JPH09329649A - Method and device for automatically analyzing lsi failure - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device for automatically detecting the part causing operation failure of LSI without target LSI circuit connection information as design data in target LSI, and the information of the logical expectation value of each part at the time when a test signal is applied. SOLUTION: A test signal supply device 1 supplies, a test signal which reproduces failure operation in failure LSI under control by a measuring control mechanism 13, LSI mounted at a position to be measured in a vacuum chamber 3 is irradiated with an electronic beam by an electronic beam test system 2. A secondary electronic amount is measured in response to irradiation for a designated part, an LSI switching mounting control mechanism 11 is mounted at the position to be measured in accordance with one of good and failure LSI 19, 8 in the vacuum chamber 3 and made a state in which the test signal is applied from the test signal supply device 1. The potential of the designated part is judged on the secondary electronic amount measure with a potential judgement mechanism 12. The measuring control mechanism 13 finally specifies the failure cause part of the failure LSI or designates the selected range of the failure cause part.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an LSI failure automatic analysis apparatus and analysis method, and more particularly, to an L method using an electron beam.
The present invention relates to an LSI failure automatic analysis device and analysis method for searching for a defective portion of SI.

[0002]

2. Description of the Related Art An electron beam tester is one of the effective means for searching for defective locations of LSIs. A test signal that reproduces a defect generated by an LSI tester or the like is applied to the defective LSI in the chamber, and then the surface of the LSI is irradiated with an electron beam, and the amount of secondary electrons is measured, thereby increasing the amount. The relative potential can be read based on whether the potential is relatively low or the potential is relatively high.

While referring to the circuit connection information of the target LSI and the information of the logic expected value at each location when a test signal is applied, the defective potential propagates inside the LSI from the abnormal output location while comparing with the potential during normal operation. It is possible to find the cause of the defect by a method such as tracing the route in reverse.

An apparatus for automatically searching for the cause of a defect using such an electron beam tester is disclosed in Japanese Patent Laid-Open No. 63-1244.
No. 38 is proposed. In this device, as the design data for the target LSI, the circuit connection information of the target LSI and the information of the logic expected value at each location when the test signal is applied are used, and the potential measured in the defective LSI is used as the logic expected value. By comparing the potentials, the potential causes of defects are pointed out by measuring the potentials at various places.

[0005]

The problem of such a conventional device is that the circuit connection information of the target LSI and the information of the logic expected value at each point when the test signal is applied are required as the design data of the target LSI. It is that you are. LSI
These data may not be obtained depending on the progress of correction at the time of design / manufacture and the situation at the time of failure analysis.
In such a case, such a device cannot be used.

Therefore, the present invention has been made to solve the above-mentioned problems of the prior art, and its purpose is to provide circuit connection information as design data for the target LSI and various points when a test signal is applied. It is an object of the present invention to provide an LSI failure automatic analysis apparatus and analysis method that can automatically search for the cause of an LSI operation failure without requiring the logical expected value information and the like.

[0007]

An LSI automatic analysis apparatus according to the present invention is a test signal for alternately supplying a test pattern signal to a good product and a defective product LSI which have been tested beforehand to determine good and defective products. Supply means, means for measuring the secondary electron amount of the LSI to which the test pattern signal is supplied while irradiating the electron beam, and the secondary electron amount of the non-defective and defective LSIs by the measuring means. And an analyzing unit for analyzing a defect occurrence state of the defective LSI based on the difference.

The analyzing means determines the potential of the irradiation position according to the measured secondary electron amount, and analyzes the defective state based on the difference in the potential. .

Further, the analysis means selects the pattern of the test pattern signal in which an abnormality is observed at an output terminal as a target pattern and the output terminal having the abnormality as an initial difference range, and the target pattern is selected. In the above, the difference between the potentials of the non-defective product and the defective LSI is sequentially detected, and these different parts are added to the initial difference range to update it as the latest difference range. It is characterized in that a defect occurrence candidate range is obtained based on.

Further, the analysis means switches to the pattern before the pattern of interest, sequentially detects the potential difference portions between the non-defective product LSI and the defective product LSI in this pattern, and updates the latest difference range. Then, these processes are repeated while sequentially switching the pattern of interest to the previous pattern, and the difference range finally obtained is set as the defect occurrence candidate range.

The LSI failure automatic analysis method according to the present invention comprises:
Alternately supplying a test pattern signal to a good product and a defective product LSI which have been judged as good or defective by being tested in advance, and irradiating the LSI to which the test pattern signal is supplied with an electron beam And a step of measuring the secondary electron amount while analyzing the defect occurrence state of the defective product LSI based on the difference in secondary electron amount of the non-defective product and defective product LSI by this measurement. I am trying.

Then, in the analyzing step, the potential of the irradiation position is determined according to the measured secondary electron amount,
It is characterized in that the defective state is analyzed based on the difference in the potential.

In the analysis step, the pattern of the test pattern signal in which an abnormality is observed at an output terminal is selected as a pattern of interest, and the output terminal having the abnormality is selected as an initial difference range, and the pattern of interest is selected. In the above, the latest difference range obtained by sequentially detecting the points where the potentials are different between the non-defective item and the defective item LSI and adding these different points to the initial difference range to update the latest difference range It is characterized in that a defect occurrence candidate range is obtained based on.

Further, in the analyzing step, the pattern is switched to the pattern before the pattern of interest, the portions where the potential is different between the non-defective and defective LSIs in this pattern are sequentially detected, and the latest difference range is updated. Then, these processes are repeated while sequentially switching the pattern of interest to the previous pattern, and the difference range finally obtained is set as the defect occurrence candidate range.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The operation of the present invention will be described. The test pattern signals are selectively applied alternately to the good and defective LSIs that have been previously determined, and the LSI to which the test pattern signals are applied is irradiated with an electron beam, whereby the amount of secondary electrons is increased. Is measured to detect the potential of the electron beam irradiation location. Good product and defective product L
Based on the difference in the potential between the SI and the SI, the defective cause portion or the defective cause portion candidate range, which is the defective state of the defective LSI, is specified.

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention. Referring to FIG. 1, the test signal supply device 1 has a function of generating a test signal that causes a defective LSI to perform a defective operation, and is capable of performing a control operation under the control of the measurement control mechanism 13. I will keep it. It is convenient to use a device that has a built-in memory for a test pattern and has a mechanism for generating a registered defect reproduction test pattern as a test signal.

The electron beam test system 2 is provided with an electron beam source 4 and a secondary electron detector 5 for measuring the amount of secondary electrons in a vacuum chamber 3. An electron beam is emitted from an electron beam generation source 4 to an LSI or the like mounted on a measured position to measure the amount of secondary electrons. A device that can be controlled by execution from another device and can incorporate an LSI switching mounting mechanism 11 described later in the chamber is used.

Instead of outputting the secondary electron amount itself,
Generates and outputs time-series waveform data in synchronization with a test signal applied to an LSI, etc., or generates and outputs a potential distribution image in a two-dimensional area by raster scanning an electron beam in a certain two-dimensional area. The measurement control mechanism 1
The configuration is such that it operates based on the control from 3.

The LSI switching mounting mechanism 11 is provided with a mechanism for mounting one of the two LSIs 8 and 9 in the vacuum chamber at the measured position under the control of the measurement control mechanism 13. The test signal from the test signal supply device 1 is electrically connected to the mounted LSI.

The potential measuring mechanism 12 determines the potential based on the amount of secondary electrons measured by the electron beam test system 2. The measurement control mechanism 13 determines the LSI and the location to be measured next based on the measured potentials of the non-defective product 9 and the defective product 8 obtained from the potential determination mechanism 12, and further, the signal supply device 1, the electron beam test system 2, LSI switching mounting control mechanism 1
1 and the potential determination mechanism 12 to control the required potential measurement, and identify the defective product cause location in the defective LSI based on the obtained potentials of each LSI and each location, or determine the defective cause candidate range. Point out.

Next, the operation of the present invention will be described. Under the control of the measurement control mechanism 13, in the vacuum chamber 3 of the electron beam test system 2, one of the good product and the defective product is mounted at the measured position according to the control by the LSI switching mounting mechanism 11. Further, the test signal supply device 1 supplies a test signal for reproducing a defective operation, and the test signal is applied to the LSI on which it is mounted.

With the electron beam test system 2, the LSI mounted at the measured position in the vacuum chamber 3
The electron beam is irradiated, and the amount of secondary electrons at the designated location is measured. 2 measured by the potential determination mechanism 12
The potential at the designated location is determined based on the amount of secondary electrons, and this is sent to the measurement control mechanism 13. By such a series of operations, the potential of the designated location is measured for the designated LSI under the control of the measurement control mechanism 13.

The measurement control mechanism 13 repeats such a measurement function and measures the potential of the defective product while comparing it with the potential of the non-defective product, thereby focusing on the different portion and tracing the defective cause portion. By continuing this, the location of the defect cause is finally specified or the defect cause candidate range is pointed out.

Further, an embodiment of the present invention will be described in detail with reference to FIG. In FIG. 1, as the test signal supply device 1, a normal LSI logic tester capable of execution control under the control of another device may be used.
A program that generates a test pattern that causes a malfunction may be prepared and executed. In order to average the secondary electron amount measurement and improve the accuracy, it is preferable to repeatedly generate the test pattern.

As the electron beam test system 2, if a normal electron beam tester capable of execution control under the control of other equipment and capable of incorporating the LSI switching mounting mechanism 11 in the chamber 3 is used. good.

The LSI mounted inside the chamber
On the other hand, the electron beam generated by the electron beam generation source 4 is irradiated, and the amount of secondary electrons corresponding thereto is detected by the secondary electron detector 5. The physical observation position is moved according to the designation of the measurement point from the measurement control mechanism 13. The load module 7 has a shape in which connection pins are arranged on the back surface for the LSI switching mounting mechanism 11.

As the LSI switching mounting mechanism 11, a publicly known one can be used, but for example, a structure shown in Japanese Patent Application No. 6-301630 may be used. In the vacuum chamber, a selected one of good and defective LSIs is mounted on the device under test, that is, at the position where the electron beam is irradiated, and the supplied test signal is applied. The test signal supplied from the test signal supply device 1 is sent via the connection cable 10 to the load module 7 having the connection pins on the back surface for the LSI switching mounting mechanism, and the top surface is electrically connected to the load module side. Connection pins for contact are arranged, and on the lower surface, through a movable chip mounting base 6 with an IC socket prepared, a non-defective LSI 9 or defective LS
A test signal is sent up to I8.

The potential determination mechanism 12 determines the potential based on the amount of secondary electrons measured by the electron beam test system 2. When the potential of the point irradiated with the beam is low, the amount of secondary electrons is large, and when the potential is high, the amount of secondary electrons is small. Therefore, binarization may be performed based on the amount.

When an ordinary electron beam tester is used as the electron beam test system 2, the data (waveform data) obtained by time-sequentially acquiring the secondary electron amount for each test signal timing with respect to the test signal repeatedly input. Is generated. Alternatively, in the specified test pattern, the electron beam is raster-scanned to the two-dimensional observation area, and 2
A potential distribution image is generated by measuring the amount of secondary electrons.

The potential determination mechanism 12 uses these waveforms or potential distribution image data, and binarizes them based on the data to obtain time-series data at the same point and two-dimensional distribution data at the same timing. To be

In the measurement control mechanism 13, the LSI and the location to be measured next are determined based on the obtained measured potentials of the non-defective product and the defective product, and based on the finally obtained potentials of each LSI and each location. The defective cause point in the non-defective LSI is specified or the defective cause point candidate range is pointed out. As one of the methods, there is a method in which a potential difference between a good product and a defective product is investigated while taking a potential distribution image. Figure 2 shows the method
Shown in

Referring to FIG. 2, there is shown an operation flowchart for searching a defect occurrence candidate range of a defective LSI. As the initial pattern of the test signal pattern, select the pattern in which an abnormality is observed at the output terminal as the pattern of interest, and use the abnormal output terminal as the initial range of the difference in measured potential between the good product and the defective product. Select (step 21).

The new difference range in the pattern of interest is obtained as follows, but it is obtained by sequentially repeating the processing of steps 22 to 25. That is, in a non-measured portion around the initial difference range, the potentials of the target pattern of the non-defective product and the defective product are checked (step 22,
23) If they are different (step 24), the different part is added to the initial difference range and updated as the latest difference range (step 25).

The above processing (steps 22 to 25) is repeated until no difference occurs. After that, the pattern is switched to the pattern before the target pattern, and it is checked whether or not there is a difference in potential between the non-defective product and the defective product at each location within the latest difference range (steps 26 and 27). If there is a difference, the pattern at that time is set as the pattern of interest, the difference range is set as a portion where the difference exists (step 28), and the above-described processing (steps 22 to 25) is repeated.

If there is no difference, the latest difference range obtained at that time is set as the defect occurrence candidate range and the process is terminated (step 29). In this way, the defect cause candidate range can be specified for the sequential circuit of the LSI.

As another method, the mask data of LSI /
There is a method to refer to layout data. Since the mask data of this LSI is always prepared for mask creation regardless of any design changes, as described in the section of the prior art, each circuit connection information of the target LSI and each test signal application time are applied. This method is also advantageous due to the fact that it is easier to obtain than data such as logical expected value information at a location.

As an example of this method, there is a method shown in "Application of charged particle beams to industry") LSI Testing Symposium / 1995 in 132nd Research Material, 132nd Committee, Japan Society for the Promotion of Science, 1995. . In that method, first, circuit connection information at the transistor level is extracted and used based on the layout data of the LSI. Using the layout information and the circuit connection information, the defect cause candidate range is searched by comparing with the potential of the non-defective product.
Basically, by referring to the connection relationship of MOS transistors, the measurement points are set and the potentials of good and defective products are measured, and the points where the potentials of good and defective products differ are traced back to the upstream of the circuit. This is a method of measuring and finally pointing out the location where the defect occurs. This method further utilizes the hierarchical structure of the layout to improve efficiency.

Further, the measurement control mechanism 13 controls the necessary potential measurement by controlling the signal supply device 1, the electron beam test system 2, the LSI switching mounting control mechanism 11 and the potential determination mechanism 12.

[0040]

The device of the present invention can be used even when the circuit connection information of the target LSI or the information of the logic expected value of each portion at the time of applying the test signal, which cannot be conventionally applied, cannot be obtained.

[Brief description of drawings]

FIG. 1 is a diagram for explaining an example of the present invention.

FIG. 2 is a flowchart showing a method of searching for a cause of failure and pointing out a candidate of a failure cause in the measurement control mechanism.

[Explanation of symbols]

 1 Test Signal Supply Device 2 Electron Beam Test System 3 Vacuum Chamber 4 Electron Beam Source 5 Secondary Electron Detector 6 Chip Mount 7 Load Module 8 Defective LSI 9 Good LSI 10 Signal Line (Cable) 11 LSI Switching Mounting Control Mechanism 12 Potential determination mechanism 13 Measurement control mechanism

Claims (8)

[Claims] 1. A test signal supply means for alternately supplying a test pattern signal to a non-defective product and a defective product LSI, which have been judged as good or defective by being tested in advance, and this test pattern signal is supplied. A means for measuring the secondary electron amount of the LSI while irradiating it with an electron beam, and a defect occurrence state of the defective LSI based on the difference in the secondary electron amount of the non-defective and defective LSIs by the measuring means. An LSI failure automatic analysis device, comprising: 2. The analyzing means determines the potential of the irradiation position according to the measured secondary electron amount, and analyzes the defective state based on the difference in the potential. The LSI failure automatic analysis device according to claim 1. 3. The analysis means selects a pattern of the test pattern signal in which an abnormality is observed at an output terminal as a target pattern and an output terminal having the abnormality as an initial difference range, and the target pattern. In the above, the difference between the potentials of the non-defective product and the defective LSI is sequentially detected, and these different parts are added to the initial difference range to update it as the latest difference range. 3. The LSI failure automatic analysis apparatus according to claim 2, wherein the failure occurrence candidate range is obtained based on the above. 4. The analysis means switches to a pattern before the pattern of interest and sequentially detects a portion of the potential difference between the non-defective item and defective item LSI in this pattern and updates the latest difference range. Then, these processes are repeated while sequentially switching the pattern of interest to the previous pattern,
4. The LSI failure automatic analysis device according to claim 3, wherein the difference range finally obtained is set as the failure occurrence candidate range. 5. A step of alternately supplying a test pattern signal to a non-defective product and a defective product LSI, which have been judged as good or defective by being tested in advance, and to the LSI to which this test pattern signal is supplied. Measuring the secondary electron amount while irradiating an electron beam with the electron beam, and analyzing the defect occurrence state of the defective product LSI based on the difference in the secondary electron amount of the non-defective product and defective product LSI by this measurement. An LSI failure automatic analysis method comprising: 6. The analyzing step is characterized in that the potential of the irradiation position is determined according to the measured secondary electron amount, and the defective state is analyzed based on the difference in the potential. The LSI failure automatic analysis method according to claim 5. 7. The analysis step includes selecting a pattern of the test pattern signal in which an abnormality is observed in an output terminal as a target pattern and an output terminal in which the abnormality is detected as an initial difference range, In the above, the difference between the potentials of the non-defective product and the defective LSI is sequentially detected, and these different parts are added to the initial difference range to update it as the latest difference range. 7. The LSI failure automatic analysis method according to claim 6, wherein the failure occurrence candidate range is obtained based on the above. 8. The analysis step switches to a pattern before the pattern of interest and sequentially detects a portion of the potential difference between the non-defective and defective LSIs in this pattern and updates the latest difference range. 8. The LSI failure automatic analysis according to claim 7, wherein the pattern of interest is sequentially switched to the previous pattern and these processes are repeated to set the finally obtained difference range as the failure occurrence candidate range. Method.