System correction from long timescales

Abstract

Aberrations of a projection lens for microlithography can be subdivided into two classes: a first class of aberrations, which are distinguished by virtue of the fact that their future size increases by a non-negligible value after a constant time duration, independently of their current size, and a second class of aberrations, which, after reaching a threshold, only increase by a negligible value after each further time duration. An adjustment method is proposed, which adjusts these two classes of aberrations in parallel in time with one another.

Claims

1. A method of adjusting a microlithographic projection exposure apparatus comprising a projection lens, the method comprising: performing a first partial adjustment which comprises: prescribing a first temporal adjustment duration t.sub.1; establishing a first aberration b.sub.1 of the projection lens; adjusting b.sub.1 during t.sub.1; and performing a second partial adjustment which comprises: prescribing a second temporal adjustment duration t.sub.2, establishing a second aberration b.sub.2 of the projection lens, adjusting b.sub.2 during t.sub.2, wherein: t.sub.1<t.sub.2; b.sub.1 changes more quickly than b.sub.2 during t.sub.2; b.sub.1 and b.sub.2 are scalar aberrations; b.sub.1 has a lower waviness than b.sub.2; and the first partial adjustment is performed in parallel with the second partial adjustment.

2. The method of claim 1, comprising performing the first partial adjustment repeatedly while performing the second partial adjustment.

3. The method of claim 2, wherein: the projection exposure apparatus comprises a plurality of manipulators comprising one or more manipulators of a first class of manipulators and one or more manipulators of a second class of manipulators; the one or more manipulators in the first class of manipulators are quicker manipulators than the one or more manipulators in the second class of manipulators; and the method comprises: using the one or more manipulators of the first class of manipulators during the first partial adjustment; and using one or more manipulators of a class of the second manipulators of the manipulators during the second partial adjustment.

4. The method of claim 3, comprising using equivalent manipulators in terms of design during the first and second partial adjustments.

5. The method of claim 4, comprising using equivalent thermal manipulators in terms of design during the first and second partial adjustments.

6. The method of claim 5, comprising using a manipulator with a greater accuracy setting during the second partial adjustment than during the first partial adjustment.

7. The method of claim 1, wherein t.sub.1<100 milliseconds, and t.sub.2 is at least twice t.sub.1.

8. The method of claim 1, comprising: using a first algorithm to determine manipulator travel for the first partial adjustment; and using a second algorithm to determine manipulator travel for the second partial adjustment, wherein the first algorithm has a quicker run time than the second algorithm.

9. The method of claim 1, comprising: using a first algorithm to determine manipulator travel for the first partial adjustment; and using a second algorithm to determine manipulator travel for the second partial adjustment, wherein the first algorithm has a lower accuracy than the second algorithm.

10. The method of claim 1, using a measurement to establish at least one aberration selected from the group consisting of b.sub.1 and b.sub.2.

11. The method of claim 1, using a prediction based on a prediction model to establish at least one aberration selected from the group consisting of b.sub.1 and b.sub.2.

12. The method of claim 11, comprising adjusting b.sub.1 repeatedly while adjusting b.sub.2.

13. The method of claim 12, wherein: the projection exposure apparatus comprises a plurality of manipulators comprising one or more manipulators of a first class of manipulators and one or more manipulators of a second class of manipulators; the one or more manipulators in the first class of manipulators are quicker manipulators than the one or more manipulators in the second class of manipulators; and the method comprises: using the one or more manipulators of the first class of manipulators while adjusting b.sub.1; and using one or more manipulators of a class of the second manipulators of the manipulators while adjusting b.sub.2.

14. The method of claim 13, wherein t.sub.1<100 milliseconds, and t.sub.2 is at least twice t.sub.1.

15. A method, comprising: adjusting a first aberration b.sub.1 of a projection lens of a microlithographic projection exposure apparatus during a first temporal adjustment duration t.sub.1; and simultaneously adjusting a second aberration b.sub.2 of the projection lens of the microlithographic projection exposure apparatus during a second temporal adjustment duration t.sub.2, wherein: t.sub.1<t.sub.2; during t.sub.2, b.sub.1 changes more quickly than b.sub.2; the method comprises: using a first algorithm to determine manipulator travel for adjusting b.sub.1; and using a second algorithm to determine manipulator travel for adjusting b.sub.2; and the first algorithm has a quicker run time than the second algorithm, and/or the first algorithm has a lower accuracy than the second algorithm.

16. The method of claim 15, using a measurement to establish at least one aberration selected from the group consisting of b.sub.1 and b.sub.2, and/or using a prediction based on a prediction model to establish at least one aberration selected from the group consisting of b.sub.1 and b.sub.2.

17. A method of adjusting a microlithographic projection exposure apparatus comprising a projection lens, the method comprising: performing a first partial adjustment which comprises: prescribing a first temporal adjustment duration t.sub.1; establishing a first aberration b.sub.1 of the projection lens; adjusting b.sub.1 during t.sub.1; and performing a second partial adjustment which comprises: prescribing a second temporal adjustment duration t.sub.2, establishing a second aberration b.sub.2 of the projection lens, adjusting b.sub.2 during t.sub.2, wherein: t.sub.1<t.sub.2; b.sub.1 changes more quickly than b.sub.2 during t.sub.2; the first partial adjustment is performed in parallel with the second partial adjustment; and the method comprises using equivalent manipulators in terms of design during the first and second partial adjustments.

18. A method of adjusting a microlithographic projection exposure apparatus comprising a projection lens, the method comprising: performing a first partial adjustment which comprises: prescribing a first temporal adjustment duration t.sub.1; establishing a first aberration b.sub.1 of the projection lens; adjusting b.sub.1 during t.sub.1; and performing a second partial adjustment which comprises: prescribing a second temporal adjustment duration t.sub.2, establishing a second aberration b.sub.2 of the projection lens, adjusting b.sub.2 during t.sub.2, wherein: t.sub.1<t.sub.2; b.sub.1 changes more quickly than b.sub.2 during t.sub.2; the first partial adjustment is performed in parallel with the second partial adjustment; the method comprises: using a first algorithm to determine manipulator travel for the first partial adjustment; and using a second algorithm to determine manipulator travel for the second partial adjustment; and the first algorithm has a quicker run time than the second algorithm, and/or the first algorithm has a lower accuracy than the second algorithm.

19. The method of claim 18, wherein the first algorithm has a quicker run time than the second algorithm.

20. The method of claim 18, wherein the first algorithm has a lower accuracy than the second algorithm.

Description

(1) FIG. 1 shows the temporal behavior of various aberrations

(2) FIG. 2 shows the one assignment of manipulators and adjustment times

(3) FIG. 3a-d show explanations of relatively quick and relatively slow, thermal manipulators

(4) FIG. 4a-c show subdivisions into relatively quick and relatively slow manipulators

(5) FIG. 5 shows a lens with the regulation unit for the parallel first and second partial adjustment

(6) FIG. 6 shows a flowchart of the parallel first partial adjustment and second partial adjustment

(7) FIG. 1 illustrates the typical time profile of some aberrations, which are generated by the heating experienced by the optical elements of the lens as a result of the absorption of projection light. The lens is a catadioptric lens, designed for scanning operation, with an extent of the image field of 25 mm orthogonal to the scanning direction. The temporal aberration profiles for three Zernike coefficients, Z5, Z12 and Z21, are compared; cf. DE102008042356A1 and DE102004035595A1 for the definition of the Zernike coefficients. The three Zernike coefficients in the three illustrated rows are three astigmatic terms with (radial) waviness R of second, fourth and sixth order, increasing from Z5 to Z21, and constant azimuthal waviness θ of second order, wherein (R, θ) are the normalized pupil coordinates. To be more precise: Z5=√{square root over (6)}R.sup.2 cos 2θ, Z12=√{square root over (10)}(4R.sup.2−3)R.sup.2 cos 2θ, Z21=√{square root over (14)}(15R.sup.4−20R.sup.2+6)R.sup.2 cos 2θ. The diagrams, respectively situated five in a row, correspond to a subdivision into five equidistantly distributed field points orthogonal to the scanning direction.

(8) The general definition of Zernike polynomials and their coefficients can be taken from Herbert Gross, ed., Vol. 1: Fundamentals of Technical Optics, Wiley, 2005. Each Zernike polynomial p depends from an radial argument r and an azimuthal argument φ and is given by a product p(r, φ)=p.sub.1(r).Math.p.sub.2(φ) consisting of a radial factor p.sub.1 which is a polynomial and an azimuthal factor p.sub.2 which is a trigonometric polynomial. Both polynomials, respectively, depending from the radial argument r and azimuthal argument φ only. The radial and azimuthal waviness of the Zernike polynomial p is given by the polynomial order of the radial factor p.sub.1 and the polynomial order of the azimuthal factor p.sub.2, respectively. A Zernike polynomial p is said to have higher radial waviness than a Zernike polynomial q provided that the radial factor of p has a higher polynomial order than the radial factor of q. A Zernike polynomial p is said to have higher azimuthal waviness than a Zernike polynomial q provided that the azimuthal factor of p has a higher polynomial order than the azimuthal factor of q. A Zernike polynomial p is said to have higher waviness than a Zernike polynomial q provided that p has a higher radial waviness than q and the azimuthal waviness of q is not greater than that of p or p has a higher azimuthal waviness than q has and the radial waviness of q is not greater than that of p.

(9) The respective individual diagrams show the temporal developments of the corresponding Zernike coefficient at the relevant field point of the image field. Respectively one simulation value and one measured value are depicted for two different projection lenses. In each image, FIG. 1 respectively shows five measured profiles from five different lenses.

(10) The abscissa is plotted in seconds and the ordinate is plotted in nanometers. The function values correspond to the amplitude (or synonymously the value) of the corresponding Zernike coefficient in the series expansion of the deviation of the wavefront from a spherical wave according to Zernike polynomials as illustrated in DE102008042356A1 and DE102004035595A1.

(11) Thus, for example, the first diagram in the first row shows the temporal development of the aberration Z5 at the outermost left-hand field edge. The loading case is an actuation of the projection lens with projection light an operating wavelength of 193 nanometers, the setting of a 35° Y-dipole and with a ratio of sigma-in to sigma-out of 0.7-0.9. In approximately the first 1800 seconds, the amplitude of initial value defined as a zero point increases like a root function up to an amplitude of approximately 25 nm. The actuation with projection light ends after these 1800 seconds and the optical elements cool down again, leading to a decay of the amplitude.

(12) It is possible to gather from the first row that the behavior of Z5 is relatively independent of the observed field point.

(13) If the aberrations in the second and then in the third row of FIG. 1 are observed, the temporal developments are likewise seen to be approximately independent of the observed field point. However, the shape of the increase (in terms of absolute value) is less steep and transitions more quickly into saturation with ultimately smaller amplitude (in terms of absolute value).

(14) In general, it is possible to determine that the aberrations or Zernike coefficients with a higher (radial) waviness, such as e.g. Z21, exhibit, compared to an aberration with a lower (radial) waviness, such as e.g. Z5, a behavior of, in relative terms, lower amplitudes (in terms of absolute value) and of reaching saturation more quickly.

(15) If, within the scope of operating a projection exposure apparatus, a comparison is made between the period of time of actuating the lens with projection light and the heating accompanying this and the time without actuating the lens with projection light and the cooling accompanying this, then this ratio is significantly greater than one. Thus, in general, it is the case in all three aberrations Z5, Z12 and Z21 that the amplitudes thereof have grown so far after a certain amount of time that, without an adjustment, the lens would depart from the specification.

(16) In general, such a specification is provided by upper limits for aberrations such as the three aforementioned Zernike coefficients Z5, Z12, Z21. Such an adjustment is therefore prescribed to be repeating and the repetition is prescribed with a high frequency. In this respect, cf. DE102008042356A1. Here, the available adjustment time can be several seconds to a few milliseconds. The predetermined frequency for the adjustment is set by the fastest-growing aberration. This is Z5 in FIG. 1. From the view of this frequency predetermined thereby, (radial) higher-order aberrations, such as e.g. Z21, exhibit the behavior of an initial strong increase with subsequently negligible growth in relation to the specification. An aberration such as Z21 so to speak changes its behavior.

(17) The aberration Z5 will therefore be adjusted in each adjustment cycle since it continuously threatens to depart from the specification. Initially, the aberration Z21 will likewise have to be adjusted with the same frequency or a frequency of the same order of magnitude as the aberration Z5. However, after approximately 100 seconds, the growth in the amplitude of Z21 is so small that this growth remains within the scope of the specification. Then, as seen relative to Z5, the aberration Z21 becomes a slowly changing aberration.

(18) The aberrations can therefore generally be divided into two classes, defined relative to one another, of quickly changing and slowly changing aberrations. This subdivision can change in respect of time and with respect to the maximum predetermined adjustment time. By way of example, the classification of the aberration Z12 depends precisely on these two parameters and the predetermined specifications.

(19) This different behavior of the aberration growth now results in the following problem: if use is always made of the same adjustment concept with a high frequency, the class of slowly changing aberrations will be largely ignored by the adjustment algorithm for such a long time until the amplitudes thereof threaten to violate a specification. In such a situation, the predetermined adjustment algorithm is then forced, within its predetermined, possibly very short maximum adjustment time, to also adjust the slowly changing aberration Z21, which has now departed from the specification, in addition to the quickly changing aberration Z5. In general, an adjustment algorithm addresses all errors at the same time. However, if the relatively quickly changing Z5 is once again dominant in the meantime, the adjustment may continue to ignore the aberration Z21. However, the problem is that, in particular, the (radial) higher-order Zernike coefficients resist a fast adjustment. This will now be explained.

(20) The manipulation options, mentioned at the outset, of the optical elements of the projection exposure apparatus and, in particular, of the lens can likewise be subdivided into classes.

(21) FIG. 2 shows such a class subdivision using an illustrative selection of manipulator types. Manipulators which displace an optical element of the projection lens relative to other optical elements of the projection lens can generally be employed for adjustment times of 100 milliseconds. Depending on the actuation of such a displacing manipulator, it is also possible to achieve adjustment times of 10 milliseconds. However, this is generally to the detriment of the accuracy of the deflection of the manipulator or to the detriment of the possible deflection of the manipulator relative to the accuracy. Thus, it is still possible to use a control loop in the case of 100 milliseconds, but is dependent on pure control in the case of 10 milliseconds. Thus, a displacing manipulator can satisfy different maximum adjustment times while having the same design, dependent on its closed-loop or open-loop control.

(22) An Alvarez manipulator, as shown in EP851305B1, can be deflected within an adjustment time of 100 milliseconds. However, if the plate pair of the Alvarez manipulator should be replaced, cf. e.g. DE102007058158A1, this may require an adjustment time of 100 seconds. Such an Alvarez manipulator can therefore likewise be categorized into two classes: a first class, in which the Alvarez manipulator can quickly be deflected and has one degree of freedom, which is given by the relative position of its two plates in relation to one another, and a second class, in which the plate pair of the Alvarez manipulator is replaced by an alternative pair and this alternative pair is then deflected relative to its parts. In the first class, the Alvarez manipulator has one degree of freedom; in the second class, the Alvarez manipulator has as many degrees of freedom as there are plate pairs available for replacement.

(23) Such an Alvarez manipulator can also, in accordance with the demands of a customer, be provided with a plate pair yet to be produced (customization). This then requires an adjustment time of several days. The degrees of freedom of such customization are then merely determined by the position of the Alvarez manipulator in the projection lens.

(24) FIGS. 3a-3d explain the temporal behavior of a thermal manipulator as described in WO2009026970A1. This is a plane plate (2) pervaded by wires (1) in the horizontal and vertical direction, which wires are numbered from 1 to 16 and numbered from R1 to R16, respectively. A resistance density is applied to the wires and the latter can be actuated with different voltages at the ends (3) thereof. The optical effect is brought about by the local change in the refractive index of the plane plate accompanying the temperature change. In order to be able to set a zero deflection of this thermal manipulator, the latter is provided with a heatsink, such as e.g. a continuously cooling airflow. The latter is not illustrated here.

(25) FIGS. 3a-3d show the plane plate from the direction in which the projection light passes through the plane plate.

(26) For simplifying the illustration, the assumption is made here and shown that the heat actuation of each individual one of the illustrated pixels can be addressed in isolation from the others.

(27) FIG. 3a shows the temperature distribution in the plate one second after a constant, continuous heat actuation of the pixels (4) with the coordinates [5, 12]×[R5, R12]. What is shown here are an inner region and an outer region with pixels in each case having a common temperature, and, compared thereto, a very narrow intermediate region with a temperature gradient.

(28) FIG. 3b shows the equilibrium state for the same actuation, which is achieved after approximately five seconds. The temperature in the inner region of the plane plate has remained unchanged, the outer region with the temperature gradient has become wider, but it will no longer change as long as the heat actuation remains unchanged. Overall, in the equilibrium state, the same amount of heat dissipates from the whole plane plate (2) as is added by the pixels (4) actuated by heat.

(29) FIGS. 3c and 3d show the behavior if the heat actuation is selected in such a way that an individual pixel (5) with the coordinates [5, 6]×[R5, R6] now experiences a different heat actuation compared to its neighbors.

(30) Since the change in the refractive index of the plane plate (2) in a small region depends strictly monotonically on the change in the temperature, the temperature differences of the plane plate (2) for small temperature differences can be seen as being proportional to a phase error of the wavefront. Hence, the equilibrium state achievable as per FIG. 3b can be considered to be a first manipulator degree of freedom of a thermal manipulator and the equilibrium state achievable as per FIG. 3d can be considered to be a second manipulator degree of freedom of this thermal manipulator.

(31) If an adjustment time of five seconds or more is scheduled, this thermal manipulator has the degrees of freedom corresponding to FIGS. 3b and 3d.

(32) However, if an adjustment time of e.g. at most one second is scheduled, the degree of freedom in accordance with FIG. 3d cannot be achieved. By contrast, the temperature difference between FIGS. 3a and 3b appears to be qualitatively relatively low and one is inclined to tolerate the corresponding phase difference and to consider the degree of freedom in accordance with FIG. 3b to belong to this thermal manipulator with its adjustment time of one second. Quantitatively, this initially qualitative point of view is expressed by virtue of the fact that the ratio of the area of pixels with constant temperature to the area of pixels which experience a temperature change is significantly lower in the case of FIGS. 3a and 3b than in the case of FIGS. 3c and 3d.

(33) The actuation of a smaller region of the plane plate (2) with heat generally results in a higher waviness of the corresponding phase error.

(34) It is therefore possible to say that a thermal manipulator belongs to two classes: a class of relatively quick manipulators if the manipulator should bring about wavefront changes with low waviness and a class of relatively slow manipulators if the manipulator should additionally bring about wavefront changes with high waviness.

(35) In FIGS. 4a-4c, the above-discussed classifications of the aberrations and manipulators are carried out in an exemplary fashion as partial adjustments carried out in parallel. FIGS. 4a-4b illustrate a first partial adjustment and a second partial adjustment carried out in parallel thereto.

(36) The first partial adjustment in accordance with FIG. 4a should address those aberrations b.sub.1 which threaten to depart from the specification within 1 s as a result of the heating of the lens connected to the actuation with projection light. By way of example, for a specification of 5 nm, this is the astigmatism b.sub.1=Z5, which leads to an aberration Z5 of more than 5 nm directly with the start of the actuation. In this respect, cf. once again FIG. 1. Hence, an adjustment time of t.sub.1=1 s is provided in this first partial adjustment. The manipulators utilized in the first partial adjustment are highlighted in FIG. 4a via a bold type. These are slowly displacing manipulators, changes in the setting, changes in the wavelength of the projection light, deforming manipulators, the deflection of an Alvarez manipulator, low-resolution, thermal manipulators and a change in the pressure conditions in the lens.

(37) The second partial adjustment in accordance with FIG. 4b should address those aberrations b.sub.2 which threaten to depart from the specification within 100 s as a result of the heating of the lens connected to the actuation with projection light. By way of example, for a specification of 5 nm, this is the astigmatism b.sub.2=Z12, which threatens to exceed an amplitude of 5 nm after approximately 100 s. In this respect, cf. once again FIG. 1. Hence, an adjustment time of t.sub.2=100 s is provided in this second partial adjustment. The utilized manipulators are slowly displacing manipulators, changes in the setting, a change in the wavelength of the projection light, deforming manipulators, the deflection of an Alvarez manipulator, high-resolution, thermal manipulators, a change in the pressure conditions in the lens and a change of the stop of the lens.

(38) As a second example, FIGS. 4b-4c now illustrate a first partial adjustment and a second partial adjustment carried out in parallel thereto.

(39) The first partial adjustment corresponds to the aforementioned second partial adjustment and addresses the corresponding aberrations and specifications. The second partial adjustment now addresses aberrations with a very high waviness, which generally can no longer be compensated for by manipulators which are fixedly installed into the projection lens. By way of example, this includes the compaction of optical materials such as glass or layers, mentioned at the outset, if these are impinged upon by projection light. This impingement can be of a very local nature if, for example, a so-called free-form setting is employed. In this respect, cf. WO09080279A1 or Illumination Optics for Source-Mask Optimization, Mizunoi, et al., Nikon Corp., 201-9 Miizugahara, Kumagaya, Saitama, Japan 360-8559. In this case, the second partial adjustment contains a replacement or post-processing of an optical element of the projection lens such as e.g. the stop, a lens element, a mirror or one or more aspherized plane plates. At the time of such replacement, a parallel first partial adjustment no longer makes sense. The parallel nature of first and second partial adjustment in this second example should be understood in such a way that the slowly growing aberration b.sub.2 is compensated for by the continuously operating manipulators from FIG. 4c. At first glance, there does not appear to be a difference from the manipulators printed in bold type in FIG. 4b, but one should consider that the predetermined adjustment times are different and different aberrations are compensated for. Thus, for example, the manipulator which influences the setting will, in the partial adjustment in accordance with FIG. 4b, compensate for an aberration, caused by heating of optical elements near the pupil, with a relatively high intensity and a low resolution. However, the same manipulator will, in the partial adjustment in accordance with FIG. 4c, address an aberration caused by layer degradation with a relatively low intensity and a high resolution.

(40) FIG. 5 has in part been taken from FIG. 1 of DE102008042356A1.

(41) FIG. 5 shows an exemplary embodiment of a projection exposure apparatus 100 for microlithography for imaging an object field 101 onto an image field 102. The projection apparatus 100 contains a lens 110. Two field points 103 and 104 situated in the object field, which are imaged in the image plane 102 by the lens, are illustrated in an exemplary manner. The beam path is delimited by a stop, which is situated in the vicinity of a pupil plane of the lens.

(42) The lens contains optical elements such as lens elements 111, mirrors 112 and plane plates 113. A manipulator 121 acts on one of the lens elements, which manipulator can displace, bend, heat and/or cool the lens elements or can locally heat a plane plate, as discussed above, with a high or low resolution. A second manipulator 122 acts on the mirror 112 in the same manner and a third manipulator 123 serves for replacing an individual plane plate (not illustrated here) with a further individual plane plate (likewise not illustrated here), which is aspherized. Alternatively, this third manipulator can be an Alvarez manipulator 123 with replaceable plate pairs 113, or a manipulator which changes stops or manipulates the stop shape or the stop diameter.

(43) Light beams which are subsequently restricted by the stop are emitted by the two field points 103 and 104. The outermost rays of which light beams are illustrated here in a dashed manner. These outermost rays delimit the wavefronts respectively belonging to the field points 103 and 104. For the purposes of illustrating the invention, these wavefronts are assumed to be spherical. A wavefront sensor and/or further sensors and/or a prediction model form(s) a determination unit 150, which supplies information in respect of aberrations established from the determined wavefronts. These further sensors are e.g. air pressure sensors, sensors for measuring the temperature in the lens or sensors which measure the temperature on lens elements or on the rear side of mirrors.

(44) The manipulators 121, 122, 123 can be controlled by a regulation unit 130.

(45) The regulation unit 130 can be used to obtain upper limits and manipulator specifications from a memory 140 and also information in respect of the measured aberrations or wavefronts from the determination unit 150.

(46) In contrast to DE102008042356A1, the regulation unit 130 contains a first adjustment unit 132 and a second adjustment unit 133, which regulate or control the first and the second partial adjustment, respectively, and a lookup table 131, which provides information in respect of the subdivision of the determined aberrations into relatively slow and relatively quick aberrations and which contains information in respect of the relatively slow manipulators to be used for the first partial adjustment and of the relatively quick manipulators to be used for the second partial adjustment.

(47) FIG. 6 illustrates the principle of the parallel first partial adjustment and second partial adjustment.

(48) The wavefront errors or aberrations are established by the determination unit 150 in a first step 201. They are subsequently, via a lookup table 131, subdivided in a second step 202 into two classes of quickly changing aberrations b.sub.1 and slowly changing aberrations b.sub.2 respectively relative to one another. In parallel with this, the manipulators are subdivided into the classes of relatively slow and relatively quick manipulators according to a further lookup table, which is likewise denoted by 131. The next steps 203 and 204 are carried out in parallel: step 203 denotes the first partial adjustment, in which the relatively quickly changing aberrations b.sub.1 are adjusted by the relatively quick manipulators, and step 204 denotes the second partial adjustment, in which the relatively slowly changing aberrations b.sub.2 are adjusted by the relatively slow manipulators. Here, step 203 is carried out repeatedly, wherein, after each run through, the quickly changing aberrations are re-measured in an additional step 205 by the determination unit 150 or, for reasons of a tightly allocated first adjustment time t.sub.1, predicted by the determination unit 150. By way of example, this prediction can be undertaken by a model which is calibrated automatically.

(49) The totality of steps 201-205 can likewise be carried out repeatedly once the second partial adjustment has been completed after the second adjustment time t.sub.2.