Abstract
A method for bonding at least three substrates to form a substrate stack, wherein the substrate stack has at least one lowermost substrate a middle substrate, and an upper substrate. The method includes the following steps: aligning the middle substrate to the lowermost substrate and bonding the middle substrate to the lowermost substrate, then aligning the upper substrate and bonding the upper substrate to the middle substrate, wherein the upper substrate is aligned to the lowermost substrate.
Claims
1. A method for bonding at least three substrates to form a substrate stack, wherein the substrate stack has at least one lowermost substrate, a middle substrate, and an upper substrate, said method comprising: aligning the middle substrate to the lowermost substrate and bonding the middle substrate to the lowermost substrate, and aligning the upper substrate and bonding the upper substrate to the middle substrate, wherein the upper substrate is aligned to the lowermost substrate.
2. The method according to claim 1, wherein each of the at least three substrates has a plurality of optical lenses.
3. The method according claim 1, wherein the upper substrate is aligned to alignment marks of the lowermost substrate.
4. The method according to claim 1, wherein each subsequent substrate is aligned to the lowermost substrate.
5. The method according to claim 2, wherein optical axes, of the plurality of lenses, which are arranged on top of one another, are aligned congruently.
6. The method according to claim 1, wherein at least four substrates are bonded to one another.
7. The method according to claim 1 wherein the method includes: arranging and fixing the lowermost substrate to a lower substrate holder, detecting alignments marks on the lowermost substrate in fields of view of lenses, arranging and fixing the middle substrate on an upper substrate holder, detecting alignment marks on the lowermost substrate by use of the lenses, aligning the middle substrate to the lowermost substrate, bonding the middle substrate to the lowermost substrate, wherein the bonded substrates remain on the lower substrate holder, arranging and fixing the upper substrate on the upper substrate holder, detecting alignment marks on the upper substrate by use of the lenses, aligning the upper substrate to the lowermost substrate, bonding the upper substrate to the middle substrate to produce the substrate stack.
8. The method according to claim 7, wherein the alignment marks are detected by lenses (7ul, 7ur), which are arranged below the lowermost substrate, through recesses in the lower substrate holder.
9. The method according to claim 8, wherein the recesses are embodied as continuous holes in the lower substrate holder.
10. The method according to claim 8, wherein the recesses are embodied as elongated holes in the lower substrate holder.
11. The method according to claim 7, wherein the lower substrate holder is moveable such that the lower lenses are arranged inside recesses.
12. The method according to claim 1, wherein the lower substrate holder is moved in the Z direction by use of a Z-positioning unit such that a deviation in an X and Y direction is minimized.
13. The method according to claim 1, wherein a readjusting unit corrects a deviation of the lower substrate holder in the X and/or Y direction in response to a movement in the Z direction.
14. A substrate holder for a method according to claim 1, wherein the substrate holder has recesses for accommodating lenses and fixing elements for fixing a substrate.
15. The substrate holder according to claim 14, wherein the fixing elements include vacuum openings.
Description
[0092] Further advantages, features and details of the invention result from the following description of preferred exemplary embodiments as well as from the drawings. Schematically:
[0093] FIG. 1a shows three different optionspositions ptical axes of two lenses relative to one another.
[0094] FIG. 1b shows three further options of positions of optical axes of two lenses relative to one another,
[0095] FIG. 2a shows a first step of a first method according to the invention,
[0096] FIG. 2b shows a second step of the first method according to the invention,
[0097] FIG. 2c shows a third, optional step of the first method according to the invention,
[0098] FIG. 2d shows a fourth step of the first method according to the invention,
[0099] FIG. 2e shows a fifth step of th.e first method accordingto th invention,
[0100] FIG. 2f shows the sixth step of the first method according to the invention,
[0101] FIG. 2g shows a seventh step of the first method according to the invention
[0102] FIG. 2h shows an eighth step of the first method according to the invention.
[0103] FIG. 2i shows a ninth step of the first method according to the invention.
[0104] FIG. 3a shows a first step of a second method according o the invention,
[0105] FIG. 3b shows a second step of the second method according to the invention,
[0106] FIG. 3c shows a third, optional step of the second method according to the invention,
[0107] FIG. 3d shows a fourth step of the second method according to the invention.
[0108] FIG. 3e shows a fifth step of the second method according to the invention,
[0109] FIG. 3f shows a sixth step of the second method according to the invention,
[0110] FIG. 3g shows a seventh step of the second method according to the invention,
[0111] FIG. 3h shows an eighth step of the second method according to the invention,
[0112] FIG. 3i shows a ninth step of the second method according to the invention,
[0113] FIG. 4a shows a schematic, not to scale upper view of a first substrate holder according to the invention,
[0114] FIG. 4b shows a schematic, not to scale upper view of a second substrate holder according to the invention,
[0115] FIG. 5a shows a first flow chart,
[0116] FIG. 5b shows a second flow chart,
[0117] FIG. 6a shows a substrate stack produced by means of a method according to the prior art, and
[0118] FIG. 6b shows a substrate stack produced by means of a method according to the invention.
[0119] Identical components or components with the identical function are identified with identical reference numerals in the figures.
[0120] Publication U.S. Pat. No. 6,214,692 B1 discloses an exemplary device for carrying out the following method. The disclosure of this publication is thus included in this description.
[0121] A left detection unit 81 has a lower left lens 7ul and an upper left lens 7ol. A right detection unit 8r has a lower right lens 7ur and an upper right lens 7or.
[0122] FIG. 1a shows three different options of positions of optical axes 10ol, 10u1 of the left lenses 7u1, 7o1 relative to one another. In the first case (left), there is no point of intersection of the optical axes 10ul, 10ol. In the second case (middle), the optical axes 10ol intersect in a left alignment mark 51 on a substrate. This is the most frequently occurring case. In the third case (right), the two optical axes 10ul, 10ol are collinear (congruent) and centrically pierce the left alignment mark 51. This is the optimal case. Prior to the method according to the invention, the lenses 7o1, 7ul are calibrated in such a way that at least the middle (second) case is realized, the right (third) case is to be realized, if possible.
[0123] FIG. 1b shows three different options of positions of optical axes 10or, 10ur of the right lenses 7ur, 7or relative to one another. In the first case (left), there is no point of intersection. In the second case (middle), the optical axes 10or, 10ur intersect in a right alignment mark 5r. This is the most frequently occurring case. in the third case (right), the two optical axes 10or, 10ur are collinear and centrically pierce the right alignment mark 5r. This is the optimal case. Prior to the method according to the invention, the lenses 7or, 7ur are calibrated in such a way that at least the middle case is realized, the right case is to be realized, if possible.
[0124] FIG. 2a shows a first process step 100 of a first exemplary method according to the invention.
[0125] A lower substrate 4u has a lower, left alignment mark 5ul and a lower, right alignment mark 5ur, and is fixed to a lower substrate holder 1u in such a way that the alignment marks 5ul, 5ur for the lower left lens 7u1 and the lower right lens 7ur can be detected and measured through recesses 3 in the substrate holder 1u (see upper illustration of FIG. 2a). In general, the alignment marks 5ul, 5ur are not yet in a depth of field t of the lenses 7ul, 7ur at that point in time, whereby the alignment marks 5u1, 5ur appear to be correspondingly out of focus. The alignment marks 5u1, 5ur, however, should preferably already at least be in fields of view 6ul, 6ur of the lenses 7ul, 7ur (see lower illustrations). This can be attained by means of a mechanical preadjustment (suggested by means of the arrow). If the alignment marks 5ul, 5ur are not yet in the fields of view 6ul, 6ur, the detection units 81, 8r must at least be moved in the x and/or y direction until the alignment marks 5ul, 5ur become visible for the lower lenses 7ul, 7ur. The exact alignment of the detection units 81, 8r on the alignment marks 5ul, 5ur only occurs in FIG. 2c. The lenses 7ul, 7ur must not be moved individually, because previously performed calibration would otherwise be lost.
[0126] FIG. 2b shows a second process step 101 of the method. The lower substrate holder 1u is moved in the Z direction (suggested by means of the arrow), until the two alignment marks 5ul, 5ur of the lower substrate 4u are located in the depth of field t. If necessary, a wedge error compensation can also be performed in this process step. A wedge error can be detected in that one of the two alignment marks 5ul, 5ur is depicted out of focus, because it is located outside of the depth of field t, while the second one of the alignment marks 5ul, 5ur is depicted in focus, because it is still located inside the depth of field t. The wedge error, however, is preferably detected by means of more precise measuring devices, in particular by means of interferometers, and is corrected by tilting the lower substrate holder 1u accordingly.
[0127] FIG. 2e shows a third, optional process step 102. The detection units 81, 8r are thereby moved in the X and/or Y direction (suggested by means of the arrow), until the alignment marks 5ul, 5ur have been centered by the lower lenses 7ul, 7ur, or until the alignment marks 5ul, 5ur are illustrated in a centered manner in the lenses 7ul, 7ur. The alignment marks 5ul, 5ur then coincide as accurately as possible, preferably exactly, with the points of intersection of the optical axes of the lenses 7ul, 7ol or 7ur, 7or, respectively. The detection units 81, 8r with the lenses 7ul, 7ur, 7ol, 7or are then fixed and are no longer moved during the process according to the invention. The translational positions and/or rotational positions of the lower substrate holder 1u are stored.
[0128] FIG. 2d shows a fourth process step 103. The lower substrate lu is moved in the negative z direction (suggested by means of the arrow), to clear a process region for an upper substrate holder 1o (not illustrated). It is necessary thereby to move the lower substrate holder 1u as precisely as possible. As precisely as possible means that the deviation of the lower substrate holder 1u in the x and/or y direction with movement in the z direction needs to be minimal. In art alternative embodiment according to the invention, the substrate holder iu can also be moved out of the process region. The substrate holder 1u must then be moved into the process region again prior to the process step 105 from FIG. 2f.
[0129] FIG. 2e shows a fifth process step 104. The upper substrate holder to positions (suggested by means of the arrow) a loaded and fixed upper substrate 4o (see lower illustration) in such a way that upper alignment marks 5ol, 5or on the upper side of the upper substrate 4o are located in fields of view 6ol, 6or of the upper lenses 7ol, 7or and are centered correctly (see upper illustrations). The upper substrate holder 1 o is preferable constructed in such a way that only a movement in the x and y direction, but not in the z direction is possible. By loading the upper substrate 4o on the upper substrate holder lo, the upper alignment marks 5ol, 5or are thus preferably already in the depth of field range t. However, a device, in which the upper substrate holder 1o can in fact be moved in the z direction, at least across short distances, to compensate for small height errors, is also conceivable.
[0130] The process step 103 from FIG. 2d and the process step 104 from FIG. 2c in particular occur simultaneously.
[0131] FIG. 2f show a sixth process step 105. In this process step, the fixing and/or bonding process of the two substrates 4u, 4o occurs. In general, a relative approach of the two substrates 4u and 4o occurs. In a specific, illustrated embodiment according to the invention, the lower substrate holder 1u moves the lower substrate 4u in the z direction on contact with the upper substrate 4o (suggested by means of the arrow). A very important aspect is that the deviation of the lower substrate holder 1u in the x and/or y direction in relation to the ideal x and/or y position is minimal at the time of the contact between an upper substrate surface 4u of the lower substrate 4u and a lower substrate surface 4os of the upper substrate 4o. In another embodiment according to the invention, it would also be possible in this process step, to bring the upper substrate holder to closer to the lower substrate holder lu. Analogous considerations then apply for the highly precise positioning in the z direction. In a very specific embodiment according to the invention, both substrate holders 1u and 1o would approach one another.
[0132] FIG. 2g shows a seventh process step 106, in which the lower substrate holder 1u is moved in the negative z direction (suggested by means of the arrow) with a formed substrate stack 9, which consists of the first two bonded substrates 4u, 4o, to clear the process region for the upper substrate holder 1o. It is also conceivable that the substrate holder 1u leaves the process region and returns at a point in time prior to the next bonding process. The process steps 104-106 can now be performed several times.
[0133] FIG. 2h show an eighth process step 107, in which a third substrate 4o with alignment marks 5ol, 5or is fixed on its upper side by the upper substrate holder, and is arranged above the substrate stack 9. Analogous to FIG. 2e or step 104, respectively, the upper substrate holder 1o (suggested by means of the arrow) positions the loaded and fixed further upper substrate 4o (see lower illustration) in such a way that the upper alignment marks 5ol, 5or are located on the upper side of the upper substrate 4o in the fields of view 6ol, 6or of the upper lenses 7ol, 7or and are centered correctly (see upper illustrations). The upper substrate holder 1o is preferably constructed in such a way that only a movement in the x and y direction, but not in the z direction is possible. By loading the upper substrate 4o on the upper substrate holder lo, the upper alignment marks 5ol, 5or are thus preferably already in the depth of field range t. However, a device, in which the upper substrate holder 1o can in fact be moved in the z direction, at least across short distances, to compensate for small height errors, is also conceivable.
[0134] The bonding process follows analogously to FIG. 2f or step 105, respectively, as well as the clearing of the process region analogously to FIG. 2g or step 106, respectively.
[0135] Further substrates can then be aligned and bonded analogously to steps 107, 105, 106. According to the invention, the substrate n+1, which is to be bonded, is aligned to the first substrate and not to the n-th substrate, so that the alignment error can he minimized.
[0136] FIG. 3a shows a first method step 200 of a second exemplary method according to the invention. A left lower alignment mark 5ul and a right lower alignment mark 5ur are arranged on the upper side of the lower substrate 4u (see lower illustration). In general, the alignment marks 5ul, 5ur are not yet in a depth of field t of the lenses 7ol, 7or at that point in time, whereby the alignment marks 5ul, 5ur appear to be correspondingly out of focus (see upper illustrations). The alignment marks 5ul, 5ur, however, should preferably already at least be in fields of view 6ol, 6or. This can always be attained by means of a mechanical preadjustment. If the alignment marks 5ul, 5ur are not yet located in the fields of view 6ol, 6or, the detection units 81, 8r must be moved in the x and/or y direction until the alignment marks 5ul, 5ur for the upper lenses 7ol, 7or become visible. In contrast to the first process, only the upper lenses 7ol and 7or are preferably used. The detection units 81, 8r are nonetheless moved as a whole. The lower lenses 7ul, 7ur could be used to detect the alignment marks 5ul, 5ur, provided that the substrate 4u is transparent.
[0137] In this case, an observation of all alignment marks 5ul, 5ur, 5ol, 5or from the upper side, and an observation of the alignment marks 5ul, 5ur from the lower side would be possible. However, this would not change anything about the process according to the invention, because the calibration of each substrate also occurs to the respective first substrate in this case. Even though the lower lenses 7ul, 7or are still illustrated in the further figures, they are not described in detail.
[0138] FIG. 3b shows a second method step 201 of the second method according to the invention. The lower substrate holder 10 is moved in the Z direction, until the two alignment marks 5ul, 5ur of the lower substrate 4u are located in the depth of field t. If necessary, a wedge error compensation can also be performed in this process step. A wedge error can be detected in that one of the two alignment marks 5ul, 5ur is depicted out of focus, because it is located outside of the depth of field t, while the second one of the alignment marks 5u, 5ur is illustrated in focus, because it is still located within the depth of field t. The wedge error, however, is preferably detected by means of more precise measuring devices, in particular by means of interferometers, and is corrected by tilting the lower substrate holder 1u accordingly,
[0139] FIG. 3e shows a third, optional process step 202. The detection units 81, 8r are thereby moved in the Z direction, until the alignment marks 5ul, 5ur have been centered by the upper lenses 7ol, 7or. The detection units 81 and 8r are then fixed and are no longer moved during the process according to the invention. The translational positions and/or rotational positions of the lower substrate holder 1u are stored.
[0140] FIG. 3d shows a fourth process step 203. The lower substrate lu is moved in the negative z direction, in order to clear the process region for the upper substrate holder 1o. It is necessary thereby to move the lower substrate holder 1u as precisely as possible. As precisely as possible means that the deviation of the lower substrate holder 1u in the x and/or y direction with movement in the z direction needs to be minimal. In an alternative embodiment according to the invention, the substrate holder 1u can also be moved out of the process region. The substrate holder 1u must then be moved into the process region again prior to the process step 205 from FIG. 3f.
[0141] FIG. 3e shows a fifth process step 204. The upper substrate holder 1o positions the loaded and fixed upper substrate 4o in such a way that upper alignment marks 5ol, 5or are located in the fields of view 6ol, 6or of the upper lenses 7ol, 7or and are centered correctly. The upper substrate holder 1o is preferable constructed in such a way that only a movement in the x and y direction, but not in the z direction is possible. By loading the upper substrate 4o on the upper substrate holder lc, the upper alignment marks 5ol, 5or are thus preferably already located in the depth of field range t. However, a device, in which the upper substrate holder 4o can in fact be moved in the z direction, at least across short distances, to compensate for small height errors, is also conceivable.
[0142] The process step 203 from FIG. 3d and the following process step 204 from FIG. 3f in particular take place simultaneously.
[0143] FIG. 3f show a sixth process step 205. In this process step, the fixing and/or bonding process of the two substrates 4u, 4o occurs. In general, a relative approach of the two substrates 4u and 4o occurs. In a specific, illustrated embodiment according to the invention, the lower substrate holder 1u moves the lower substrate 4u on contact with the upper substrate 4o. A very important aspect according to the invention is that the deviation of the lower substrate holder 1u in the x and/or y direction in relation to the ideal x and/or y position is minimal at the time of the contact between the upper substrate surface 4us and the substrate surface 4os. In another embodiment according to the invention, it would also be possible in this process step, to bring the upper substrate holder 1o closer to the lower substrate holder 1u. Analogous considerations then apply for the highly precise positioning in the z direction. In a very specific embodiment according to the invention, both substrate holders 1u and 1o would approach one another.
[0144] FIG. 3g shows a seventh process step 206, in which the formed substrate stack 9. which was bonded from the first two substrates 4u, 4o, is moved in the negative z direction, to clear the process region for the upper substrate holder 1o. It is also conceivable that the substrate holder 1u leaves the process region and returns at a point in time prior to the next bonding process. The process steps 204-206 can now be performed several times.
[0145] FIG. 3h shows an eighth process step 207, in which a third substrate 4o with alignment marks 5ol, 5or is fixed on its upper side by the upper substrate holder, and is arranged above the substrate stack 9, Analogous to FIG. 3e or step 204, respectively, the upper substrate holder 1o (suggested by means of the arrow) positions the loaded and fixed further upper substrate 4o (see lower illustration) in such a way that the upper alignment marks 5ol, 5or are located on the upper side of the upper substrate 4o in the fields of view 6ol, 6or of the upper lenses 7ol, 7or and are centered correctly (see upper illustrations). The upper substrate holder 1o is preferably constructed in such a way that only a movement in the x and y direction, but not in the z direction is possible. By loading the upper substrate 4o on the upper substrate holder 1o, the upper alignment marks 5or are thus preferably already located in the depth of field range t. However, a device, in which the upper substrate holder to can in fact be moved in the z direction, at least across short distances, so as to compensate for small height errors, is also conceivable.
[0146] FIG. 3i shows an eighth process step 207, in which the complete substrate stack 9 is unloaded from the lower substrate holder 1u.
[0147] Apart from that, the statements made with regard to the first process apply for the second process.
[0148] FIG. 4a shows a schematic, not to scale upper view of a first substrate holder 1 according to the invention, consisting of a substrate holder plate, on which, in particular a plurality of fixing elements 2 are present to fix a substrate 4 (only illustrated partially). The fixing elements 2 are preferably vacuum fixations, in particular vacuum openings, as part of vacuum channels. A plurality, in particular at least two recesses 3, which are preferably arranged at the same distance from the respective edge is of the substrate holder 1, are incorporated in the substrate holder plate. The recesses 3 are preferably continuous holes, which are embodied in a circular manner in this embodiment. The recesses 3, however, can have any shape, in particular a rectangular shape. Recesses 3 with a complicated milled geometry are also possible.
[0149] FIG. 4b shows a schematic upper view of a second substrate holder 1 according to the invention comprising a substrate holder plate, on which, in particular a plurality of fixing elements 2 are present to fix a substrate 4 (only illustrated partially). The fixing elements 2 are preferably vacuum fixations, in particular vacuum openings, as part of vacuum channels. A plurality, in particular at least two recesses 3, are incorporated in the substrate holder plate, The recesses 3 are preferably elongated recesses 3, which are embodied by an edge 1s of the substrate holder 1 and which lead into the interior and which were in particular created by means of a milling process. Due to the embodiment of the recesses 3 according to the invention, it is advantageously not necessary to lift the substrate holder 1 over the lower lenses 7ul, 7ur.
[0150] The mentioned substrate holders 1, 1 can be used on the lower side and/or upper side of a system according to the invention.
[0151] FIG. 5a shows a flow chart of the first method according to the invention. The process steps 100 to 103 are only performed once, while the process steps 104 to 106 can be performed a total of n-times. For the sake of clarity, the process step 107 was illustrated in a separate FIG. 21i and is equivalent to the process step 104, with the difference that a third substrate 4o is already bonded here on a substrate stack 9, which has already been created.
[0152] FIG. 5b shows a flow chart of the second method according to the invention. The process steps 200 to 203 are only performed once, while the process steps 204 to 206 can be performed a total of n-times. For the sake of clarity, the process step 2107 was illustrated in a separate FIG. 3h and is equivalent to the process step 204, with the difference that a third substrate 4o is already bonded here on a substrate stack 9, which has already been created.
[0153] FIG. 6a shows the prior art, namely a partial section of a substrate stack 9, consisting of four substrates 4, 4, 4, 4, which contain an increasing number of alignment errors and which are stacked on top of one another. In an exemplary manner, the substrates 4, 4, 4, 4 are monolithic lens wafers. Optical axes 12, 12, 12, 12 of optical elements 13, 13, 13, 13, in particular lenses, are not collinear. It can in particular be seen that the distance between two optical axes of two consecutive optical elements 13, 13, 13, 13 increases.
[0154] FIG. 6b shows a partial section of a substrate stack 9, produced by means of the method according to the invention, consisting of four substrates 4, 4, 4, 4, which contain a consistent amount of alignment errors and which are stacked on top of one another. In an exemplary manner, the substrates 4, 4, 4, 4 are monolithic lens wafers. The optical axes 12 of the optical elements 13, 13, 13, 13, in particular lenses, are not collinear. It can be seen, however, that the distance between the optical axes of the n-th substrate 4, 4, 4 and the first substrate 4 is approximately identical, because according to the invention, the n-th substrate 4, 4, 4 is aligned to the first substrate 4. An optimal state would be a stacking and alignment of the substrates 4, 4, 4, 4 in such a way that the optical axes 12, 12, 12, 12 of all optical elements 13, 13, 13, 13 are collinear.
LIST OF REFERENCE NUMERALS
[0155] 1, 1, 1u, 1o substrate holder [0156] 1s, 1s edge [0157] 2, 2 fixations [0158] 3, 3 recesses [0159] 4o, 4o, 4u, 4, 4, 4, 4 substrate. [0160] 5ul, 5ur, 5ol, 5or, 5ul, 5ur alignment marks [0161] 6ul, 6ur, 6ol, , 6or, field of view [0162] 7ul, 7ur, 7or lens [0163] 81, 8r detection unit [0164] 9, 9, 9, 9 substrate stack [0165] 10ul, 10ol, 10ur, 10or optical axes [0166] 12, 12, 12, 12 optical axis [0167] 13, 13, 13, 13 optical element [0168] t depth of field
