Optical arrangement, EUV lithography apparatus and method for configuring an optical arrangement

Abstract

The invention relates to an optical arrangement comprising: at least one optical element comprising an optical surface and a substrate, wherein the substrate is formed from a material whose temperature-dependent coefficient of thermal expansion at a zero crossing temperature T.sub.ZC=T.sub.ZCT.sub.ref related to a reference temperature T.sub.ref is equal to zero, wherein the optical surface has, during the operation of the optical arrangement, a location-dependent temperature distribution T(x, y) that is dependent on a local irradiance (5a), is related to the reference temperature T.sub.ref and has an average temperature T.sub.av, a minimum temperature T.sub.min and a maximum temperature T.sub.max, wherein the average temperature T.sub.av is less than the average value 1/2 (T.sub.max+T.sub.min) formed from the minimum temperature T.sub.min and the maximum temperature T.sub.max, and wherein the zero crossing temperature T.sub.ZC is greater than the average temperature T.sub.av.

Claims

1. An optical arrangement, comprising: an optical element, comprising: an optical surface; and a substrate which comprises a material having a temperature-dependent coefficient of thermal expansion that is zero at a zero crossing temperature, wherein: during use of the optical arrangement, the optical surface has a location-dependent temperature distribution that: a) is dependent on a local irradiance; and b) has an average temperature, a minimum temperature and a maximum temperature; the average temperature of the location-dependent temperature distribution is less than half of a sum of the minimum temperature of the location-dependent temperature distribution and the maximum temperature of the location-dependent temperature distribution; the zero crossing temperature is greater than the average temperature; the difference between the zero crossing temperature and the average temperature of the location-dependent temperature distribution is <T.sup.3>/<T.sup.2>; and T is a deviation of the location-dependent temperature distribution from the average temperature of the location-dependent temperature distribution.

2. The optical arrangement of claim 1, further comprising: a device configured to regulate a temperature of the optical element; and a temperature control device configured to set the average temperature of the optical surface.

3. The optical arrangement of claim 2, wherein the temperature control device is configured to set, for closed-loop control, the average temperature of the optical surface.

4. The optical arrangement of claim 2, wherein the temperature control device is configured to set, in a manner dependent on the local irradiance, a difference between the zero crossing temperature and the average temperature of the location-dependent temperature distribution.

5. The optical arrangement of claim 4, wherein the average temperature location-dependent temperature distribution is such that it minimizes a measure of the wavefront aberration at the optical surface.

6. The optical arrangement of claim 4, wherein the optical arrangement comprises a plurality of optical elements, and, for each of the optical elements, an average temperature location-dependent temperature distribution is such that it minimizes a measure of the wavefront aberration of the optical arrangement.

7. The optical arrangement of claim 6, wherein the measure of the wavefront aberration is selected from the group consisting of a RMS value, an overlay error, a scale error, a telecentricity error, a depth of focus, a best focus, coma and an astigmatism.

8. The optical arrangement of claim 2, wherein the temperature control device is configured to adapt a heating power of the temperature regulating device to a radiation power absorbed by the substrate so the average temperature of the location-dependent temperature distribution is constant.

9. The optical arrangement of claim 1, wherein the zero crossing temperature is at least 0.1 K greater than the average temperature of the location-dependent temperature distribution.

10. The optical arrangement of claim 1, wherein: the optical surface has a first area portion at which a temperature of the optical surface is greater than the average temperature of the location-dependent temperature distribution; a second area portion at which a temperature of the optical surface is less than the average temperature of the location-dependent temperature distribution; and the first area portion is smaller than the second area portion.

11. The optical arrangement of claim 1, wherein the optical surface is reflective to EUV radiation.

12. A lens, comprising: an optical arrangement according to claim 1, wherein the lens is an EUV projection lens.

13. An apparatus, comprising: a projection lens comprising an optical arrangement according to claim 1, wherein the apparatus is an EUV lithography apparatus.

14. The apparatus of claim 13, wherein the optical element is in the vicinity of a pupil plane.

15. The apparatus of claim 13, further comprising an illumination system.

16. The apparatus of claim 15, wherein the illumination system is configured to provide an illumination ray with an illumination pupil having a pupil filling of less than 50%.

17. A method for configuring an optical arrangement which comprises an optical element comprising an optical surface and a substrate, the substrate comprising a material having a temperature-dependent coefficient of thermal expansion that is zero at a zero crossing temperature, the method comprising: determining a local irradiance at the optical surface of the optical element during use of the optical arrangement; determining a location-dependent temperature distribution that results from the irradiance at the optical surface, the location-dependent temperature distribution having an average temperature, a minimum temperature and a maximum temperature; determining a difference between the average temperature at the optical surface and the average value of the minimum temperature and the maximum temperatures; and based on the difference, producing the optical element from a substrate whose zero crossing temperature is greater than the average temperature.

18. The method of claim 17, further comprising: determining deformations of the optical surface caused by the location-dependent temperature distribution; and choosing the zero crossing temperature to minimize the wavefront aberration at the optical surface.

19. The method of claim 18, further comprising measuring the wavefront aberration based on a parameter selected from the group consisting of a RMS value, an overlay error, a scale error, a telecentricity error, a depth of focus, a best focus, coma and an astigmatism.

20. The method of claim 18, wherein the temperature distribution at the optical surface is time-dependent, and a temperature distribution at a point in time at which the measure of the wavefront aberration is maximized is used for the choice of the zero crossing temperature.

21. The method of claim 17, further comprising determining deformations of the optical surface caused by the location-dependent temperature distribution, wherein: determining the local irradiance, the location-dependent temperature distribution and deformations of the optical surface caused by the temperature distribution is performed for all optical elements of the optical arrangement; and a respective optical element is produced from a substrate having a zero crossing temperature chosen to minimize a measure of the wavefront aberration of the optical arrangement.

22. The method of claim 17, further comprising choosing the zero crossing temperature so that the zero crossing temperature has a predefined difference with respect to the average temperature, and the difference is dependent on the local irradiance.

23. The method of claim 22, wherein the predefined difference between the zero crossing temperature and the average temperature is <T.sup.3>/<T.sup.2>, wherein T designates the deviation of the location-dependent temperature distribution from the average temperature of the optical surface.

24. The method of claim 17, wherein the substrate is selected so that the difference between the zero crossing temperature and the average temperature is <T.sup.3>/<T.sup.2>, and T designates the deviation of the location-dependent temperature distribution from the average temperature of the optical surface.

Description

DRAWING

(1) Exemplary embodiments are illustrated in the schematic drawing and are explained in the description below. In the figures:

(2) FIG. 1 shows a schematic illustration of an EUV lithography apparatus comprising an illumination system and a projection lens,

(3) FIG. 2 shows a schematic illustration of an EUV mirror for the projection lens from FIG. 1,

(4) FIGS. 3a-c show schematic illustrations of a location-dependent temperature distribution and of deformations resulting therefrom at an optical surface of the EUV mirror from FIG. 2,

(5) FIG. 4 shows a schematic illustration of a frequency distribution of the temperature values at the surface of the EUV mirror from FIG. 2, and

(6) FIG. 5 shows an illustration of the time-dependent temperature profile during the heating of the optical surface of the EUV mirror from FIG. 2 to its steady-state operating temperature.

(7) FIG. 1 schematically shows an EUV lithography apparatus 1. The latter comprises an EUV light source 2 for generating EUV radiation having a high energy density in an EUV wavelength range of less than 50 nm, in particular between approximately 5 nm and approximately 15 nm. The EUV light source 2 can be embodied for example in the form of a plasma light source for generating a laser induced plasma or as a synchrotron radiation source. In the former case, in particular, as shown in FIG. 1, a collector mirror 3 can be used in order to concentrate the EUV radiation of the EUV light source 2 to form an illumination ray 4 and, in this way, to increase the energy density further. The illumination ray 4 serves for illuminating a patterned object M via an illumination system 10, which comprises four reflective optical elements 13 to 16 in the present example.

(8) The patterned object M can be a reflective mask, for example, which has reflective and non-reflective or at least less-reflective regions for producing at least one structure on the object M. Alternatively, the patterned object M can be a plurality of micromirrors which are arranged in a one- or multidimensional arrangement and which, if appropriate, are moveable about at least one axis in order to set the angle of incidence of the EUV radiation 4 on the respective mirror.

(9) The patterned object M reflects part of the illumination ray 4 and shapes a projection ray 5, which carries the information about the structure of the patterned object M and which is radiated into a projection lens 20, which generates an imaging of the patterned object M or of a respective partial region thereof on a substrate W. The substrate W, for example a wafer, comprises a semiconductor material, e.g. silicon, and is arranged on a mount, also designated as a wafer stage WS.

(10) In the present example, the projection lens 20 comprises four reflective optical elements 21 to 24 (mirrors) in order to generate an image of the structure present on the patterned object M on the wafer W. Typically, the number of mirrors in a projection lens 20 is between four and eight, but if appropriate just two mirrors can also be used.

(11) In order to achieve a high imaging quality during the imaging of a respective object point OP of the patterned object M onto a respective image point IP on the wafer W, extremely stringent requirements have to be imposed with regard to the surface shape of the mirrors 21 to 24 and the position or the orientation of the mirrors 21 to 24 with respect to one another or relative to the object M and to the substrate W also requires a precision in the nanometers range. In particular, a diffraction-limited imaging enabling the maximum possible resolution can be generated only when the wavefront aberrations of the projection lens 20 are sufficiently small. In the case of a diffraction-limited projection lens 20, the RMS value (root mean square) of the wavefront aberrations should be less than 1/14 of the wavelength of the imaging light. In order to achieve this, the surface shape of the mirrors 21 to 24 has to be set with high precision and the mirrors 21 to 24 likewise have to be positioned very precisely.

(12) During the operation of the projection lens 20, the problem occurs that a proportion of the radiation of the projection ray 5, which can be up to approximately 70%, is absorbed by a respective optical element 21 to 24. Depending on the quantity of absorbed radiation, heating occurs in a respective mirror 21 to 24 and, as a result, a thermal expansion occurs which leads to deformations of the reflective surfaces of the respective mirrors 21 to 24, which can alter the orientation or the surface shape of the mirrors 21 to 24 in an undesirable manner. One possibility for combating this problem is to use an open-loop or closed-loop control device 30 for setting the operating temperature or the (average) temperature of the individual mirrors 21 to 24. Changes in the expansion of a respective mirror 21 to 24 or of the associated substrate which are caused by fluctuations in temperature can be kept small in this way.

(13) In the case of the projection lens 20 shown in FIG. 1, all four mirrors 21 to 24 comprise TiO.sub.2-doped quartz glass (ULE) as substrate material. FIG. 2 shows by way of example the first mirror 21 of the projection lens 20 in a schematic illustration. The first mirror 21 comprises a substrate 32 composed of ULE, the TiO.sub.2 proportion of which is chosen such that the substrate 32 has a desired zero crossing temperature T.sub.ZC (which is as constant as possible over the substrate volume). For the following considerations, the zero crossing temperature T.sub.ZC and also further temperature-dependent variables T.sub.a are related to a reference temperature T.sub.ref (that is to say that T.sub.ZC=T.sub.ZCT.sub.ref and T.sub.a=T.sub.aT.sub.ref). The reference temperature T.sub.ref denotes a (steady-state) temperature state that is present in the substrate material 32 or in the projection lens 20 when no illumination ray 4 is fed to the EUV lithography apparatus 1. Typically, the reference temperature T.sub.ref corresponds to the ambient temperature and can be e.g. room temperature (approximately 22 C.).

(14) A reflective coating 31 is applied to the substrate 32, the reflective coating comprising a plurality of individual layers (not designated in more specific detail) which consist alternately of materials having different reflective indices. If EUV radiation at a wavelength in the range of 13.5 nm is used in the projection lens 20, then the individual layers usually consist of molybdenum and silicon. Other material combinations such as e.g. molybdenum and beryllium, ruthenium and beryllium or lanthanum and B.sub.4C are likewise possible. In addition to the individual layers, a reflective coating can also comprise intermediate layers for preventing diffusion and also a capping layer for preventing oxidation and/or corrosion. The top side of the substrate 32 is designated hereinafter as reflective or optical surface 31a, even though strictly speaking the reflective coating 31 as a whole brings about the reflection of the EUV radiation.

(15) The substrate 32 is applied to a carrier 33, in which a plurality of heating/cooling elements 33a in the form of Peltier elements are provided, which serve for heating, if appropriate also cooling, the substrate 32 as homogeneously as possible to a working temperature, which is also designated as average temperature T.sub.av. As a result of the projection ray 5, to put it more precisely as a result of the local irradiance 5a of the projection ray, the local irradiance being illustrated for a dipole illumination in FIG. 2, a location-dependently varying temperature distribution T(x, y)=T(x,y)T.sub.ref arises at the optical surface 31a, the temperature distribution being illustrated in a plan view and respectively in a sectional illustration along the X-direction of an XYZ coordinate system in FIGS. 3a,b. In order to simplify the illustration, a planar optical surface 31a was assumed in this case, but it goes without saying that the optical surface 31a typically has an (e.g. spherical) curvature.

(16) It goes without saying that, as an alternative or in addition to the Peltier elements 33a, it is also possible to provide other devices for regulating the temperature of the substrate 32 and/or of the optical surface 31a, for example heating wires. Moreover, temperature regulation can be effected by applying thermal radiation to the optical surface 31a. The thermal radiation can be generated e.g. by infrared radiation-emitting diodes or with the aid of IR lasers which are arranged at a distance from the optical surface. The IR radiation can be guided if appropriate with the aid of optical fibres or light guiding rods to the optical surface 31a and/or into the vicinity of the substrate 32. In this case, the thermal radiation can be introduced into the substrate 32 from below (from the carrier 33) but it is also possible, if appropriate, to radiate the thermal radiation onto the optical surface 31a directly from outside (from a location outside the projection ray 5).

(17) The location-dependent temperature distribution T(x, y) at the surface of the mirror 21 is related to the reference temperature T.sub.ref (which is constant over the surface), which, in the present example, is the ambient temperature of the mirror 21, corresponding to room temperature (T.sub.ref=22 C.). The temperature distribution T(x, y) can be represented as the sum of a value averaged over the surface <T(x, y)>=T.sub.av=<T>=const. (typically obtained by integration of the temperature distribution T(x, y) over all locations of the surface 31a, divided by the total area) and a (location-dependent) deviation T(x, y) from the average value T.sub.av:
T(x,y)=T.sub.av+T(x,y)=<T>+T(x,y).

(18) In this case, the value of the deviation T(x, y) that is averaged over the surface vanishes by definition, that is to say that <T>=0 holds true.

(19) Ideally, the zero crossing temperature T.sub.ZC=T.sub.ZCT.sub.ref related to the reference temperature T.sub.ref is constant over the substrate volume and thus over the reflective surface 31a. A power series expansion of a surface deformation D(x, y) resulting from the location-dependent variation of the temperature distribution T(x, y) depending on the temperature deviation T(x, y) and the average value <T> yields:
D(x,y)=D.sub.hom+(<T>T.sub.ZC)T(x,y)+T.sup.2(x,y),(1)
wherein denotes the (constant) gradient of the coefficient of thermal expansion at the zero crossing temperature T.sub.ZC.

(20) The homogeneous thermal expansion D.sub.hom of the surface of the mirror can typically be corrected well (e.g. with the aid of manipulators), and so this is not discussed in any greater detail here. Initially it appears plausible that the optimum average temperature <T> for the operation of the mirror 21 corresponds to the zero crossing temperature T.sub.ZC since the linear term in Equation (1) is omitted in this case.

(21) It is subsequently shown, however, that for minimizing the wavefront aberrations of the mirror (expressed by the RMS (root mean square) value in the present example) it is more advantageous in specific cases if the average value <T> of the temperature distribution at the mirror surface does not correspond to the zero crossing temperature T.sub.ZC. The RMS value (or the square thereof, also designated by RMS.sup.2) is dependent on the deformations D(x,y) at the reflective surface 31a as follows:
RMS.sup.2=<(D<D>).sup.2>=<D.sup.22D<D>+<D>.sup.2>=<D.sup.2>2<D.sup.2>+<D>.sup.2, i.e. RMS.sup.2=<D.sup.2><D>.sup.2(2)

(22) The RMS.sup.2 value is a measure of the deformation of the surface and corresponds to the variance of the distribution of the deformations D(x,y) at the surface, while the RMS value corresponds to the standard deviation.

(23) Via averaging, the following arises from Equation (1):
<D>=(<T>T.sub.ZC)<T>+<T.sup.2>=<T.sup.2>, i.e. the following holds true:
<D>.sup.2=1/4<T.sup.2>.sup.2

(24) Taking account of the homogeneous contribution D.sub.hom was omitted here and use was made of the fact that <T>=0 holds true (see above).

(25) Squaring Equation (1) and averaging yields:
<D.sup.2>=.sup.2(<T>T.sub.ZC)).sup.2<T.sup.2>+.sup.2(<T>T.sub.ZC))<T.sup.3>+.sup.2<T.sup.4>

(26) For the optimization (determination of the extreme value), the RMS value (or RMS.sup.2=<D.sup.2><D>.sup.2) is differentiated with respect to the zero crossing temperature T.sub.ZC and the result is set to be equal to zero. The following should hold true:
dRMS.sup.2/dT.sub.ZC=2.sup.2(<T>T.sub.ZC))<T.sup.2>.sup.2<T.sup.3>=0
Solving for the zero crossing temperature T.sub.ZC yields:
T.sub.ZC=<T>+<T.sup.3>/<T.sup.2>(3)

(27) The correction term <T.sup.3>/<T.sup.2> takes account of the asymmetry in the frequency distribution of the temperature values at the reflective surface 31a. If the temperature distribution is a (for example Gaussian) distribution that is symmetrical with respect to the average value <T>, the correction term vanishes since in this case <T.sup.3>=0 holds true for reasons of symmetry.

(28) In the case of EUV mirrors, however, the temperature distribution is generally highly asymmetrical, wherein it holds true, in particular, that |T.sub.MIN|<|T.sub.Max|, as is illustrated by way of example on the basis of a frequency distribution P(T) in FIG. 4. In the case of the distribution shown in FIG. 4, <T.sup.3> is greater than zero and the optimum zero crossing temperature T.sub.ZC is therefore above the average temperature <T>. Such an asymmetrical form of the frequency distribution in which the optimum zero crossing temperature T.sub.ZC is above the average temperature <T> is systematically provided when the average temperature <T> or T.sub.av is less than the average value (T.sub.max+T.sub.min) formed from the maximum temperature T.sub.max and the minimum temperature T.sub.min, cf. FIG. 3b.

(29) This condition can also be expressed on the basis of the location-dependent temperature distribution T(x, y) at the reflective surface 21a, such as is illustrated in FIG. 3a, and in which a first area portion A.sub.1 (illustrated in a hatched fashion), at which the temperature T(x, y) is greater than the average temperature T.sub.av, has a smaller surface area than a second area portion A.sub.2, at which the temperature T(x, y) is less than the average temperature T.sub.av, that is to say that A.sub.1<A.sub.2 holds true.

(30) The temperature distribution T(x, y) at the optical surface 31a, as illustrated in FIGS. 3a, b, substantially corresponds to the angular distribution of the illumination ray 4 upon entry into the projection lens 20, since the first EUV mirror 21 is arranged in proximity to a pupil plane 25, at which the location-dependent illuminance corresponds substantially (convolved with the diffraction structures on the mask M) with the field distribution in the pupil plane of the illumination system 10.

(31) The choice of a zero crossing temperature T.sub.ZC above the average temperature T.sub.av is therefore advantageous, in particular on optical elements 21 in proximity to the pupil, if the illumination system 10 generates an illumination ray 4 with an illumination pupil which has a pupil filling of less than 50%, preferably of less than 30%, particularly preferably of less than 15%, in particular of less than 1%, that is to say if only a corresponding area portion of the illumination pupil is illuminated. In this case, the condition A.sub.1<A.sub.2 is typically met at the optical surface 31a of an optical element in proximity to the pupil. This condition may, if appropriate, also be met at optical elements which are arranged in proximity to a field plane, if the local irradiance impinging there produces an asymmetrical temperature distribution in which the condition A.sub.1<A.sub.2 is met. In this case, depending on the degree of asymmetry of the temperature distribution, the zero crossing temperature T.sub.ZC can be chosen to be at least 0.1 K, if appropriate at least 0.2 K, in particular at least 0.4 K, greater than the average temperature T.sub.av at the optical surface 31a. Conversely, if A.sub.2<A.sub.1 (or A.sub.1=A.sub.2) at an optical element, then it is also possible, if appropriate, to choose an average temperature T.sub.av which is greater than (or equal to) the zero crossing temperature T.sub.ZC of the respective substrate.

(32) It goes without saying that besides optimization or minimization of the wavefront aberration at each individual mirror 21 to 24, it is also possible to effect an optimization of the aberrations of the entire projection lens 20, that is to say of wavefront aberrations or of image aberrations which are produced by the projection lens 20 in the image plane. For the optimization of the entire wavefront aberrations of the projection lens, at individual mirrors 21 to 24 it is also possible, if appropriate, to deviate from an average temperature T.sub.av, which minimizes the wavefront aberration at the respective mirror 21 to 24, provided that the wavefront aberration of the entire projection lens 20 is improved by this deviation. As a measure of the wavefront aberration in the image plane of the projection lens 20, as an alternative or in addition to the RMS value, it is possible to use other image aberrations, e.g. overlay, depth of focus, best focus, etc., or specific aberrations such as coma, astigmatism, etc. These wavefront aberrations can be measured or simulated in the aerial image, for example, and the dependence of these aberrations on the temperature-dictated deformations of the individual mirrors 21 to 24 can be determined. Via the temperature control device 30, on the basis of this known dependence it is possible to set a suitable difference between average temperature T.sub.av and zero crossing temperature T.sub.ZC at a respective mirror 21 to 24, in order to minimize the used measure of the wavefront aberration in the image plane.

(33) In order to set the desired average temperature T.sub.av even in the case of a temporally variable intensity of the projection ray 5 on the reflective surface 31a of the optical element 21 or of all the optical elements 21 to 24 of the projection lens 20 or in order to keep the average temperature T.sub.av constant, it is likewise possible to use the temperature control device 30 shown in FIG. 1, which serves for driving the heating device 33a (and/or further heating devices not shown) for the further mirrors 22 to 24). In order that the temperature of the substrate 32 can be controlled to the desired average temperature T.sub.av by closed-loop control, a temperature sensor 35 is provided laterally on the substrate 32 in the example shown in FIG. 2, the temperature sensor being connected to the control device 30 via a connecting line (not shown). Alternatively or additionally, one or more temperature sensors (not shown) can also be embedded into the substrate 32 or into the volume of the substrate 32 in order to detect the temperature at different locations below the optical surface 31a. In this case, the temperature sensors can be read via connecting lines that are led out from the substrate 32. If appropriate, the read-out can also be effected contactlessly via an optical interface or the like.

(34) Via the temperature control device 30, it is also possible to set a desired difference between the average temperature T.sub.av and the zero crossing temperature T.sub.ZC, the difference being dependent on the local irradiance 5a or on the respective application, wherein the difference can be determined in accordance with Equation (3), for example. In particular, it is possible to determine or simulate in advance the expected local irradiance or irradiation intensity 5a for specific operating parameters (e.g. dipole illumination, annular illumination, etc.) and to determine an appropriate difference T.sub.avT.sub.ZC on the basis of Equation (3), for example. The assignment between the operating parameters and the difference respectively to be chosen can be stored in the temperature control device 30 in order to be able to select or set the desired difference depending on the operating parameters. However, it is also possible, on the basis of a location-dependent temperature distribution measured or simulated during operation at the optical surface 31a, to determine or calculate the frequency distribution of the temperature at the surface 31a and also the effects thereof on the form of the surface 31a or the wavefront aberrations. This information can be used by the temperature control device 30 for setting the desired difference.

(35) The temperature control device 30 can in particular also be used in the transient case, that is to say directly after illumination radiation 4 has been applied to the projection lens 20, in which the average temperature T.sub.av and also the maximum and minimum temperatures T.sub.max, T.sub.min at the surface 31a of the mirror 21 (without additional heating) vary in a time-dependent manner until reaching a steady-state temperature state having a constant average temperature T.sub.av,s and respectively constant maximum and minimum temperatures T.sub.max,s, T.sub.min,s cf. FIG. 5.

(36) In order to reach the steady-state temperature state even without irradiation, the temperature control device 30 can be used for heating the mirrors 21 to 24 to the desired (steady-state) average temperature T.sub.av,s before illumination radiation is applied to the projection lens 20. In this case, the heating power of the heating device 33a in the transient case, in which the illumination radiation is additionally absorbed by the respective substrate 32, should be adapted such that the total thermal power (sum of the radiation power and heating power) taken up by the substrate and thus the average temperature of the substrate 32 and/or of the optical surface 31a remain constant, that is to say that the heating power should be gradually reduced in the transient case in order to keep the average temperature T.sub.av as constant as possible.

(37) As explained above, it is possible to minimize wavefront aberrations at the individual mirrors 21 to 24 or of the entire projection lens 20 by the average temperature T.sub.av of the mirror surface and the zero crossing temperature T.sub.ZC being suitably adapted to one another. As explained above, such an adaptation can be effected with the aid of a temperature control device 30 which sets the average temperature T.sub.av at the respective reflective surface 31a in a suitable manner. However, the zero crossing temperature T.sub.ZC can be set only during the production or configuration of the projection lens 20, but not during operation. For configuring the projection lens 20 or for choosing a suitable zero crossing temperature T.sub.ZC of the substrate material 32 of a respective mirror 21 to 24, the following procedure can be adopted:

(38) Firstly, the local irradiance 5a to be expected in the operating case at the optical surface 31a of the mirror 21 is determined, for which purpose a computer simulation of the illumination ray 4 entering into the projection lens 20 of the projection ray 5 is typically used. The temperature distribution at the optical surface 31a is then determined from the locally varying irradiance 5a, for which purpose, typically all heat conduction mechanisms (heat transfer into the substrate 32, but also thermal radiation at the surface 31a and heat emission to the residual gas) are taken into consideration. If it holds true for the temperature distribution T(x, y) that A.sub.1<A.sub.2 or that the average temperature T.sub.av is below the average value (T.sub.max+T.sub.min) formed from the minimum and maximum temperatures T.sub.min, T.sub.max, the material of the substrate is chosen such that the zero crossing temperature T.sub.ZC thereof is greater than the average temperature T.sub.av corresponding to the expected operating temperature produced by the irradiation.

(39) The average temperature T.sub.av in the operating case is the static temperature T.sub.av,s established after the heating of the mirrors 21 to 24. It goes without saying however, that the average temperature T.sub.av varies in a time-dependent manner in the transient case (cf. FIG. 5) if the temperature control device 30, in the transient case, does not ensure that the average temperature T.sub.av(t) at the optical surface 31a remains substantially constant. For the average temperature T.sub.av(t) present at a respective point in time t, a likewise varying optimum (that is to say aberration-minimizing) zero crossing temperature T.sub.ZC(t) arises in this case. Since the zero crossing temperature T.sub.ZC is finally defined after the production of the substrate 32, it is necessary to make a selection from the different zero crossing temperatures T.sub.ZC(t) in the transient case. This selection can be made for example in such a way that the zero crossing temperature T.sub.ZC(t) is selected at a point in time t* at which the resulting wavefront aberration is the greatest. This makes it possible to ensure that the wavefront aberrations are still within the specification even under the worst-case conditions.

(40) By way of example, for checking whether the RMS value of the wavefront aberrations fulfils the specification, it is possible to compare the RMS value with a suitable threshold value, e.g. with a fraction (e.g. 1/14, see above) of the wavelength of the EUV radiation, in order to ensure that the projection lens 20 is diffraction-limited. Of course, it is also possible to compare other measures of the wavefront aberration of the projection lens 20 with corresponding threshold values in order to determine whether they fulfil the specification.