Method for predicting at least one illumination parameter for evaluating an illumination setting

Abstract

In a method for predicting at least one illumination parameter for evaluating an illumination setting for illuminating an object field of a projection exposure apparatus, illumination parameters are measured at a number of calibration settings, correction terms for prediction values of the illumination parameters are determined from the measured values, and then at least one illumination parameter of at least one illumination setting, which is not contained in the set of n calibration settings, is predicted.

Claims

1. A method for predicting an illumination parameter for evaluating an illumination setting for illuminating an object field of a projection exposure apparatus, the projection exposure apparatus comprising an illumination optical unit comprising a facet element for generating a plurality of illumination settings with a plurality of illumination channels, the method comprising: calibrating a model for predicting a plurality of illumination parameters depending on the illumination setting by: determining a first set of correction terms on the basis of system data of the illumination optical unit; using the first set of correction terms for prediction values of a first subset of the illumination parameters; measuring actual values of the illumination parameters for a number of calibration settings; determining a second set of correction terms on the basis of the measured actual values; using the second set of correction terms for prediction values of a second subset of the illumination parameters; and updating the model taking into account the first and second sets of correction terms; and using the model to predict an illumination parameter for an illumination setting which is not contained in the set of calibration settings.

2. The method of claim 1, wherein the illumination parameters which are predictable by the model comprise parameters selected from the group consisting of intensity centroids of an illumination setting and integral intensities of an illumination setting.

3. The method of claim 1, wherein the first set of correction terms consists of correction terms for predicting intensity centroids of illumination settings.

4. The method of claim 1, wherein the second set of correction terms consists of correction terms for predicting intensities.

5. The method of claim 1, wherein the number of calibration settings for determining the correction terms is at most 50.

6. The method of claim 1, further comprising re-determining the second set of correction terms regularly on a timescale of at least a few hours.

7. A method of using an illumination optical unit to illuminate a field with an illumination setting, the illumination optical unit comprising a facet element for generating a plurality of illumination settings with a plurality of illumination channels, the method comprising: calibrating a model for predicting a plurality of illumination parameters depending on the illumination setting by: providing a first set of correction terms for prediction values of a first subset of the illumination parameters; measuring actual values of illumination parameters for a number of calibration settings; determining a second set of correction terms for prediction values of a second subset of the illumination parameters taking into account the measured actual values; and updating the model taking into account correction terms for prediction values of a second subset of the illumination parameters.

8. The method of claim 7, further comprising using the model to predict an illumination parameter of an illumination setting which is not contained in the set of calibration settings.

9. The method of claim 7, wherein the illumination optical unit is a unit of a projection exposure apparatus, and the field is an object field of the projection exposure apparatus.

10. The method of claim 7, wherein the illumination parameters which are predictable by the model comprise parameters selected from the group consisting of intensity centroids of an illumination setting and integral intensities of an illumination setting.

11. The method of claim 7, wherein the first set of correction terms consists of correction terms for predicting intensity centroids of illumination settings, and the set of correction terms consists of correction terms for predicting intensities.

12. The method of claim 7, wherein the number of calibration settings for determining the correction terms is at most equal to a number of calibration settings performed to provide the correction terms.

13. The method of claim 7, further comprising re-determining the second set of correction terms regularly on a timescale of at least a few hours.

14. A method for evaluating an illumination setting for illuminating an object field of a projection exposure apparatus, the projection exposure apparatus comprising an illumination system for generating a plurality of illumination settings with a plurality of illumination channels, the method comprising: providing correction terms for prediction values of a subset of the illumination parameters; measuring actual values of the illumination parameters at a number of calibration settings; determining correction terms for prediction values of a subset of the illumination parameters taking into account the measured values; updating the model taking into account the correction terms; predicting the illumination parameters for establishing a functional value of an evaluation function for an illumination setting depending on the illumination setting, the evaluation function being a function of predetermined illumination parameters; evaluating the illumination settings via the evaluation function on the basis of the predicted illumination parameters; and iteratively varying the illumination settings until a predetermined termination criterion is achieved, wherein the method further comprises at least one of the following: establishing the correction terms on the basis of system data of the illumination optical unit; and establishing the correction terms on the basis of measurements.

15. The method of claim 14, further comprising: establishing the correction terms on the basis of system data of the illumination optical unit; and establishing the correction terms on the basis of measurements.

16. The method of claim 1, further comprising storing the first set of correction terms in a memory of the illumination optical unit.

17. The method of claim 16, wherein the first set of correction terms comprises correction terms for correcting geometric deviations of the illumination optical unit.

18. The method of claim 1, wherein the first set of correction terms is disjoint from the second set of correction terms.

19. The method of claim 1, wherein the first set of correction terms comprises correction terms for correcting geometric deviations of the illumination optical unit.

20. The method of claim 7, further comprising storing the first set of correction terms in a memory of the illumination optical unit.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Further details and specifics of the disclosure emerge from the description of exemplary embodiments on the basis of the drawings. In detail:

(2) FIG. 1 shows a schematic illustration of the components of a microlithographic projection exposure apparatus,

(3) FIG. 2 schematically shows a flowchart of a method for predicting an illumination parameter for evaluating an illumination setting,

(4) FIGS. 3A to 3G show exemplary diagrams for elucidating the correlations between measurements and radiation intensities, predicted via the model, of individual illumination channels for different illumination settings, wherein, for elucidation purposes, correlations of the non-calibrated model (open circles) and correlations of the model calibrated according to the disclosure (crosses) are depicted in each case, and

(5) FIG. 4 schematically shows a flowchart of a method for optimizing an illumination setting.

DETAILED DESCRIPTION

(6) Below, the basic design of a microlithographic projection exposure apparatus 1 is first of all described in an exemplary manner. The disclosure is not restricted to the specific details of this embodiment. In particular, it can also be applied in the case of alternative embodiments of projection exposure apparatuses 1, in particular illumination optical units. The projection exposure apparatus 1 substantially corresponds to the one known from DE 10 2012 220 596 A1, which is referred to herewith. In particular, DE 10 2012 220 596 A1 is herewith integrated in its entirety into the present application as a component thereof.

(7) A microlithographic projection exposure apparatus 1 serves to produce a microstructured or nanostructured electronic semiconductor component. A radiation source 2 emits used EUV radiation in the wavelength range of e.g. between 5 nm and 30 nm for illumination purposes. The radiation source 2 can be a GDPP (gas discharge produced plasma) source or an LPP (laser produced plasma) source. A radiation source based on a synchrotron can also be used for the radiation source 2. A person skilled in the art finds information about such a radiation source in e.g. U.S. Pat. No. 6,859,515 B2. EUV illumination light or illumination radiation 3 is used for illumination and imaging within the projection exposure apparatus 1. After the radiation source 2, the EUV illumination light 3 initially passes through a collector 4, which may be e.g. a nested collector with a multi-shell design known from the prior art or, alternatively, an ellipsoidally formed collector. A corresponding collector is known from EP 1 225 481 A2. After the collector 4, the EUV illumination light 3 initially passes through an intermediate focal plane 5, which can be used to separate unwanted radiation or particle components from the EUV illumination light 3. After passing through the intermediate focal plane 5, the EUV illumination light 3 initially impinges on a field facet mirror 6 with a multiplicity of field facets 7.

(8) In order to simplify the description of positional relationships, a global Cartesian xyz coordinate system is respectively plotted in the drawing. In FIG. 1, the x-axis extends perpendicular to the plane of the drawing and out of the latter. In FIG. 1, the y-axis extends to the right. In FIG. 1, the z-axis extends upward.

(9) In the region of the object plane and the image plane, the y-direction corresponds to the scanning direction of the reticle and of the wafer.

(10) After reflection at the field facet mirror 6, the EUV illumination light 3 split into bundles of rays or partial beams which are assigned to the individual field facets 7 impinges on a pupil facet mirror 10 with a multiplicity of pupil facets 11.

(11) The field facets 7 of the field facet mirror 6 are tiltable between a plurality of illumination tilting positions, such that a beam path of the illumination light 3 reflected by the respective field facet 7 can thereby be altered in terms of its direction and the impingement point of the reflected illumination light 3 on the pupil facet mirror 10 can thus be altered. Corresponding field facets that can be displaced between different illumination tilting positions are known from U.S. Pat. Nos. 6,658,084 B2 and 7,196,841 B2.

(12) Each partial beam of the EUV illumination light 3 which is reflected by one of the field facets 7 is assigned at least one pupil facet 11 in such a way that a respective impinged-upon facet pair including one of the field facets 7 and one of the pupil facets 11 predefines an object field illumination channel for the associated partial beam of the EUV illumination light 3. The object field illumination channel is also denoted as illumination channel below. The channel-wise assignment of the pupil facets 11 to the field facets 7 is effected in a manner dependent on a desired illumination by the projection exposure apparatus 1. The totality of the correspondingly selected illumination channels is also referred to as illumination setting.

(13) By way of the respective illumination tilting positions of the respective field facet 7, the disjoint set of pupil facets 11 of the pupil facet mirror 10 is assigned to this field facet 7. Each one of the pupil facets 11 from one of these sets is impinged upon by illumination light 3 by way of exactly one of the various tilting positions of the associated field facets 7 such that, depending on the tilting position of the field facet 7, a specific illumination channel is formed between the field facet 7 and one of the pupil facets 11 of the pupil facet set. The illumination channels which can be used depending on the tilting position of exactly one of the field facets 7, that is to say by way of which the pupil facets 11 of the disjoint set of pupil facets 11 assigned to this field facet 7 can be impinged upon with the illumination light partial beam, form an illumination channel group. A field facet 7 can have more tilting positions which can be set via an actuator connected therewith than tilting positions which lead to the formation of an illumination channel. Only a tilting position which leads to the formation of an illumination channel is intended to be referred to as tilting position below.

(14) The field facet mirror 6 has several hundred field facets 7, for example 300 field facets 7. The pupil facet mirror 10 has a number of pupil facets 11 which is at least as large as the sum of the tilting positions of all field facets 7 of the field facet mirror 6. In this case, some of the pupil facets are not used for the employed assignment of pupil facets to field facets. In particular, it is advantageous if the sum of the tilting positions of all field facets 7 of the field facet mirror 6 is equal to the number of pupil facets 11. In a variant not depicted here, the pupil facet mirror 10 is constructed as MEMS mirror array with a multiplicity of tiltable individual mirrors, wherein each one of the pupil facets 11 is formed by a plurality of such individual mirrors. Such a design of the pupil facet mirror 10 is known from US 2011/0001947 A1.

(15) The field facets 7 are imaged in an object plane 16 of the projection exposure apparatus 1 by way of the pupil facet mirror 10 (cf. FIG. 1) and a subsequent transmission optical unit 15 consisting of three EUV mirrors 12, 13, 14. The EUV mirror 14 is configured as a mirror for grazing incidence (grazing incidence mirror). Arranged in the object plane 16 is an object in the form of reticle 17, of which an illumination region in the form of an illumination field is illuminated using the EUV illumination light 3, which illumination region coincides with an object field 18 of a projection optical unit 19 of the projection exposure apparatus 1 disposed downstream thereof. The object field illumination channels are superposed in the object field 18. The EUV illumination light 3 is reflected by the reticle 17.

(16) The projection optical unit 19 images the object field 18 in the object plane 16 into an image field 20 in an image plane 21. Arranged in this image plane 21 is a wafer 22, which carries a light-sensitive layer which is exposed during the projection exposure using the projection exposure apparatus 1. During the projection exposure, both the reticle 17 and the wafer 22 are scanned in the y-direction in a synchronized manner. The projection exposure apparatus 1 is configured as a scanner. The scanning direction y is also referred to as object displacement direction below.

(17) The field facet mirror 6, the pupil facet mirror 10 and the mirrors 12 to 14 of the transmission optical unit 15 are components of an illumination optical unit 23 of the projection exposure apparatus 1. In one variant of the illumination optical unit 23, which is not depicted in FIG. 1, it is also possible to partly or completely dispense with the transmission optical unit 15 such that no further EUV mirror, exactly one further EUV mirror or else exactly two further EUV mirrors may be arranged between the pupil facet mirror 10 and the object field 18. The pupil facet mirror 10 can be arranged in an entry pupil plane of the projection optical unit 19.

(18) Together with the projection optical unit 19, the illumination optical unit 23 forms an optical system of the projection exposure apparatus 1.

(19) The illumination optical unit 23 is a component of an illumination system 24, which moreover includes the radiation source 2.

(20) The field facet mirror 6 constitutes a first facet mirror of the illumination optical unit 23. The field facets 7 constitute first facets of the illumination optical unit 23.

(21) The pupil facet mirror 10 constitutes a second facet mirror of the illumination optical unit 23. The pupil facets 11 constitute second facets of the illumination optical unit 23.

(22) A method for predicting at least one illumination parameter for evaluating an illumination setting for illuminating the object field 18, in particular the reticle 17 arranged in the object field 18, during the operation of the projection exposure apparatus 1 is described below. A model, in particular a simulation model, serves to predict the illumination parameters. The model includes design data of the projection exposure apparatus 1, in particular of the illumination optical unit 23. The model can also include further design data, in particular of the radiation source 2 and/or of the collector 4. A corresponding model is known e.g. from DE102012220596A1, to which reference is hereby made.

(23) In particular, the model renders possible the prediction of intensity centroids rhx.sup.PP, rhy.sup.PP and integral intensities I.sup.PP of the illumination spots in an illumination pupil. This should be understood to mean, in particular, the field point-resolved angle distribution of the illumination radiation on the reticle.

(24) What was identified is that there may be errors in the prediction of the illumination quality (imaging performance) of the different illumination settings due to deviations of the real system from the design, i.e. from the setpoint state thereof. In particular, the illumination quality can be characterized by parameters such as uniformity, telecentricity or ellipticity of the illumination. Alternatively, it is possible to use a structure resolution achievable by the illumination when imaging the illuminated object field 18 into the image field 20 as an illumination parameter. It is also possible to use a variation of a line width of an imaged structure over the image field 20 as a parameter for evaluating an illumination setting. In respect of further details, reference is made to DE 10 2012 220 596 A1 and the prior art cited therein.

(25) According to the disclosure, a real illumination pupil is measured. It is chopped into individual illumination spots. The energetic centroid rhx.sup.MESS, rhy.sup.MESS and the total intensity I.sup.Mes contained therein is determined for each spot. The illumination parameters which reflect the imaging performance, such as e.g. uniformity, telecentricity or ellipticity, can be determined from these measurement data. Accordingly, the model serves to predict the illumination parameters, in particular to predict the illumination centroids of the pupil spots and the intensity contained therein, i.e. the prediction of the illumination pupil. This information serves as input of the further models, by which the imaging performance can be determined and/or predicted.

(26) What was identified according to the disclosure is that the measured illumination pupils can be used to improve the prediction for further illumination settings. In particular, it is possible to measure a relatively small number of illumination settings, for example at most 10, in particular at most 5, in particular at most 3, in particular at most 2, in particular merely 1 illumination setting, in order thereby to improve the prediction accuracy of the model for a multiplicity of other illumination settings, in particular for all illumination settings generable by the illumination optical unit.

(27) The model for predicting the illumination parameters serves in particular to predict the intensity centroids rhx.sup.PP.sub.ijk, rhy.sup.PP.sub.ijk and the intensities I.sup.PP.sub.ijk, where i denotes the index of the illumination spot, in particular of the illumination spot in a specific illumination setting, in the pupil, j indexes the illumination setting and k characterizes the field point. In particular, the index i of the illumination channel runs from 1 to the number of spots in a pupil, i.e. up to 100, in particular up to 300, up to 500, up to 1000 or more; the index k of the field point runs e.g. from 1 to the number of measured field points, i.e. up to 3, in particular to 5, in particular up to 10, up to 20 or more. Other values are likewise possible.

(28) The actual values of these parameters emerging from measurements are denoted by rhx.sup.mess.sub.ijk, rhy.sup.mess.sub.ijk and I.sup.Mess.sub.ijk.

(29) For the purposes of comparability of different field points and illumination settings, the overall intensities of a pupil, i.e. the overall intensities of an illumination setting, can be normalized to 1:

(30) $\underset{i}{.Math.}{I}_{\mathrm{ijk}}^{\mathrm{PP}}=1$$\underset{i}{.Math.}{I}_{\mathrm{ijk}}^{\mathrm{Mess}}=1$

(31) Using this normalization, the following correction terms emerge for the intensities I.sub.ijk and centroids rhx.sub.ijk, rhy.sub.ijk, taking into consideration the overall intensities I.sub.tot:

(32) $v_{\mathrm{ijk}}=\frac{I_{\mathrm{ijk}}^{\mathrm{Mess}}.Math.I_{\mathrm{tot},\mathrm{jk}}^{\mathrm{Mess}}}{I_{\mathrm{ijk}}^{\mathrm{PP}}.Math.I_{\mathrm{tot},\mathrm{jk}}^{\mathrm{PP}}}$${\mathrm{rhx}}_{\mathrm{ijk}}={\mathrm{rhx}}_{\mathrm{ijk}}^{\mathrm{Mess}}-{\mathrm{rhx}}_{\mathrm{ijk}}^{\mathrm{PP}}$${\mathrm{rhy}}_{\mathrm{ijk}}={\mathrm{rhy}}_{\mathrm{ijk}}^{\mathrm{Mess}}-{\mathrm{rhy}}_{\mathrm{ijk}}^{\mathrm{PP}}$

(33) Here, I.sup.mess.sub.tot,jk and I.sup.PP.sub.tot,jk denote the overall intensities of the respective pupils prior to the normalization under the assumption that the radiation source maintains its emission characteristic during the measurements and the measurement time of each measurement is constant. In this manner, two pupils with different settings and field points become comparable to one another. If this were not the case, it would be unclear for a specific factor in the case of an intensity difference between two spots as to whether this actually is a property of the illuminator or whether the exposure times of the measurements differed by this factor.

(34) What was identified according to the disclosure is that the number of illumination settings for establishing correction terms for each illumination channel is generally less than the overall number of all possible illumination settings. FIGS. 3A to 3G depict, in an exemplary manner, the effect of taking into account the above-described correction terms on the prediction of the intensities of individual illumination channels. In each case, the prediction value (y-axis) is plotted against the actual measured value (x-axis).

(35) The prediction values without taking into account the correction terms are denoted by circular symbols. The prediction values taking into account the aforementioned correction terms are characterized by + symbols.

(36) Seven different illumination settings are depicted in an exemplary manner, wherein the illumination settings depicted in FIGS. 3A and 3B were used for establishing the correction terms. The prediction of the intensities of the illumination settings depicted in FIGS. 3C to 3G was carried out taking into account the correction values established on the basis of the illumination settings underlying FIGS. 3A and 3B. As can clearly be identified qualitatively, only taking into account the correction terms leads to a clear correlation between the prediction values and the measured values.

(37) What has emerged furthermore is that taking into account the correction terms also has a significant influence on predicting the conventional pupil parameters such as e.g. telecentricity and ellipticity.

(38) By way of example, it was possible to improve the prediction error for the telecentricity by a factor of more than 2. In some illumination settings, the prediction error for telecentricity could also be improved by a factor of more than 3, in particular of more than 5, in particular of up to 10.

(39) What was identified according to the disclosure is that significant measurement time can be saved when determining the correction terms. In particular, this can be traced back to the fact that a subset of the correction terms, which are denoted by s.sub.i below, is substantially constant in time. By way of example, these include the correction terms rhx.sub.ijk and rhy.sub.ijk. This results primarily from geometric deviations of the illumination optical unit. It is sufficient to determine these correction terms only once, in particular when putting the illumination optical unit 23 into operation. Then, they can be stored in a memory of the illumination system. They can be updated where desired.

(40) In particular, the correction terms s.sub.i can be established on the basis of system data of the illumination optical unit 23. In particular, they can be established on the basis of a selection of the individual part acceptance data, overall system acceptance data and calibration data described in DE 10 2012 220 596 A1.

(41) Another subset of the correction terms, which is denoted by f.sub.i below, is traced back to dynamic variations. By way of example, these include the corrections v.sub.ijk.

(42) The intensity deviations v.sub.ijk are dominated by variations in the radiation source. What was identified is that it is therefore sufficient to establish the corresponding correction terms for a small number of illumination settings, in particular at most 10, in particular at most 5, in particular at most 3, in particular at most 2, in particular exactly 1 illumination setting. The corresponding correction terms v.sub.ijk can then be adopted for the remaining illumination settings. A significant amount of measurement time can be saved thereby.

(43) In particular, the correction terms f.sub.i can be established on the basis of a selection of the online measurement data described in DE 10 2012 220 596 A1.

(44) The concept according to the disclosure is advantageous, in particular, for illumination optical units 23 with a large number of different switching positions.

(45) Below, the method for predicting at least one illumination parameter p* for evaluating an illumination setting is described once again in other words, with reference being made to FIG. 2.

(46) Initially, m calibration settings j.sub.Kal generated by the illumination system 24 are measured in a first measurement step 31.

(47) In a first determination step 32, the correction terms s.sub.i for the prediction values of the first subset p.sub.1 of the illumination parameters p are determined. In particular, the correction terms s.sub.i are determined on the basis of the measured values from the first measurement step 31.

(48) Then, a check is carried out in a first query 33 as to whether the correction terms f.sub.i are already determined.

(49) If this is not the case, nm calibration settings j.sub.Kal generated by the illumination system are measured in a second measurement step 34.

(50) The correction terms f.sub.i are determined in a second determination step 35 on the basis of the measured values from this second measurement step 34.

(51) Once both the correction terms s.sub.i and the correction terms f.sub.i are determined, it is possible to wait until a prediction of an illumination parameter becomes desirable. This is depicted in FIG. 2 as calibrated basic state 36. To the extent that what emerges in the first query 33 is that the correction terms f.sub.i are already determined, it is possible to transition immediately to the basic state 36. Initially, it is not necessary to once again carry out the second measurement step 34 and the second determination step 35.

(52) Together, the steps 31 to 35 form a calibration part 37 for calibrating the model for predicting the illumination parameters. Advantageously, as few calibration settings j.sub.Kal as possible are measured in the calibration part 37. The number of calibration settings is less than 50, in particular less than 10, in particular less than 5, in particular less than 3, in particular less than 2. In particular, it is merely 1.

(53) After calibrating the model in the calibration part 37, an illumination parameter of an illumination setting j* can be predicted in a prediction part 38. To this end, a check is initially carried out in a second query 39 as to whether the correction terms s.sub.i are still current. To the extent that this is not the case, there is a return to the first measurement step 31. To the extent that the correction terms s.sub.i are still current, a check is carried out in a third query 40 as to whether the correction terms f.sub.i are still current. To the extent that this is not the case, there is a return to the second measurement step 34.

(54) By way of example, updating the correction terms can be triggered by measuring the radiation source to the extent that the latter shows a strong change in same. Furthermore, it is desirable to carry out an update when a component of the illumination system is exchanged. In particular, then it is desirable to update the geometric correction terms.

(55) Whether the second query 39 is carried out can, in particular, be decided and triggered individually by the user of the system. In particular, provision can be made for starting the second query 39 only upon a user trigger.

(56) To the extent that all correction terms s.sub.i and f.sub.i are current, it is possible to predict an illumination parameter of an illumination setting j* in a prediction step 41. Here, in particular, the current correction terms s.sub.i and f.sub.i are taken into account.

(57) In particular, the illumination setting j* can be an illumination setting which is not contained in the set of calibration settings j.sub.kal.

(58) Using the method in accordance with FIG. 2 it is possible, in principle, to evaluate all illumination settings j which can be generated by the illumination system 24. On the other hand, it is only necessary to measure a relatively small number of calibration settings j.sub.Kal for calibrating the model in the calibration part 37. This leads to a significant saving of time. In particular, the method according to the disclosure renders it possible to use the measurement of the few illumination settings in order thereby to improve the prediction accuracy for all illumination settings.

(59) Below, a method for optimizing an illumination setting is described in note form, with reference being made to FIG. 4. Initially, a calibration of the model for predicting the parameters for evaluating the illumination settings is carried out. To this end, provision is made for a method in accordance with the calibration part 37, the description of which is hereby referred to.

(60) Thereupon, in a first definition 41, an evaluation function is defined for evaluating an illumination setting. In particular, the evaluation function is a function of predetermined illumination parameters, in particular illumination parameters which can be predicted with the aid of the method described above.

(61) Moreover, a termination criterion is defined in a second definition 42.

(62) Furthermore, an illumination setting j.sub.start is defined as start illumination setting in a third definition 43. The start illumination setting serves as starting point for an iterative optimization of the illumination setting.

(63) Then, the predetermined illumination parameters, which form the arguments for the evaluation function, are predicted in the prediction part 38 in accordance with the preceding description, to which reference is made hereby.

(64) The illumination setting is evaluated in an evaluation step 44 on the basis of the predicted illumination parameters with the aid of the evaluation function defined in the first definition 41.

(65) A check is carried out in a subsequent query 45 as to whether the termination criterion defined in the second definition 42 has been reached. To the extent that this is the case, an acceptable illumination setting was found. It can be used, in particular, to illuminate a reticle 17 arranged in the object field 18, in particular to image the reticle 17 on the wafer 22 arranged in the image field 20. This is depicted in summary in FIG. 4 as projection exposure 46.

(66) To the extent that the termination criterion has not been reached, the switching position of at least one of the field facets 7 is varied in a variation step 47. It is also possible to vary the switching position of a plurality of field facets 7, in particular of at least two, in particular of at least three, in particular of at least four, five, six, eight or ten field facets. In principle, it is even possible to vary the switching positions of all field facets 7.

(67) A new illumination setting emerges from varying at least one of the switching positions of the field facets 7, which new illumination setting in turn is analysed in the prediction part 38 and evaluated in the evaluation step 44. This is repeated until the predetermined termination criterion has been reached.