Substrate bonding apparatus and substrate bonding method

Abstract

To provide a substrate bonding device capable of bonding a plurality of pairs of first substrates and second substrates in a shorter time.SOLUTION: A substrate bonding device includes an actuator mechanism that bonds a first substrate W1 held by a first chuck 11 and a second substrate W2 held by a second chuck 21 by relatively moving the first chuck 11 and the second chuck 21, and an inspection camera 43 that photographs the first substrate W1 and the second substrate W2 held by the second chuck 21 after the first substrate W1 and the second substrate W2 are bonded, and a transport robot TR that, when the first substrate W1 and the second substrate W2 held by the second chuck 21 are photographed by the inspection camera 43, transports the second first substrate W1 to the first chuck 11.SELECTED DRAWING: Figure 31G

Description

Technical Field

[0006] , , ,

[0005] , ,

[0001] The present invention relates to a substrate bonding device and a substrate bonding method for bonding two substrates. The substrates include, for example, semiconductor wafers, substrates for FPD (Flat Panel Display) such as liquid crystal display devices and organic EL (electroluminescence) display devices, substrates for optical disks, substrates for magnetic disks, substrates for magneto-optical disks, substrates for photomasks, ceramic substrates, substrates for solar cells, and the like.

Background Art

[0007] In the above embodiment, at least one of the following features may be added to the substrate bonding apparatus.

[0008] The actuator mechanism includes an inversion actuator that is transported to the first chuck by the transport robot and inverts the second first substrate held in the first chuck when the inspection camera is photographing the first substrate and the second substrate held in the second chuck.

[0009] The actuator mechanism further includes a horizontal actuator that moves the second chuck horizontally relative to the first chuck between a joining position in which the first substrate held by the first chuck and the second substrate held by the second chuck are joined, and a shooting position in which the inspection camera photographs the first substrate and the second substrate held by the second chuck.

[0010] Another embodiment of the present invention provides a substrate joining method comprising the steps of: holding a first substrate with a first chuck; holding a second substrate with a second chuck; joining the first substrate held in the first chuck and the second substrate held in the second chuck by moving the first chuck and the second chuck relative to each other; releasing the first substrate from the first chuck; having an inspection camera photograph the first substrate and the second substrate held in the second chuck after the first substrate and the second substrate have been joined; and transporting a second first substrate to the first chuck by a transport robot while the inspection camera is photographing the first substrate and the second substrate held in the second chuck. The step of releasing the first substrate from the first chuck is a step of changing the first substrate from a state in which the first substrate is fixed to the first chuck to a state in which the first substrate is movable relative to the first chuck. The release of the first substrate may be performed after the bonding of the first substrate and the second substrate is completed, or it may be performed while the bonding of the first substrate and the second substrate is in progress. At least one of the above-described features relating to the substrate bonding apparatus may be added to the substrate bonding method. [Brief explanation of the drawing]

[0011] [Figure 1] This is a schematic plan view of a substrate bonding apparatus according to the first embodiment of the present invention. [Figure 2] This is a process diagram illustrating an example of a substrate bonding method performed by a substrate bonding apparatus. [Figure 3] This is a schematic diagram showing an example of cross-sections of the first and second substrates before and after bonding. [Figure 4] This is a schematic plan view of the first substrate to illustrate the first alignment mark on the first substrate. [Figure 5] This is a schematic plan view of the second substrate to illustrate the second alignment mark on the second substrate. [Figure 6]It is a schematic diagram showing the first alignment mark and the second alignment mark before and after alignment adjustment of the first substrate and the second substrate. [Figure 7] It is a perspective view of the bonding unit. [Figure 8] It is a plan view of the bonding unit. [Figure 9] It is a view of the bonding unit horizontally viewed in the Y direction. [Figure 10A] It is a schematic plan view of the first chuck. [Figure 10B] It is a schematic plan view of the first chuck. [Figure 11] It is a view of the first bonding module horizontally viewed in the X direction. [Figure 12] It is a plan view of the first bonding module. [Figure 13] It is a plan view of the first stage. [Figure 14] It is a view of the first stage horizontally viewed in the direction of arrow XIV shown in FIG. 13. [Figure 15] It is a view of the second bonding module horizontally viewed in the X direction. [Figure 16] It is an enlarged view of a part of FIG. 15. [Figure 17] It is a plan view of the second bonding module. [Figure 18] It is a view of the second bonding module horizontally viewed in the Y direction. [Figure 19] It is a plan view of the second stage. [Figure 20] It is a view of the second stage horizontally viewed in the direction of arrow XX shown in FIG. 19. [Figure 21] It is a plan view of the bonding unit. [Figure 22] It is a view of the second camera photographing the second substrate horizontally viewed in the Y direction. [Figure 23] It is a view of the first camera photographing the first substrate horizontally viewed in the Y direction. [Figure 24] It is a view of the alignment camera photographing the first fixed reference and the second fixed reference horizontally viewed in the Y direction. [Figure 25]Figure 24 shows the alignment camera viewed horizontally in the direction of arrow XXV. [Figure 26] This is a schematic diagram showing an example of an image of a cross-shaped first alignment mark generated by the first camera. [Figure 27] This is a block diagram of the displacement detector. [Figure 28] This is a diagram showing a vertical cross-section of the displacement detector. [Figure 29] Figure 28 shows the displacement detector viewed vertically in the direction of arrow XXIX. [Figure 30] This is a block diagram showing the electrical configuration of a substrate bonding device. [Figure 31A] This is a schematic diagram illustrating a first example of bonding a first substrate and a second substrate performed by a bonding unit. [Figure 31B] This is a schematic diagram illustrating a first example of bonding a first substrate and a second substrate performed by a bonding unit. [Figure 31C] This is a schematic diagram illustrating a first example of bonding a first substrate and a second substrate performed by a bonding unit. [Figure 31D] This is a schematic diagram illustrating a first example of bonding a first substrate and a second substrate performed by a bonding unit. [Figure 31E] This is a schematic diagram illustrating a first example of bonding a first substrate and a second substrate performed by a bonding unit. [Figure 31F] This is a schematic diagram illustrating a first example of bonding a first substrate and a second substrate performed by a bonding unit. [Figure 31G] This is a schematic diagram illustrating a first example of bonding a first substrate and a second substrate performed by a bonding unit. [Figure 31H] This is a schematic diagram illustrating a first example of bonding a first substrate and a second substrate performed by a bonding unit. [Figure 32A] This is a schematic diagram illustrating a second example of bonding a first substrate and a second substrate by a bonding unit. [Figure 32B] This is a schematic diagram illustrating a second example of bonding a first substrate and a second substrate by a bonding unit. [Modes for carrying out the invention]

[0012] Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0013] Figure 1 is a schematic plan view of a substrate bonding apparatus 1 according to the first embodiment of the present invention. Figure 2 is a process diagram illustrating an example of a substrate bonding method performed by the substrate bonding apparatus 1. Figure 3 is a schematic diagram showing an example of cross-sections of the first substrate W1 and the second substrate W2 before and after bonding.

[0014] The substrate bonding apparatus 1 is a device for bonding two disc-shaped substrates W. As shown in Figure 1, the substrate bonding apparatus 1 includes a plurality of load ports LP, each containing one of a plurality of carriers CA that accommodate multiple substrates W, such as FOUPs (Front-Opening Unified Pods); a plurality of processing units 2 that process the substrates W transported from the plurality of load ports LP; and a transport system TS that transports the substrates W between the load ports LP and the processing units 2, or between two processing units 2. The substrate bonding apparatus 1 further includes an outer wall 1a that forms a sealed space housing the plurality of processing units 2 and the transport system TS, and a control device 3 that controls the substrate bonding apparatus 1.

[0015] Figure 1 shows an example where three load ports LP are provided. The three load ports LP include a first load port LP1 where a carrier CA containing a first substrate W1 is placed, a second load port LP2 where a carrier CA containing a second substrate W2 is placed, and a third load port LP3 where a carrier CA containing the bonded first substrate W1 and second substrate W2 is placed. The first load port LP1 and the second load port LP2 are input ports where carrier CAs containing substrates W to be bonded by the substrate bonding apparatus 1 are placed. The third load port LP3 is an output port where a carrier CA containing the two substrates W bonded by the substrate bonding apparatus 1 is placed.

[0016] The transport system TS transports the first substrate W1 and the second substrate W2 to be joined from the first load port LP1 and the second load port LP2 to a plurality of processing units 2, and transports the joined first substrate W1 and the second substrate W2 from the plurality of processing units 2 to the third load port LP3. The transport system TS may also include at least one transport robot TR that transports one or more substrates W in a horizontal position on the transport path TP.

[0017] The transport robot TR includes at least one hand TH that holds a single substrate W in a horizontal position. The transport robot TR moves along the transport path TP while holding the substrate W horizontally with the hand TH. The thick line in Figure 1 represents the transport path TP. Figure 1 shows an example in which the transport path TP extends from a first load port LP1 and a second load port LP2 to multiple processing units 2, and returns from the multiple processing units 2 to a third load port LP3.

[0018] The multiple processing units 2 include at least one cleaning unit 2c for cleaning the substrates W to be joined, and at least one hydrophilization unit 2h for hydrophilizing the joining surfaces WA1 and WA2 (see Figure 3) of the substrates W to be joined. The multiple processing units 2 further include an inversion unit for inverting one of the two substrates W to be joined, a joining unit 2b for bringing the joining surfaces WA1 and WA2 of the two substrates W into contact, and an inspection unit for checking the joining accuracy of the two substrates W. Figure 1 shows an example where the inversion unit and the inspection unit are part of the joining unit 2b. The inversion unit may be a unit independent of the joining unit 2b. The same applies to the inspection unit.

[0019] The processing units 2 may further include a heating unit that increases the bonding strength of the two substrates W by heating the two substrates W joined by the bonding unit 2b. The processing units 2 may further include at least one pre-aligner that reduces the difference between the actual position and angle of the substrates W and the ideal position and angle of the substrates W by adjusting the position and angle of the substrates W before they are fed into the bonding unit 2b. The pre-aligner may adjust the position and angle of the substrates W with respect to the outer circumference of the substrates W, or it may adjust the position and angle of the substrates W with respect to alignment marks described later. The thickness of the substrates W to be joined may be measured by the pre-aligner or another processing unit 2.

[0020] An example of the procedure for joining two substrates W using the substrate joining apparatus 1 is as follows. Specifically, a pre-cleaning step (step S1 in Figure 2) is performed to wash and dry the first substrate W1 taken from the carrier CA on the first load port LP1 and the second substrate W2 taken from the carrier CA on the second load port LP2. Then, a hydrophilization step (step S2 in Figure 2) is performed to make the joining surfaces WA1 and WA2 of the two substrates W hydrophilic, and a post-cleaning step (step S3 in Figure 2) is performed to wash and dry the two hydrophilized substrates W.

[0021] After cleaning the two hydrophilized substrates W, an inversion step (step S4 in Figure 2) is performed to invert one of the two substrates W. Then, an alignment confirmation step (step S5 in Figure 2) is performed to check the alignment of the two substrates W to be joined, an alignment adjustment step (step S6 in Figure 2) is performed to adjust the alignment of the two substrates W based on the confirmed alignment, and a substrate contact step (step S7 in Figure 2) is performed to bring the two aligned substrates W into contact. After that, an inspection step (step S8 in Figure 2) is performed to check the joining accuracy of the two substrates W, that is, the amount of misalignment of the two joined substrates W and the amount of misalignment of the angle (angle around the center of the substrates W) of the two joined substrates W. Finally, the two joined substrates W are placed in the carrier CA on the third load port LP3.

[0022] The pre-cleaning step may be a single-wafer cleaning method in which a processing liquid such as a cleaning solution is supplied to multiple substrates W one by one, or a batch cleaning method in which a processing liquid such as a cleaning solution is supplied simultaneously to multiple substrates W arranged parallel to each other with intervals in between. The same applies to the post-cleaning step. Either the pre-cleaning step or the post-cleaning step may be a single-wafer cleaning method, and the other may be a batch cleaning method.

[0023] The pre-cleaning and post-cleaning processes may be performed in the same cleaning unit 2c or in separate cleaning units 2c. The cleaning unit 2c that performs the pre-cleaning process for the first substrate W1 may be the same as or different from the cleaning unit 2c that performs the pre-cleaning process for the second substrate W2. The same applies to the post-cleaning process. Figure 1 shows an example in which the pre-cleaning and post-cleaning processes for the first substrate W1 and the pre-cleaning and post-cleaning processes for the second substrate W2 are performed in separate cleaning units 2c.

[0024] The hydrophilization process may be a plasma treatment in which plasma, such as oxygen plasma, is irradiated onto the bonding surfaces WA1 and WA2 of the substrate W. In this case, moisture in the air and moisture supplied to the substrate W in the post-cleaning process come into contact with the bonding surfaces WA1 and WA2 of the substrate W irradiated with plasma, and hydrophilic groups such as hydroxyl groups (OH groups) are formed on the bonding surfaces WA1 and WA2 of the substrate W. The hydrophilization process may also be a wet treatment in which a hydrophilization liquid that forms hydrophilic groups on the bonding surfaces WA1 and WA2 of the substrate W is supplied to the bonding surfaces WA1 and WA2 of the substrate W. In this case, the hydrophilization unit 2h may perform at least one of a pre-cleaning process and a post-cleaning process in addition to the hydrophilization process. If the hydrophilization unit 2h performs both the pre-cleaning and post-cleaning processes, the cleaning unit 2c may be omitted.

[0025] The hydrophilization process may be a single-wafer hydrophilization process in which multiple substrates W are hydrophilized one by one, or a batch hydrophilization process in which multiple substrates W arranged parallel to each other with spacing in between are hydrophilized simultaneously. The hydrophilization unit 2h that performs the hydrophilization process on the first substrate W1 may be the same as or different from the hydrophilization unit 2h that performs the hydrophilization process on the second substrate W2. Figure 1 shows an example of the latter.

[0026] In the example shown in Figure 1, the alignment confirmation process, alignment adjustment process, and substrate contact process are performed by the bonding unit 2b. In this example, the inversion process and inspection process are also performed by the bonding unit 2b. The alignment confirmation process is a process of confirming the alignment of the two substrates W to be bonded based on images of the alignment marks (see Figure 4) of the two substrates W. The substrate contact process is a process of bonding the two substrates W by bringing the two aligned substrates W into contact. The allowable bonding accuracy (tolerance of the two bonded substrates W) is, for example, ±100 nm, and in some cases ±10 nm.

[0027] The substrate contact process may be a process of directly joining two substrates W in ambient air at room temperature. The substrate contact process may also be a face-to-face joining process in which the surfaces of the two substrates W are joined so that they face each other. In this case, the surfaces of the two substrates W correspond to the joining surfaces WA1 and WA2 of the two substrates W. The substrate contact process may also be a process of joining two substrates W by not pressing one of the two substrates W against the other, or by pressing one of the two substrates W against the other with a pressure that does not damage the devices formed on the two substrates W.

[0028] Figure 3 shows cross-sections of the first substrate W1 and the second substrate W2, cut in a plane perpendicular to the substrates W1 and W2. The ratio of the thickness of the device layers WC1 and WC2 to the thickness of the bonding layers WB1 and WB2 shown in Figure 3 may not be the same as the actual ratio. The left side of Figure 3 shows an example where a hydroxyl group, an example of a hydrophilic group, is formed on the bonding surface WA1 of the first substrate W1 and the bonding surface WA2 of the second substrate W2 before bonding. In this example, the oxygen atom (O) in the hydroxyl group is bonded to the silicon atom (Si) in the bonding layers WB1 and WB2.

[0029] The bonding surfaces WA1 of the first substrate W1 and WA2 of the second substrate W2, before bonding, are terminated with multiple hydroxyl groups. When the bonding surfaces WA1 of the first substrate W1 and WA2 of the second substrate W2 are brought into contact, the first substrate W1 and the second substrate W2 are bonded together by intermolecular forces acting between the two hydroxyl groups. In some cases, water molecules detach from the two hydroxyl groups, and silicon atoms in the bonding layer WB1 of the first substrate W1 and silicon atoms in the bonding layer WB2 of the second substrate W2 are bonded together via oxygen atoms. In this way, the first substrate W1 and the second substrate W2 are bonded together.

[0030] Next, we will explain the substrate W and alignment marks.

[0031] Figure 4 is a schematic plan view of the first substrate W1 to illustrate the first alignment mark AM1 of the first substrate W1. Figure 5 is a schematic plan view of the second substrate W2 to illustrate the second alignment mark AM2 of the second substrate W2.

[0032] In Figures 4 and 5, the first alignment mark AM1 and the second alignment mark AM2 are exaggerated for ease of understanding. The size of the alignment marks on the substrate W is not limited to the examples shown in Figures 4 and 5. The arrangement of the alignment marks on the substrate W and the number of alignment marks on a single substrate W are also not limited to the examples shown in Figures 4 and 5.

[0033] As shown in Figures 4 and 5, the first substrate W1 and the second substrate W2 are flat discs with equal diameters. The diameters of the first substrate W1 and the second substrate W2 may be 300 mm or other. The thermal expansion coefficient of the first substrate W1 may be equal to or different from that of the second substrate W2. The first substrate W1 contains a disc-shaped substrate WD1, and the second substrate W2 contains a disc-shaped substrate WD2 (see Figure 3). Substrates WD1 and WD2 are made of semiconductors such as silicon single crystals. Substrates WD1 and WD2 may be made of materials other than semiconductors.

[0034] Substrates WD1 and WD2 each include a circular front and back surface parallel to each other, and an annular end surface connecting the outer edges of the front and back surfaces. The front and back surfaces of substrates WD1 and WD2 are flat surfaces parallel to each other. The front surfaces of substrates WD1 and WD2 are device-forming surfaces on which devices are formed. The back surfaces of substrates WD1 and WD2 are non-device-forming surfaces on which devices are not formed. Both the front and back surfaces of substrate WD1 or substrate WD2 may be device-forming surfaces.

[0035] The outer periphery of substrate WD1 forms a V-shaped notch that opens at the end face of substrate WD1 when viewed perpendicular to the surface of substrate WD1. The outer periphery of substrate WD1 may also form an orientation flat (so-called orientation flat) instead of a notch. The notch and orientation flat indicate the crystal orientation of substrate WD1. The first substrate W1 is positioned in the circumferential direction of the first substrate W1 with reference to the notch or orientation flat of the first substrate W1. The same applies to the second substrate W2.

[0036] The first substrate W1 includes at least one first alignment mark AM1 which serves as a reference when aligning the first substrate W1. The first substrate W1 further includes at least one device formed on the surface of the substrate WD1, such as a transistor or a capacitor. The first alignment mark AM1 may be part of the device or it may be a structure unrelated to the device. The position of the first substrate W1 in a direction parallel to the surface of the first substrate W1 and the angle of the first substrate W1 around a perpendicular line perpendicular to the center of the surface of the first substrate W1 are determined based on the at least one first alignment mark AM1 provided on the first substrate W1. The shape and number of the first alignment marks AM1 are arbitrary as long as the position of the first substrate W1 and the angle of the first substrate W1 can be determined. The same applies to the second alignment mark AM2.

[0037] The first substrate W1 includes a device layer WC1 (see Figure 3) covering the surface of the base material WD1, and a junction layer WB1 (see Figure 3) covering the surface of the device layer WC1. The first alignment mark AM1 is located on the device layer WC1. Devices such as transistors are also located on the device layer WC1. The first alignment mark AM1 and the devices are covered by the junction layer WB1. The junction layer WB1 is a transparent or translucent insulating layer. The junction layer WB1 may be a silicon oxide film or a thin film of a material other than silicon oxide. In the former case, the junction layer WB1 may be a silicon oxide film made using TEOS (tetraethoxysilane). The same applies to the second substrate W2.

[0038] The front and back surfaces of the first substrate W1 are two planes parallel to each other. The back surface of the first substrate W1 is the plane opposite to the front surface of the first substrate W1. The bonding surface WA1 is the surface of the first substrate W1 that is in contact with the atmosphere in the space in which the first substrate W1 is placed. When the bonding surface WA1 of the first substrate W1 is facing upward, the bonding surface WA1 of the first substrate W1 corresponds to the upper surface of the first substrate W1, and the back surface of the first substrate W1 corresponds to the lower surface of the first substrate W1. The surface of the bonding layer WB1 corresponds to the bonding surface WA1 of the first substrate W1. The bonding surface WA1 may be a circular plane parallel to the surface of the base material WD1 that covers the entire surface of the base material WD1, or it may be a plane parallel to the surface of the base material WD1 that covers only a part of the surface of the base material WD1. The same applies to the second substrate W2.

[0039] Figures 4 and 5 show an example in which four first alignment marks AM1 of the same shape and size are provided on the first substrate W1, and four second alignment marks AM2 of the same shape and size are provided on the second substrate W2. Figure 4 shows an example in which the first alignment marks AM1 are cross-shaped. Figure 5 shows an example in which the second alignment marks AM2 include four L-shaped figures, one at each of the four corners of a rectangle, along any two of the four sides of the rectangle.

[0040] Alignment marks may consist of a single continuous shape, as shown in Figure 4, or they may consist of multiple shapes separated from each other, as shown in Figure 5. When multiple alignment marks are formed on a single substrate W, the alignment marks may be circles or rings.

[0041] The alignment marks are preferably shaped in a way that allows for the identification of mutually orthogonal height and width directions. The height and width directions of the alignment marks are parallel to the surface of the substrate W and are mutually orthogonal. If the alignment marks are cross-shaped, the direction parallel to the vertical lines of the alignment marks is the height direction, and the direction parallel to the horizontal lines of the alignment marks is the width direction. In Figure 4, the vertical direction of the paper is the height direction of the alignment marks, and the horizontal direction of the paper is the width direction of the alignment marks. As shown in Figure 5, if the alignment marks include four L-shaped figures placed at the four corners of a horizontally elongated rectangle, the direction parallel to the shorter side of the rectangle is the height direction of the alignment marks, and the direction parallel to the longer side of the rectangle is the width direction of the alignment marks.

[0042] The centroids AG1 of the four first alignment marks AM1 shown in Figure 4 are located at 90-degree intervals on a circle concentric with the first substrate W1. Similarly, the centroids AG2 of the four second alignment marks AM2 shown in Figure 5 are located at 90-degree intervals on a circle concentric with the second substrate W2. The vertical axis AV1, horizontal axis AH1, and reference point AP1 of the first substrate W1 are identified based on the four first alignment marks AM1 shown in Figure 4. The vertical axis AV2, horizontal axis AH2, and reference point AP2 of the second substrate W2 are identified based on the four second alignment marks AM2 shown in Figure 5.

[0043] The vertical axis AV1, horizontal axis AH1, and reference point AP1 of the first substrate W1 are all virtual lines or points set on the surface of the first substrate W1. The vertical axis AV1 and horizontal axis AH1 are reference lines indicating the angles of the first substrate W1. The vertical axis AV1 is a line passing through the centroids AG1 of two first alignment marks AM1 that differ in angle by 180 degrees around the center of the first substrate W1. The horizontal axis AH1 is a line passing through the centroids AG1 of the remaining two first alignment marks AM1. The vertical axis AV1 and horizontal axis AH1 are mutually orthogonal lines. The reference point AP1 is the intersection of the vertical axis AV1 and horizontal axis AH1. The center of the first substrate W1 is determined based on the contour of the first substrate W1. The reference point AP1 may or may not coincide with the center of the first substrate W1.

[0044] The vertical axis AV2, horizontal axis AH2, and reference point AP2 of the second substrate W2 are all virtual lines or points set on the surface of the second substrate W2. The vertical axis AV2 and horizontal axis AH2 are reference lines indicating the angles of the second substrate W2. The vertical axis AV2 is a line passing through the centroids AG2 of two second alignment marks AM2 that differ in angle by 180 degrees around the center of the second substrate W2. The horizontal axis AH2 is a line passing through the centroids AG2 of the remaining two second alignment marks AM2. The vertical axis AV2 and horizontal axis AH2 are mutually orthogonal lines. The reference point AP2 is the intersection of the vertical axis AV2 and horizontal axis AH2. The center of the second substrate W2 is determined based on the contour of the second substrate W2. The reference point AP2 may or may not coincide with the center of the second substrate W2.

[0045] When joining two substrates W, namely the first substrate W1 and the second substrate W2, alignment adjustment is performed to minimize the misalignment between the first substrate W1 and the second substrate W2, with the joining surface WA1 of the first substrate W1 and the joining surface WA2 of the second substrate W2 facing each other vertically. In the example shown in Figures 4 and 5, the relative positions of the first substrate W1 and the second substrate W2 are adjusted so that the distance between the reference point AP1 of the first substrate W1 and the reference point AP2 of the second substrate W2 decreases, the angle between the horizontal axis AH1 of the first substrate W1 and the horizontal axis AH2 of the second substrate W2 decreases, and the angle between the vertical axis AV1 of the first substrate W1 and the vertical axis AV2 of the second substrate W2 decreases.

[0046] Furthermore, minimizing the positional misalignment of the first substrate W1 and the second substrate W2 can be determined by comprehensively considering the distance between two reference points, reference point AP1 and reference point AP2, the angle between two vertical axes, vertical axis AV1 and vertical axis AV2, and the angle between two horizontal axes, horizontal axis AH1 and horizontal axis AH2. In other words, when the positional misalignment of the first substrate W1 and the second substrate W2 is minimized, it is not necessarily required that at least one of the two reference points, the distance between two reference points, reference point AP1 and reference point AP2, the angle between two vertical axes, vertical axis AV1 and vertical axis AV2, and horizontal axes, horizontal axis AH1 and horizontal axis AH2, is minimized.

[0047] The method for adjusting the alignment of the first substrate W1 and the second substrate W2 is not limited to the above. For example, if the same number of alignment marks (three or more) are provided on each of the first substrate W1 and the second substrate W2, the relative positions of the first substrate W1 and the second substrate W2 may be adjusted so as to minimize the sum or average value of the distances between the centroid AG1 of the first alignment mark AM1 on the first substrate W1 and the centroid AG2 of the second alignment mark AM2 on the second substrate W2, without using reference lines such as vertical axes AV1 and AV2, and horizontal axes AH1 and AH2. Alternatively, the relative positions of the first substrate W1 and the second substrate W2 may be adjusted so as to minimize the distance between the approximate curve or straight line obtained from the centroid AG1 of the multiple first alignment marks AM1 on the first substrate W1 using a calculation method such as the least squares method, and the approximate curve or straight line obtained from the centroid AG2 of the multiple second alignment marks AM2 on the second substrate W2 using the same calculation method.

[0048] Figure 6 is a schematic diagram showing the first alignment mark AM1 and the second alignment mark AM2 before and after alignment adjustment of the first substrate W1 and the second substrate W2. The left side of Figure 6 shows the first alignment mark AM1 and the second alignment mark AM2 before alignment adjustment of the first substrate W1 and the second substrate W2. In this example, the center of gravity AG1 of the first alignment mark AM1 is misaligned with the center of gravity AG2 of the second alignment mark AM2, and the first alignment mark AM1 is tilted relative to the second alignment mark AM2. As shown on the right side of Figure 6, after alignment adjustment, the center of gravity AG1 of the first alignment mark AM1 coincides with the center of gravity AG2 of the second alignment mark AM2, and the direction of the first alignment mark AM1 coincides with the direction of the second alignment mark AM2.

[0049] Next, we will explain the joining unit 2b.

[0050] Figure 7 is a perspective view of the joining unit 2b. Figure 8 is a plan view of the joining unit 2b. Figure 9 is a view of the joining unit 2b horizontally in the Y direction.

[0051] In the following example, the first chuck 11 is rotatable around a horizontal line. Unless otherwise specified, the first chuck 11 in the loading / unloading state will be described below. In the loading / unloading state of the first chuck 11, the center line 11c of the first chuck 11 (see Figure 9) is vertical and the first suction surface 11s of the first chuck 11 (see Figure 9) is facing upward.

[0052] The X and Y directions are horizontal directions that are perpendicular to each other. The Z direction is a vertical direction that is perpendicular to the X and Y directions. The X, Y, and Z directions correspond to the left-right, front-back, and up-down directions of the joining unit 2b, respectively. The positive sides of the X, Y, and Z directions correspond to the right, front, and up directions of the joining unit 2b, respectively.

[0053] As shown in Figures 7 to 9, the bonding unit 2b includes a first bonding module 2b1 including a first chuck 11 for holding a first substrate W1, a second bonding module 2b2 including a second chuck 21 for holding a second substrate W2, an actuator mechanism AC for moving at least one of the first chuck 11 and the second chuck 21, and a base 20 for supporting the first bonding module 2b1 and the second bonding module 2b2.

[0054] The first chuck 11 and the second chuck 21 are positioned above the base 20. The base 20 is positioned between the floor of the cleanroom where the substrate bonding apparatus 1 is installed and the first chuck 11 and the second chuck 21. The base 20 is a common component for the first bonding module 2b1 and the second bonding module 2b2. The base 20 may be a single, integrated component or may consist of multiple components fixed to each other.

[0055] As shown in Figure 9, the joining unit 2b may include a vibration isolation table 10 that reduces vibrations of the base 20. The vibration isolation table 10 may include a vibration sensor that detects vibrations of at least one of the cleanroom floor and the base 20, and a vibration isolation actuator that reduces vibrations of the base 20 by moving the base 20 based on the vibration sensor's detection value. The vibration isolation table 10 is positioned between the cleanroom floor and the base 20. The base 20 is supported by the vibration isolation table 10.

[0056] The bonding unit 2b may include a chamber 9 that houses the first chuck 11 and the second chuck 21. The chamber 9 includes a box-shaped partition wall 9p with an inlet / outlet through which the substrate W passes, and a shutter 9s for opening and closing the inlet / outlet. The partition wall 9p may have one inlet / outlet through which both the first substrate W1 and the second substrate W2 pass, or it may have an inlet / outlet through which the first substrate W1 passes and an inlet / outlet through which the second substrate W2 passes. In the latter case, a shutter 9s can be provided for each inlet / outlet. Equipment inside the chamber 9, such as the first chuck 11, is separated by the chamber 9 from processing units 2 other than the bonding unit 2b, such as the cleaning unit 2c (see Figure 1).

[0057] At least one of the first chuck 11 and the second chuck 21 is movable in the Y direction relative to the base 20. Similarly, at least one of the first chuck 11 and the second chuck 21 is movable in the X direction relative to the base 20. At least one of the first chuck 11 and the second chuck 21 is movable in the Z direction relative to the base 20. Figures 7 to 9 show an example in which the first chuck 11 is movable in the Z direction relative to the base 20, and the second chuck 21 is movable in the X and Y directions relative to the base 20.

[0058] As long as the first chuck 11 and the second chuck 21 are movable relative to each other in the X, Y, and Z directions, the directions in which the first chuck 11 and the second chuck 21 can move relative to the base 20 are not limited to those described above. For example, if one of the first chuck 11 and the second chuck 21 is movable relative to the base 20 in the X, Y, and Z directions, the other of the first chuck 11 and the second chuck 21 does not need to be movable relative to the base 20 in the X, Y, and Z directions.

[0059] The first chuck 11 and the second chuck 21 are vacuum chucks that hold the first substrate W1 with suction force generated by gas attraction, or electrostatic chucks that hold the first substrate W1 with electrical suction force. The first chuck 11 and the second chuck 21 may be other types of chucks as long as they can fix the first substrate W1 and the second substrate W2 to the first chuck 11 and the second chuck 21.

[0060] As shown in Figure 8, the first chuck 11 includes a first suction surface 11s that contacts the lower surface of the first substrate W1. Similarly, the second chuck 21 includes a second suction surface 21s that contacts the lower surface of the second substrate W2. The center line 11c of the first chuck 11 is a straight line perpendicular to the first suction surface 11s, passing through the center of the first suction surface 11s. The center line 21c of the second chuck 21 is a straight line perpendicular to the second suction surface 21s, passing through the center of the second suction surface 21s.

[0061] Figures 7 to 9 show an example where the first adsorption surface 11s and the second adsorption surface 21s are circular, horizontal planes with diameters greater than or equal to the diameters of the first substrate W1 and the second substrate W2. In this example, the first substrate W1 is placed on the first adsorption surface 11s such that the center of the first substrate W1 coincides with the center of the first adsorption surface 11s. Similarly, the second substrate W2 is placed on the second adsorption surface 21s such that the center of the second substrate W2 coincides with the center of the second adsorption surface 21s. As a result, the bottom surface of the first substrate W1 is in contact with the first adsorption surface 11s, and the bottom surface of the second substrate W2 is in contact with the second adsorption surface 21s.

[0062] The first suction surface 11s may have a shape other than a circle, or a shape other than a horizontal plane. The same applies to the second suction surface 21s. For example, when the first chuck 11 is in the loading / unloading state, the first suction surface 11s may be a spherical curved surface that is convex upward or a circular curved surface that is curved in an arc shape, or it may include a horizontal part and a non-horizontal part. The same applies to the second suction surface 21s. The non-horizontal part may be a plane inclined at a constant angle with respect to the horizontal plane, or a curved surface in which the angle of inclination with respect to the horizontal plane changes continuously, or it may be something other than these. One of the first suction surface 11s and the second suction surface 21s may be a horizontal plane, and the other of the first suction surface 11s and the second suction surface 21s may have a shape other than a horizontal plane.

[0063] When the first substrate W1 and the second substrate W2 are loaded into the bonding unit 2b, they are flat. Even if the first substrate W1 and the second substrate W2 are warped, the amount of warping is small enough that the first substrate W1 and the second substrate W2 can be considered substantially flat. If the first suction surface 11s has a shape other than a horizontal plane, when the first substrate W1 is placed on the first chuck 11, the first suction surface 11s partially contacts the lower surface of the first substrate W1, and a gap is formed between the lower surface of the first substrate W1 and the first suction surface 11s of the first chuck 11. When the first chuck 11 picks up the first substrate W1, the upper and lower surfaces of the first substrate W1 elastically deform to the same or nearly the same shape as the first suction surface 11s, and the contact area between the lower surface of the first substrate W1 and the first suction surface 11s of the first chuck 11 increases. The same applies when the second suction surface 21s has a shape other than a horizontal plane.

[0064] The first adsorption surface 11s may include a plurality of adsorption regions that independently generate and stop the adsorption force that holds the first substrate W1 on the first adsorption surface 11s. The same applies to the second adsorption surface 21s. An adsorption region is a region that generates an attractive force regardless of whether other adsorption regions are generating an adsorption force or not. The plurality of adsorption regions of the first adsorption surface 11s may extend radially from the center of the first adsorption surface 11s, as shown in Figure 10A, or they may extend along a plurality of parallel lines crossing the first adsorption surface 11s, as shown in Figure 10B. The arrangement and shape of the plurality of adsorption regions of the first adsorption surface 11s may be other than those described above. The same applies to the plurality of adsorption regions of the second adsorption surface 21s.

[0065] Figures 10A and 10B are schematic plan views of the first chuck 11. Figures 10A and 10B show an example where the first chuck 11 is a vacuum chuck and the first adsorption surface 11s includes multiple adsorption regions. Multiple suction ports 11p for drawing in gas are open on the first adsorption surface 11s. In the example shown in Figure 10A, the multiple suction ports 11p are arranged on multiple straight lines extending radially from the center of the first adsorption surface 11s. In the example shown in Figure 10B, the multiple suction ports 11p are arranged on multiple parallel straight lines crossing the first adsorption surface 11s. Figure 10B shows an example where the multiple suction ports 11p are arranged at equal intervals in the longitudinal and transverse directions.

[0066] If the first chuck 11 or the second chuck 21 is a vacuum chuck, one or more suction ports 11p and a valve 11v should be provided for each suction area. If the first chuck 11 or the second chuck 21 is an electrostatic chuck, multiple electrodes should be provided that correspond one-to-one with multiple suction areas. When the valve 11v is opened, one or more suction ports 11p that have opened in the suction area corresponding to this valve 11v begin to draw in gas. When the valve 11v is closed, one or more suction ports 11p that have opened in the suction area corresponding to this valve 11v stop drawing in gas. When voltage is applied to an electrode, the suction area corresponding to this electrode begins to generate an electrical attractive force. When the application of voltage to an electrode is stopped, the suction area corresponding to this electrode stops generating an electrical attractive force. As a result, multiple suction areas generate and stop suction forces independently of each other.

[0067] Next, we will describe the first junction module 2b1.

[0068] Figure 11 is a horizontal view of the first joining module 2b1 in the X direction. Figure 12 is a plan view of the first joining module 2b1. As shown in Figures 11 and 12, the first joining module 2b1 includes a first stage 13 that is movable in the Z direction relative to the base 20 together with the first chuck 11. The first joining module 2b1 further includes a pair of support shafts 14, a pair of support bases 15, and a first frame 19 that support the first chuck 11 via the first stage 13. The first stage 13 may be a horizontal plate.

[0069] The first frame 19 includes a pair of first side frames 19s extending upward from the upper surface of the base 20, and a first upper frame 19u extending from one first side frame 19s to the other first side frame 19s, separated upward from the base 20. Figure 11 shows an example in which the first frame 19 is gantry-shaped. In this example, the first upper frame 19u connects the upper ends of the pair of first side frames 19s.

[0070] A pair of first side frames 19s are fixed to the base 20. A first upper frame 19u is fixed to the pair of first side frames 19s. The pair of first side frames 19s are spaced apart and face each other in the Y direction. The first upper frame 19u is spaced apart and face the top surface of the base 20 in the Z direction. The pair of first side frames 19s are positioned on both sides of the first chuck 11, first stage 13, pair of support shafts 14, and pair of support bases 15 in the Y direction. The first upper frame 19u is positioned above the first chuck 11, first stage 13, pair of support shafts 14, and pair of support bases 15 so as to overlap them in a plan view.

[0071] The first chuck 11, the first stage 13, a pair of support shafts 14, and a pair of support bases 15 are supported on the first frame 19, separated above the base 20. The pair of support bases 15 are positioned between a pair of first side frames 19s in the Y direction. The pair of support shafts 14 are positioned between the pair of support bases 15 in the Y direction. The first stage 13 is positioned between the pair of support shafts 14 in the Y direction. The pair of support bases 15 are supported on the pair of first side frames 19s. The pair of support shafts 14 are supported on the pair of support bases 15 in a horizontal position with the centerlines of the pair of support shafts 14 parallel to the Y direction. The first stage 13 is supported on the pair of support shafts 14 with the centerline 11c of the first chuck 11 in a vertical position.

[0072] The first joint module 2b1 includes a plurality of linear guides 17 that linearly guide a pair of support bases 15 in the Z direction relative to the first frame 19. At least one linear guide 17 is provided for each support base 15. Figure 12 shows an example where two linear guides 17 are provided for each support base 15. The linear guide 17 includes a rail 17L that extends linearly in the longitudinal direction, a slide block 17b that moves along the rail 17L in the longitudinal direction of the rail 17L, and a plurality of balls interposed between the rail 17L and the slide block 17b. The rail 17L is fixed to the first side frame 19s in a position parallel to the Z direction. The slide block 17b is fixed to the support base 15.

[0073] The actuator mechanism AC includes at least one Z actuator 18 that moves the first chuck 11 and the second chuck 21 relative to each other in the Z direction. Figure 11 shows an example in which the first substrate W1 held by the first chuck 11 is moved in the Z direction by at least one Z actuator 18 moving a pair of support bases 15 in the Z direction. The Z actuator 18 may be provided on only one of the support bases 15, or at least one may be provided on each support base 15. Figure 11 shows an example in which one Z actuator 18 is provided on each support base 15.

[0074] An actuator is a device that converts electrical, fluid, magnetic, thermal, or chemical energy into mechanical work. Actuators include electric motors (rotary motors), linear motors, air cylinders, and other devices. The Z actuator 18 may be an electric motor, a linear motor, or something else. Figures 11 and 12 show an example where the Z actuator 18 is an electric motor.

[0075] Power from the Z actuator 18 is transmitted to the first chuck 11 via a transmission mechanism including a drive element and a driven element. Figures 11 and 12 show an example where the drive element is a ball screw 18s and the driven element is a ball nut 18n. The drive element and driven element may be other components besides the ball screw 18s and ball nut 18n, such as a rack and pinion.

[0076] As shown in Figure 12, the ball screws 18s and ball nuts 18n are positioned between the first side frame 19s and the support base 15. The ball nuts 18n are fixed to the support base 15. The ball nuts 18n are connected to the ball screws 18s via multiple balls. The two ball screws 18s are positioned at two locations, each at an angle of 180 degrees different around the centerline 11c of the first chuck 11. The centerlines 18c of the two ball screws 18s extend vertically within a single plane that spans the Y and Z directions.

[0077] When an electric motor, which is an example of a Z actuator 18, rotates a ball screw 18s, the ball nut 18n moves along the ball screw 18s in the Z direction. The ball nut 18n and the support base 15 move in the Z direction by an amount of movement corresponding to the rotation angle of the Z actuator 18. When two Z actuators 18 rotate in the same direction at the same angle, the two ball nuts 18n move by the same amount in the Z direction. As a result, the first substrate W1 held in the first chuck 11 rises or falls in the Z direction.

[0078] The actuator mechanism AC includes a first rotation motor 12 that rotates the first chuck 11 relative to the first stage 13 around the center line 11c of the first chuck 11, which corresponds to the axis of rotation. The first rotation motor 12 is an electric motor. The first rotation motor 12 is positioned between the first chuck 11 and the first stage 13. The first rotation motor 12 is connected to the first chuck 11 and the first stage 13. The first substrate W1 is held in the first chuck 11 such that the center of the first substrate W1 lies on the center line 11c of the first chuck 11. When the first rotation motor 12 rotates in this state, the first substrate W1 and the first chuck 11 rotate relative to the first stage 13 in the same direction and at the same angle as the first rotation motor 12.

[0079] The actuator mechanism AC includes at least one inversion actuator 16 that rotates the first chuck 11 around a horizontal inversion center 16c. Figures 11 and 12 show an example in which the inversion actuator 16 is located inside the support shaft 14. In this example, the inversion actuator 16 is located above the base 20 so as to overlap the base 20 in a plan view. The inversion actuator 16 may be located outside the support shaft 14. There may be one inversion actuator 16 on only one support shaft 14, or there may be one in each support shaft 14. Figures 11 and 12 show an example of the latter.

[0080] The reversing actuator 16 is an electric motor. The reversing actuator 16 may be an actuator other than an electric motor. In the example shown in Figures 11 and 12, the reversing actuator 16 rotates the first chuck 11 by rotating the first stage 13 around the centerlines of a pair of support shafts 14. The centerlines of the pair of support shafts 14 correspond to the reversal center 16c. Therefore, the first chuck 11 and the first stage 13 rotate around a horizontal reversal center 16c extending in the Y direction. The reversal center 16c is a horizontal straight line passing through the centerlines 18c of the two ball screws 18s (see Figure 12). When the reversing actuator 16 rotates 180 degrees, the first chuck 11 and the first stage 13 rotate 180 degrees and face downwards. This reverses the first substrate W1 held in the first chuck 11.

[0081] The first stage 13 may be fixed to a pair of support shafts 14, or it may be rotatable relative to the pair of support shafts 14 around the reversal center 16c. In the former case, the reversal actuator 16 rotates the first stage 13 and the first chuck 11 around the reversal center 16c by rotating the pair of support shafts 14 relative to a pair of support bases 15. In the latter case, the reversal actuator 16 rotates the first chuck 11 around the reversal center 16c by rotating the first stage 13 relative to the pair of support shafts 14.

[0082] The first chuck 11 and the first stage 13 are movable in the Z direction relative to the base 20 and the first frame 19. The range of movement of the first chuck 11 and the first stage 13 in the Z direction includes a position where the rotating object, such as the first stage 13, does not collide with the base 20 and the first frame 19 even if the first chuck 11 and the first stage 13 rotate 180 degrees around the inversion center 16c. The size of the first stage 13 and the first frame 19 may be set so that the rotating object does not collide with the base 20 and the first frame 19 regardless of the position in the Z direction to which the first chuck 11 and the first stage 13 rotate.

[0083] The inversion actuator 16 is part of the inversion unit. The inversion actuator 16 inverts the first substrate W1 by rotating the first chuck 11 180 degrees around the inversion center 16c while the first substrate W1 is held by the first chuck 11. As will be described later, the first chuck 11 holds the first substrate W1 when it is joined to the second substrate W2. Therefore, the first substrate W1 is held by the first chuck 11 before inversion and is also held by the first chuck 11 when it is joined to the second substrate W2.

[0084] The inverting actuator 16 is an example of a posture-changing actuator that changes the orientation of the substrate W. The two Z actuators 18 mentioned above are also examples of posture-changing actuators. When the two Z actuators 18 position the two ball nuts 18n at different heights relative to each other within the range of mechanical play of the mechanism supporting the first chuck 11, the centerline 11c of the first chuck 11 tilts in the Y direction. This changes the orientation of the first substrate W1 held by the first chuck 11.

[0085] Next, I will explain the first fixed criterion 31r.

[0086] Figure 13 is a plan view of the first stage 13. Figure 14 is a horizontal view of the first stage 13 in the direction of arrow XIV shown in Figure 13. As shown in Figures 13 and 14, the first bonding module 2b1 includes a first fixed reference 31r which serves as an indirect reference when aligning the first substrate W1. The first fixed reference 31r includes at least one first reference mark 31m which serves as an indirect reference when aligning the first substrate W1, and a first display board 31d which is marked with at least one first reference mark 31m.

[0087] Figure 13 shows an example in which two first reference marks 31m of the same shape and size are provided. In this example, the first reference marks 31m are cross-shaped. The first reference marks 31m may have a shape other than a cross. The first reference marks 31m may be formed on the first display panel 31d by photolithography, or by other methods such as laser marking.

[0088] In the example shown in Figure 13, the first display panel 31d is a horizontal rectangular flat plate extending in the Y direction, with two first reference marks 31m marked on each end of the first display panel 31d in the Y direction. The two long sides of the first display panel 31d extend in the Y direction, and the two short sides of the first display panel 31d extend in the X direction. As shown in Figure 14, in this example, the first display panel 31d is supported on the first stage 13 via two first support blocks 31b. The top and bottom surfaces of the first display panel 31d are two horizontal planes parallel to each other. The bottom surface of the first display panel 31d faces the horizontal, flat top surface of the first stage 13 with a gap in the Z direction. The bottom surface of the first display panel 31d may be in contact with the top surface of the first stage 13.

[0089] The shape and arrangement of the first display board 31d, and the number and arrangement of the first reference marks 31m are not limited to the examples shown in Figures 13 and 14. For example, the first display board 31d may be rotated 90 degrees in the horizontal plane so that its long side extends in the X direction. The first display board 31d may include a horizontal rectangular vertical plate extending in the Y direction and a horizontal rectangular horizontal plate extending in the X direction. The vertical plate and the horizontal plate may be touching each other or separated from each other. If the first display board 31d includes a vertical plate and a horizontal plate, these first reference marks 31m may be arranged so that three or more first reference marks 31m are aligned in the X and Y directions. For example, the second first reference mark 31m may be placed on a horizontal line parallel to the X direction passing through the first first reference mark 31m, and the third first reference mark 31m may be placed on a horizontal line parallel to the Y direction passing through the first first reference mark 31m.

[0090] At least one first reference mark 31m is used to set a first fixed point 31p and a first fixed line 31L, which serve as an indirect reference when aligning the first substrate W1. The first fixed point 31p and the first fixed line 31L are virtual points and lines. In the example shown in Figure 13, a horizontal line passing through the centroids of two cross-shaped first reference marks 31m is set as the first fixed line 31L of the first stage 13. The first fixed point 31p of the first stage 13 may be the centroid of one of the first reference marks 31m or any point on the first fixed line 31L, or it may be any other point.

[0091] As shown in Figure 13, the first fixed reference 31r is positioned above the first stage 13 such that, in a plan view, the first fixed reference 31r overlaps with the first stage 13. The first fixed reference 31r is fixed to the first stage 13. The upper end of the first fixed reference 31r is positioned below the first substrate W1 on the first chuck 11. When the first substrate W1 is held by the first chuck 11, the first fixed reference 31r does not overlap with the first substrate W1 held by the first chuck 11 when viewed in the direction of the center line 11c of the first chuck 11. Therefore, when viewing the first stage 13 in the direction of the center line 11c of the first chuck 11 in this state, not only the first reference mark 31m but also the first alignment mark AM1 on the first substrate W1 can be photographed with the camera.

[0092] As described later, at least one first alignment mark AM1 on the first substrate W1 held in the first chuck 11 is photographed by a camera. This camera also photographs at least one first reference mark 31m on the first fixed reference 31r. The position and angle of the first substrate W1 in the horizontal plane are determined based on the image of the at least one first alignment mark AM1 captured by the camera. The position and angle of the first fixed reference 31r in the horizontal plane are also determined based on the image of the at least one first reference mark 31m captured by the camera. Furthermore, the relative position and angle between the first substrate W1 and the first fixed reference 31r are also determined based on the images of the first alignment mark AM1 and the first reference mark 31m. Therefore, if the position and angle of the first fixed reference 31r are known, the position and angle of the first substrate W1 held in the first chuck 11 can also be determined. This information is stored in the control device 3 (see Figure 1).

[0093] Next, we will explain the second junction module 2b2.

[0094] Figure 15 is a horizontal view of the second joint module 2b2 in the X direction. Figure 16 is an enlarged view of a part of Figure 15. Figure 17 is a plan view of the second joint module 2b2. Figure 18 is a horizontal view of the second joint module 2b2 in the Y direction.

[0095] As shown in Figures 15 to 17, the second joining module 2b2 includes a second stage 23 that is movable in the X and Y directions relative to the base 20 together with the second chuck 21. The second stage 23 includes an upper stage 23u positioned between the second chuck 21 and the base 20, and a lower stage 23L positioned between the upper stage 23u and the base 20. Figures 15 to 17 show an example where the upper stage 23u is a Y-stage movable in the Y direction relative to the base 20, and the lower stage 23L is an X-stage movable in the X direction relative to the base 20. The lower stage 23L is larger than the upper stage 23u. One or both of the upper stage 23u and the lower stage 23L may be horizontal flat plates.

[0096] The second joining module 2b2 includes at least one linear guide 26 that linearly guides the lower stage 23L in the X direction relative to the base 20. Figure 15 shows an example in which two linear guides 26 are provided. The two linear guides 26 are positioned on the base 20. The lower stage 23L is positioned on the two linear guides 26. As shown in Figure 16, the linear guide 26 includes a rail 26L that extends linearly in the longitudinal direction, a slide block 26b that moves along the rail 26L in the longitudinal direction of the rail 26L, and a plurality of balls interposed between the rail 26L and the slide block 26b. The rail 26L is fixed to the base 20 in a position parallel to the X direction. The slide block 26b is fixed to the lower stage 23L.

[0097] The actuator mechanism AC includes at least one X actuator 27 that moves the first chuck 11 and the second chuck 21 relative to each other in the X direction. Figure 15 shows an example in which at least one X actuator 27 moves the lower stage 23L in the X direction, thereby moving the second substrate W2 held in the second chuck 21 in the X direction. In this example, two X actuators 27 are provided, each X actuator 27 being a linear motor. As shown in Figure 16, the linear motor includes a fixed magnet 27f that extends linearly in the longitudinal direction and a movable magnet 27m that moves along the fixed magnet 27f in the longitudinal direction of the fixed magnet 27f. The fixed magnet 27f is fixed to the base 20 in a position parallel to the X direction. The movable magnet 27m is fixed to the lower stage 23L.

[0098] As shown in Figures 17 and 18, the second joining module 2b2 includes at least one linear guide 24 that guides the upper stage 23u linearly in the Y direction relative to the lower stage 23L. Figure 17 shows an example in which two linear guides 24 are provided. The two linear guides 24 are positioned on the lower stage 23L. The upper stage 23u is positioned on the two linear guides 24. The linear guide 24 includes a rail 24L that extends linearly in the longitudinal direction, a slide block 24b that moves along the rail 24L in the longitudinal direction of the rail 24L, and a plurality of balls interposed between the rail 24L and the slide block 24b. The rail 24L is fixed to the lower stage 23L in a position parallel to the Y direction. The slide block 24b is fixed to the upper stage 23u.

[0099] The actuator mechanism AC includes at least one Y actuator 25 that moves the first chuck 11 and the second chuck 21 relative to each other in the Y direction. Figure 18 shows an example in which at least one Y actuator 25 moves the upper stage 23u in the Y direction, thereby moving the second substrate W2 held in the second chuck 21 in the Y direction. In this example, two Y actuators 25 are provided, each Y actuator 25 being a linear motor. The linear motor includes a fixed magnet 25f that extends linearly in the longitudinal direction and a movable magnet 25m that moves along the fixed magnet 25f in the longitudinal direction of the fixed magnet 25f. The fixed magnet 25f is fixed to the lower stage 23L in a position parallel to the Y direction. The movable magnet 25m is fixed to the upper stage 23u.

[0100] X actuator 27, Y actuator 25, and Z actuator 18 are examples of translation actuators that move at least one of the first chuck 11 and the second chuck 21 in a parallel direction. Y actuator 25 is an example of a horizontal actuator that moves at least one of the first chuck 11 and the second chuck 21 horizontally. X actuator 27 is also an example of a horizontal actuator. Z actuator 18 is an example of a vertical actuator that moves at least one of the first chuck 11 and the second chuck 21 vertically.

[0101] The X actuator 27 moves the second chuck 21 in the X direction relative to the first chuck 11 between the joining position (the position shown in Figure 31E) and the imaging position (the position shown in Figure 31G). The joining position is the position where the first substrate W1 held by the first chuck 11 and the second substrate W2 held by the second chuck 21 are joined. The imaging position is the position where the camera images the first substrate W1 and the second substrate W2 held by the second chuck 21.

[0102] The imaging position may be a range that extends in at least one of the X and Y directions. For example, the imaging position may include a pre-bonding imaging position in which the first narrow-field camera 41n and the first wide-field camera 41w, described later, are used to image the second substrate W2, etc., held in the second chuck 21, and a post-bonding imaging position in which the inspection camera 43, described later, is used to image the first substrate W1 and the second substrate W2, held in the second chuck 21, after the first substrate W1 and the second substrate W2 have been bonded.

[0103] The actuator mechanism AC includes a second rotation motor 22 that rotates the second chuck 21 relative to the upper stage 23u around the center line 21c of the second chuck 21, which corresponds to the axis of rotation. The second rotation motor 22 is an electric motor. The second rotation motor 22 is positioned between the second chuck 21 and the upper stage 23u. The second rotation motor 22 is connected to the second chuck 21 and the upper stage 23u. The second substrate W2 is held in the second chuck 21 such that the center of the second substrate W2 lies on the center line 21c of the second chuck 21. When the second rotation motor 22 rotates in this state, the second substrate W2 and the second chuck 21 rotate relative to the upper stage 23u in the same direction and at the same angle as the second rotation motor 22.

[0104] As shown in Figure 15, the second joining module 2b2 includes a second frame 28 that forms a space through which the second chuck 21 and the second stage 23 can move in and out in the X direction. The second frame 28 includes a pair of second side frames 28s extending upward from the top surface of the base 20 and a second upper frame 28u extending from one second side frame 28s to the other second side frame 28s, separated upward from the base 20. Figure 15 shows an example where the second frame 28 is gantry-shaped. In this example, the second upper frame 28u connects the upper ends of the pair of second side frames 28s.

[0105] A pair of second side frames 28s are fixed to the base 20. The second upper frame 28u is fixed to the pair of second side frames 28s. The pair of second side frames 28s are spaced apart and face each other in the Y direction. The second upper frame 28u is spaced apart and face the top surface of the base 20 in the Z direction. The pair of second side frames 28s are positioned on both sides of the second chuck 21 and the second stage 23 in the Y direction. The second upper frame 28u is positioned above the second chuck 21 and the second stage 23 so as to overlap them in a plan view.

[0106] The second frame 28 is separated from the first frame 19 in the X direction (see Figure 7). When the X actuator 27 moves the lower stage 23L in the X direction, the second chuck 21 and the second stage 23 move in and out of the space inside the second frame 28 in the X direction. Similarly, the second chuck 21 and the second stage 23 move in and out of the space inside the first frame 19 in the X direction. The second chuck 21 and the second stage 23 can enter between the first stage 13 and the base 20. When the first chuck 11 is facing downwards, the second chuck 21 and the second stage 23 can enter between the first chuck 11 and the base 20.

[0107] With the second chuck 21 and the second stage 23 positioned between a pair of second side frames 28s, even if the Y actuator 25 moves the upper stage 23u in the Y direction relative to the lower stage 23L, the second chuck 21 and the upper stage 23u will not collide with the second frame 28. Similarly, with the second chuck 21 and the second stage 23 positioned between a pair of first side frames 19s, even if the Y actuator 25 moves the upper stage 23u in the Y direction relative to the lower stage 23L, the second chuck 21 and the upper stage 23u will not collide with the first frame 19.

[0108] The bonding unit 2b may be equipped with a plurality of distance meters for measuring the vertical distance from the bonding surface WA1 of the downward-facing first substrate W1 held by the first chuck 11. The distance meters are non-contact distance meters that measure the distance to an object such as the first substrate W1 without contact. The distance meters may be optical distance meters such as laser distance meters, or other types of distance meters.

[0109] Multiple distance meters are positioned below the downward-facing first substrate W1 held by the first chuck 11. Multiple distance meters may be fixed to the upper stage 23u. In this case, multiple distance meters move horizontally in the same direction, speed, and amount of movement as the upper stage 23u. Multiple distance meters measure vertical distances to multiple positions within the bonding surface WA1 of the downward-facing first substrate W1. This makes it possible to measure the flatness of the bonding surface WA1 of the first substrate W1.

[0110] If the bonding surface WA1 of the first substrate W1 is horizontal before inversion, then when the inversion actuator 16 rotates exactly 180 degrees, the bonding surface WA1 of the downward-facing first substrate W1 will be positioned horizontally. This can be confirmed by measuring the distance to the bonding surface WA1 of the downward-facing first substrate W1 from multiple distance meters. If the bonding surface WA1 of the first substrate W1 is not horizontal, the inversion actuator 16 should be rotated until the bonding surface WA1 of the first substrate W1 is positioned horizontally. In this way, the parallelism between the first substrate W1 and the second substrate W2 can be increased.

[0111] Next, I will explain the second standard mark of 32m.

[0112] Figure 19 is a plan view of the second stage 23. Figure 20 is a horizontal view of the second stage 23 in the direction of arrow XX shown in Figure 19. As shown in Figures 19 and 20, the second bonding module 2b2 includes a second fixed reference 32r which serves as an indirect reference when aligning the second substrate W2. The second fixed reference 32r includes at least one second reference mark 32m which serves as an indirect reference when aligning the second substrate W2, and a second display plate 32d on which at least one second reference mark 32m is marked.

[0113] Figure 19 shows an example in which two second reference marks 32m of the same shape and size are provided. In this example, the second reference marks 32m are rectangular (including four L-shaped figures). The second reference marks 32m may be of a shape other than a rectangle. The second reference marks 32m may be formed on the second display panel 32d by photolithography, or by other methods such as laser marking.

[0114] In the example shown in Figure 19, the second display panel 32d is a horizontal rectangular flat plate extending in the Y direction, and two second reference marks 32m of equal shape and size are marked on each end of the second display panel 32d in the Y direction. The two long sides of the second display panel 32d extend in the Y direction, and the two short sides of the second display panel 32d extend in the X direction. As shown in Figure 20, in this example, the second display panel 32d is supported on the upper stage 23u via two second support blocks 32b. The top and bottom surfaces of the second display panel 32d are two horizontal planes parallel to each other. The bottom surface of the second display panel 32d faces the horizontal, flat top surface of the upper stage 23u with a gap in the Z direction. The bottom surface of the second display panel 32d may be in contact with the top surface of the upper stage 23u.

[0115] At least one second reference mark 32m is used to set a second fixed point 32p and a second fixed line 32L, which serve as an indirect reference when aligning the second substrate W2. The second fixed point 32p and the second fixed line 32L are virtual points and lines. In the example shown in Figure 19, a horizontal line passing through the centroids of two rectangular second reference marks 32m is set as the second fixed line 32L of the upper stage 23u. The second fixed point 32p of the upper stage 23u may be the centroid of one of the second reference marks 32m, any point on the second fixed line 32L, or other locations.

[0116] As shown in Figure 19, the second fixed reference 32r is positioned above the upper stage 23u such that, in a plan view, the second fixed reference 32r overlaps with the upper stage 23u. The second fixed reference 32r is fixed to the upper stage 23u. The upper end of the second fixed reference 32r is positioned below the second substrate W2 on the second chuck 21. When the second substrate W2 is held in the second chuck 21, the second fixed reference 32r does not overlap with the second substrate W2 held in the second chuck 21 when viewed in the direction of the center line 21c of the second chuck 21. Therefore, when viewing the upper stage 23u in the direction of the center line 21c of the second chuck 21 in this state, not only the second reference mark 32m but also the second alignment mark AM2 of the second substrate W2 can be photographed with the camera.

[0117] As described later, at least one second alignment mark AM2 on the second substrate W2 held in the second chuck 21 is photographed by a camera. This camera also photographs at least one second reference mark 32m on the second fixed reference 32r. The position and angle of the second substrate W2 in the horizontal plane are determined based on the image of the at least one second alignment mark AM2 captured by the camera. The position and angle of the second fixed reference 32r in the horizontal plane are also determined based on the image of the at least one second reference mark 32m captured by the camera. Furthermore, the relative position and angle between the second substrate W2 and the second fixed reference 32r are also determined based on the images of the second alignment mark AM2 and the second reference mark 32m. Therefore, if the position and angle of the second fixed reference 32r are known, the position and angle of the second substrate W2 held in the second chuck 21 can also be determined. This information is stored in the control device 3 (see Figure 1).

[0118] Next, we will describe the camera provided in the joining unit 2b.

[0119] Figure 21 is a plan view of the bonding unit 2b. Figure 22 is a view of the second camera 42, which is photographing the second substrate W2, horizontally in the Y direction. Figure 23 is a view of the first camera 41, which is photographing the first substrate W1, horizontally in the Y direction. Figure 24 is a view of the alignment camera 45, which is photographing the first fixed reference 31r and the second fixed reference 32r, horizontally in the Y direction. Figure 25 is a view of the alignment camera 45 horizontally in the direction of arrow XXV shown in Figure 24.

[0120] As shown in Figures 21 and 22, the bonding unit 2b includes at least one first camera 41 for photographing the first substrate W1 and the first fixed reference 31r held in the first chuck 11, and at least one second camera 42 for photographing the second substrate W2 and the second fixed reference 32r held in the second chuck 21. The bonding unit 2b further includes an inspection camera 43 for photographing the first substrate W1 and the second substrate W2 held in the second chuck 21 after bonding. Figures 21 and 22 show an example in which there are two first cameras 41 and two second cameras 42. The two first cameras 41 are a first narrow-field camera 41n and a first wide-field camera 41w. The two second cameras 42 are a second narrow-field camera 42n and a second wide-field camera 42w.

[0121] The first narrow-field camera 41n and the first wide-field camera 41w are fixed to the upper stage 23u of the second stage 23. The first narrow-field camera 41n and the first wide-field camera 41w move in the X and Y directions with the upper stage 23u at the same speed and distance. The positions of the first narrow-field camera 41n and the first wide-field camera 41w in the X and Y directions relative to the first stage 13 are adjusted by moving the second stage 23 using at least one of the X actuator 27 and the Y actuator 25.

[0122] The second narrow-field camera 42n, the second wide-field camera 42w, and the inspection camera 43 are fixed to the second frame 28 via a bracket 44. The second narrow-field camera 42n, the second wide-field camera 42w, and the inspection camera 43 are positioned above the base 20 so as to overlap the base 20 in a plan view. The inspection camera 43 is part of the inspection unit. The second substrate W2 is held in the second chuck 21 before it is bonded to the first substrate W1, and is also held in the second chuck 21 when the bonded first substrate W1 and second substrate W2 are photographed by the inspection camera 43.

[0123] Figure 21 shows an example in which the second narrow-field camera 42n, the second wide-field camera 42w, and the inspection camera 43 are inserted into three through-holes 44h that penetrate the bracket 44 in the Z direction and are positioned on the first chuck 11 side (right side in Figure 21) relative to the second frame 28. The positions of the second narrow-field camera 42n, the second wide-field camera 42w, and the inspection camera 43 in the X and Y directions relative to the second stage 23 are adjusted by moving the second stage 23 with at least one of the X actuator 27 and the Y actuator 25.

[0124] A narrow-field camera is a camera with a narrower field of view than a wide-field camera. The shooting range of the first narrow-field camera 41n may be part of the shooting range of the first wide-field camera 41w, or it may be outside of that shooting range. The same applies to the second narrow-field camera 42n and the second wide-field camera 42w. The resolution of the narrow-field camera may be equal to or different from the resolution of the wide-field camera. The narrow-field camera and the wide-field camera may be equipped with lenses of different magnifications and may be two cameras with the same specifications other than the shooting magnification, such as resolution.

[0125] The control device 3 (see Figure 1) receives electronic data of still or moving images generated by the first wide-field camera 41w. The control device 3 analyzes and makes judgments about the images based on a program stored in memory 3c. If the control device 3 determines, as a result of the analysis, that the first alignment mark AM1 is included in the image, it may store the position of the first alignment mark AM1. In this way, the shooting range of the first narrow-field camera 41n can be quickly moved to the range where the first alignment mark AM1 is located. Even if the first narrow-field camera 41n loses sight of the first alignment mark AM1, that is, if the shooting range of the first narrow-field camera 41n moves to a range where the first alignment mark AM1 is not located, the shooting range of the first narrow-field camera 41n can be quickly returned to the range where the first alignment mark AM1 is located.

[0126] Similarly, the control device 3 may store the position of the first reference mark 31m if it determines that the image generated by the first wide-field camera 41w includes the first reference mark 31m. Similarly, the control device 3 may store the position of at least one of the second alignment mark AM2 and the second reference mark 32m if it determines that the image generated by the second wide-field camera 42w includes at least one of the second alignment mark AM2 and the second reference mark 32m. In this way, the acquisition time of the first narrow-field camera 41n and the second narrow-field camera 42n can be shortened.

[0127] The lenses of the first narrow-field camera 41n and the first wide-field camera 41w are pointed upwards. When the first narrow-field camera 41n and the first wide-field camera 41w are taking pictures, the first stage 13 rotates 180 degrees around the inversion center 16c while the first substrate W1 is held in the first chuck 11, as shown in Figure 23. The first narrow-field camera 41n and the first wide-field camera 41w take pictures of the downward-facing first substrate W1 and the downward-facing first fixed reference 31r. At least one of the X actuator 27 and the Y actuator 25 moves the first narrow-field camera 41n and the first wide-field camera 41w in at least one of the X and Y directions while the first narrow-field camera 41n and the first wide-field camera 41w are taking pictures, so that the first narrow-field camera 41n and the first wide-field camera 41w take pictures of the entire range to be photographed. This captures the first alignment mark AM1 on the first substrate W1 and the first reference mark 31m on the first fixed reference 31r.

[0128] The lenses of the second narrow-field camera 42n and the second wide-field camera 42w are pointed downwards. When the second narrow-field camera 42n and the second wide-field camera 42w are taking pictures, as shown in Figure 22, the second substrate W2 is held in the second chuck 21, and at least one of the second substrate W2 and the second fixed reference 32r is positioned within the range that the second narrow-field camera 42n and the second wide-field camera 42w can photograph. In this state, the second narrow-field camera 42n and the second wide-field camera 42w photograph the upward-facing second substrate W2 and the upward-facing second fixed reference 32r. While the second narrow-field camera 42n and the second wide-field camera 42w are taking pictures, at least one of the X actuator 27 and the Y actuator 25 moves the second substrate W2 and the second fixed reference 32r in at least one of the X and Y directions. As a result, the second narrow-field camera 42n and the second wide-field camera 42w can capture the entire area to be photographed, and the second alignment mark AM2 on the second substrate W2 and the second reference mark 32m on the second fixed reference 32r are captured.

[0129] Similar to the second narrow-field camera 42n and the second wide-field camera 42w, the lens of the inspection camera 43 is pointed downwards. When the inspection camera 43 takes a picture, the first substrate W1 and the second substrate W2 are held in the second chuck 21, and the first substrate W1 and the second substrate W2 are positioned within the range that the inspection camera 43 can photograph. In this state, the inspection camera 43 photographs the downward-facing first substrate W1 and the upward-facing second substrate W2. At least one of the X actuator 27 and the Y actuator 25 moves the first substrate W1 and the second substrate W2 in at least one of the X and Y directions while the inspection camera 43 is taking a picture. This ensures that the entire range to be photographed is captured by the inspection camera 43.

[0130] The first narrow-field camera 41n, the first wide-field camera 41w, the second narrow-field camera 42n, and the second wide-field camera 42w are visible-light cameras that generate electronic data of still images or videos by converting visible light into electrical signals. The inspection camera 43 is an infrared camera that generates electronic data of still images or videos by converting infrared light into electrical signals. Visible-light cameras can produce clearer images than infrared cameras, but they cannot photograph a subject if there is an object that does not transmit visible light between the visible-light camera and the subject. Even if there is an object that does not transmit visible light between the infrared camera and the subject, if this object is made of a material that transmits infrared light, such as silicon, the infrared camera can photograph the subject through that object.

[0131] When the inspection camera 43 photographs the first substrate W1 and the second substrate W2, the first substrate W1 and the second substrate W2 are stacked on top of each other with the first substrate W1 facing downwards and the second substrate W2 facing upwards. Even if the first substrate W1 and the second substrate W2 in this state are photographed from above with a visible light camera, the first alignment mark AM1 on the first substrate W1 and the second alignment mark AM2 on the second substrate W2 will not be visible in the captured image. The inspection camera 43 is an infrared camera. At least one first alignment mark AM1 on the first substrate W1 is captured by the inspection camera 43 via the substrate WD1 (see Figure 3) of the first substrate W1. At least one second alignment mark AM2 on the second substrate W2 is captured by the inspection camera 43 via the first substrate W1. As a result, with the first substrate W1 and the second substrate W2 joined together, the first alignment mark AM1 and the second alignment mark AM2 are photographed by the inspection camera 43.

[0132] As shown in Figures 24 and 25, the joining unit 2b includes, in addition to the first camera 41, at least one alignment camera 45 that photographs the first fixed reference 31r and the second fixed reference 32r when they are spaced apart and facing each other in the Z direction. Figures 24 and 25 show an example in which two alignment cameras 45 are provided. In Figures 24 and 25, the lower surface of the first stage 13 is facing upward. The alignment camera 45 protrudes from the lower surface of the first stage 13. The alignment camera 45 is fixed to the first stage 13. The alignment camera 45 moves in the Z direction together with the first stage 13 and rotates together with the first stage 13 around the inversion center 16c.

[0133] At least a portion of the alignment camera 45 is positioned directly below the first fixed reference 31r (or directly above the first fixed reference 31r in the configuration shown in Figures 24 and 25). The lens of the alignment camera 45 faces the first fixed reference 31r directly in the Z direction. Figure 25 shows an example in which the alignment camera 45 is inserted into a through-hole 13h that penetrates the first stage 13 in the Z direction. The positioning of the alignment camera 45 is not limited to this, as long as the alignment camera 45 and the first fixed reference 31r face each other directly in the Z direction. In any case, the first reference marks 31m are positioned within the range that the alignment camera 45 can capture. The number of first reference marks 31m captured by one alignment camera 45 may be two or more.

[0134] When performing alignment adjustment to minimize the misalignment of the first substrate W1 and the second substrate W2, as shown in Figures 24 and 25, the first stage 13 is rotated 180 degrees around the inversion center 16c so that the first substrate W1 held by the first chuck 11 faces downwards. At this time, the first fixed reference 31r is also turned downwards. When the second chuck 21 is positioned below the first chuck 11 so that the downward-facing first substrate W1 held by the first chuck 11 and the upward-facing second substrate W2 held by the second chuck 21 face directly opposite each other in the Z direction, the first fixed reference 31r and the second fixed reference 32r face directly opposite each other in the Z direction with a gap between them. The alignment camera 45 photographs the first fixed reference 31r and the second fixed reference 32r in this state.

[0135] The alignment camera 45 is a visible light camera. When the alignment camera 45 is photographing the first fixed reference 31r and the second fixed reference 32r, the first fixed reference 31r, which corresponds to a mask, is positioned between the alignment camera 45 and the second fixed reference 32r. The first display panel 31d of the first fixed reference 31r is a transparent plate that transmits visible light. The second display panel 32d of the second fixed reference 32r is also a transparent plate. The alignment camera 45 simultaneously photographs both the first reference mark 31m and the second reference mark 32m when the first fixed reference 31r and the second fixed reference 32r are spaced apart and directly facing each other in the Z direction. If the alignment camera 45 can photograph the second reference mark 32m in this state, the second display panel 32d of the second fixed reference 32r may be an opaque plate that does not transmit visible light.

[0136] After the first camera 41 photographs the first alignment mark AM1 and the first reference mark 31m, not only the position and angle of the first substrate W1 in the horizontal plane, but also the relative position and angle between the first substrate W1 and the first fixed reference 31r can be determined. Similarly, after the second camera 42 photographs the second alignment mark AM2 and the second reference mark 32m, not only the position and angle of the second substrate W2 in the horizontal plane, but also the relative position and angle between the second substrate W2 and the second fixed reference 32r can be determined. When the alignment camera 45 photographs the first reference mark 31m and the second reference mark 32m, the relative position and angle between the first fixed reference 31r and the second fixed reference 32r can be determined. Once this is known, the relative position and angle between the first substrate W1 and the second substrate W2 can also be indirectly determined.

[0137] The control device 3 adjusts the relative position and angle of the first fixed reference 31r and the second fixed reference 32r so as to minimize the misalignment of the first substrate W1 and the second substrate W2, while having the alignment camera 45 photograph the first reference mark 31m and the second reference mark 32m.

[0138] Specifically, the control device 3 moves one or both of the first substrate W1 and the second substrate W2 in the horizontal plane using at least one of the first rotation motor 12, the second rotation motor 22, the Y actuator 25, and the X actuator 27. This minimizes the misalignment of the first substrate W1 and the second substrate W2. The control device 3 monitors the change in the misalignment value, and when the misalignment value converges to a certain value or less, it determines that the alignment adjustment is complete and the alignment is complete, stops the processing command for the alignment adjustment, and starts the processing command for the next process, the substrate contact process.

[0139] In the substrate contact process, the control device 3 lowers the first substrate W1 using the Z actuator 18, bringing the downward-facing bonding surface WA1 of the first substrate W1 into contact with the upward-facing bonding surface WA2 of the second substrate W2. This joins the first substrate W1 and the second substrate W2. The operation parameters, such as the distance traveled during the lowering operation, are based on a recipe stored in memory 3c. The control device 3 determines that the bonding is complete when the operations defined in the substrate contact process are completed.

[0140] After the first substrate W1 and the second substrate W2 are bonded together, the inspection camera 43 photographs the first substrate W1 and the second substrate W2 while they are held in the second chuck 21. As a result, the first alignment mark AM1 and the second alignment mark AM2 are captured by the inspection camera 43 while the first substrate W1 and the second substrate W2 are bonded together. The amount of misalignment between the two bonded substrates W and the amount of misalignment between the two bonded substrates W are measured based on the images of the first alignment mark AM1 and the second alignment mark AM2 generated by the inspection camera 43.

[0141] The position captured by the first camera 41 is changed by the X actuator 27 and Y actuator 25 moving the first camera 41. The position captured by the second camera 42 is changed by the X actuator 27 and Y actuator 25 moving the upper stage 23u. The same applies to the alignment camera 45 and inspection camera 43 as to the second camera 42. The X actuator 27 and Y actuator 25 also function as shooting position adjustment actuators that change the position captured by cameras such as the first camera 41. The joining unit 2b may also include shooting position adjustment actuators that move the second camera 42 and inspection camera 43 relative to the base 20.

[0142] Figure 26 is a schematic diagram showing an example of an image of the cross-shaped first alignment mark AM1 generated by the first camera 41.

[0143] The first camera 41 generates a rectangular or square two-dimensional image composed of multiple pixels aligned vertically and horizontally in mutually orthogonal directions. The control device 3 receives the electronic data of the still or moving image generated by the first camera 41. The control device 3 analyzes and makes decisions about the image based on a program stored in memory 3c. The same applies to other cameras such as the second camera 42.

[0144] If the first reference mark 31m has a shape that allows for the identification of mutually orthogonal height and width directions, when the first camera 41 photographs the first reference mark 31m, the height and width directions of the first reference mark 31m are determined based on the image of the first reference mark 31m. This allows for the determination of the tilt angle of the first reference mark 31m with respect to the image generated by the first camera 41 (for example, the tilt angle of the first reference mark 31m in the width direction with respect to the width direction of the image).

[0145] The height and width directions of the first reference mark 31m are recorded in memory 3c in correspondence with the shape of the first reference mark 31m being used. The control device 3 identifies the height and width directions of the first reference mark 31m in the captured image by matching the shape of the first reference mark 31m recorded in memory 3c with the first reference mark 31m in the image generated by the first camera 41.

[0146] If the first reference mark 31m has a shape that allows for the identification of mutually orthogonal height and width directions, the first fixed reference 31r may be positioned such that the inclination angle of the first reference mark 31m with respect to the vertical or horizontal direction of the image generated by the first camera 41 is less than or equal to the upper limit. For example, if a horizontal line parallel to the width direction is included in the contour of the first reference mark 31m, the first fixed reference 31r may be positioned such that the inclination angle of the horizontal line with respect to the horizontal direction of the image generated by the first camera 41 is less than or equal to the upper limit. The same applies to the positioning of the second fixed reference 32r with respect to the second reference mark 32m.

[0147] Similar to the first reference mark 31m, if the first alignment mark AM1 has a shape that allows for the identification of mutually orthogonal height and width directions, when the first camera 41 photographs the first alignment mark AM1, the height and width directions of the first alignment mark AM1 are determined based on the image of the first alignment mark AM1. This allows for the determination of the tilt angle of the first alignment mark AM1 with respect to the image generated by the first camera 41 (for example, the tilt angle of the first alignment mark AM1 in the width direction with respect to the lateral direction of the image).

[0148] If the first alignment mark AM1 has a shape that allows for the identification of mutually orthogonal height and width directions, the control device 3 may determine whether the tilt angle of the first alignment mark AM1 with respect to the vertical or horizontal direction of the image generated by the first camera 41 exceeds an upper limit. If the control device 3 determines that the tilt angle of the first alignment mark AM1 exceeds an upper limit, it may rotate the first chuck 11 with the first rotation motor 12 so that the tilt angle decreases to or below the upper limit, and then have the first camera 41 photograph the first alignment mark AM1.

[0149] The left side of Figure 26 shows an example of an image generated by the first camera 41. The two grids on the right side of Figure 26 are enlarged portions of the image on the left side of Figure 26. Each square block in each grid corresponds to a pixel. The boundary between the black and white blocks corresponds to the outline of the first alignment mark AM1. In the upper right grid of Figure 26, the outline of the first alignment mark AM1 is tilted with respect to the horizontal direction of the image (left-right direction of the paper). In contrast, in the lower right grid of Figure 26, the outline of the first alignment mark AM1 is parallel to the horizontal direction of the image.

[0150] The control device 3 may, for example, rotate the first chuck 11 with the first rotation motor 12 so that the image generated by the first camera 41 changes from the state of the grid in the upper right of Figure 26 to the state of the grid in the lower right of Figure 26. In this way, the shape of the first alignment mark AM1 can be identified with higher accuracy from the image generated by the first camera 41, and the position and angle of the first substrate W1 can be identified with higher accuracy. The same applies when capturing the second alignment mark AM2 with the second camera 42, or when capturing the first alignment mark AM1 and the second alignment mark AM2 with the inspection camera 43.

[0151] Next, we will describe the displacement detector 46.

[0152] Figure 27 is a block diagram of the displacement detector 46. Figure 28 is a diagram showing a vertical cross-section of the displacement detector 46. Figure 29 is a view of the displacement detector 46 vertically in the direction of arrow XXIX shown in Figure 28.

[0153] As shown in Figure 27, the joining unit 2b includes a displacement detector 46 that detects the horizontal relative displacement between the base 20 and the second chuck 21. Figures 28 and 29 show an example in which the displacement detector 46 detects the horizontal relative displacement between the base 20 and the second chuck 21 by detecting the displacement of the second chuck 21 relative to the base 20 in the X and Y directions.

[0154] The displacement detector 46 includes an X-displacement detector that detects the relative displacement between the base 20 and the second chuck 21 in the X direction, and a Y-displacement detector that detects the relative displacement between the base 20 and the second chuck 21 in the Y direction. The Y-displacement detector may be integrated with the X-displacement detector, or it may be a displacement detector independent of the X-displacement detector. Figure 27 shows an example of the former.

[0155] The X displacement detector may be an optical displacement detector or another type of displacement detector such as a magnetic one. In the former case, the X displacement detector may be a grid interferometer linear encoder or another type of displacement detector such as an interferometer. Compared to interferometers such as laser interferometers, grid interferometer linear encoders are less affected by environmental changes such as temperature and humidity changes and can stably detect displacement. The X displacement detector may be an absolute position detector that detects the absolute position of an object such as a second substrate W2, or a relative position detector that detects the relative position of an object. The contents of this paragraph are also similar for the Y displacement detector. If the Y displacement detector is independent of the X displacement detector, the Y displacement detector may be a different type of displacement detector than the X displacement detector.

[0156] Figure 27 shows an example including a grid-interference linear encoder in which a displacement detector 46 detects the displacement of the second substrate W2 in the X and Y directions. The grid-interference displacement detector 46 includes a scale 47 that moves in the X and Y directions together with an object such as the second chuck 21, and a detection head 48 that detects the displacement of the scale 47 in the X and Y directions. The detection head 48 is a non-contact displacement sensor that detects the displacement of the scale 47 in the X and Y directions without contacting the scale 47.

[0157] The detection head 48 includes an X detection head 48x that detects the displacement of the scale 47 in the Y direction, and a Y detection head 48y that detects the displacement of the scale 47 in the Y direction. Both the X detection head 48x and the Y detection head 48y include a light source 48s that emits light reflected by the scale 47, a photodetector 48r that converts the light reflected by the scale 47 into an electrical signal, and an interpreter 48i that calculates the amount and direction of horizontal movement of the scale 47 based on the electrical signal input from the photodetector 48r. The displacement of the second substrate W2 detected by the displacement detector 46 is input to the control device 3 through the interpreter 48i.

[0158] Scale 47 is a reflective diffraction grating that reflects light emitted from a light source 48s, such as a semiconductor laser. Scale 47 may be a holographic diffraction grating or another type of diffraction grating, such as a blazed diffraction grating. Scale 47 may be a planar diffraction grating or a volume diffraction grating. Figure 29 shows an example where scale 47 is a 2D scale that reflects light emitted from an X detection head 48x and a Y detection head 48y. In this example, scale 47 includes a square reflective surface 47s that reflects light emitted from the light source 48s.

[0159] The 2D scale includes an X diffraction grating containing a surface with regularly alternating convex and concave portions in the X direction, and a Y diffraction grating containing a surface with regularly alternating convex and concave portions in the Y direction. The Y diffraction grating moves horizontally in the same direction, velocity, and amount as the X diffraction grating. The tips of the convex portions of the X diffraction grating and the tips of the convex portions of the Y diffraction grating are aligned on a single horizontal plane. The surface of the X diffraction grating and the surface of the Y diffraction grating are part of the reflective surface 47s of scale 47. Light emitted from the light source 48s is reflected by the surface of the X diffraction grating or the surface of the Y diffraction grating.

[0160] The recesses of the X diffraction grating may be a continuous groove extending in the Y direction, where the length is greater than the width, or a plurality of holes arranged in the Y direction, or they may include a groove and a plurality of holes. The recesses of the Y diffraction grating may be a continuous groove extending in the X direction, where the length is greater than the width, or a plurality of holes arranged in the X direction, or they may include a groove and a plurality of holes. The vertical cross-sections of the convex and recesses of the X diffraction grating may be an arc, a square, or a rectangle, or other shapes. The same applies to the vertical cross-sections of the convex and recesses of the Y diffraction grating.

[0161] The scale 47 is fixed to the underside of the upper stage 23u with its light-reflecting surface 47s facing downwards. When the upper stage 23u moves in the Y direction relative to the base 20, the scale 47 moves horizontally in the same direction, with the same velocity and amount of movement as the upper stage 23u. When the lower stage 23L moves in the X direction relative to the base 20, the scale 47 moves horizontally in the same direction, with the same velocity and amount of movement as the lower stage 23L. When the second substrate W2, held by the second chuck 21, moves horizontally in at least one of the X and Y directions relative to the base 20, the scale 47 moves horizontally in the same direction, with the same velocity and amount of movement as the second substrate W2. Even when the upper stage 23u moves in the Y direction relative to the lower stage 23L, the scale 47 does not collide with the lower stage 23L.

[0162] The detection head 48 is positioned below the scale 47 and the upper stage 23u. The detection head 48 faces the scale 47 directly in the Z direction. The detection head 48 is inserted into a through-window 23w that penetrates the lower stage 23L vertically. The detection head 48 is fixed to the base 20. The detection head 48 includes a light-emitting part 48e that faces the scale 47 directly in the Z direction, and a body 48m that supports the light-emitting part 48e. The body 48m is inserted into the through-window 23w of the lower stage 23L. The light-emitting part 48e is positioned above the upper surface of the lower stage 23L. Regardless of the position of the second chuck 21, the detection head 48 does not come into contact with the lower stage 23L.

[0163] Regardless of the position of the second chuck 21, the light-emitting unit 48e faces the scale 47 directly in the Z direction. Light from the light source 48s (see Figure 27) travels from the light-emitting unit 48e to the scale 47, is reflected by the scale 47, and then enters the detection head 48. The detection head 48 detects the displacement of the scale 47 in the X and Y directions relative to the base 20. This allows the X and Y displacements of the second chuck 21 relative to the base 20 to be detected. The detection results for the X and Y displacements are transmitted to the control device 3.

[0164] Next, the electrical configuration of the substrate bonding apparatus 1 will be described.

[0165] Figure 30 is a block diagram showing the electrical configuration of the substrate bonding apparatus 1. The control device 3 controls the electrical and electronic equipment installed in the substrate bonding apparatus 1 by transmitting control signals. The control device 3 also receives signals from various cameras, detectors, etc., installed in the substrate bonding apparatus 1 and stores and processes the signals. The control device 3 is programmed to execute each of the processes described later.

[0166] The control device 3 includes at least one computer. The computer includes a computer body 3a and peripheral devices 3d connected to the computer body 3a. The computer body 3a includes a CPU 3b (central processing unit) that executes various instructions and a memory 3c that stores information. The peripheral devices 3d include a storage device 3e that stores information such as a program P, a reader 3f that reads information from removable media RM, and a communication device 3g that communicates with other devices such as a host computer.

[0167] The control device 3 is connected to an input device and a display device. The input device is operated when an operator, such as a user or maintenance personnel, inputs information into the substrate bonding device 1. The information is displayed on the screen of the display device. The input device may be a keyboard, a pointing device, or a touch panel, or it may be a device other than these. A touch panel display that serves as both an input device and a display device may be provided in the substrate bonding device 1.

[0168] The CPU 3b executes the program P stored in the storage 3e. The program P in the storage 3e may be pre-installed on the control unit 3, sent to the storage 3e from the removable media RM via the reader 3f, or sent to the storage 3e from an external device such as a host computer via the communication device 3g.

[0169] Storage 3e and removable media RM are non-volatile memories that retain data even without power supply. Storage 3e is, for example, a magnetic storage device such as a hard disk drive. Removable media RM is, for example, an optical disc such as a compact disc or a semiconductor memory such as a memory card. Removable media RM is an example of a computer-readable recording medium on which a program P is recorded. Removable media RM is a non-transitory tangible recording medium.

[0170] Next, a first example of bonding the first substrate W1 and the second substrate W2 will be described.

[0171] Figures 31A, 31B, 31C, 31D, 31E, 31F, 31G, and 31H are schematic diagrams illustrating a first example of bonding a first substrate W1 and a second substrate W2 performed by bonding unit 2b. Figures 31A to 31H show the first chuck 11 and the second chuck 21, etc., viewed horizontally in the Y direction. Figures 32A and 32B are similar.

[0172] When joining the first substrate W1 and the second substrate W2 by the joining unit 2b, as shown in Figure 31A, the transport robot TR holds the first substrate W1 horizontally with the hand TH, with the joining surface WA1 facing upwards, and places the first substrate W1 in the space above the first chuck 11 in the loading / unloading state. Similarly, the transport robot TR holds the second substrate W2 horizontally with the hand TH, with the joining surface WA2 facing upwards, and places the second substrate W2 in the space above the second chuck 21 in the loading / unloading state.

[0173] The loading / unloading state of the first chuck 11 is when the first chuck 11 is maintained in a position and orientation that allows the first substrate W1 to be moved between the transport robot TR and the first chuck 11, the rotation angle of the first chuck 11 is zero, and the center line 11c of the first chuck 11 is vertical. The loading / unloading state of the second chuck 21 is when the second chuck 21 is maintained in a position and orientation that allows the second substrate W2 to be moved between the transport robot TR and the second chuck 21, and the rotation angle of the second chuck 21 is zero.

[0174] The transport robot TR may place the first substrate W1 on the first chuck 11 in the loading / unloading state, or it may place the first substrate W1 on a plurality of lifting pins that move up and down while horizontally supporting the first substrate W1 between a lower position where the first substrate W1 is placed on the first chuck 11 and an upper position where the first substrate W1 is separated from the first chuck 11. The same applies to the second substrate W2. In either case, the first substrate W1 is placed on the first chuck 11 in the loading / unloading state with the bonding surface WA1 facing upwards and is attracted to the first chuck 11. Similarly, the second substrate W2 is placed on the second chuck 21 in the loading / unloading state with the bonding surface WA2 facing upwards and is attracted to the second chuck 21.

[0175] If the first suction surface 11s of the first chuck 11 in the loading / unloading state is a horizontal plane, the upper surface (bonding surface WA1) and lower surface of the first substrate W1 remain horizontal planes even after the first substrate W1 is adsorbed onto the first chuck 11. If the first suction surface 11s of the first chuck 11 in the loading / unloading state has a shape other than a horizontal plane, when the first substrate W1 is adsorbed onto the first chuck 11, the upper and lower surfaces of the first substrate W1 elastically deform from a horizontal plane to a shape identical or nearly identical to the first suction surface 11s of the first chuck 11, and maintain that shape. The same applies to the shapes of the upper and lower surfaces of the second substrate W2.

[0176] Figure 31A shows an example where the first suction surface 11s of the first chuck 11 in the loading / unloading state is a horizontal plane, and the second suction surface 21s of the second chuck 21 in the loading / unloading state is a horizontal plane. Therefore, in this example, even after the first substrate W1 is adsorbed onto the first chuck 11, the top and bottom surfaces of the first substrate W1 are maintained as horizontal planes. Similarly, even after the second substrate W2 is adsorbed onto the second chuck 21, the top and bottom surfaces of the second substrate W2 are maintained as horizontal planes.

[0177] The first substrate W1 may be placed on the first chuck 11 before or after the second substrate W2 is placed on the second chuck 21, or it may be placed on the first chuck 11 at the same time as the second substrate W2 is placed on the second chuck 21. The hand TH that transports the first substrate W1 may be the same as or different from the hand TH that transports the second substrate W2. In the latter case, two hands TH of one transport robot TR may transport the first substrate W1 and the second substrate W2, or two transport robots TR may transport the first substrate W1 and the second substrate W2.

[0178] When the first substrate W1 is placed on the first chuck 11, as shown in Figure 31B, the inversion actuator 16 rotates the first chuck 11 180 degrees so that the bonding surface WA1 of the first substrate W1 faces downward. Before or after this, the upper stage 23u is positioned in a shooting area where the first narrow-field camera 41n and the first wide-field camera 41w, mounted on the upper stage 23u, can photograph at least one of the first substrate W1 and the first fixed reference 31r. In this state, the first narrow-field camera 41n and the first wide-field camera 41w begin shooting. If necessary, the first narrow-field camera 41n and the first wide-field camera 41w may be moved horizontally relative to the first substrate W1 during shooting.

[0179] The first narrow-field camera 41n and the first wide-field camera 41w capture images of at least one first alignment mark AM1 on the first substrate W1 and at least one first reference mark 31m on the first fixed reference 31r. After that, the first narrow-field camera 41n and the first wide-field camera 41w terminate their imaging. By analyzing the captured image of at least one first alignment mark AM1, the position and angle of the first substrate W1 are determined. Similarly, by analyzing the captured image of at least one first reference mark 31m, the position and angle of the first fixed reference 31r are determined. This allows the position and angle of the first substrate W1 relative to the first fixed reference 31r to be determined.

[0180] For example, the coordinate position of the first narrow-field camera 41n when the first alignment mark AM1 is imaged may be calculated based on the design information of the device and the output of the displacement detector 46 described later. Based on the calculated coordinate position of the first narrow-field camera 41n and the position of the first alignment mark AM1 in the captured image, the coordinate position of the first alignment mark AM1 may be identified. By similarly identifying the coordinate position of the first reference mark 31m, the relative positions of the first alignment mark AM1 and the first reference mark 31m may be calculated, and the position of the first substrate W1 with respect to the first fixed reference 31r may be determined.

[0181] When the second substrate W2 is placed on the second chuck 21, the upper stage 23u is positioned in a shooting area where the second narrow-field camera 42n and the second wide-field camera 42w can photograph at least one of the second substrate W2 and the second fixed reference 32r, as shown in Figure 31C. In this state, the second narrow-field camera 42n and the second wide-field camera 42w begin shooting. If necessary, the upper stage 23u may be moved horizontally relative to the second narrow-field camera 42n and the second wide-field camera 42w during shooting.

[0182] The second narrow-field camera 42n and the second wide-field camera 42w capture images of at least one second alignment mark AM2 on the second substrate W2 and at least one second reference mark 32m on the second fixed reference 32r. After that, the second narrow-field camera 42n and the second wide-field camera 42w terminate their imaging. By analyzing the captured image of at least one second alignment mark AM2, the position and angle of the second substrate W2 are determined. Similarly, by analyzing the captured image of at least one second reference mark 32m, the position and angle of the second fixed reference 32r are determined. This allows the position and angle of the second substrate W2 relative to the second fixed reference 32r to be determined.

[0183] For example, the position of the second chuck 21 when the second alignment mark AM2 is imaged can be obtained based on the output of the displacement detector 46 described later, and the coordinate position of the second alignment mark AM2 linked to the position of the second chuck 21 can be determined based on the position of the second alignment mark AM2 in the imaged image. The position of the second reference mark 32m can also be determined by a similar method linked to the position of the second chuck 21. By calculating the coordinate position of the second alignment mark AM2 when the second reference mark 32m is imaged based on the position of the second chuck 21 when the second reference mark 32m is imaged and the position of the second chuck 21 when the second alignment mark AM2 is imaged, the relative position between the second alignment mark AM2 and the second reference mark 32m can be calculated. This allows the position of the second substrate W2 relative to the second fixed reference 32r to be determined.

[0184] At least a portion of the time during which the second narrow-field camera 42n and the second wide-field camera 42w photograph the second substrate W2 and the second fixed reference 32r may or may not coincide with the time during which the first narrow-field camera 41n and the first wide-field camera 41w photograph the first substrate W1 and the first fixed reference 31r. At least a portion of the time during which the second alignment mark AM2 of the second substrate W2 held in the second chuck 21 is photographed may or may not coincide with the time during which the second reference mark 32m is photographed. The same applies to the time during which the first alignment mark AM1 and the first reference mark 31m are photographed.

[0185] The control device 3, after determining the position and angle of the first substrate W1 relative to the first fixed reference 31r and the position and angle of the second substrate W2 relative to the second fixed reference 32r, operates the actuator mechanism AC to change the first substrate W1 and the second substrate W2 from a non-facing state to a facing state. The control device 3 then operates the actuator mechanism AC to change the first substrate W1 and the second substrate W2 from a facing state to an alignment completed state. The control device 3 then operates the actuator mechanism AC to change the first substrate W1 and the second substrate W2 from an alignment completed state to a bonding completed state.

[0186] The non-facing state is when the first substrate W1 and the second substrate W2 do not overlap when viewed vertically. The facing state and the alignment completed state are when the entire bonding surface WA1 of the first substrate W1 and the entire bonding surface WA2 of the second substrate W2 are separated from each other, and the first substrate W1 and the second substrate W2 face each other vertically. The facing state is the state before alignment of the first substrate W1 and the second substrate W2, and the alignment completed state is the state after alignment of the first substrate W1 and the second substrate W2. The facing state is the non-aligned state. The bonding completed state is when the entire bonding surface WA1 of the first substrate W1 and the entire bonding surface WA2 of the second substrate W2 are in contact.

[0187] When changing the first substrate W1 and the second substrate W2 from a non-facing state to a facing state, with the bonding surface WA1 of the first substrate W1 held by the first chuck 11 facing downwards, the upper stage 23u and the lower stage 23L are moved in the X direction, thereby moving the second substrate W2 held by the second chuck 21 in the X direction relative to the first substrate W1 held by the first chuck 11. As a result, as shown in Figure 31D, the first substrate W1 and the second substrate W2 face each other vertically, with the entire bonding surface WA1 of the first substrate W1 and the entire bonding surface WA2 of the second substrate W2 separated from each other.

[0188] After the first substrate W1 and the second substrate W2 are changed to a facing state, with the bonding surface WA1 of the first substrate W1 held by the first chuck 11 facing downwards, the first substrate W1 and the second substrate W2 are moved relative to each other in at least one of the X direction, Y direction, first θ direction (circumferential direction of the first substrate W1), and second θ direction (circumferential direction of the second substrate W2). This changes the first substrate W1 and the second substrate W2 from a facing state to an alignment completed state. Subsequently, by moving the first stage 13 in the Z direction, the first substrate W1 held by the first chuck 11 is lowered relative to the second substrate W2 held by the second chuck 21. As a result, as shown in Figure 31E, the entire bonding surface WA1 of the first substrate W1 comes into contact with the entire bonding surface WA2 of the second substrate W2, and the first substrate W1 and the second substrate W2 change from an alignment completed state to a bonding completed state.

[0189] The bonded state is a state in which the first substrate W1 and the second substrate W2 are aligned with respect to at least one first alignment mark AM1 and at least one second alignment mark AM2. The alignment-completed state may differ from the bonded state only in whether or not the first substrate W1 and the second substrate W2 are separated from each other. In other words, the alignment-completed state may be a state in which the first substrate W1 and the second substrate W2 are separated from each other and the misalignment between the first substrate W1 and the second substrate W2 is minimized.

[0190] As shown in Figure 31D, the opposing state of the first substrate W1 and the second substrate W2 is when the first fixed reference 31r and the second fixed reference 32r overlap when viewed vertically. Before or after the first substrate W1 and the second substrate W2 are in the opposing state, the alignment camera 45 attached to the first stage 13 starts taking images. As a result, the alignment camera 45 captures images of the first fixed reference 31r and the second fixed reference 32r overlapping in a plan view.

[0191] The control device 3 determines the position and angle of the first fixed reference 31r by analyzing the image of the first fixed reference 31r captured by the alignment camera 45. Similarly, the control device 3 determines the position and angle of the second fixed reference 32r by analyzing the image of the second fixed reference 32r captured by the alignment camera 45. Since the first narrow-field camera 41n and the first wide-field camera 41w have already captured images of the first substrate W1 and the first fixed reference 31r, the position and angle of the first substrate W1 can be indirectly determined from the position and angle of the first fixed reference 31r. Similarly, since the second narrow-field camera 42n and the second wide-field camera 42w have already captured images of the second substrate W2 and the second fixed reference 32r, the position and angle of the second substrate W2 can be indirectly determined from the position and angle of the second fixed reference 32r. Therefore, the first substrate W1 and the second substrate W2 can be indirectly aligned by adjusting the relative position and angle between the first fixed reference 31r and the second fixed reference 32r.

[0192] The control device 3 moves the first substrate W1 and the second substrate W2 relative to each other in at least one of the X, Y, first θ, and second θ directions, while having the alignment camera 45 capture images of the first fixed reference 31r and the second fixed reference 32r. As a result, the first substrate W1 and the second substrate W2 are aligned so that they are separated from each other, and the alignment is completed. After the control device 3 determines that the first substrate W1 and the second substrate W2 have changed to the alignment completed state, the alignment camera 45 stops capturing images.

[0193] After the first substrate W1 and the second substrate W2 change to the alignment-completed state, the control device 3 moves the first substrate W1 and the second substrate W2 relative to each other at least in the Z direction, thereby bringing the entire bonding surface WA1 of the first substrate W1 into contact with the entire bonding surface WA2 of the second substrate W2. As a result, the first substrate W1 and the second substrate W2 change from the alignment-completed state to the bonding-completed state. When the first substrate W1 and the second substrate W2 change to the bonding-completed state, the first chuck 11 releases its grip on the first substrate W1, and the second chuck 21 continues to grip the second substrate W2. Therefore, the bonded first substrate W1 and the second substrate W2 are held by the second chuck 21.

[0194] After the first substrate W1 and the second substrate W2 have changed to the bonded state, the control device 3 raises the first chuck 11 and retracts the second chuck 21 (moves it to the negative side in the X direction), as shown in Figure 31F. After the first chuck 11 has risen, the reversing actuator 16 rotates the first chuck 11 so that its centerline 11c is vertical, as shown in Figure 31G, so that the first chuck 11 faces upward. If necessary, the first rotation motor 12 rotates the first chuck 11 to return its rotation angle to zero. This returns the first chuck 11 to the loading / unloading state. The reversing actuator 16 may start rotating the first chuck 11 before or after the second chuck 21 retracts, or it may start rotating the first chuck 11 at the same time as the second chuck 21 retracts. The same applies to the first rotation motor 12.

[0195] After the first substrate W1 and the second substrate W2 have been bonded together, the upper stage 23u is positioned in a shooting area where the inspection camera 43 can photograph the first substrate W1 and the second substrate W2 on the second chuck 21, as shown in Figure 31G. The inspection camera 43 starts taking pictures before or after this. This captures at least one first alignment mark AM1 and at least one second alignment mark AM2 with the first substrate W1 and the second substrate W2 bonded together. If necessary, the upper stage 23u may be moved horizontally relative to the inspection camera 43 during this photography.

[0196] Images of the first alignment mark AM1 and the second alignment mark AM2 captured by the inspection camera 43 are transmitted to the control device 3 for analysis. This analysis detects the bonding accuracy of the first substrate W1 and the second substrate W2. Furthermore, based on the detected bonding accuracy of the first substrate W1 and the second substrate W2, changes in the alignment of the first substrate W1 and the second substrate W2 between the alignment-completed state and the bonding-completed state are detected. When bonding another first substrate W1 and second substrate W2, the alignment of the first substrate W1 and the second substrate W2 in the alignment-completed state is corrected based on the detected changes in alignment. This ensures that subsequent first substrates W1 and second substrates W2 are bonded with higher accuracy.

[0197] After the inspection camera 43 photographs the first substrate W1 and the second substrate W2 on the second chuck 21, the second chuck 21 is moved in the X direction while holding the bonded first substrate W1 and the second substrate W2 in the second chuck 21 until the second chuck 21 returns to the loading / unloading state. If necessary, the second rotation motor 22 rotates the second chuck 21 to return its rotation angle to zero. Then, as shown in Figure 31H, the second chuck 21 releases its grip on the second substrate W2, and the hand TH of the transport robot TR receives the first substrate W1 and the second substrate W2 directly or indirectly from the second chuck 21. As a result, the bonded first substrate W1 and the second substrate W2 are held horizontally by the hand TH of the transport robot TR and transported to the next destination.

[0198] When joining the next first substrate W1 and second substrate W2, the series of steps described above are repeated. The next first substrate W1 may be placed in the first chuck 11 before or after removing the previous first substrate W1 and second substrate W2 from the second chuck 21, or it may be placed in the first chuck 11 at the same time as removing the previous first substrate W1 and second substrate W2 from the second chuck 21. For example, as shown in Figure 31G, the next first substrate W1 may be placed in the first chuck 11 while the inspection camera 43 is photographing the joined first substrate W1 and second substrate W2. In this case, as shown in Figure 31H, the inversion actuator 16 may invert the next first substrate W1 before removing the previous first substrate W1 and second substrate W2 from the second chuck 21.

[0199] Next, a second example of joining the first substrate W1 and the second substrate W2 will be described.

[0200] Figures 32A and 32B are schematic diagrams illustrating a second example of bonding the first substrate W1 and the second substrate W2 by the bonding unit 2b.

[0201] As described above, the first substrate W1 and the second substrate W2 change from a non-facing state to a facing state and then to an alignment-completed state, and finally to a bond-completed state. The bond-completed state is a state in which the entire bond surface WA1 of the first substrate W1 and the entire bond surface WA2 of the second substrate W2 are in contact. The alignment-completed state may be a state in which the entire bond surface WA1 of the first substrate W1 and the entire bond surface WA2 of the second substrate W2 are parallel or non-parallel, or a part of the bond surface WA1 of the first substrate W1 and a part of the bond surface WA2 of the second substrate W2 are parallel, and the remaining part of the bond surface WA1 of the first substrate W1 and the remaining part of the bond surface WA2 of the second substrate W2 are non-parallel.

[0202] In other words, the alignment completion state may differ from the bonding completion state in at least one of the following: the position of the first substrate W1 in the horizontal direction relative to the second substrate W2, the orientation of the first substrate W1 relative to the second substrate W2, the relative angle between the first substrate W1 and the second substrate W2 around a vertical line, the shape of the bonding surface WA1 of the first substrate W1, and the shape of the bonding surface WA2 of the second substrate W2, in addition to whether the first substrate W1 and the second substrate W2 are separated from each other (in addition to the position of the first substrate W1 in the vertical direction relative to the second substrate W2).

[0203] Figures 31A to 31H show an example of joining the first substrate W1 and the second substrate W2 by full-surface simultaneous bonding, in which the entire bonding surface WA1 of the first substrate W1 and the entire bonding surface WA2 of the second substrate W2 are brought into contact simultaneously or nearly simultaneously. However, the first substrate W1 and the second substrate W2 may also be joined by boundary shift bonding, which expands the area in which the bonding surface WA1 of the first substrate W1 and the bonding surface WA2 of the second substrate W2 are joined.

[0204] Boundary shift bonding is a bonding method in which the boundary between the contact area within the bonding surface WA1 of the first substrate W1 that is in contact with the bonding surface WA2 of the second substrate W2 and the non-contact area within the bonding surface WA1 of the first substrate W1 that is away from the bonding surface WA2 of the second substrate W2 is moved within the bonding surface WA1 of the first substrate W1 until the entire bonding surface WA1 of the first substrate W1 and the entire bonding surface WA2 of the second substrate W2 are in contact.

[0205] In boundary shift bonding, the first substrate W1 and the second substrate W2 are changed from an alignment-completed state to a bond-completed state, passing through a bonding start state and a bonding expansion state. Even when bonding the first substrate W1 and the second substrate W2 in full-surface simultaneous bonding, there are strictly speaking periods of bonding start state and bonding expansion state, but since these periods are extremely short, it can be considered that the first substrate W1 and the second substrate W2 change from an alignment-completed state to a bond-completed state without passing through a bonding start state and bonding expansion state.

[0206] The bonding start state and the bonding expansion state are both states in which the first substrate W1 and the second substrate W2 are in the process of changing from the alignment complete state to the bonding complete state. The bonding start state is the state in which contact has begun between the bonding surface WA1 of the first substrate W1 and the bonding surface WA2 of the second substrate W2. The bonding expansion state is the state in which the area in contact between the bonding surface WA1 of the first substrate W1 and the bonding surface WA2 of the second substrate W2 is expanded until the entire area of ​​the bonding surface WA1 of the first substrate W1 and the entire area of ​​the bonding surface WA2 of the second substrate W2 are in contact. In both the bonding start state and the bonding expansion state, a part of the bonding surface WA1 of the first substrate W1 and a part of the bonding surface WA2 of the second substrate W2 are in contact, while the remaining part of the bonding surface WA1 of the first substrate W1 and the remaining part of the bonding surface WA2 of the second substrate W2 are separated.

[0207] The alignment completion state, bonding start state, bonding expansion state, and bonding completion state are such that at least one of the following is different from each other: the position of the first substrate W1 in the vertical direction relative to the second substrate W2, the position of the first substrate W1 in the horizontal direction relative to the second substrate W2, the orientation of the first substrate W1 relative to the second substrate W2, the relative angle between the first substrate W1 and the second substrate W2 around a vertical line, the shape of the bonding surface WA1 of the first substrate W1, and the shape of the bonding surface WA2 of the second substrate W2.

[0208] The bonding start state, bonding expansion state, and bonding completion state are defined as having at least one of these conditions, plus the areas of the contact regions between the bonding surface WA1 of the first substrate W1 and the bonding surface WA2 of the second substrate W2 being different from each other. The bonding expansion state represents all consecutive states from the bonding start state to the bonding completion state of the first substrate W1 and the second substrate W2. During the process of the first substrate W1 and the second substrate W2 changing from the bonding start state to the bonding completion state via the bonding expansion state, the entire area of ​​the bonding surface WA1 of the first substrate W1 and the entire area of ​​the bonding surface WA2 of the second substrate W2 are bonded while the first substrate W1 and the second substrate W2 are aligned.

[0209] For example, as shown in Figure 32A, the alignment completion state may be a state in which the entire bonding surface WA1 of the first substrate W1 is positioned in a single plane inclined with respect to the horizontal plane, and the entire bonding surface WA2 of the second substrate W2 is positioned in a single horizontal plane. Specifically, the inversion actuator 16 may adjust the angle of the first stage 13 so that the downward bonding surface WA1 of the first substrate W1 is inclined with respect to the horizontal plane. In this case, after performing alignment adjustment of the first substrate W1 and the second substrate W2 with the bonding surface WA1 of the first substrate W1 in a horizontal state, the bonding surface WA1 of the first substrate W1 may be inclined, and the relative positions of the first substrate W1 and the second substrate W2 in the horizontal direction may be adjusted according to the inclination angle of the bonding surface WA1 of the first substrate W1 with respect to the horizontal plane.

[0210] When the downward-facing bonding surface WA1 of the first substrate W1 is inclined with respect to the horizontal plane, a portion of the outer circumference of the bonding surface WA1 of the first substrate W1 corresponds to the lower end of the first substrate W1, which is positioned lower than the rest of the bonding surface WA1 of the first substrate W1. The lower end of the first substrate W1 corresponds to the bonding start position where it first contacts the bonding surface WA2 of the second substrate W2. When changing the first substrate W1 and the second substrate W2 from the alignment completed state to the bonding start state, the lower end of the first substrate W1 is brought into contact with the bonding surface WA2 of the second substrate W2, as shown in Figure 32B. Then, the first substrate W1 and the second substrate W2 are moved relative to each other in the Z and X directions while the first substrate W1 is rotated around the reversal center 16c so that the first substrate W1 rotates relative to the second substrate W2 around a horizontal line parallel to the reversal center 16c of the first chuck 11 and passing through the lower end of the first substrate W1.

[0211] The area of ​​the region where the bonding surface WA1 of the first substrate W1 and the bonding surface WA2 of the second substrate W2 are in contact increases continuously as the inclination angle of the bonding surface WA1 of the first substrate W1 with respect to the horizontal plane decreases. When the bonding surface WA1 of the first substrate W1 becomes horizontal, the entire area of ​​the bonding surface WA1 of the first substrate W1 and the entire area of ​​the bonding surface WA2 of the second substrate W2 come into contact. As a result, the first substrate W1 and the second substrate W2 change from a bonding start state to a bonding expansion state and then to a bonding completion state. By bonding the first substrate W1 and the second substrate W2 in this way, it is possible to eliminate or reduce air bubbles (also called voids) remaining between the bonded first substrate W1 and the second substrate W2.

[0212] The first chuck 11 releases its grip on the first substrate W1 after the first substrate W1 and the second substrate W2 have changed to a bonded state. This is not limited to cases where the first chuck 11 includes multiple suction regions that generate suction forces independently of each other. The control device 3 may reduce the number of suction regions that generate suction forces as the area of ​​the contact region increases, so that the grip on the first substrate W1 is released in the contact region where the bonding surface WA1 of the first substrate W1 and the bonding surface WA2 of the second substrate W2 are in contact.

[0213] Specifically, if the multiple adsorption regions are arranged and shaped as shown in Figure 10B, the control device 3 may stop the generation of suction force in each adsorption region, one by one, from the adsorption region at the right end of Figure 10B to the adsorption region at the left end of Figure 10B. In this way, a boundary extending vertically in Figure 10B is formed between the adsorption regions that are generating suction force and those that are not, and this boundary moves from the right end of the first adsorption surface 11s to the left end of the first adsorption surface 11s. This makes it possible to synchronize the decrease in the area in contact between the first chuck 11 and the first substrate W1 with the increase in the area in contact between the first substrate W1 and the second substrate W2.

[0214] The alignment completion state may be a state in which the bonding surface WA1 of the first substrate W1 is elastically deformed into a shape such as a downwardly convex spherical crown, and the entire area of ​​the bonding surface WA2 of the second substrate W2 is positioned within a single horizontal plane. In this case, the alignment completion state may be a state in which the misalignment of the first substrate W1 and the second substrate W2 is minimized when the entire bonding surface WA1 of the first substrate W1 returns to a flat plane.

[0215] Specifically, if the bonding surface WA1 of the first substrate W1 is elastically deformed and the shape of the bonding surface WA1 of the first substrate W1 in the alignment-completed state can be estimated, the position of the first substrate W1 in the alignment-completed state may be corrected according to the estimated shape. Regardless of whether or not the shape of the bonding surface WA1 of the first substrate W1 in the alignment-completed state can be estimated, the shape of the bonding surface WA1 of the first substrate W1 in the alignment-completed state may be measured, and the position of the first substrate W1 in the alignment-completed state may be corrected according to the measured shape. For example, a distance meter such as a laser distance meter may be used to measure the vertical distance from one horizontal plane to multiple positions within the bonding surface WA1 of the first substrate W1, and the position of the first substrate W1 in the alignment-completed state may be corrected according to the values ​​obtained from the measurement.

[0216] When changing the first substrate W1 and the second substrate W2 from the alignment completed state to the bonding start state, the lowest part of the bonding surface WA1 of the first substrate W1 is brought into contact with the bonding surface WA2 of the second substrate W2. Then, the distance between the first substrate W1 and the second substrate W2 is reduced while changing at least one of the following: the horizontal position of the first substrate W1 relative to the second substrate W2, the orientation of the first substrate W1 relative to the second substrate W2, the relative angle between the first substrate W1 and the second substrate W2 around a vertical line, the shape of the bonding surface WA1 of the first substrate W1, and the shape of the bonding surface WA2 of the second substrate W2. As a result, the entire area of ​​the bonding surface WA1 of the first substrate W1 and the entire area of ​​the bonding surface WA2 of the second substrate W2 come into contact, and the first substrate W1 and the second substrate W2 change from the bonding start state to the bonding expansion state and then to the bonding completed state.

[0217] Next, the effects of this embodiment will be described.

[0218] In this embodiment, the bonding unit 2b of the substrate bonding apparatus 1 not only bonds the first substrate W1 and the second substrate W2, but also inspects the bonding accuracy of the bonded first substrate W1 and second substrate W2. Therefore, it is not necessary to move the first substrate W1 and the second substrate W2 from the bonding unit 2b before inspecting the bonding accuracy of the first substrate W1 and the second substrate W2. This makes it possible to shorten the transport time of the first substrate W1 and the second substrate W2.

[0219] In this embodiment, the second substrate W2 is held in the second chuck 21 before it is bonded to the first substrate W1, and is also held in the second chuck 21 when the bonded first substrate W1 and second substrate W2 are photographed by the inspection camera 43. Therefore, it is not necessary to move the second substrate W2 from the second chuck 21 before inspecting the bonding accuracy of the first substrate W1 and second substrate W2. This makes it possible to shorten the transport time of the first substrate W1 and second substrate W2.

[0220] In this embodiment, the bonding unit 2b is located inside the outer wall 1a of the substrate bonding apparatus 1, and the first chuck 11, the second chuck 21, and the inspection camera 43 are located inside the chamber 9 of the bonding unit 2b. Therefore, it is not necessary to move the first substrate W1 and the second substrate W2 outside the chamber 9 before inspecting the bonding accuracy of the first substrate W1 and the second substrate W2. This makes it possible to shorten the transport time of the first substrate W1 and the second substrate W2.

[0221] In this embodiment, the first chuck 11, the second chuck 21, and the inspection camera 43 are positioned above the base 20 of the bonding unit 2b and overlap the base 20 in a plan view. In other words, the first chuck 11, the second chuck 21, and the inspection camera 43 are positioned directly above the base 20. Before inspecting the bonding accuracy of the first substrate W1 and the second substrate W2, it is not necessary to move the first substrate W1 and the second substrate W2 from the space directly above the base 20. This makes it possible to shorten the transport time of the first substrate W1 and the second substrate W2.

[0222] In this embodiment, the relative displacement between the base 20 and the second chuck 21 is detected by a grid interferometer-type displacement detector 46, rather than a laser interferometer. Compared to interferometers such as laser interferometers, the grid interferometer-type displacement detector 46 is less affected by environmental changes such as temperature and humidity changes, and can stably detect displacement. Furthermore, the displacement detector 46 detects not only the relative displacement between the base 20 and the second chuck 21 in the Y direction, but also the relative displacement between the base 20 and the second chuck 21 in the X direction. Therefore, compared to the case where separate detectors are provided for the Y direction and the X direction, the relative displacement between the base 20 and the second chuck 21 can be detected with higher accuracy.

[0223] In this embodiment, the displacement detector 46 detects the displacement of the second chuck 21 in the X and Y directions. The first chuck 11 is supported by a first frame 19 fixed to the base 20. By stopping the operation of the inversion actuator 16, the first chuck 11 can be held in a fixed position in the X and Y directions. Therefore, the relative positions of the first chuck 11 and the second chuck 21 in the X and Y directions can be adjusted without detecting the displacement of the first chuck 11 in the X and Y directions.

[0224] In this embodiment, the displacement of the scale 47, which moves in the X and Y directions together with the second chuck 21, is detected by the detection head 48. This makes it possible to detect the displacement of the second chuck 21. Furthermore, since the detection head 48 is not in contact with the scale 47, the amount of particles can be reduced compared to when there is contact. In addition, since the scale 47 moves together with the second chuck 21 rather than the detection head 48, structural complexity can be prevented or reduced compared to when the detection head 48 moves together with the second chuck 21.

[0225] In this embodiment, the scale 47 overlaps the second chuck 21 in a plan view. Therefore, the footprint (area of ​​the object when viewed from directly above) of the second chuck 21 and the scale 47 can be reduced. Furthermore, the detection head 48 overlaps the scale 47 in a plan view. Therefore, the footprint of the second chuck 21, the scale 47, and the detection head 48 can be reduced. This makes it possible to miniaturize the substrate bonding apparatus 1.

[0226] In this embodiment, power from two Z actuators 18 is transmitted to the first chuck 11 via two driven bodies. Two ball nuts 18n are an example of two driven bodies. The two driven bodies correspond one-to-one with the two Z actuators 18 and move in the Z direction together with the first chuck 11 along two parallel and vertical lines. The two driven bodies can move independently in the Z direction. By adjusting the positions of the two driven bodies in the Z direction using the two Z actuators 18, the orientation of the first substrate W1 held in the first chuck 11 can be controlled. This stabilizes the orientation of the first substrate W1.

[0227] In this embodiment, the inversion actuator 16 inverts the first substrate W1, which is held in the first chuck 11. Therefore, the first substrate W1 is held in the first chuck 11 before inversion and is also held in the first chuck 11 when it is joined with the second substrate W2. Thus, it is not necessary to move the first substrate W1 from the first chuck 11 between inversion and joining with the second substrate W2. Furthermore, the two Z actuators 18 move not only the two driven bodies and the first chuck 11 but also the inversion actuator 16 in the Z direction. Therefore, the orientation of the inverted first substrate W1 can be adjusted.

[0228] In this embodiment, the two driven bodies move in the Z direction together with the first chuck 11 along two parallel and vertical lines. The reversing actuator 16 rotates the first substrate W1 around a horizontal line passing through these two lines. The two Z actuators 18 tilt the centerline 11c of the first chuck 11 in the Y direction. The reversing actuator 16 tilts the centerline 11c of the first chuck 11 in the X direction. Therefore, the orientation of the first substrate W1 can be adjusted more flexibly.

[0229] In this embodiment, the first chuck 11 is guided in the Z direction by at least one linear guide 17. This reduces the mechanical play of the mechanism supporting the first chuck 11 and makes the orientation of the first substrate W1 more stable.

[0230] In this embodiment, the first chuck 11 and the second chuck 21 are moved relative to each other while the alignment camera 45 photographs the first fixed reference 31r and the second fixed reference 32r. The first fixed reference 31r does not overlap with the first substrate W1 held by the first chuck 11, and the second fixed reference 32r does not overlap with the second substrate W2 held by the second chuck 21. Furthermore, the first display plate 31d of the first fixed reference 31r is a transparent plate that transmits visible light. Therefore, the alignment camera 45, which is a visible light camera, can photograph the first reference mark 31m and the second reference mark 32m by photographing the second fixed reference 32r through the first fixed reference 31r.

[0231] The first reference mark 31m is an indirect reference when aligning the first substrate W1 held in the first chuck 11. Similarly, the second reference mark 32m is an indirect reference when aligning the second substrate W2 held in the second chuck 21. The first reference mark 31m moves with the first substrate W1, and the second reference mark 32m moves with the second substrate W2. Therefore, by adjusting the relative positions of the first reference mark 31m and the second reference mark 32m, the relative positions of the first substrate W1 and the second substrate W2 can be adjusted. This makes it possible to adjust the alignment of the first substrate W1 and the second substrate W2 without photographing the overlapping first substrate W1 and the second substrate W2.

[0232] In this embodiment, the first camera 41 captures not only the first reference mark 31m but also the first alignment mark AM1 of the first substrate W1 held in the first chuck 11. Similarly, the second camera 42 captures not only the second reference mark 32m but also the second alignment mark AM2 of the second substrate W2 held in the second chuck 21. The relative position and angle between the first alignment mark AM1 and the first reference mark 31m can be determined from the image captured by the first camera 41. Similarly, the relative position and angle between the second alignment mark AM2 and the second reference mark 32m can be determined from the image captured by the second camera 42. This allows the alignment of the first substrate W1 and the second substrate W2 to be adjusted using the first reference mark 31m and the second reference mark 32m.

[0233] In this embodiment, with the first substrate W1 and the second substrate W2 each in a flat state, the bonding surface WA1 of the first substrate W1 is tilted relative to the bonding surface WA2 of the second substrate W2. In this state, the bonding surface WA1 of the first substrate W1 is brought into partial contact with the bonding surface WA2 of the second substrate W2. Subsequently, by reducing the tilt angle of the bonding surface WA1 of the first substrate W1 relative to the bonding surface WA2 of the second substrate W2, the area of ​​the region in contact between the bonding surface WA1 of the first substrate W1 and the bonding surface WA2 of the second substrate W2 is increased. Air is discharged from between the bonding surface WA1 of the first substrate W1 and the bonding surface WA2 of the second substrate W2 during this process. This eliminates or reduces air bubbles remaining between the two bonded substrates W. Furthermore, since the first substrate W1 and the second substrate W2 are joined while both are flat, there is no need for a structure or mechanism to deform at least one of the first substrate W1 or the second substrate W2.

[0234] In this embodiment, the alignment adjustment of the first substrate W1 and the second substrate W2 is performed with the entire bonding surface WA1 of the first substrate W1 held by the first chuck 11 parallel to the entire bonding surface WA2 of the second substrate W2 held by the second chuck 21. Therefore, the relative position and angle of the first substrate W1 and the second substrate W2 can be adjusted more precisely than when the alignment adjustment is performed with the first substrate W1 and the second substrate W2 not parallel. After the alignment adjustment is performed, the first substrate W1 and the second substrate W2 are bonded together as described above. Therefore, the first substrate W1 and the second substrate W2 can be bonded together with higher bonding accuracy while eliminating or reducing air bubbles remaining between the two bonded substrates W.

[0235] In this embodiment, the first camera 41 captures the first alignment mark AM1 of the first substrate W1. If the first alignment mark AM1 of the first substrate W1 has a shape that allows for the identification of mutually orthogonal height and width directions, it is determined whether the tilt angle of the first alignment mark AM1 with respect to the vertical or horizontal direction of the image generated by the first camera 41 exceeds a first upper limit value (for example, a value greater than 0 and less than or equal to 5). If it is determined that the tilt angle of the first alignment mark AM1 exceeds the first upper limit value, the first substrate W1 and the first chuck 11 are rotated by the first rotation motor 12 so that the tilt angle of the first alignment mark AM1 decreases to or less than the first upper limit value. In this way, the shape of the first alignment mark AM1 can be identified with higher accuracy from the image generated by the first camera 41, and the position and angle of the first substrate W1 can be identified with higher accuracy. Therefore, alignment adjustment of the two substrates W can be performed with higher accuracy.

[0236] In this embodiment, the first camera 41 captures the first alignment mark AM1 of the first substrate W1, and the second camera 42 captures the second alignment mark AM2 of the second substrate W2. The control device 3 receives the electronic data of the captured images and analyzes and makes decisions about the images based on a program stored in memory 3c. If the second alignment mark AM2 of the second substrate W2 has a shape that allows for the identification of mutually orthogonal height and width directions, the control device 3 determines whether the tilt angle of the second alignment mark AM2 with respect to the vertical or horizontal direction of the image generated by the second camera 42 exceeds a second upper limit value (for example, a value greater than 0 and less than or equal to 5). When the control device 3 determines that the tilt angle of the second alignment mark AM2 exceeds the second upper limit value, the control device 3 rotates the second substrate W2 and the second chuck 21 with the second rotation motor 22 so that the tilt angle of the second alignment mark AM2 decreases to less than or equal to the second upper limit value. In this way, not only the position and angle of the first substrate W1, but also the position and angle of the second substrate W2 can be determined with higher accuracy.

[0237] In this embodiment, the inversion actuator 16 inverts the first substrate W1 held in the first chuck 11. Furthermore, at least one Z actuator 18 moves the first chuck 11 and the second chuck 21 relative to each other in the Z direction, thereby joining the first substrate W1 held in the first chuck 11 and the second substrate W2 held in the second chuck 21. The first substrate W1 is held in the first chuck 11 before inversion and is still held in the first chuck 11 when it is joined with the second substrate W2. Therefore, it is not necessary to move the first substrate W1 from the first chuck 11 between inversion and joining with the second substrate W2. This makes it possible to shorten the transport time of the first substrate W1 and the second substrate W2.

[0238] In this embodiment, horizontal actuators such as the X actuator 27 and the Y actuator 25 move the second chuck 21 horizontally relative to the first chuck 11. The horizontal actuators move not only the second chuck 21 but also the first camera 41 horizontally. The first camera 41 photographs the first substrate W1, which is held inverted by the first chuck 11. The horizontal actuators also serve as shooting position adjustment actuators that change the position photographed by the first camera 41. Therefore, the number of components in the substrate bonding device 1 can be reduced compared to the case where a separate shooting position adjustment actuator is provided in addition to the horizontal actuators.

[0239] In this embodiment, after joining the first substrate W1 held in the first chuck 11 and the second substrate W2 held in the second chuck 21, the first substrate W1 and the second substrate W2 are held in the second chuck 21, and the first substrate W1 and the second substrate W2 held in the second chuck 21 are photographed by the inspection camera 43. While this photography is taking place, the second first substrate W1 is transported to the first chuck 11 by the transport robot TR. Therefore, multiple pairs of first substrates W1 and second substrates W2 can be joined in a shorter time compared to the case where the joined first substrates W1 and second substrates W2 are unloaded from the second chuck 21 before the second first substrate W1 is transported to the first chuck 11.

[0240] In this embodiment, not only is the second first substrate W1 transported to the first chuck 11 while the inspection camera 43 is photographing the first substrate W1 and the second substrate W2, but the second first substrate W1 held in the first chuck 11 is also inverted while the inspection camera 43 is photographing the first substrate W1 and the second substrate W2. Therefore, multiple pairs of first substrates W1 and second substrates W2 can be joined in a shorter time compared to the case where the second first substrate W1 is inverted after the inspection camera 43 has finished photographing.

[0241] In this embodiment, horizontal actuators such as the X actuator 27 and the Y actuator 25 move the second chuck 21 horizontally relative to the first chuck 11 between the joining position and the imaging position. The joining position is the position where the first substrate W1 held by the first chuck 11 and the second substrate W2 held by the second chuck 21 are joined. The imaging position is the position where the inspection camera 43 images the first substrate W1 and the second substrate W2 held by the second chuck 21. The horizontal actuator also serves as an imaging position adjustment actuator that changes the position captured by the inspection camera 43. Therefore, the number of components in the substrate joining device 1 can be reduced compared to the case where a separate imaging position adjustment actuator is provided in addition to the horizontal actuator.

[0242] Next, other embodiments will be described.

[0243] The first substrate W1 may be the lower substrate W, not the upper substrate W. In other words, the upper substrate W may be the second substrate W2, and the lower substrate W may be the first substrate W1.

[0244] The inspection camera 43 may be fixed to the second frame 28 via a bracket different from the bracket 44 that holds the second narrow-field camera 42n and the second wide-field camera 42w. The inspection camera 43 may also be fixed to the base 20 via a member other than the second frame 28, such as the first frame 19. For example, the inspection camera 43 may be fixed to the first stage 13. In this case, the travel distance of the first substrate W1 and the second substrate W2 when positioning them within the range that the inspection camera 43 can capture can be shortened compared to when the inspection camera 43 is fixed to the second frame 28.

[0245] The first camera 41 does not need to photograph the first fixed reference 31r every time it photographs the first substrate W1 held in the first chuck 11. For example, the position and angle of the first substrate W1 relative to the first fixed reference 31r may be determined using an image of the first fixed reference 31r that has been photographed in advance. However, since the conditions (temperature, humidity, etc.) when the first fixed reference 31r is photographed may differ from the conditions when the first substrate W1 is photographed, it is preferable to photograph the first fixed reference 31r every time the first substrate W1 held in the first chuck 11 is photographed in order to detect the position and angle of the first substrate W1 relative to the first fixed reference 31r with greater precision. The same applies to photographing the second fixed reference 32r.

[0246] If the first alignment mark AM1 is photographed before the first substrate W1 is held in the first chuck 11, it is not necessary to photograph the first alignment mark AM1 while the first substrate W1 is held in the first chuck 11. In other words, if the position and angle of the first substrate W1 are known in advance and the first substrate W1 is held in the first chuck 11, it is not necessary to photograph the first alignment mark AM1. In this case, the first camera 41 may be omitted from the bonding unit 2b. The same applies to the second camera 42.

[0247] The actuator mechanism AC is provided on the first stage 13 and may further include one or more piezo actuators that move the first chuck 11 in the Z direction. In this case, the control device 3 may lower the first substrate W1 with the Z actuator 18 and hold the first substrate W1 in a state where the downward-facing bonding surface WA1 of the first substrate W1 maintains a small gap with the upward-facing bonding surface WA2 of the second substrate W2. Subsequently, the control device 3 may drive one or more piezo actuators to bring the bonding surface WA1 of the first substrate W1 into contact with the bonding surface WA2 of the upward-facing second substrate W2.

[0248] The substrate bonding apparatus 1 is not limited to an apparatus for bonding two disc-shaped substrates W, but may also be an apparatus for bonding two polygon-shaped substrates W.

[0249] You may combine two or more of the aforementioned components. You may also combine two or more of the aforementioned processes.

[0250] The base 20 is an example of a base member having a main surface that extends along the horizontal direction. The first frame 19 is an example of a support part that is fixed to the base member and supports the first chuck 11. The first camera 41 is an example of a first imaging unit that moves in the X and Y directions together with the second chuck 21 and images the first substrate W1. The second camera 42 is an example of a second imaging unit that is fixed in the horizontal direction and images the second substrate W2. The pair of support bases 15 are examples of a first connection part and a second connection part that are provided sandwiching the first substrate W1 in the Y direction perpendicular to the Z direction.

[0251] Although embodiments of the present invention have been described in detail, these are merely specific examples used to clarify the technical content of the invention, and the present invention should not be construed as being limited to these examples. The spirit and scope of the present invention are limited only by the appended claims. [Explanation of symbols]

[0252] 1: Substrate bonding apparatus, 1a: Outer wall, 2: Processing unit, 2b: Bonding unit, 2b1: First bonding module, 2b2: Second bonding module, 2c: Cleaning unit, 2h: Hydrophilization unit, 3: Control device, 3a: Computer main unit, 3b: CPU, 3c: Memory, 3d: Peripheral device, 3e: Storage, 3f: Reader, 3g: Communication device, 9: Chamber, 9p: Partition wall, 9s: Shutter, 10: Vibration isolation table, 11: First chuck, 11c: Centerline, 11p: Suction port, 11s: First suction surface, 11v: Valve, 12: First rotation motor, 13: First stage, 13 h: Through hole, 14: Support shaft, 15: Support base, 16: Reversing actuator, 16c: Reversing center, 17: Linear guide, 17L: Rail, 17b: Slide block, 18: Z actuator, 18c: Centerline, 18n: Ball nut, 18s: Ball screw, 19: First frame, 19s: First side frame, 19u: First upper frame, 20: Base, 21: Second chuck, 21c: Centerline, 21s: Second suction surface, 22: Second rotation motor, 23: Second stage, 23L: Lower stage, 23u: Upper stage, 23w: Through window, 2 4: Linear guide, 24L: Rail, 24b: Slide block, 25: Y actuator, 25f: Fixed magnet, 25m: Movable magnet, 26: Linear guide, 26L: Rail, 26b: Slide block, 27: X actuator, 27f: Fixed magnet, 27m: Movable magnet, 28: Second frame, 28s: Second side frame, 28u: Second upper frame, 31L: First fixed line, 31b: First support block, 31d: First display board, 31m: First reference mark, 31p: First fixed point, 32L: Second fixed line, 32b: Second support block, 32d: 2 display board, 32m: second reference mark, 32p: second fixed point, 41: first camera, 41n: first narrow-field camera, 41w: first wide-field camera, 42: second camera, 42n: second narrow-field camera, 42w: second wide-field camera, 43: inspection camera, 44: bracket, 44h: through hole, 45: alignment camera, 46: displacement detector, 47: scale, 47s: reflective surface, 48: detection head, 48e: light-emitting part, 48i: intermediate, 48m: main unit, 48r: light-receiving element, 48s: light source, 48x: X detection head, 48y: Y detection head, AC: actuator mechanismAG1: Center of gravity, AG2: Center of gravity, AH1: Horizontal axis, AH2: Horizontal axis, AM1: First alignment mark, AM2: Second alignment mark, AP1: Reference point, AP2: Reference point, AV1: Vertical axis, AV2: Vertical axis, CA: Carrier, LP: Load port, LP1: First load port, LP2: Second load port, LP3: Third load port, P: Program, RM: Removable media, S1~S8: Step, TH: Hand, TP: Transport path, TR: Transport robot, TS: Transport system, W: Substrate, W1: First substrate, W2: Second substrate, WA1: Bonding surface, WA2: Bonding surface, WB1: Bonding layer, WB2: Bonding layer, WC1: Device layer, WC2: Device layer, WD1: Substrate, WD2: Substrate

Claims

1. A first chuck that holds the first substrate, A second chuck that holds the second substrate, An actuator mechanism that moves the first chuck and the second chuck relative to each other to join the first substrate held by the first chuck and the second substrate held by the second chuck, An inspection camera for photographing the first and second substrates held in the second chuck after the first and second substrates have been joined together, A substrate bonding apparatus comprising: a transport robot that transports a second first substrate to the first chuck while the inspection camera is photographing the first substrate and the second substrate held in the second chuck.

2. The substrate bonding apparatus according to claim 1, wherein the actuator mechanism includes a reversing actuator that is transported to the first chuck by the transport robot and reverses the second first substrate held in the first chuck when the inspection camera is photographing the first substrate and the second substrate held in the second chuck.

3. The substrate bonding apparatus according to claim 1 or 2, wherein the actuator mechanism further includes a horizontal actuator that moves the second chuck horizontally relative to the first chuck between a bonding position in which the first substrate held by the first chuck and the second substrate held by the second chuck are joined, and a shooting position in which the inspection camera photographs the first substrate and the second substrate held by the second chuck.

4. A step of holding the first substrate with a first chuck, A step of holding the second substrate with a second chuck, A step of joining a first substrate held by the first chuck and a second substrate held by the second chuck by moving the first chuck and the second chuck relative to each other, A step of releasing the first substrate from the first chuck, A step of having the first substrate and the second substrate, which are held in the second chuck after the first substrate and the second substrate are joined together, photographed by an inspection camera, A substrate bonding method comprising the step of transporting a second first substrate to the first chuck by a transport robot while the inspection camera is photographing the first substrate and the second substrate held in the second chuck.

Images