Alignment mark and method for using same

Abstract

This alignment mark comprises a first mark part and a second mark part. The mark parts have line segment elements having the same length and extending in a direction inclined relative to a first direction in a virtual plane when viewed in the direction perpendicular to the virtual plane. The second mark part is formed such that the line segment element thereof when viewed in the direction perpendicular to the virtual plane has a shape obtained by inverting the line segment element of the first mark part with a virtual line extending in a second direction perpendicular to the first direction in the virtual plane as an axis. The first mark part and the second mark part are formed such that the middle points of the line segment elements mutually coincide when the relative positional relationship between two workpieces becomes a prescribed positional relationship. Each of the line segment elements is formed such that, in the case where a component in the first direction is defined as a first component and a component in the second direction is defined as a second component, the length Lb of the second component is greater than the length La of the first component.

Description

Alignment Mark and Method of Using the Same

[0001] The present invention relates to an alignment technique for chips and substrates.

[0002] Patent Document 1 discloses a technique for joining a chip (such as an LSI or IC) and a substrate (such as a wafer). After activating the bonding surfaces of the chip and the substrate with plasma, the chip is pressed against the surface of the substrate for bonding.

[0003] Japanese Patent Application Laid-Open No. 2018-160687

[0004] In recent years, with the progress of miniaturization technology in the semiconductor field and the progress of integration technology for achieving high capacity beyond the limits of miniaturization, in alignment technology, higher precision (on the order of 0.01 μm or more) has been required for the alignment accuracy of the chip with respect to the bonding position than before. For this reason, it has become necessary to detect the amount of deviation of the chip with respect to the bonding position with high precision.

[0005] On the other hand, in conventional alignment technology, a vernier formed by using photolithography technology on a chip to be joined and a substrate (the portion of the substrate corresponding to the bonding position of the chip) has been used to detect the relative amount of deviation between them. However, in such a vernier, the resolution (detection accuracy of the amount of deviation) depends on the accuracy of photolithography, and due to this, the limit of the resolution becomes on the order of 0.1 μm, so it is difficult to meet the accuracy (on the order of 0.01 μm or more) required in recent years.

[0006] Therefore, an object of the present invention is to provide a new technique that enables alignment with higher precision than a vernier.

[0007] The alignment mark according to the present invention is an alignment mark that, for a first workpiece and a second workpiece that overlap each other when viewed from a direction perpendicular to a two-dimensional imaginary plane, enables the relative positional relationship between the workpieces to be adjusted to a predetermined positional relationship and enables detection of a relative misalignment amount ΔL in a first direction within the imaginary plane, and has the following configuration (Aspect 1). The alignment mark includes a first mark portion formed on the first workpiece and a second mark portion formed on the second workpiece. The first mark portion and the second mark portion each have line segment elements of the same length that extend in a direction inclined from the first direction within the imaginary plane when viewed from a direction perpendicular to the imaginary plane. The second mark portion is formed so that the shape of the line segment element when viewed from a direction perpendicular to the imaginary plane is a shape obtained by inverting the line segment element of the first mark portion around a virtual line that extends in a second direction perpendicular to the first direction within the imaginary plane. The first mark portion and the second mark portion are formed so that the midpoints of the line segment elements coincide with each other when the relative positional relationship between the first workpiece and the second workpiece is a predetermined positional relationship. Furthermore, each line segment element is formed so that, when a component in the first direction is defined as a first component and a component in the second direction is defined as a second component, the length Lb of the second component is greater than the length La of the first component.

[0008] According to the above-mentioned aspect 1, it is possible to use the alignment mark in the following method (aspect 2). First, of the two ends of the line element of the first mark portion and the line element of the second mark portion, the ends that approach each other when the first workpiece is displaced in the first direction relative to the second workpiece from a predetermined positional relationship are set as the reference end. Next, for the portion of the line element of the first mark portion from the intersection with the line element of the second mark portion when viewed from a direction perpendicular to the imaginary plane to the reference end, the component in the second direction is set as the third component, and the length Lc of the third component can be used to calculate the relative displacement ΔL in the first direction by the formula ΔL = (Lb / 2 - Lc) × La / Lb.

[0009] According to this method of use (aspect 2), it is possible to measure the misalignment amount ΔL in the first direction by expanding the width within the alignment mark by a factor of Lb / La, and as a result, it is possible to detect the misalignment amount ΔL with a resolution of a factor of Lb / La. In other words, it is possible to detect the positional misalignment in the first direction with high sensitivity.

[0010] In the alignment mark according to aspect 1 above, each line element may be formed so that the length Lb of the second component is 10 times or more the length La of the first component, depending on the resolution required to detect the deviation amount ΔL in the first direction (aspect 3).

[0011] According to the above-mentioned aspect 3, while the limit of resolution (detection accuracy of the amount of deviation ΔL) of a vernier formed using photolithography technology is on the order of 0.1 μm, it is possible to detect the amount of deviation with even higher resolution (on the order of 0.01 μm or more) that exceeds that order.

[0012] In the method of use according to the above aspect 2, the positional relationship between the first workpiece and the second workpiece in the first direction may be adjusted to approach a predetermined positional relationship based on the deviation amount ΔL calculated using the above formula (aspect 4).

[0013] According to the above-mentioned aspect 4, it becomes possible to feed back the deviation amount ΔL when adjusting (aligning) the position of the second workpiece relative to the first workpiece, and as a result, it becomes possible to efficiently bring the positional relationship of the works closer to a predetermined positional relationship.

[0014] The method of use according to Aspect 2 above may have the following configuration (Aspect 5). Two axial directions constituting an imaginary plane may be defined as the X-axis direction and the Y-axis direction. Alignment marks may be formed on the first workpiece and the second workpiece, with the X-axis direction being the first direction, and additional alignment marks may be formed separately with the Y-axis direction being the first direction. These alignment marks may then be used to calculate both the amount of deviation ΔL in the X-axis direction and the amount of deviation ΔL in the Y-axis direction.

[0015] According to the above-mentioned aspect 5, it is possible to perform position adjustment (alignment) in a two-dimensional plane (virtual plane) with high precision.

[0016] In the method of use according to the above-mentioned aspect 5, the positional relationship between the first workpiece and the second workpiece in the virtual plane may be adjusted to approach a predetermined positional relationship based on both the shift amount ΔL in the X-axis direction and the shift amount ΔL in the Y-axis direction (aspect 6).

[0017] According to the above-mentioned aspect 6, when adjusting (aligning) the position of the second workpiece relative to the first workpiece, it is possible to feed back the deviation amount ΔL in two directions (X-axis direction and Y-axis direction), and as a result, it is possible to efficiently and accurately bring the positional relationship of the works in a two-dimensional plane (virtual plane) closer to a predetermined positional relationship.

[0018] According to the present invention, alignment with higher accuracy than that achieved by a vernier is possible.

[0019] FIG. 1 is a conceptual diagram showing a bonding apparatus according to an embodiment. FIG. 2 is a conceptual diagram showing the configuration of each point of an attitude adjustment mechanism that supports a piezo stage at three points. FIG. 3(A) is a conceptual diagram illustrating the configuration of an alignment mark used in the embodiment, and FIG. 3(B) is a conceptual diagram illustrating the state of the alignment mark when the alignment target is in a predetermined positional relationship. FIG. 4(A) is a conceptual diagram illustrating the state of the alignment mark when the first workpiece is relatively displaced in a first direction (positive direction), and FIG. 4(B) is a conceptual diagram illustrating the state of the alignment mark when the first workpiece is relatively displaced in the direction opposite to the first direction (negative direction). FIG. 5(A) is a conceptual diagram showing a modified shape of the alignment mark, and FIG. 5(B) is a conceptual diagram showing a method of using the alignment mark. FIGS. 6(A) and 6(B) are conceptual diagrams showing two example arrangements of the alignment mark. 7A and 7B are conceptual diagrams showing other modified shapes of alignment marks. 8A and 8B are conceptual diagrams showing changes in the state of the bonding device caused by the execution of a bonding process. 9A and 9B are conceptual diagrams showing changes in the state of the bonding device caused by the execution of a bonding process.

[0020] 1 is a conceptual diagram showing a bonding apparatus according to an embodiment. This bonding apparatus presses a chip (such as an LSI or IC, hereinafter referred to as a "first workpiece W1") against the surface of a substrate (such as a wafer, hereinafter referred to as a "second workpiece W2") to bond them together, and includes a holding mechanism 1, a pressing mechanism 2, an alignment stage 3, a detection sensor 4, and a control device 5.

[0021] The holding mechanism 1 includes a sheet holding unit 11 , a table 12 , and a drive unit 13 .

[0022] The sheet holding unit 11 is a part that holds the peripheral edge of an elastically deformable sheet Sw (such as a dicing tape) in a tensioned state. In this embodiment, such a sheet Sw is prepared with a plurality of first workpieces W1 attached to one side of the sheet Sw, and is held by the sheet holding unit 11 with the one side facing downward.

[0023] The table 12 is a portion that holds the second workpiece W2 below the sheet Sw held by the sheet holding unit 11, and is equipped with a chuck function (such as vacuum suction or electrostatic chuck) for fixing the second workpiece W2. A circuit is formed on the second workpiece W2, and a joining position Pt for the first workpiece W1 is set on the surface of the second workpiece W2 at a predetermined position corresponding to the circuit (specifically, a predetermined position that can be connected to the circuit). Therefore, the second workpiece W2 is held on the table 12 with the joining surface with the first workpiece W1 (the surface where the joining position Pt of the first workpiece W1 is provided) facing upward.

[0024] The drive unit 13 is a mechanism that enables the table 12 to move up and down and adjust its position in a horizontal plane, and is composed of, for example, a ball screw and a motor. By adjusting the position of the table 12 using the drive unit 13 in this manner, when joining one workpiece (hereinafter referred to as the "target workpiece Wt") selected from the multiple first workpieces W1 attached to the sheet Sw at the corresponding joining position Pt, it is possible to perform a position adjustment (position adjustment that does not require high precision) to position the joining position Pt opposite the target workpiece Wt. After such a position adjustment, the position of the target workpiece Wt is adjusted (aligned) with high precision relative to the joining position Pt using a piezo stage 31, which will be described later.

[0025] In this embodiment, the joining device is configured to be able to press a plurality of first workpieces W1 attached to a sheet Sw onto the surface of a second workpiece W2 one by one, and join them together. Specifically, this is as follows.

[0026] The pressing mechanism 2 includes a pressing portion 21 , a peripheral holding portion 22 , and driving portions 23 and 24 .

[0027] The pressing unit 21 is a part that presses the target workpiece Wt attached to the sheet Sw against the surface of the second workpiece W2. Specifically, the pressing unit 21 has an adsorption surface 21p at its tip (lower end). In this embodiment, this adsorption surface 21p is configured to adsorb the back surface of the sheet Sw by vacuum suction. The pressing unit 21 is also configured to be able to move up and down by the power of the drive unit 23. More specifically, the pressing unit 21 is configured to be able to move up and down independently of the peripheral holding unit 22 described below.

[0028] With this pressing unit 21, the back surface of the area of ​​the sheet Sw where the target workpiece Wt is attached is attached to the suction surface 21p, so that the target workpiece Wt can be held by the suction surface 21p via the sheet Sw (see FIG. 8B). Also, by lowering the pressing unit 21 to bring the suction surface 21p closer to the surface of the second workpiece W2, the target workpiece Wt can be pressed against the surface of the second workpiece W2 (see FIG. 9B).

[0029] The peripheral holding unit 22 is a portion that holds the peripheral area of ​​the area of ​​the sheet Sw to which the target workpiece Wt is attached when the target workpiece Wt is pressed. Specifically, the peripheral holding unit 22 has an adsorption surface 22p at its tip (lower end). In this embodiment, this adsorption surface 22p is configured to adsorb the back surface of the sheet Sw by vacuum suction. In addition, the peripheral holding unit 22 is configured to be able to move up and down by the power of the drive unit 24.

[0030] With this peripheral holding unit 22, when the pressing unit 21 presses the target workpiece Wt, the periphery of the area of ​​the sheet Sw to which the target workpiece Wt is attached can be held in place by the suction surface 22p (see FIG. 9B). This allows the sheet Sw to bend appropriately inside the suction surface 22p when the target workpiece Wt is pressed, which in turn makes it easier to peel the target workpiece Wt from the sheet Sw.

[0031] The alignment stage 3 is a combination of multiple stages to enable high-speed and high-precision position adjustment of the pressing mechanism 2. Specifically, the alignment stage 3 includes a piezo stage 31 and a wide-area stage 32 that allows position adjustment over a wider range than the piezo stage 31. The piezo stage 31 is mounted on the wide-area stage 32, and the pressing mechanism 2 is mounted on the piezo stage 31. The piezo stage 31 uses a piezo element as its stage drive source, allowing for high-precision (nanometer-order precision) stage position adjustment. On the other hand, the wide-area stage 32 uses a ball screw and a motor as its stage drive source, allowing for high-speed movement of the stage over a wide range.

[0032] With this alignment stage 3, the position of the pressing mechanism 2 can be adjusted in two stages using the wide-area stage 32 and the piezo stage 31. Specifically, the wide-area stage 32 allows the pressing mechanism 2 to be moved at high speed over a wide range without requiring high accuracy, and then the piezo stage 31 allows the position of the pressing mechanism 2 to be adjusted with high accuracy (nanometer-order accuracy).

[0033] Therefore, when aligning the target workpiece Wt with the joining position Pt provided on the surface of the second workpiece W2, the wide stage 32 enables the pressing mechanism 2 to be moved at high speed to a position where the target workpiece Wt can be held from the back side (in other words, to a position above the corresponding joining position Pt).

[0034] Furthermore, because the sheet Sw is elastically deformable, even after the target workpiece Wt is held on the suction surface 21p via the sheet Sw, the position of the pressing mechanism 2 (i.e., the position of the target workpiece Wt) can be adjusted by the piezo stage 31 while the sheet Sw is elastically deforming. Therefore, the target workpiece Wt can be aligned with respect to the joining position Pt with high precision by the piezo stage 31 without being affected by the sheet Sw. This makes it possible to significantly improve the accuracy of alignment with respect to the joining position Pt.

[0035] In this way, the joining device of this embodiment can achieve both high-speed movement of the pressing mechanism 2 over a wide range and high-precision position adjustment of the target workpiece Wt relative to the joining position Pt. As a result, even if the joining positions Pt are scattered over a wide range, it is possible to quickly and highly accurately align the target workpiece Wt relative to the joining position Pt one by one.

[0036] In this embodiment, the alignment stage 3 further includes an attitude adjustment mechanism 33 (see FIG. 2) that enables adjustment of the attitude of the piezo stage 31 relative to the wide stage 32. Specifically, the attitude adjustment mechanism 33 is configured to support the piezo stage 31 relative to the wide stage 32 at three points.

[0037] 2 is a conceptual diagram showing the configuration at each point of the attitude adjustment mechanism 33 that supports the piezo stage 31 at three points. At each point, the attitude adjustment mechanism 33 is configured by erecting a piezo actuator 331, which can adjust the protrusion amount of a movable part 331A, on the wide stage 32, and supporting the piezo stage 31 by bringing the tip of the movable part 331A into contact with the back surface of the piezo stage 31. The example of FIG. 2 also shows a case where a main body 331B of the piezo actuator 331 is connected to the piezo stage 31 via an annular leaf spring 332 in order to stabilize the support state of the piezo stage 31 by the piezo actuator 331.

[0038] Such an attitude adjustment mechanism 33 makes it possible to adjust the attitude of the piezo stage 31 so that the piezo stage 31 is parallel to the wide stage 32 .

[0039] Furthermore, when the target workpiece Wt is pressed against the surface of the second workpiece W2, a force equivalent to the pressure generated at that time (a reaction force acting on the piezo stage 31 at that time) is concentrated at three points (piezo actuators 331) supporting the piezo stage 31. Therefore, the posture adjustment mechanism 33 is equipped with pressure sensors 333 that measure the force (a force equivalent to the pressure) applied to the piezo stage 31 at three points (see FIG. 2). This allows the pressure sensors 333 to measure the pressure generated when the target workpiece Wt is pressed against the surface of the second workpiece W2.

[0040] The detection sensor 4 is a sensor that detects alignment marks 6 formed on the first workpiece W1 and the second workpiece W2.

[0041] In this embodiment, each first workpiece W1 and a portion of the second workpiece W2 where the joining position Pt of the first workpiece W1 is located constitute a pair of joining targets to be joined together. A pair of joining targets (the target workpiece Wt and the portion of the second workpiece W2 where the joining position Pt of the target workpiece Wt is located) selected from all pairs of joining targets is subjected to alignment before joining. During alignment, the relative positional relationship between the first workpiece W1 and the second workpiece W2 is adjusted (position adjustment that does not require high precision) by the drive unit 13 of the holding mechanism 1 so that the workpieces overlap (face each other) when viewed from a direction perpendicular to a two-dimensional virtual plane (a horizontal plane in this embodiment). Then, the piezo stage 31 performs high-precision position adjustment (alignment) to achieve a predetermined positional relationship (a state in which the position of the target workpiece Wt coincides with the joining position Pt when viewed from the vertical direction).

[0042] In this embodiment, an alignment mark 6 is formed on each pair of objects to be joined so that the above-described high-precision alignment can be performed for each pair of objects to be joined. The alignment mark 6 not only makes it possible to adjust the positional relationship between the two objects to a predetermined positional relationship in each pair, but also makes it possible to detect the relative misalignment ΔL between the two objects to be joined in the first direction Da (see FIG. 3A) in the virtual plane. Specifically, the alignment mark 6 has the following configuration.

[0043] 3A is a conceptual diagram illustrating the alignment mark 6 used in this embodiment. The alignment mark 6 includes a first mark portion 61 formed on the first workpiece W1 and a second mark portion 62 formed on the second workpiece W2.

[0044] Here, the first mark portion 61 has a shape of a right-angled triangle Gp when viewed from a direction perpendicular (in this embodiment, a horizontal plane; in the example of Figure 3 (A) , the plane of the paper) to an imaginary plane (in this embodiment, a horizontal plane), and is formed so that the direction in which the first side Ka of the two sides (hereinafter, these sides will be referred to as the "first side Ka" and the "second side Kb") that form the right angle of the triangle extends coincides with the first direction Da. As an example, the first mark portion 61 is formed by using photolithography technology to fill with metal a groove of a predetermined shape (the same shape as the right-angled triangle Gp) recessed into the joining surface of the first workpiece W1.

[0045] As a result, the hypotenuse of the right triangle Gp is ​​a line segment extending in a direction inclined from the first direction Da in the imaginary plane, and further, when the component of this hypotenuse in the first direction Da is defined as the first component and the component in the second direction Db perpendicular to the first direction Da in the imaginary plane is defined as the second component, the length La of the first component is equal to the length of the first side Ka, and the length Lb of the second component is equal to the length of the second side Kb. The first mark portion 61 has such a hypotenuse as the line segment element Q1 necessary for high-precision alignment.

[0046] Furthermore, the right-angled triangle Gp is ​​formed so that the length of the second side Kb is greater than the length of the first side Ka. In other words, the line segment element Q1 (the hypotenuse of the right-angled triangle Gp) is formed so that the length Lb of the second component (= the length of the second side Kb) is greater than the length La of the first component (= the length of the first side Ka).

[0047] The second mark portion 62 is formed so that its shape when viewed from a direction perpendicular to the imaginary plane is the shape of the first mark portion 61 inverted about an imaginary line Cs extending in the second direction Db (a right-angled triangle Gq obtained by inverting the front and back of the right-angled triangle Gp). As an example, the second mark portion 62 is formed by using photolithography technology to fill, with metal, a groove of a predetermined shape (the same shape as the right-angled triangle Gq) recessed in the joining surface of the second workpiece W2.

[0048] The second mark portion 62 has the hypotenuse of the right-angled triangle Gq as the line segment element Q2 necessary for high-precision alignment. In other words, the line segment element Q2 of the second mark portion 62 is formed so that its shape when viewed from a direction perpendicular to the imaginary plane is the inverted shape of the line segment element Q1 of the first mark portion 61 about the imaginary line Cs. Therefore, the line segment element Q2 is arranged in a twisted positional relationship with the line segment element Q1 (in other words, so as to intersect with the line segment element Q1 when viewed from a direction perpendicular to the imaginary plane), and the total length of the line segment element Q2 is equal to that of the line segment element Q1. Furthermore, the length La of the first component and the length Lb (>La) of the second component are also equal to those of the line segment element Q1.

[0049] Furthermore, the first mark portion 61 and the second mark portion 62 are formed so that when the relative positional relationship of the alignment objects (a pair of joining objects) becomes a predetermined positional relationship, the midpoints Pg of the line segment elements Q1 and Q2 coincide with each other as shown in Figure 3 (B).

[0050] According to the configuration of the alignment mark 6, when the first work W1 is relatively displaced in the first direction Da (positive direction) (see FIG. 4A) or in the opposite direction (negative direction) (see FIG. 4B) from a predetermined positional relationship (see FIG. 3B) with respect to the second work W2, the relative displacement ΔL in the first direction Da can be calculated using the formula ΔL = ΔLb × La / Lb = (Lb / 2 - Lc) × La / Lb.

[0051] This formula utilizes the fact that when the first workpiece W1 is shifted relatively in the first direction Da or the opposite direction, in the second direction Db, the intersection Ph of the line segment elements Q1 and Q2 shifts from the midpoint Pg by an amount of shift ΔLb that has a sensitivity of Lb / La times the amount of shift ΔL in the first direction Da.

[0052] Here, Lc is a length set as follows. First, of the two ends (both ends of the hypotenuse) of the line segment elements Q1 and Q2, the ends that approach each other when the first workpiece W1 is displaced relative to the second workpiece W2 in the first direction Da (positive direction) from a predetermined positional relationship (see FIG. 4A) are set as the reference ends e1 and e2. Next, for the portion of the line segment element Q1 of the first mark portion 61 from the intersection Ph with the line segment element Q2 of the second mark portion 62 when viewed from a direction perpendicular to the imaginary plane to the reference end e1, the component in the second direction Db is set as the third component. Then, the length of the third component is set as Lc. Then, the amount of deviation ΔLb in the second direction Db is expressed as ΔLb = Lb / 2 - Lc.

[0053] Therefore, when the first workpiece W1 is shifted relatively in the first direction Da, the shift amount ΔLb becomes a positive value and the shift amount ΔL also becomes a positive value (ΔLb > 0, ΔL > 0), and when the first workpiece W1 is shifted relatively in the opposite direction, the shift amount ΔLb becomes a negative value and the shift amount ΔL also becomes a negative value (ΔLb < 0, ΔL < 0).

[0054] According to this method of using the alignment mark 6, it becomes possible to measure the amount of misalignment ΔL in the first direction Da by expanding the width by Lb / La within the alignment mark 6, and as a result, it becomes possible to detect the amount of misalignment ΔL with a resolution that is Lb / La times greater. In other words, it becomes possible to detect positional misalignment in the first direction Da with high sensitivity.

[0055] In this embodiment, the length Lb of the second component is set to be 10 times or more the length La of the first component (Lb / La≧10). This allows for detection of the misalignment ΔL at a higher resolution (on the order of 0.01 μm or greater), exceeding the resolution limit (0.1 μm or greater) of a vernier formed using photolithography. This allows for alignment with higher accuracy than a vernier. Furthermore, this resolution factor (Lb / La) can be changed simply by adjusting at least one of the length La of the first component and the length Lb of the second component, and can therefore be appropriately set depending on the resolution required to detect the misalignment ΔL.

[0056] More specifically, the alignment mark 6 may have the following configuration. Fig. 5A is a conceptual diagram showing a modified example of the shape of the alignment mark 6. The first mark portion 61 may be formed so that its shape, when viewed from a direction perpendicular to the imaginary plane, is an isosceles triangle formed by combining a right triangle Gp and a right triangle Gp', which is an inverted version of the original triangle, at the first side Ka (see Fig. 3A). The second mark portion 62 may be formed so that its shape, when viewed from a direction perpendicular to the imaginary plane, is an inverted version of the first mark portion 61 about the imaginary line Cs (an inverted version of the isosceles triangle).

[0057] With this configuration, the alignment mark 6 has two pairs of line segment elements Q1 and Q2 on the left and right sides that intersect when viewed perpendicular to the imaginary plane (in FIGS. 5A and 5B, the pair on the right side has a dash (') next to the symbol to distinguish between the left and right pairs). Therefore, when the first workpiece W1 is displaced relative to the second workpiece W2 from a predetermined positional relationship in the first direction Da or the opposite direction (FIG. 5B illustrates a case where the first workpiece W1 is displaced in the first direction Da), the relative displacement ΔL in the first direction Da can be calculated from each pair of line segment elements Q1 and Q2. The two calculated displacement amounts ΔL are then averaged to determine the final displacement amount ΔL, thereby reducing the detection error of the displacement amount ΔL.

[0058] Furthermore, as a method for arranging the alignment marks 6 when forming the alignment marks 6 on the bonding objects of each set, the following arrangement example can be proposed.

[0059] 6(A) and 6(B) are conceptual diagrams showing two examples of the arrangement of the alignment mark 6. As shown in these figures, the two axial directions constituting an imaginary plane are the X-axis direction and the Y-axis direction, and alignment marks 6 are formed on the first workpiece W1 and the second workpiece W2 with the X-axis direction as the first direction Da, and separately, alignment marks 6 are further formed with the Y-axis direction as the first direction Da.

[0060] 6A and 6B show an example in which the shape of each pair of joining objects when viewed from a direction perpendicular to the imaginary plane is a rectangle, with the leftward direction on the paper being the X-axis direction and the upward direction being the Y-axis direction. The example in Fig. 6A shows a case in which the alignment mark 6 having the X-axis direction as the first direction Da is formed near the right side, and the alignment mark 6 having the Y-axis direction as the first direction Da is formed near the bottom side. The example in Fig. 6B shows a case in which the alignment mark 6 having the X-axis direction as the first direction Da is formed at two positions, one near the right side and one near the left side, and the alignment mark 6 having the Y-axis direction as the first direction Da is formed at two positions, one near the bottom side and one near the top side.

[0061] According to this method of arranging the alignment marks 6, it is possible to detect (calculate) the amount of deviation ΔL in the X-axis direction from the alignment marks 6 whose first direction Da is the X-axis direction, and further, it is possible to detect (calculate) the amount of deviation ΔL in the Y-axis direction from the alignment marks 6 whose first direction Da is the Y-axis direction. Therefore, it is possible to perform position adjustment (alignment) with high precision within a two-dimensional plane.

[0062] 7(A) and 7(B) are conceptual diagrams showing other modified examples of the shape of the alignment mark 6. As shown in these figures, the first mark portion 61 and the second mark portion 62 may be formed to have a shape obtained by combining a triangle (a right triangle in FIG. 7(A) (see also FIG. 3(A)), and an isosceles triangle in FIG. 7(B) (see also FIG. 5(A))) with a rectangle in order to widen the acute angles of the triangle. This makes it possible to form all corners at obtuse angles, making it less likely that the metal (the metal filled in to form the mark portion) will peel off at the corners.

[0063] The detection sensor 4 (see FIG. 1) is an image sensor that detects light from a light source 41 (not shown) to obtain an image including an image or shadow of the alignment mark 6 (an image or shadow of the first mark portion 61 and the second mark portion 62 overlapping each other). In this embodiment, IR light (infrared light) is used as the light from the light source 41, and materials that can transmit IR light are used for the sheet Sw and the table 12. Note that the IR light may be extracted by passing the light from the light source 41 through a bandpass filter.

[0064] The detection sensor 4 is also installed on the wide stage 32 together with the piezo stage 31. Therefore, when it is desired to obtain an image including an image of the alignment mark 6 with the detection sensor 4, the light source 41 is positioned so that light is irradiated onto the alignment mark 6 from the rear side (upper side in FIG. 1 ) of the sheet Sw. This allows the detection sensor 4 to detect light reflected from the alignment mark 6. On the other hand, when it is desired to obtain an image including the shadow of the alignment mark 6 with the detection sensor 4, the light source 41 is positioned so that light is irradiated onto the alignment mark 6 from the rear side (lower side in FIG. 1 ) of the table 12. This allows the detection sensor 4 to detect transmitted light that casts a shadow on the alignment mark 6.

[0065] Furthermore, in this embodiment, light incident on the suction surface 21p of the pressing unit 21 from the alignment mark 6 side is used to acquire an image (an image including an image or a shadow of the alignment mark 6) by the detection sensor 4. Specifically, this is as follows: The pressing mechanism 2 is formed with an optical path Rm that transmits light incident on the suction surface 21p of the pressing unit 21 from the alignment mark 6 side and guides the light to the detection sensor 4 (see FIG. 1 ). To achieve this, in this embodiment, the pressing unit 21 is made of a material (such as quartz or acrylic) that can transmit IR light.

[0066] With this detection sensor 4 and related configuration, alignment (correction of misalignment) using the alignment mark 6 can be performed while observing the alignment mark 6 with the detection sensor 4 provided on the wide stage 32. Therefore, alignment using the alignment mark 6 can be performed while the target workpiece Wt is in close proximity to the second workpiece W2 (see FIG. 9A). As a result, after alignment, the target workpiece Wt can be pressed against the surface of the second workpiece W2 without causing misalignment (see FIG. 9B). This significantly improves the accuracy of alignment with respect to the joining position Pt.

[0067] The control device 5 is composed of a processing device (such as a CPU) and a storage device (such as a RAM or a ROM), and controls each part of the bonding apparatus (such as the holding mechanism 1, the pressing mechanism 2, the alignment stage 3, and the detection sensor 4) in accordance with a control program installed in the bonding apparatus. In this embodiment, the control device 5 can execute the bonding process described below.

[0068] Here, before being installed in the bonding device, the control program may be stored in a readable state on a portable storage medium (for example, a flash memory, etc.), or may be stored in a downloadable state on another server, etc. The bonding process described below is not limited to being realized by software through the execution of a program, but may also be realized by hardware using a processing circuit built in the bonding device.

[0069] [1-2] Bonding Process FIGS. 8A to 9B are conceptual diagrams sequentially showing the state changes of the bonding device that occur as the bonding process is performed.

[0070] When the joining process is started, the control device 5 first controls the wide stage 32 to move the pressing mechanism 2 at high speed to a position where the target workpiece Wt can be held from the back side (in other words, a position above the corresponding joining position Pt) in order to align the target workpiece Wt with the joining position Pt provided on the surface of the second workpiece W2 (step S101, see FIG. 8A). At this time, the control device 5 adjusts the position of the table 12 holding the second workpiece W2 by controlling the drive unit 13 while observing the alignment marks 6 (first mark portion 61 and second mark portion 62) provided on the target workpiece Wt and the joining position Pt with the detection sensor 4, thereby also performing position adjustment (position adjustment that does not require high precision) to make the joining position Pt face the target workpiece Wt.

[0071] After step S101, the control device 5 controls the drive units 23 and 24 of the pressing mechanism 2 to lower the pressing unit 21 and the peripheral holding unit 22, thereby causing the back surface of the area of ​​the sheet Sw to which the target workpiece Wt is attached to be adsorbed onto the suction surface 21p of the pressing unit 21, and the back surface of the surrounding area to be adsorbed onto the suction surface 22p of the peripheral holding unit 22 (step S102, see FIG. 8B). As a result, the target workpiece Wt is held by the suction surface 21p of the pressing unit 21 via the sheet Sw, and the area of ​​the sheet Sw surrounding the area to which the target workpiece Wt is attached is held by the suction surface 22p of the peripheral holding unit 22.

[0072] After step S102, the control device 5 controls the piezo stage 31 to adjust the position of the pressing mechanism 2 with high precision (nanometer-order precision) (step S103; see FIG. 9A).

[0073] Specifically, the control device 5 acquires an image including an image or shadow (an image or shadow of the first mark portion 61 and the second mark portion 62 overlapping) of the alignment mark 6 provided on the target workpiece Wt and the joining position Pt using the detection sensor 4, and derives the length Lc (see FIGS. 4A and 4B) by analyzing the image. The control device 5 then uses the length Lc to calculate the relative misalignment ΔL in the first direction Da using the formula: ΔL = (Lb / 2 - Lc) × La / Lb. The control device 5 then controls the piezo stage 31 based on the calculated misalignment ΔL to adjust the positional relationship between the first workpiece W1 and the second workpiece W2 in the first direction Da to approach a predetermined positional relationship. Thereafter, the control device 5 continues to adjust the positional relationship between the first workpiece W1 and the second workpiece W2 in the first direction Da by controlling the piezo stage 31 while feeding back the calculated misalignment ΔL as needed. This makes it possible to efficiently bring the positional relationship of the workpieces closer to a predetermined positional relationship.

[0074] 6A and 6B , the control device 5 detects (calculates) a misalignment amount ΔL in the X-axis direction from an image of the alignment mark 6 in which the X-axis direction is the first direction Da, and detects (calculates) a misalignment amount ΔL in the Y-axis direction from an image of the alignment mark 6 in which the Y-axis direction is the first direction Da. The control device 5 then controls the piezo stage 31 based on both the misalignment amount ΔL in the X-axis direction and the misalignment amount ΔL in the Y-axis direction, thereby adjusting the positional relationship between the first workpiece W1 and the second workpiece W2 in a two-dimensional plane to approach a predetermined positional relationship. More specifically, the control device 5 controls the piezo stage 31 while constantly feeding back the calculated misalignment amounts ΔL in the two directions (X-axis and Y-axis), thereby adjusting the positional relationship between the first workpiece W1 and the second workpiece W2 in the two-dimensional plane so that the midpoints Pg of the two line segment elements Q1 and Q2 in each alignment mark 6 coincide with each other. This makes it possible to bring the positional relationship of the workpieces closer to a predetermined positional relationship efficiently and with high precision.

[0075] After step S103, the control device 5 controls the drive unit 23 of the pressing mechanism 2 to lower the pressing unit 21, thereby pressing the target workpiece Wt against the surface of the second workpiece W2 (step S104, see FIG. 9B). At this time, the control device 5 lowers the pressing unit 21 until the output of the pressure sensor 333 reaches the target value (in other words, until the pressing force reaches the target value).

[0076] In step S104, the control device 5 keeps the peripheral holding unit 22 in the same position when lowering the pressing unit 21. This allows the sheet Sw to bend appropriately inside the suction surface 22p when pressing the target workpiece Wt, and as a result, makes it easier to peel the target workpiece Wt from the sheet Sw.

[0077] The control device 5 performs such joining processing (steps S101 to S104) on each of the plurality of first workpieces W1 attached to the sheet Sw.

[0078] [2] Modifications [2-1] First Modification In the control process described above, the control device 5 may perform the following process. The control device 5 may measure the pressure generated during pressing in step S104 using pressure sensors 333 at three points supporting the piezo stage 31, and feed back the three measured values ​​obtained thereby to the control of the attitude adjustment mechanism 33, thereby adjusting the attitude of the piezo stage 31 so that all three measured values ​​obtained during the next pressing are values ​​within a predetermined range. Here, the predetermined range is a range set so that it is permissible to estimate that the piezo stage 31 is parallel to the wide stage 32 when all three measured values ​​are within that range.

[0079] Specifically, the control device 5 uses the three measurement values ​​to calculate values ​​such as the deviation of the measurement value from a reference value (such as a central value) within a specified range and the difference between the measurement values, and based on the calculated values, adjusts the protrusion amount of the movable part 331A of each piezo actuator 331 so that all three measurement values ​​obtained at the next pressing are within the specified range.

[0080] According to this control process, each time the target workpiece Wt is pressed against the surface of the second workpiece W2, the three measured values ​​obtained at that time are fed back to the control of the attitude adjustment mechanism 33, making it possible to adjust the attitude of the piezo stage 31 so that the piezo stage 31 is parallel to the wide stage 32. This makes it possible to always keep the piezo stage 31 parallel to the wide stage 32.

[0081] [2-2] Second Modification The above-described joining device may be appropriately modified to join the first workpiece W1 and the second workpiece W2 with their joining surfaces facing sideways (horizontally).

[0082] [2-3] Third Modification The above-described joining device may be appropriately modified to hold the target workpiece Wt on the suction surface 21p of the pressing portion 21, and then adjust the position of the target workpiece Wt (transport and alignment) in two stages using the wide stage 32 and the piezo stage 31.

[0083] [2-4] Fourth Modification The above-mentioned joining device and the detection and position adjustment (alignment) of the misalignment amount ΔL using the alignment mark 6 are not limited to the case of joining a chip (first work W1) to a substrate (second work W2), but can also be applied to the case of joining a chip (first work W1) to another chip (second work W2).

[0084] The above-described embodiments and modifications should be considered to be illustrative in all respects and not restrictive. The scope of the present invention is defined not by the above-described embodiments and modifications, but by the claims. Furthermore, the scope of the present invention is intended to include all modifications that are equivalent to the scope of the claims and fall within the scope thereof.

[0085] Furthermore, from the above-described embodiments and variant examples, the subject matter of the invention is not limited to the alignment mark 6 and its method of use, but may also include the configuration of a bonding device or part of it, or part or all of the bonding process (including the corresponding control method and bonding method), or even programs for executing them.

[0086] REFERENCE SIGNS LIST 1 Holding mechanism 2 Pressing mechanism 3 Alignment stage 4 Detection sensor 5 Control device 6 Alignment mark 11 Sheet holding section 12 Table 13 Driving section 21 Pressing section 21p Suction surface 22 Periphery holding section 22p Suction surface 23, 24 Driving section 31 Piezo stage 32 Wide stage 33 Attitude adjustment mechanism 41 Light source 61 First mark section 62 Second mark section Cs Virtual line Da First direction Db Second direction Gp, Gq Right triangle Ka First side Kb Second side La, Lb, Lc Length ΔL, ΔLb Deviation amount Pg Midpoint Ph Intersection Pt Joint position Q1, Q2 Line element Rm Light path Sw Sheet W1 First workpiece W2 Second workpiece Wt Target workpiece e1, e2 Reference end 331 Piezo actuator 331A Movable part 331B Main body part 332 Leaf spring 333 Pressure sensor

Claims

1. An alignment mark that targets a first workpiece and a second workpiece that overlap each other when viewed from a direction perpendicular to a two-dimensional imaginary plane, enables the relative positional relationship between the workpieces to be adjusted to a predetermined positional relationship, and enables detection of a relative deviation ΔL in a first direction within the imaginary plane, the alignment mark comprising: a first mark portion formed on the first workpiece; and a second mark portion formed on the second workpiece, wherein the first mark portion and the second mark portion each have line segment elements of the same length that extend in a direction inclined from the first direction within the imaginary plane when viewed from a direction perpendicular to the imaginary plane, an alignment mark in which the second mark portion is formed so that the shape of the line element when viewed from a direction perpendicular to the imaginary plane is the shape of the line element of the first mark portion inverted around a virtual line extending in a second direction perpendicular to the first direction within the imaginary plane, and the first mark portion and the second mark portion are formed so that the midpoints of the line element coincide with each other when the relative positional relationship between the first workpiece and the second workpiece becomes the predetermined positional relationship, and each line element is formed so that when the component in the first direction is defined as a first component and the component in the second direction is defined as a second component, the length Lb of the second component is greater than the length La of the first component.

2. An alignment mark as described in claim 1, wherein each line element is formed so that the length Lb of the second component is 10 times or more the length La of the first component, depending on the resolution required to detect the deviation amount ΔL.

3. A method of using an alignment mark according to claim 1 or 2, wherein the end of each of the two ends of the line segment elements that approaches each other when the first workpiece shifts relatively with respect to the second workpiece in the first direction from the predetermined positional relationship is used as a reference end, and for the portion of the line segment element of the first mark part from the point of intersection with the line segment element of the second mark part when viewed from the perpendicular direction to the reference end, the component in the second direction is set as a third component, and the length Lc of the third component is used to calculate the relative shift amount ΔL in the first direction using the formula ΔL = (Lb / 2 - Lc) × La / Lb.

4. A method of using an alignment mark as described in claim 3, wherein the positional relationship between the first workpiece and the second workpiece in the first direction is adjusted to approach the specified positional relationship based on the deviation amount ΔL calculated using the formula.

5. A method of using an alignment mark as described in claim 3, wherein the two axial directions constituting the virtual plane are the X-axis direction and the Y-axis direction, and the alignment mark is formed on the first workpiece and the second workpiece with the X-axis direction as the first direction, and separately, the alignment mark is further formed with the Y-axis direction as the first direction, and these alignment marks are used to calculate both the amount of deviation ΔL in the X-axis direction and the amount of deviation ΔL in the Y-axis direction.

6. A method of using an alignment mark as described in claim 5, wherein the positional relationship between the first workpiece and the second workpiece in the virtual plane is adjusted to approach the specified positional relationship based on both the deviation amount ΔL in the X-axis direction and the deviation amount ΔL in the Y-axis direction.

Images