This application is based upon and claims the benefit of priority from the prior Japanese Patent Application No. 2013-187896, filed on Sep. 11, 2013, the entire contents of which are incorporated herein by reference.
1. Field
Embodiments of the present invention relate to a nonvolatile semiconductor memory device.
2. Description of the Related Art
Currently, semiconductor memories are utilized in any number of places, from large-scale computers, personal computers, and household electrical goods to mobile phones, and the like. Receiving particular attention among these semiconductor memories is flash memory. Flash memory is utilized in many information devices such as mobile phones or digital cameras, for reasons such as it being a nonvolatile memory or it having a structure suitable for a high degree of integration, and so on.
A nonvolatile semiconductor memory device according to an embodiment comprises: a memory string including a first memory cell and a second memory cell; a first word line connected to a gate of the first memory cell; a second word line connected to a gate of the second memory cell; and a peripheral circuit configured to control a write sequence and a read sequence, the peripheral circuit, during the write sequence or the read sequence on the first memory cell, executing a first operation on the condition that a positive first pass voltage is applied to the first word line and the second word line. Here, the word “connected” of “a first word line connected to a gate of the first memory cell” and “a second word line connected to a gate of the second memory cell” means a physical connection or an electrical connection, and includes a case of being electrically connected through a current path of a transistor.
A semiconductor memory device according to embodiments will be described below with reference to the drawings.
<Overall Configuration>
First, an overall configuration of a nonvolatile semiconductor memory device according to a first embodiment will be described.
A NAND flash memory being the nonvolatile semiconductor memory device according to the present embodiment comprises a cell array 1 and a peripheral circuit. The peripheral circuit includes a row decoder/word line driver 2a and column decoder 2b, a page buffer 3, a row address register 5a and column address register 5b, a logic control circuit 6, a sequence control circuit 7, a high voltage generating circuit 8, an I/O buffer 9, and a controller 11.
The cell array 1 has a so-called BiCS (Bit-Cost-Scalable) structure. The cell array 1 includes a plurality of memory strings similarly to a cell array of a NAND flash memory of planar structure. Each of the memory strings includes a plurality of cells connected in series. Each of the cells is configured by a transistor (hereafter, referred to as “cell transistor” or “memory cell”) having a charge accumulation layer. The cell array 1 will be described in detail later.
The row decoder/word line driver 2a drives a word line and a select gate line of the cell array 1. The page buffer 3 comprises a one-page portion of sense amplifiers and a data holding circuit, and controls read of data of the cell array 1 in page units. The page buffer 3 performs column selection sequentially by the column decoder 2b to output a one-page portion of read data to an external I/O terminal via the I/O buffer 9. Every one page of write data supplied from the I/O buffer 9 is selected by the column decoder 2b to be loaded into the page buffer 3. A row address signal and a column address signal are inputted via the I/O buffer 9 to be respectively transferred to the row decoder/word line driver 2a and the column decoder 2b. The row address register 5a holds an erase block address during an erase sequence, and holds a page address in a write sequence or a read sequence. The column address register 5b has inputted thereto a lead column address required in load of write data prior to start of the write sequence or a lead column address required in the read sequence. The column address register 5b holds an inputted column address until a write enable signal/WE or read enable signal/RE are toggled by a certain condition. The logic control circuit 6 controls input of a command or address and input/output of data based on control signals such as a chip enable signal/CE, a command enable signal CLE, an address latch enable signal ALE, the write enable signal/WE, and the read enable signal/RE. The sequence control circuit 7 receives a command from the logic control circuit 6 to control the erase sequence, the read sequence, or the write sequence. That is, the sequence control circuit 7 controls the row address register 5a, the column address register 5b, the row decoder/word line driver 2a, and so on, thereby controlling the erase sequence, the read sequence, or the write sequence. The high voltage generating circuit 8 is controlled by the sequence control circuit 7 to generate a certain voltage required in various kinds of operations. The controller 11 controls the write sequence, and so on, by conditions appropriate to a current read state, and so on. Note that as required, the page buffer 3 may comprise a data latch DL for holding open failure information, which will be described later.
<Cell Array>
Next, a specific example of the cell array 1 will be described.
This cell array 1 has formed sequentially therein, on a semiconductor substrate: a plurality of source lines SL that are aligned in the Y direction and extend in the X direction; a plurality of source side select gate lines SGS that are aligned in the Y direction and extend in the X direction; a plurality of word lines WL that are planar extending in the X direction and the Y direction; a plurality of drain side select gate lines SGD that are aligned in the Y direction and extend in the X direction; and a plurality of bit lines BL that are aligned in the X direction and extend in the Y direction. Moreover, formed between each of the plurality of source lines SL and the plurality of bit lines BL is a pillar that penetrates the source side select gate line SGS, the plurality of word lines WL, and the drain side select gate line SGD. This pillar configures part of a memory string MS.
Next, another specific example of the cell array 1 will be described.
This cell array 1 includes, on a semiconductor substrate: a plurality of word lines WL (first lines) that are aligned in a two-dimensional matrix in the Y direction and the Z direction, and extend in the X direction; a plurality of select gate lines that are aligned in the Y direction and extend in the X direction; a plurality of source lines SL that are aligned in the Y direction and extend in the X direction; and a plurality of bit lines BL that are aligned in the X direction and extend in the Y direction. Note that the plurality of select gate lines have the source side select gate lines SGS and the drain side select gate lines SGD aligned alternately two at a time in the Y direction. Moreover,
The memory string MS shown in
For convenience of description, the cell array 1 having a BiCS structure shown in
As shown in
The back gate layer 130 includes a back gate conductive layer 131 formed on the semiconductor substrate 110 via the insulating layer 120. The back gate conductive layer 131 functions as the back gate line BG and as a gate of the back gate transistor BTr. In addition, the back gate layer 130 includes a back gate trench 132 formed so as to dig into the back gate conductive layer 131.
The memory transistor layer 140 includes a plurality of word line conductive layers 141 formed in the Z direction while interposing an insulating layer 142. The word line conductive layer 141 functions as the word line WL and as a gate of the memory transistor MTr. In addition, the memory transistor layer 140 includes a memory hole 143 formed so as to penetrate the plurality of word line conductive layers 141 and a plurality of the insulating layers 142.
Moreover, the back gate layer 130 and the memory transistor layer 140 include a memory gate insulating layer 144 and a semiconductor layer 145. As shown in
The select transistor layer 150 includes a drain side conductive layer 151 and a source side conductive layer 152 that are formed in the same layer. The drain side conductive layer 151 functions as the drain side select gate line SGD and as a gate of the drain side select transistor SDTr. The source side conductive layer 152 functions as the source side select gate line SGS and as a gate of the source side select transistor SSTr. In addition, the select transistor layer 150 includes a drain side hole 153, a source side hole 154, a drain side gate insulating layer 155, a source side gate insulating layer 156, a drain side columnar semiconductor layer 157, and a source side columnar semiconductor layer 158. The drain side columnar semiconductor layer 157 functions as a body of the drain side select transistor SDTr. The source side columnar semiconductor layer 158 functions as a body of the source side select transistor SSTr.
The wiring line layer 160 includes a first wiring line layer 161, a second wiring line layer 162, and a plug layer 163. The first wiring line layer 161 functions as the source line SL. The second wiring line layer 162 functions as the bit line BL.
<Open Failure Detection of Memory String>
When employing the cell array 1 of BiCS structure, the following problem occurs.
It is conceivable that when manufacturing the cell array 1 of BiCS structure, a plurality of conductive layers and insulating films are stacked to process the memory hole 143 in one shot.
However, this memory hole 143 has a depth of as much as, for example, about 1.5 μm and has a high aspect ratio, hence processing is difficult. As a result, sometimes, the memory hole 143 stops midway and a so-called open failure occurs in the memory hole 143 (Hereafter, a memory string in which there is an open failure in the memory hole will sometimes be referred to simply as an “open failure memory string”. Moreover, a memory transistor included in an open failure memory string will sometimes be referred to simply as an “open failure memory transistor”). In this case, write and read of data to/from that open failure memory transistor MTr cannot be performed. Moreover, if many open failure memory strings MS exist in the memory block MB, relief by ECC also becomes impossible, leading to the entire memory block MB being treated as a failure.
Accordingly, in the present embodiment, a countermeasure for the open failure memory string MS is adopted during the write sequence. Note that write sequence refers to a series of processings of data write on the memory transistor MTr.
First, as a precondition to describing the write sequence of the present embodiment, a relationship between a threshold voltage and data of the memory transistor MTr will be simply described.
Set as a threshold voltage Vth of the memory transistor MTr are, in order of increasing voltage, four voltage ranges, namely level E, level A, level B, and level C. Neighboring levels are separated from each other by a certain margin. Moreover, for example, four data values, namely “11”, “01”, “00”, and “10”, correspond to level E, level A, level B, and level C. The nonvolatile semiconductor memory device stores four different data by causing the threshold voltage Vth of the memory transistor MTr to undergo transition to a desired level.
Next, as a precondition to describing the write sequence of the present embodiment, a read operation from the memory transistor MTr will be simply described. Note that this read operation is employed as an “ordinary read operation” in step S201 of a read sequence shown in
The read operation is executed targeting a selected memory transistor MTr of a selected memory string MS of a selected memory block MB. First, the page buffer 3 charges the bit line BL to “H” level, and the row decoder/word line driver 2a applies a reference voltage Vrf to a selected word line WL and a read voltage Vread to an unselected word line WL. Now, the reference voltage Vrf is, for example, any of a voltage Vra between level E and level A, a voltage Vrb between level A and level B, and a voltage Vrc between level B and level C, shown in
Then, the current flowing in this bit line BL is detected by the sense amplifier included in the page buffer 3, whereby data of the selected memory transistor MTr is determined. Specifically, when reading data of a lower bit of the selected memory transistor MTr, the selected memory string MS is put in the above-described bias state, and then the row decoder/word line driver 2a applies the reference voltage Vrf=Vrb to the selected word line WL. If, as a result, the selected memory transistor MTr is turned on and a current flows in the bit line BL, then the threshold voltage Vth of that selected memory transistor MTr is at level E or level A, hence the lower bit is understood to be “1”.
Next, the write sequence of the present embodiment will be described.
The write sequence according to the present embodiment includes two operations, that is, a memory hole detection operation (first operation) that detects presence/absence of an open failure of a memory hole in the memory block MB, and a write operation that writes data to the memory transistor MTr.
First, when an instruction of data write is inputted from the controller 11 via I/O (this instruction is, for example, a command, such as “80h-Add-10h” of
When the memory hole detection operation is entered, the page buffer 3 charges the bit line BL to “H” level, and the row decoder/word line driver 2a applies the read voltage Vread to all of the word lines WL, including selected/unselected word lines WL (step S104). Then, the row decoder/word line driver 2a applies the select gate voltage Vsg to the source side select gate line SGS and the drain side select gate line SGD of the selected memory string MS (step S105). At this time, the source side select gate line SGS and the drain side select gate line SGD of the unselected memory string MS are maintained unchanged at a ground voltage Vss.
If it is assumed there is an open failure in the selected memory string MS, then the selected memory string MS never conducts even if the read voltage Vread is applied to the selected word line WL. This is because the memory hole of the selected memory string MS is an open failure and a current does not flow from the bit line BL to the source line SL. Since the selected memory string MS does not conduct regardless of a voltage applied to the selected word line WL, then as shown in
On the other hand, if it is assumed there is not an open failure in the selected memory string MS, then the read voltage Vread which is higher than the highest level C is applied to the gates of all of the memory transistors MTr, hence all of the memory transistors MTr attain an on state regardless of what data the memory transistor MTr itself is storing. In this case, the selected memory string MS conducts, hence a current flows from the bit line BL to the source line SL, and the bit line BL lowers to “L” level. In this case, the page buffer 3 determines there is at least not an open failure in this selected memory string MS.
The memory hole detection operation of the above kind is performed on all of the memory strings MS in the memory block. Subsequently, when, for example, there is/are a certain number or more of the open failure memory strings MS in the memory block, then as shown by the dotted line of
Here, reference will be made to the case where the write operation includes a verify operation (read operation) that verifies whether write of data has been completed or not.
In such a case, if there is an open failure memory string MS, then the following problem occurs. In other words, since write of data cannot be performed in the open failure memory transistor MTr, then, depending on the data it is desired to write, verify cannot be passed. This case results in programming on the memory transistor MTr being repeated, and processing of the write sequence being delayed.
Accordingly, if it is desired to solve such a problem, then the write operation is processed making the open failure memory transistor MTr non-target of programming. For example, in the verify operation (read operation), when the open failure memory transistor MTr maybe regarded as storing a data value “1”, then, regardless of input data, the open failure memory transistor MTr is made non-target of programming (programming prohibited), after which the verify operation is processed assuming write data was “1”. In this way, the open failure memory transistor MTr can immediately pass verify, hence a wasted verify operation on the open failure memory transistor MTr can be reduced. Note that in the case of this method, an error is sometimes also included in stored data, but in this case all that is required is to perform error correction by the likes of an external ECC system.
Next, reference will be made also to a voltage application method on the word line WL during switching from the memory hole detection operation to the write operation.
In each of the examples of
The voltage application methods on the word line WL in the examples of
Note that in the memory hole detection operation, the number of open failure memory strings MS may be counted, or only a non-write and open failure memory string MS, in other words, a memory string MS whose status is read failure, may be detected.
In addition, an open failure detection result (hereafter, referred to as “memory hole data”) may be configured to, for example, be held in a data latch DL provided in the likes of the page buffer 3 and be readable from the controller 11.
In the case of a cell array of BiCS structure, lithography of a critical layer performed in each layer is unnecessary, hence the cell array of BiCS structure is advantageous over a conventional stacked-structure cell array in terms of cost. On the other hand, in the case of the cell array of BiCS structure, it has been a problem that it is required to form a memory hole having a depth of about 1.5 μm in the stacking direction, and an open failure occurs frequently in the memory hole.
Regarding this frequent occurrence of open failure, in the present embodiment, detection of open failure is performed before the write operation, hence write to an open failure memory string can be avoided. As a result, data reliability can be improved even when employing a cell array of BiCS structure.
In the first embodiment, a countermeasure for the open failure memory string MS during the write sequence was described, but in a second embodiment, a countermeasure for the open failure memory string MS during the read sequence will be described.
In the read sequence of the present embodiment, first, the sequence control circuit 7 executes the ordinary read operation described in the first embodiment (step S201). Next, the sequence control circuit 7 executes ECC on read data (step S202). Now, if an error of the read data is in a range of correction capability of ECC and error correction by ECC has succeeded, then the read sequence is finished adopting that corrected data as output data (step S203). On the other hand, if an error of the read data has exceeded the range of correction capability of ECC, then processing is shifted to step S204. In step S204, after execution of the ordinary read operation, the sequence control circuit 7 executes the memory hole detection operation similar to that of the first embodiment. Now,
Subsequently, ECC is executed on the read data that has undergone the computational processing in step S204 (step S205). Now, in the case where error correction by ECC has succeeded, the read sequence is finished adopting the post-computational processing data as output data (step S203). On the other hand, in the case where error correction by ECC has been unable to be performed, the read sequence is finished as a read failure (step S206). Note that correction capability may be changed between ECC in step S204 and ECC in step S202.
Next, the computational processing in step S204 will be described.
Now,
In the case shown in
Now, if, for example, error correction capability of ECC was two bits, then in the case of read data generated by the processing of A in
As described above, the present embodiment enables an apparent effect of open failure to be reduced by half, hence enables reliability of data to be improved. In addition, products having an amount of open failures up to twice the conventional tolerance number of open failures at shipment can be made shipment-target products, hence leading also to yield improvement.
[Other]
While certain embodiments of the inventions 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.
Number | Date | Country | Kind |
---|---|---|---|
2013-187896 | Sep 2013 | JP | national |