Method and device for contactlessly measuring an object

Abstract

The invention relates to a method for measuring an object (12; 12'), in which method an optical coherence tomograph (22) directs a measurement light beam (24) onto at least one point (39; 39', 39'') on a surface (40; 40') of an object (12; 12') via a scanning device (26) and via an optical system (38), thereby obtaining a plurality of measurement values. The angular position of an object table (16) on which the object has been placed is altered depending on the measurement values in such a way that, after the change in angular position, the surface (40; 40') of the object (12; 12'), or at least one region of the surface, is oriented, on average, more perpendicularly to the optical axis (OA) of the optical system (38) than before the change in angular position.

Description

[0001] Method and device for non-contact measurement of an object

[0002] BACKGROUND OF THE INVENTION

[0003] 1. Field of the invention

[0004] The invention relates to a method and a device for non-contact measurement of an object. The object can, in particular, be a wafer whose geometric properties, such as TTV, deflection, or warping, are to be determined.

[0005] 2. Description of the state of the art

[0006] For measuring the surface and thickness profiles of workpieces and other objects, devices based on the principle of SD OCT have been used for several years. The acronym OCT stands for optical coherence tomography, and SD stands for spectral domain. These devices measure the distance to a scattering or at least partially reflective surface of the object at individual measurement points without contact, provided they have a reference arm. Repeating this measurement at a large number of measurement points distributed across a grid-like area yields a two-dimensional surface profile of the object.

[0007] If the object is sufficiently transparent to the measuring light, the thicknesses between scattering or reflecting interfaces that define the object or are located within the object can also be measured with high accuracy. A measurement at multiple points then provides a corresponding thickness profile. Combining distance and thickness measurements allows, for example, verification that the surface of a coated workpiece has the desired shape and that the coating thickness is constant throughout.

[0008] The first devices of this type could only determine the distance for a single measuring point. To scan an area, the measuring head had to be moved relative to the workpiece (or vice versa). DE 10 2022 104416 A1 discloses a device in which a scanning unit guides the measuring beam across the workpiece. The lateral distance between adjacent measuring points depends on the required resolution and can be on the order of approximately 20 pm. The applicant markets such devices under the name "Flying Spot Scanner".

[0009] Such devices are particularly suitable for measuring wafers. Wafers are circular or square discs, approximately one millimeter thick, that serve as substrates for integrated circuits, micromechanical components, or photoelectric coatings. Wafers are produced from monocrystalline or polycrystalline ingots, which are sawn perpendicular to their longitudinal axis into individual wafers. In most cases, wafers consist of silicon, but other materials are also used, such as glass for the production of microlens arrays. Increasingly, two or more very thin wafers made of the same or different materials are bonded together in a bonding process to form a composite structure.

[0010] The shape of the wafers must conform to strict geometric specifications. These specifications include the TTV (total thickness variation), which is the maximum difference between the thickest and thinnest points of a wafer. Bow is defined as the maximum deviation of the wafer's median surface from a reference plane. Warp, as understood by those skilled in the art, is the deviation of the wafer's median surface from the reference plane after correcting for bow across the entire wafer surface.

[0011] In addition to these common definitions of TTV, warping, and buckling, similar but slightly different definitions are sometimes used to characterize wafers and describe their deviation from the ideal shape. These include total thickness variation as well as global and local warping.

[0012] To comply with the strict specifications, these geometric dimensions must be measured at least on a sample basis and preferably during the regular production process. As mentioned above, the devices known from DE 10 2022 104416 A1 are particularly well suited for this purpose.

[0013] Especially when measuring thicker wafers or composite structures made up of several individual wafers, the intensity of the light scattered or reflected by the different interfaces can vary considerably. In the worst case, the intensity is so low that the interference contrast is insufficient for a reliable measurement.

[0014] This problem is solved by using an optical system with a large depth of field to focus the measuring light. This large depth of field is achieved through a small image-side aperture angle of the optical system. In simpler terms, this results in a very narrow beam of measuring light; that is, the measuring light strikes the wafer either parallel to the optical axis or at very small angles, typically no more than 2°. If the focus is then positioned approximately midway between the two outermost interfaces of the wafer, these interfaces, despite any warping, etc., are generally located within a region where the measuring light is sufficiently focused. This is advantageous because the reflection, and consequently the interference signal, is stronger the better the light is focused at an interface.

[0015] However, the small aperture angle of the optical system also has disadvantages. If the interfaces are not aligned perpendicular to the optical axis of the optical system, some of the reflected light cannot be captured by the optical system and is lost to interference in the coherence tomograph. The greater the deviation from perpendicular alignment, the lower the interference contrast. If the surface normal of an interface deviates from the optical axis by more than approximately 4° in such devices, no reflected light reaches the coherence tomograph. Depending on the measurement system, the limits of measurability are reached even at smaller angles, e.g., 1.5°. Measurement of the wafer is therefore not possible at such locations.

[0016] Similar problems arise when measuring not wafers, but other essentially flat objects whose thickness requires small opening angles.

[0017] SUMMARY OF THE INVENTION

[0018] The object of the invention is to provide a method and a device that enable the measurement of wafers and other objects even when the optical interfaces of the object to be measured are more inclined or have areas with a greater inclination.

[0019] With regard to the method, this problem is solved according to the invention by a method for measuring an object comprising the following steps: a) an optical system having an optical axis and a scanning device are provided; b) an object stage is provided which defines a reference plane whose angle to the optical axis is adjustable; c) the object is placed on the object stage such that a surface of the object faces the optical system; d) an optical coherence tomograph directs a measuring light beam via the scanning device and the optical system to at least one point on the surface of the object, thereby obtaining several measured values;e) The angular position of the object stage is changed depending on the measured values ​​obtained in step d) such that the surface of the object, or at least a region of the surface, is oriented more perpendicular to the optical axis of the optical system on average after the change in angular position than before the change in angular position; and f) the distances to an optical interface of the object and / or between two optical interfaces of the object are measured by guiding the measuring light beam from the scanning device over the surface of the object or over the region of the object.

[0020] The invention utilizes the fact that an undesired tilt of an object's surface can be compensated for by changing the angular position of the stage on which the object is placed, thus compensating for the tilt. "Compensated" in this context means that, after the change in angular position, the object's surface—or at least a region thereof—is oriented more perpendicular to the optical axis of the optical system than before. With the stage optimized, the entire surface, or at least parts of it, can be measured more effectively, i.e., with higher interference contrast.

[0021] In principle, tilting the optical system relative to the stage to compensate for inclinations of the object surface would also be an option. However, with extended objects such as large-diameter wafers, the optical system becomes bulky and correspondingly heavy. Furthermore, connecting the beam paths of the optical system and the coherence tomograph is more complex if the optical system is to be tilted. Since the objects to be measured are usually considerably lighter than the optical system, a tiltable stage can be implemented mechanically with minimal effort.

[0022] The invention is based on the consideration that at least a rough determination of the inclination of the object surfaces can in most cases be carried out with the coherence tomograph that is already available, so that no external additional measuring devices are required.

[0023] A first and particularly simple approach to improving alignment is to repeatedly direct the measuring light beam onto the same point on the object's surface, but with the stage angles set differently, in step d). The optical coherence tomograph then detects the intensity of a portion of the measuring light that is reflected at that point and returns to the optical coherence tomograph. In step e), the angular position of the stage is changed so that the intensity of the portion of the measuring light detected by the optical coherence tomograph is increased and preferably maximized.

[0024] This approach, which can also be implemented manually, utilizes the fact that the intensity of the reflected measuring light decreases with increasing inclination of the object surface, as explained above. At maximum intensity, the surface is aligned perpendicular to the optical axis, as desired.

[0025] Preferably, these steps are performed automatically by changing the angular position of the stage using servomotors, depending on the detected intensity. This automatic adjustment of the stage can, for example, be achieved by first tilting the stage over a wide angular range around a first axis using the servomotors, and then tilting it around a second axis. Simultaneously, the intensity of the reflected light is measured at each angular position. Since the measurement time is very short, it can be performed during a continuous tilting movement of the stage. The stage is then moved to the angular position at which the highest intensities were measured.

[0026] If reflections from different optical interfaces interfere in the optical coherence tomograph, the intensity of the reflected component of the measurement light can be determined indirectly via the interference contrast.

[0027] If the intensity measurement is to be performed with the object stationary and only a few angular positions are to be reached to reduce measurement time, interpolating the dependence of the intensity on the angles can also be considered. In this way, the angular position to which the stage is ultimately moved can be one at which no intensity measurement has previously taken place.

[0028] The point at which the intensity measurement is performed at different angular positions of the stage can lie on the optical axis of the optical system. Since the object in step c) is often positioned on the stage so that it is centered with respect to the optical axis, this means a measurement at the center of the object's surface, which is convenient for many types of objects. Furthermore, the two axes around which the stage can be tilted are usually arranged so that the optical axis intersects both axes. A measurement on the optical axis also has the advantage that when the stage is tilted, its z-position (i.e., height) changes only slightly. This eliminates the need for height adjustments that might be required when taking measurements at the edge of the stage.For some objects, it may be more advantageous to take the measurement at a point outside the optical axis or even at the edge of the object, which may require corresponding compensatory movements.

[0029] Performing the intensity measurement described above at only one point is generally sufficient if the object surface is largely flat but inclined, i.e., has a surface normal that forms a significant angle with the optical axis. If the inclination is known at one location on the surface, it will not differ significantly at other locations on the surface.

[0030] If these conditions are not met, it is usually more practical to perform intensity measurements at several points with the stage at different angles. Since a curved surface, or one that deviates from a plane in any way, can never be oriented perfectly perpendicular to the optical axis everywhere, an orientation should be chosen where the average deviations from perpendicularity are as small as possible. This average need not be an arithmetic mean. A quadratic mean or an even higher-order mean is usually more suitable, as smaller deviations from perpendicularity can be easily tolerated. Particularly large deviations, however, can prevent any measurements of distances or thicknesses at those points.

[0031] When measuring at multiple points on the surface, it is most advantageous to perform the intensity measurement at all desired points with the stage at a specific angle, then change the angle of the stage, measure the intensity again at all desired points with the stage at the changed angle, and so on.

[0032] With particularly curved object surfaces, it may occur that there is no orientation that allows for distance or thickness measurements across the entire object surface, or—if the entire surface is not to be measured—across the entire desired measurement area. In these cases, the object cannot be measured with the same angular position of the object table. Instead, the angular position of the object table must be changed at least once during the measurement. This change in angular position during the measurement must be taken into account when calculating the desired surface or thickness profile by combining the results of the partial measurements with appropriate modifications.

[0033] Finding one (or more) suitable angular positions of the stage as described above is reliable but relatively slow, since the intensity measurements must be performed at different angular positions. A second approach is faster, in which the coherence tomograph is not used solely for intensity measurements. Instead, the object is roughly measured to approximate the shape of its surface, and the object is then optimally aligned using the stage based on this measured shape. The actual measurement, in which the measurement points are distributed in a grid pattern or along straight or curved lines, is only performed after this alignment.

[0034] In this second approach, in step d), the measuring light beam is successively directed at different points on the object's surface, and the distance to the surface is measured at each point. From these distances, an approximate shape of the object's surface is calculated. This approximate shape is sufficient to find, in step e), an angular position of the stage at which the object's surface, or at least a region of the surface, is oriented more perpendicular to the optical axis of the optical system after the change in angular position than before the change in angular position.

[0035] Preferably, the points in this preliminary measurement are distributed over the entire surface of the object. If, for example, experience with comparable objects has shown that the surface deviates significantly from a flat shape only in a specific area (e.g., near the edge), this preliminary measurement can also be limited to the corresponding area of ​​the surface.

[0036] To determine the shape of the surface approximately, for typical dimensions of the object (e.g. a wafer with a diameter of 20 or 30 cm) the number of different points can be between 5 and 1000, preferably between 50 and 500.

[0037] Similar to the first approach described above, it may occur that during the preliminary measurement, the distance to the interface cannot be measured for one or more of the different points. In this case, the angle of the stage can be adjusted so that a distance measurement at these points becomes possible. The angle of the stage can be automatically adjusted using servo motors until the distance to the interface can be measured for one of the different points.

[0038] Since changing the orientation of the measuring light beam is generally much faster than adjusting the angle of the stage, it is advantageous to first perform a preliminary measurement at a given stage angle, then change the angle and perform another preliminary measurement. This is preferably repeated until an angle setting has been determined for all areas of the object or for all grid points of the preliminary measurement at which the distance to the interface can be measured. The procedure can be accelerated by performing further preliminary measurements only in areas for which no distances to the interface could yet be measured. The angle settings are either systematically tested (preferably from smaller to larger angles), or an inter- or inter-measurement is performed as described below.Extrapolation is performed from the already known areas. Once measurement values ​​of sufficient quality are available for all grid points, the preliminary measurements are completed.

[0039] The same applies to the actual high-resolution measurement after optimized alignment of the object using the stage. If, despite this alignment, strongly inclined areas of the surface remain where no distance or thickness measurement is possible, the measurement must be divided into different phases, between which the orientation of the object is adjusted using the stage.

[0040] The angular settings of the object table during the measurement phases are expediently selected based on data obtained from one or more preliminary measurements. An algorithm – for example, based on machine learning or neural networks – can be advantageously used here to ensure an optimal balance between speed and accuracy of the measurement when choosing the angular settings and number of phases.

[0041] Changes to the angular position of the object table made during the surveying process must also be taken into account in this second approach when determining the profiles of the optical interfaces or their distances from each other.

[0042] There are also cases where, despite changing the angle of the stage, areas remain on the object's surface where no distances can be measured. In such cases, one could try numerous different angles until a suitable one is found. It is more efficient to approximate the shape of the affected areas by extrapolation or interpolation, since information is usually available for points on the edge of these areas, allowing the local inclination of the surface at those points to be determined. Using a camera that captures an image of the object after step c) can increase reliability in many cases. An image processing algorithm then calculates, based on the captured image, how the object's surface is inclined at several locations relative to the optical axis.

[0043] With regard to the device, the aforementioned problem is solved by a device for measuring an object, comprising an optical system having an optical axis, a scanning device, and a stage defining a reference plane whose angle to the optical axis is adjustable, wherein the object can be placed on the stage such that a surface of the object faces the optical system. The device further comprises an optical coherence tomograph configured to direct a measuring light beam via the scanning device and the optical system onto at least one point on the surface of the object, thereby obtaining multiple measurement values.A control device is configured to (a) change the angular position of the object stage depending on the measured values ​​obtained, such that the surface of the object or at least a region of the surface is oriented more perpendicular to the optical axis of the optical system on average after the change in angular position than before the change in angular position, and (b) after the change in angular position of the object stage, guide the measuring light beam from the scanning device over the surface of the object or over the region of the object in order to measure the distances to an optical interface of the object and / or between two optical interfaces of the object.

[0044] The control unit can be configured to repeatedly direct the measuring light beam onto the same point on the object's surface, but with the stage angles set differently. The optical coherence tomograph is then configured to detect the intensity of a portion of the measuring light that is reflected at that point and returns to the optical coherence tomograph. The control unit is configured to change the angular position of the stage so that the intensity of the portion of the measuring light detected by the optical coherence tomograph is increased. The device can include actuators configured to automatically change the angular position of the stage depending on the detected intensity.

[0045] The actuators are preferably designed so that the stage can be tilted by up to 15° from the neutral position in all directions. For many applications, 10° is sufficient.

[0046] The control unit can be configured to direct the measuring light beam successively onto different points on the object's surface. The optical coherence tomograph is then configured to measure the distance to the object's surface at each point. An evaluation unit is configured to calculate an approximate shape of the object's surface from these distances.

[0047] The control device can be configured to adjust the angular position of the object table if the distance to the interface cannot be measured for one of the different points.

[0048] In this case, the control device can also be configured to automatically adjust the angular position of the object table using actuators until the distance to the interface can be measured for one of the different points.

[0049] The device can be configured to measure the distances to the optical interface of the object across the entire surface of the object, wherein an evaluation unit of the device is configured to calculate a shape of the surface from the measured distances and to computationally take into account changes in the angular position of the object stage made during the measurement.

[0050] In one embodiment, the device has a camera which is set up to take an image of the object and, using an image processing algorithm, to calculate how the surface of the object is inclined at several locations relative to the optical axis based on the image taken.

[0051] The invention further relates to a measuring system comprising a device according to the invention and an airtight, sealable cabin in which the device is mounted in a vibration-isolated manner. The vibration isolation ensures that vibrations of the cabin, such as those caused by impacts or similar events, are not transmitted to the device and thus do not distort the measurement result. The airtight seal ensures that dust or other contaminants do not impair the measurements. Additional measures can be taken to create cleanroom conditions inside the cabin to further reduce the risk of contamination.

[0052] BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Exemplary embodiments of the invention are explained in more detail below with reference to the drawings. These show:

[0054] Figure 1 shows a wafer to be measured in a perspective and not-to-scale representation in its real and ideal form;

[0055] Figure 2 shows an exemplary embodiment of the measuring device according to the invention in a schematic representation;

[0056] Figure 3 shows important parts of a scanning device that is part of the measuring device shown in Figure 2, in a simplified perspective view;

[0057] Figure 4 shows a mechanism for changing the angular position of the object stage in a schematic side view;

[0058] Figure 5a shows a section of the measuring device during the measurement of a wafer with a wedge-shaped cross-section in an initial angular position of the object table;

[0059] Figure 5b shows the spectrum obtained by measurement for the constellation shown in Figure 5a;

[0060] Figure 6a shows the section according to Figure 5a, but after a first tilting of the object table;

[0061] Figure 6b shows the spectrum obtained by measurement for the constellation shown in Figure 6a;

[0062] Figure 7a shows the section shown in Figure 5a, but after a second tilting of the object stage; Figure 7b shows the spectrum obtained by measurement for the constellation shown in Figure 7a;

[0063] Figure 8 shows a top view of a wafer with a uniform grid of measuring points where distances are measured as part of a preliminary measurement;

[0064] Figure 9a shows a section of the measuring device based on Figure 5a, with a more complexly deformed wafer in a first angular position of the object stage;

[0065] Figure 9b shows the section according to Figure 9a in a second angular position of the object stage;

[0066] Figure 10 shows a top view of a wafer similar to Figure 8, where measurements are not possible at all measuring points;

[0067] Figures 11a and 11b show a section of the measuring device based on Figure 5a, with a particularly strongly curved wafer in two different angular positions of the object stage;

[0068] Figure 12 shows a section of a measuring device with a camera, similar to Figure 5a, wherein the object to be measured has an irregularly shaped surface;

[0069] Figure 13 shows an image of the object taken by the camera shown in Figure 12; and

[0070] Figure 14 shows a measuring system with a cabin and a measuring device included therein in a simplified perspective view.

[0071] DESCRIPTION OF PREFERRED EXAMPLES

[0072] 3. Measurement of wafers

[0073] Figure 1 shows a wafer 10 in a perspective, but not definitive, view. In the illustrated embodiment, the wafer 10 has the shape of a right circular cylinder, with the thickness being significantly exaggerated. Real wafers 10, for example, have a diameter of up to 300 mm, while the thickness is only about 1 mm. Occasionally, wafers 10 are also used whose surface is not circular, but square.

[0074] The ideal circular cylindrical shape of wafer 10 is indicated by dashed lines 12. Due to manufacturing tolerances, deviations from this ideal shape can occur. In Figure 1 and all subsequent figures, these deviations are greatly exaggerated for clarity. To determine the deviations, wafer 10 must be measured. By measuring the topography of both wafer surfaces, all common geometric specifications of the wafer, such as TTV, warpage, and curvature, can be derived.

[0075] The distribution of the measurement points at which the topography is measured is adapted to the specific measurement task. In the embodiment shown in Figure 1, it is assumed that the measurement points are arranged along two lines 11, 13, which are perpendicular to each other, intersect in the center of the wafer 10, and each extend to the circumferential edge of the wafer 10. Other measurement patterns, e.g., spirals or grid patterns, are of course also possible. The measurement points can be very close together and, for example, have a distance of a few micrometers. In other measurement patterns, the distances are in the range of 1 mm.

[0076] 4. Setup of the measuring device

[0077] Figure 2 shows a schematic representation of a measuring device, designated 14, according to an embodiment of the invention. The measuring device 14 serves to measure the wafer 12 in the manner described above. The wafer 12 is placed on a stage 16, which is height-adjustable and has two orthogonal axes A extending along the x and y directions. x and A y The object stage 16 can be tilted. In this way, not only the z-position but also the angular position of the object stage 16 can be changed. In Figure 2, the object stage 16 and the wafer 12 supported by it are indicated by dashed lines in a different angular position, with a tilt about the axis A. y was assumed.

[0078] In the illustrated embodiment, the height of the specimen stage 16 is adjusted by a hydraulically driven telescopic cylinder 18, which is supported on a base 20. A mechanism for changing the angular position of the specimen stage 16 is explained below with reference to Figure 4.

[0079] The measuring device 14 comprises an optical coherence tomograph 22, which generates a measuring light beam 24 and whose construction is also explained in more detail below.

[0080] A scanning device, indicated by 26, deflects the measuring light beam 24 variably in two orthogonal scanning directions x, y. For this purpose, the scanning device 26 has a first scanning mirror 28, which is rotatably mounted about a first axis of rotation 30. A second scanning mirror 32 is rotatably mounted about a second axis of rotation 34, which is oriented perpendicular to the first axis of rotation 30. The scanning mirrors 28, 32 are driven by galvanometer drives (not shown), which are controlled by a control unit 36. Figure 3 shows the scanning device 26 with the scanning mirrors 28, 32 (shown here as rectangular) enlarged in a perspective and highly schematic representation.

[0081] The measuring device 14 also includes an image-side telecentric optical system 38, which is indicated in Figure 2 by three lenses L1, L2, and L3. The optical system 38 focuses the measuring light beam 24 deflected by the scanning device 26 such that it always strikes the surface 40 of the wafer 12 facing the optical system 38 at approximately a perpendicular angle. The respective point of impact of the measuring light beam 24, which depends on the position of the scanning mirrors 28, 32, defines a measuring point 39, 39'.

[0082] The optical coherence tomograph 22 includes a light source 42, a first beam splitter 44 which splits the light generated by the light source into the measuring light beam 24 and a reference light beam 46, a reference arm 48 for guiding the reference light beam 46, and an object arm 50 which uses the optical system 38 and the scanning device 26 and in which the measuring light beam 24 is guided.

[0083] During a measurement, the measuring light beam 24, propagating in the object arm 50, is focused onto the surface 40 of the wafer 12, where it is reflected predominantly specularly and to a lesser extent diffusely, and travels along the same light path back through the object arm 50 to the first beam splitter 44. There, the reflected portion of the measuring light beam 24 is superimposed with the reference light beam 46, which is guided in the reference arm 48 and reflected there by a mirror 52. Both light components are directed by a second beam splitter 54 onto a detector 56, which converts the optical interference signal into electrical signals.

[0084] In the illustrated embodiment, the optical coherence tomograph 22 is designed as an SD OCT (SD stands for spectral domain). The detector 56 therefore contains a spectrometer that records the spectral intensity distribution of the interference signal. From this, an evaluation unit 57 connected to the detector 56 can calculate, in a manner known per se, the distance of an optical interface, e.g., the surface 40 of the wafer 12, to the measuring device 14 (e.g., the lens L3) at the measuring point 39. For further details on the optical coherence tomograph, reference is made to DE 10 2017 128 158 A1 (corresponding to US 2018 / 0164089 A1).

[0085] The wavelength range of the light generated by the light source 42 can be selected such that the measurement beam 24 can penetrate at least partially into the wafer 12. In particular, the wavelength range lies in the infrared or near-infrared range. This corresponds to wavelengths between 950 nm and 2000 nm, with the wavelength range from 1000 nm to 1300 nm being preferred. Reflections then occur at both interfaces of the wafer 12, which are detected by the optical coherence tomograph 22. A spectral peak is obtained for each interface, from which the distance to the respective interface can be determined.

[0086] In addition, the optical coherence tomograph 22 can be used in a thickness mode. In this case, the evaluation unit 57 only evaluates interference signals generated by interference between measurement light components reflected at the two interfaces. In thickness mode, it is important to prevent light from propagating into the reference arm 48, as this would also interfere and thus generate unwanted noise signals. For this purpose, the reference arm 48 contains a switchable diaphragm, indicated at 60, which can be designed, for example, as a central or focal-plane shutter. When switching from distance mode to thickness mode, the switchable diaphragm 60 is automatically closed, preventing any light from the reference arm 48 from contributing to the interference on the detector 56.When switching back to distance mode, the switchable dimming device 60 clears the path for the reference light beam 46. In the embodiment shown in Figure 2, the measuring light 24 propagates completely in free space. In other embodiments, the light is guided partially in optical fibers.

[0087] Figure 4 schematically shows a possible mechanism for changing the angular position of the specimen stage 16. A first servo motor 62 drives a first shaft 66 extending along the x-direction via a first bevel gear 64. This shaft is guided in a bearing block 68 and is rotationally fixed to a saddle 70. This allows the saddle 70 to be moved about the axis A by actuating the first servo motor 62. x The first servo motor 62 and the bearing block 68 are supported by a carrier plate 71, on the underside of which the telescopic cylinder 18 engages.

[0088] The saddle 70 carries a second servomotor 72, which, like the first servomotor 62, is controlled by the control unit 36 ​​and drives a second shaft 76 extending along the y-direction via a second bevel gear 74. The second shaft 76 is guided in the saddle 70 and non-rotatably connected to a flange 80, which supports the stage 16. This allows the flange 80 to be rotated about the axis A by actuating the second servomotor 72. y be swerved.

[0089] 5. Alignment of the object stage - intensity measurement

[0090] Reliable measurement of the wafer 12 using the measuring device 14 is only possible if the measuring light beam 24 strikes the surface 40 of the wafer 12 at least approximately perpendicularly. If this is not the case, too little light returns to the coherence tomograph 22 to obtain a sufficiently strong interference contrast. The signal peaks in the Fourier spectrum then no longer stand out sufficiently from the background noise and therefore cannot be reliably identified.

[0091] Figure 5a shows a section of Figure 2 in a further simplified representation. It depicts the flange 80, which can be tilted about two axes, with the attached object stage 16 and a wafer 12 supported by it. In Figure 5a, the object stage 16 is aligned using the servomotors 62, 72 such that its reference plane 84, defined by the flat support surface for the wafer 12, is perpendicular to the optical axis OA of the optical system 38. Here, it is assumed that the wafer 12 has a TTV that gives it an approximately wedge-shaped cross-section. The wedge angle α is approximately 2°, so that a surface normal 86 of the surface 40 of the wafer 12 to the optical axis OA of the optical system 38 also forms an angle α of approximately 2°.When the axially parallel measuring light beam 24 strikes the surface 40 of the wafer 12, it is not reflected back onto itself due to the inclination of the surface 40, but is deflected laterally according to the law of reflection, as indicated by the dashed lines 24' in Figure 5a. Due to the small opening angle of the optical system 38, no light propagating at angles larger than the measuring light beam 24 can pass through the optical system 38. Consequently, in the configuration shown in Figure 5a, approximately half of the measuring light is lost for measurement due to the inclined surface 40.

[0092] Figure 5b schematically shows the spectrum l(f) obtained by the Fourier transform of the measurement signal generated by the detector. Since the measurement light beam 24 is reflected not only partially from the surface 40 of the wafer facing the optical system 38, but also from the lower surface 88, i.e., facing away from the optical system 38, two peaks PI and P2 are formed in the spectrum, with the first peak P1 corresponding to surface 40 and the second peak P2 to surface 88. In this example, the second peak P2 is somewhat weaker than the first peak PI because the intensity of the measurement light beam 24 is lower when it strikes the lower surface 88 than when it strikes the upper surface 40 due to reflection and absorption losses. If, as can be seen in Figure 5a, the focus 90 of the measurement light beam 24 lies on the lower surface 88, this effect can be attenuated or even compensated.However, due to the small opening angle, the depth of field of the optical system 38 is so large that the intensity of the measuring light beam 24 varies only slightly over the axial measuring range.

[0093] For a reliable evaluation, it is crucial how well the peaks PI and P2 stand out from the signal noise 92. In the situation indicated in Figure 5b, the peaks P1 and P2 are just barely reliably recognizable. As the angle α increases, the peaks PI and P2 eventually become lost in the signal noise 92.

[0094] To enable unambiguous identification of peaks PI and P2, the stage 16 is tilted according to the invention so that the measuring light beam 24 strikes the surface 40 of the wafer as perpendicularly as possible. Figure 6a shows the configuration shown in Figure 5 after a tilt of the stage 16 by 2°, indicated by an arrow. The tilting here occurs about the y-axis and is effected by actuating the second servomotor 72. In Figure 6b, a comparison with Figure 5b shows that the two peaks P1 and P2 now stand out more clearly from the signal noise 92.

[0095] After a further tilt of 2°, the configuration shown in Figure 7a is obtained. The surface 40 is now aligned perpendicular to the optical axis OA of the optical system 38, so that the peaks P1 and P2 stand out optimally from the signal noise 92.

[0096] Since the deviations of the wafer 12's shape from the ideal circular cylindrical shape are unknown and generally not visible to the naked eye, the angular position of the object stage 16 shown in Figure 7a is determined automatically in this embodiment. For this purpose, the control unit 36 ​​first activates one of the two servo motors 62, 72 so that it slowly rotates the object stage 16 around the corresponding axis A. x or A yThe system is tilted. During the tilting process, the distances to surfaces 40 and 88 are measured at a predetermined frequency, yielding a multitude of results, as exemplified in Figures 5b, 6b, and 7b. In this way, the angle at which peaks PI and P2 have the greatest intensity is determined. This angle is then set by the corresponding servo motor, and the measurement is repeated for the other axis by controlling the other servo motor. The result is two angles with respect to axes A. x , A y , which define an angular position of the object stage 16 for which the surface 40 is aligned perpendicular to the optical axis OA and thus the interference contrast is maximized.

[0097] With the stage 16 at its optimal angle, determined automatically or manually, the wafer 12 can now be fully measured using the measuring device 14. The measuring light beam 24 scans the surface 40 along a predefined path. To determine geometric properties such as TTV, warpage, or curvature, it is generally sufficient to scan the surface only along individual lines, as explained above with reference to Figure 1. However, if sufficient measuring time is available, the wafer 12 can also be scanned with a uniformly dense grid of measuring points, thus obtaining a complete surface or thickness profile.

[0098] Wafers 12 are received.

[0099] In the case of the wafer 12 shown in Figures 5 to 7, which has a flat surface 40 and a wedge-shaped cross-section, the change in the angular position of the stage 16 described above ensures that the measuring light beam strikes the surface 40 perpendicularly at all measuring points. However, the surface 40 of the wafer is usually not perfectly flat, but rather curved to a greater or lesser degree, as is also indicated in Figure 1. Therefore, a change in the angular position of the stage 16 generally only results in the surface 40 of the wafer 12 being oriented, on average, more perpendicular to the optical axis OA of the optical system 38 after the change in angular position than before the change in angular position.

[0100] In many cases, it is sufficient to perform the measurement described above at only a single measuring point, which may, for example, lie on the optical axis 38. If the surface 40 is inclined differently in a small area surrounding this single measuring point than outside this area, this would lead to an angular position of the stage 16 in which the aforementioned area can be measured optimally, but the measuring points lying outside this area cannot.

[0101] For more strongly curved surfaces, it may therefore be necessary to repeat the measurement at several measuring points. For this purpose, the control unit 36 ​​simply needs to control the scanning unit 26 accordingly. Since changing the angular position is much slower than moving the measuring point using the scanning unit 26, the preferred procedure is to perform the intensity measurement at all desired points with the object stage 16 at a specific angular position, then change the angular position of the object stage 16, measure the intensity again at all desired points with the object stage at the changed angular position, and so on.

[0102] 6. Alignment of the object table - preliminary distance measurement

[0103] The alignment of the object stage 16 described above is simple and reliable and can also be carried out manually, i.e., without control by the control unit 36, if necessary. However, this approach is comparatively time-consuming, as the angular position of the object stage 16 must be changed several times during the process.

[0104] If the measurement of wafer 12 is to be carried out as part of a production process, this time may not be available.

[0105] Based on Figures 9 to 11, a second approach is described below, which allows for a faster finding of an optimal angular position of the object stage 16.

[0106] For this purpose, the wafer 12 is first roughly measured in a preliminary measurement to approximately determine the shape of the surface 40. Once the approximate shape of the surface 40 is known, an angular position for the stage 16 can be calculated such that the surface 40 is oriented as perpendicular as possible to the optical axis OA of the optical system 38. The actual measurement with a denser grid of measuring points is only carried out after this alignment.

[0107] Figure 8 shows the wafer 12 with a uniform grid of measuring points 39 covering the entire surface e40 of the wafer 12. 50 to 500 measuring points, as in the illustrated embodiment, are sufficient in many cases, allowing the preliminary measurement to be carried out very quickly.

[0108] Figure 9a illustrates, in a schematic representation based on Figure 5a, a situation in which the shape of the wafer 12 deviates in a complex manner from the ideal circular cylindrical shape. Even in such cases, an angular position of the stage 16 can often be found using an optimization algorithm, which leads to a significantly improved (in the sense of being more perpendicular to the optical axis OA) orientation of the surface 40 on average.

[0109] In the case shown, however, this is not possible. If the larger central area of ​​the wafer 12 around the axial measurement point 39 is oriented approximately perpendicular to the optical axis OA, there are areas near the edges for which this is not even approximately true. When measuring at the edge-adjacent measurement points 39', 39" with dashed measurement light beams 24' and 24" respectively, the interference contrast can be so low that the peaks PI, P2 do not stand out clearly enough from the signal noise 92.

[0110] This problem can be solved by tilting the wafer 12 at least once during the high-resolution measurement using the stage 16. Figure 9b shows the wafer 12 from Figure 9a after the central area of ​​the wafer 12 has first been measured and then the angular position of the stage 16 has been changed. It can be seen that in this angular position, the surface 40 of the wafer 12 near the two measurement points 39', 39" is oriented approximately perpendicular to the optical axis OA. The measurement point 39" shown on the right and the measurement points in its vicinity can now be reliably measured.

[0111] However, this does not apply to measuring point 39' on the left side. Due to the tilting of the stage 16, the surface 40 of the wafer 12 is now too close to the optical system 38. Here, an additional axial compensation movement in the z-direction is required, which can be achieved with the help of the telescopic cylinder 18.

[0112] If the angular position of the object stage 16 and, if applicable, its z-position are changed during the measurement process, this must be taken into account when calculating the desired surface or thickness profile. For this purpose, the results of the partial measurements must be combined with the appropriate modifications.

[0113] The data is combined using well-known methods. In the simplest case, the known changes in position and inclination caused by the angle setting are simply compensated for mathematically. Alternatively, a more refined method is used. For this, the partial measurements must overlap slightly, and it is checked whether the measurement results largely agree in the overlapping areas, for example, by calculating a correlation factor. If necessary, the partial measurements are modified, for example, by shifting them or adding a correction value, to bring them into agreement.

[0114] Figure 10 illustrates a case where the surface 40 of the wafer 12 is so strongly curved that pre-measurement is not possible over large areas. These areas are shown in white in Figure 10, i.e., without measuring points 39. Similar to the procedure described above for the final high-resolution measurement, the angular position of the object stage 16 can also be changed during the pre-measurement so that successive distance measurements at all points are possible. Preferably, the angular position of the object stage 16 is changed automatically using the actuators 62, 72 until a distance can be measured for all measuring points intended for the pre-measurement. Figures 11a and 11b illustrate this process in diagrams similar to Figures 9a and 9b.In the angular position of the specimen table 16 shown in Figure 11a, measurements can only be taken for a central area, not for the outer measuring points 39', 39". After tilting the specimen table 16 to the right, it becomes possible to measure measuring point 39" and its surroundings. To also reach measuring point 39' indicated on the left, the specimen table 16 must be tilted accordingly to the left. The measurement results obtained in this way must also be combined with appropriate modifications during the preliminary measurement.

[0115] As an alternative to the tilt described above, the shape of the areas where measurement is not possible can often be determined by extrapolation or interpolation. For example, if a wafer has a rotationally symmetrical shape with a concave curvature, the tilt at the edge of the wafer may be so large that distance measurement is impossible. In this case, the shape in the edge region can easily be extrapolated.

[0116] An example of interpolation would be a central area where no measurement is possible. If the surrounding measurable area can be described, for example, as a paraboloid of revolution, the shape of the surface in the unmeasurable area can be easily interpolated. In other cases, it may suffice to consider only the just-measurable boundary adjacent to the unmeasurable area and to approximate the surface in the unmeasurable area using straight lines connecting opposite points on the boundary. Based on these lines, a plane can then be defined and aligned perpendicular to the optical axis OA using the stage 16 before the actual measurement begins.

[0117] Especially when the objects to be measured are not wafers, the preliminary measurement process can take too long because measurements must be taken in virtually all mechanically possible angular positions. It must be taken into account that the two axes A x A yare independent, so that for a specific angular position about one axis, all angular positions of the other axis must be "tried through". Figure 12 shows, in a representation based on Figure 9a, the object stage 16 with a highly curved object 12', whose surface is to be measured. The shape of the surface 40' is not determined here, or at least not solely, by means of a preliminary measurement, but rather on the basis of an image taken by a camera 96 ​​of the measuring device 14. In the illustrated embodiment, the camera 96 ​​is arranged laterally and captures the space between the object stage 16 and the optical system 38. However, it is also possible to design the beam path of the camera 96 ​​coaxially with the beam path of the measuring device 14, as disclosed in the unpublished German patent application with file number 10 2024 120424.9.

[0118] Figure 13 shows image 98 of object 12' taken by camera 96. Using image processing algorithms, the approximate shape of surface 40' can be derived from image 98. During the subsequent measurement, the angular position of the object stage 16 is changed several times so that the measuring light beam 24 can strike all areas of the surface approximately perpendicularly. In general, the changes in angular position are linked to corresponding adjustments in the z-position of the object stage, as already explained above.

[0119] 7. Fourth embodiment - Complete measuring system

[0120] Figure 14 shows a complete measuring system with a measuring device 14 according to the invention, which is vibration-isolated and housed in a preferably airtight sealed cabin 100. The cabin 100 is equipped with a door 102 through which an operator or a robot can place wafers 12 or other objects on the object table 16 and remove them again after measurement. It is also possible to connect the cabin 100 to an existing transport system of a production facility to place objects on the object table 16.

[0121] Also visible are an externally mounted control panel 104, which also serves to display recorded measurement data, and inside the cabin 100 several parts of the measuring device 14, identified by reference numerals and described in more detail above, including the optical system 38, which directs the measuring light beam 24 towards the stage 16. Arrows indicate the tiltability and height adjustability of the stage 16. The cabin 100 and dampers 106 for vibration-isolated mounting of the measuring device 14 within the cabin 100 protect the measuring device 14 from external influences and vibrations. If necessary, cleanroom conditions can be created inside the cabin 100 using known methods to prevent airborne particles from distorting the measurements.

[0122] In a preferred embodiment, the following procedure is used when measuring a wafer 12 or another object:

[0123] The wafer 12 is placed on the object table 16 and the door 102 is closed. Subsequently, an automatic two-stage adjustment of the object table 16 takes place to transfer the wafer into an optimized starting position for its measurement.

[0124] In a first step, the height of the object stage 16 is adjusted, taking two aspects into account. Firstly, the wafer 12 is positioned relative to the focus of the measuring light beam 24 in such a way as to obtain the strongest possible reflections and correspondingly strong interference signals.

[0125] Secondly, the wafer 12 is positioned relative to a plane whose axial position in the object arm 50 is determined by the optical path length in the reference arm of the coherence tomograph. This positioning is such that the wafer 12 is located entirely on one side of this plane (i.e., either above or below it). This significantly facilitates the unambiguous assignment of peaks in the spectrum of the signals generated by the coherence tomograph to interfaces and thus improves the reliability of the topography measurement. If both criteria cannot be met simultaneously, consideration can be given to either changing the position of the focus, e.g., using a liquid lens in the optical system 38, or changing the position of the plane, e.g., using an optical path length modulator arranged in the reference arm 48.

[0126] Details of this height adjustment can be found in the German patent application entitled "Method and device for measuring transparent multilayer objects with a coherence tomograph," filed by the applicant on the same day. In a second stage, the angle adjustment is carried out according to the inventive method described above. Preferably, the two stages are carried out in the sequence described above, i.e., beginning with the height adjustment. However, the reverse sequence is also possible.

[0127] The progress of the automated height and / or angle adjustment is preferably displayed on the control panel 104. Optionally, the setting can be confirmed or corrected by an operator.

[0128] After automated height and / or angle adjustment, the actual measurement of the object being measured takes place. This can involve surveying the topography, measuring the thickness of layers of wafer 12 (or another object), or both. In the case of a wafer 12, parameters such as deflection, warping, or TTV (thickness variation) are preferably determined and displayed on the control panel. Furthermore, defects in the object can be detected and displayed. The entire topography and / or thickness distribution can also be shown. Defects can be highlighted on the control panel display.

[0129] In a further step, it is possible to measure specific "regions of interest" (ROIs) even more precisely with higher resolution and / or different height and angle settings. The ROIs are selected either by an operator or by an algorithm, particularly through artificial intelligence or machine learning.

[0130] It is also possible to read identification information applied or burned onto the wafer 12 and display it on the control panel 104. For this purpose, either a separate camera can be used or the coherence tomograph can be employed, as disclosed in the German patent application filed by the applicant on October 24, 2024, under file number 10 2024 131 114.2.

Claims

PATENT CLAIMS 1. Method for measuring an object (12; 12'), comprising the following steps: a) providing an optical system (38) having an optical axis (OA) and a scanning device (26); b) providing an object stage (16) defining a reference plane (84) whose angle to the optical axis (38) is adjustable; c) placing the object (12; 12') on the object stage (16) such that a surface (40) of the object (12; 12') faces the optical system (38); d) An optical coherence tomograph (22) directs a measuring light beam (24) via the scanning device (26) and the optical system (38) onto at least one point (39; 39', 39") on the surface (40; 40') of the object (12; 12'), thereby obtaining several measured values; e) the angular position of the object stage (16) is changed depending on the measured values ​​obtained in step d) so that the surface (40; 40') of the object (12;12') or at least a region of the surface is oriented more perpendicular to the optical axis (OA) of the optical system (38) on average after the change in angular orientation than before the change in angular orientation; and f) the distances to an optical interface (40; 40'; 88) of the object (12; 12') and / or between two optical interfaces of the object are measured by guiding the measuring light beam (24) from the scanning device (26) over the surface (40; 40') of the object or over the region of the object.; 2. Method according to claim 1, wherein in step d) the measuring light beam (24) is directed several times to the same point (39) on the surface (40) of the object, but with differently set angles of the object table (16); The optical coherence tomograph (22) detects the intensity of a portion of the measuring light that was reflected at point (39) and returns to the optical coherence tomograph (22); and in step e) the angular position of the stage (16) is changed so that the intensity of the portion of the measuring light detected by the optical coherence tomograph (22) is increased.

3. Method according to claim 2, wherein the angular position of the object table (16) is automatically changed using actuators (62, 72) depending on the detected intensity.

4. Method according to claim 1, wherein in step d) the measuring light beam (24) is successively directed to different points (39) on the surface (40) of the object (12); - the distance to the surface (40) of the object (12) is measured in each case; and an approximate shape of the surface of the object (12) is calculated from the distances.

5. Method according to claim 4, wherein, if the distance to the interface cannot be measured for one of the different points (39', 39"), the angular position of the object table (16) is adjusted.

6. Method according to claim 5, wherein the angular position of the object table (16) is automatically adjusted by means of actuators (62, 72) until the distance to the interface can be measured for one of the different points (39', 39").

7. Method according to claim 5, wherein, after the angular position of the object table (16) has been adjusted, step d) is repeated by directing the measuring light beam (24) successively to different points (39) on the surface (40) of the object (12), and step d) is repeated with different angular settings of the object table (16) until the distance to the interface can be measured for each of the different points (39', 39").

8. Method according to one of claims 5 to 7, wherein in step f) the distances to the optical interface of the object (12) are measured over the entire surface of the object, and wherein a shape of the surface is calculated from the measured distances, taking into account computational changes in the angular position of the object table (16) made during the measurement.

9. Method according to one of the preceding claims, wherein a camera (96) after step c) takes a picture (98) of the object (12') and an image processing algorithm calculates, on the basis of the taken picture (96), how the surface of the object (12') is inclined at several locations relative to the optical axis (OA).

10. Device (14) for measuring an object (12; 12'), comprising: a) an optical system (38) having an optical axis (OA), b) a scanning device (26), c) a stage (16) defining a reference plane (84) whose angle to the optical axis (OA) is adjustable, wherein the object (12; 12') can be placed on the stage (26) such that a surface (40; 40') of the object faces the optical system (38), d) an optical coherence tomograph (22) configured to direct a measuring light beam (24) via the scanning device (26) and the optical system (38) to at least one point (39) on the surface of the object, thereby obtaining multiple measured values, e) a control device (36) configured to change the angular position of the stage (16) depending on the measured values ​​obtained, such that the surface (40; 40') of the object (12;12') or at least an area of ​​the surface is oriented more perpendicular to the optical axis (OA) of the optical system (38) after the change in angular position than before the change in angular position, and; After changing the angular position of the object stage (16), the measuring light beam (24) from the scanning device (26) is directed over the surface (40; 40') of the object (12; 12') or over the area of ​​the object in order to measure the distances to an optical interface of the object and / or between two optical interfaces of the object.

11. Measuring system with a device (14) according to claim 10 and with an airtight sealed cabin (100) in which the device (14) is held in a vibration-isolated manner.

Images