatus for a display device such as a liquid crystal display element formed on a square glass plate or a plasma display, It can also be widely applied to an exposure apparatus for manufacturing various devices such as an image sensor (CCD or the like), a micromachine, a thin film magnetic head, and a DNA chip. Furthermore, the present invention can also be applied to an exposure process (exposure apparatus) when manufacturing a mask (photomask, reticle, etc.) on which mask patterns of various devices are formed using a photolithography process.

In the above-described embodiment, ArF excimer laser light (wavelength: 193 nm) or KrF excimer laser light (wavelength: 248 nm) is used as the exposure light. However, the present invention is not limited to this, and other suitable pulse lasers are used. A light source, for example, an F ₂ laser light source that supplies laser light with a wavelength of 157 nm, a Kr ₂ laser light source that supplies laser light with a wavelength of 146 nm, an Ar ₂ laser light source that supplies laser light with a wavelength of 126 nm, or the like can be used. It is also possible to use a CW (Continuous Wave) light source such as an ultrahigh pressure mercury lamp that emits bright lines such as g-line (wavelength 436 nm) and i-line (wavelength 365 nm). A harmonic generator of a YAG laser or the like can also be used. In addition, as disclosed in, for example, US Pat. No. 7,023,610, a single wavelength laser beam in an infrared region or a visible region oscillated from a DFB semiconductor laser or a fiber laser is used as vacuum ultraviolet light. For example, a harmonic that is amplified by a fiber amplifier doped with erbium (or both erbium and ytterbium) and wavelength-converted into ultraviolet light using a nonlinear optical crystal may be used.

  In the above-described embodiment, a so-called immersion method is applied in which the optical path between the projection optical system and the photosensitive substrate is filled with a medium (typically liquid) having a refractive index larger than 1.1. You may do it. In this case, as a technique for filling the liquid in the optical path between the projection optical system and the photosensitive substrate, a technique for locally filling the liquid as disclosed in International Publication No. WO99 / 49504, a special technique, A method of moving a stage holding a substrate to be exposed as disclosed in Kaihei 6-124873 in a liquid bath, or a predetermined stage on a stage as disclosed in JP-A-10-303114. A method of forming a liquid tank having a depth and holding the substrate therein can be employed. Here, the teachings of International Publication No. WO99 / 49504, JP-A-6-124873 and JP-A-10-303114 are incorporated by reference.

  In the above-described embodiment, a so-called polarization illumination method disclosed in US Publication Nos. 2006/0170901 and 2007/0146676 can be applied. Here, the teachings of US Patent Publication No. 2006/0170901 and US Patent Publication No. 2007/0146676 are incorporated by reference.

  In the above-described embodiment, the present invention is applied to the illumination optical system that illuminates the mask in the exposure apparatus. However, the present invention is not limited to this, and the present invention generally relates to an illumination optical system for supplying illumination light to an imaging optical system that forms an image of an object disposed on the first surface on the second surface. Can be applied.

DESCRIPTION OF SYMBOLS 1 Light source 2 Beam transmission part 3 Spatial light modulation unit 30 Spatial light modulator 30a Mirror element 30c Drive part 32 K prism 4 Relay optical system 5 Micro fly eye lens 6 Condenser optical system 7 Illumination field stop (mask blind)
8 Imaging optical system IL Illumination optical system CR Control unit DT Pupil distribution measuring device M Mask PL Projection optical system W Wafer

Claims (30)

A method of setting a pupil intensity distribution to be formed on an illumination pupil of an illumination optical system that supplies illumination light to an imaging optical system that forms an image of a pattern arranged on a first surface on a second surface,
Calculating a process window using a predetermined pattern arranged on the first surface and a pupil intensity distribution represented by a plurality of light intensity elements distributed two-dimensionally along the illumination pupil;
As the acceptable luma disk error enhancement-factor constraints, the process window is characterized in that it comprises, and setting the pupil intensity distribution using a linear programming so that a predetermined value or more How to set the pupil intensity distribution. Further comprising setting an allowable line width error;
Setting the pupil intensity distribution is used as a mask error enhancement factor constraints that the are acceptable linewidth error and the allowable setting the pupil intensity distribution such that the process window is equal to or greater than a predetermined value setting method according to claim 1, characterized in that. A method of setting a pupil intensity distribution to be formed on an illumination pupil of an illumination optical system that supplies illumination light to an imaging optical system that forms an image of a pattern arranged on a first surface on a second surface,
Calculating a process window using a predetermined pattern arranged on the first surface and a pupil intensity distribution represented by a plurality of light intensity elements distributed two-dimensionally along the illumination pupil;
Setting the allowable line width error;
Setting the pupil intensity distribution so that the process window is equal to or greater than a predetermined value, with the allowable line width error and the allowable mask error enhancement factor as constraints. How to set the pupil intensity distribution . 4. The setting method according to claim 3 , wherein the allowable line width error is an allowable error from a desired value of a line width of a pattern image formed on the second surface . As constraints on the allowable line width error and the allowable mask error enhancement factor, individual differences of devices including the imaging optical system and the illumination optical system and / or variations of apparatus parameters were introduced. The setting method according to claim 4 , wherein a constraint condition is used . The setting of the pupil intensity distribution includes setting the pupil intensity distribution using a convex programming method so that the process window is equal to or greater than a predetermined value. The setting method described in the section . A method of setting a pupil intensity distribution to be formed on an illumination pupil of an illumination optical system that supplies illumination light to an imaging optical system that forms an image of a pattern arranged on a first surface on a second surface,
Calculating a process window using a predetermined pattern arranged on the first surface and a pupil intensity distribution represented by a plurality of light intensity elements distributed two-dimensionally along the illumination pupil;
Using a convex programming method to set a pupil intensity distribution so that the process window is equal to or greater than a predetermined value using an allowable mask error enhancement factor as a constraint. Distribution setting method. The setting of the pupil intensity distribution includes setting the pupil intensity distribution so that an area of an ellipse representing the process window approaches a maximum value. The setting method described. The setting method according to claim 1, wherein setting the pupil intensity distribution includes obtaining a pupil intensity distribution set from a standard pupil intensity distribution . Calculating a mask error enhancement factor based on the set pupil intensity distribution;
Determining whether a difference between the calculated mask error enhancement factor and the allowable mask error enhancement factor is within an allowable range. Item 10. The setting method according to any one of Items 1 to 9 . When the difference between the calculated mask error enhancement factor and the allowable mask error enhancement factor is not within the allowable range, the calculated mask error enhancement factor and the Further comprising setting the mask error enhancement factor by repeating setting the pupil intensity distribution until a difference from an allowable mask error enhancement factor falls within the allowable range. The setting method according to claim 10. Repeatedly setting the mask error enhancement factor for different DOFs;
The setting method according to claim 11 , further comprising: obtaining a set DOF value when an area of an ellipse representing the process window is closest to a maximum value . Obtaining a shift amount from a desired position of an evaluation position of a pattern image formed on the second surface using a pupil intensity distribution corresponding to the set DOF;
The setting method according to claim 12, further comprising: determining whether or not the shift amount is larger than a predetermined value . Repeatedly setting the mask error enhancement factor for a plurality of different predetermined focus positions;
The setting method according to claim 11, further comprising: obtaining a predetermined focus position that is set when an area of an ellipse representing the process window is closest to a maximum value. . Obtaining a shift amount from a desired position of an evaluation position of a pattern image formed on the second surface using a pupil intensity distribution corresponding to the set predetermined focus position;
The setting method according to claim 14, further comprising: determining whether the shift amount is greater than a predetermined value . The method further includes setting the evaluation position while adjusting the evaluation position so that an area of an ellipse representing the process window converges toward a maximum value when the shift amount is larger than the predetermined value. The setting method according to claim 13 or 15, characterized in that 17. The setting method according to claim 16, wherein when the evaluation position is set, the shift amount calculated according to a plurality of defocus positions on the second surface is used as the adjustment amount of the evaluation position . In the adjustment method of the illumination optical system that illuminates the illuminated surface with light from the light source,
Using the setting method according to any one of claims 1 to 17, setting a pupil intensity distribution to be formed on the illumination pupil of the illumination optical system;
Adjusting the pupil intensity distribution formed on the illumination pupil using the set pupil intensity distribution as a target . When adjusting the pupil intensity distribution formed on the illumination pupil with the set pupil intensity distribution as a target, the light intensity on a surface optically conjugate with the illumination pupil based on the light passing through the irradiated surface The adjustment method according to claim 18, wherein the distribution is measured . A spatial light modulator provided in the illumination optical system and controlling the spatial light modulator for forming a pupil intensity distribution in the illumination pupil, thereby setting the set pupil intensity distribution as a target to the illumination pupil. 20. The adjustment method according to claim 18 or 19, wherein the pupil intensity distribution to be formed is adjusted . 21. An illumination optical system that is adjusted by the adjustment method according to claim 18 . In the illumination optical system that illuminates the illuminated surface with light from the light source,
A pupil distribution measuring device that measures a pupil intensity distribution formed on the illumination pupil of the illumination optical system;
A pupil adjustment device for adjusting a pupil intensity distribution formed on the illumination pupil;
Control for controlling the pupil adjustment device to adjust the pupil intensity distribution formed on the illumination pupil with the pupil intensity distribution set by using the setting method according to claim 1 as a target. And an illumination optical system. The illumination optical system according to claim 22 , wherein the pupil distribution measurement device measures a light intensity distribution on a surface optically conjugate with the illumination pupil based on light passing through the irradiated surface . 24. The illumination optical system according to claim 23, wherein the pupil distribution measurement device measures a light intensity distribution on a surface optically conjugate with the illumination pupil based on light traveling toward the irradiated surface . The pupil adjustment device includes a plurality of optical elements arranged on a predetermined plane and individually controlled, and includes a spatial light modulator that variably forms a pupil intensity distribution on the illumination pupil,
25. The illumination optical system according to claim 22, wherein the control unit controls the plurality of optical elements of the spatial light modulator . The spatial light modulator includes a plurality of mirror elements arranged two-dimensionally within the predetermined plane, and a drive unit that individually controls and drives the postures of the plurality of mirror elements. The illumination optical system according to 25 . 27. An exposure apparatus comprising the illumination optical system according to claim 21 for illuminating a predetermined pattern, and exposing the predetermined pattern onto a photosensitive substrate . 28. A projection optical system that forms an image of the predetermined pattern on the photosensitive substrate, wherein the illumination pupil is at a position optically conjugate with an aperture stop of the projection optical system. The exposure apparatus described. Using the exposure apparatus according to claim 27 or 28, exposing the predetermined pattern to the photosensitive substrate;
Developing the photosensitive substrate having the predetermined pattern transferred thereon, and forming a mask layer having a shape corresponding to the predetermined pattern on the surface of the photosensitive substrate;
Processing the surface of the photosensitive substrate through the mask layer. A device manufacturing method comprising: A program for causing a computer to execute the setting method according to any one of claims 1 to 17.