JP3370379B2 - Method for manufacturing semiconductor device - Google Patents

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention relates to a method of manufacturing a semiconductor device.
Method , particularly, a defect inspection technique for a semiconductor wafer (hereinafter referred to as a wafer) in a semiconductor device manufacturing line, and a defect analysis technique based on this inspection result, for example, based on a foreign substance inspection, an appearance defect inspection, a probe inspection The present invention relates to a wafer failure analysis technology. The present invention can also be applied to other defect analysis techniques for magnetic disks, liquid crystal displays, printed circuit boards, and photomasks.

[0002]

2. Description of the Related Art In the process of manufacturing a semiconductor device, foreign matter and defects on the surface of a wafer as a work cause defective products. Therefore, foreign substances and appearance defects (hereinafter sometimes referred to as defects) are quantitatively inspected to constantly monitor whether or not there is a problem in the manufacturing apparatus and its surrounding environment. Then, it is necessary to grasp the influence of the foreign matter and the defect on the yield and take a countermeasure against the foreign matter and the defect effective for improving the yield as soon as possible.

By the way, a product defect can be recognized by fail chip data (hereinafter referred to as FC data) obtained by a probe inspection. Therefore, by investigating the correlation between the foreign matter inspection data or the defect inspection data and the FC data, the cause of the product defect can be analyzed.

Conventionally, the following method has been tried as a wafer failure analysis method by investigating the correlation between foreign substance inspection data and defect inspection data and FC data. (1) Defective cause analysis by a superposition map of foreign matter inspection data and appearance defect inspection data. (2) Cause analysis by correlation between the chip-level probe FC data and the foreign matter inspection data and the appearance defect inspection data. (3) Cause analysis by matching chip level probe FC data with foreign matter inspection data and appearance defect inspection data.

An example of the inspection data analysis system of this type is disclosed in Japanese Patent Laid-Open No. 3-44054.

[0006]

However, in the conventional semiconductor wafer defect analysis method, the defect analysis up to the chip level is limited.

A first object of the present invention is to provide a semiconductor wafer defect analysis technique capable of realizing defect analysis in cell units and improving the accuracy of defect analysis from macro analysis to micro analysis. is there.

A second object of the present invention is to provide a semiconductor wafer which enables early investigation of the cause of adhesion of foreign matter or appearance defect defect and clarifies the causal relationship between the foreign substance or appearance defect and the product defect. To provide defect analysis technology.

A third object of the present invention is to provide a semiconductor wafer defect analysis technique capable of early investigating the causal relationship between a foreign substance and a fail bit and the causal relationship between an appearance defect and a fail bit. .

The above and other objects and novel features of the present invention will be apparent from the description of this specification and the accompanying drawings.

Means for solving the above problems are
Occurred on a semiconductor wafer in the conductor device manufacturing process
Inspect for foreign objects or defects and at least locate and
Storing first information representing the memory, the semiconductor
Each of a plurality of semiconductor chips formed on the wafer
The electrical characteristics of the
Inspect the electrical characteristics of each of several cells to
At least the position of the semiconductor chip or each cell and
The step of storing the second information indicating the quality and the cause of the occurrence
And the second information indicating a product defect
And is compared, and there is a causal relationship in the combination of both information.
Determine the allowable error range of both position information to determine that
Determining, and the first information and the second information
By comparing and collating the combination of
Determine whether the paired error is within the allowable error range
Determining whether there is a causal relationship, and
According to the information determined to be
And a step of taking measures. this
The means shall be realized by the following system as follows:
You can That is, the foreign substance inspection device, the appearance defect inspection device, and the probe inspection device are connected to the appearance defect analysis device via each analysis station, and the product design support system and the data input terminal are connected to the appearance defect analysis device. Has been done.

The foreign matter inspection data is fetched from the foreign matter inspection apparatus through the foreign matter analysis station to the appearance defect analysis apparatus and registered in the database. The appearance defect inspection data is fetched from the appearance defect inspection device to the appearance defect analysis device via the appearance defect analysis station and registered in the database. The probe FC inspection data for each chip is fetched from the probe inspection device to the appearance defect analysis device via the probe data analysis station and registered in the database. The probe fail bit (hereinafter referred to as probe FB) inspection data in units of cells is fetched from the probe inspection device to the appearance defect analysis device via the fail bit data analysis station and registered in the database.

The design layout data is fetched from the product design support system into the appearance defect analysis device. The measurement condition data is fetched from the data input terminal into the appearance defect analysis device.

Probe FC inspection data and probe F
Since the B inspection data is electrical characteristic data and is a logical one-dimensional array, it cannot be directly matched with the foreign matter inspection data and the appearance defect inspection data. Therefore, the data is based on the design layout data and the measurement condition data. Array conversion and data coordinate conversion are executed. The data array conversion is to convert logical one-dimensional array data into physical array data, and the data coordinate conversion is to convert the physical array data into physical probe FB data (coordinate data) containing dimension information. .

Since the probe FB inspection data is data in units of cells, the data capacity is enormous, and if the database is directly created, the database becomes large in capacity and
The access speed will decrease. Therefore, the probe FB
It is desirable that the inspection data be compressed in the state of the logical one-dimensional array data and the physical array data before being made into a database so that the data is decompressed when it is used.

When investigating the cause of a defect in a semiconductor element formed on a semiconductor wafer, data relating to the same wafer in the same lot is selected from among the data for which data collection and data conversion have already been performed, and the actual probe is selected. The FB data, the foreign substance inspection data, and the appearance defect inspection data are matched with each other and logically operated (comparative collation) to clarify the causal relationship with each other.

In the matching logic operation, a matching position correction calculation process is executed before the matching analysis logic calculation process in order to cancel the influence of the positional deviation caused by the alignment accuracy of each inspection device.

As one method of the butt position correction calculation process, a plurality of positions having different data (for example, foreign matter inspection data and FB inspection data) in one arbitrarily specified area are selected and the distances between them are selected. Can be measured and the average value thereof can be determined as the amount of deviation.

As one method of the butt position correction calculation process, different data (for example, foreign matter inspection data and FB inspection data) are shifted and matched while being logically operated in an arbitrarily specified area to overlap each other. It is possible to employ a method in which the position at which is maximized is determined and the shift amount (coordinates) is determined as the shift amount.

As a method of the butt analysis logic operation process, when the foreign matter adhesion position and the appearance defect position and the fail bit data are butted, a circle having a radius of an expected deviation amount around one coordinate of the foreign matter data is drawn. ,
When even one probe FB data exists within the range, a method can be adopted in which the probe FB data is determined to be an FB defect due to a foreign substance.

As a method of the match analysis logic operation process, when the foreign matter adhesion position and the appearance defect position are matched with the probe FB data, a method of expanding the fail bit data to an arbitrary size and matching them is adopted. it can.

It is preferable that the matching analysis logic operation process is configured so that the matching target range can be freely set such as wafer, shot, chip, mat, and arbitrary designation.

It is desirable that the matching analysis logic operation process is configured so that the matching target range can be freely set such as by foreign particle size, appearance defect category, fail bit category, and arbitrary designation.

The match analysis logic operation process can be configured by a map match analysis function and a defect factor match analysis function. The match map analysis is a match logic operation of different data, and a simple match function that takes the logical sum of data, a defect factor emphasis function that takes the logical value of data, and a defect factor mask function that takes the exclusive OR of data. It consists of and.

The defect factor matching analysis is a matching of different data. For example, the foreign substance inspection data of the previous process which becomes the FB defective factor is traced back and logically operated to clarify the causal relationship with the foreign substance inspection data of each process. It is composed of a defect factor trace function for increasing the number of defects and a defect rate factor matching analysis function for matching the ratio of the foreign matter inspection data in each process to the FB defect.

[0026]

[Function] By connecting the foreign matter inspection device, the appearance defect inspection device, and the probe inspection device to the appearance defect analysis device via each analysis station, and by connecting the product design support system and the data input terminal to the appearance defect analysis device. It is possible to automatically collect foreign matter inspection data, appearance defect inspection data, probe FC inspection data in chip units, probe FB inspection data in cell units, design layout data, measurement condition data, etc. online to the appearance defect analysis device. Becomes

Probe FC inspection data and probe F
By converting the B inspection data into the data array conversion and the data coordinate conversion based on the design layout data and the measurement condition data, the probe FC inspection data and the probe FB inspection data, which are the electrical characteristics of the logical one-dimensional array, are used as the foreign material data and the appearance The data can be converted into physical probe FC data (coordinate data) and probe FB data (coordinate data) that can be directly matched with the data.

The probe FB inspection data is compressed into a database in the state of logical one-dimensional array data and physical array data, and the compressed data is decompressed when the data is utilized to reduce the size and access of the database. The speed can be increased.

Data relating to the same wafer in the same lot is selected from the data collected by the above-mentioned method and the data obtained by the data conversion, and the substance probe F is selected.
The B data, the foreign substance inspection data, and the appearance defect inspection data are collated on a cell-by-cell basis, and the causal relationship between them is clarified, so that the cause of the defect of the semiconductor element formed on the semiconductor wafer can be clarified at an early stage. It will be possible.

Of the probe FB data, there are especially many defective fail bit defects caused by foreign substances. Therefore, as one method of the butt position correction calculation processing, different data (for example, foreign substance inspection data) in an arbitrarily specified area is used. , FB inspection data), one position is selected, a distance between them is measured, and an average value thereof is used as a deviation amount. It is possible to cancel the influence of the positional deviation.

Since many fail bit defects in the probe FB data are caused by foreign substances and appearance defects, as a method of the butt position correction calculation process, different data (for example, foreign substance inspection data,
The method in which the FB inspection data) is subjected to a butt logic operation while being shifted to obtain a position where the mutual data overlap is maximum, and the shift amount (coordinates) is used as the shift amount is particularly realistic. It is possible to cancel the influence of misalignment caused by alignment accuracy.

The resolution of fail bit data in cell units increases in proportion to the miniaturization of products, and is always
Since the alignment accuracy of the foreign matter inspection device or the appearance defect inspection device is several tens of times, the fail bit data is arbitrarily set when the foreign substance adhesion position and the appearance defect position are matched with the probe FB data as one method of the matching analysis logic operation processing. The method of enlarging the size to match the size is particularly realistic, and it is possible to cancel the influence of the positional deviation caused by the alignment accuracy of each inspection device.

The simple matching map of the matching analysis function can be expressed by displaying the result of the logical sum (OR) of the data of each other.

The defect factor emphasis map of the matching analysis function can be expressed by displaying the result of taking the logical product (AND) of mutual data.

The defect factor mask map of the matching analysis function can be expressed by displaying the result of exclusive OR (EXOR) of mutual data.

[0036]

FIG. 1 shows a semiconductor device according to an embodiment of the present invention.
It is a block diagram which shows the failure analysis apparatus of the semiconductor wafer used for the failure analysis method in a manufacturing method . FIG. 2 and subsequent figures are explanatory views for explaining the failure analysis method of the semiconductor wafer.

As shown in FIG . 1, a semiconductor device manufacturing line is an assembly of semiconductor manufacturing devices.
Are arranged in large numbers. A wafer as a workpiece goes through each manufacturing process 1 from left to right as the processing progresses, and a probe inspection is performed by a probe inspection device 5 in the final process of the wafer. In FIG. 1, reference numeral 2 denotes the flow of wafers.

Further, during each step 1, the foreign matter inspection device 3,
An appearance defect inspection device 4 and the like are provided, and these inspection devices perform a foreign matter inspection and an appearance defect inspection on a wafer as a work processed in each step 1.

The foreign matter inspection device 3 has a function of measuring the foreign matter adhesion position (coordinates) on the wafer, the number of foreign matters, and the foreign matter size, and has an external communication function for transmitting the data obtained by the inspection to the outside. have. Here, the foreign substance inspection data is collected by the appearance defect analysis device 13 via the A line 6, the foreign substance analysis station 7, and the B line 245 and registered in the database.

The appearance defect inspection apparatus 4 has a function of measuring the defect position (coordinates), the number of defects, and the defect category on the appearance on the wafer, and external communication for transmitting the data obtained by the inspection to the outside. Have a function. Here, the appearance defect inspection data is the C line 246,
It is collected by the appearance failure analysis device 13 via the appearance analysis station 8 and the D line 247 and registered in the database.

The probe inspecting device 5 has a logical position for each chip formed on the wafer, the number of non-defective products, the number of defective products,
It has a function to measure electrical characteristics and has an external communication function to send the data obtained by the inspection to the outside. Here, the probe FC data subjected to the probing inspection is collected by the appearance defect analysis device 13 via the E line 248, the probe data analysis station 9, and the F line 249, and registered in the database.

Further, the probe inspection device 5 has a function of measuring the logical position of each cell formed on the wafer, the number of non-defective products, the number of defective products, and the electrical characteristics, and the data obtained by the inspection is externally read. Has an external communication function to send to. Here, the probe FB data subjected to the probing inspection is collected by the appearance defect analysis device 13 via the G line 250, the FB data analysis station 10 and the H line 251, and registered in the database.

The product design support system 11 has a function of generating design layout data (data relating to a physical position and size from a chip to a cell on the wafer) necessary for forming a semiconductor element on the wafer. , It has an external communication function to send the data to the outside. Here, the design layout data is I line 2
It is collected by the appearance failure analysis device 13 via 52 and registered in the database.

The data input terminal 12 has a function of generating the measurement conditions (the measurement start position and the measurement direction of the chip and the cell) of the inspection in the probe inspection device 5, and the external communication for transmitting the data to the outside. Have a function. Here, the measurement condition data is collected by the appearance defect analysis device 13 via the J line 253 and registered in the database.

The appearance defect analysis device 13 includes a data coordinate conversion unit 14, a matching position correction calculation unit 15, and a matching logic calculation unit 16.

The data coordinate conversion unit 14 is configured to convert the probe FC data and the probe FB data from logical array data into physical array data and physical entity coordinate data.

The butt position correction calculation unit 15 is configured to check each of the inspection devices 3, 4 when the different types of data are matched.
The position correction is performed so as to absorb the variation in alignment accuracy among the five.

The matching analysis logical operation unit 16 is configured to quantitatively express the causal relationship of the different data by performing a logical operation (logical sum, logical result, exclusive logical sum) of different data.

FIG. 2 shows a semiconductor device according to an embodiment of the present invention.
5 is a flowchart showing the positioning of a failure analysis method in the manufacturing method of FIG.

As shown in FIG. 2, in a semiconductor device manufacturing line, defect analysis for a wafer as a work is performed by (1) state change investigation step 1 of defect factors.
7, (2) The first stage of the failure cause investigation step 18,
The work is carried out in each step of (3) second step 19 of defect cause investigation step, (4) countermeasure step 23, and (5) countermeasure effect confirmation step 24. 20 is a high-precision analysis step, 21 is a heterogeneous data matching analysis unit, and 22 is a map.

The wafer failure analysis method according to this embodiment is
Of these steps, it is used in the second stage 19 of the defect cause investigation step and is executed. The second stage 19 of the defect cause investigating step corresponds to a microscopic analysis in contrast to the macro analysis in the first stage 18 of the defect cause investigating step, and the defective portion and the defective cause are directly matched. It will be.

FIG. 3 shows a semiconductor device according to an embodiment of the present invention.
FIG. 6 is an explanatory diagram showing the function and target data of the wafer failure analysis method in the manufacturing method of FIG.

In FIG. 3, the heterogeneous data matching analysis unit 21 is broadly divided into a map matching analysis unit 25 that displays a wafer or chip map and performs defect analysis, and a matching result is processed and displayed quantitatively and visually. And a failure factor matching analysis unit 26 for 27
Is a conceptual diagram of analysis target data.

The map matching analysis unit 25 does not overlap the heterogeneous data with the simple matching map function unit that simultaneously displays different types of data, the defect factor emphasis matching map function unit that emphasizes and displays only the data where different types of data overlap. It is composed of three components: a defect factor mask matching map function unit that displays only data.

The defect factor matching analysis unit 26 clarifies the causal relationship by sequentially comparing different types of data in each process and clarifying the causal relationship between them, and clarifies the ratio of a certain defect factor to the defect rate. It is composed of two parts: a failure rate factor matching analysis section.

The data to be matched is of four types: (1) foreign matter and appearance defect, (2) foreign matter and probe FC data, (3) appearance defect and probe FC data, (4) foreign matter, appearance defect, probe FC data. is there.

FIG. 4 is an explanatory view showing the map matching analysis unit.

The map matching analysis unit 25 includes a particle / defect matching map unit 28, a particle / FB matching map unit 258, a defect / FB matching map unit 259, and a particle / defect / FB matching map unit 260.
In each map portion, as matching target data, (1) foreign matter and appearance defect, (2) foreign matter and probe FB data,
(3) Appearance defect and probe FB data, (4) Foreign matter, appearance defect, probe FB data are transmitted.

Each map unit in the map matching analysis unit 25 serves as a matching function unit as a simple matching map unit 29, a defect factor emphasis matching map unit 30, a defect factor mask matching first map unit 31, a defect factor mask matching second map unit. It has 32 four functional parts.

The simple matching map unit 29 calculates a logical sum (O of the matching source data and the matching destination data).
R) is taken and the result is displayed as a wafer map. With this wafer map, it is possible to analyze the degree of overlapping of data.

The defect factor highlighting matching map section 30
The logical product (AND) of the matching source data and the matching destination data is obtained, and the result is displayed as a wafer map. With this wafer map, it is possible to analyze the causal relationship of the overlapping data.

Defect Factor Mask Matching First Map Part 3
1 is configured to obtain the exclusive OR (EXOR) of the matching source data and the matching destination data, and as a result, display the matching source data as a wafer map. With this wafer map, factors other than the overlapping data can be analyzed.

Defect Factor Mask Matching Second Map Part 3
2 is configured to obtain an exclusive OR (EXOR) between the matching source data and the matching destination data, and as a result, display the matching destination data as a wafer map. With this wafer map, factors other than the overlapping data can be analyzed.

Each of the map units 28 to 32 in the map matching analysis unit 25 has a function of selecting target data, and selection by total number, category (grain size), and process is possible.

Further, each of the map units 28 to 32 has a matching and display range selecting function, and is capable of selecting a wafer map, a shot map, a chip map, a mat map and an arbitrarily designated map.

Here, the map analysis in the foreign substance and defect matching map unit 28 of the map matching analysis unit 25 according to the present embodiment will be outlined with reference to FIGS. 11 and 12.

FIG. 11A shows a foreign matter wafer map 75 in the first step and a foreign matter wafer map 76 in the second step.
And the enlarged chip map display. In each map, the foreign matter adhesion position and the foreign matter particle size display are displayed in different sizes. 78
Is the Y coordinate axis on the wafer, 79 is the X coordinate axis on the wafer, 8
Reference numeral 0 is a chip matrix display (scribe line) on the wafer, 81 is a position indication mark for large foreign matter, 82 is a position indication mark for medium foreign matter, and 83 is a position indication mark for small foreign matter.

FIG. 11B shows the appearance defect wafer map 77 and the enlarged chip map display. In the map, the display of appearance defects by category is displayed with different patterns. 84 is a position display mark of the A category defect, 85 is a position display mark of the B category defect, and 86 is a position display mark of the C category defect.

FIG. 12 is a map diagram showing a case where a match map analysis of a foreign substance and a defect is executed using each of the above-mentioned data. Reference numeral 87 is a simple matching map of foreign matters and defects, 88 is a defect factor emphasis matching map, and 89.
Is a first map of defective factor mask matching, and 90 is a second map of defective factor mask matching.

Simple match map 87 of foreign matter and defect
Is a logical sum (OR) of the matching source data (defect) and the matching destination data (foreign matter), and displays the result as a simple matching map of the foreign matter and the defect.

The defect factor emphasizing matching map 88 takes a logical product (AND) of the data of the matching source (defect) and the data of the matching destination (foreign matter), and displays the result as a defect factor emphasizing matching map of the foreign matter and the defect. There is something to do.

Defect factor mask matching first map 89
Takes the exclusive OR (EXOR) of the matching source data (defects) and the matching destination data (particles), and matches the result (matching source data (defects)) with the defect factor mask of the particles. Some are displayed as a map.

Defect factor mask matching second map 90
Takes the exclusive OR (EXOR) of the matching source data (defect) and the matching destination data (foreign matter), and compares the result (matching destination data (foreign matter)) with the defect factor mask of the foreign matter and the defect. Some are displayed as a map.

Next, the map analysis in the match map unit 29 between the foreign matter and the FB in the map match analysis unit 25 will be outlined with reference to FIGS. 13 and 14.

FIG. 13A shows a foreign matter wafer map 75 in the first step and a foreign matter wafer map 76 in the second step,
The enlarged chip display is shown. In each map, the foreign matter adhesion position and the foreign matter particle size display are displayed with different display sizes.

FIG. 13B shows the probe FB wafer map 91 and the enlarged chip map display. On the map 91, the inspection results in cell units are displayed as white squares for good products (pass) and black squares for defective products (fail). Reference numeral 92 is a bit matrix display, and 93 is an FB position display mark.

FIG. 14 is a map diagram showing a case where the FB and foreign matter matching map analysis is executed using the above-mentioned data. Reference numeral 94 is a simple matching map between the foreign matter and FB, 95 is a defect factor emphasized foreign matter FB matching map, 96 is a defect factor mask matching first map, 97
Is a second map of defective factor mask matching.

Simple matching map 94 of foreign matter and defect
Is a logical sum (OR) of the matching source data (FB) and the matching destination data (foreign matter), and the result is displayed as a simple matching map of the foreign matter and FB.

The defect factor emphasizing matching map 95 takes the logical product (AND) of the matching source data (FB) and the matching destination data (foreign substance), and displays the result as a defect factor emphasizing matching map of the foreign substance and FB. To do.

Defect factor mask matching first map 96
Takes the exclusive OR (EXOR) of the matching source data (FB) and the matching destination data (foreign matter), and obtains the result (matching source data (FB)) as the foreign matter and FB.
It is displayed as a defect factor mask matching map of.

Defect factor mask matching second map 97
Takes the exclusive OR (EXOR) of the matching source data (FB) and the matching destination data (foreign matter), and obtains the result (matching destination data (foreign matter)) as FB
It is displayed as a defect factor mask matching map of.

Next, the map analysis in the match map unit 30 between the defect of the map match analysis unit 25 and the FB will be outlined with reference to FIGS. 15 and 16.

A defective wafer map 77 is shown in FIG.
And the enlarged chip map display. On the map, the display of defects by category is displayed with different display patterns.

FIG. 15B shows a probe FB wafer map 91 and a chip map display which is an enlarged display thereof. In the map 91, a cell unit inspection result is displayed as a white square when it is a non-defective product (pass) and a black square when it is a defective product (fail).

FIG. 16 is a map diagram showing a case where the FB and defect matching map analysis is executed using the above-mentioned data. Reference numeral 98 is a simple match map between defects and FBs, 99 is a defect factor emphasis defect FB match map, 100 is a defect factor mask match first map, 1
Reference numeral 01 indicates a second map for defective factor mask matching.

Simple match map 98 of defect and FB
Is a logical sum (OR) of the matching source data (FB) and the matching destination data (defect), and the result is displayed as a simple matching map of the defect and FB.

The defect factor emphasizing matching map 99 takes the logical product (AND) of the matching source data (FB) and the matching destination data (defect), and displays the result as a defect factor emphasizing matching map of defect and FB. There is something to do.

Defect factor mask matching first map 10
0 is the exclusive OR (EXOR) of the matching source data (FB) and the matching destination data (defect),
As a result (matching source data (FB)
The defect cause is displayed as a mask matching map.

Defect factor mask matching second map 10
1 takes the exclusive OR (EXOR) of the matching source data (FB) and the matching destination data (defect),
As a result (matching destination data (defect)),
There is one displayed as the defect factor mask matching map 2 of B.

Next, the map analysis of the foreign substance / defect / FB matching map unit 31 of the map matching analysis unit 25 will be outlined with reference to FIGS. 17 and 18.

FIG. 17A shows a foreign matter wafer map 75 in the first step and a foreign matter wafer map 76 in the second step.
And the enlarged chip map display. In each map, the foreign matter adhesion position and the foreign matter particle size display are displayed with different display sizes.

In FIG. 17B, the defective wafer map 7 is shown.
7 and the enlarged chip map display are shown. On the map 77, the display of defects by category is displayed with the display pattern changed.

FIG. 17C shows the wafer map 91 of the probe FB inspection data and the display of the chip map as the enlarged data thereof. On the map 91, the inspection results in cell units are displayed as white squares for good products (pass) and black squares for defective products (fail).

FIG. 18 is a map diagram showing a case where a butt map analysis of foreign matters, defects and FBs is executed using the above-mentioned data. Reference numeral 102 shows a simple match map of foreign matter, defect and FB, and 103 shows a defect factor emphasized foreign matter defect FB match map.

The simple matching map 102 includes the matching source data (FB) and the matching destination data (foreign matter,
The defect is ORed (OR) and the result is displayed as a simple matching map of the foreign matter, the defect and the FB.

The defect factor emphasis matching map 103 is
The logical product (AND) of the data of the matching source (FB) and the data of the matching destination (foreign matter, defect) is taken and the result is displayed as a defect factor emphasized matching map of the foreign matter, the defect and the FB.

FIG. 19 is a map diagram showing a case where a butt map analysis of foreign matters, defects and FBs is executed using the above-mentioned data. Reference numeral 104 is a first map of defective factor mask matching, 105 is a second map of defective factor mask matching, and 106 is a third map of good factor mask matching.

Defect factor mask matching first map 10
4 obtains the exclusive OR (EXOR) of the matching source data (FB) and the matching destination data (foreign matter, defect), and obtains the result (matching source data (FB)) for the foreign matter, the defect and the FB. It is displayed as a defect factor mask matching map.

Defect factor mask matching second map 10
5 obtains the exclusive OR (EXOR) of the matching source data (FB) and the matching destination data (foreign matter, defect), and obtains the result (the matching destination data (foreign matter)) of the foreign matter, the defect and the FB. The defect factor mask matching map 2 is displayed.

Defect Cause Mask Matching Third Map 10
6 obtains the exclusive OR (EXOR) of the matching source data (FB) and the matching destination data (foreign matter, defect), and obtains the result (matching destination data (defect)) for the foreign matter, defect and FB. It is displayed as a defect factor mask matching map.

Next, the function of the defect factor matching analysis unit 26 of the map matching analysis unit 25 according to the present invention will be outlined with reference to FIG.

The defect factor matching analysis unit 26 determines the ratio of a certain defect factor to the defect rate, and the defect factor trace analysis unit 33 that sequentially compares different types of data in each process to clarify the causal relationship between them. And a failure rate factor matching analysis unit 34.

The defect factor trace analysis unit 33 is composed of four functional units: a defect defect foreign substance factor trace unit 35, an FB defect defect factor trace unit 36, an FB defective foreign substance factor trace unit 37, and an FB defective defect foreign substance factor trace unit 38. Has been done.

The defect / defective foreign matter factor trace unit 35 is configured to match the foreign matter data of each step as a defective defect factor one after another and to clarify the causal relationship between the defect and the foreign matter. The logical product (AND) of the matching source data (defect) and the matching destination data (dust) is obtained, and the result is displayed in a graph as a defect defect foreign matter factor trace.

The FB defect defect factor trace unit 36 is
The appearance data of each process is sequentially matched as a failure factor, and the function of clarifying the causal relationship between the FB failure and the defect is executed, and the matching source data (FB) and the matching destination data are configured. A logical product (AND) with (defect) is taken and the result is displayed as a FB defect defect factor trace on a graph.

The FB defective foreign matter factor trace unit 37
It is configured to match the foreign matter data of each process as a failure factor one after another, and to execute the function of clarifying the causal relationship between the FB failure and the foreign matter, and the matching source data (FB) and the matching destination data. The logical product (AND) with (foreign matter) is taken and the result is displayed on the graph as an FB defect foreign matter factor trace.

The FB defect defect foreign matter factor trace unit 38 is
As the FB defect factor, the defect data and the foreign substance data of each process are sequentially matched, and the function of clarifying the causal relationship between the three elements of the FB defect, the defect, and the foreign substance is executed. Logical product (AN) of (FB) and the data (defect, particle) at the matching destination
D) is taken, and the result is displayed on a graph as a trace of the FB defect defect foreign matter factor.

The defect rate factor matching analysis unit 34 includes a defect analysis unit 39 for comparing the defect defect rate and the number of foreign particles, an analysis unit 40 for comparing the FB defect ratio and the number of defects, and an analysis unit 41 for comparing the FB defect ratio and the number of foreign particles. It is composed of three.

Defect defect rate and foreign matter number matching analysis unit 39
Is configured to perform a function of sequentially comparing defects and foreign matter data of each process, and clarifying a ratio of foreign matter as a defect defect factor to the defect defect rate,
The logical product (AND) of the data of the matching source (defect) and the data of the matching destination (foreign matter) is taken, and the result is totaled for each defect category and displayed in a graph.

FB defect rate / defect number matching analysis unit 40
Is configured to perform a function of sequentially comparing FB and defect data of each process and clarifying the ratio of the FB defect factor defects to the FB defect rate.
The logical product (AND) of the matching source data (FB) and the matching destination data (defect) is taken, and the result is F
B categories are totaled and displayed in a graph.

FB defective rate and foreign matter number matching analysis unit 41
Is configured to compare the FB and the foreign matter data of each process one after another, and to clarify the ratio of the foreign matter of the FB failure factor to the FB failure rate.
The logical product (AND) of the matching source data (FB) and the matching destination data (foreign matter) is taken, and the result is F
B categories are totaled and displayed in a graph.

The trace units 35 to 38 and the matching units 39 to 41 are the processing target data selection function unit 4 respectively.
2 is provided for each, and selection by total number, category (particle size), and process is possible.

Further, each of the trace units 35 to 38 and the matching units 39 to 41 described above is provided with a processing target range selection function unit 43 of the matching and display range, respectively, and a wafer map, shot map, chip map and mat map. It is possible to select an optional map.

Here, the defect defect foreign matter factor trace unit 35
The function of is outlined using FIG.

The defect / defective foreign matter factor trace unit 35 sequentially checks the foreign matter data of each process as a defective defect factor to clarify the causal relationship between the defect and the foreign matter. Then, the logical product (AND) of the data (defects) at the matching source and the data (particulates) at the matching destination is obtained, and the result is displayed in the graph shown in FIG.

The defect defect foreign matter factor trace graph is shown in FIG.
As shown in, trace analysis target defect display axis 1
The defect NO. The X-axis, which is the trace analysis target foreign matter adhesion process display axis 108, displays the foreign matter adhesion history in each process of the foreign matter that is the factor. Reference numeral 109 is a mark representing a process having a foreign matter adhesion history, and 110 is a legend display thereof.

That is, as a result of matching (logical operation), a process in which a foreign substance and a defect are present at the same position has a white circle mark 109 in the process corresponding to the defect.
Is displayed, and processes that are not displayed are displayed with no mark. Hereinafter, the same process is repeatedly performed on the foreign substance data of each process.

Next, the function of the FB defect defect factor trace unit 36 will be outlined with reference to FIG.

The FB defect defect factor trace unit 36 sequentially compares defect data of each process as an FB defect factor to clarify the causal relationship between FB and defects. Then, the logical product (AND) of the matching source data (FB) and the matching destination data (defect) is obtained, and the result is displayed as a FB defective foreign matter factor trace in the graph shown in FIG.

As shown in FIG. 21, the trace graph of the FB defect foreign matter factor is displayed on the Y axis which is the FB display axis 111 of the trace analysis target. The defect history in each process of the defect that causes the trace analysis target defect finding process display axis 112 is displayed on the X axis. Reference numeral 113 is an FB defect history display, and 114 is a legend display thereof.

That is, as a result of the matching (logical operation), a process in which the defect and the FB exist at the same position is a white triangle mark 11 in the corresponding process of the corresponding FB.
3 is displayed, and processes that are not displayed are displayed with no mark. Hereinafter, the same process is repeatedly performed on the foreign substance data of each process.

Next, the function of the FB defective foreign matter factor trace unit 37 will be outlined with reference to FIG.

The FB defect foreign matter factor trace unit 37 sequentially compares the foreign substance data of each process as the FB defect factor to clarify the causal relationship between FB and the foreign substance. Then, the logical product (AND) of the matching source data (FB) and the matching destination data (foreign matter) is obtained, and the result is displayed as a FB defective foreign matter factor trace in the graph shown in FIG.

As shown in FIG. 22, this FB defective foreign matter factor trace graph shows that the FBNO. The X-axis, which is the trace analysis target foreign matter adhesion process display axis 108, displays the foreign matter adhesion history in each process of the foreign matter that is the factor. Reference numeral 115 is an FB foreign matter adhesion history display, and 116 is a legend display thereof.

That is, as a result of the matching (logical operation), a process in which the foreign matter and the FB exist at the same position is indicated by a white circle mark 115 in the corresponding process of the corresponding FB.
Is displayed, and the process which is not displayed is displayed with no mark. Hereinafter, the same process is repeatedly performed on the foreign substance data of each process.

Next, the FB defect defect foreign matter factor trace unit 3
The functions of No. 8 will be outlined with reference to FIG.

FB defect defect foreign matter factor trace unit 38 is F
As the B defect factor, the defect and foreign substance data of each process are sequentially compared, and the causal relationship between FB and the defect and foreign substance is clarified. Then, the logical product (AN) of the matching source data (FB) and the matching destination data (defect, foreign material)
D) is taken, and the result is displayed in the graph shown in FIG. 23 as the FB defect defect foreign matter factor trace.

As shown in FIG. 23, the trace graph of the FB defect defect foreign matter factor indicates that FBNO. The trace analysis target foreign matter adhesion, defect finding step display axis 117 is displayed on the X-axis, which is the cause of the defect, the defect in each step, and the foreign matter adhesion history. Reference numeral 118 is a legend display of FB, foreign matter adhesion, and defect history display.

As a result of the matching (logical operation), a process in which the defect and the FB exist at the same position is displayed with a white triangle mark 113 in the process at the corresponding position, and a process other than that is displayed with no mark. .

Further, in the process in which the foreign matter and the FB exist in the same position, the white circle mark 115 is displayed in the process of the corresponding portion, and the process other than that is displayed in the no mark. Hereinafter, the same process is repeatedly performed on the foreign substance data of each process.

Next, the function of the defect defect rate foreign matter number butt matching portion 39 will be outlined with reference to FIG.

Defect defective rate and foreign matter number matching The matching section 39 sequentially compares the defect and foreign matter data of each process to clarify the ratio of the defect foreign matter number to the defect defective rate. Then, the logical product (AND) of the data of the matching source (defect) and the data of the matching destination (foreign matter) is taken, and the result is totaled for each defect category and displayed in the graph shown in FIG. .

As shown in FIG. 24, the defect defect rate and the foreign matter number match analysis graph show the defect rate on the Y axis, which is the defect rate display axis 119 that reaches the foreign matter. The X-axis, which is the adhesion process display axis 120, is used for the measurement process of the foreign matter that causes it.

Reference numeral 121 is a display of the ratio of the foreign matter that has reached the defect included in the other categories of the foreign matter deposited in the process, and 122 is the category N of the foreign matter deposited in the process.
O. Display of the proportion of foreign matter that has led to defects 51 to 75, 12
No. 3 of the foreign substances adhered in the process is Cacodery No. 31
Display of the ratio of the foreign matter that has reached the defect of 50 to 124, and 124 is the category No. of the foreign matter adhered in the process. A display of the ratio of the foreign matter that has led to the defects of 01 to 30, 125 is a legend indicating the defect category display type, and 126 is a display of the particle diameter of the foreign matter to be analyzed.

The bar graph for each step shows the ratio of the number of foreign particles that have resulted in defective defects to the total number of foreign particles. Further, as a breakdown of the bar graph, aggregate display for each defect category is performed.

Thereafter, the same process is repeatedly executed for the foreign substance data of each process.

Next, the FB defect rate defect number matching section 40.
The function of is outlined with reference to FIG.

The FB defective rate and the defect number matching section 40 are
FB and the defect data of each process are matched one after another,
The ratio of the number of defects due to FB defects to the FB defect rate will be clarified. And the matching source data (FB)
AND of the data (defect) at the matching destination
The results were aggregated by FB category,
It is displayed in the graph shown in FIG.

As shown in FIG. 25, this FB defective rate and defect number matching analysis graph shows the ratio of reaching the FB on the Y axis, which is the axis 127 for displaying the ratio of reaching the FB of defects. On the X axis, which is the defect generation process display axis 128, a defect measurement process that causes the defect is taken.

In FIG. 25, reference numeral 129 indicates the proportion of defects that have reached the FB included in other categories among the defects generated in the process, and 130 indicates category NO. Display of the percentage of defects that have reached the FB of 51 to 75, 131 is a category No. among the defects generated in the process. Display of the percentage of defects that have reached the FB of 31 to 50, and 132 is the defect generated in the process.
Category NO. The display of the percentage of defects that have reached the FB of 01 to 30, 133 is a legend display showing the FB category display type, and 134 is a legend display showing the analysis target defect category.

The bar graph for each step shows the ratio of the number of defects that resulted in FB defects to the total number of defects. Further, as a breakdown of the bar graph, a total display for each FB category is performed.

Thereafter, the same process is repeatedly executed for the defect data of each process.

Next, the FB defect rate foreign matter number butting portion 41
The function of will be outlined with reference to FIG.

The FB defective rate and the foreign matter number matching section 41 are
FB and foreign material data of each process are matched one after another,
We will clarify the ratio of the number of foreign substances that are the cause of FB failure to the FB failure rate. And the matching source data (FB)
AND (AND) with the matching data (foreign matter)
The results were aggregated by FB category,
It is displayed in the graph shown in FIG.

As shown in FIG. 26, the ratio of reaching FB is taken on the Y-axis 127, and the step of measuring the foreign matter causing the FB is taken on the X-axis 120. .

In FIG. 26, 135 is a display of the ratio of the foreign matters that have reached the FB included in the other categories among the foreign matters that have been attached in the process, and 136 is the category NO. The display of the ratio of the foreign matter that reaches the FB of 51 to 75, 137 is the category NO. 31 to 50 is a display of the ratio of the foreign matter reaching the FB, and 138 is the category No. of the foreign matter attached in the process. It is a display of the ratio of the foreign matter which reached FB of 01-30.

The bar graphs for each step show the ratio of the number of foreign substances that resulted in FB failure to the total number of foreign substances.
Further, as a breakdown of the bar graph, a total display for each FB category is performed.

Hereinafter, the same processing is repeatedly executed for the foreign substance data of each process.

Next, the appearance defect analysis apparatus 1 according to the present embodiment.
The defect analysis process flow of No. 3 will be outlined with reference to FIG.

As shown in FIG. 6, the inspection data of each inspection device 3, 4, 5 is collected by the appearance defect analysis device 13 via each station and terminal, and stored in each database. It In order to match various collected data, it is necessary to integrate various data levels, and the function as the integration is the data coordinate conversion function shown below.

The foreign matter inspection data stored in the foreign matter inspection database unit 44 includes coordinate data based on the wafer origin and coordinate data based on the chip origin.
Here, coordinate data based on the chip origin is used as matching data. Therefore, the coordinate data based on the wafer origin is coordinate-converted into chip coordinate data based on the chip origin by the chip coordinate conversion unit 263 based on the data of the design layout database unit 49.

Since the appearance defect inspection data stored in the appearance defect inspection database unit 45 is coordinate data based on the chip origin, no data coordinate conversion is performed.

The probe FC data stored in the probe FC database unit 46 is logical array data in chip units. Therefore, the probe FC data is array-converted by the data array converter 53 from the logical array data to the physical array data based on the measurement condition data stored in the measurement condition database unit 47.

Next, the probe FC data converted into the physical array data is subjected to data coordinate conversion in the physical physical coordinate data section 14 based on the layout data stored in the design layout database section 49, and the tip origin is used as a reference. Are converted into physical substance coordinate data.

The probe FB data stored in the probe FB database unit 42 is logical array data in chip units and cell units. Therefore, the probe FB
The data 42 is array-converted from the logical array data to the physical array data by the data array conversion unit 54 for FB data based on the measurement condition data.

Next, the FB converted into physical array data
The data is subjected to data coordinate conversion based on the design layout data 49, and is converted into physical physical coordinate data based on the chip origin in the physical physical data coordinate conversion unit 57.

Here, since the probe FB data is data on a cell-by-cell basis, the data capacity is enormous, and it is conceivable that the memory capacity and the processing speed will be reduced due to the increased capacity of the database. Therefore, the compression processing by the data compression unit 55 is executed in the logical array data and physical array data states and stored as the FB compressed data 50. Then, when the matching process or the like is utilized, the data decompression unit 56 decompresses the data and uses it as the FB coordinate data 51.

By the data array conversion and data coordinate conversion described above, all data are integrated in a level,
As described above, the matching analysis logic operation unit 16 is ready for processing. Then, the calculation result is stored in the calculation result database unit 262.

By the way, each of the inspection devices 3, 4, and 5 has a wafer alignment function, and the alignment accuracy is also different. Therefore, the inspection devices 3, 4,
If the matching is executed with the data from 5 being as they are, it is affected by the difference in alignment accuracy, and normal matching becomes impossible.

Therefore, in this embodiment, before the matching analysis calculation processing, the matching position correction calculation for absorbing the variation in the alignment accuracy in each inspection device 3, 4, 5 is executed by the matching position correction calculation unit 15. To be done.

Of the butt position correction calculation, the deviation amount measurement position correction method will be outlined below with reference to FIG.
Here, a case where it is applied to the butt position correction calculation before the butt analysis calculation process of the foreign substance inspection data and the probe FB data described above with reference to FIGS. 13 and 14 will be described as an example.

In FIG. 7, (a) is a flow chart and (b) is an explanatory view schematically showing the flow.

(1) First, in the flowchart 67 of the displacement amount measurement position correction method, as shown in FIG. 7B, the position correction calculation is performed from the foreign matter coordinate data 58 and the probe FB coordinate data 59. The data range R to be used is designated as the position correction calculation data comparison area 60.

(2) Based on the alignment accuracy of the foreign substance inspection device 3 and the probe inspection device 5, the predicted area 63 in which the position where the FB factor foreign substance is present is predicted is designated, and the predicted deviation amount D is set in advance. In FIG. 7B, 64 is a vector indicating the deviation between the foreign substance position 62 and the FB position 61 which is considered to be the FB factor.

(3) The logical sum (OR) of the foreign substance data in the designated data range R and the probe FB data is obtained.

(4) In the resulting data, a circle centered on the coordinates of the foreign substance data and having the radius of the predicted deviation amount D is drawn, and an area in which only one probe FB data exists is searched. To be extracted.

(5) With respect to all the collected data, the deviation amount between the foreign substance position 62 and the FB position 61 is measured, a distribution chart 65 for the vector of the deviation is created, and the average value 66 thereof is obtained.

(6) The average value 66 is the foreign matter inspection device 3
And the probe inspection device 5 are set to the position correction amount, and the position correction processing is performed before the butt analysis logical operation.

Next, the coordinate shift position correction method of the butt position correction calculation will be outlined with reference to FIG. Also here, an application example of the butt position correction calculation before the butt analysis calculation process of the foreign substance inspection data and the probe FB data will be described.

In FIG. 8, (a) is a flow chart and (b) is an explanatory view schematically showing the flow.

(1) First, in the flowchart 71 of the coordinate shift correction method, as shown in FIG. 8B, the foreign matter inspection data 58 and the probe FB data 5
The data range R used for the position correction calculation is designated as the overlap determination processing region 68 from among the nine.

(2) The shift amount and the number of shifts in the X direction, Y direction, and θ direction for the foreign matter inspection data 58 are designated in advance.

(3) The logical product (AND) of the foreign substance data 58 and the probe FB data 59 in the designated data range R is calculated, and the result is stored.

(4) The foreign matter inspection data 58 is the above (2).
Is shifted in the X direction by a predetermined shift amount.

(5) The current number of shifts in the X direction is compared with a preset number of shifts, and if the current number of shifts is within the designated number of times, the process returns to the process (3).

(6) When the current number of shifts exceeds the designated number of shifts, the X direction of the foreign substance inspection data 58 within the designated data range R is returned to the origin.

(7) Subsequently, the foreign matter inspection data 58 is shifted in the Y direction by the designated shift amount.

(8) The current number of shifts in the Y direction is compared with the designated number of shifts. If the current number of shifts is within the designated number of times, the process returns to the process (3).

(9) When the current number of shifts exceeds the designated number of shifts, the Y direction of the foreign matter inspection data 58 is returned to the origin.

(10) Foreign matter data 58 in the data range R
And the probe FB data 59 are ANDed and the result is stored.

(11) Next, the foreign matter inspection data 58 is shifted in the θ direction by the designated shift amount.

(12) The current number of shifts in the θ direction is compared with the designated number of shifts, and if the current number of shifts is within the designated number of times, the process returns to the process (10).

(13) After that, as shown in FIG. 8B, the overlapping degree totaling graph 69 is created by the calculation result stored by the above processing, and the overlapping degree MAX that maximizes the matching amount is obtained. Point 70 is required.

(14) The shift amount up to the MAX point 70 is set as the position correction amount of the foreign substance inspection device 3 and the probe inspection device 5, and the position correction processing is performed before the butt analysis logical operation.

Next, of the matching position correction calculation, the predicted deviation amount specifying matching method will be outlined with reference to FIG. Also here, an application example of the foreign matter inspection data and the probe FB data to the matching analysis calculation processing will be described.

In FIG. 9, (a) is a flow chart and (b) is an explanatory view schematically showing the flow.

(1) First, in the flowchart 254 of the predicted displacement amount designating matching method, as shown in FIG. 9B, from the foreign matter inspection data 58 and the probe FB data 59, a matching analysis logical operation is performed. The data range R to be used is designated as the matching range 261.

(2) The predicted deviation amount D is preset based on the alignment accuracy of the foreign substance inspection device 3 and the probe inspection device 5.

(3) Foreign matter data 5 in the designated data range R
8 and the probe FB data 59 are ORed (OR). At this time, centering on the coordinate 62 of one point of the foreign matter data,
When a circle having a radius of the predicted deviation amount D is drawn and even one probe FB data 61 exists within the range 63, the probe FB data is determined to be an FB defect due to a foreign substance.

Next, of the matching position correction calculation, the data expansion matching method will be outlined with reference to FIG. Also here, an application example of the foreign matter inspection data and the probe FB data to the matching analysis calculation processing will be described.

In FIG. 10, (a) is a flow chart, and (b) is an explanatory view schematically showing the flow.

(1) First, in the flow chart 257 of the data expansion matching method, as shown in FIG. 10B, the foreign matter inspection data 58 and the probe F are detected.
From the B data 59, the data range R used for the matching analysis logical operation is designated as the matching range 261.

(2) The FB enlargement amount D is preset based on the alignment accuracy of the foreign substance inspection device 3 and the probe inspection device 5.

(3) Probe FB of designated data range R
The data 61 is expanded according to the designated FB expansion amount D.

(4) Foreign object data 6 in the designated data range R
2 is ANDed with the expanded probe FB data 256.

(5) As a result, the probe FB data 256 having the logical product of 1 is determined to be the FB defect due to the foreign matter.

According to the above-mentioned embodiment, the following effects can be obtained. (1) Since the causal relationship between the foreign matter and the appearance defect is clarified by matching the foreign matter inspection data and the appearance defect inspection data in various processes, the cause of the defect can be quickly investigated.

(2) By comparing the foreign matter inspection data of various processes with the probe FB inspection data,
Since the mutual causal relationship with is clear, it is possible to quickly investigate the cause of the defect.

(3) Since the causal relationship between the appearance defect and the FB is clarified by matching the appearance defect inspection data of various processes with the probe FB inspection data, it is possible to quickly investigate the cause of the defect. it can.

(4) By enabling the cell-based micro analysis, the analysis accuracy can be significantly improved as compared with the conventional chip-based macro analysis.

(5) During the matching operation of various data, by correcting the variation in alignment accuracy between various inspection devices, it is possible to absorb the variation in alignment accuracy between various inspection devices. The accuracy of the butt work can be improved.

(6) Because of the above (5), the accuracy of the failure analysis work following the matching work of various data can be significantly improved, so that the cause investigation speed of the failure can be substantially improved. As a result, the yield and productivity of semiconductor devices can be improved.

Although the invention made by the present inventor has been specifically described based on the embodiments, the present invention is not limited to the embodiments and various modifications can be made without departing from the scope of the invention. Needless to say.

For example, the handling of various data is not limited to be performed online, but may be performed offline.

When converting FC data and FB data into physical coordinate data, not only the measurement condition data and design layout data are used, but other data may be used.

The probe FB data is not limited to being compressed and decompressed, and may be used as it is.

[0207]

The effects obtained by the typical ones of the inventions disclosed in the present application will be briefly described as follows.

For example, it is possible to find out the cause of failure (fail bit) in probe inspection from among the huge causes of defects that occur in the manufacturing process of semiconductor devices having huge cells such as 16-megabit DRAMs. ,
The accuracy of defect cause analysis can be improved.

Since it is possible to investigate in real time the cause of the defect in the probe inspection, which changes moment by moment, it is possible to take measures against the defect as soon as possible.

Due to the above effects, the yield can be improved, the new product start-up period can be shortened, and the number of analysis steps can be reduced.

[Brief description of drawings]

FIG. 1 is a method for manufacturing a semiconductor device according to an embodiment of the present invention.
2 is a block diagram showing a semiconductor wafer defect analysis apparatus used in the defect analysis method in FIG.

FIG. 2 is a method of manufacturing a semiconductor device according to an embodiment of the present invention.
5 is a flowchart showing the positioning of the failure analysis method in FIG.

FIG. 3 is a method for manufacturing a semiconductor device according to an embodiment of the present invention.
FIG. 5 is an explanatory diagram showing the function and target data of the wafer failure analysis method in FIG.

FIG. 4 is an explanatory diagram showing a map matching analysis unit.

FIG. 5 is an explanatory diagram showing a defect factor matching analysis unit.

FIG. 6 is a defect analysis process flow chart of the appearance defect analysis apparatus.

7A and 7B are diagrams showing a butt position correction calculation of a displacement amount measurement position correction method, in which FIG. 7A is a flowchart and FIG.
FIG. 4 is an explanatory diagram schematically showing the flow.

8A and 8B are diagrams showing a butt position correction calculation of a coordinate shift position correction method, where FIG. 8A is a flowchart and FIG.
FIG. 4 is an explanatory diagram schematically showing the flow.

FIG. 9 is a diagram showing a butt position correction calculation of an expected deviation amount designated butt method, in which (a) is a flowchart;
(B) is an explanatory view schematically showing the flow.

FIG. 10 is a diagram showing a butt position correction calculation of a data expansion butt method, FIG.
(B) is an explanatory view schematically showing the flow.

11A and 11B are explanatory diagrams showing a foreign substance and defect matching map analysis, wherein FIG. 11A is a foreign substance wafer map diagram and a chip map diagram, and FIG. 11B is an external defect wafer map diagram and a chip map diagram.

FIG. 12 is a map diagram showing a case where a butt map analysis of a particle and a defect is executed.

13A and 13B are explanatory diagrams showing a butt map analysis of foreign matter and FB, wherein FIG. 13A is a foreign matter wafer map diagram and a chip map diagram, and FIG. 13B is a probe FB wafer map diagram and a chip map diagram.

FIG. 14 is a map diagram showing a case where an FB and foreign matter butt map analysis is executed.

15A and 15B are explanatory diagrams showing a butt map analysis of defects and FBs, FIG. 15A is a defective wafer map diagram and a chip map diagram, and FIG. 15B is a probe FB wafer map diagram and a chip map diagram.

FIG. 16 is a map diagram showing a case where a FB and defect matching map analysis is executed.

17A and 17B are explanatory diagrams showing a butt map analysis of a foreign substance, a defect, and an FB, wherein FIG. 17A is a foreign substance wafer map diagram and a chip map diagram, FIG. 17B is a defective wafer map diagram and a chip map diagram, and FIG. FIG. 4A is a wafer map diagram and a chip map diagram of probe FB inspection data.

FIG. 18 is a map diagram showing a case where a butt map analysis of a particle, a defect, and an FB is executed.

FIG. 19 is a map diagram showing a case where a butt map analysis of a particle, a defect, and an FB is executed.

FIG. 20 is a defect defect foreign matter factor trace graph.

FIG. 21 is an FB defective foreign matter factor trace graph.

FIG. 22 is an FB defective foreign matter factor trace graph.

FIG. 23 is an FB defective defect foreign matter factor trace graph.

FIG. 24 is a graph summarizing defect defect rates by the number of foreign matter butting defect categories.

FIG. 25 is an aggregate graph for each FB category in which the FB defect rate and the number of defects are compared.

FIG. 26 is an aggregate graph for each FB category in which the FB defect rate is matched with the number of foreign matters.

[Explanation of symbols]

1 ... Manufacturing process, 2 ... Wafer flow, 3 ... Foreign matter inspection device,
4 ... Appearance inspection device, 5 ... Wafer probe inspection device, 6 ...
A line, 7 ... Foreign matter analysis station, 8 ... Appearance analysis station, 9 ... Probe data analysis station, 10
… Fail bit data analysis station, 11… Product design support system, 12… Data input terminal,
Reference numeral 13 ... Appearance failure analysis device, 14 ... Data coordinate conversion processing section, 15 ... Butt position correction calculation processing section, 16 ... Butt logic operation processing section, 17 ... Defect cause state change investigation step, 18 ... Defect cause investigation step 1 stage, 1
9 ... Second step of defect cause investigation step, 20 ... High accuracy analysis step, 21 ... Different data matching analysis step, 22 ... Map / factor matching analysis step, 23
… Countermeasure step, 24… Countermeasure effect confirmation step, 25…
Map matching analysis unit, 26 ... Defect factor matching analysis unit, 27 ... Analysis target data conceptual diagram, 28 ... Foreign matter and defect matching map unit, 29 ... Simple matching map unit,
30 ... Defect factor emphasis matching map part, 31 ... Defect factor mask matching first map part, 32 ... Defect factor mask matching second map part, 33 ... Defect factor trace analysis part, 34 ... Defect ratio factor match analysis part, 35 ... Defect / defective foreign matter factor trace section, 36 ... FB defective defect factor trace section, 37 ... FB defective foreign matter factor trace section, 38 ... F
B defective defect foreign matter factor trace section, 39 ... Defect defective rate foreign matter number matching analysis section, 40 ... FB defective rate defect number matching analysis section, 41 ... FB defective rate foreign matter number matching analysis section,
42 ... Process target data selection function unit, 43 ... Process target range selection function unit, 44 ... Foreign substance inspection database unit, 45 ... Appearance inspection database unit, 46 ... Probe inspection FC database unit, 47 ... Measurement condition database unit, 48 ... Probe Inspection FB database unit, 49 ... Design layout database unit, 50 ... FB compression database unit, 51 ...
FB coordinate database section, 52 ... FC coordinate database section, 53 ... FC data array conversion section, 54 ... FB data array conversion section, 55 ... FB data compression section, 56 ... FB data decompression section, 57 ... FB data coordinate conversion section, 58 ... Matching foreign object coordinate data, 59 ... Matching target FB coordinate data, 60 ... Position correction calculation data comparison area, 61 ...
FB position, 62 ... Foreign matter position, 63 ... FB factor Area in which presence of foreign matter is expected, 64 ... Deviation vector between foreign matter position and FB position that is considered to be FB factor, 65 ... Distribution diagram of deviation vector, 66 ... Average value of distribution , 67 ... Flowchart of displacement amount measurement position correction method, 68 ... Overlap determination processing area, 69
… Overlap degree aggregation graph, 70 ... Overlap degree MAX points, 71 ... Flow chart of coordinate shift correction method, 75
Foreign matter wafer map in the first step, 76 ... Foreign matter wafer map in the second step, 77 ... Exterior defect wafer map, 78 ... Y coordinate axis on wafer, 79 ... X coordinate axis on wafer, 80 ...
Chip matrix display (scribe line) on wafer, 81 ... Large foreign matter position indication mark, 82 ... Medium foreign matter position indication mark, 83 ... Small foreign matter position indication mark, 84 ... A category defect position indication mark, 85 ...
B category defect position display mark, 86 ... C category defect position display mark, 87 ... Simple butting map of foreign matter and defect, 88 ... Defect factor emphasized foreign matter defect map,
89 ... Defect factor mask matching first map, 90 ... Defect factor mask matching second map, 91 ... Probe F
B wafer map, 92 ... Bit matrix display, 93 ...
FB position display mark, 94 ... Foreign matter and FB simple matching map, 95 ... Defect factor emphasized foreign matter FB matching map, 96 ... Defect factor mask matching first map, 97
Defect factor mask matching second map, 98 Defect and FB simple matching map, 99 Defect factor emphasized defect F
B match map, 100 ... Defect factor mask match first map, 101 ... Defect factor mask match second
Map, 102 ... Foreign matter / defect / FB simple matching map, 103 ... Defect factor emphasis foreign matter defect FB matching map, 104 ... Defect factor mask matching first map, 1
05 ... Defect factor mask matching second map, 106 ...
Defect factor mask matching third map, 107 ... Trace analysis target defect display axis, 108 ... Trace analysis target foreign matter adhesion process display axis, 109 ... Defect foreign matter adhesion history display mark, 110 ... Defect foreign matter adhesion history display legend display, 111 ...
Trace analysis target FB display axis, 112 ... Trace analysis target defect finding process display axis, 113 ... FB defect history display, 1
14 ... FB defect history legend display, 115 ... FB foreign matter adhesion history display, 116 ... FB foreign matter adhesion history legend display, 117 ...
Trace analysis target Foreign matter adhesion, defect detection process display axis, 11
8 ... FB, foreign matter adhesion, defect history legend display, 119 ... foreign matter defect rate display axis, 120 ... analysis object foreign matter adhesion process display axis, 121 ... defects included in other categories of foreign matter adhered in the process Display of foreign matter ratio, 12
2 ... Out of the category NO. 51
Display of the proportion of foreign matter that has resulted in defects of ~ 75, 123 ... 31 to 50, a display of the proportion of foreign matter that has led to defects, 124 ... Display of the ratio of foreign particles that lead to defects of 01 to 30, 125 ... Legend display showing the defect category display type, 126 ... Display of particle size of foreign particles to be analyzed,
127 ... Defect to FB rate display axis, 128 ... Analysis target defect occurrence process display axis, 129 ... Display of ratio of defects reaching FB included in other categories of defects generated in the process, 130 ... In the process Of the defects that occurred, the category No. Display of the percentage of defects that led to FB of 51 to 75, 1
31 ... Category No. of defects generated in the process. Three
Display of the percentage of defects that have reached the FB of 1 to 50, 132 ... Category No. of defects generated in the process. 01-30
Display of percentage of defects that have reached FB, 133 ... Legend display showing FB category display type, 134 ... Legend display showing analysis target defect category, 135 ... FB included in other categories of foreign matter adhered in the process Display of the ratio of foreign matter reaching 136, ... In the category No. of foreign matter adhered in the process. Display of the ratio of foreign matter that reaches FB of 51 to 75, 137 ... Display of the proportion of foreign matter that reaches 31 to 50 FB,
138 ... Out of the category NO.
Display of the proportion of foreign matter that has reached the FB of 01 to 245 ...
Line B, 246 ... Line C, 247 ... Line D, 248 ... Line E, 249 ... Line F, 250 ... Line G, 251, ... Line H, 252 ... Line I, 253 ... Line J, 254 ... Method flow chart, 255 ... Pass bit (non-defective cell), 256 ... Expand fail bit,
257 ... Flow chart of data expansion matching method,
258 ... Foreign matter and FB matching map portion, 259 ... Defect and P inspection FB matching map portion, 260 ... Foreign material and defect and FB matching map portion, 261 ... Matching range, 2
62 ... Analysis logic operation result database unit, 263 ... Chip coordinate conversion unit.

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yoshiyuki Miyamoto 5-20-1 Kamimizumoto-cho, Kodaira-shi, Tokyo Hitachi Ltd. Semiconductor Division (72) Inventor Taizou Hashimoto 5-chome, Mizumizumoto-cho, Kodaira-shi, Tokyo No. 20-1 Hitachi, Ltd. Semiconductor Division (72) Inventor Shimoo Sadao 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Hitachi, Ltd. Production Technology Research Institute (72) Inventor Jun Nakazato Totsuka-ku, Yokohama-shi, Kanagawa 292 Yoshida-cho, Hitachi, Ltd., Production Engineering Laboratory (72) Inventor, Seiji Ishikawa, 292, Yoshida-cho, Totsuka-ku, Yokohama, Kanagawa Prefecture, Hitachi, Ltd., Production Engineering Laboratory (72) Inventor, Koichi Kubouchi, Tokyo 5-20-1 Mizumizuhonmachi Hitachi, Ltd. Semiconductor Division (72) Inventor Osamu Sato Tokyo Metropolitan Elementary School 5-20-1 Ichijomizuhonmachi, Hitachi Ltd., Semiconductor Business Division (56) Reference JP-A-63-174330 (JP, A) JP-A-49-91658 (JP, A) JP-A-58-207647 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H01L 21/66

Claims (5)

(57) [Claims] 1. Manufacturing of a semiconductor deviceProcessAt the semiconductor
Foreign matter generated on the roofOrInspect for defectsFewAt least
RankLocation and categoryStores the first information representingStep
TheWhen, A plurality of semiconductor chips formed on the semiconductor wafer
Each ofElectrical characteristics ofOr inside each semiconductor chip
Inspect the electrical characteristics of each of the formed cells
At least each of the semiconductor chips or each of the cells
positionandMemorize the second information indicating the qualityStep
When,The first information indicating the cause of occurrence and the first information indicating a product defect
2 information CompareVerificationdo it,Causal to the combination of both information
Tolerance error of both position information to determine that there is a relationship
Determining the range, A combination of the first information and the second information is appropriately set.
After comparing and collating, the relative error between the position information of
Whether or not there is a causal relationship is determined by determining whether it is included in the difference range.
A determining step, According to the information determined to have the causal relationship,
Take measures against the manufacturing processStepCharacterized by having and
And a method for manufacturing a semiconductor device. 2. A permissible error range of the position information is determined.
Step and step of determining the presence or absence of the causal relationship
The process in the step of aligning each inspection device is centered on the position of the first information.
The radius of the expected deviation determined based on the
A circle to be determined, and there is at least one position of the second information within that range.
In this case, the cause of the second information is determined to be the first information.
The method for manufacturing a semiconductor device according to claim 1, wherein the method is a predetermined process. 3. An allowable error range of the position information is determined.
Step and step of determining the presence or absence of the causal relationship
In the process in step 1, the size of the second information depends on the alignment precision of each inspection device.
The size of the first information is expanded to a size determined based on the degree, and the position of the first information is within the range of the expanded size.
If there is at least one, its first information is
The method for manufacturing a semiconductor device according to claim 1 , wherein the process is a process of determining that the information is generated . 4. An allowable error range of the position information is determined.
Step and / or determine the presence or absence of the causal relationship
In the step, the position that occurs due to the alignment accuracy of each inspection device
Position misalignment correction calculation processing that cancels this effect is executed
4. The method of manufacturing a semiconductor device according to claim 1, 2 or 3 . 5. The method of the misalignment correction calculation processing is optional.
Alignment of each inspection device in the area
The expected deviation amount determined by the measurement accuracy is the radius,
Centering on any one position information included in the second information
And the circle included in the first information
If there is location information that
The distance is measured and the average value of the distance is calculated as the deviation amount.
The method for manufacturing a semiconductor device according to claim 4 , wherein the determination is made.