Semiconductor storage device

This application is based upon and claims the benefit of priority from the prior PCT Application No. PCT/JP2013/073812, filed on Sep. 4, 2013, the entire contents of which are incorporated herein by reference.

Embodiments of the present invention relate to a semiconductor storage device that can store defect information.

A ROM FUSE block to store defect information of memory cells in a block unit or a column unit may be provided in a memory cell array.

The ROM FUSE block has a size equal to a size of a normal block and includes a plurality of memory cells. When a defect beyond an allowable limit occurs in the memory cells of the ROM FUSE block, ROM FUSE data itself stored in the ROM FUSE block may become defective. In addition, when a defect occurs in the ROM FUSE data, it takes time to correct an error, which may result in decreasing an operation speed of a semiconductor storage device. Particularly, when the memory cells are miniaturized, the memory cells are likely to become defective. For this reason, a defect check of the ROM FUSE block becomes important.

The defect check of the ROM FUSE block is generally executed during a power-on read period after power is supplied.

However, when the memory cells are miniaturized, an erasure level is likely to change, a difference of the erasure level and a read level decreases, and erroneous read is more likely to occur.

If the memory cell is miniaturized and a memory capacity increases, the number of blocks also increases and defect information to be recorded in the ROM FUSE block also increases. For this reason, it takes time to read the defect information from the ROM FUSE block.

FIG. 1 is a block diagram illustrating a schematic configuration of a semiconductor storage device 1 according to a first embodiment;

FIG. 2 is a block diagram illustrating a detailed configuration of a peripheral portion of a memory cell array 2 ;

FIG. 3 is a flowchart illustrating a defect check process of a ROM FUSE block according to the first embodiment;

FIG. 4 is a flowchart illustrating a defect check process of a ROM FUSE block according to a second embodiment;

FIG. 5 is a timing diagram when defect information is rewritten at the time of a refresh operation;

FIG. 6 is a flowchart illustrating a defect check process of a ROM FUSE block according to a third embodiment;

FIG. 7 is a timing diagram illustrating process timing of steps S 31 to S 38 of FIG. 6 ;

FIG. 8 is a diagram illustrating a memory map of a memory cell array 2 including a ROM FUSE block 31 ;

FIG. 9A is a diagram illustrating a data configuration of bad block information stored in each data area of the ROM FUSE block 31 ;

FIG. 9B is a diagram illustrating a correspondence relation of a quotient Block Add X and 8-bit data Block IO;

FIG. 10 is a flowchart illustrating an example of a process sequence to transfer defect block information; and

FIG. 11 is a flowchart illustrating an example of a process sequence to store defect information in the ROM FUSE block 31 .

According to one embodiment, a semiconductor storage device has:

a memory cell array that has a plurality of memory cells, a plurality of word lines connected to at least some memory cells of the plurality of memory cells, a plurality of bit lines connected to at least some memory cells of the plurality of memory cells, and a plurality of blocks including a group of at least some memory cells of the plurality of memory cells;

a defect information storage block that is at least one of the plurality of blocks and stores defect information in the memory cell array;

a first defect detection circuitry that reads data of at least some memory cells in the defect information storage block, verifies the data, and determines whether there is a defect in the defect information storage block;

a second defect detection circuitry that, when it is determined by the first defect detection circuitry that there is the defect, changes a read voltage level for reading the data of the memory cells, rereads data of at least some memory cells in the defect information storage block, verifies the data, and determines whether there is the defect in the defect information storage block; and

a defect determination circuitry that, when it is determined by the second defect detection circuitry that there is the defect, determines the defect information storage block as a defective block.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

First Embodiment

FIG. 1 is a block diagram illustrating a schematic configuration of a semiconductor storage device 1 according to a first embodiment. The semiconductor storage device 1 of FIG. 1 is an example of a NAND-type flash memory.

The semiconductor storage device 1 of FIG. 1 includes a memory cell array 2 , a row decoder 3 , a word line driver 4 , a column decoder 5 , a sense amplifier (S/A) 6 , a data latch circuit 7 , a controller 8 , a high voltage generator 9 , an address register 10 , a command decoder 11 , and an I/O buffer 12 .

The memory cell array 2 has NAND strings 20 that are obtained by connecting a plurality of memory cells in series. FIG. 2 is a block diagram illustrating a detailed configuration of a peripheral portion of the memory cell array 2 . As illustrated in FIG. 2 , the memory cell array 2 is divided into a plurality of blocks BLKO to BLKn. In each block, the plurality of NAND strings 20 are arranged in a column direction. Each NAND string 20 has a plurality of memory cells 21 that are connected in series, a selection gate transistor S 1 that is connected to one end side of the memory cell 21 , and a selection gate transistor S 2 that is connected to the other end side.

Gates of the individual memory cells 21 in the NAND string 20 are connected to corresponding word lines WL 0 to WLn+1 (n is an integer of 0 or more). A gate of the selection gate transistor S 1 is connected to a selection gate line SGD. A gate of the selection gate transistor S 2 is connected to a selection gate line SGS. The individual NAND strings 20 are connected to a common cell source line via the corresponding selection gate transistors S 1 . In addition, the individual NAND strings 20 are connected to corresponding bit lines BL 0 to BLn via the corresponding selection gate transistors S 2 .

The individual word lines WL 0 to WLn+1 connected to the gates of the individual memory cells 21 in the NAND string 20 are connected to the row decoder 3 . The row decoder 3 decodes a row address transferred from the address register 10 . The word line driver 4 is disposed in the vicinity of the row decoder 3 . The word line driver 4 generates a voltage to drive each word line, on the basis of decoded data.

The bit lines BL 0 to BLn connected to the individual NAND strings 20 are connected to the sense amplifiers 6 via bit line selection transistors Q 0 . Read data detected by the sense amplifier 6 is held as binary data in the data latch circuit 7 .

The column decoder 5 illustrated in FIG. 1 decodes a column address from the address register 10 . In addition, the column decoder 5 determines whether the data held in the data latch circuit 7 is transferred to a data bus, on the basis of a decoding result.

The I/O buffer 12 buffers an address, data, and a command input from an I/O terminal. In addition, the I/O buffer 12 transfers the address to the address register 10 , transfers the command to the command decoder 11 , and transfers the data to the data bus.

The controller 8 identifies the address and the command and controls a defect check of a ROM FUSE block to be described below or an operation of the sense amplifier 6 .

As illustrated in FIG. 2 , the memory cell array is divided into the plurality of blocks. However, two blocks among the plurality of blocks are a ROM FUSE block (defect information storage block) and a spare ROM FUSE block (spare storage block). In the ROM FUSE block and the spare ROM FUSE block, defect information in the memory cell array is stored. The spare ROM FUSE block is used when the ROM FUSE block becomes defective. When new defect information is stored in the ROM FUSE block, the same defect information is stored in the spare ROM FUSE block. By setting, it can be switched whether the spare ROM FUSE block is used.

The defect information stored in the ROM FUSE block and the spare ROM FUSE block is bad block information showing whether each block in the memory cell array is defective or bad column information showing whether each column in the memory cell array is defective. In addition, parameter information to adjust a voltage level of an internal voltage may be stored in the ROM FUSE block.

The controller 8 of FIG. 1 executes a process for checking whether there is a defect in the ROM FUSE block at the time of a power-on read (POR) operation. FIG. 3 is a flowchart illustrating a defect check process of the ROM FUSE block according to the first embodiment.

As write methods to the memory cell in the NAND-type flash memory, there are methods corresponding to both a single level cell (SLC) method and a multi-level cell (MLC) method. In this embodiment, SLC is adopted. This is because MLC executes multivalued write, a read margin is small, and erroneous read is likely to occur. Therefore, the defect check of the ROM FUSE block is executed using SLC enabling high-reliability read.

First, data stored in a plurality of memory cells on a specific word line among a plurality of word lines in the ROM FUSE block is read (step S 1 ). The specific word line may be basically any word line. However, a word line to read data with high reliability among the plurality of word lines in the ROM FUSE block is preferably set as the specific word line. In the present specification, an example of the case in which the specific word line is set as a word line WL 15 will be described.

Next, a check (hereinafter, referred to as E-to-A detection) on whether normal read is disabled due to a shift of a threshold voltage distribution of an erasure level of the memory cell in a direction of a threshold voltage distribution of an A level close to the erasure level is executed using the data read by step S 1 (step S 2 ). Steps S 1 and S 2 described above correspond to a first defect detection unit (first defect detection circuitry).

Three methods (E-to-A detection methods 1 to 3) to be described below are considered as a specific method of the E-to-A detection and any method may be used.

Bit lines corresponding to 16 Kbytes are connected to each word line in the ROM FUSE block. These bit lines are divided in segments for every bit lines corresponding to 2 Kbytes. In the E-to-A detection, a read check of the memory cells is executed using a non-used segment among the segments.

In the E-to-A detection method 1, in a state in which all memory cells of the non-used segment are erased, all of the memory cells of the non-used segment are read. If the number of data read as data “0” instead of data “1” to be an erasure state is more than a predetermined value, it is determined that the E-to-A detection has failed in.

In the E-to-A detection method 2, in a state in which all memory cells on a specific bit line in the non-used segment in the ROM FUSE block are erased, all of the memory cells are read. If the number of data read as data “0” is more than the predetermined value, it is determined that the E-to-A detection has failed in.

In the E-to-A detection method 3, complementary data is written to the memory cells of the non-used segment in the ROM FUSE block. If the written complementary data is read and the number of changes of bit logic of the complementary data is more than a predetermined value, it is determined that the E-to-A detection has failed in.

When the E-to-A detection of step S 2 executed by any one of the E-to-A detection methods 1 to 3 has failed in, it is determined that the threshold voltage distribution of the E level has been shifted in the direction of the threshold voltage distribution of the A level, a read voltage level for reading the memory cells is increased, and reread (hereinafter, referred to as ROM reread) from the memory cells is executed (step S 3 ). In the ROM reread, a voltage applied between a gate and a source of the memory cell is increased and the reread from the memory cell is executed.

Next, the E-to-A detection is executed again using the data read by step S 3 (step S 4 ). In this case, in the E-to-A detection, any one of the detection methods 1 to 3 may be adopted. Steps S 3 and S 4 described above correspond to a second defect detection unit (second defect detection circuitry).

If the E-to-A detection of step S 4 is failed in, it is determined that the check of the ROM FUSE block has failed in (step S 5 : defect determination unit or defect determination circuitry). In this case, the NAND-type flash memory including the ROM FUSE block is assumed as a defective product.

Meanwhile, when the E-to-A detection of step S 2 has passed in, a block check executed by a normal power-on read operation is executed (step S 6 : third defect detection unit or third defect detection circuitry). In the block check, for example, when bad block information or bad column information is stored in the ROM FUSE block, the bad block information or the bad column information is stored as the complementary data, the stored complementary data is read, two kinds of data configuring the complementary data are compared in a bit unit, and it is determined whether there is a read error. The block check does not need to be executed on all memory cells in the ROM FUSE block and may be executed on only the memory cells on the specific word line.

When the block check of step S 6 has failed in, the ROM reread of step S 3 is executed and when the block check has passed in, it is determined that the ROM FUSE block is normal (step S 7 ).

Likewise, even when the E-to-A detection of step S 4 has passed in, the block check is executed (step S 8 : third defect detection unit or third defect detection circuitry). When the block check has failed in, a fail is determined (step S 5 ) and when the block check has passed in, it is determined that the check of the ROM FUSE block has passed in (step S 7 ).

In the flowchart of FIG. 3 , the example of the case in which the E-to-A detection and the ROM reread in which the read voltage level has changed are combined has been described. However, only the ROM reread may be executed without executing the E-to-A detection. In this case, instead of the E-to-A detection of steps S 2 and S 4 , an existing block check may be executed.

In addition, only the E-to-A detection may be executed without executing the ROM reread. In this case, it is determined whether the ROM FUSE block is normal, using a result of one E-to-A detection. For this reason, a level of defectiveness of the ROM FUSE block may increase. Therefore, when the level of defectiveness of the ROM FUSE block needs to be decreased, the ROM reread is preferably executed.

As such, in the first embodiment, when it is determined by the E-to-A detection or the block check that the ROM FUSE block is defective, the read voltage level is increased and it is predetermined whether the ROM FUSE block is defective. Therefore, even though the threshold voltage distribution of the E level of the memory cell is slightly shifted to the side of the threshold voltage distribution of the A level due to a manufacturing variation, it is not determined that the ROM FUSE block is defective and the level of defectiveness of the ROM FUSE block can be reduced.

Second Embodiment

In a second embodiment to be described below, when it is determined that a ROM FUSE block is defective, a refresh operation of the ROM FUSE block is executed.

A NAND-type flash memory according to the second embodiment includes the same block configuration as the block configuration of FIG. 1 and a process operation of a controller 8 is different from a process operation of the first embodiment. Therefore, the process operation of the controller 8 will be described hereinafter.

FIG. 4 is a flowchart illustrating a defect check process of the ROM FUSE block according to the second embodiment. First, a defect check of the ROM FUSE block is executed (step S 11 : first defect detection unit or first defect detection circuitry). In the defect check of step S 11 , for example, the existing block check is executed on memory cells on a specific word line in the ROM FUSE block. Alternatively, any one of the E-to-A detection methods 1 to 3 may be adopted.

When it is determined in step S 11 that the ROM FUSE block is normal, it is determined that the check of the ROM FUSE block has passed in (step S 19 ) and the process of FIG. 4 ends. In this case, the refresh operation of the ROM FUSE block is not executed.

Meanwhile, when it is determined in step S 11 that the ROM FUSE block is defective, an address is issued to have access to a spare ROM FUSE block (step S 12 ) and a defect check of the spare ROM FUSE block is executed (step S 13 : second defect detection unit or second defect detection circuitry).

When it is determined that the spare ROM FUSE block is defective, defect information in the spare ROM FUSE block cannot be effectively used. For this reason, it is determined that the refresh operation of the ROM FUSE block does not need to be executed and it is determined that the check of the ROM FUSE block has failed in (step S 14 ). In this case, the NAND-type flash memory including the ROM FUSE block is assumed as a defective product.

When it is determined that the spare ROM FUSE block is normal, all memory cells in the ROM FUSE block are erased (step S 15 : data erasure unit or data erasure circuitry). As a result, data “1” is written to all of the memory cells in the ROM FUSE block.

After a process of step S 15 , entire defect information stored in the spare ROM FUSE block is latched temporarily by a data latch circuit 7 in a sense amplifier 6 (step S 16 : temporary storage).

Next, the defect information in the spare ROM FUSE block latched temporarily by step S 16 is written to the ROM FUSE block (step S 17 : defect information write unit or defect information write circuitry). Steps S 15 to S 17 correspond to a refresh process unit or refresh process circuitry.

In the write of step S 17 , defect information may be written to memory cells on a word line determined as a defect by step S 11 in the ROM FUSE block and the defect information may be written to memory cells on word lines other than the word line determined as the defect. When the defect occurs due to a structural failure in a specific memory cell, it is preferable that the memory cell be not used. However, when the defect occurs due to temporary read disturbance, rewriting to the memory cell having caused the defect may not cause a problem. Therefore, in this embodiment, rewriting to the memory cell determined as the defect is allowed.

If the refresh process of steps S 15 to S 17 ends, it is assumed that there is no defect in the ROM FUSE block and a pass is determined (step S 18 ).

FIG. 5 is a timing diagram when the defect information is rewritten to the memory cell determined as the defect in the past at the time of the refresh operation, in step S 17 of FIG. 4 . FIG. 5 illustrates signal waveforms of a ready/busy signal RB, a word line WL 15 PB 0 of the ROM FUSE block, a substrate well region CPWELL PB 0 of a memory cell connected to the word line, and a word line WL 15 PB 1 of the spare ROM FUSE block.

A power-on read operation starts after the ready/busy signal RB changes from a high level to a low level and the refresh operation is executed during the power-on read operation. First, data is read from memory cells on a specific word line (for example, a word line WL 15 ) in the ROM FUSE block (time t 1 ). At the time t 1 , the word line WL 15 of the ROM FUSE block is set to a read voltage VCGRV.

When it is determined by the read data that the ROM FUSE block is defective, data is read from memory cells on a specific word line (for example, a word line WL 15 ) in the spare ROM FUSE block (time t 2 ) and it is determined whether there is a defect in the spare ROM FUSE block.

When it is determined that the spare ROM FUSE block is normal, data of all memory cells in the ROM FUSE block determined as the defect is erased (time t 3 ). The memory cells in which the data has been erased are verified and it is confirmed whether the data of the memory cells has been erased correctly (time t 4 ). The data erasure and the verification of the memory cells in the ROM FUSE block are repeated until data erasure and verification for all of the memory cells in the ROM FUSE block end.

If the data erasure of the ROM FUSE block ends, the defect information from the spare ROM FUSE block is read and is held temporarily by the data latch circuit in the sense amplifier. Then, the held defect information is written to the ROM FUSE block (time t 5 ). Next, it is verified whether the written defect information is normally read (time t 6 ). The processes of the times t 5 and t 6 are continuously executed until the entire defect information in the spare ROM FUSE block is written to the ROM FUSE block.

Information showing whether the refresh operation of the ROM FUSE block has been executed may be stored in a register circuit 8 a (refer to FIG. 1 ) in the controller 8 . For example, the register circuit (refresh history information storage) 8 a composed of a non-volatile memory such as an SRAM capable of storing data of at least one byte may be provided, the information showing whether the refresh operation of the ROM FUSE block has been executed may be stored in one bit of one byte, and the data of one byte of the register circuit 8 a may be output to the outside according to a specific command from the outside. As a result, history information showing whether the refresh operation of the ROM FUSE block is executed in the past can be detected from the outside of the NAND-type flash memory and reliability of the ROM FUSE block can be evaluated by the history information.

As such, in the second embodiment, when it is determined that the ROM FUSE block is defective, the refresh operation to write the defect information in the spare ROM FUSE block after the data of all of the memory cells in the ROM FUSE block is erased is executed. Therefore, reliability of the defect information in the ROM FUSE block can be improved and a level of defectiveness of the ROM FUSE block can be decreased.

Third Embodiment

In a third embodiment to be described below, it is automatically determined whether a refresh operation of a ROM FUSE block is executed in the past and the refresh operation is not executed again on the ROM FUSE block on which the refresh operation is executed in the past.

A NAND-type flash memory according to the third embodiment includes the same block configuration as the block configuration of FIG. 1 and a process operation of a controller 8 is different from the process operations of the first and the second embodiments. Therefore, the process operation of the controller 8 will be described hereinafter.

FIG. 6 is a flowchart illustrating a defect check process of the ROM FUSE block according to the third embodiment. First, data of memory cells on a specific word line in the ROM FUSE block is read (step S 31 ).

Data of all memory cells of a non-used segment on the specific word line are all “1”. Meanwhile, if “0” is initially written to all of the memory cells of the non-used segment on the specific word line in the ROM FUSE block on which a refresh operation is not executed, “0” is maintained in all of the memory cells.

Therefore, it is determined whether data of the memory cells of the non-used segment on the specific word line in the ROM FUSE block are all “1” (step S 32 : refresh determination unit or refresh determination circuitry). When the data of the memory cells are all “1”, it is determined that the refresh operation of the ROM FUSE block is executed in the past, data of memory cells on a predetermined word line other than the specific word line in the ROM FUSE block is read, and it is determined whether there is a defect (step S 33 ). Here, bad block information or bad column information is stored previously as complementary data in the ROM FUSE block, the complementary data is read, two kinds of data configuring the complementary data are compared, and it is determined whether there is the defect.

In step S 33 , when data read from the ROM FUSE block has been normally executed, it is determined that a check of the ROM FUSE block has passed in (step S 34 ), and a process of FIG. 6 ends.

Meanwhile, when it is determined in step S 33 that the ROM FUSE block is defective, the refresh operation of the ROM FUSE block is not executed and a change of an address to have access to a spare ROM FUSE block is executed (step S 35 ). In addition, memory cells on a specific word line in the spare ROM FUSE block are read and it is checked whether the spare ROM FUSE block is defective (step S 36 ). Here, similar to step S 33 , the complementary data is read and it is determined whether there is a defect. As a result, when it is determined that the spare ROM FUSE block is defective, it is determined that the check of the ROM FUSE block has failed in (step S 37 ). As a result, the NAND-type flash memory including the ROM FUSE block is assumed as a defective product.

When it is determined in step S 36 that the spare ROM FUSE block is not defective, it is determined that the check of the ROM FUSE block has passed in (step S 34 ). In this case, the ROM FUSE block is not used and the defect information in the spare ROM FUSE block is used.

Meanwhile, in step S 32 , when the data of all of the memory cells of the non-used segment on the specific word line in the ROM FUSE block are all “0”, it is determined that the refresh operation of the ROM FUSE block has not been executed and an existing block check is executed (step S 38 ). In step S 38 , the complementary data of the memory cells on the specific word line in the ROM FUSE block is read and it is determined whether there is a defect.

When it is determined in step S 38 that the ROM FUSE block is normal, it is assumed that the check of the ROM FUSE block has passed in (step S 34 ). Meanwhile, when it is determined in step S 38 that the ROM FUSE block is defective, the refresh process of steps S 12 to S 20 of the flowchart of FIG. 4 is executed.

As such, in the process of FIG. 6 , it is confirmed whether the refresh operation is executed on the ROM FUSE block in the past. If the refresh operation is executed on the ROM FUSE block in the past, the spare ROM FUSE block is used without executing the refresh operation again, even though the ROM FUSE block becomes defective again. Meanwhile, if the refresh operation is not executed on the ROM FUSE block in the past, at the time of becoming defective, the refresh operation is executed on the ROM FUSE block, the defect information in the spare ROM FUSE block is written to the ROM FUSE block, and the ROM FUSE block is continuously used.

FIG. 7 is a timing diagram illustrating process timing of a power-on read operation after the ROM refresh operation when the process changes to steps S 31 , S 32 , S 33 , and S 35 of FIG. 6 . First, data is read from memory cells on a specific word line (for example, a word line WL 15 ) in the ROM FUSE block (time t 11 ). When the read data are all “1”, it is determined that the refresh operation is executed on the ROM FUSE block in the past, the refresh operation of the ROM FUSE block is not executed gain, and data of memory cells on a predetermined word line other than the specific word line in the ROM FUSE block is read (time t 12 ). In addition, the block check of the ROM FUSE block is executed on the basis of the read data. When it is determined that the ROM FUSE block is defective, data of memory cells on a specific word line (for example, a word line WL 15 ) in the spare ROM FUSE block is read (time t 13 ).

As such, in the third embodiment, the data of the memory cells on the specific word line in the ROM FUSE block is read and it can be accurately determined whether the refresh operation is executed on the ROM FUSE block in the past, according to whether the read data are all “1”.

In addition, the refresh operation is not executed again on the ROM FUSE block on which the refresh operation is executed in the past. Therefore, a ROM FUSE block having low reliability can be prevented from being continuously used.

When it is determined that the ROM FUSE block on which the refresh operation is not executed in the past is defective, similar to the second embodiment, the refresh operation of the ROM FUSE block is executed. Therefore, a level of defectiveness of the ROM FUSE block can be reduced.

Fourth Embodiment

A fourth embodiment to be described below is characterized by a method of storing defect information in a ROM FUSE block.

FIG. 8 is a diagram illustrating a memory map of a memory cell array 2 including a ROM FUSE block 31 . As illustrated in FIG. 8 , the ROM FUSE block 31 is one of a plurality of blocks provided in the memory cell array 2 . The ROM FUSE block 31 is divided into four data areas d 0 to d 3 . Information stored in the four data areas d 0 to d 3 are the same. The information stored in the individual data areas includes bad column information d 00 , parameter information d 01 , and bad block information d 02 . When it is determined that one data area is defective, the same type of information can be read from other data area.

FIG. 9A is a diagram illustrating a data configuration of the bad block information stored in each data area of the ROM FUSE block 31 . In the bad block information of FIG. 9A , defect information of 8 blocks is stored in 4 bytes. More specifically, address information of 8 bits showing a quotient obtained by dividing a block address by 8 is stored in the first byte. Inversion data of the address information of the 8 bits of the first byte is stored in the second byte. Flag information showing whether each of the 8 blocks is defective is stored in one bit in the third byte. Inversion data of 8-bit data of the third byte is stored in the fourth byte.

In FIG. 9A , the quotient obtained by dividing the block address by 8 is represented as Block Add X and the 8-bit data including the flag information of the 8 bits showing whether each of the 8 blocks is defective is represented as Block IO.

FIG. 9B is a diagram illustrating a correspondence relation of each block in the memory cell array 2 , the quotient Block Add X obtained by dividing the block address of each block by 8, and the 8-bit data Block IO showing whether each block is defective.

As known from FIGS. 9A and 9B , in this embodiment, the 8 block addresses are converted into one address information Block Add X, the flag information showing whether each block is defective is represented by one bit, and the flag information of the 8 blocks is stored as one byte in the ROM FUSE block 31 . In addition, in this embodiment, when there is no defective block in the 8 blocks represented by Block Add X, information is not stored in the ROM FUSE block 31 . As a result, defect information of a large amount of blocks can be stored in the ROM FUSE block 31 .

The defect information stored in the ROM FUSE block 31 according to FIGS. 9A and 9B is latched by a latch circuit (defect information holding unit or defect information holding circuitry) 3 a provided in a block selection circuit (not illustrated in the drawings) in a row decoder 3 . The block selection circuit refers to the data latched by the latch circuit 3 a , so that the block selection circuit selects whether or not to have access to each block.

FIG. 10 is a flowchart illustrating an example of a process sequence to transfer defective block information from the ROM FUSE block 31 to the latch circuit 3 a in the block selection circuit. The flowchart is executed by a controller 8 , for example. First, the defective block information is read from the ROM FUSE block 31 (step S 61 ). The read defective block information is stored temporarily. Next, the address information Block Add X and the inversion data thereof stored in the first and second bytes of FIG. 9A are compared and the flag information Block IO and the inversion data thereof stored in the third and fourth bytes of FIG. 9A are compared (step S 62 ).

When the comparison of step S 62 has passed in, it is determined that there is no defect in the ROM FUSE block 31 and the address information Block Add X and Block IO are converted into an original block address (step S 63 : address conversion unit or address conversion circuitry). Defect information of a corresponding block is stored in the latch circuit in the block selection circuit corresponding to the block address converted by step S 63 (step S 64 ).

Next, it is determined whether data has been read until final data in the ROM FUSE block 31 (step S 65 ). When non-read data exists, next 4-byte data in the ROM FUSE block 31 is read (step S 66 ) and the process proceeds to step S 62 .

FIG. 11 is a flowchart illustrating an example of a process sequence to store (register) defect information in the ROM FUSE block 31 at the time of a test before a shipment. First, a block 0 in the memory cell array 2 is selected (step S 71 ). In this embodiment, block addresses of 8 blocks in which at least one block is defective is stored in the ROM FUSE block 31 as one address information. For this reason, an existing block check is executed on the 8 blocks, sequentially from the block 0 , and it is checked whether there is a defect in the blocks (step S 72 ).

When it is determined that there is a defect in the selected block, the address information Block Add X represented by the quotient obtained by dividing the block address corresponding to the block by 8 is latched by a data latch circuit in a sense amplifier (step S 73 ). Next, inversion data of the address information Block Add X latched by step S 73 is latched by a data latch circuit different from the data latch circuit used in step S 73 (step S 74 ).

Next, flag information showing whether which of the 8 blocks is defective is held as 8-bit data in a data latch circuit different from the data latch circuits used in steps S 73 and S 74 (step S 75 ). Next, inversion data of the data held by step S 75 is held in a data latch circuit different from the data latch circuits used in steps S 73 to S 75 (step S 76 ).

Next, it is determined whether the selected block is a final block of the 8 blocks stored collectively (step S 77 ). When the selected block is not the final block, a block number is increased by 1 (step S 78 ) and the processes after step S 72 are repeated. When it is determined in step S 77 that the selected block is the final block, the latch data of the 4 data latch circuits held by steps S 73 to S 76 are transferred as data corresponding to 4 column addresses to the ROM FUSE block 31 , such that data of all defective blocks are collected and transferred to the ROM FUSE block 31 (step S 79 ). When it is determined in the defect check of step S 72 that there is no defect, the process of step S 77 is executed. The above processes are executed on all blocks of the memory cell array 2 .

As such, in the fourth embodiment, the quotient obtained by dividing each block address or each column address by 8 is set as the address and the data in which the defect information of each block or each column is represented by one bit is stored in the ROM FUSE block 31 . Therefore, the defect information that can be stored in the ROM FUSE block 31 can be increased. In addition, in this embodiment, because information of a non-defective block or a non-defective column is not stored in the ROM FUSE block 31 , storage areas of the ROM FUSE block 31 can be effectively used. Therefore, when the power-on read operation is executed, read of the defect information from the ROM FUSE block 31 can be executed at a high speed. According to this embodiment, the defect information of the 8 blocks can be stored in only 4 bytes and defect information of blocks of an amount 8 times larger than an amount in a comparative example in which defect information of one block is stored in 4 bytes can be stored.

The data configuration of the ROM FUSE block 31 according to the fourth embodiment can be applied to the ROM FUSE blocks 31 according to the first to third embodiments. As a result, even in the first to third embodiments, a large amount of defect information can be stored in the ROM FUSE block 31 and it is possible to correspond to an increase in the capacity of the NAND-type flash memory.

In the embodiments described above, the example of the case in which the present invention is applied to the NAND-type flash memory has been described. However, the present invention can be applied to various non-volatile semiconductor storage devices having a ROM FUSE block, such as a NOR-type flash memory, an MRAM, and a ReRAM.

While certain embodiments have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. Indeed, the novel methods and systems described herein may be embodied in a variety of other forms; furthermore, various omissions, substitutions and changes in the form of the methods and systems described herein may be made without departing from the spirit of the inventions. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the inventions.