JP2773576B2 - Design support system for semiconductor device manufacturing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a design support system for a method of manufacturing a semiconductor device, and more particularly to a pressure bonding technique.

[0002]

2. Description of the Related Art Conventionally, Tape Automated
Bonding (TAB) is designed and manufactured as follows.

[0003] A copper tape is laminated on a polyimide tape having sprocket holes, and a lead pattern corresponding to a bonding wire is formed on the polyimide tape by photo-etching.
A semiconductor device having a protruding electrode (hereinafter referred to as a bump) formed on an external lead electrode pad is prepared.
The lead of the B tape was superimposed, the required pressing force and heating temperature based on the past results, the respective operation times, and the amount of press-down (hereinafter, the amount of reduction) due to the reduction were set in the manufacturing apparatus, and bonding was performed. Then, to determine the bondability, a method of measuring and confirming the final deformation of the bumps and leads by a tension test with a tension gauge and a visual inspection with a microscope, and a bump / lead structure by accelerated reliability tests such as a temperature cycle test -The method of judging the quality of the design of the dimensions and the manufacturing process was used. For new products that are difficult to infer from past performance, stress analysis is performed by simulations based on computational dynamics methods applied to fields such as mechanical, structural, and material processing represented by the finite element method.・ We performed thermal analysis and used it as a reference for designing the structure of bumps, the structure of film carrier tapes, material properties, and processing power.

[0004]

In the above-described conventional semiconductor device and the design method of the method of manufacturing the same, the actual product or
Since it is necessary to judge the quality of the design by reliability evaluation tests etc. after undergoing the entire manufacturing process using prototypes of a form similar to that, feedback to the design guidelines must be passive, at the same time, It takes a long time (Turn Around Time: TAT) from arranging and purchasing materials to producing prototypes of products and creating samples for evaluation, as well as costs for trial production, and the evaluation results are only good or bad. Therefore, it was difficult to obtain a theoretical, qualitative and quantitative design guideline such as what kind of improvement design should be performed, specific information on the design method and the improvement part.

[0005] Further, even after an optimized result is finally obtained after a trial and error by a skilled engineer, there is no methodology necessary for systematizing the process, and there are support tools for the design. I didn't.

More specifically, after obtaining a defect determination, the designer must try to make an improvement by analogy from some past results, and it is necessary to change the design of the semiconductor device, We had to worry about design changes, material changes, and so on.

[0007] When the design is changed, the evaluation is restarted from the beginning, and the evaluation using the actual product described above is repeated. Speaking of material changes, it is not in line with the design orientation of selecting from materials prepared by material manufacturers, and when evaluating them, it is necessary to evaluate all prepared materials uniformly. Was. In other words, the design environment had to spend a lot of man-hours, manpower, time and money, relying on some past achievements.

Recently, simulation tools and CAD systems for mechanical design, structural design, material processing process design and the like, which have been researched and developed in the field of semiconductors and other industrial fields, have been diverted to, for example, TAB. Regarding, there is an example in which the stress distribution at the time of pressurization, the temperature distribution at the time of heating, heat flux, thermal stress, etc. are obtained by numerical calculation by the finite element method (hereinafter, FEM), and the design is performed with reference to the results Now you can see it. These simulation tools include the general-purpose structural analysis program “ABAQUA
S "(plasticity and processing, Vol. 31, No. 350, 19)
90-3, pp. 347-349) or forging three-dimensional simulation “FORGE3” (plasticity and processing Vo)
l. 31, No. 350,1990-3, pp. 292
297) or “MARC” (ibid., Pp. 350)
To 353), and efforts are being made to apply these to obtain design guidelines. However, TAB
As in the case of the bonding process of (1), the target is an extremely small scale, and the unsteady deformation process (Metal flo
There is no simulation tool useful in production technology that can describe a process of joining in a three-dimensional manner with w) and with the contact conditions and boundary conditions of the joining surface and the tool surface changing in time series. In this field of research, it is necessary to link the deformation behavior to the joining process, and in addition to mechanical and applied mathematical problems, physical and applied problems such as contact problems and material deformation mechanisms.
Since it also includes many problems related to metallurgy, research is ongoing to solve them, and even if only analysis of deformation behavior is taken, it can be determined according to the analysis target and purpose. In fact, they have set their own subtle processes and restrictions, and selected and combined analysis methods. Therefore, it is expected that an analysis method for obtaining a complete solution that can realize the reality or an approximate solution close thereto will be widely available for a while (Plasticity and Processing, Vol. 3).
1, No. 350, p. p282-283 and pp. 2
84-291 and pp. 292-297).

[0009] Even from the viewpoint of hardware, which has been making remarkable progress, a supercomputer capable of high-speed operation is still a valuable resource, and at the same time, the calculation time for FEM analysis is only several times that of a general-purpose large-sized machine. Report that it cannot be shortened (plasticity and processing, Vol. 31, No. 35)
0, (1990-3) p. 296), and the effect of lowering the calculation cost may not be so remarkably expected because of the high CPU cost. In addition, there is a problem in practicality that the output time (image processing) of the peripheral device may take 24 hours or more. Also, Engineer
The ing Work Station (hereinafter referred to as EWS) also has a calculation speed comparable to that of a conventional minicomputer, but it is difficult to solve a three-dimensional analysis having nodes at several thousand points in minutes to tens of minutes. (Plasticity and Processing, Vol. 37, No. 350, (1990-
3), p. 296). Therefore, for production engineers,
There was also the background that the emergence of a low-cost design support system that could withstand practical use was awaited.

[0010]

SUMMARY OF THE INVENTION A design support system for a method of manufacturing a semiconductor device according to the present invention includes the following: A: Power by a gear such as a rack and pinion to which a cylinder / piston mechanism or a speed-controllable servo mechanism is added. A press mechanism unit such as a transmission mechanism or a feed screw mechanism; B: a press control unit for controlling the magnitude of driving power of the press mechanism unit and a pressing speed; C: an accelerometer added to the press mechanism unit; Load cell,
A press sensor unit for measuring various mechanical and geometrical quantities associated with a pressing process such as a displacement gauge; and D: crimping of an upper die punch corresponding to a bonding tool of a semiconductor manufacturing apparatus which is also added to a part of the press mechanism. An upper punch having a flat, uneven or tapered surface, E: a stopper that can arbitrarily change the pressing amount of the upper punch, and F: an upper punch is pressed and deformed by a pressing process. Giving a sample made of a deformed model material in which a lattice line is drawn on an arbitrary cross section that is made by imitating a bump and a lead of a film carrier tape, and G: supporting the sample made of the model material and forming an upper punch. The pressure receiving sensor that detects the surface pressure distribution generated on the bottom surface of the sample when the sample is pressed is deployed in a lattice point or is arranged one-dimensionally on a straight line. And H: a stage in which the pressure receiving sensor plate is incorporated and a guide bar is provided so that the upper die presses vertically downward; I: output from the press sensor portion and the pressure receiving sensor plate An instrumentation unit that reads and stores in real time each voltage change under a transient phenomenon that is performed, and J: controls the instrumentation unit and converts the voltage change that indicates a transient phenomenon in the pressing process into actual mechanical quantities and geometric quantities. A series of data that has been calibrated and stored on a peripheral device such as a floppy disk, graphed on a display screen or a printer, and graphically output to quantitatively reproduce and edit press characteristics. And a central processing unit such as a personal computer in which the procedures are incorporated in internal and external storage units and the like are managed and executed. A series of various procedure means for interactively designing the press characteristics interactively on the display screen and transmitting the information to the press control unit, which is executed in the central processing unit; Copy photographic means in which the pre-pressed pattern of the grid lines and the pattern of the grid lines in any pressing step are copied; and M: the grid line pattern is copied from the copy photographic means to the xy orthogonal coordinate system. N: controlling the image capturing device or the pointing device such as an image scanner or a digitizer which reads as coordinate values of the image scanner; and N: controlling the image capturing device or the pointing device by using a grid line analysis method from data of each coordinate value of a grid point of a grid line. A numerical analysis program for performing an unsteady deformation analysis.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described with reference to the drawings.

FIG. 1 is an overall system configuration diagram of a first embodiment of the present invention.

An air actuator is used as an actuator, and a cylinder 11, a piston 12, an upper punch 13
The press mechanism unit 10 is constituted. The press mechanism 10 is connected through a pipe to an electromagnetic valve 51 and an electropneumatic regulator 53 having a gauge 52 for digitally displaying air pressure, so that high-pressure air can be sent into the cylinder 11 at a required set pressure. The air pressure of the electropneumatic regulator 53 can be arbitrarily set by transmitting a pressure setting signal from the press control controller 54 and performing valve control of the regulator by voltage control. At the same time, the press start signal is
4, the gate circuit is turned on, and the solenoid valve 5
Open the 1 valve. FIG. 14 shows a detailed piping diagram when an air actuator is used.

When the solenoid valve 51 is opened and the piston 12 and the upper punch 13 start to descend, the displacement meter 21 starts to output, and this is used as a trigger to start all measurements. First, the voltage output of the displacement meter 21 is recorded as a position change (movement) of the upper die 13 through the amplifier 61 into the instrumentation system 62 over the entire pressing process. Load cell 2 at the same time
2 records the total pressure drop (voltage output) of the upper die 13 as a transient phenomenon in the instrumentation system 62 through the amplifier 61 over the entire pressing process. Also, during this process, the pressure receiving sensor plate 41 detects the surface pressure distribution generated on the bottom surface of the bump model sample 32, and this is also recorded as a voltage output through the amplifier 61 in the instrumentation system 62.

The bump model sample 32 appearing in the above description is, together with the lead model sample 31, the lead 3 of the film carrier tape of the semiconductor device shown in FIGS. This is a model material sample 30 created by imitating the bump 4 and imitating it.
The same sample is used because the unsteady process of deformation / joining cannot be visually recognized, and detailed and direct strain analysis inside the joint cannot be performed. This point also differs from the conventional design method in the present invention. This is the part that has superiority. The detailed description of the sample creation will be described later. As described above, the process of applying the pressing and deformation to the model material sample 30 by the press mechanism unit 10 includes the press sensor unit 20, the pressure receiving sensor plate 41 of the stage unit 40, and the instrumentation unit 6.
0, and the computer 71 of the central processing unit 70, the measurement / control / processing program 90, and the floppy disk device 72 are combined to perform a series of operations and processing, thereby proceeding. During the same process, the measured displacement / load (total rolling force) and each voltage change indicating the surface pressure distribution are represented as a voltage output-time curve on a graphic display of the computer 71 or a peripheral device such as an xy plotter 74. It is displayed in real time in the form of a graph. The above experimental results can be reproduced any number of times after the experiment.
Further, the results are interactively calibrated into various mechanical quantities and geometric quantities from the above-described voltage output-time curves on the computer 71 by using the measurement / control / processing program 90. That is, a press load-time curve and a displacement (a reduction amount) corresponding to a bonding process of a semiconductor device.
It can be known as a time curve and a surface pressure distribution generated below the bump 4 of the semiconductor device, which cannot be detected in the manufacturing process of the actual product, that is, a surface pressure distribution generated on the bottom surface of the bump model sample 32 versus time curve. If necessary, conversion, editing, and output of the press load-displacement curve, which is important for knowing the operating characteristics of the upper punch 13, can be automatically performed, and the reproducibility of the experiment can be confirmed while viewing the result on the display screen And dynamic phenomena at the time of pressing (ie, at the time of bonding) can be compared and studied. All of the above processes are automated, reducing the workload of the experimenter,
There is an advantage that the time required for the analysis process can be shortened and the experimental data can be managed in an electronic file.

Next, unsteady deformation information of the model material sample 30 is obtained as a change in the pattern of a grid line previously drawn on an arbitrary cross section of the sample material. The change of the grid line is copied to a copy photograph 33, digitized by a digitizer 73, and read as xy coordinate values. The read coordinate values are sequentially written to the floppy disk device 72 through the computer 71 and recorded and held. Based on the data, the distribution of deformation rate (velocity vector), strain rate, and strain in the joint area, which indicates the aspect of the deformation process accompanying the joining phenomenon, was calculated, and quantitative information necessary for joining was obtained. It can be applied to design. This series of analysis processes is also performed in an interactive manner using the central processing unit 70 and the unsteady deformation analysis program 80.
It is possible to enlarge and visualize the deformation process during the bonding process occurring inside the bump 4 and the inside of the bump 4, perform highly accurate strain analysis and mechanical analysis, and grasp the correlation between the bonding process and the deformation behavior. FIG. 15 shows a detailed system configuration of the sensor unit, the instrumentation unit, and the central processing unit.

FIG. 2 is a block diagram of the entire system according to the second embodiment of the present invention. The difference from the first embodiment shown in FIG. 1 is that an electric actuator is used as the actuator, and the pulse motor 1 is used as the press mechanism 10 ′.
1 ', piston 12', upper punch 13 ', pinion 1
4 'and a rack 15', and a servo mechanism is used in the press control unit 50 ', so that the pressing speed of the upper punch 13' can be controlled with high accuracy, and the transmission of the pressing force is also performed by the pinion 14. '· The reliability is increased by the gear transmission mechanism of the rack 15'. Further, on the computer 71 of the central processing unit 70 using the press dynamic characteristic generation program 91, a displacement-time curve (formula 1) and a press speed-time curve (second equation) showing the characteristics in an interactive manner. Equation) and press load-time curve (Equation 3) can be designed, which is equivalent to designing the manufacturing process conditions in actual products.

The above first to third formulas are as follows.

[0019]

[0020]

[0021]

This information is sent to a variable frequency pulse oscillator 55 '.
And the pulse motor 11 'is driven at a required speed according to the pulse frequency and the pulse number. At the same time, the resolver 51 'detects the accuracy of the drive by the rotation angle, and this is detected by A
The signal is converted into a digital signal by a / D converter 53 ', compared with an initial set value by a comparator 54', and if a difference occurs, it is corrected by an acceleration / deceleration circuit 52 'to
Send information to 1 '. The subsequent analysis process using the model material sample 30, the unsteady deformation analysis program 80, and the measurement / control / processing program 90 is the same as in the first embodiment.

The above is a description of the flow of a series of operations, analysis, and processing when the systems of the first embodiment and the second embodiment of the present invention are used. This will be described in more detail with reference to FIG. In addition, FIG.
This is a diagram showing the flow of the conventional design optimization, and the basis of the optimization is the same for the collective bonding type shown in FIG. 4 and the single point bonding type shown in FIG. Most of the information is based on empirical information from a skilled technician, and is a part that is compared with the superiority of the present invention.

FIGS. 6 to 9 are flowcharts showing the process of analysis and experiment when the system of the present invention is used.

First, tempering of the deformed model material is performed.
As a deformation model material, a clay-like material called trade name plasticine is used. Even if it is made of gold, copper, aluminum, etc. like lead or actual product, it can be replaced, but these metals are disadvantageous in that the processing power required for pressing is very large, so the above clay-like material Is preferred.

The quality of plasticine is adjusted by adding and drying water (for a certain period of time), cooling and heating (for a certain period of time) so that the deformation characteristics and hardness of the leads 3 and bumps 4 of the actual film carrier tape are similar to those of the plastic carrier. These are also controlled for a certain period of time) or by adding a solid substance.

Subsequently, in order to quantitatively know the deformation characteristics of the same material after tempering, the press mechanisms 10, 10 'of the present system are used.
Perform a uniaxial compression test using. At this time, the model material is used after being formed into a cylindrical shape. This is exactly the same as the means for knowing the deformation characteristics of other industrial materials. The outputs of the load cells 22 and 22 'and the displacement meters 21 and 21' at the time of the test are measured and processed by the measurement control processing program 90 and the central processing unit 70 to calculate a deformation resistance curve, which is transferred to the floppy disk drive 72. Record the numerical data of the same curve.

Next, a lead model sample 37 and a bump model sample 32, which are experimental samples, are placed in a Teflon mold so that the above-described model materials are similar in shape and size to the bumps and leads 3 of the actual product. Is molded. After forming, for grid line analysis (for analysis used for transient deformation analysis)
Is transferred to an arbitrary cross section of the sample by a rubber stamp or the like. The cross section to be transferred is obtained by cutting the same sample with a cutter knife or the like, and obtaining the cut surface. FIG. 11 shows a state of drawing. In FIG. 11, two cross sections orthogonal to each other are illustrated. After the grid lines are drawn, the cut surfaces are realigned to form an integrated sample, which is pressed down and deformed by the upper punch 13 of the press mechanism shown in detail in FIG. In order to visualize the unsteady rolling process, the stopper 13d shown in the figure is prepared by various methods, and the stopper 13d corresponding to a required deformation amount is used. In other words, the rolling process in several stages from the initial stage of rolling and deformation to the final rolling is divided according to the difference in the total length of the stopper, and the deformation information of each time is obtained, and the analysis of the unsteady deformation process is applied. become.

The progress of the deformation step is shown in FIG.

Subsequently, an unsteady deformation analysis program 80 is prepared, and the digitizer 73 reads the change of the grid line at each deformation stage from the enlarged copy photograph 33. The portion to be read is the intersection of the free surface contour of the cross section of the model material sample 30 shown in the copy photograph 33 with the grid line drawn on the same cross section and the grid point which is the cross point of each grid line in the same cross section. Each position. Thereby, an arbitrary deformation state can be digitized. Since the position of the same grid point changes (some points do not change) as the deformation progresses, the amount of change, that is, the displacement is calculated, and the deformation speed of each corresponding deformation stage is calculated using the time required for the deformation. (Speed vector) is calculated. Next, since the displacement of each lattice point is calculated, the change in the contact area of the joint where the lead model sample 31 and the bump model sample 32 are in contact with each other can also be calculated. This is a result of an effect of increasing the surface area due to the exposure of the new surface due to the deformation, and it is possible to estimate the effect of improving the bondability by knowing the amount of the enlargement. The state of this deformation is graphically displayed on the graphic screen of the computer 71.

Subsequently, the velocity vector of each grid point obtained in this way is calculated by interpolation on the characteristic curve. The characteristic curve group is a mathematical square grid line group provided so as to be superimposed on the cross section of the lead model sample 31 and the bump model sample 32, and is a grid line transferred by a rubber stamp or the like on the cross section of the sample. Is a different group of grid lines. Therefore, hereinafter, for convenience of description, a grid line transferred on the sample cross section (a grid line for obtaining deformation information) is referred to as a cross section grid line, and a newly superimposed square grid line group is referred to as a square analysis grid line. (Grid lines for performing numerical analysis). The former changes its shape / pattern as the deformation progresses (FIG. 12), while the latter does not change its shape irrespective of the deformation because it is a grid provided mathematically for numerical analysis. Since the free surface contour of the cross section must be superimposed so as to surround it, as the deformation progresses, the free surface contour of the sample cross section of the area to be surrounded also changes every moment, so the size of the square analysis grid frame also changes I have to do it. One of the points that requires an improvement in analysis is to make the unsteady deformation process of the free surface contour mathematically recognized. However, the unsteady deformation analysis program 80 of the present system uses the digitizer 73 to read the sample cross section. From the coordinate value indicating the change of the free surface and the position of the grid point determined by the grid line interval of the square analysis grid line described later, the position of the free surface is searched and recognized in which square analysis grid at that time. Solved using the method.

The square analysis grid line is mathematically expressed by the following fourth
Equation (5) In an xy orthogonal coordinate system,

[0033]

Therefore, the grid line spacing is

[0035]

ΔC is an arbitrary constant, and the value of ΔC can be set large or small according to the analysis accuracy in the unsteady deformation analysis program. Mathematically, it corresponds to the difference interval of the difference scheme for performing the difference calculation. The use of such a square analysis grid line, that is, a group of characteristic curves, is based on the fact that the strain velocity is originally given by a first-order partial differential system of the same position variable of a velocity vector at a certain point, and is converted into an ordinary differential form. This is for the purpose of conversion to facilitate numerical calculation. In other words, when the velocity vector of each grid point of the cross-sectional grid line at any deformation stage is interpolated onto each grid point on the square analysis grid line, a square is obtained from the value of the speed vector after interpolation and the position of the same grid point. The strain rate can be obtained by the difference calculation along the above equation of the analysis grid line, that is, the characteristic curve group. The ordinary differential equation obtained at that time is the following equation (6).

[0037]

Further, when an equivalent strain rate equivalently expressing these as one parameter is obtained, the following equation (7) is obtained.

[0039]

This equation (7) is equally applied to all grid points.

Then, the equivalent strain (ε bar), which is an important parameter linking the deformation behavior and the joining process, is calculated from the equivalent strain rate in the equation (7). When a considerable strain is generated particularly at the joint surface of the joint, the joint strength at that portion is also high. In other words, a large number of new faces required for bonding appear, indicating that the highly active faces are in an environment where they are strongly linked to each other by mass transfer.
Therefore, the equivalent strain has a meaning as a parameter that can be uniquely and quantitatively compared with this under any material and under any condition. This equivalent strain is
In any deformation step of interest, it can be obtained by the following eighth equation.

[0042]

Actually, the above equation is performed by numerical integration along the time axis of the deformation stage. Then, it moves to the calculation of the deformation resistance.

First, it is selected whether the object is a deformed model material or an actual material. If you select a deformed model material,
When an experiment was performed using this system, the data of the deformation resistance curve of the same material, which was measured in advance, was selected from the floppy disk device 72, and the equivalent strain rate and the equivalent strain distribution described above were used using the data (numerical data). Then, if necessary, the deformation resistance distribution of the joint region corresponding to the temperature T at the time of the experiment is obtained. At the same time, a surface pressure distribution generated on the bottom surface of the bump model sample 32 is calculated based on the deformation resistance distribution. The surface pressure distribution can be easily obtained from the fact that it acts as a reaction of the deformation resistance distribution. Further, it is determined whether or not the relationship with the load at the time of pressing (total rolling force) p satisfies the following equation (9). You can check the balance of power.

[0045]

The distribution pattern described above is graphically displayed on the computer 71.

Next, when a real material is selected, a general single-axis compression tester for the real material or a micro-load compression test device capable of applying a micro load to a micro area and detecting the micro displacement at that time is used. The mechanical properties and deformation resistance curves of thin film forming materials such as plated films that form microstructures such as bumps 4 and leads 3 are determined, deformation resistance formulas are determined, and these formulas are incorporated into an analysis program. From the changes in the strain rate, strain, and temperature T in the joining region in the joining process, the change in the deformation resistance is calculated using the deformation resistance estimation method based on the deformation resistance equation and the work hardening rate equation of the actual material to obtain the same distribution. Hereinafter, the surface pressure distribution is obtained, the state is graphically displayed, and the examination is performed.

According to the above-described analysis process, it is possible to simultaneously control the generation of equivalent strain contributing to bonding and control the mechanical damage to the substrate (pellet) constituting the semiconductor device. It will be possible to obtain guidelines for the design of materials, the design of jigs and tools, and the design of deformation characteristics of materials.

Therefore, next, the flow of design and evaluation using the present system will be described. FIG. 10 shows the flow.

First, the design of the dynamic characteristics of the upper die punch is performed by the press dynamic characteristic generation program 91 and the central processing unit 70 by using the displacement (reduction amount) -time curve (formula 1) and the press speed-time curve (the first curve). 2), press load-time curve (3rd
The profile is designed while graphing the expression. This corresponds to the design of the bonding process conditions of the actual product.

Next, the shape and properties of the upper punch are designed. The upper punch 13 is fixed to the upper die 13a at the press mechanism shown in the front view (A) and the side view (B) in FIG. Since it has a freely detachable structure, the experiment can be performed by changing the upper die 13 in various ways. Front view (A) showing some examples in each of FIGS.
And in the side view (B). In each case, the part to be the crimping surface is devised. FIG. 21 particularly shows a capillary of wire bonding, and shows that the bonding process can also be tested. These are all
It is designed for the purpose of controlling the metal flow that affects the bonding property in the bonding region and controlling the deformed shape and amount of deformation, and the shape and properties are exemplified in the figure. It is not limited to. Needless to say, the design of the upper die 13 corresponds to the design of a bonding tool / tool.

Next, the deformation characteristic of the deformation model material (plasticine in this example) is designed. The design method uses the method described in the description of the analysis process. The quantitative understanding of the characteristics is also as described above. This corresponds to the design of the deformation characteristics in the actual material.

Next, the structures, dimensions and shapes of the bump model sample 32 and the lead model sample 31 are designed. This also contributes to the metal flow control and deformation shape control that contribute to the bondability, so that the damage to the semiconductor device is reduced as much as possible, and the processing power applied during manufacturing is reduced (for example, lower load and lower temperature heating). ). This can be directly fed back to the design of the structure and dimensions of the bumps 4 and leads 3 in the actual product.

Lastly, it is necessary to set the range of the allowable condition at the time of actual manufacturing. In FIG. 4 and FIG. 5, changes in the degree of contact between the bonding tools 1 and 1 'and the areas where the leads 3 and the bumps 4 are overlapped due to relative positional deviations are described. The quality of the setting will be examined based on the state of occurrence of the equivalent strain and the difference in the aspect of the surface pressure distribution.

Under the above-described design specifications, the above-mentioned analysis experiment is applied, and its quality is examined. This corresponds to an evaluation process in which there was no other way but to make decisions while making prototypes by using conventional actual products.In addition, objective and quantitative information that could not be obtained by conventional prototype evaluation was obtained. This has the advantage that the design can be performed according to the design orientation.

Lastly, the structure of the pressure receiving sensor plate 41 for measuring the surface pressure distribution will be described.

FIG. 22 is a top view (A) and a front view (B).
, The pressure receiving sensor plate 41 is attached to the pressure receiving sensor embedded long plate 41C using a pressure receiving sensor 41a (for example, a micro pressure transducer PS-B manufactured by Kyowa Dengyo Co., Ltd.).
b is embedded so as to be flush with the surface of the long plate. Several long plates are arranged on the same long plate 41C so as to form a one-dimensional array in the longitudinal direction, and a long plate having the same dimensions is prepared. Prepare several sets embedded at positions shifted by the diameter of the minute.
The long plate 41C has a long plate fixing key and a key groove 71d.
Are fixed to each other by the key and the key groove, and are also fixed to the long fixing frame 41e.
Since it is fixed by the key and the key groove, it can be freely removed by sliding it as a guide. Therefore, by replacing the position where the long plate 41C is inserted with another long plate, the surface on the pressure receiving sensor plate 41 can be moved so as to scan.
Each time, by repeating the experiment under the same conditions, two-dimensional information of the surface pressure distribution can be obtained. The reason for adopting the same structure is to reduce the cost of the pressure receiving sensor plate 41a and to make it possible to perform maintenance conveniently.

[0058]

As described above, according to the present invention, it is possible to minimize the evaluation process which needs to go through the entire manufacturing process using an actual product or a prototype in a similar form by the above-described configuration. It is effective in reducing the cost of the prototype and the TAT, and at the same time, it is possible to theoretically, qualitatively, and quantitatively present specific information on a design method and an improved portion of the semiconductor device and its manufacturing method. In addition, since the process of design optimization is systematized, the process of trial and error by an expert is unnecessary, and a design guideline in accordance with the design direction can be obtained. In addition, by using this system, it is possible to solve the problems of practicality and applicability associated with the conventional computational dynamics method (such as FEM analysis), and it becomes a source of design for production engineers. There is an effect that a design support tool can be realized.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram showing a first embodiment of the present invention.

FIG. 2 is a system configuration diagram showing a second embodiment of the present invention.

FIG. 3 is a flowchart showing a flow of a conventional design process.

FIG. 4 is a schematic diagram showing batch bonding.

FIG. 5 is a schematic diagram showing single point bonding.

FIG. 6 is a flowchart describing an analysis / experiment method according to the present invention along the process.

FIG. 7 is a flowchart continued from FIG. 6;

FIG. 8 is a flowchart continued from FIG. 7;

FIG. 9 is a flowchart continued from FIG. 8;

FIG. 10 is a flowchart describing a method of optimizing design / evaluation in the present invention performed while applying the method of analysis / experiment in the present invention.

FIG. 11 is a view showing a model sample of a lead bump formed by a deformed model material (plasticine in the case of the present embodiment) used in the above-described method, and an arbitrary cross section thereof (two mutually separated in the figure). FIG. 4 is a diagram illustrating that a lattice pattern is drawn in an orthogonal cross section.

FIG. 12 is a diagram schematically illustrating a state in which the lead model sample 31 and the bump model sample 32 are pressed down by the upper die 13 and the deformation progresses, and the state of the change of the grid lines accompanying the deformation.

FIG. 13 is a view mainly showing a press mechanism in the first embodiment of the present invention shown in FIG. 1, (A) is a front view,
(B) is a side view.

FIG. 14 is a piping diagram when an air actuator is used as the actuator according to the first embodiment of the present invention shown in FIG. 1;

FIG. 15 is a detailed system configuration diagram of the sensor unit 20, the instrumentation unit 67, and the central processing unit 71 in the first embodiment of the present invention shown in FIG.

FIG. 16 is a view showing an example of a crimping surface shape of the upper die punch, (A) is a front view, and (B) is a side view.

17A and 17B are diagrams showing examples of the crimping surface shape of the upper punch, wherein FIG. 17A is a front view and FIG. 17B is a side view.

FIG. 18 is a view showing an example of a crimping surface shape of the upper die punch, (A) is a front view, and (B) is a side view.

FIG. 19 is a view showing an example of a crimping surface shape of the upper die punch, (A) is a front view, and (B) is a side view.

FIG. 20 is a view showing an example of a crimping surface shape of the upper die punch, (A) is a front view, and (B) is a side view.

21A and 21B are diagrams showing examples of the crimping surface shape of the upper punch, wherein FIG. 21A is a front view and FIG. 21B is a side view.

FIG. 22 is a structural view showing a pressure receiving sensor plate, and FIG.
Is a top view, and (B) is a front view.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Batch bonding tool 1 'Single point bonding tool 2 Crimp surface of batch bonding tool 3 Lead of film carrier tape (TAB tape) 4 Bump 5 Pellet 6 Bonding stage 10, 10' Press mechanism 11 Cylinder (air actuator) 11a Piston position Sensor 11 'Pulse motor (electric actuator) 12, 12' Piston 13, 13 'Upper punch 14' Pinion 15 'Rack 13a Upper die 13b Guide groove 13c Bush 13d Stopper 13e Stopper receiver 20, 20' Press sensor 21, 21 ′ Displacement type, displacement sensor 22, 22 ′ Load cell 22a Load cell attachment 23 ′ Speed sensor 30 Model material sample 31 Lead model sample 32 Bump model Sample 33 Copy photo of sample cross section 40 Stage part 41 Pressure receiving sensor plate 41a Pressure receiving sensor 41b Sensor pressure receiving surface 41c Pressure receiving sensor embedded long plate 41d Long plate fixing key 41e Long fixing frame 41f key 41g Key groove 50, 50 'Press control Part 51 Solenoid valve 51a Direct acting valve 51b Flow control valve with check valve 51c High relief regulator 52 Gauge 53 Electropneumatic regulator 54 Press control controller 51 'Resolver 52' Acceleration / deceleration circuit 53 'A / D converter 54' Comparator 55 ′ Variable frequency pulse oscillator 60 Instrumentation unit 61 Amplifier 61a Signal conditioner 61b Displacement amplifier 61c Multi-channel operational amplifier 62 Instrumentation system 62a Digital memory 62b Storage scope 70 Central processing unit 71, 71a Yuta, PC-9801 personal computer 72 a floppy disk device 73,73a digitizer, KL-4000 74,74a x-y plotter, MP-4300

Continuation of the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) H01L 21/603 H01L 21/60 311 H01L 21/60 321

Claims (1)

(57) [Claims] 1. A semiconductor device which is bonded to a semiconductor device having a protruding electrode at one of a lead end portion of a film carrier tape and an external lead electrode pad provided on a semiconductor element by a thermocompression bonding process. A power transmission mechanism such as a gear such as a rack and pinion to which a cylinder / piston mechanism or a servo mechanism capable of speed control is added, or a press mechanism unit such as a feed screw mechanism; A press controller for controlling the magnitude of driving power of the press mechanism and the speed at the time of pressing; an accelerometer, a load cell,
A press sensor unit for measuring various mechanical and geometrical quantities associated with a pressing process such as a displacement gauge; and a flat or uneven surface or a crimping surface corresponding to a bonding tool of a semiconductor manufacturing apparatus added to a part of the press mechanism. An upper punch portion that has been subjected to processing such as a taper, a stopper portion that can arbitrarily change the press amount of the upper punch, and the protrusion-type electrode, where the upper punch gives a pressure reduction and deformation by a pressing process. A sample made of a deformed model material in which lattice lines are drawn on an arbitrary cross section made by imitating a lead of a film carrier tape, and a sample made of the model material is supported, and the sample is pressed by the upper punch. The pressure receiving sensor that detects the surface pressure distribution generated on the bottom surface of the sample when the sample is drawn is developed in a lattice point or is arranged one-dimensionally on a certain straight line. A pressure receiving sensor plate, a stage unit in which the pressure receiving sensor plate is incorporated, and a guide bar provided so that the upper punch presses vertically downward, and a transient output from the press sensor unit and the pressure receiving sensor plate. An instrumentation unit that reads and stores each voltage change under the phenomenon in real time, and controls the instrumentation unit to calibrate the voltage change indicating a transient phenomenon in the pressing process to actual various mechanical quantities and geometric quantities. In addition, a series of procedures for quantitatively reproducing and editing the press characteristics by storing the data after the calibration on a peripheral device such as a floppy disk, graphing it on a display screen or a printer and performing graphic output, and A central processing unit such as a personal computer in which the procedures are incorporated in internal and external storage devices and the like and in which various procedures are managed and executed; A series of procedural means for arbitrarily designing the press characteristics interactively on a display screen and transferring the information to the press control unit; and Copy photographic means which captures the pattern of the grid lines before pressing and the pattern of the grid lines in any of the pressing steps; and
An image capturing device such as an image scanner or a pointing device such as a digitizer that reads as coordinate values in a rectangular coordinate system, and the image capturing device or a pointing device is controlled, and a grid line is formed from data of each coordinate value of a grid point of the grid line. A design support system for a method of manufacturing a semiconductor device, comprising: a numerical analysis program for performing an unsteady deformation analysis by an analysis method.