Method and system for aligning two oppositely arranged optical subsystems and camera chip

Abstract

Method for aligning two oppositely arranged optical subsystems (12u,12u',12u'',12o,12o',12o'') of an optical system (11,11',11"), comprising the following steps: - Projection of alignment marks (9l, 9r, 9l', 9r') into an image plane (5u) of the first optical subsystem (12u, 12u', 12u''), - Projection of the alignment marks (9l,9r,9l',9r') from the first image plane (5u) onto a sensitive surface (14so) of the second optical subsystem (12o, 12o', 12o''), - Alignment of the optical subsystems (12u,12u',12u'',12o,12o',12o'') to each other, such that projections (9pl,9pr) of the alignment marks (9l,9r,9l',9r') are imaged at ideal positions (9il,9ir) in a depth of field area of ​​the sensitive surface (14so), characterized in that the alignment marks (9l, 9r) are applied directly to a first sensitive surface (14su) of a first camera chip (2u), wherein the alignment marks (9l, 9r) have a collection of non-functional pixels.

Description

[0001] In the semiconductor industry, aligners are used to align substrates, especially wafers, so that they can be bonded together in a subsequent process step. The substrates are aligned using alignment markers located on their surfaces. These alignment markers, located on opposite sides of the substrates, are complementary to each other.

[0002] US5940528A discloses a positioning method in which multicolored light can be used as a light source for alignment purposes. US5100237A discloses a device for projecting a mask pattern onto a substrate using a projection lens system. DE69012874T2 discloses a device for projecting a mask pattern onto a substrate.

[0003] US 2008 / 0292177 A1 discloses a method for aligning the top side of a substrate with alignment marks on the back side of the substrate. DE102005013755A1 discloses a method for manufacturing composite lens systems by calculating the centering errors. DE102013004738A1 discloses a method for the targeted adjustment of optical components to a reference axis.

[0004] Alignment processes in which the alignment marks are located on the substrate surfaces to be bonded are referred to as face-to-face alignments. If the substrates are not transparent to the electromagnetic radiation being measured, methods must be developed to measure the marks before the substrates are brought into close proximity.

[0005] A long-established method involved inserting a camera between the substrates, which was then removed before the two substrates were brought into close proximity. This had several disadvantages. First, the camera inevitably resulted in particle contamination on at least the lower substrate. Second, the distance between the substrates had to be relatively large to accommodate the camera. This large distance meant that the path required to move the two substrates when they were brought into close proximity after the camera was removed was very long, and the substrates shifted laterally as they approached each other, thus deviating from their previous, optimal alignment.

[0006] An improvement in face-to-face alignment is represented by the alignment system described in US6214692B1. In this disclosed alignment system, the use of a camera that would need to be inserted between the substrates was completely eliminated. Instead, two optical groups, each with two opposing optics, were used to create a system with two reference points relative to which the substrates were positioned. The reference points were the intersections of the optical axes of the two opposing optics. The corresponding process for calibrating such optics was described extensively in the Fig. 2c and Fig. 2e of the printed document WO 2014 / 202106 A1 or the Fig. 5a and Fig. 5b of the printed document WO 2015 / 082020 A1 is disclosed and described.

[0007] The problem with the state of the art, in particular the calibration procedure according to the Fig. 2c and Fig. 2e of the printed document WO 2014 / 202106 A1 or according to the Fig. 5a and Fig. Section 5b of publication WO 2015 / 082020 A1 stipulates that a calibration substrate is required against which the optical axes of the optics are calibrated. The upper and lower optics are calibrated to an alignment mark in the substrate, which is typically transparent. This aligns the left optical axes to a left focus point and the right optical axes to a right focus point.

[0008] It is therefore the object of the present invention to provide a method and a system which do not have the disadvantages of the prior art and in particular with which two optics can be aligned or calibrated to each other in a simplified manner.

[0009] This problem is solved by the features of the dependent claims. Advantageous embodiments of the invention are specified in the sub-claims.

[0010] The invention describes in particular how the two image planes of the two opposing optics and / or the optical axes of the two optics can be aligned relative to each other.

[0011] The use of a calibration substrate according to publication WO 2014 / 202106 A1 can be advantageously dispensed with.

[0012] According to the invention, a method for aligning two optical subsystems of an optical system arranged opposite each other, in particular one above the other, is provided, comprising the following steps: - Projection of alignment marks into an image plane of the first, especially lower, optical subsystem, - Projection of the alignment marks from the first image plane onto a sensitive surface of the second, especially upper, optical subsystem, - Alignment of the optical subsystems with each other, so that projections of the alignment marks are imaged at ideal positions within a depth-of-field area of ​​the sensitive surface.

[0013] The depth of field and / or the depth of focus ranges of all optical elements, in particular the optics, are between 1 nm and 10 mm, preferably between 10 nm and 1 mm, more preferably between 50 nm and 500 µm, most preferably between 500 nm and 250 µm, and most preferably between 1 µm and 100 µm.

[0014] In a preferred embodiment, light for illuminating the alignment marks is coupled into the first and / or second optical subsystem via at least one light source, in particular a lamp, preferably a halogen lamp, more preferably an LED, most preferably a laser light source, and at least one mirror. With such an external light source, a particularly bright illumination of the alignment marks can advantageously be achieved.

[0015] According to another preferred embodiment, ambient light is coupled in at least one location of the optical subsystems, in particular at a first camera and / or at a mirror, to illuminate the alignment marks, preferably without shielding the optical subsystems. This advantageously makes it possible to illuminate the alignment marks without additional external light sources.

[0016] It is disclosed that the alignment marks are projected by an alignment mark projection system onto a first sensitive surface of a first camera chip of a first camera, wherein the alignment marks in particular comprise an LED array and / or are configured as masks. This advantageously eliminates the need to integrate the alignment marks into the camera chip.

[0017] According to another preferred embodiment, the optical subsystems comprise two opposing optics whose image planes and / or optical axes are aligned with each other.

[0018] This allows for a particularly efficient alignment of the optical subsystems or optics.

[0019] According to another preferred embodiment, at least the optics are provided for as part of a system for aligning substrates, in particular wafers. This advantageously allows for a significant improvement of a system for aligning substrates.

[0020] According to another preferred embodiment, the method according to the invention has the following sequence: - Arrangement of alignment marks on a first sensitive surface of a first camera chip of a first camera and / or projection of alignment marks onto a first sensitive surface of a first camera chip of a first camera, - Projection of the alignment marks from the first camera chip via the first optical subsystem into the first image plane, wherein the first optical subsystem comprises the first optics, and wherein the first image plane is arranged between the optics, - Projection of the alignment marks from the first image plane via the second optical subsystem onto the second sensitive surface of a second camera chip of a second camera, wherein the second optical subsystem comprises the second optics, - in particular, alignment of the optics to each other, so that projected alignment marks of the alignment marks are sharply imaged on the second camera and the positions of the projected alignment marks correspond to desired and / or ideal positions.

[0021] According to another preferred embodiment, the optics are moved translationally and / or rotationally relative to each other for alignment. This advantageously allows for particularly precise alignment.

[0022] According to another preferred embodiment, the optics are positioned so close together for alignment that the first image plane of the first optic lies within the depth of field of the second object plane of the second optic. This advantageously allows for a particularly sharp image of the alignment marks.

[0023] According to another preferred embodiment, the alignment marks are applied directly to a first sensitive surface of a first camera chip of a first camera, particularly by means of lithographic methods. This advantageously enables a particularly efficient generation of the alignment marks.

[0024] According to the invention, the alignment marks are applied directly to a first sensitive surface of a first camera chip, wherein the alignment marks comprise a collection of non-functional pixels, in particular in a cross shape. This advantageously enables a particularly simple generation of the alignment marks.

[0025] It is disclosed that the alignment marks are applied to a separate marking plate in a first camera chip of a first camera, wherein the distance between the alignment marks and a first sensitive surface of the first camera chip is less than 1 mm, preferably less than 100 µm, more preferably less than 10 µm, most preferably less than 1 µm, and most preferably less than 0.1 µm. This advantageously eliminates the need for any machining of the camera chip to generate the alignment marks.

[0026] Another object of the present invention relates to a system for aligning two optical subsystems of the optical system arranged opposite each other, in particular one above the other, in particular with a method according to one of the preceding claims, comprising: - A first camera with a first camera chip for projecting alignment marks into an image plane of the first, in particular lower, optical subsystem, - Means for projecting the alignment marks from the first image plane onto a sensitive surface of the second, especially upper, optical subsystem, - Means for aligning the optical subsystems to each other, so that projections of the alignment marks can be imaged at ideal positions within a depth-of-field range of the sensitive surface.

[0027] The embodiments relating to the method also relate to the system according to the invention.

[0028] Another aspect of the present invention relates to a camera chip with alignment marks near and / or on its sensitive surface, particularly for a system and / or method according to one of the preceding claims. The description of the method also relates to the camera chip according to the invention. The optical system

[0029] The embodiment according to the invention relates to an optical system comprising at least a first, lower optical subsystem and a second, upper optical subsystem. In the remainder of this publication, in consideration of the preferred structure, all entities will be referred to only by the words "upper" or "lower". The lower optical subsystem and the upper optical subsystem each have at least one optical element, and in particular several optical elements. The optical elements are, in particular, □ Mirrors, especially ◯ Plan mirror ◯ Convex mirror ◯ Concave mirror □ Lenses ◯ Convex lenses ▪ Biconvex ▪ Planar-arconvex ▪ Concave-convex ◯ Concave lenses ▪ Biconcave ▪ Planar concave ▪ Convex-concave ◯ Fresnel lenses □ Prisms □ Diffraction elements ◯ Diffraction grating □ etc.

[0030] The mirrors are preferably cold-light mirrors (hereinafter also referred to as cold mirrors) that filter out infrared light from the coupled light in order to prevent unnecessary and undesirable heating of the optical system. In particular, light with a wavelength between 1000 µm and 0.8 µm, preferably between 750 µm and 0.8 µm, even more preferably between 500 µm and 0.8 µm, most preferably between 100 µm and 0.8 µm, and most preferably between 50 µm and 0.8 µm, is filtered.

[0031] The embodiment according to the invention is described by way of example such that the objects to be projected, in particular alignment marks, are projected from a lower camera chip, in particular from its sensitive surface, onto an upper camera chip, in particular its sensitive surface.

[0032] It is also conceivable that the alignment marks are projected from an upper camera chip, particularly its sensitive surface, onto the lower camera chip, particularly its sensitive surface. For the sake of completeness, the extreme case is also mentioned in which alignment marks are located on both camera chips, particularly their sensitive surfaces, and are projected alternately onto the opposite camera chip or its sensitive surface. It will be shown later in the description that only one projection direction is relevant to generate the effect according to the invention.

[0033] In the remainder of this publication, only the exemplary case will be discussed, in which lower alignment marks are projected onto the sensitive surface of an upper camera chip.

[0034] The embodiment according to the invention requires a lower camera chip that has alignment marks near its sensitive surface, preferably directly on the sensitive surface. The alignment marks can either be physical or they can be projected onto the lower camera chip. In the latter case, these are projected onto the sensitive surface of the lower camera chip by an alignment mark projection system. This embodiment is the most preferred of all embodiments, since it does not require camera chips that already have corresponding alignment marks, but rather allows alignment marks to be projected onto any commercially available chip.

[0035] In a first embodiment of the invention, the alignment marks are manufactured directly on the lower camera chip. Specifically, the alignment marks are located on the sensitive surface of the lower camera chip. The production of the alignment marks on the lower camera chip should preferably be carried out by the chip manufacturer and is hardly feasible for customers purchasing such a chip. One possibility would be to disassemble the purchased lower camera chip so that its sensitive surface is completely exposed. Then, using well-known lithographic methods, appropriate lower alignment marks could be applied, and the disassembled lower camera chip could subsequently be reassembled.However, removing the protective glass for the sensitive surface often results in the destruction of the camera chip, especially its sensitive surface itself, and is therefore usually not possible without causing damage.

[0036] In a second, preferred embodiment of the invention, the lower alignment marks are not manufactured directly on the lower camera chip, but on a carrier (also called a marking plate) which can be mounted over the encapsulated lower camera chip using fixing means. Care must be taken to ensure that the lower alignment marks are located as close as possible to the surface of the sensitive area of ​​the lower camera chip so that they are always within the depth of field of the recording lower camera, which is located on the opposite side of the optical path.

[0037] The distance between the alignment marks and the sensitive surface should be less than 1 mm, preferably less than 100 µm, more preferably less than 10 µm, most preferably less than 1 µm, and most preferably less than 0.1 µm. Encapsulated camera chips can usually be disassembled relatively easily, except for the last cover plate of the photosensitive area, so that positioning a carrier according to the invention with alignment marks close to the photosensitive surface is feasible.

[0038] In a third, less preferred embodiment of the invention, the lower alignment marks are projected onto the photosensitive surface of a lower camera chip. This has the advantage that any lower camera chip can be provided with lower alignment marks. To ensure that the lower alignment marks projected onto the lower camera chip are in the ideal positions, they must be calibrated by the alignment mark projection system before being projected onto the upper camera chip via the optical path of the optical system to align the two camera chips relative to each other.

[0039] The essential aspect of the invention is to project the lower alignment marks of the lower camera chip onto the photosensitive surface of the upper camera chip, which is opposite the lower camera chip in the optical path. To achieve a correct, sharp projection of the lower alignment marks onto the upper camera chip, i.e., to obtain sharp projected upper alignment marks, the optical elements, in particular the lower optical subsystem and the upper optical subsystem, must be calibrated relative to each other. The calibration is complete when the projected upper alignment marks are located at the ideal positions on the upper camera chip.

[0040] Each of the two optical subsystems has an object plane and an image plane that are conjugate to each other. Conjugate planes are planes that are mathematically related to each other via the mapping rule of the Fourier transform. Two planes are conjugate to each other if they are perpendicular to the optical axis of an optical ray path and map objects from the object space to the image space and vice versa.

[0041] Since we are considering, as an example, the projection of lower alignment marks of the lower camera chip onto the upper camera chip, the lower photosensitive surface of the lower camera chip is the first object plane.

[0042] The real or projected alignment marks of the alignment mark projection system on the lower photosensitive surface of the lower camera chip are preferably located in the object plane. Generally, the real alignment marks only need to be within the depth of field around the object plane.

[0043] The lower optical subsystem projects these alignment marks onto its image plane. This image plane is a first, lower image plane. It is always located between the two optics whose optical paths must be calibrated relative to each other. The second optical subsystem also has an optical plane between the optics. This optical plane acts as the object plane and is responsible for projecting the alignment marks of the photosensitive surface of the lower camera chip, which are projected onto the lower image plane, onto the second, upper image plane of the upper optical subsystem. The upper image plane therefore coincides with the sensitive surface of the upper camera chip.In order for the image of the alignment marks projected onto the lower image plane to be mapped onto the upper image plane with as little loss as possible and, above all, with sharpness, the lower image plane and the upper object plane must be as congruent as possible.

[0044] Due to the fact that the lower image plane of the lower optical subsystem and the upper object plane of the upper optical subsystem are generally not congruent or even lie within the depth of field of each other's optical plane, a modification of the optical subsystems, in particular the optical elements, is necessary. This modification requires, in particular, a translational and / or rotational adjustment of the optical elements of the lower and / or upper optical subsystem, especially a translational and / or rotational modification of the optics. The aim of the embodiment according to the invention is not only to project the lower alignment marks of the lower camera chip sharply onto the photosensitive surface of the upper camera chip, but also to adjust the optical elements in the optical system until the positions of the projected alignment marks correspond to the theoretically desired, ideal positions.Preferably, the same camera chips are used in the lower and upper optical subsystems. This means that the positions of the ideally projected alignment marks correspond to the same detector coordinates as the alignment marks to be projected on the lower camera chip. The optical systems are thus modified and adjusted so that the marks lie within a freely selectable tolerance range with respect to the ideal position. process

[0045] The process according to the invention can be used, in particular, for calibrating the optics in the systems described in publications US6214692B1, WO 2011 / 042093 A1, WO 2014 / 202106 A1 and WO 2015 / 082020 A1. However, this does not limit the generality of the process. In general, the process according to the invention can be used for the calibration, in particular fully automatic, of any two optics.

[0046] In a first process step according to the invention, the two optical planes are brought closer together so that the lower image plane is within the depth of field of the upper object plane, or vice versa. In this first process step, the projected left upper alignment mark or the right upper alignment mark is not necessarily in focus. In particular, both upper alignment marks do not necessarily have to be within the field of view of the upper camera. The fields of view of the cameras are between 10 µm and 50 mm, preferably between 50 µm and 25 mm, more preferably between 100 µm and 15 mm, most preferably between 250 µm and 10 mm, and most preferably between 300 µm and 5 mm.

[0047] In a second process step according to the invention, the lower and / or upper optical subsystem is modified such that at least both projected alignment marks are within the field of view of the upper camera. A modification of the lower and / or upper optical subsystem is understood to mean, in particular, a translational and / or rotational and / or functional modification of at least one optical element in the lower and / or upper optical subsystem. In particular, this refers to a translational and / or rotational movement of the lower and / or upper optics.

[0048] In a third process step according to the invention, wedge error compensation is performed between the two optical planes, or the angle between the optical axes is reduced. Wedge error compensation is carried out by modifying at least one optical element in the lower and / or upper optical subsystem such that the inclination between the two optical axes of the two optical subsystems is reduced. In particular, the wedge error is reduced by rotating at least one of the two optics.

[0049] In a fourth process step according to the invention, after all alignment marks are located within the depth of field and the field of view of the upper camera chip and the wedge error has been largely reduced, fine adjustment is carried out between the alignment marks by moving the projected alignment marks on the upper camera chip into the position of the ideal positions. This is done in particular by translational displacement and / or rotation of the lower and / or upper optics.

[0050] In particular, process steps one to four according to the invention can be carried out in any order and / or simultaneously. Simultaneous execution of all process steps according to the invention is ensured by appropriate firmware and / or hardware and / or software that can calibrate the two optical subsystems to each other, in particular fully automatically by means of an algorithm. In particular, the calibration can be carried out automatically by means of a control loop.

[0051] The process steps one to three according to the invention are, in particular, coarse calibration process steps that only very rarely need to be carried out completely, especially after maintenance, assembly, or relocation of the embodiment according to the invention. After a coarse calibration has been performed, usually only a fine calibration according to process step four is necessary for an extended period.

[0052] The optical system calibrated according to the invention can then be used for measuring surfaces. In particular, the system is suitable for aligning two substrates relative to each other. In a particularly preferred embodiment, the system is used for face-to-face alignment of two substrates relative to each other.

[0053] Another possible use of the optical system calibrated according to the invention is in measuring the top and bottom surfaces of a substrate or substrate stack. In this specific case, the optical system is used as part of a metrology tool. Simultaneous measurement of the top and bottom surfaces of a substrate or substrate stack is usually performed to measure the horizontal distance between two surface features. camera chip

[0054] The camera chip is configured as a planar detector. Preferably, the camera chip is a CCD detector, a CMOS detector, an analog detector, a four-quadrant detector, or a so-called position sensing device (PSD). The camera chip has a readout frequency between 1 Hz and 1 MHz, preferably between 10 Hz and 100,000 Hz, more preferably between 20 Hz and 10,000 Hz, most preferably between 30 Hz and 1,000 Hz, and most preferably between 40 and 100 Hz. Readout frequency refers to the number of complete interference patterns that the camera chip can read out per second.

[0055] The horizontal pixel resolution of the camera chip is in particular more than 10 pixels / cm, preferably more than 100 pixels / cm, more preferably more than 1000 pixels / cm, most preferably more than 10000 pixels / cm, and most preferably more than 100000 pixels / cm.

[0056] The vertical pixel resolution of the camera chip is in particular more than 10 pixels / cm, preferably more than 100 pixels / cm, more preferably more than 1000 pixels / cm, most preferably more than 10000 pixels / cm, and most preferably more than 100000 pixels / cm.

[0057] The distance between two pixels is between 0.1 µm and 100 µm, preferably between 0.5 µm and 50 µm, more preferably between 1 µm and 25 µm, most preferably between 2.5 µm and 10 µm, and most preferably 5 µm.

[0058] In a particular embodiment of the invention, the optical system can have two cameras with two different camera chips. The camera chips could, for example, differ in size. In this case, the use of alignment marks can even be dispensed with. The calibration according to the invention is then carried out by projecting the contours of one camera chip onto the other. By means of optical elements that enable scaling, the camera chip to be projected can be mapped perfectly onto the surface of the other camera chip. In general, an affine transformation of one camera chip onto the other camera chip takes place, which is characterized by the fact that the surfaces of the camera chips are congruent.Such an affine transformation is always achievable through the mathematical operations of translation, rotation, and scaling, which are physically implemented in the optical system by the optical elements. The two differently sized chips preferably have the same pixel size and / or the same pixel resolution, i.e., the same distance between the pixels. This embodiment of the invention is particularly preferred because it eliminates the need to project alignment marks onto the chip surfaces, or because the contours of the camera chips can be considered alignment marks. The chip, and in particular its contour, itself serves as the alignment mark, resulting in further cost savings in the construction of the device according to the invention, as no alignment mark projection system is required.Furthermore, the device according to the invention becomes completely independent of the chip manufacturer, since the latter does not need to provide any alignment marks, while the contour of a chip is an inherent part of its structure. If the contour of a chip is used as an alignment mark, the chip should be detected as centrally as possible by the optical systems, i.e., the center of gravity of the chip should be as close as possible to the optical axis. In particular, the distance between the optical axis and the center of gravity of the chip is less than 1 mm, preferably less than 0.5 mm, even more preferably less than 0.1 mm, most preferably less than 0.01 mm, and most preferably less than 0.001 mm.

[0059] For all the aforementioned embodiments of the invention for calibrating two optical subsystems relative to each other, it is true that in practice, perfect calibration is either not possible or, particularly for economic reasons, not desirable. Extremely precise calibration can take a considerable amount of time, during which other processes performed with the system cannot take place. Generally, a calibration process can be aborted as soon as a defined limit value is reached or exceeded. In a further extension of the invention, it is conceivable that the coordinates of the two alignment marks, which are not perfectly calibrated relative to each other, are stored with respect to a common coordinate system. This common coordinate system could, for example, be the lower detector coordinate system or the upper detector coordinate system.It is also conceivable that the coordinates of the two alignment marks, which are not perfectly calibrated to each other, are stored in any other coordinate system. Due to the failure to achieve and / or the undesired perfect alignment of the alignment marks, they are generally still slightly shifted and / or rotated relative to each other after the calibration process. This shift and / or rotation can be measured, stored, and used in a later process, particularly an alignment process of two substrates, to correct the determined coordinates of the alignment marks on the substrates.

[0060] Further advantages, features, and details of the invention will become apparent from the following description of preferred embodiments and from the drawings. These show: Fig. 1a a schematic, not to scale sketch of a first optical system according to the invention with a first light coupling, Fig. 1b a schematic, not to scale sketch of a second optical system according to the invention with a second light coupling, Fig. 1c a schematic, not to scale sketch of a third optical system according to the invention with a third light coupling, Fig. 2a a schematic, not to scale simplified sketch of a first process step according to the invention, Fig. 2b a schematic, not to scale simplified sketch of a second process step according to the invention, Fig. 2c a schematic, not to scale simplified sketch of a third process step according to the invention, Fig. 2d a schematic, not to scale simplified sketch of a fourth process step according to the invention, Fig. 3a a schematic, not to scale simplified, enlarged sketch of a lower sensitive surface, Fig. 3b a schematic, not to scale simplified, enlarged sketch of a fourth process step according to the invention in the initial state, Fig. 3c a schematic, not to scale simplified, enlarged sketch of a fourth process step according to the invention in the final state, Fig. 4a a schematic, not to scale simplified isometric exploded view of a first camera chip according to the invention and Fig. 4b a schematic, not to scale simplified isometric exploded view of a second camera chip according to the invention.

[0061] In the figures, identical components or components with the same function are marked with the same reference symbols.

[0062] In the figures shown, all optical paths 7, 7', 8 are depicted schematically and are not to be understood as paths strictly conforming to the laws of optics. Furthermore, the optical paths 7, 7', 8 and the optical subsystems 12u, 12u', 12u'', 12o, 12o', 12o'' are represented in such a way that no mutual, mirror-image projection of camera chips 2u, 2o occurs. This particularly facilitates the understanding of the coincident detector coordinate systems. It is clear to those skilled in the art that in a real embodiment, the mutual imaging of the camera chips 2u, 2o can be, and preferably is, mirror-image.

[0063] The Fig. Figure 1a shows a schematic, not-to-scale sketch of a first optical system 11 according to the invention, comprising two optical subsystems 12u, 12o with two light paths 7, 8. Light path 7 represents an illumination path for a lower camera chip 2u ​​of a lower camera 1u. Light path 8 is the imaging path of the alignment marks 9l, 9r.

[0064] The optical system 11 comprises several mirrors 3 and optical elements 4, 4', such as mirrors, lenses, prisms, etc. Light 7 from a light source 10 is coupled into the optical system 11. The light 7 passes through the optical system 11 with the mirrors 3 and optical elements 4, 4' and exits the optical system 11, striking a photosensitive surface 14su of the lower camera chip 2u ​​of the lower camera 1u.

[0065] In this first embodiment according to the invention, the light from the illumination path 7 is coupled into the optical system 11 via the source 10, in particular a lamp, more preferably a halogen lamp, and a mirror 3. The light from the illumination path 7 passes through two optics 60, 6u and is projected onto the camera chip 2u ​​of the lower camera 1u via further optical elements, in particular further mirrors 3.

[0066] The coupling of light into the optical system 11 is of course also conceivable from any other mirror 3 in the optical system 11, especially when it comes to intensely illuminating a camera chip 20 of an upper camera 10. It is also conceivable to couple light into the optical system 11 via multiple sources 10 and multiple mirrors 3.

[0067] The Fig. Figure 1a also represents the imaging path 8, which maps the surface of the camera chip 2u ​​or its surroundings onto the camera chip 2o via optical elements 4, 4', 3. In this specific case, the camera chip 2u ​​represents the object plane and the camera chip 2o the image plane of the entire optical system 11. In the sectional view AA, two alignment marks 9l, 9r can be seen.

[0068] In the first embodiment according to the invention, these are, for example, alignment marks 91, 9r according to Fig. 4a, which were applied directly to the photosensitive surface 14su of the lower camera chip 2u. The alignment marks 91, 9r are preferably a collection of non-functional pixels, particularly in a cross shape. These pixels are non-functional because they have been coated, not connected, destroyed, inactivated, etc.

[0069] It is also conceivable that these are alignment marks 91, 9r according to Fig. 4b. In this case, the alignment marks 91, 9r are not located directly on the photosensitive surface 14su of the lower camera chip 2u, but at least in its vicinity. In any case, the alignment marks 9l, 9r are real physical entities.

[0070] The alignment marks 9l, 9r are now projected via the imaging path 8 onto the upper camera chip 2o, resulting in projected alignment marks 9pl, 9pr on the photosensitive surface 14so of the upper camera chip 2o. In the Fig. 1a The cameras 1u, 1o and the optical elements in projection path 8 are already perfectly calibrated to each other, so that the projected alignment marks 9pl, 9pr appear sharp at the correct, ideal positions in section BB. In particular, it can be seen that the image plane 5u of the lower optical subsystem 12u and the object plane 5o of the upper optical subsystem 12o are at a distance d from each other. For clarity, the distance d was chosen to be large enough to clearly distinguish the image plane 5u and the object plane 5o from each other. To produce sharp projected alignment marks 9pl, 9pr, the image plane 5u and the object plane 5o must each lie within the depth of field of the other plane.

[0071] The Fig. Figure 1b shows a schematic, not-to-scale sketch of a second optical system 11' according to the invention, comprising two optical subsystems 12u', 12o' with two light paths 7', 8. The light path 7' represents the illumination path for the lower camera chip 2u ​​of the lower camera 1u. In this embodiment of the invention, the light from the illumination path 7' is not coupled into the optical system 11' via a source 10, but simply from the environment, in particular directly. It is therefore scattered light, which is granted access to the optical system 11' by intentionally omitting corresponding shields of the optical system 11', in particular at least one location, and more preferably at the camera 1u. The coupling of light into the optical system 11' is also conceivable from any other location in the optical system 11', in particular at the mirrors 3. It is also conceivable to couple light into the optical system 11' at several locations.Alternatively, lighting could be provided in the vicinity of the camera, preferably using LEDs as light sources.

[0072] The Fig. Figure 1b also represents the imaging path 8, which maps the surface of the camera chip 2u ​​or its surroundings onto the camera chip 2o via optical elements 4, 4', 3. In this specific case, the camera chip 2u ​​represents the object plane and the camera chip 2o the image plane. Sections AA and BB show, as already described in the Fig. 1a describes an optimal imaging situation.

[0073] The Fig. Figure 1c shows a schematic, not to scale sketch of an optical system 11'', consisting of two optical subsystems 12u'', 12o'' with two light paths 7'', 8.

[0074] In a first embodiment of the invention, the light path 7'' is generated by alignment marks 91, 9r on a surface 19, wherein the alignment marks 91, 9r also simultaneously function as light sources 10 (view AA). It would be conceivable, for example, that the alignment marks 91, 9r consist of an LED array, in particular a cross-shaped one.

[0075] In a further, second embodiment according to the invention, the surface 19 on which the alignment marks 91, 9r are located emits light and the alignment marks 91, 9r do not function as active light sources 10 but only as masks. In this case, the alignment marks 91, 9r would be projected as shadow marks over the illumination path 7''.

[0076] The embodiment according to the invention projects the alignment marks 91, 9r of the surface 19 onto the camera chip 2u ​​(view BB). The photosensitive surface 14su of the camera chip 2u ​​in turn represents the object plane for the camera 1o. The alignment marks 9l', 9r' projected by an alignment mark projection system 21 are thus projected via the optical system 11" as alignment marks 9pl, 9pr onto the photosensitive surface 14so of the upper camera chip 2o of the upper camera 1o.

[0077] The difference to the embodiments according to the Fig. 1a and Fig. 1b consists in the fact that neither of the two camera chips 2u, 2o themselves needs to have alignment marks 91, 9r, but that these are projected onto the photosensitive surface 2u of the lower camera 1u via the alignment mark projection system 21 according to the invention, as projected alignment marks 9l' and 9r', and thus introduced into the optical system 11''.

[0078] The depictions in the Fig. 1a-1c are ideal in that the optical elements 3, 3o, 3u, 4, 4', 6u, 6o, 1u, 1o have already been perfectly aligned with each other. In the further Fig. 2a-2d as well as in the Fig. 3a, Fig. Section 3b explains the individual process steps that must be carried out to achieve this ideal state and to calibrate the optical system 11, 11', 11". In particular, for the sake of clarity, the representation of the light paths, especially the illumination path 7, 7' and the imaging path 8, and the optical elements is largely omitted. Fig. 2a-2d represent an optical system 11 as an example and representative. Fig. 1a. The Fig. Figure 2a shows a schematic, not to scale, simplified sketch of the optical system 11 according to the invention in a first process step 101 according to the invention, consisting of the two optical subsystems 12u, 12o, according to Fig. 1a. The characteristic feature of this representation consists primarily in a distance d between the lower image plane (also called optical plane) 5u of the lower optics 6u and the upper image plane 5o of the upper optics 6o.

[0079] The alignment marks 9l, 9r are located on the photosensitive surface 14su at a distance dx in the X direction from each other and at a distance dy in the Y direction.

[0080] The distance d between the optical planes 5u and 5o is such that the upper optical plane 5o does not intersect the lower depth of field tu of the lower optical plane 5u, and vice versa. The optical plane 5u is the image plane of the lower optical subsystem 12u.

[0081] The alignment marks 9l, 9r projected onto the image plane 5u must be projected onto the photosensitive surface 14so of the upper camera chip 2o via the upper optical subsystem 12o. However, in order to sharply image the alignment marks 9l, 9r projected onto the image plane 5u of the lower optical system 12u onto the photosensitive surface 14so, the optical plane 5u must lie within the depth of field t0 of the upper optical plane 5o, which simultaneously represents the object plane for the upper optical subsystem 12o. If this is not the case, as in the Fig. As shown in Figure 2a, the projected alignment marks 9pl, 9pr are blurred on the photosensitive surface 14so (Figure BB). The process step of Fig. 2a therefore consists of bringing the two optical planes 5u and 5o closer together.

[0082] The Fig. Figure 2b shows a schematic, not to scale, simplified sketch of the optical system 11 in a second process step 102 according to the invention, consisting of the two optical subsystems 12u, 12o, in which the optical planes 5u and 5o have been largely brought close to each other, while the optical axes 13u and 13o of the optics 6u and 6o are shifted so far from each other that they do not intersect between the optics 6u, 6o.

[0083] This large deviation of the optical axes 13u and 13o means that only the alignment mark 9pl is within the field of view of the upper camera 1u (Figure BB). This mark is sharply imaged because the part of the optical plane 5u onto which the alignment mark 91 is intermediately projected lies within the depth of field of the upper optical plane 5o.

[0084] The displacement of the optical axes 13u and 13o relative to each other results in a displacement of the projected image with the alignment marks 9pl and 9pr on the upper photosensitive surface 14so of the upper camera chip 2o. This displacement can be so large that one of the alignment marks 9pl, 9pr is not visible because it is not projected onto the photosensitive surface 14so. In this case, it is the alignment mark 9pr. The process step of Fig. 2b therefore consists of bringing the two alignment marks 9pl, 9pr into the field of view of the upper camera 10.

[0085] The Fig. Figure 2c shows a schematic, not-to-scale, simplified sketch of the optical system 11 in a third process step 103 according to the invention, consisting of the two optical subsystems 12u, 12o, in which the two projected alignment marks 9pl, 9pr are located in the field of view of the upper camera 1o, but the two optical planes 5u and 5o are still strongly inclined to each other. Part of the lower optical plane 5u is located in the upper depth of field t0 of the upper optical plane 5o, while the other part is located outside of it.

[0086] By analogy to the description of the Fig. 2b therefore follows that the alignment mark 91 is sharply imaged as alignment mark 9pl, while the alignment mark 9r is blurred as alignment mark 9pr. By appropriately moving the lower optical subsystem 12u and / or the upper optical subsystem 12o, in particular the optics 6u, 6o, the optical planes 5u and 5o can be aligned relative to each other so that the projected alignment marks 9pl and 9pr are both sharply imaged.

[0087] From what has been said, it can be mathematically derived how to calculate the minimum wedge error that is just measurable and thus adjustable using the method according to the invention. For the sake of simplicity, it is assumed that the two optical planes 5u, 5o are tilted exclusively about a tilting axis 20 that lies parallel to the x-axis. Let to be the depth of field of the upper optical plane 5o and dx the distance between the alignment marks 9l, 9r. Then the following applies: tanα=todx

[0088] The minimum correctable wedge error is therefore α=arctantodx

[0089] It can be seen that the correctable wedge error is smaller the greater the distance dx between the alignment marks 91, 9r and the smaller the depth of field to of the upper optical plane 50. The process step of Fig. 2c therefore consists of minimizing the wedge error between the two optical planes 5u, 5o as much as possible.

[0090] The Fig. Figure 2d shows a schematic, not-to-scale, simplified sketch of the optical system 11 in a fourth process step 104 according to the invention, consisting of the two optical subsystems 12u and 12o, in an optimal state in which the optical plane 5u, which represents the image plane for the projected alignment marks 91 and 9r, is located in the depth of field 0 of the upper optical plane 5o, which simultaneously serves as the object plane for the upper optical subsystem 12o and thus for the projected alignment marks 9pl and 9pr. Preferably, the two planes 5u and 5o are parallel to each other. In reality, however, the two planes will always have a non-zero angle to each other. The angle between the two planes 5u and 5o is less than 5°, preferably less than 1°, more preferably less than 0.1°, most preferably less than 0.01°, and most preferably less than 0.001°.The expert knows that the angle between two planes is identical to the angle between the two plane normals.

[0091] Since the lower optical plane 5u is located in the depth of field range to the upper optical plane, the alignment marks 91 and 9r are imaged as sharp alignment marks 9pl and 9pr (see illustration BB).

[0092] In this process step, the alignment marks 9pl, 9pr are now fine-tuned in relation to the desired, ideal positions 9il, 9ir (see Fig. 3b). The process step of Fig. 2d therefore consists of performing the fine calibration of the alignment marks 9pl, 9pr, especially in the x and / or y direction.

[0093] The Fig. Figure 3a shows a schematic, not-to-scale, simplified, enlarged sketch of the photosensitive area 14su of the lower camera chip 2u. The corresponding alignment marks 91, 9r are visible. The position of the alignment marks 91, 9r is best specified in relation to a lower detector coordinate system (xdu, ydu). The origin of the lower detector coordinate system (xdu, ydu) is located, for example, in the upper left corner of the photosensitive surface 14su.

[0094] The Fig. Figure 3b shows a schematic, not to scale, simplified, enlarged sketch of the photosensitive area 14so of the upper camera chip 2o at the beginning of the fourth process step 104 according to the invention.

[0095] It is evident that the projected alignment marks 9pl, 9pr are not congruent with the ideal positions 9il, 9ir. The upper photosensitive surface 14so has an upper detector coordinate system (xdo, ydo). The origin of the upper detector coordinate system (xdo, ydo) lies in the upper left corner of the photosensitive surface 14so.

[0096] The coordinates of the ideal positions 9il, 9ir in the detector coordinate system (xdo, ydo) preferably correspond to the coordinates of the alignment marks 91, 9r in the lower detector coordinate system (xdu, ydu), provided that the photosensitive areas 14su, 14so have the same properties such as length, width, resolution in the x-direction, resolution in the y-direction, etc., or in short, that the same camera chips 2u, 2o are used. A further condition is, of course, the correct and consistent choice of the origin of the detector coordinate systems (xdu, ydu) and (xdo, ydo).

[0097] By further adjusting the optical elements in the optical systems 11, 11', 11" the alignment marks 9pl, 9pr can be calibrated with respect to the ideal positions 9il, 9ir. This may result in translational shifts of the alignment marks 9pl, 9pr and / or a rotation with respect to the detector coordinate system (xdo, ydo).

[0098] If perfect calibration is not fully possible and / or economically undesirable because it would take too long, the alignment marks 9pl, 9il and 9pr, 9ir will not perfectly coincide but will be translationally and / or rotationally displaced relative to each other. The alignment process would then end at this step. According to the invention, the system stores the coordinates of the alignment marks 9pl, 9il, 9pr, 9ir with respect to the detector coordinate system (xdo, ydo). If the optical system thus calibrated is used in a subsequent process, for example, to align the first alignment marks of a first substrate surface of a first substrate with the second alignment marks of a second substrate surface of a second substrate, then the measured and stored translational and / or rotational displacements must be used for correction.

[0099] The Fig. Figure 3c shows a schematic, not to scale, simplified, enlarged sketch of the photosensitive area 14so of the upper camera chip 2o at the end of the fourth process step 104 according to the invention. The alignment marks 9pl, 9pr are located at the ideal positions 9il, 9ir. The optical system 11, 11', 11" is thus calibrated. Although the stated and preferred goal of the process according to the invention is to align the projected alignment marks 9pl, 9pr with the ideal positions 9il, 9ir, the possibility is disclosed of not aligning the projected alignment marks 9pl, 9pr with the ideal positions 9il, 9ir and instead storing the rotational rotation and / or translational displacement between the projected alignment marks 9pl, 9pr and the ideal positions 9il, 9ir. This stored information allows for coordinate and / or image correction at any time.Performing such coordinate and / or image correction requires computing power. It is also conceivable that the alignment marks 9pl, 9pr cannot be perfectly aligned with the ideal positions 9il, 9ir, particularly due to tolerances in the optical and / or mechanical components of the optical system 11, 11', 11". In this case, the rotational rotation and / or translational displacement are preferably also recorded. Coordinate and / or image correction should preferably always be performed to obtain the most accurate results possible.

[0100] Fig. Figure 4a shows a schematic, not-to-scale, isometric exploded view of a disassembled camera chip 2u ​​according to a first embodiment of the invention. The camera chip 2u ​​consists of a chip carrier 14 having a photosensitive surface 14su.

[0101] The chip carrier 14 is usually sealed with a protective plate 15, in particular a glass plate. Breaking this seal requires enormous force and regularly results in the destruction of the chip carrier 14, especially the photosensitive surface 14su. Therefore, the alignment marks 91, 9r must preferably be produced by the chip manufacturer before sealing with the protective plate 15.

[0102] Additional glass plates 16, in particular optical filters, can be mounted and attached to the protective plate 15. Most often, all components 14, 15, 16 are fixed by means of a frame 17.

[0103] Fig.Figure 4b shows a schematic, not-to-scale, isometric exploded view of a disassembled camera chip 2u ​​according to another embodiment of the invention, in which a separate marking plate 18 according to the invention with alignment marks 91, 9r is fixed, in particular directly after the protective plate 15, i.e., as close as possible to the photosensitive surface 14su. The alignment marks 91, 9r are preferably located on the side of the marking plate 18 that faces the photosensitive surface 14su, in order to further reduce the distance between the alignment marks 91, 9r and the photosensitive surface 14su.

[0104] Inserting the marking plate 18 is relatively easy with most 2u camera chips, as all components above the protective plate 15 can be easily removed, even intentionally. This allows such 2u camera chips to be expanded, particularly by filters, and in this specific case by a marking plate 18. Reference symbol list 10, 1u camera 20, 2u camera chip 3, 30, 3u mirror 4.4 lenses 5o, 5u Optical plane, in particular object or image plane 6o, 6u Optics 7, 7', 7'' lighting path 8 Image path 9l, 9r, 9l', 9r' alignment mark 9pl,9pr Projected alignment mark 9il,9ir ideal alignment brands 10 Source 11, 11',11" Optical System 12u, 12u', 12u'' Optical subsystem 12o, 12o', 12o Optical subsystem 13u, 13o Optical axis 14 chip carriers 14so, 14su Photosensitive surface 15 Protective plate 16 Glass plate, especially filter 17 frames 18 Marking plate 19 surface 20 Tilting axle 21 Alignment Mark Projection System α Angle between optical planes AA Sectional View BB Sectional View d distance dx,dy distance depth of field range xdu,ydu coordinates

Claims

[1] Method for aligning two oppositely arranged optical subsystems (12u,12u',12u'',12o,12o',12o'') of an optical system (11,11',11"), comprising the following steps: - Projection of alignment marks (9l, 9r, 9l', 9r') into an image plane (5u) of the first optical subsystem (12u, 12u', 12u''), - Projection of the alignment marks (9l,9r,9l',9r') from the first image plane (5u) onto a sensitive surface (14so) of the second optical subsystem (12o, 12o', 12o''), - Alignment of the optical subsystems (12u,12u',12u'',12o,12o',12o'') to each other, so that projections (9pl,9pr) of the alignment marks (9l,9r,9l',9r') are imaged at ideal positions (9il,9ir) in a depth of field range of the sensitive surface (14so), characterized by, that the alignment marks (9l, 9r) are applied directly to a first sensitive surface (14su) of a first camera chip (2u), wherein the alignment marks (9l, 9r) have a cluster of non-functional pixels. [2] Method according to claim 1, wherein light (7) for illuminating the alignment marks (9l, 9r) is coupled into the first and / or second optical subsystem (12u,12u',12u'',12o,12o',12o'') via at least one light source (10), in particular a lamp, preferably a halogen lamp, and at least one mirror (3). [3] Method according to one of the preceding claims, wherein ambient light (7') is coupled in at least one place of the optical subsystems (12u,12u',12u'',12o,12o',12o''), in particular at a first camera (1u) and / or at a mirror (3,3u,3o) to illuminate the alignment marks (9l, 9r), wherein preferably a shielding of the optical subsystems (12u,12u',12u'',12o,12o',12o'') is not installed. [4] Method according to any of the preceding claims, wherein the optical subsystems (12u,12u',12u'',12o,12o',12o'') comprise two optics arranged opposite each other (6o,6u) whose image planes (5o, 5u) and / or optical axes (13o, 13u) are aligned to each other. [5] Method according to claim 4, wherein at least the optics (6o, 6u) are part of a system for aligning substrates, in particular wafers. [6] Method according to claim 4 or 5, wherein the optics (6o,6u) are moved translationally and / or rotationally relative to each other for alignment. [7] Method according to claim 4, 5 or 6, wherein the optics (6o, 6u) are brought close together for alignment such that the first image plane (5u) of the first optic (6u) is in a depth of field area (to) of the second object plane (5o) of the second optic (6o). [8] Method according to any of the preceding claims, wherein the assembly of alignment markers has a cross shape. [9] System for aligning two oppositely arranged optical subsystems (12u,12u',12u'',12o,12o',12o'') of the optical system (11,11',11'') using a method according to one of the preceding claims, comprising: - A first camera (1u) with a first camera chip (2u) for projecting alignment marks (9l,9r,9l',9r') into an image plane (5u) of the first optical subsystem (12u, 12u', 12u''), - Means for projecting the alignment marks (9l,9r,9l',9r') from the first image plane (5u) onto a sensitive surface (14so) of the second optical subsystem (12o,12o',12o''), - Means for aligning the optical subsystems (12u,12u',12u'',12o,12o',12o'') to each other, so that projections (9pl,9pr) of the alignment marks (9l,9r,9l',9r') can be imaged in a depth-of-field area of ​​the sensitive surface (14so) at ideal positions (9il,9ir), characterized by , that the alignment marks (9l, 9r) are applied directly to a first sensitive surface (14su) of the first camera chip (2u), wherein the alignment marks (9l, 9r) have a cluster of non-functional pixels. [10] Camera chip (2u) with alignment marks (9l, 9r) on its sensitive surface (14su) for a system and / or method according to one of the preceding claims, characterized by , that the alignment marks (9l, 9r) are applied directly to the sensitive surface (14su) of the camera chip (2u), wherein the alignment marks (9l, 9r) have a cluster of non-functional pixels.

Images