Method For Adjusting A Projection Objective

Abstract

A projection objective having a number of adjustable optical elements is optimized with respect to a number of aberrations by specifying a set of parameters describing imaging properties of the objective, each parameter in the set having an absolute value at each of a plurality of field points in an image plane of the projection objective. At least one of the optical elements is adjusted such that for each of the parameters in the set, the field maximum of its absolute value is minimized.

Claims

1-19. (canceled)

20. A microlithography projection exposure machine, comprising: an illumination device for illuminating a reticle; a projection objective comprising a plurality of optical elements for projecting an image of the reticle onto a substrate in an image field at an image plane of the projection objective; a plurality of manipulators each configured to move either the reticle or a corresponding one of the optical elements of the projection objective; and a control unit configured to adjust the projection exposure machine between different configurations by causing one or more of the manipulators to adjust a position of the reticle and/or at least one of the optical elements of the projection objective, the different configurations comprising an initial configuration and a second configuration, wherein, in the initial configuration, a maximum value at least one field point among a plurality of different field points in the image field for an absolute value of a parameter indicative of a performance of the projection exposure machine exceeds a specified field maximum value, the parameter being a lithographic process parameter or an image-forming parameter of the projection objective, and in the second configuration, a largest absolute value for every one of the plurality of different field points is less than the specified field maximum value.

21. The microlithography projection exposure machine of claim 20, wherein the field at the image plane is defined by a scanner slit.

22. The microlithography projection exposure machine of claim 21, wherein the plurality of different field points lie along a line oriented in a scanning direction of the projection exposure machine.

23. The microlithography projection exposure machine of claim 20, wherein field maximum values are specified for more than one parameter indicative of the performance of the projection exposure machine.

24. The microlithography projection exposure machine of claim 20, wherein the lithographic process parameter is related to the image-forming parameter.

25. The microlithography projection exposure machine of claim 20, wherein the image-forming parameter of the projection objective is selected from the group consisting of: individual Zernike coefficients describing wave aberrations of an objective pupil of the projection objective; a linear combinations of Zernike coefficients; and an average of Zernike coefficients over a plurality of the different points.

26. The microlithography projection exposure machine of claim 20, wherein the lithographic process parameter is selected from the group consisting of a horizontal offset of a structure from its ideal position and a deviation from an ideal line width.

27. The microlithography projection exposure machine of claim 20, wherein the parameter indicative of a performance of the projection exposure machine describes a centrable aberration, and the manipulators are configured to tilt the reticle of the projection objective to adjust for the centrable aberration.

28. The microlithography projection exposure machine of claim 27, wherein the manipulators are configured to: (i) displace at least one of the optical elements in a direction perpendicular to an optical axis of the projection objective, and (ii) tilt at least one of the optical elements in a direction perpendicular to said optical axis of the projection objective.

29. The microlithography projection exposure machine of claim 28, wherein the parameter indicative of a performance of the projection exposure machine describes a tunable aberration, and the manipulators are configured to adjust at least one of the optical elements to adjust for the tunable aberration and the centrable aberration jointly.

30. The microlithography projection exposure machine of claim 29, wherein the manipulators are configured to: (i) displace at least one of the optical elements in a direction along the optical axis of the projection objective; (ii) change a wavelength of illumination of the projection exposure machine; (iii) change a temperature within the projection exposure machine; (iv) change an air pressure within the projection exposure machine, and (v) change a composition of a purge gas surrounding the optical elements.

31. The microlithography projection exposure machine of claim 20, wherein the control unit is configured to adjust the projection exposure machine according to a nonlinear method.

32. The microlithography projection exposure machine of claim 31, wherein the nonlinear method comprises a nonlinear optimization of the parameter values.

33. The microlithography projection exposure machine of claim 32, wherein the nonlinear optimization comprises a nonlinear min-max optimization.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0013] FIG. 1 shows a schematic representation of a projection exposure machine for microlithography which can be used to expose structures on wafers coated with photosensitive materials;

[0014] FIG. 2 shows an illustration of a scanner slit within a full image plane of a projection objective;

[0015] FIG. 3 shows a graph of a Zernike coefficient Z.sub.7 after a tilt optimization with the aid of z-manipulators;

[0016] FIG. 4 shows a graph of a profile of an optimization of an astigmatism aberration;

[0017] FIG. 5 shows a graph of a profile of a Zernike coefficient Z.sub.2 with and without joint optimization of reticle tilt and xy-manipulation;

[0018] FIG. 6 shows a graph of a profile of a Zernike coefficient Z.sub.7 with and without joint optimization of reticle tilt and xy-manipulation; and

[0019] FIG. 7 shows a graph of a distortion for annular and for coherent illumination setting.

DETAILED DESCRIPTION

[0020] FIG. 1 illustrates a projection exposure machine 1 for microlithography. This serves for exposing structures on a substrate coated with photosensitive materials and which in general overwhelmingly comprises silicon and is denoted as a wafer 2 for fabricating semiconductor components such as, for example, computer chips.

[0021] The projection exposure machine 1 essentially comprises in this case an illumination device 3, a device 4 for accommodating and exactly positioning a mask provided with a grid-like structure, a so-called reticle 5 which is used to determine the later structures on the wafer 2, a device 6 for holding, moving and exactly positioning this very wafer 2, and an imaging device, specifically a projection objective 7 with a number of optical elements such as, for example, lenses 8, which are supported by mounts 9 and/or manipulators 9 in an objective housing 10 of the projection objective 7.

[0022] The fundamental functional principle provides in this case that the structures introduced into the reticle 5 are imaged in a demagnified fashion onto the wafer 2.

[0023] After exposure has been performed, the wafer 2 is moved on so that a multiplicity of individual fields each having the structure prescribed by the reticle 5 are exposed on the same wafer 2.

[0024] The illumination device 3 provides a projection beam 11, for example light or a similar electromagnetic radiation, required for imaging the reticle 5 onto the wafer 2. A laser or the like can be used as source for this radiation. The radiation is shaped in the illumination device 3 via optical elements (not illustrated) such that when impinging onto the reticle 5 the projection beam 11 has the desired properties as regards diameter, polarization, coherence and the like. The spatial coherence is in this case a measure of the angular spectrum of the radiation in the reticle plane. This parameter can be varied by the setting of various illumination settings.

[0025] An image of the structures of the reticle 5 which are introduced is produced via the projection beam 11 and transferred onto the wafer 2 in an appropriately demagnified fashion by the projection objective 7, as already explained above. The projection objective 7 has a multiplicity of individual refractive, diffractive and/or reflective optical elements such as, for example, lenses 8, mirrors, prisms, plane-parallel plates and the like, only the lens 8 being illustrated.

[0026] After the fabrication, a concluding optimization of the manipulators of the optical elements, in particular the lenses 8, the reticle tilt/reticle height and the wavelength is essential in deciding the final image-forming quality of the projection objective 7. In this case, the image-forming quality of the projection objective is optimized, inter alia, taking account of the following aberrations: distortion, field curvature, astigmatism, coma, spherical aberration and wavefront errors of higher order.

[0027] It is known from the prior art to introduce slight changes in the optical properties by measuring parameters in the case of which the effects of the manipulation of the optical elements on these parameters are known, whereupon optimization of the parameters is carried out. As a rule, Zernike coefficients which describe the image-forming properties of the projection objective are determined for this purpose from measured values. This is achieved, for example, by measurements at a number of field points via an imaging scanner slit (=field in the image plane which is relevant to lithographical imaging). As described, for example, in EP 1 231 516 A2, Zernike coefficients with designations Z.sub.2 to Z.sub.37 are determined in this way, after which the optimization is performed. For this purpose, the average root-mean-square deviation of all the measured field points from 0 is minimized for each Zernike coefficient (so-called least square optimization). Subsequently, the Zernike coefficients Z.sub.2 to Z.sub.9 are represented by their corresponding function terms. Zernike coefficients of higher order are described in EP 1 231 516 A2.

TABLE-US-00001 Zernike coefficient Z.sub.n Function term f.sub.n Z.sub.2 cos () Z.sub.3 sin () Z.sub.4 2 .sup.2 1 Z.sub.5 .sup.2 cos(2) Z.sub.6 .sup.2 sin(2) Z.sub.7 (3.sup.3 2) cos () Z.sub.8 (3.sup.3 2) sin () Z.sub.9 6.sup.4 6.sup.2 + 1

[0028] It is likewise known to use these Zernike coefficients to find the relationship between the objective properties and the lithographic process. Assuming sufficiently small aberrations, this can be accomplished in many cases with the aid of a linear model:

L.sub.i=a.sub.2Z.sub.2(i)+a.sub.3Z.sub.3(i)+ . . . +a.sub.nZ.sub.n(i)

[0029] The weighted sum can be truncated after a sufficient number of terms, since in most cases the weighting factors become small very rapidly with rising Zernike number n. Of course, it is also possible to include square terms or terms of even higher order. The weighting factors a.sub.n can be determined experimentally or by simulation.

[0030] The variable L.sub.i describes a parameter of the lithographic process at the field point i. L.sub.i can be, for example, a horizontal offset of a structure relative to its ideal position (distortion), or else the deviation from an ideal line width.

[0031] The fabrication or optimization of a projection objective 7 firstly requires knowledge of the critical lithographic process for which the projection objective 7 is later to be used. It is then possible to calculate the appropriate weighting factors a.sub.n for various Zernike coefficients from this information. Maximum absolute values can then be derived for various Zernike coefficients from the prescription as to how far various L.sub.i may be maximized.

[0032] During optimization of the projection objective 7, various Zernike coefficients are then minimized at various field points, it being possible for these also to be various L.sub.i in a specific instance. Projection objectives 7 are then also usually specified such that various Zernike coefficients and/or various L may not exceed a maximum absolute value for a specific number of field points. It is thereby ensured that the image-forming properties of the projection objective 7 suffice for a representative selection of lithographic processes.

[0033] FIG. 2 shows the round full image field 20 of the projection objective 7. All the points on the reticle 5 which lie within this region can be imaged onto the wafer 2 with the aid of the projection objective 7. When the projection exposure machine 1 is used as scanner, it is only a rectangular section, the so-called scanner slit 21, from the full image field 20 that is used. During a lithographic exposure with the aid of a scanner, the reticle 5 and the wafer 2 are moved simultaneously during imaging in a plane perpendicular to the optical axis. The consequence of this is that a point on the reticle 5 is imaged by various field points of the projection objective 7. The aberrations relevant to this imaging are therefore the aberrations of all the field points which lie on a straight line in the scanner slit 21 which is orientated in the scanning direction (indicated by arrow 22). In order to describe the image-forming properties of the projection objective 7 in a scanner, the relevant parameters such as, for example, Zernike coefficients, are not specified for individual field points, but averaged over all the field points in the scanning direction. It is also possible to introduce a weighting during this averaging in order to take account of different intensities of illumination in the course of the scanning operation. The parameters averaged in such a way are referred to as being scanner integrated.

[0034] FIGS. 3 to 7 illustrate profiles of Zernike coefficients Z.sub.2 and Z.sub.7 over a number of field points which describe the aberrations and which, for their part, are determined at various field points in the scanner slit 21 of the projection objective 7. Plotted respectively on the x-axis is the x-position in the scanner slit 21, while the y-axis respectively specifies the y-deviation of the respective Zernike coefficient from 0 in nm.

[0035] Various manipulators of the projection objective 7 are moved during the optimization of the image-forming properties. These manipulators can be subdivided into two classes with the aid of the symmetry of the induced aberrations:

[0036] 1. Manipulators for optimizing tunable aberrations: Tunable aberrations are changes in various Zernike coefficients at various field points, the induced changes being invariant under an arbitrary rotation about the optical axis (z-axis).

[0037] The following come into consideration in this case as manipulators: [0038] displacement of lenses 8 or reticle 5 along the optical axis; [0039] change in temperature and atmospheric pressure; [0040] change in wavelength; and [0041] change in the composition of the purge gas surrounding the lenses 8.

[0042] 2. Manipulators for optimizing centrable aberrations:

Centrable aberrations are changes in various Zernike coefficients at various field points, the induced changes in the plane perpendicular to the optical axis having a marked axis of symmetry. The following come into consideration in this case as manipulators: [0043] displacement of lenses 8 perpendicular to the optical axis; and [0044] tilting of lenses 8 or reticle 5 about an axis perpendicular to the optical axis.

[0045] It is known from the Seidel aberration theory that small changes in tunable aberrations always have the same field distribution for a specific Zernike coefficient. This fundamental shape of the aberrations is independent of the type of manipulator. An equivalent theoretical model exists for centrable aberrations.

[0046] The following table shows the tunable and centrable aberrations of lowest order, the tunable aberrations presented here corresponding to the third order Seidel aberrations. The centrable aberrations refer in this case to an x-decentering, this corresponding, however, to a displacement of the lens 8 along the x-axis. For decenterings along another axis, it is necessary to rotate the field profiles (with coordinates r, in the field for lithographic imaging) correspondingly.

TABLE-US-00002 Zernike Type of Tunable Centrable coefficient Z.sub.n aberration profile profile Z.sub.2 distortion r cos() r.sup.2 Z.sub.3 distortion r sin() Z.sub.4 image surface r.sup.2 Z.sub.5 astigmatism r.sup.2 cos() r cos() Z.sub.6 astigmatism r.sup.2 sin() r sin() Z.sub.7 coma r cos() r.sup.0 Z.sub.8 coma r sin() Z.sub.9 spherical r.sup.0 r cos() aberration

[0047] As an example, FIG. 3 shows a profile of a Zernike coefficient Z.sub.7 of a projection objective 7. The tilt of this profile can be set by z-manipulators. In accordance with the prior art, the optimum tilt is achieved by means of a least square optimization, that is to say the root-mean-square value of the Zernike coefficient Z.sub.7 is minimized over all the field points. When, however, the field maximum (=maximum absolute value of a Zernike coefficient at all the field points in the scanner slit 21) is used to specify the projection objective 7, it is advantageous for precisely this field maximum to be set as small as possible. This is achieved by means of a nonlinear min-max optimization.

[0048] As may be seen from FIG. 3, the field maximum in accordance with a least square optimization (curve 12a) can be substantially larger than in accordance with a min-max optimization (curve 12b). The inventor has established that performance with regard to the Zernike coefficients Z.sub.7 and Z.sub.9 can be improved by more than half a nanometer in a significant number of cases for the projection objectives 7 solely by changing from least square optimization to min-max optimization. It was possible-here for the field maximum to be lowered from 4.06 nm (least square optimization) to 2.92 nm (min-max optimization).

[0049] FIG. 4 shows a joint optimization of centrable and tunable aberrations. The joint optimization of centrable and tunable aberrations on the projection objective 7 instead of a sequential procedure additionally permits a more accurate and speedy optimization of the optical image-forming properties of the projection objective 7. A profile of a Zernike coefficient Z.sub.7 (coma) whose tunable component (tilt) and centrable component (offset) are minimized with the aid of a min-max optimization was selected as an example. The simultaneous optimization of tilt and offset here delivers a substantially better result than the sequential min-max optimization of tunable aberrations with the aid of z-manipulators, and subsequent optimization of centrable aberrations with the aid of xy-manipulators. A curve 13a shows the uncorrected Z.sub.7 profile here. A curve 13b is the Z.sub.7 profile with optimized tilt. A curve 13c shows the Z.sub.7 profile with optimized tilt and offset (optimized sequentially one after another). As may further be seen from FIG. 4, a simultaneous optimization of the field maximum with the aid of z-manipulators (tilt) and xy-manipulators (offset) is the most advantageous (curve 13d), in which case the field maximum of 8.2 nm can be lowered to 5.6 nm in contrast with the sequential method.

[0050] FIG. 5 shows the effect of a reticle tilt (curve 14a) or of a movement of the xy manipulator (curve 14b) on a profile of the Zernike coefficient Z.sub.2. FIG. 6 shows the effect of a reticle tilt (curve 15b) or of a movement of the xy manipulator (curve 15a) on a Z.sub.7 profile. A Z.sub.2 offset, which corresponds to the curve 14a, could be removed in the case of the above-described scenario by means of a reticle tilt. However, when the Z.sub.7 offset corresponding to the curve 15a is removed by the xy-manipulator in a second step, a Z.sub.2 error is introduced again into the objective (curve 14b). In the case of a 10 nm Z.sub.7 offset, this would result nevertheless in an additional Z.sub.2 error of more than 3 nm. This could be avoided by means of an orthogonalized concept of xy-manipulators and reticle tilt, for which purpose it would be necessary to treat the reticle tilt like any other xy-manipulator.

[0051] It is advantageous to apply a distortion optimization dependent on the illumination setting of the projection objective 7. The distortion values can be substantially improved in specific cases by tracking the manipulators during changing of the illumination setting (for example from annular to coherent). A geometrical distortion (Z.sub.2) is usually not optimized, but a combination of geometrical distortion and coma-induced distortion. All the scanner-integrated Zernike coefficients vanish here, with the exception of Z.sub.7, the Z.sub.7 profile including higher tunable components. In the case of the present optimization with the aid of z-manipulators, a Z.sub.2-component in the projection objective 7 is then increased so that the resulting distortion results in an annular illumination setting of 0 (curve 16a in FIG. 7). In the event of a change in setting from annular to coherent, however, the result in this case is distortion values of up to approximately 15 nm (curve 16b). It is therefore proposed according to the invention to track the xy- and z-manipulators (including wavelength and reticle) during each change in illumination setting.