JP5157364B2 - Alignment method for joining objects, component joining method and component joining apparatus using the same - Google Patents

Description

  The present invention relates to a method for aligning an object to be joined, in which mainly an opaque joining component and a component to be joined are aligned, a component joining method using the same, and a component joining apparatus.

Conventionally, as an alignment method of this kind of bonding object, a method of aligning a semiconductor chip with a substrate to which the semiconductor chip is bonded is known (see Patent Document 1). In this alignment method (positioning method), the substrate recognition camera faces the substrate set on the bonding stage from the upper side to recognize the image of the substrate, and the semiconductor chip adsorbed to the bonding tool is fixedly arranged from the upper side. The semiconductor chip image is recognized by facing the semiconductor recognition camera, and the recognition result of the substrate and the recognition result of the semiconductor chip are combined to obtain the position correction amount of the semiconductor chip with respect to the substrate, and based on this position correction amount The semiconductor chip is aligned with the substrate.
Japanese Patent Laid-Open No. 9-232363

  In such a conventional alignment method, even if the position correction amount is obtained with high accuracy, there is a limit to high-precision alignment due to the mechanical movement accuracy of the bonding tool that moves from the position of the semiconductor recognition camera to the bonding stage. There was a problem that occurred. In addition, when a semiconductor chip is bonded to the substrate, there is a problem that a certain pressing force acts, and this pressing force impairs alignment accuracy.

  The present invention relates to a method of aligning objects to be joined, which can perform high-precision alignment between a joined part and a joined part regardless of the mechanical accuracy of the moving system, a part joining method and a part joining using the same. An object is to provide an apparatus.

An alignment method for a bonding object according to the present invention includes an alignment mask formed of a light-transmitting substrate on which a reference mark is formed, a first image recognition unit that recognizes a positioning reference and a reference mark of a bonding component, and a first image recognition unit. And a second image recognition means for recognizing the positioning reference and reference mark of the part to be joined from the opposite direction, and the joining part in which the joining positioning reference is formed on the joining surface and the joining position on the joining surface. A method for aligning a joining object, wherein a joining target formed with a reference is aligned with a joining surface facing the joining surface with reference to a reference mark of an alignment mask. the criteria are formed surface, comprising the steps of and pressing superposed in facing the surface on which the reference mark is formed of an alignment mask, bonding parts and a In the stacked state and Lee instrument mask reference marks and bonding positioning reference junction part of the alignment mask from the side facing the bonding component across the alignment mask, a first recognition step position recognizing by the first image recognition device After the joining component is separated from the alignment mask , the second image recognition means is inserted at a position between the joining component and the alignment mask, and the reference mark of the alignment mask is formed using the second image recognition means. from the side, located recognizing the reference mark of the alignment mask, and a second recognition step for taking the second image recognition means with reference to the reference mark position data of the joining positioning reference recognized in the first recognition step, the second recognition a position where there is alignment mask in the step, a step of moving the bonding component, the second image recognition The stage, and a third recognition step position recognizes the joined positioning reference of the joint parts from the side of the joined positioning reference is formed to be joined parts, the position of the joint positioning reference incorporated in the second image recognition device The mutual displacement amount acquisition step of acquiring the displacement amount between the data and the bonding positioning reference as the mutual displacement amount of the bonded component and the bonded component, and the bonded component based on the acquired mutual displacement amount A position correction step for correcting the position of the component to be bonded relatively, and each position recognition in the first recognition step, the second recognition step, and the third recognition step is a bonding position where the bonding component is bonded to the bonding component. It is characterized by being implemented in .

In this case, it is preferable that both the joining component and the joined component include a light shielding material.

On the other hand, the component joining apparatus of the present invention is characterized by executing the above-described method of aligning the joining objects, and joining the joined components in a state where they are superposed and pressed against the joined components after the position correction step. .

  According to these configurations, in the first recognition process (first recognition operation), the position of the reference mark and the joint positioning reference is simultaneously recognized by the first image recognition means in a state where the joint component is superimposed and pressed on the alignment mask. Therefore, the recognition result can reflect the condition for joining the joined component to the joined component. Next, in the second recognition step (second recognition operation), the position of the reference mark of the alignment mask is recognized again by the second image recognition means. Thereby, the position of the joining component, which is the recognition result of the first recognition step, can be moved from the coordinate system of the first image recognition means to the coordinate system of the second image recognition means. Next, in the third recognition step (third recognition operation), the second image recognition unit recognizes the position of the bonding positioning reference of the bonded parts. Thereby, the position of the part to be joined can be taken into the coordinate system of the second image recognition means. Therefore, finally, based on the recognition result of the second image recognition means, it is possible to accurately grasp the amount of misalignment between the position of the joined part and the position of the joined part. In addition, since these recognition operations are performed at the joining position where the joining component is joined to the joined component, the alignment is not affected by the mechanical movement accuracy in the moving system of the joining component and the joined component. In this way, it is possible to align the joined component and the joined component that are joined together in a pressed state with high accuracy.

In these cases, the reference marks, joint positioning reference and the joined positioning criterion, nor any configured respectively as a pair of disposed apart from each other, the first image recognition means and the second image recognition means, It is preferable that each of them is composed of a pair of ones spaced apart from each other.

  According to these configurations, image recognition by the first image recognition unit and the second image recognition unit can be easily performed in a short time, and a calibration error can be suppressed.

  The component joining method of the present invention preferably includes the above-described method of aligning the joining objects, and a joining step of joining the joined components in a state of being overlapped and pressed against the joined components after the position correction step. .

According to this configuration, the joined component and the joined component can be joined in a highly aligned state, and the joining accuracy and, consequently, the yield of the joining object can be improved.

  Hereinafter, a component joining apparatus to which an alignment method of the present invention is applied will be described with reference to the accompanying drawings. This component joining apparatus joins a flexible substrate (FPC) to a small liquid crystal panel by thermocompression bonding via an anisotropic conductive film (AFC). Specifically, a liquid crystal panel and a flexible substrate with an AFC are supplied in a pre-aligned state, both are aligned on the apparatus, and then the terminal portion of the flexible substrate with an AFC is superimposed on the terminal portion of the liquid crystal panel. And thermocompression bonding. For ease of explanation, in the following explanation, a flexible substrate as a joining member is indicated as a joining work, and a liquid crystal panel as a joining member is indicated as a joining work, and both are illustrated as plate-like members. .

  Here, before explaining the component joining apparatus, a joined work, a work to be joined, and an alignment mask will be briefly explained. As shown in FIG. 2, the alignment mask M is a transparent (translucent) glass mask made of quartz glass or the like, and has a pair of reference marks spaced apart from each other on the upper surface (surface). Ma is formed (see FIG. 5A). Each reference mark Ma is composed of four marking points Mb positioned so as to form a cross, and the center position (the center position of the cross) of the four marking points Mb is substantially the reference mark (reference point Mc) Ma. . Similarly, a pair of bonding positioning references Aa and Aa corresponding to the pair of reference points Mc and Mc of the alignment mask M are marked on the lower surface of the bonding workpiece A (see FIG. 4B). Furthermore, on the upper surface of the workpiece B to be joined, a pair of to-be-joined positioning references Ba and Ba corresponding to the pair of reference points Mc and Mc of the alignment mask M are marked (see FIG. 10C).

  That is, the pair of reference points Mc and Mc of the alignment mask M, the pair of joint positioning references Aa and Aa of the joined workpiece A, and the pair of joined positioning references Ba and Ba of the workpiece B to be joined are the same. The separation distance is provided. In this case, the pair of bonding positioning references Aa and Aa are formed on the lower surface of the bonded work A, that is, the bonding surface, and the pair of bonded positioning references Ba and Ba are formed on the upper surface of the bonded work B, that is, the bonding surface. ing. Accordingly, the workpiece A and the workpiece B to be joined are aligned by superimposing the workpiece A on the workpiece B and matching the pair of bonding positioning references Aa and Aa with the pair of workpiece positioning references Ba and Ba. Will be. However, since the workpiece A and the workpiece B to be joined are opaque members and cannot be directly aligned by overlapping the workpiece A and the workpiece B to be joined, the alignment is performed with the alignment mask M interposed. Like to do.

  As shown in FIG. 1, the component joining apparatus 1 includes a holding head 2 that sucks and holds a workpiece A, a lifting mechanism 3 that lifts and lowers the holding head 2, a mask stage 4 on which an alignment mask M is set, and a workpiece to be joined. The alignment mask M and the workpiece B to be joined are passed through the θ table 5 on which the work B is set, the Y-axis table 6 on which the θ table 5 is mounted, and the mask stage 4, θ table 5 and Y-axis table 6. The alignment mask M and the workpiece A are imaged from the lower side of the X-axis table 7 on the X-axis table (moving table) 7 mounted side by side and on the axis of the apparatus (on the same axis as the holding head 2 and the lifting mechanism 3). A pair of first image recognition cameras (only one shown: first image recognition means) 11 faces the axis of the device from the side of the device and aligns from the top A pair of second image recognition cameras (only one is shown: second image recognition means) 12 that images the mask M and the workpiece B to be joined and a pair of second image recognition cameras 12 are moved forward and backward so as to freely advance and retract on the axis. A mechanism 13 and a control device 14 that controls these constituent devices are provided.

  The pair of first image recognition cameras 11 are disposed with the same separation distance as the separation distance of the pair of reference points Mc, Mc and the like. Similarly, the pair of second image recognition cameras 12 are also arranged with the same separation distance as the separation distance of the pair of reference points Mc, Mc and the like. Thus, one reference mark Ma, the joining positioning reference Aa, and the joined positioning reference Ba are imaged by the first image recognition camera 11 and the second image recognition camera 12, and the other reference mark Ma, the joining positioning reference Aa. And the to-be-joined positioning reference | standard Ba is imaged with the other 1st image recognition camera 11 and the 2nd image recognition camera 12. FIG. Needless to say, the pair of first image recognition cameras 11 and the pair of second image recognition cameras 12 are calibrated with each other.

  The holding head 2 is a so-called suction head, and has a heater and a spring built therein although not shown in the drawing. The holding head 2 thermally press-bonds the adsorbed bonded workpiece A to the workpiece B to be bonded, but the pressure-bonding force is maintained by a spring, and thermal bonding is performed by a heater. The elevating mechanism 3 supports the holding head 2 and raises and lowers the bonded workpiece A between the bonding position and the rising standby position via the holding head 2. Although details will be described later, the bonded workpiece A that has moved to the bonding position is pressed against the alignment mask M during alignment and is pressed against the workpiece B during bonding. In addition, the second image recognition camera 12 faces the position immediately above the joining position from the side while the joining workpiece A has moved to the ascending standby position.

  The mask stage 4 supports the alignment mask M so that the upper surface of the set alignment mask M and the upper surface of the workpiece B set on the θ table 5 are located in the same horizontal plane. That is, the mask stage 4 supports the alignment mask M so that the reference mark Ma and the bonded positioning reference Ba are positioned at the same height. The mask stage 4 has a stage opening 21 at the center so that the first image recognition camera 11 facing from below can capture the reference mark Ma of the alignment mask M.

  The θ table 5 rotates the set work B to be joined by a minute angle in the horizontal plane. The Y-axis table 6 moves the work B to be joined in the Y-axis direction in the horizontal plane via the θ table 5. Similarly, the X-axis table 7 moves the work B to be joined in the X-axis direction in the horizontal plane via the θ table 5 and the Y-axis table 6. The workpiece B and the alignment mask M are set side by side in the X-axis direction on the X-axis table 7, and the X-axis table 7 moves the workpiece B and the alignment mask M to the bonding position and the respective movements. It is alternately moved in the X-axis direction with respect to the standby position. FIG. 1 shows a state in which the workpiece B is moved to the bonding position and the alignment mask M is moved to the movement standby position.

  On the other hand, mutual alignment (position correction) between the workpiece A and the workpiece B is performed by minutely moving the workpiece B with respect to the workpiece A. This correction means is a cooperating θ table. 5 and a Y-axis table 6 and an X-axis table 7. The X-axis table 7 is formed with a table opening 22 that coincides with the stage opening 21 so that the first image recognition camera 11 can image the reference mark Ma in the same manner as the mask stage 4.

  Each first image recognition camera 11 is constituted by a so-called CCD camera, and is provided upward and fixedly on an axis below the X-axis table 7. Although the details will be described later, the first image recognition camera 11 presses the joining workpiece A against the alignment mask M moved to the joining position, and the reference mark Ma of the alignment mask M and the joining positioning reference Aa of the joining workpiece A are Simultaneous position recognition from the bottom. In a state where the bonded workpiece A is pressed against the alignment mask M, the reference mark Ma formed on the upper surface of the alignment mask M and the bonded positioning reference Aa formed on the lower surface of the bonded workpiece A are positioned in the same horizontal plane. The single-image recognition camera 11 captures the reference mark Ma and the joint positioning reference Aa within the field of view. In addition, it is preferable that the pressing force when pressing the workpiece A against the alignment mask M is the same as the pressing force when pressing the workpiece A against the workpiece B in thermocompression bonding.

  Each second image recognition camera 12 is also composed of a CCD camera, and is configured to be movable in the X-axis direction between the position just above the joint position and the retracted position by the advance / retreat mechanism 13. Although the details will be described later, the second image recognition camera 12 recognizes the position of the reference mark Ma of the alignment mask M at the joining position and the joining positioning reference Ba of the work B to be joined moved to the joining position from above. The second image recognition camera 12 that has moved to the position immediately above is individually within the field of view of the reference mark Ma and the position reference Ba to be bonded with respect to the alignment mask M and the work B to be bonded that are moved immediately below by the X-axis table 7. Capture and image this. The first image recognition camera 11 and the second image recognition camera 12 may be a single unit and may be moved in the Y-axis direction by a moving mechanism.

  Although not shown in the drawings, the control device 14 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like, and includes a first image recognition camera 11 and a second image recognition. In addition to the camera 12, the driving of the holding head 2, the lifting mechanism 3, the θ table 5, the Y axis table 6, the X axis table 7, and the advance / retreat mechanism 13 is controlled. Further, the mutual displacement amount between the workpiece A and the workpiece B is calculated from the recognition results by the first image recognition camera 11 and the second image recognition camera 12.

  Next, with reference to FIG. 3 and FIG. 4, the alignment operation | movement and joining operation | movement of the workpiece | work A and the workpiece | work B to be joined are demonstrated. In this alignment operation, the alignment mask M faces the axis of the apparatus (on the coaxial line with the holding head 2 (joining workpiece A), the lifting mechanism 3 and the pair of first image recognition cameras 11), and a pair of second images. The recognition camera 12 starts from a state where it is retracted (retracted position) from the axis.

  First, the holding head 2 is lowered by the elevating mechanism 3 and moved to the joining position, and the joining work A is pressed and brought into contact with the alignment mask M on the mask stage 4 (FIG. 3A). When the alignment mask M comes into contact, the pair of reference marks Ma, Ma positioned on the upper surface of the alignment mask M and the pair of bonding positioning references Aa, Aa positioned on the lower surface of the bonded workpiece A are positioned at substantially the same position. The pair of first image recognition cameras 11 simultaneously recognizes the pair of reference marks Ma, Ma and the pair of joint positioning references Aa, Aa from the stage opening 21 and the table opening 22 (first image recognition step). FIG. 4A is an example of the image recognition result at this time, and each joint positioning reference Aa is misaligned with respect to each reference point Mc.

  Next, after the holding head 2 is lifted by the lifting mechanism 3 and moved to the rising standby position, the pair of second image recognition cameras 12 are caused to face on the axis by the advance / retreat mechanism 13 (FIG. 3B). ). As a result, the pair of second image recognition cameras 12 faces the alignment mask M (directly above position), and images the pair of reference marks Ma, Ma on the alignment mask M (second image recognition step). FIG. 4B is an example of the image recognition result at that time, and the pair of joint positioning references Aa and Aa whose positions are recognized by the first image recognition camera 11 are taken into the second image recognition camera 12 in the data. This is the state (displayed with “+”).

  Thereafter, the alignment mask M and the workpiece B to be joined are moved by the X-axis table 7 so that the workpiece B is exposed on the axis (FIG. 3C). At this time, the pair of second image recognition cameras 12 faces the workpiece B to be welded, and images the to-be-joined positioning standards Ba and Ba of the workpiece B to be joined (third image recognition step). FIG. 4C is an example of superimposed images taken in the second image recognition process and the third image recognition process.

  Based on the captured image of the first image recognition camera 11 in the first image recognition step and the captured image of the second image recognition camera 12 in the second image recognition step and the third image recognition step, the bonded workpiece B is bonded. A misregistration amount with respect to A is calculated.

  First, using a pair of reference marks Ma, Ma of the alignment mask M as a reference, the captured images of the pair of first image recognition cameras 11 in the first image recognition step, and the pair of pairs in the second image recognition step and the third image recognition step. The captured image of the second image recognition camera 12 is overlaid (it is only necessary to overlap in calculation). Based on FIG. 4C, the positional deviation (ΔX, ΔY) in the XY-axis direction between the midpoint of the pair of joining positioning references Aa, Aa and the midpoint of the pair of joined positioning references Ba, Ba is calculated. The angle deviation (Δθ) between the line segment of the pair of bonding positioning references Aa and Aa and the line segment of the pair of bonded positioning references Ba and Ba is calculated. The positional deviation amounts (ΔX, ΔY, Δθ) of the joining positioning references Aa, Aa and the joining positioning references Ba, Ba become the positional deviation amounts of the work B to be joined with respect to the joining work A as they are. Therefore, the X-axis table 7, the Y-axis table 6, and the θ table 5 correct the position of the workpiece B to be bonded (in the XYθ direction) (alignment).

Then, after retracting the pair of second image recognition cameras 12 by the advancing / retreating mechanism 13, the holding head 2 is lowered by the elevating mechanism 3 so that the bonded work A is pressed into contact with the work B to be bonded. thermocompression bonding by a built-in heater (FIG. 3 (d)). Thereby, the workpiece A is joined to the workpiece B.

  The alignment process is preferably performed once for each bonding process.

  Further, in this embodiment, after the position correction, the bonded workpiece A and the workpiece B to be bonded are pressed and contacted, but in the state in which the bonded workpiece A and the workpiece B to be bonded are in contact (or pressed contact), Correction may be performed.

  As described above, according to the present embodiment, the reference mark Ma of the alignment mask M, the bonding positioning reference Aa of the bonded workpiece A, and the bonded position of the workpiece B to be bonded at the bonding position on the component bonding apparatus 1. The position of the reference Ba is recognized, and the position is recognized in a state where the bonded work A is pressed against the alignment mask M. That is, the position recognition of the workpiece A and the workpiece B to be joined is performed under the same conditions as in the thermocompression bonding operation in the position, environment, pressing condition, and the like. For this reason, mutual alignment of the workpiece A and the workpiece B can be performed with extremely high accuracy. This alignment method is suitable for alignment of opaque workpieces.

  In this embodiment, the positioning between the flexible substrate and the liquid crystal panel has been described as an example. However, the present invention can be applied to the positioning (mounting) between the IC chip and the circuit board, the positioning between the inkjet head and the carriage, and the like. Needless to say. Therefore, the present invention is not limited to thermocompression bonding and can be applied to ultraviolet bonding, solder bonding, and the like using an ultraviolet curable adhesive.

1 is an overall schematic diagram of a component bonding apparatus according to an embodiment. It is the top view (a) of an alignment mask, the back view (b) of a joining workpiece | work, and the top view of a to-be-joined workpiece | work (c). It is explanatory drawing which shows the alignment operation | movement in a component bonding apparatus. It is explanatory drawing showing the imaging result.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1: Component joining apparatus, 2: Holding head, 3: Elevating mechanism, 5: (theta) table, 6: Y-axis table, 7: X-axis table, 11: 1st image recognition camera, 12: 2nd image recognition camera, 14 : Control device, A: Joined workpiece, Aa: Joining positioning reference, B: Joined workpiece, Ba: Joined positioning reference, M: Alignment mask, Ma: Reference mark

Claims (5)

An alignment mask made of a translucent substrate on which fiducial marks are formed;
First image recognition means for recognizing the positioning reference of the joining component and the reference mark;
Second image recognition means for recognizing the positioning reference of the part to be joined and the reference mark from the opposite direction to the first image recognition means;
Use
The joining part in which the joining positioning reference is formed on the joining surface, and the joined part in which the joining positioning reference is formed on the joining surface, and the joining surface with respect to the reference mark of the alignment mask An alignment method of an object to be bonded that is aligned in a state in which the bonded surfaces are opposed to each other,
A step of superimposing and pressing the surface of the bonding component on which the bonding positioning reference is formed, facing the surface of the alignment mask on which the reference mark is formed;
In a state where the joining component and the alignment mask are overlapped, the reference mark of the alignment mask and the joining positioning reference of the joining component from the side facing the joining component with the alignment mask interposed therebetween, A first recognition step for recognizing the position by the image recognition means;
After separating the joining component from the alignment mask, inserting the second image recognition means at a position between the joining component and the alignment mask;
Using the second image recognition means, the position of the reference mark of the alignment mask is recognized from the surface side of the alignment mask on which the reference mark is formed , and the bonding positioning reference recognized in the first recognition step A second recognition step of taking the position data of the second image recognition means with reference to the reference mark ;
Moving the part to be joined to a position where the alignment mask was present in the second recognition step ;
A third recognizing step of recognizing the position of the to-be-joined part of the part to be joined from the side of the part to be joined of the part to be joined by the second image recognition unit;
A mutual misalignment for acquiring a displacement amount between the position data of the joining positioning reference taken into the second image recognition means and the joined positioning reference as a mutual displacement amount of the joining component and the joined component. A quantity acquisition process;
A position correction step for relatively correcting the position of the bonded component and the bonded component based on the acquired mutual displacement amount;
Equipped with a,
Alignment of bonding objects , wherein each position recognition in the first recognition step, the second recognition step, and the third recognition step is performed at a bonding position where the bonding component is bonded to the bonded component. Way . 2. The method of aligning bonding objects according to claim 1, wherein both the bonded component and the bonded component include a light-shielding material. Said reference mark, the joint positioning reference and the object to be bonded positioned criteria, nor any configured respectively as a pair of disposed apart from each other,
The joining object according to claim 1 or 2, wherein each of the first image recognizing unit and the second image recognizing unit is composed of a pair of components disposed apart from each other. Alignment method. An alignment method for bonding objects according to any one of claims 1 to 3 ,
And a joining step of joining the joined parts in a state of being overlapped and pressed against the parts to be joined after the position correcting step. Execute the method for aligning bonding objects according to any one of claims 1 to 3,
A component joining apparatus that joins the joined component in a state of being overlapped and pressed against the joined component after the position correcting step.

Images