JPH0376243A - Test of semiconductor wafer formed with a plurality of ic chips - Google Patents

Abstract

PURPOSE:To facilitate the stepup of a prober by a method wherein which contact pad among contact pads in directions X and Y on an IC chip is shifted from the probe of a probe card is detected and not only the amounts of the shifts in the directions X and Y but also the amount of the shift in a direction thetais calculated from information on these shifts. CONSTITUTION:An additional correcting device 1 is between an LSI tester 10 and a control device 2. Contact failure pin information from the tester 10 is inputted in a test data loading register 12 through an exclusive bus 10a. A CPU (a central processing unit) 13 controls the whole device 1 and executes a correction algorithm 14 as well. Correction instructions 15, 16 and 17 in directions X, Y and theta, which are outputted from the algorithm 14, are given to the device 2.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a method for testing a plurality of IC (integrated circuit) chips formed on a semiconductor wafer, and in particular, to a method for testing a plurality of IC (integrated circuit) chips formed on a semiconductor wafer.
This invention relates to improving the alignment method between the tip and the probe needle.

[Prior Art] FIG. 11 is a block diagram showing a conventional prober device. This blow μ consists of an LSI tester main body 10, a measuring section 6 of the tester, a probe card 11 having a probe needle 11a for measuring IC chips, a chuck 9 for placing a wafer 8 on which a plurality of IC chips are formed, and a probe card 11 for placing the probe card 11 on a wafer. A probe card drive device 4 that rotates the chuck 9 in the θ direction in a plane parallel to the wafer, a chuck drive device 5 that moves the chuck 9 in the orthogonal X, Y directions, and the θ rotation direction in the city parallel to the wafer, and a probe card drive device 4 and a control device 2 that controls the chuck drive device 5, a capacitive sensor and alignment detection device 7 that senses the thickness of the wafer 8, and a position detection bag f3 that recognizes the position by the sensor and the detection device 7.

FIG. 12 is a flowchart showing the flow of a conventional wafer test performed using the prober of FIG. 11.

First, in step 24, information about the IC chips on the wafer to be measured is input to the prober as set values. These set values include chip size (X, Y), wafer size, and fail count that allows you to stop measurement depending on the number of failures. In step 25, a probe card, which is a measurement jig, is attached. No angle adjustment of the probe card is performed at this stage. Step 26
At , a wafer to be measured is inserted.

Usually, measurement is performed in lots (8 to 24 wafers), so all of the wafers are placed in a cassette, and one of the wafers is placed on the chuck 9. In step 27, the chuck 9
The first wafer placed on top is aligned. Alignment is the horizontal
and aligning the direction of the scribe line for cutting the wafer 8 into IC chips in the vertical Y direction. After the alignment is completed, the chuck 9 moves directly below the probe card.

Thereafter, as shown in FIGS. 13A, 13B, and 13C, fine adjustments are made manually using the chuck drive device 5 in order to align the needles 11a of the probe card 11 with the contact pads 8b of the chip to be measured 8a. be done.

Here, FIG. 13B is an enlarged view of one IC chip 8a in the wafer 8 of FIG. 13A, and FIG. 13C is an enlarged view of one IC chip 8a in the wafer 8 of FIG.
3B is a further enlarged view of a part of FIG. 3B. By the way, the probe card needle 118 cannot be adjusted only by adjusting the chuck 9.
The θ direction of is shifted from the θ direction of the chuck 9. This means that the horizontal and vertical directions X and Y of the chuck 9 are the horizontal and vertical directions between the probe needles due to the alignment function of the prober.
This is because while the probe card 11 matches Y, it does not match the probe card 11. Therefore, the angle θ of the probe card 11 is adjusted at this stage. This adjustment is performed manually by visually dropping the needle 11a onto the contact pad 8b and checking the needle mark 11b using the card drive bag f14. In step 29, a contact test is performed on the chip 8a using the LSI tester 10,
If this passes, the setting of the wafer 8 is completed.

Fig. ff113B and Fig. 13C show this visually observed state. With the above steps, the setup of the prober device is completed.

After setting up the blower device, a wafer test is started in step 30. In step 32, the chip 8a on the first wafer 8 used when setting up the apparatus is measured. According to the measurement result of this chip, if it is a fail, the fail count is increased in step 33 and the process moves to the next chip in step 35. Conversely, if it is a bus, the process directly proceeds to step 35 and moves to the next chip. Then, in step 36, the fail count number is compared with a set value, and if it reaches the set value, the wafer to be measured is rejected as some kind of abnormality, and step 37
The wafer is stored in a cassette. Naturally, even if there are no unmeasured chips on the wafer, the wafer is stored in the cassette. In step 38, the number of wafers after the second one is determined, and if there are wafers after the second one, alignment is performed in step 39, and the process proceeds to wafer measurement in step 32, and if there are no wafers, the process ends. The flow is as follows.

[Problem to be solved by the invention] In the conventional test method as described above, the probe card 1
The θ adjustment of 1 relies on manual visual inspection. However,
As the number of probe needles increases due to the increase in the number of pins on the chip to be measured, and as the distance between the probe needles (depending on the pad size) decreases, it becomes difficult to visually observe the needle marks. Additionally, the possibility of needle misalignment (defective contact) occurring is low for the first wafer used when setting up the equipment, but for subsequent wafers, a certain level of pin contact testing is required without performing a manual pin contact test. Since the test starts automatically after alignment, the probability of needle misalignment increases.

Therefore, there is a problem in that a simple fail count function results in a large number of wafers being rejected.

In view of these problems in the prior art, an object of the present invention is to provide a method that facilitates the setup of a blow μ device, automatically corrects needle misalignment, and performs a wafer test without contact failure. That's true.

[Means for Solving the Problems] According to the present invention, a method for testing a semiconductor wafer on which a plurality of IC chips are formed is provided such that any contact pad in the X direction on the IC chip is misaligned with the probe needle of the probe card. Similarly, which contact pad in the Y direction on the IC chip is misaligned with the probe needle is detected as deviation information in the Y direction from the tester measurement results. By relatively moving the wafer and probe card based on the Y-direction misalignment information, it not only automatically corrects the X-direction and Y-direction misalignment, but also automatically corrects the X-direction and Y-direction misalignment based on the X-direction and Y-direction misalignment information. The method includes a step of also calculating a certain deviation in the θ direction and automatically correcting the deviation in the θ direction as well.

[Operation] According to the method of the present invention, which contact pad on the IC chip is misaligned with the probe needle of the probe card in the X direction and the Y direction is detected as misalignment information from the measurement results of the tester, and this misalignment information is detected. Since it calculates not only the deviation in the X and Y directions but also the deviation in the θ direction, the components in each deviation direction are automatically adjusted, making it possible to efficiently test IC chips without contact failure due to needle deviation can do well.

[Embodiments of the Invention] FIG. 1 is a block diagram showing probe traces used in the method of the present invention. The blow μ in FIG. 1 is similar to that in FIG. 11, but the LSI tester 10 and the control unit! 2 includes an additional correction device 1. FIG. 2 is a block diagram showing the internal structure of the correction device 1. Contact failure bin information from the LSI tester 10 is manually entered into the test data import register 12 via the dedicated bus 10a. A CPU (central processing unit) 13 controls the entire correction device 1 and also executes a correction algorithm 14. X output from the correction algorithm 14.

Correction commands 15, 16° 17 in the Y and θ directions are given to the control device 2.

FIG. 3A, FIG. 3B, and FIG. 3C show the IC chip 8a.
This shows the state of misalignment between the contact pad 8b and the probe needle mark 11b. First of all, the deviation to be corrected by the correction device 1 must be defined. FIG. 3A shows the deviation in the θ direction, and some bins in the X or Y direction have poor contact. FIG. 3B shows the deviation in the Y direction, and all the pins in the X direction have contact defects. FIG. 3C shows the deviation in the X direction, and all the pins in the Y direction have contact defects. In reality, these deviations in the θ, y, and X directions combine to cause contact failure. Therefore, the actual deviation can be corrected by breaking it down into deviations in the θ, Y, and X directions.

Furthermore, considering the alignment accuracy of ±20 μm and the pad spacing of 100 μm (which is also the minimum needle spacing), the degree of possible deviation is considered to be about 50 μm, and the degree that can be corrected is about one pad spacing.

FIG. 4 shows files prepared for CP013 before wafer testing. Since the present invention corrects needle misalignment based on contact failure bin information, information necessary for analyzing needle misalignment must be provided in advance. That is, the needle stand coordinate list in the X and Y directions, the tester bin list PxP7, and the size of the contact pad 8a (h,
l1l), number of corrections N1 and number of alignments M
Enter the following into the file.

FIG. 5 is a flowchart showing an example of a wafer test procedure according to the present invention. The flowchart of FIG. 5 is similar to that of FIG. 12, but additionally includes steps 14a and 14b of the correction algorithm.

FIG. 6 shows the internal routine of the correction algorithm 14. The probe card correction by this correction algorithm 14 is performed in step 14a when the contact test in the two steps when setting up the probe card 11 fails, and in step 31 when the pin contact is defective during each chip measurement on the wafer. This is executed in step 14b when this occurs.

In this case, as initial settings, the correction number counter 11, the alignment number counter J1, and the θ, Y, and X flags Iθ*IY+'X are all set to "0".

As a result of the pin contact test executed in the two steps, if it is a bus, the process proceeds to step 30 to start the wafer test, and if it is a fail, the process proceeds to step 14a and enters the defective pin analysis routine 41 in the correction algorithm 14. In the defective pin analysis routine 41, the process proceeds to correction routines 17, 15, and 16 in the θ, A decision is made whether to proceed and reject the wafer and proceed to test the next wafer. If you proceed to correction routines 17, 15, and 16 in the θ, x, and y directions, corrections will be made by each routine, and the process will return to step 2 to perform pin contact again, and repeat these corrections until a pass is achieved. . Similarly, when proceeding to the realignment in step 42, the number of times J-J+1 is counted in step 43,
A pin contact test is performed again in step 29.

That is, the maximum number of times this algorithm 14 is executed is determined by the number of corrections N and the number of alignments M set previously. The above correction algorithm 14 is executed in exactly the same way in step 14b after the contact test in step 31.

FIG. 7 shows the contents of the defective pin analysis routine 41. The defective pin analysis routine 41 takes in defective pin information F (n) from the LSI tester 10 in step 44 . In step 45, the information F (n) is compared with the above-mentioned analysis target pin Px (n) PY (n), and information Fx (n) and F
Extract Y (n). Then, in step 46, in order to detect the deviation in the X direction, it is determined whether all the pins in the Y direction are defective FY (n)-PY (n). If all the pins are defective, the correction counter is counted up (I-1+1)L in step 48, and if it is determined in step 51 that the number of corrections is less than the set number of corrections (I≦N), the process proceeds to the X-direction correction routine 15. Conversely, if it is determined in step 51 that the set number of corrections has been exceeded (1>N),
Proceed to step 54. In step 54, the number of alignments J is compared with the set value M, and it is determined whether to proceed to realignment in step 42 or to reject in step 37.

Similarly, in order to detect deviation in the Y direction, it is determined in step 47 whether all the pins in the X axis direction are defective, and the process proceeds to the Y direction correction routine 16 through steps 49 and 52, which are the same routes as described above. .

If all the pins are not defective in either the X or Y direction, it is considered that the deviation is in the θ direction, and the process passes through steps 50 and 53 to proceed to the θ direction correction routine 17.

FIG. 8 shows the contents of the X-direction correction routine 15.

In this routine, it is not known at first whether the deviation is in the positive or negative direction of the In order to move, flag (2) is first determined in step 55.

If Ix-〇, it means the first correction, and the correction device 2 issues a command to move the chuck 9 (that is, the chip 8a) in the positive direction of the X axis by a distance of half h/2 of one contact pad.
Take it out. Then, in step 57, the flag Ix is set to 1, and in step 58, the chuck 9 is moved in the positive direction of the X axis, and then the process proceeds to step 2 or 32 of the main routine, where a pin contact test is performed.

Here, if the correction direction is wrong, then of course Y
All pins in the direction are defective, and the X direction correction routine 15 is executed again.
It will come back to. At that time, the flag is set to lX-1, so in step 59, it is sufficient to move the chuck 9 in the negative direction of the X axis by a distance twice that of the first correction. . This completes the correction in the X direction.

FIG. 9 shows the contents of the Y direction correction routine 16.

This shows a procedure similar to the X-direction correction routine in FIG. 8, except that the X-direction is replaced by the Y-direction.

FIG. 10 shows the contents of the correction routine 17 in the θ direction. This routine has a flag Iθ, and in step 65 it is determined whether it is the first correction or the second correction. In the θ direction correction, the angle θ at which the probe card 11 is rotated must be calculated from the defective pin information Fx(n).

Therefore, first in step 66, the defective pin information F
F'x(n) is obtained by rearranging x(n) in the order of pads. In step 67, the rearranged defective pin information Fx (n) and the X direction pin information Px (n)
The information on the passing pin, S,(n), is obtained by comparing the S,(n). In step 68, this pass pin information Sx (
n) with the X coordinate X (xn) to find the distance #iX of the passing pin in the X direction (trailing passing pin coordinate - leading passing pin coordinate - distance). Since it is known that the probe needle is in contact with the pad of the chip for this distance X, in step 69, θ-tan-' (m/X) is calculated from the pad length of the chip and the distance X. The subsequent steps are the same as the X direction correction, and step 70
At this point, a command is issued to the control device 2 to rotate the probe card 11 by an angle θ in the positive direction.

Then, in step 71, the flag Iθ-1 is set, and in step 72, the probe card 11 is rotated by θ in the forward direction. Thereafter, the process returns to step 29 or 32 of the main routine and a contact test is executed. If there is still a contact failure, the process returns to the θ direction correction routine 17 again, and since the flag is 2 θ-1, the process proceeds to step 73. Therefore, the probe card 11 is rotated by 2θ in the negative direction and the process returns to step 29 or 32 of the main routine again.

Correction routine 15 in the X, Y, and θ directions as described above
．． If you go around 16.17, you will definitely be able to make contact even if the deviations in the x, y, and θ directions are compounded.

Furthermore, even if there is an abnormality in the chip or probe card, the wafer is rejected based on the correction counter setting value N and the realignment counter setting value M, allowing efficient wafer testing.

In the above-described embodiment, the probe card 11 was rotated to perform the correction in the θ direction, but it will be clear to those skilled in the art that the chuck 9 may be rotated to perform the correction.

[Effects of the Invention] As described above, according to the present invention, X, Y on an IC chip
Check which contact pad in the direction (X, Y) is misaligned with the probe needle of the probe card from the tester measurement results.
Not only can it detect deviation information in the X and Y directions and automatically correct deviations in the It is also possible to automatically correct any deviations. Therefore, efficient and accurate wafer testing can be performed.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing a prober device used in the method of the present invention. FIG. 2 is a block diagram showing the contents of a correction device used in the method of the present invention. FIGS. 3A, 3B, and 3C are diagrams showing the state of misalignment between the contact pad and the probe needle. FIG. 4 is a diagram showing the contents of the preparation file. FIG. 5 is a flowchart showing the procedure of a wafer test according to an embodiment of the present invention. FIG. 6 is a flowchart showing the internal routine of the correction algorithm. FIG. 7 is a diagram showing the contents of the defective pin analysis routine. FIG. 8 is a diagram showing the contents of the X-direction correction routine. FIG. 9 is a diagram showing the contents of the Y-direction correction routine. FIG. 10 is a diagram showing the contents of the θ direction correction routine. FIG. 11 is a block diagram showing a prober device used in conventional wafer testing. FIG. 12 is a flowchart showing the procedure of a conventional wafer test. FIG. 13A is a top view showing a wafer on which a plurality of IC chips are formed. FIG. 13B is an enlarged view showing one IC chip on the wafer set in the blow μ. FIG. 13C is a further enlarged view of a part of FIG. 13B. In the figure, 1 is a correction device, 2 is a control device, 3 is a position detection device, 4 is a card drive device, 5 is a chuck drive device,
6 is a test head, 7 is a capacitance sensor and alignment detection device, 8 is a wafer, 9 is a chuck, 10 is an LSI tester, 11 is a probe card, lla is a probe needle, 1
2 is a test data acquisition register, 13 is a CPU, 14 is a correction algorithm, 15 is an X-direction correction instruction, 16 is a Y-direction correction instruction, and 17 is a θ-direction correction instruction. In each figure, the same reference numerals indicate the same contents or corresponding parts.

Claims (1)

[Claims] A method for testing a semiconductor wafer on which a plurality of IC chips are formed, the tester determining which contact pad in the X direction on the IC chip is misaligned with the probe needle of a probe card. Similarly, which contact pad in the Y direction on the IC chip is misaligned with the probe needle is detected as misalignment information in the X direction, and Based on the misalignment information, not only the misalignment in the X and Y directions is automatically corrected by relatively moving the wafer and the probe card, but also the rotation is performed based on the misalignment information on the X and Y directions. A method for testing a semiconductor wafer, comprising the step of also calculating a deviation in the θ direction and automatically correcting the deviation in the θ direction.