METHOD FOR ALIGNING TWO SUBSTRATES
Patent Information
- Application Number
- DE502017017150
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2017-03-20
- Publication Date
- 2025-12-11
- Estimated Expiration
- 2037-03-20
AI Technical Summary
Existing alignment methods for substrates are limited by imperfect alignment marks, which are distorted, have rough surfaces, and minimal deviations, leading to inaccuracies in positioning and bonding processes.
A method involving the detection and storage of multiple alignment marks on each substrate surface, followed by a statistical averaging process to improve alignment accuracy, using interferometers and optics to minimize errors and correct for curvature, with alignment marks positioned to optimize the field of view and reduce movement of optical systems.
Enhances alignment precision, allowing substrates to be bonded with an accuracy of 50-100 nm, even with imperfect alignment marks, by averaging out quality deficiencies and minimizing optical movement.
Description
[0001] Substrates are oriented relative to each other using alignment systems, according to the state of the art. Alignment marks are located on the substrates and are aligned relative to each other by the alignment systems. Numerous methods exist in the state of the art for measuring the alignment marks, for correctly positioning the substrates on which the alignment marks are located, and for the subsequent bonding step.
[0002] Publication WO 2010 / 023935 A1 discloses a device for aligning substrates, wherein a holding surface is displaced in the surface direction of one substrate, while one is held by a pair of opposing substrates.
[0003] The biggest problem with measuring alignment marks is that the marks themselves are never perfect. They are distorted, have a rough surface, poorly manufactured edges, etc. Very often, contrast or optical effects are the main reason why the optics and cameras are unable to obtain an optimal image of the alignment mark. Since the alignment marks are detected and measured by software, the accuracy of the position determination depends heavily on their quality. However, the quality of the alignment mark cannot be improved indefinitely, so one must always assume that imperfect alignment marks are being measured.
[0004] Another problem with the prior art is that there can be minimal deviations from the target position of the alignment marks. While the positions of the alignment marks are defined with the greatest possible accuracy in the computer, the production of the alignment marks is limited by the accuracy of the processes and equipment used to create the alignment marks on the substrate surface. Even if the alignment marks were to exhibit no deviation from their ideal geometry, their actual position can deviate minimally from the ideal position. This deviation is approximately between 1 and 100 nm, i.e., in the nanometer range.
[0005] It is therefore the object of the present invention to provide a method by which the accuracy of the alignment of two substrates relative to each other can be improved.
[0006] This problem is solved by the features of the dependent claims. Advantageous embodiments of the invention are specified in the sub-claims. For value ranges, values lying within the stated limits should also be disclosed as limit values and be claimable in any combination.
[0007] A method for aligning two substrates to be bonded is proposed, wherein, among other things, the first substrate has a first substrate surface with a first substrate surface side (hereinafter also referred to as the left substrate surface side) and a second substrate surface side opposite the first substrate surface side (hereinafter also referred to as the right substrate surface side), wherein the second substrate has a second substrate surface to be bonded to the first substrate surface, wherein the second substrate surface has a third substrate surface side (hereinafter also referred to as the left substrate surface side) and a fourth substrate surface side opposite the third substrate surface side (hereinafter also referred to as the right substrate surface side), wherein the method comprises at least the following steps: -- Detection and storage of first positions of a first pair of alignment markers on the first substrate surface of the first substrate, wherein a first alignment marker of the first pair of alignment markers is located on the first substrate surface side and a second alignment marker of the first pair of alignment markers is located on the second substrate surface side, -- Detection and storage of second positions of a second pair of alignment markers on the second substrate surface of the second substrate, wherein a third alignment marker of the second pair of alignment markers is located on the third substrate surface side and a fourth alignment marker of the second pair of alignment markers is located on the fourth substrate surface side, -- Detection and storage of third positions of a third pair of alignment markers on the second substrate surface of the second substrate,wherein a fifth alignment mark of the third alignment mark pair is arranged on the third substrate surface side and a sixth alignment mark of the third alignment mark pair is arranged on the fourth substrate surface side, -- detection and storage of fourth positions of a fourth alignment mark pair on the first substrate, wherein a seventh alignment mark of the fourth alignment mark pair is arranged on the first substrate surface side and an eighth alignment mark of the fourth alignment mark pair is arranged on the second substrate surface side, -- alignment of the two substrates relative to each other depending on the detected first positions, second positions, third positions and fourth positions.
[0008] The procedure preferably comprises the following steps, in particular the following sequence: -- Arranging and fixing the substrates on substrate holders, -- Moving the first substrate holder with the first substrate into a first detection position, -- Detection and storage of the first positions of the first alignment mark pair on the first substrate surface of the first substrate in the first detection position, wherein the first alignment mark of the first alignment mark pair is located on the first substrate surface side and the second alignment mark of the first alignment mark pair is located on the second substrate surface side, -- Moving the second substrate holder with the second substrate into a second detection position and, in particular, simultaneously moving the first substrate holder with the first substrate into a first waiting position, -- Detection and storage of the second positions of the second alignment mark pair on the second substrate surface of the second substrate in the second detection position,wherein the third alignment mark of the second pair of alignment marks is located on the third substrate surface side and the fourth alignment mark of the second pair of alignment marks is located on the fourth substrate surface side, -- detecting and storing the third positions of the third pair of alignment marks on the second substrate surface of the second substrate in the second detection position of the second substrate holder, wherein the fifth alignment mark of the third pair of alignment marks is located on the third substrate surface side and the sixth alignment mark of the third pair of alignment marks is located on the fourth substrate surface side, -- moving the first substrate holder with the first substrate into the first detection position and, in particular, simultaneously moving the second substrate holder with the second substrate into a second waiting position,-- Detection and storage of the fourth positions of the fourth alignment mark pair on the first substrate in the first detection position, wherein the seventh alignment mark of the fourth alignment mark pair is located on the first substrate surface side and the eighth alignment mark of the fourth alignment mark pair is located on the second substrate surface side, -- Alignment of the two substrates relative to each other depending on the detected first positions, second positions, third positions and fourth positions.
[0009] In a preferred embodiment, the positions of at least three alignment marks are detected on each substrate surface side, preferably at least four, more preferably at least five, and most preferably at least six. This advantageously improves the accuracy of the alignment.
[0010] In a preferred embodiment, the alignment marks of a substrate surface are arranged within a circle K with a radius of less than 100 mm, preferably less than 50 mm, more preferably less than 25 mm, most preferably less than 10 mm, and most preferably less than 1 mm. This minimizes the travel distance of the optics for detecting the alignment marks, so that only minimal errors arise from the movement of the optics. Preferably, the optics do not need to be moved for detection.
[0011] According to the invention, several, in particular even all, alignment marks on the same substrate surface side are located in the field-of-view of an optic.
[0012] In a particularly preferred embodiment of the invention, several, in particular all, alignment marks on the same substrate surface side are located within the field of view of an optic and are centered relative to each other. For this to be possible, it is necessary that the alignment marks become progressively larger from the center outwards and do not overlap.
[0013] In the embodiments according to the invention, in which several alignment marks are located within the field of view of the optics, it is of particular interest that the alignment marks are located as far as possible from the edge of the substrate in order to increase the alignment accuracy. According to the invention, the alignment marks are then located less than 30 mm, preferably less than 20 mm, more preferably less than 10 mm, most preferably less than 5 mm, and most preferably less than 11 mm from the edge of the substrate.
[0014] According to the invention, the positions of the alignment marks are optically detected. This advantageously enables a comparatively precise determination of the position.
[0015] In a preferred embodiment, the positions of the substrate holders are detected by interferometers and / or optics, which are independent of detection devices for determining the positions of the alignment marks. This advantageously improves the position determination of the substrate holders. The positions of the optics can also be determined by detection devices, particularly interferometers.
[0016] In a preferred embodiment, the substrate holders are fixed in the detection positions. This allows for better determination of the substrate holder positions. A device for aligning two substrates to be bonded is also disclosed, comprising a method comprising: -- Substrate holders for receiving and fixing the substrates, -- movement devices for moving the substrate holders, -- optics for detecting positions of alignment marks on the substrates, -- a storage device for storing the positions of the alignment marks.
[0017] The characteristics of the process also apply analogously to the device.
[0018] The present disclosure also relates to a product with two substrates, having at least two alignment marks on each side, wherein this product is not part of the claimed subject matter.
[0019] The core of the invention consists in particular in demonstrating a statistical method by which two substrates can be better aligned and bonded with respect to their alignment marks by using the determination of several alignment marks located at different positions, especially at different radial distances from the center of the substrate. The problem of substrate alignment is improved in particular by using statistically averaged positions, calculated from several alignment marks, rather than the alignment marks themselves, to align the substrates relative to each other.
[0020] The invention advantageously allows the alignment of two substrates with increased precision, even when the alignment marks are not ideally created with respect to their shape, position, and / or orientation. The achievable alignment accuracy depends primarily on the number of alignment marks used in the alignment process. By using multiple alignment marks, especially those of lower quality, quality deficiencies are averaged out from the final result. For an alignment process that relies on averaging multiple alignment marks, it is particularly advantageous if all marks on one side are simultaneously within the field of view of an optical system.
[0021] If the process according to the invention is intended to correct errors resulting from local and / or global curvature of the substrate, then concentrating the production of alignment marks in an area covered by the optical field of view is rather detrimental. The alignment accuracy would then increase with the inventive measurement of multiple alignment marks that are further apart.
[0022] According to the invention, a suitable solution is preferred depending on the specific problem. If it is known that the alignment marks are manufactured with inferior quality, they are preferably produced close together. If higher alignment accuracy is to be achieved for a highly distorted and / or curved substrate using the alignment marks, the alignment marks are further apart on each side of the substrate. The aforementioned possibility of statistical error averaging can therefore be applied to both methods.
[0023] In a particularly preferred embodiment of the invention, it is conceivable that there are several areas, each with several alignment marks, each of which can be completely covered by the viewing area of the optics. In this way, according to the invention, all alignment marks of one area are used to compensate for quality deficiencies in the alignment marks, while the alignment marks of different areas can be used to improve the alignment accuracy with respect to curvatures. Substrates and alignment marks
[0024] The substrates have a primary substrate surface and a secondary substrate surface opposite the primary substrate surface. The substrate surface that forms part of the bond interface is called the inner substrate surface, and the substrate surface opposite the inner substrate surface is called the outer substrate surface.
[0025] In the remainder of this document, only the inner substrate surfaces are relevant. Unless otherwise stated, all substrate surfaces mentioned in this document are inner substrate surfaces.
[0026] A substrate surface has a left substrate surface side and a right substrate surface side. The left substrate surface side comprises the entire area of the substrate surface to the left of a plane that is normal to the substrate surface and preferably passes through the notch and the substrate center, or preferably is normal to the flat side of the substrate and passes through the substrate center. The right substrate surface side comprises the entire area of the substrate surface to the right of the plane that is preferably normal to the substrate surface and passes through the notch, or preferably is normal to the flat side of the substrate.
[0027] For the sake of completeness, it should be mentioned that the plane does not necessarily have to pass through the notch or be perpendicular to the flat surface. However, considerations of symmetry usually lead to such substrates, and in particular the functional units on the substrates, being symmetrical with respect to a plane that passes through the notch or is perpendicular to the flat surface. Corresponding to the symmetrical distribution of the functional units on the substrate surface, the alignment marks are then also symmetrically distributed.
[0028] The substrates to be aligned according to the invention have at least two alignment marks per substrate surface side. More than two alignment marks per substrate surface side are also conceivable, in particular more than three alignment marks, more preferably more than four alignment marks, most preferably more than five alignment marks, and most preferably more than six alignment marks. The alignment marks on one substrate surface side are preferably located within a circle K with a radius of less than 100 mm, more preferably less than 50 mm, more preferably less than 25 mm, most preferably less than 10 mm, and most preferably less than 1 mm.
[0029] In a particularly preferred embodiment, the alignment marks of one substrate surface side are located on the same radius of the substrate. In a particularly preferred embodiment according to the invention, all alignment marks of both substrate surface sides are located on the same diameter of the substrate.
[0030] The alignment marks are further classified according to their position relative to the center of the substrate. The alignment marks closest to the center of the substrate are called inner alignment marks. The alignment marks furthest from the center are called outer alignment marks. To keep the description of the process according to the invention as simple as possible, we will henceforth only refer to two alignment marks per substrate surface. Extending the embodiment according to the invention to more than two (n) alignment marks, or extending it to a repeated measurement of the same alignment marks per substrate surface, will then be achieved by analogy.An extension of the embodiment according to the invention to a repeated measurement of the same alignment marks per substrate surface side is also graphically revealed by the loop in the specified flowchart. Calibration process
[0031] Prior to the actual process according to the invention, the optics of the alignment device are calibrated. Corresponding systems for aligning substrates are disclosed in publications US6214692B1, PCT / EP2013 / 075831, WO2011042093A1 and WO2014202106A1 and are therefore not described in detail here.
[0032] The method according to the invention operates with at least four optics, which are preferably grouped into two optic groups. Each optic within an optic group can be moved individually, in particular translationally and / or rotated, but is fixed relative to the opposite optic after calibration. Therefore, after fixation, both optics move simultaneously due to a translational and / or rotational movement of the optic group. This rigid fixation of the optics relative to each other is preferably not released after the respective calibration process, which is explained in more detail in the following section.
[0033] The calibration of the optics within an optical group aims to define the intersection point of the optical axes. The depth of field range also includes the area of the subsequent focal planes.
[0034] The optics are calibrated using alignment marks. The alignment mark is preferably located in a calibration substrate, particularly a transparent one. However, alignment marks located on the surface of a calibration substrate can also be used.
[0035] In a first calibration step, the first optic of the first optic group of a first side is focused on a first alignment mark. The optic is preferably calibrated so that the first alignment mark is located in the center of the depth of field.
[0036] In a second calibration step, the second optic of the first optical group on the first side is focused on the same first alignment mark. The optic is preferably calibrated such that the first alignment mark is located in the center of the depth of field. In particular, the optic moves in the x and / or y and / or z direction or rotates about the x and / or y and / or z axis. After the calibration process, the alignment mark is preferably located exactly in the center of the field of view.
[0037] Generally, the optical axes of the first two optics are not perpendicular to the focusing plane. Therefore, the optical axes preferably intersect at the first alignment mark. After this calibration step, the two optics of the first optical group are no longer moved relative to each other.
[0038] In a third calibration step, the first optic of the second optic group on the second side is focused onto a second alignment mark. The optic is preferably calibrated so that the alignment mark is located in the center of the depth of field.
[0039] In a fourth calibration step, the second optic of the second optic group on the second side is focused on the same, second alignment mark. The optic is preferably calibrated so that the alignment mark is located in the center of the depth of field.
[0040] Generally, the optical axes of the two second optical elements are not perpendicular to the focusing plane. Therefore, the optical axes preferably intersect at the second alignment mark. After this calibration step, the two optical elements of the second optical group are no longer moved relative to each other.
[0041] It is also conceivable to use a calibration substrate that has only one alignment mark, which then needs to be moved between the calibration steps of the first side and the calibration steps of the second side. During this translational movement of the calibration substrate, the alignment mark must maintain a z-position that is as constant as possible, particularly assuming a flat substrate. In general, the z-distances to the substrate surface must be constant at all substrate positions. If the substrate is curved, for example, a translational movement of the substrate relative to the optics must preferably be followed by a correction of the optics and / or the substrate in the z-direction. Such corrections can be omitted for negligible curvatures. After calibration, the focal plane is within the depth-of-field ranges of all four optics.The focusing plane and the subsequent bonding plane preferably coincide. After the optics have been calibrated, the actual process according to the invention can begin. The calibration process is described in WO2014202106A1.
[0042] The resolution of the optics is between 10000 nm and 10 nm, preferably between 5000 nm and 50 nm, more preferably between 1000 nm and 100 nm, most preferably between 750 nm and 380 nm.
[0043] The resolution is preferably defined according to the Rayleigh criterion.
[0044] For the sake of completeness, it is disclosed that the method according to the invention also works with alignment systems in which the optics on one side are not mechanically coupled to each other in an optical group. Such an alignment system is mentioned, for example, in publication WO2011042093A1. In such an alignment system, an upper and a lower optic do not necessarily have to be positioned opposite each other, but can also be spaced apart from each other in the xy plane. However, such an alignment system can also be used to carry out the process according to the invention. Surveying process
[0045] Several systems with different alignment concepts exist in the prior art. The systems most relevant to this invention are disclosed in publications US6214692B1 and WO2014202106A1.
[0046] Document US6214692B1 describes a system in which the substrates are alternately pushed between the optics located in front of or behind the substrate stack.
[0047] In contrast, publication WO2014202106A1 describes a system in which the substrates are alternately pushed between optics located laterally to the substrate stack. The improved alignment process disclosed according to the invention can be applied to all the aforementioned alignment systems, but especially also to WO2014202106A1. However, the improved alignment process disclosed according to the invention is illustrated by way of example using the disclosed alignment system in patent US6214692B1.
[0048] The method according to the invention can in principle be applied to all known alignment methods and carried out by almost all alignment systems.
[0049] The substrates to be aligned are fixed to sample holders. The sample holders have fixings. The fixings serve to hold the substrates in place. The fixings can be... Mechanical fixations, in particular clamps; vacuum fixations, in particular with individually controllable vacuum tracks or interconnected vacuum tracks or as part of a pin chuck (see, for example, WO2015113641A1); electrical fixations, in particular electrostatic fixations; magnetic fixations; adhesive fixations, in particular gel-pak fixations or fixations with adhesive, in particular controllable, surfaces The fixations are particularly electronically controllable. Vacuum fixation is the preferred fixation method. The sample holder preferably consists of several vacuum channels that emerge from its surface. The vacuum channels are preferably individually controllable. In a more technically feasible application, several vacuum channels are combined into vacuum channel segments that can be individually controlled and therefore evacuated or flooded. Each vacuum segment is independent of the other vacuum segments. This allows for the construction of individually controllable vacuum segments. The vacuum segments are preferably ring-shaped. This enables targeted, radially symmetrical fixation and / or release of a substrate from the sample holder, particularly from the inside out. In a highly preferred embodiment, the sample holder is designed as a pin chuck.This is a sample holder in which numerous, particularly symmetrically arranged, protrusions, especially needles, are formed from the surface, the surrounding area of which can be evacuated. The substrate rests only on the heads of the protrusions. The contact area is correspondingly small. An embodiment of a needle sample holder is described in the publication . WO2015113641A1 revealed.
[0050] In the further course of the embodiment according to the invention, a displacement of the optics from the inner to the outer alignment marks is described. However, the process according to the invention also works if the alignment marks are measured from the outside in, or even in any sequence. It is conceivable, although not preferred, to measure the alignment marks from the inside out on the first substrate surface side and from the outside in on the second substrate surface side.
[0051] The embodiment according to the invention consists in the fact that several alignment marks are located in the field of view of an optic on each substrate surface side, so that a displacement of the optics outwards or inwards is not necessary at all. According to the invention, this even makes it possible to measure several alignment marks without having to move the optic groups.
[0052] In a far more preferred embodiment of the invention, not only are several alignment marks located on each substrate surface side within the optical field of view, but they are even nested within one another without touching each other. This allows for a particularly efficient placement of several alignment marks per unit area. As a result, it becomes particularly efficient to measure multiple alignment marks according to the invention without having to move the optical groups.
[0053] In a first process step according to the invention, the lower, inner alignment marks of the lower substrate are measured. For this purpose, the upper substrate is moved out of the field of view of the upper optics, if necessary, so that the upper optics have a clear view of the lower, inner alignment marks of the lower substrate. Preferably, the lower, inner alignment marks are already within the field of view of the upper optics, so that the optic groups do not need to be moved. Moving the individual optics is no longer required at this point. From now on, only the optic groups are moved. After the lower, inner alignment marks have been measured, there is a unique correlation between the lower sample holder position and the lower, inner alignment marks.
[0054] In a second process step according to the invention, the upper sample holder with the upper substrate moves over the lower substrate. Simultaneously, the lower sample holder with the lower substrate is moved back. Preferably, a position error correction (PEC) is performed during this process. This refers to the measurement of the position of the lower sample holder by measuring devices, in particular interferometers and / or other special optics, which are hereinafter referred to as PEC optics. Position error correction is necessary because the lower substrate cannot be observed by the upper optics during its movement out of their field of view, as the upper substrate has generally already moved into their field of view.To ensure that the lower substrate can be returned to its original position with high precision in subsequent process steps, its position must be measured exactly during movement from the field of view of the optics. Preferably, interferometers are used for the continuous measurement of the lower sample holder's position. Even more preferably, additional optics, PEC optics, are used that are completely independent of the aforementioned optics for detecting the alignment marks. These PEC optics measure markings on the lower sample holder and can thus indicate the sample holder's position, particularly at any given time. Preferably, the position of the lower substrate holder is measured at least before the translational movement. This ensures that the position can be restored at any time.The measured sample holder positions allow the lower sample holder, and thus the lower substrate, to be returned to its original position at any time.
[0055] The accuracy with which the position detection of the sample holders for position error correction is performed is better than 1 µm, preferably better than 500 nm, even more preferably better than 100 nm, most preferably better than 50 nm, and most preferably better than 10 nm.
[0056] In a third process step according to the invention, the upper, inner alignment marks of the upper substrate are measured. Preferably, the upper, inner alignment marks are already within the field of view of the lower optics, so that the optic groups do not need to be moved. Moving the individual optics is no longer permitted at this point. After the upper, inner alignment marks have been measured, there is a unique correlation between the upper sample holder position and the upper, inner alignment marks. In a further step, the upper sample holder is fixed in place.
[0057] In a fourth process step, the optical groups are moved so that the upper, outer alignment marks are within the field of view of the lower optics. After the upper, outer alignment marks have been measured, a unique correlation exists between the upper sample holder position and the upper, outer alignment marks. To maintain this correlation with the previously recorded upper, outer alignment marks, the position of the lower optics is preferably monitored precisely, ideally using optical measuring instruments, and even more preferably by interferometry. This allows the path traveled by the optics to be recorded. In a fifth process step, the lower substrate is positioned. The positions measured in process step two, which were used for position correction, are then used to correctly reposition the lower substrate in its original position.
[0058] In particular, the upper substrate moves out of the field of view of the upper optics at the same time.
[0059] In a sixth process step, the optical groups are moved so that the lower, outer alignment marks are within the field of view of the upper optics. After the lower, outer alignment marks have been measured, a unique correlation exists between the lower sample holder position and the lower, outer alignment marks. To maintain this correlation with the previously recorded lower, outer alignment marks, the position of the upper optics is preferably monitored precisely, ideally using optical measuring instruments, and even more preferably by interferometry. This allows the path traveled by the optics to be recorded.
[0060] According to the invention, the positions of the upper and lower, outer and inner alignment marks exist at this point. The process according to the invention can now be carried out on further alignment marks. In this case, the number of data available for statistics and error correction increases. The number of alignment marks per substrate surface side can generally be arbitrarily large. The number of alignment marks per substrate surface side is greater than 1, preferably greater than 2, even more preferably greater than 5, most preferably greater than 10, and most preferably greater than 20.
[0061] It is also conceivable to repeat the measurements at the alignment marks. By continuously switching between inner and outer alignment marks, a statistical evaluation of the alignment positions is possible, thus enabling the determination of an even more optimal, especially imaginary, alignment mark.
[0062] According to the invention, the optical groups do not need to be moved after calibration, since at least two alignment marks per substrate surface side are located within the field of view of each optical element. Therefore, only steps one to three of the previous process need to be performed to measure at least two alignment marks per substrate surface side. Calculation process
[0063] In short, the calculation process is based on error minimization between the corresponding alignment marks of the top and bottom surfaces. Preferably, the least squares method is used. This error minimization results in an ideal XY position and / or rotational orientation of the lower sample holder, as well as an ideal XY position and / or rotational orientation of the upper sample holder, where the substrates are optimally aligned and can be contacted, particularly bonded, in a subsequent process step. Contacting process / Bonding process
[0064] After calculating the statistically averaged positions, the two substrates are bonded together. The substrates are aligned with respect to these statistically averaged positions. They are then moved into the designated positions. Following this, the two substrates are brought close together in the z-direction. In particular, a fusion bond, or more preferably a hybrid bond, is performed. In a fusion bond, the upper substrate, in particular, is curved by a pin into a convex shape as viewed from the bonding interface. The curved substrate then contacts the second, preferably flat, substrate, preferably centrally.
[0065] The alignment and bonding process can be carried out with a reproducibility between 150-250 nm, preferably between 125-225 nm, more preferably between 100-200 nm, most preferably between 75-150 nm, and most preferably between 50-100 nm.
[0066] Further advantages, features, and details of the invention will become apparent from the following description of preferred embodiments and from the drawings. These show: Figure 1a a schematic, not to scale, top view of an ideal alignment mark in the field of view, Figure 1b a schematic, not to scale, top view of several ideal alignment marks in the field of view, Figure 1c a schematic, not to scale, top view of several ideal, centered alignment marks in the field of view, Figure 2 a schematic, not to scale, top view of a non-ideal alignment mark, Figure 3 a schematic, not to scale, top view of a superposition of an ideal alignment mark with a non-ideal alignment mark, Figure 4 a schematic, not to scale, top view of a first substrate type with a notch, Figure 5 a schematic, not to scale, top view of a second substrate type with a flat side, process flow, Figure 6 a representation of a Figure 7a a schematic, not to scale, top view of a first process step according to the invention, Figure 7b a schematic,Figure 8a: a schematic, non-scale side view of a first process step according to the invention; Figure 8a: a schematic, non-scale bottom view of a second process step according to the invention; Figure 8b: a schematic, non-scale side view of a second process step according to the invention; Figure 9a: a schematic, non-scale bottom view of a third process step according to the invention; Figure 9b: a schematic, non-scale side view of a third process step according to the invention; Figure 10a: a schematic, non-scale bottom view of a fourth process step; Figure 10b: a schematic, non-scale side view of a fourth process step; Figure 11a: a schematic, non-scale top view of a fifth process step; Figure 11b: a schematic, non-scale side view of a fifth process step; Figure 12a: a schematic,Figure 12 shows a top view of a sixth process step (not to scale), Figure 13 a schematic side view (not to scale) of a sixth process step, and Figure 14 a graphical representation of the reduction of the alignment error as a function of the number of measured alignment marks.
[0067] In the figures, identical components or components with the same function are marked with the same reference symbols.
[0068] The following nomenclature applies in the subsequent figure descriptions. Elements on the left side of the substrate are indexed with a lowercase 1 (germ: left). Elements on the right side of the substrate are indexed with a lowercase r (germ: right). Elements on the top side are indexed with a lowercase o (germ: top). Elements on the bottom side are indexed with a lowercase u (germ: bottom).
[0069] The Figure 1aFigure 1 shows a schematic, not-to-scale representation of a single, ideal alignment mark 6i in a viewing area 10 of an optical system. The alignment mark 6i has straight edges 8i, is specifically telecentrically imaged, has the best possible contrast with its surroundings, is symmetrical, and, in particular, is not distorted. An alignment mark 6i of such high quality is difficult to produce. Software would calculate an ideal, centric position 7i from the ideal alignment mark 6i.
[0070] The Figure 1bFigure 1 shows a schematic, not-to-scale representation of two ideal alignment marks 6i, 6i' in a viewing area 10. The alignment marks 6i and 6i' are small enough and close enough to each other that they are simultaneously located in the viewing area 10 of a viewing optic 11ol, 11or, 11ul, 11ur. According to the invention, this makes it possible to avoid a process of the optic group between the two alignment marks 6i and 6i'. At the same time, both alignment marks 6i and 6i' can be measured. The alignment marks 6i and 6i' have straight edges 8i and 8i', are preferably telecentrically recorded, have the best possible contrast with their surroundings, are symmetrical, and, in particular, are not distorted. Alignment marks 6i and 6i' of such high quality are difficult to produce. Software would calculate two ideal, centric positions 7i and 7i' from the ideal alignment marks 6i and 6i'.
[0071] The Figure 1cFigure 1 shows a schematic, not-to-scale representation of two ideal alignment marks 6i, 6i' in a viewing area 10. The alignment marks 6i and 6i' are designed to be nested within each other but not touching. Both alignment marks 6i and 6i' are simultaneously located in the viewing area 10 of a viewing optic 11ol, 11or, 11ul, 11ur. According to the invention, this eliminates the need for the optics to move between the two alignment marks 6i and 6i'. At the same time, both alignment marks 6i and 6i' can be measured. The alignment marks 6i and 6i' have straight edges 8i and 8i', are telecentrically recorded, have the best possible contrast with their surroundings, are symmetrical, and, in particular, are undistorted. Alignment marks 6i and 6i' of such high quality are difficult to manufacture.Software would calculate two ideal, centric positions 7i and 7i' from the ideal alignment marks 6i and 6i'.
[0072] The representation of analogous figures 2b, 2c, 3b and 3c for the non-ideal alignment marks 6 is omitted hereafter. Those skilled in the art will understand that the considerations of the ideal alignment marks 6i, 6i' can be transferred to the non-ideal alignment marks 6.
[0073] The Figure 2Figure 6 shows a schematic, not-to-scale representation of a non-ideal alignment mark 6. The deviations from the desired, ideal shape are exaggerated to facilitate illustration. In general, a non-ideal alignment mark 6 has non-ideal, odd edges 8. Furthermore, the non-ideal alignment mark 6 may be distorted or not telecentrically captured, which can lead to a blurred contrast of the non-ideal, odd edges 8. Software would calculate a non-ideal, centric position 7 from the non-ideal alignment mark 6. Analogously to Figure 1b Two small and closely spaced alignment marks 6, 6' within a field of view 10 are also conceivable. Analogous to Figure 1c Two mutually complementary alignment marks 6, 6' within a field of view 10 are also conceivable.
[0074] The Figure 3Figure 1 shows a schematic, not-to-scale representation of the superposition of an ideal alignment mark 6i and a non-ideal alignment mark 6. At the center of each alignment mark 6i, the centric positions 7, 7i of the alignment marks 6i, calculated by software, are shown. The non-ideal centric position 7 deviates from the desired, ideal centric position 7i. This error is due to the non-ideal features of the non-ideal alignment mark 6, in particular the non-ideal, odd edges 8, and creates an error that is to be largely compensated for by the process according to the invention. The analogy considerations to the [reference to the invention] also apply here. Figures 1b and 1c .
[0075] The Figure 4Figure 1 shows a first substrate type 1, consisting of a substrate 2 with several, in particular symmetrically manufactured, functional units 3. The first substrate type 1 has a notch 4. Due to the symmetrical arrangement of the functional units 3, a symmetrical arrangement of the alignment marks 6l, 6l', 6r, and 6r' is obvious, but not essential. Preferably, the substrate type 1 is therefore divided by a plane, in particular a plane of symmetry, M into a left region L and a right region R. According to the invention, at least two alignment marks 6l and 6l' are located in the left region L and at least two alignment marks 6r and 6r' are located in the right region. The alignment marks 6l, 6l', 6r, and 6r' are located at distances rl, rl', rr, and rr' from the center Z of the substrate 2. Preferably, the alignment marks 6l, 6l', 6r, and 6r' are located on a diameter line D.
[0076] The Figure 5Figure 1 shows a second substrate type 1', consisting of a substrate 2' with several functional units 3, in particular those manufactured symmetrically. The second substrate type 1' has a flat side 5. Due to the symmetrical arrangement of the functional units 3, a symmetrical arrangement of the alignment marks 6l, 6l', 6r, and 6r' is obvious, but not essential. Preferably, the substrate type 1 is therefore divided by a plane, in particular a plane of symmetry, M into a left region L and a right region R. According to the invention, at least two alignment marks 61 and 61' are located in the left region L and at least two alignment marks 6r and 6r' are located in the right region. The alignment marks 6l, 6l', 6r, and 6r' are located at distances rl, rl', rr, and rr' from the center Z of the substrate 2'. Preferably, the alignment marks 6l, 6l', 6r and 6r' are located on a diameter line D.
[0077] The Figure 6Figure 1 shows a schematic process flow. In process step 100, the lower, inner alignment marks 6ul, 6ur of the lower substrate 2u are measured. In a further second process step 101 according to the invention, the position error of the lower sample holder 12u is corrected. In a further third process step 102 according to the invention, the upper, inner alignment marks 6ol, 6or of the upper substrate 2o are measured. In a further fourth process step 103, the upper, outer alignment marks 6ol', 6or' of the upper substrate 2o are measured. In a further fifth process step 104, the sample holders 12u, 12o are changed. In a further sixth process step 105, the lower, outer alignment marks 6ul', 6ur' are measured.In a seventh process step 106, a decision is made as to whether further alignment marks are measured according to the same process sequence, or whether the measured alignment marks are measured again. In a final, eighth process step 107 according to the invention, the alignment and bonding of the two substrates 2u, 2o then takes place.
[0078] The Figure 7a or the Figure 7bFigure 1 shows a schematic top view and side view of a first process step 100 according to the invention, in which a first, lower substrate 2u with a substrate surface 2us (s = surface) is fixed on a first, lower sample holder 12u. The second, upper sample holder 12o, on which the second, upper substrate 2o is fixed, is positioned in a waiting position in the first process step 100 according to the invention at a distance from the lower, inner alignment marks 6ul and 6ur such that the upper left optic 10ol and the upper right optic 10or (hidden in the side view) have unobstructed fields of view 10ul and 10ur of the two lower, inner alignment marks 6ul and 6ur. In this alignment step according to the invention, the positions of the lower, inner alignment marks 6ul and 6ur, as well as the position of the sample holder 12u in the measurement position for the lower substrate 2u, are first stored.
[0079] The Figure 8aor the Figure 8b Figure 1 shows a schematic bottom view or side view of a second process step 101 according to the invention, in which the second, upper sample holder 12o is moved into the measurement position for the upper alignment marks 6ol, 6or. Simultaneously, the lower sample holder 12u moves into its waiting position. During the process, continuous monitoring and / or storage and / or measurement and / or correction of the lower sample holder 12u can be performed to ensure that the previously measured position of the alignment marks 6ul and 6ur can be precisely located again later. This position error correction is performed by a PEC measuring system 13, in particular an interferometer. In even more preferred embodiments, the position error correction is performed by additional PEC optics, which, however, are located below the sample holder 12u. Such PEC optics have not been shown for the sake of clarity.
[0080] The Figure 9a or the Figure 9b Figure 1 shows a schematic bottom view or side view of a third process step 102 according to the invention, in which the upper, inner alignment marks 6ol and 6or are measured. Note that the upper, left alignment mark 6ol is shown in the bottom view of the Figure 9a is located on the top, while the lower, left alignment mark 6ul is in the top view of the Figure 7a shown on the underside. In this alignment step according to the invention, the positions of the upper, inner alignment marks 6ol and 6or, as well as the position of the upper sample holder 12o in the measurement position for the upper substrate 2o, are stored.
[0081] The Figure 10a or the Figure 10bFigure 1 shows a schematic bottom view or side view of a fourth process step 103, in which the upper, outer alignment marks 6ol' and 6or' are measured. In this alignment step, the positions of the upper, outer alignment marks 6ol' and 6or' are stored. If the position of the upper sample holder 12o in the measurement position for the upper substrate 2o has not yet been stored, it can still be done in this process step.
[0082] The Figure 11a or the Figure 11bFigure 1 shows a schematic top view or side view of a fifth process step 104, in which the sample holders 12o and 12u are changed. The lower sample holder 12u is moved back to its original position, particularly with the help of the data from the position error correction from the second process step 101. Simultaneously, the upper sample holder 12o is moved into the waiting position.
[0083] The Figure 12a or the Figure 12b Figure 105 shows a schematic top view or side view of a sixth process step, in which the lower, outer alignment marks 6ul' and 6ur' are measured. In this alignment step, the positions of the lower, outer alignment marks 6ul' and 6ur' are stored.
[0084] The Figure 13The diagram shows the 3 sigma value of the alignment error as a function of the number of measured alignment marks. As the number of alignment marks increases, the 3 sigma value of the alignment error decreases inversely proportional to the square root of the number of alignment marks used. Reference sign
[0085] 1, 1' Substrate type 2, 2', 2u, 2o Substrate 2us, 2os Substrate surface 3 Functional unit 4 Notch 5 Flat side 6, 6i, 6', 6i`, 6l, 6l', 6r, 6r' 6ul, 6ul', 6ur, 6ur', 6ol, 6ol', 6or, 6or' Alignment marks 7i, 7, 7i', 7' Centric position 8i, 8, 8i', 8' Edges 9, 9' Statistically averaged position 10, 10ul, 10ur, 10ol, 10or, 10ul', 10ur', 10ol', 10or' Viewing areas 11ol, 11or, 11ul, 11ur Optics 12u, 12o Sample holder 13PEC Measuring system rl, rl',rr, rr' Radii M Plane, in particular plane of symmetry L1, L2, L3 Lines h Intersection L Left area R Right area Z Center D Diameter line N Extended circle
Claims
1. A method for aligning two substrates (2u, 2o) to be bonded, wherein the first substrate (2u) has a first substrate surface (2us) with a first substrate surface side and a second substrate surface side lying opposite the first substrate surface side, wherein the second substrate (2o) has a second substrate surface (2os) to be bonded to the first substrate surface (2us), wherein the second substrate surface (2os) has a third substrate surface side and a fourth substrate surface side lying opposite the third substrate surface side, wherein the method comprises at least the following steps: -- detection and storage of first positions of a first alignment mark pair (6ul, 6ur) on the first substrate surface (2us) of the first substrate (2u), wherein a first alignment mark (6ul) of the first alignment mark pair (6ul, 6ur) is arranged on the first substrate surface side and a second alignment mark (6ur) of the first alignment mark pair (6ul,6ur) is arranged on the second substrate surface side, -- acquisition and storage of second positions of a second alignment mark pair (6ol, 6or) on the second substrate surface (2os) of the second substrate (2o), wherein a third alignment mark (6ol) of the second alignment mark pair (6ol, 6or) is arranged on the third substrate surface side and a fourth alignment mark (6or) of the second alignment mark pair (6ol, 6or) is arranged on the fourth substrate surface side, -- detection and storage of third positions of a third alignment mark pair (6ol', 6or') on the second substrate surface (2os) of the second substrate (2o), wherein a fifth alignment mark (6ol') of the third alignment mark pair (6ol, 6or) is arranged on the third substrate surface side and a sixth alignment mark (6or') of the third alignment mark pair (6ol', 6or') is arranged on the fourth substrate surface side, -- detection and storage of fourth positions of a fourth alignment mark pair (6ul', 6ur') on the first substrate (2u), wherein a seventh alignment mark (6ul') of the fourth alignment mark pair (6ul', 6ur') is arranged on the first substrate surface side and an eighth alignment mark (6ur') of the fourth alignment mark pair (6ul', 6ur') is arranged on the second substrate surface side, -- alignment of the two substrates (2u, 2o) with one another depending on the detected first positions, second positions, third positions and fourth positions, wherein the first alignment mark (6ul), the second alignment mark (6ur), the third alignment mark (6ol), the fourth alignment mark (6or), the fifth alignment mark (6ol'), the sixth alignment mark (6or'), the seventh alignment mark (6ul') and the eighth alignment mark (6ur') are less than 30 mm from the edge of the first substrate (2u) or the second substrate (2o), wherein the positions of the alignment marks (6ul, 6ul', 6ur, 6ur', 6ol, 6ol', 6or, 6or') are optically detected, wherein at least four optics (11ol, 11or, 11ul, 11ur ) are arranged, characterised in that a plurality of alignment marks (6ul, 6ul', 6ur, 6ur', 6ol, 6ol', 6or, 6or') of the same substrate surface side are simultaneously present and measured in the field of vision of one of the at least four optics (11ol, 11or, 11ul, 11ur ).
2. The method according to claim 1, with the following steps, in particular in the following sequence: -- arrangement and fixing of the substrates (2u, 2o) on substrate holders (12u, 12o), -- movement of the first substrate holder (12u) with the first substrate (2u) into a first detection position, -- detection and storage of the first positions of the first alignment mark pair (6ul, 6ur) on the first substrate surface (2us) of the first substrate (2u) in the first detection position, -- movement of the second substrate holder (12o) with the second substrate (2o) into a second detection position and movement of the first substrate holder (12u) with the first substrate (2u) into a first waiting position, -- detection and storage of the second positions of the second alignment mark pair (6ol, 6or) on the second substrate surface (2os) of the second substrate (2o) in the second detection position, -- detection and storage of the third positions of the third alignment mark pair (6ol', 6or') on the second substrate surface (2os) of the second substrate (2o) in the second detection position of the second substrate holder (12o), -- movement the first substrate holder (12u) with the first substrate (2u) into the first detection position and movement of the second substrate holder (12o) with the second substrate (2o) into a second waiting position, -- detection and storage of the fourth positions of the fourth alignment mark pair (6ul', 6ur') on the first substrate (2u) in the first detection position, -- alignment of the two substrates (2u, 2o) with each other depending on the detected first positions, second positions, third positions and fourth positions.
3. The method according to any one of the preceding claims, wherein the positions of at least three alignment marks (6ul, 6ul', 6ur, 6ur', 6ol, 6ol', 6or, 6or'), preferably of at least four alignment marks (6ul, 6ul', 6ur, 6ur', 6ol, 6ol', 6or, 6or'), even more preferably of at least five alignment mark (6ul, 6ul', 6ur, 6ur', 6ol, 6ol', 6or, 6or'), most preferably of at least six alignment marks (6ul, 6ul', 6ur, 6ur', 6ol, 6ol', 6or, 6or') are detected on the substrates (2u, 2o) per substrate surface side.
4. The method according to any of the preceding claims, wherein the alignment marks (6ul, 6ul', 6ur, 6ur', 6ol, 6ol', 6or, 6or') of a substrate surface side are arranged inside a circle K with a radius less than 100 mm, preferably less than 50 mm, more preferably less than 25 mm, most preferably less than 10 mm, with utmost preference less than 1 mm.
5. The method according to any of the preceding claims, wherein the positions of the substrate holders (12u, 12o) are detected by interferometers and / or optics, which in particular are independent of detection devices for the detection of the positions of the alignment marks (6ul, 6ul', 6ur, 6ur', 6ol, 6ol', 6or, 6or').
6. The method according to any one of the preceding claims, wherein the substrate holders (12o, 12u) are fixed in the detection positions.