DE102006057606A1 - Test sample's e.g. concave test sample, optically smooth surface e.g. free form surface, laminar measurement method, involves filtering object waves, which are reflected from surface, by aperture stop in Fourier plane of imaging lens - Google Patents

Abstract

The method involves illuminating an optically smooth surface (33) with discrete object waves (32). The object waves, which are reflected by the optically smooth surface, are super imposed with a reference wave (42) on a detector (38) for an interferogram. The surface is illuminated synchronously with multiple switchable object waves. The object waves, which are reflected from the surface, are filtered by an aperture stop (40) that is arranged in an optical path before the detector, where the aperture stop is provided in a Fourier plane of an imaging lens (36). An independent claim is also included for a measuring device for measuring an optically smooth surface.

Description

The The present invention relates to a method according to the preamble of claim 1 and a measuring device according to the preamble of claim 21. It is under a visually smooth surface a mirroring reflective surface understood.

Such an interferometric method and such a measuring device has already been developed at the Institute for Technical Optics of the University of Stuttgart and in the DE 103 25 601 B3 been revealed.

to Measurement of optical surfaces Interferometry is an accurate and fast method at. Interferometry uses the wave property of light. In the overlay a known reference wave and one of the surface of the Test specimens usually deformed object wave arise areas with extinction and Areas with light amplification. The resulting intensity image, the interferogram, contains the information about the deviation of the test object from the nominal shape and can be evaluated.

give way however, object wave and reference wave are strongly different from each other, so are the interference fringes are so dense that they no longer reach the detector disbanded can be. The detector is usually an imaging sensor, e.g. a camera.

The Interferometry is used, for example, for the measurement of optical Components such as lenses. The majority Contains today's optics exclusively spheres, so Lenses whose surfaces have a spherical shape. The interferometric test spherical Lenses is unproblematic and standardized. In her survey avoids excessive deviations between the object wave and the reference wave by the use of a so-called null lens, which is an initially flat Measuring wave forms into a spherical wave, which then ideally perpendicular on the sphere to be measured falls and is reflected by it as an object wave, so that in the return through the Null lens again, a plane wave is created, which as an object wave is superimposed with the also flat reference wave.

By the use of aspheres, So lenses whose surfaces deviate from the pure spherical shape, optics are more powerful and can be more compact be built. Such optics penetrate the market in areas like high performance lenses for the wafer exposure, but also lenses and photo optics more and more.

When requirement for an exact production of aspheres applies the exact measurement of their surface. In principle, aspheres can be similar to spheres with Help a zero lens, or more generally, a zero optics, measured become.

in the Contrary to the measurement of spherical surfaces However, the interferometric measurement of aspheric surfaces a task that has not been satisfactorily solved in many areas the optical metrology. The origin of the problems occurring lies in the fact that zero tests on aspheres always the production of special, to the asphere adapted refractive or diffractive optics as zero lenses require.

Nucleaves for aspheres are already manufactured by default diffractive and refractive, whereby a diffractive structure is used to reshape the spherical object wave of an interferometer objective so that a wave adapted to the specimen is formed. comparisons HJ Tiziani, S. Reichelt, C. Pruss, M. Rocktäschel, U. Hofbauer, "Testing of Aspheric Surfaces", Proc. SPIE, 4440, 109-119, 2001 ,

There aspheres are produced in an enormous variety of forms, is the production adapted zero lenses but with a great deal of time and expense connected.

It is also known, a for the test more spherical Lens used lens by a so-called. Computer generated hologram (CGH). A CGH is a diffractive optical element that matches the adaptation the measuring shaft to an aspherical examinee allows so that here too, the measuring shaft is perpendicular to the test object. Again, the preparation of such CGHs is time consuming and is expensive, so this way is efficient only at high volumes.

Tries, aspheres without interferometry measurement without zero optics in the non-zero test, often fail due to the occurrence of interferograms with too high Streak densities that are no longer resolved by the camera used. Furthermore, standard calibration methods are not sufficient. Often they leave on the surface of the test piece reflected rays even the optical system.

In the publications YM Liu, GN Lawrence, CL Koliopoulos, "Subaperture Testing of Aspheric with Anular Zone", Applied Optics, 27, 4504-4513, 1988 and M. Melozzi, L. Pezzati, A. Mazzoni, "Testing aspheric surfaces using multiple annular interferograms", Optical Engineering, 32, 1073-1079, 1993 , Measuring devices are proposed in which rotationally symmetric aspheres are measured stepwise. Either the asphere or the interferometer lens is moved along the optical axis. It will be in everyone Measurement of each ring evaluable on the asphere, on which the rays impinge approximately vertically. Finally, the individual areas must be combined to form an overall measurement taking into account propagation effects.

at the measurement of individual ring zones by movements of the asphere or However, the interferometer optics is limited to rotationally symmetric aspheres. In addition, prove here mechanical inaccuracies as problematic.

For reliable measurements in the non-zero test, ways must be found to characterize the interferometer and its deviations from the desired configuration. In this context describe JE Greivenkamp and RO Gappinger in the paper "Design of a nonnull interferometer for aspheric wavefronts", Applied Optics, 43, 5143-5151, 2004 , a method for characterizing a Mach-Zehnder interferometer using reverse engineering. Measurements are performed with a known calibration object and simulations are used to identify precisely the misalignments that reproduce the measurement results. Search algorithms are used which minimize the deviations from the real measurements in the parameter space (position and orientation of the lenses included in the structure). In addition, manufacturing errors of the individual surfaces are measured separately and are included in the simulation.

at that of Greivenkamp et al. proposed method of "reverse engineering " recommended the optical system from as few components as possible to collapse the parameter space. This restricts however the efficiency of the interferometer e.g. in terms of resolution.

Another method, which has already been used successfully in the measurement of spherical optics with large apertures and has now been transferred to the measurement of aspheres, results from the "Subaperture Stitching Interferometer" from the company QED, with the individual areas of the asphere by rotation of the Test specimens and pivoting of the interferometer are measured. comparisons J. Fleig, P. Dumas, PE Murphy, GW Forbes, "An automated subaperture stitching interferometer workstation for spherical and aspherical surfaces", Proc. SPIE, 5188, 296-307, 2003

The subaperture stitching interferometer, however, is device technology relatively expensive and still on weak aspheres limited.

A stitching method of the company QED, in which also the interferometer error, which is the same for all measurements, is determined, leads to an overall measurement, the measurement uncertainty in the measurement of spheres is even lower than in measurements without stitching. comparisons www.qedmrf.com Meanwhile, the measurement of weak aspheres should work with this method.

That from the aforementioned DE 103 25 601 B3 known method uses a tiltable reference wave, with which it is possible to reduce the stripe density in individual areas of the interferogram so far that an evaluation is easier. In this case, a phase-shifting point light source array is used, with which the tilted reference waves are generated and at the same time the phase shift is realized.

By tilting the evaluable range of the interferogram is shifted, so that the asphere can be measured in segments. In this case, a shift of a region of the interferogram is understood to mean a change of the two-dimensional fringe pattern. This change is the subject of the DE 103 25 601 as a result of a change in the angle of incidence of the reference wave on the camera. For this purpose, a phase shifting point light source array is used, which allows a single activation of individual, decentralized point light sources. A decentralized arrangement is understood to be an arrangement away from an optical axis of the system. The following collimating lens converts the spherical wave emanating from the point light source into a quasi-level tilted wave. The superposition with the object wave then yields an asymmetric interferogram, in which the areas of evaluable stripes have shifted. Without tilting arise symmetric interferograms, for example in the form of concentric strips known. In operation, some interferograms are recorded with different reference wave tilts to capture each point of the measurement field with at least one measurement. Subsequently, the evaluable ranges of the individual measurements are combined to form an overall measurement.

For the right one Put together of the areas it is necessary to know the exact influence of the tilting to determine the measurement. For this purpose, calibration measurements are made in advance with two simultaneously activated tilted reference waves and hidden object wave performed to their difference directly to measure interferometrically.

However, a disadvantage of this method is the time required. The time required results from the fact that many interferograms with different reference wave tilts must be recorded sequentially around each point of the measurement field of the at least one measurement to capture. In addition, higher aspheres again show high stripe densities, which make it difficult to measure these stronger aspheres.

Known are also non-interferometric methods such as the deflectometric Procedures in which the surface shape reconstructed from the beam deflection at reflection on the surface is, and purely mechanical Tastschnittverfahren, in which a mechanical Push button over the surface guided and whose deflection is measured. Deflectometric method measure either pointwise or areally. Profile method measure basically point by point. Point-by-point measuring methods (stylus method or point-wise deflectometry) take a relatively long time for the measurements. With the area Although measuring deflectometry can be local slopes and bends of the test piece measure well, but the absolute error of the measurements is not satisfactory.

In front In this background, the object of the invention in the specification a method and a measuring device of the beginning mentioned type, with which also stronger aspheres and / or freeform surfaces with Reduced time required without Nulloptiken accurately measured.

These Task is in a method and a measuring device respectively by the characteristics of the independent claims solved.

Thereby, that the surface simultaneously with several object waves of different point light sources is illuminated and reflected from the surface object waves in the beam path in front of the detector in a Fourier plane (Focal plane) of the imaging optics filtered aperture diaphragm filtered be the advantage that the aspherical DUT at the same time among many illuminated at different angles, each lighting angle leads to evaluable interference in one or more zones.

in the Comparison with the known interferometer with tilted reference the main advantage of this is the possibility of measuring several zones in parallel, which reduces the time required for a complete Measurement of the test object drastically reduced. It has also been shown that the invention also a measurement stronger aspheres allowed as with the interferometer with tilted reference possible is.

in the Contrary to aspheric measurements with Diffractive or refractive Nulloptiken results in a time and cost savings, because no special Nulloptiken made Need to become. Across from Interferometers, in which the specimen or the interferometer optics moved along the optical axis, there is the advantage that the results are not affected by mechanical inaccuracies become. Furthermore is a measurement non-rotationally symmetric specimens possible.

The in the following introduced interferometer with variable object wave allows not just the reduction of stripe densities in the interferometric Surveying aspherical Surfaces, but offers through a specially developed calibration strategy also the possibility a comprehensive characterization of the interferometer. Includes the calibration strategy is a definition and simulation-aided determination the waves in the test room, that belong to the different sources and the waves that come from the test room on the different camera pixels. The knowledge of this Waves leads to avoid so-called "retrace errors".

Further Advantages will be apparent from the description and the attached figures.

It it is understood that the above and the following yet to be explained features not only in the specified combination, but also in other combinations or alone, without to leave the scope of the present invention.

drawings

embodiments The invention are illustrated in the drawings and in the following description explained. In each case, in schematic form:

1 an embodiment of a measuring device according to the invention;

2 a conventionally and an interferogram taken in accordance with the invention;

3 an embodiment of a point light source array of the 1 ;

4 an embodiment of an alternative embodiment for generating object waves with different propagation direction; and

5 an embodiment for testing concave DUTs.

In detail, the shows 1 a measuring device 10 with a light source 12 , the coherent light in a Objektlichtwellenleiter 14 and a reference optical fiber 16 feeds. The measuring device 10 is also referred to below as interferons ter 10 designated. From the object fiber optic cable 14 leaking light 18 is from a first collimation optics 20 collimates and serves to illuminate a point light source array 22 , which below with reference to the 3 will be explained in more detail.

It is essential first that the point light source array 22 a matrix of elements of controllable transmission, each element in the state of controlled transmission and light source switched on 12 a point light source 23 represents, in each case a spherical wave 24 emanates. For the sake of clarity is in the 1 just a beam cone of a single spherical wave 24 shown. Part of the spherical waves 24 happens a beam splitter 26 and is through a second collimation optics 28 and an interferometer lens 30 as an object wave 32 on a visually smooth surface 33 of a test object 34 displayed. At the surface 33 it can be a spherical or aspherical, concave or convex surface or generally a free-form surface. By controlling and absteuern the transmission becomes the point light source 23 and thus also the resulting object wave switched on and off.

In this embodiment, discrete object waves 32 by a rigid arrangement of switchable point light sources 23 generated, which is approximately at a distance f in front of the collimating optics 28 where f is the focal length of the collimating optics 28 designated. Ideally, the distance is exactly equal to the focal length.

The 2-dimensional point light source array is used to generate object waves 32 different tilt relative to the optical axis of the second collimating optics 28 and the interferometer lens 30 , Becomes the central point light source of the point light source array 22 It uses a central spherical wave propagating from it through the optical system consisting of second collimation optics 28 and interferometer lens 30 , in the test room behind the interferometer lens 30 generates a converging, central spherical wave. This corresponds to the configuration of conventional interferometers. In this case, the point light source of the array is under the central point light source 22 understood, the optical axis of the optical system from the second Kollimationsoptik 28 and the interferometer lens 30 is closest.

Will, as in the 1 indicated, another, decentralized point light source 23 used, in the test room, the object wave tilted with respect to the central shaft is created 32 pointing to the examinee 34 falls and on its surface 33 is reflected. The surface 33 can be any free-form surface with or without symmetry. Depending on the shape of the test object 34 becomes the object wave 32 during reflection on the test object 34 deformed. By interferometer lens 30 , second collimation optics 28 and imaging optics 36 becomes the examinee 34 on a camera as a detector 38 displayed. An aperture stop 40 is in the beam path in front of the imaging optics 36 in the focal plane (Fourier plane) of the imaging optics 36 arranged and dimensioned so that they reflect only the proportions of the. object wave 32 lets through that little of a reference wave 42 deviate and thus lead to evaluable interference.

This becomes clear when looking at the beam path in the 1 considered. From the spherical wave 24 there is shown only one beam cone, which is bounded by an upper edge beam and a lower edge beam and has a further, central beam. Both marginal rays become after reflection on the surface 33 from the aperture stop 40 hidden and therefore do not fall on the camera 38 , By contrast, the central beam passes after its reflection on the surface 33 the aperture stop 40 and falls to the camera 38 , The surface element of the surface 33 , where the central beam is reflected, thus belongs to an illumination angle, which leads to an evaluable interference with a likewise on the camera 38 falling reference wave. The remaining rays are hidden and therefore do not contribute to a disturbing fringe pattern with an unresolvable stripe density. It is understood that the aperture can also be arranged slightly outside the focal plane. Essential is the suppression of rays, which contribute to a disturbing, non-resolvable stripe density. However, the arrangement in the focal plane is preferred.

The in the reference fiber optic cable 16 fed light is used as a reference wave 42 separately over the beam splitter 26 guided and by a reference wave lens 44 on the center of the aperture stop 40 focused. In the embodiment of 1 settles at the reference wave 42 a phase shift required by a phase shifter for a phase shift algorithm 17 in the reference optical fiber 16 Taxes.

In the embodiment of 1 can be differently oriented object waves 32 with the 2-dimensional point light source array 22 switchable light sources 23 of the point light source array 22 to reach. For every surface element on the surface 33 of the test piece 34 stands by the multitude of object waves 32 at least one illumination angle available, wherein the reflected wave on the surface element at a similar angle to the camera 38 like the time constant reference wave 42 ,

Because the aperture stop 40 only lets through the object waves that are at a similar angle to the reference wave 42 into the camera 38 come to mind, At the same time, it fades out the reflected object waves, which would not lead to any evaluable stripes. For each of a point light source 23 outgoing object wave 32 arises on the camera 38 at least one demarcated area that can be evaluated. Here, an evaluable range is understood to mean a stripe pattern with resolvable strips. The area or areas of other object waves lie elsewhere. As a result, each of these areas contains information about another part of the surface 33 of the test piece 34 , Due to the clear delimitation of the areas are preferred many object waves 32 switched on simultaneously.

To be sure that every point of the surface 33 of the test piece 34 with at least one object wave 32 is evaluated, a preferred embodiment provides a design of the interferometer, or the measuring device 10 before, in which the evaluable areas of different object waves 32 from different point light sources 23 overlap.

To avoid interfering with the overlapping areas, only a selection of object waves will be used 32 switched on at the same time, for example every fourth. That is, in the point light source array 22 becomes only a quarter of the available point light sources 23 activated simultaneously. In the next measurement, every fourth object wave, but other object waves are used again, etc. After four measurements, all object waves were used once and every point of the surface 33 of the test piece 34 was with at least one object wave 32 evaluated sampled.

A calculator 48 controls the activity of each point light source 23 of the point light source array 22 , stores the for each lighting condition of the DUT 34 and thus for each of the switched states (transparent / non-transparent) of point light sources 23 of the point light source array 22 from the camera 38 recorded interferograms, and determines the dimensions of the surface 33 of the test piece 34 by an evaluation of the intensity distributions of the stored interferograms. The computer 48 In this respect, it is set up, in particular programmed, to control the course of the method according to the invention and / or one of its embodiments. Together with the calculator 48 This results in a total of a measuring device 10 , which is set up for carrying out the method according to the invention and / or its embodiments presented here.

As a light source 12 In one embodiment, a He-Ne laser of wavelength λ = 633 nm with 10 mW power is used. This embodiment is useful for measuring optics for visible light. Depending on the application, other light sources with different powers and / or other wavelengths may be useful.

2a shows qualitatively a section of a conventional interferogram, which has been taken with a single, central point light source. A central area with a stripe pattern with small radii is separated from a peripheral stripe pattern with larger radii. In between is an area where the stripe density is so high that it can not be resolved by the camera. As a consequence, no usable information about the structure of the test object can be obtained from this intermediate area 34 be won.

2 B shows, however, also purely qualitatively, a recorded using the invention interferogram. The simultaneous use of multiple tilted object waves results in multiple sub-interferograms shifted from a central position, each of which is a particular surface element of the surface 33 , an associated object wave direction and thus a specific point light source of the point light source array 22 assigned.

The following is with reference to the 3 the operation of the phase shifting point light source array 22 explained. The primary purpose of the array 22 consists in the production of individually switchable, arranged in a predetermined grid point light sources 23 , The location of the point light sources 23 is due to the location of pinhole 50.1 . 50.2 . 50.3 . 50.4 , .... a pinhole array 50 Are defined. In the beam path in front of the pinhole 50.1 . 50.2 . 50.3 . 50.4 , there are microlenses 52.1 . 52.2 . 52.3 . 52.4 , ... a microlens array 52 to focus from the right to the pinhole 50.1 . 50.2 . 50.3 . 50.4 , .... incoming light. The third component is the array 22 a liquid crystal display 54 (LCD), that, in the propagation direction of the light 18 , even before the microlens array 52 is placed.

The liquid crystal display 54 consists of a grid of small liquid crystal cells (pixels), their transmission from the computer 48 controlled is adjustable. In the ground state of the point light source array 22 (all point sources 23 inactive) all pixels of the liquid crystal display 54 switched non-transparent. The illumination of the entire point light source array 22 occurs at a slight angle α, so that the point light sources 23 remain inactive even when a small portion of the light passes through the "intransparent" pixels of the liquid crystal display 54 penetrates. This results from the fact that the light in such a case just next to the pinhole 50.1 . 50.2 . 50.3 . 50.4 , .... is focused.

To activate a point source 23 will with the liquid crystal display 54 in front of a microlens 52.2 a local diffraction grating 56 generated. Diffraction gratings cause the light to split into different orders of diffraction, which propagate in slightly different directions. The system is adjusted so that the first diffraction order of the microlens 52.2 exactly on the pinhole 50.2 is focused (dashed lines). All point sources can be activated and deactivated independently of each other. In addition, the phase of each active point light source 23 by the lateral displacement of the diffraction grating 56 be moved. Such defined phase shifts are required in the standard evaluation algorithms of phase-shifting interferometry.

In a preferred embodiment, the phase shifting point light source array 22 a liquid crystal display 54 with a pixel pitch of 36 μm and a pixel count of 1024 × 768. It is understood, however, that the dimensions and pixel numbers are not limited to these values.

The quartz substrate, the microlenses on one side 52.1 . 52.2 . 52.3 . 52.4 , ... of the microlens array 52 and on the other side the pinholes 50.1 . 50.2 . 50.3 . 50.4 , ...., in one embodiment has a thickness of 9 mm. The diffractive microlenses 52.1 . 52.2 . 52.3 . 52.4 , ... are each 2.52 mm × 2.52 mm in size and made in gray tone lithography. The distance between the pinholes 50.1 . 50.2 . 50.3 . 50.4 , to each other corresponds to the microlens size, its diameter is 25 microns. With the realized point light source array 22 In one embodiment, 11 × 13 point light sources 23 switchable. With a second collimation optics with a focal length f = 160 mm results in a maximum tilt angle of 7 °, with which the measuring device shown 10 is working. The measuring device 10 is hereafter also called interferometer 10 designated.

4 shows an alternative embodiment of a point light source array. The light of the object wave light guide 14 becomes 1: N (N preferably = 4) switch 60 on one of the N light guides 61 connected. The ends of the N light guides 61 be in a holder ( 62 ) fixed. At N = 4, the fixation is preferably such that the ends lie on the corners of a square. The N ends of the N light guides 61 Make N point light sources 23 , After a certain propagation distance that falls from such a point light source 23 outgoing light 24 on a diffractive element 66 , preferably a cross grid, which is designed so that the input light 24 diffracted in M orders of diffraction. In this way, object waves that are initially tilted relative to one another by an object wave M are generated, which illuminate the test object 34 be used.

5 shows an embodiment for testing concave DUTs 34b , Here can on the collimation optics 28 as well as the interferometer lens 30 out 1 be waived.

The embodiment of the 5 contrary to 1 no phase actuator 17 in the reference waveguide 16 rather, the phase shift becomes direct with the phase shifting point light source array 22 accomplished.

To the measuring device 10 With variable object wave, a special calibration strategy has been developed. It first requires the parameterization of all possible object waves that are in the test room (space in which the DUT 34 is positioned) arrive. This parameterization must be able to both the object waves 32 in space, as well as the dependence on the choice of the object wave (in the example: dependence on the coordinate of the point light source). The object waves 32 can be described, for example, by optical path length functions which result in a reference plane in the test room. The path length is then dependent on spatial coordinates X, Y (in the plane of reference of the beam propagation) and the source coordinates M, N. The dependence of the path length WQ on the coordinates can be expressed in the form of polynomials:

die Summe mit den Indexvariablen i, j, k, l. Diese gehen von Null bis zu einem von der gewünschten Genauigkeit abhängigen Wert. Q ist ein 4-dimensionaler Tensor der Quellenwellen mit den Koeffizienten Q_ijkl beschreibt, die zu den Potenzenprodukten MⁱN^jX^kY^k gehören. Andere Polynomdarstellungen, z.B. Zernike-Polynome, sind denkbar.Here is the term

the sum with the index variables i, j, k, l. These range from zero to a value dependent on the desired accuracy. Q is a 4-dimensional tensor describing source _waves with the coefficients Q _ijkl belonging to the power products M ⁱ N ^j X ^k Y ^k . Other polynomial representations, eg Zernike polynomials, are conceivable.

Equivalent to the definition of the source waves Q, pixel waves P may be defined as the waves that exactly pass from the inspection room through the entire imaging system to a camera pixel. These pixel waves are also described as path length functions that depend on location coordinates x, y in a reference plane in the test space and on the camera pixel coordinates m, n. Again, polynomials are used to describe the path lengths WP of the pixel waves P

The characterization of the interferometer 10 P and Q is a "black-box" description that describes the effect, but not the internal structure of the interferometer 10 describes. If P and Q are known, the expected optical path for each object wave can be found for a known object in the test room Calculate length WP from the source to each point in the camera plane (path length of a pixel-source combination). For an ideal interferometer, P and Q can be calculated using optical simulation software and polynomial fits.

It is still unclear how unavoidable deviations of a real interferometer 10 of the ideal interferometer, which can also be identified and quantified by different coefficients in P and Q. For this purpose, a selection of pixel-source combinations are first selected for one or more known calibration objects whose optical path lengths are considered below. For each coefficient of P and Q it is then simulated how the infinitesimal change of the coefficient affects the optical path lengths of the selected pixel-source combinations.

These are stored as characteristic path length changes of the coefficient. Then be with the real interferometer 10 measured the deviations of the optical path lengths compared to the ideal interferometer for the selected pixel-source combinations. Finally, an attempt is made to represent these deviations as a linear combination of the characteristic path length changes of the coefficients. This takes place in the form of a linear system of equations, which is resolved after the coefficient deviations.

These coefficient deviations are added to the coefficients of the ideal interferometer to obtain P and Q of the real interferometer 10 to obtain. If different calibration objects or the same calibration object are used in different positions for calibration, it must be expected that inaccuracies in the relative positioning will occur. The influence of the positioning errors on the pixel-source combinations can also be simulated and included in the linear system of equations, so that the solution of the linear system of equations also reveals the positioning errors.

Is this interferometer 10 calibrated as described, so may from a measurement with the interferometer 10 With variable object wave, the shape of the specimen can be uniquely determined using the real P and Q. Similar to the calibration, the process can be carried out by solving a linear system of equations. This time, the UUT shape will also be described using polynomials and their associated coefficients. It is determined again simulatively which measurement results are to be expected from an ideal test specimen and which influence the deviation of individual coefficients (of the specimen) has on the measurement result.

The Deviation of the real measurement result from the simulated measurement result an ideal examinee becomes again as linear combination of the deviations of individual coefficients shown. The resulting linear system of equations is calculated according to Coefficient deviations resolved, which are added to the coefficients of the ideal candidate form. High-frequency residual deviations, which are described by the description by Polynomials are not recorded, taking into account the impact angle the rays on the specimen surface displayed.

A preferred embodiment is characterized in that the computer performs the described calibration procedure and therefore set up to carry it out, is programmed in particular. Alternatively or in addition the results of the calibration process stored in the computer and in the measurement of test pieces used.

Claims (27)

Method for the surface measurement of an optically smooth surface ( 33 ), the surface ( 33 ) with a plurality of discrete object waves ( 32 ) is illuminated from different directions, and those at the surface ( 33 ) reflected object waves ( 32 ) with one to one or more object waves ( 32 ) coherent reference wave ( 42 ) on a detector ( 38 ) are superimposed to an interferogram in which dimensions of the surface ( 33 ), characterized in that the surface ( 33 ) simultaneously with a plurality of switchable object waves ( 32 ) and that of the surface ( 33 ) reflected object waves ( 32 ) in the beam path in front of the detector ( 38 ) arranged aperture diaphragm ( 40 ), which are preferably in the Fourier plane of an imaging optics ( 36 ) are filtered. Method according to claim 1, characterized in that the reference wave ( 42 ) separately and approximately to the center of the aperture ( 40 ) is focused. Method according to one of the preceding claims, characterized in that the discrete object waves ( 32 ) are generated by a rigid arrangement of switchable point light sources, which are located at a distance f in front of a collimating optics ( 28 ), where f is the focal length of the collimating optics ( 28 ) designated. Method according to one of the preceding claims, characterized in that phase shifts between an object wave ( 32 ) and the reference wave ( 42 ) by shifting the phase of the reference wave ( 42 ) be generated. Method according to one of the preceding claims, characterized marked that for a calibration first all possible Object waves in a test room arrive, be parameterized. Method according to claim 5, characterized in that that the parameterization by describing the object waves as Functions of the optical path lengths carried out in a reference plane in the test room. A method according to claim 6, characterized in that the description by polynomial functions WQ of the form

with spatial coordinates X, Y in the reference plane and source coordinates M, N or by Zernike polynomials. Method according to claim 7, characterized in that that, equivalent to define the source waves, pixel waves are defined as the waves be passing out of the test room run the entire imaging system exactly to a camera pixel. Method according to claim 8, characterized in that the pixel waves are also described as path length functions are derived from location coordinates x, y in a reference plane in the test room and depend on the camera pixel coordinates m, n. Method according to claim 9, characterized in that polynomials

to describe the path lengths WP of the pixel waves. Method according to one of the preceding claims, characterized characterized in that with known P and Q for a known object in the test room to each object wave the expected optical path length of the source is calculated to every point in a camera plane. Method according to claim 11, characterized in that that P and Q for an ideal interferometer using optical simulation software and polynomial fits are calculated. Method according to claim 12, characterized in that that unavoidable deviations of a real interferometer from ideal interferometer, which is also characterized by divergent coefficients in P and Q show, thereby be recognized and quantified that first for one or a plurality of known calibration objects, a selection of pixel-source combinations chosen be, for simulates each coefficient of P and Q of the associated optical path length WP becomes how the infinitesimal change of the coefficient arises the optical path lengths the chosen one Affect pixel-source combinations and then this as a characteristic path length of the coefficient, then with the real interferometer the deviations of the optical path lengths compared to the ideal interferometer for the selected pixel source combinations be measured, and these deviations as a linear combination of characteristic path length changes of Coefficients in the form of a linear equation system shown which is resolved after the coefficient deviations, and the coefficient deviations to the coefficients of the ideal interferometer are added to obtain P and Q of the real interferometer. A method according to claim 13, characterized that if different calibration objects or the same for calibration Calibration object used in various positions, an influence of positioning errors on the pixel-source combinations as well is simulated and included in the linear system of equations. Method according to one of the preceding claims, characterized characterized in that the angle of the object wave, with which the surface is illuminated, is varied. Method according to claim 15, characterized in that that differently oriented object waves with a 2-dimensional grid be generated by switchable light sources. Method according to claim 16, characterized in that the aperture diaphragm ( 40 ) is dimensioned so that it only allows angles leading to evaluable stripes in the interferogram. Method according to one of claims 15 to 17, characterized that the interferometer is designed so that from different object waves resulting subregions of the interferogram with evaluable Overlap strips. Method according to one of claims 15 to 18, characterized that when recording a first interferogram only a first Selection of object waves is switched on at the same time. A method according to claim 19, characterized in that for recording a second interferogram only a second, from the first out Choice of different selection of object waves ( 36 ) is turned on at the same time. Method according to Claim 20, characterized in that for pairings of interferograms pairs of different selections of object waves ( 36 ), the object waves ( 36 ) of a selection are turned on simultaneously, and that the process is repeated with different selections until all object waves ( 36 ) and / or each point of the surface with at least one object wave ( 36 ) was evaluated. Measuring device ( 10 ) for measuring an optically smooth surface ( 33 ), with at least three switchable point light sources ( 23 ), an imaging optics ( 36 ) and a detector ( 38 ), wherein the measuring device ( 10 ) is adapted to the surface ( 33 ) with object waves ( 32 ) to be illuminated by the point sources ( 23 ) on the surface ( 33 ) reflected object waves ( 32 ) with one to the object waves ( 32 ) coherent reference wave ( 42 ) through the imaging optics ( 36 ) and on a detector ( 38 ) to superimpose an interferogram in which dimensions of the surface ( 33 ), characterized in that the measuring device ( 10 ) one in the beam path in front of the detector ( 38 ) preferably in a Fourier plane of the imaging optics ( 36 ) arranged aperture diaphragm ( 40 ), and is adapted to the surface ( 33 ) simultaneously with several object waves ( 32 ) of different point light sources ( 23 ) and the surface ( 33 ) reflected object waves ( 32 ) through the aperture diaphragm ( 40 ). Measuring device ( 10 ) according to claim 22, characterized in that the point light source ( 23 ) Part of an array ( 22 ) of individually switchable point light sources ( 23 ), where the array ( 22 ) so many point light sources ( 23 ) and these are arranged so that for different surface elements of the surface ( 33 ) yields at least one respective illumination angle at which the object wave reflected on the surface element ( 32 ) at a similar angle of incidence on the detector ( 38 ) hits like the time-constant reference wave ( 42 ). Measuring device ( 10 ) according to claim 22, characterized in that the measuring device 10 an object wave light guide ( 14 ) and a 1: N switch 60 having the object waves from the optical fiber ( 14 ) on one of N optical fibers 61 switches, their ends in a holder ( 62 ) and point light sources ( 23 ) whose light is transmitted through a diffractive element, 66 , Preferably, a cross grating is diffracted in M diffraction orders, so that from the object wave initially tilted object wave M to each other generated object waves are generated to illuminate the specimen 34 be used. measuring device 10 according to claim 24, characterized in that N = 4 and the 4 ends are fixed on the corners of a square. Measuring device ( 10 ) according to one of claims 22 to 25, characterized in that the aperture diaphragm ( 40 ) only object waves ( 36 ) with an angle of incidence on the detector ( 38 ), which lead to evaluable strips. Measuring device ( 10 ) according to one of claims 22 to 26, characterized in that it is adapted to carry out one of the methods according to claims 2 to 20.