Electrical computers and digital processing systems: memory – Address formation – Address mapping
Reexamination Certificate
2000-11-28
2003-09-16
Portka, Gary (Department: 2188)
Electrical computers and digital processing systems: memory
Address formation
Address mapping
C711S005000, C711S103000, C365S230030
Reexamination Certificate
active
06622230
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to the field of flash memory devices. More particularly, the invention relates to both a mechanism and method for multi-set block erase in NAND-based flash memory devices.
BACKGROUND OF THE INVENTION
The overall array architecture for memory section of a typical NAND-based flash memory device comprises a core memory accessed by an upper and lower bank of page buffers and a right and left bank of word line or X-decoders. The core memory contains information stored in blocks of memory and individual memory cells or elements within the memory blocks. The right and left word line or X-decoders are used to access specific memory cells within each memory block and the upper and lower bank of page buffers provide the input and output circuitry for each memory cell. The X-decoders contain numerous final decoder circuits, each with a high voltage generating pump. There is generally one final decoder circuit for each memory block. Each final decoder circuit is linked by various predecoder output lines to a predecoder circuit.
The architecture of one memory block in the typical NAND-based flash memory device comprises the individual memory elements and select gates. The memory elements and select gates are embodied in non-volatile, floating gate transistors that may be programmed to a logic state of 0, 1, or other states depending on the particular type of transistor and programming used. The control gates of the transistors that comprise the individual memory elements and select gates in each memory block are addressed by word lines controlled by the addressing system. The memory elements are connected in series with each other and the select gates. The select gates, at the ends of the chain of memory cells, are connected with either the array common voltage Vss or a bitline. A page buffer is connected with a memory block via a bitline. The page buffer includes transistors and supporting circuitry that regulate the flow of data into and out of the memory block and into and out of the external system.
The array architecture associated with the memory device is shown in FIG. 
1
. The array contains N individual page buffers 
110
 in each bank of page buffers 
102
,
104
; thus, as there are both an upper 
102
 and lower 
104
 bank of page buffers, one entire page or word line 
112
 contains 2N bits. Left 
106
 and right 
108
 banks of X-decoders are used to select a particular word line. The core memory 
100
 is split into a set of M memory blocks 
114
 with each memory block 
114
 being L pages wide. This results in a 2N bits/page×L pages/memory block×M memory blocks/core memory=2×L×M×N bytes/core memory. One example of an array architecture used is 256 individual page buffers, a set of 1024 memory blocks with each memory block being 16 pages wide. This results in a 2×256 bytes/page×16 pages/memory block×1024 memory blocks/core memory=8M bytes/core memory. Of course, any numbers presented here are merely illustrative of the principle of the invention as a whole. Those ordinarily skilled in the art will appreciate that the numbers associated with any of these elements, as well as the number of memory cells or elements in the overall device, may be changed without departing from the spirit and scope of the invention.
FIG. 2
 shows the architecture of one memory block 
114
 in the NAND-based flash memory device of 
FIG. 1
 along with the associated page buffer 
110
. The memory block 
114
 comprises many parallel strings like the string of 
16
 individual memory elements 
130
-
145
 and two select gates, SG
1
116
 and SG
2
118
, for example. As stated before, non-volatile transistors embody the memory elements 
130
-
145
 and select gates 
116
, 
118
. The word lines 
151
-
166
 and select gate lines 
150
, 
167
 are connected with the control gates of the memory elements 
130
-
145
 and select gates 
116
, 
118
. The memory elements 
130
-
145
 and select gates SG
1
116
 and SG
2
118
 are connected in series. Specifically, the source and drain of the memory elements 
130
-
145
 are connected to each other in series. Thus, the source of the first memory element 
130
 is tied to the drain of the memory second element 
131
, the source of the second memory element 
131
 is tied to the drain of the third memory element 
132
, etc. Accordingly, the select gates 
116
, 
118
 are connected to the ends of the chain of memory elements 
130
-
145
. The source of SG
1
116
 is connected with the drain of the first memory element 
130
, while the drain of SG
1
116
 is connected with a bitline 
170
. Similarly, the drain of SG
2
118
 is connected with the source of the last memory element 
145
 and the source of SG
2
118
 is connected with the array common (usually ground) voltage Vss 
172
.
The memory block 
114
 is connected with a page buffer 
110
 via a bitline 
170
. The page buffer 
110
 includes necessary circuitry well known in the art. The necessary circuitry may include, for example, a latch for latching data onto and out of the bitline 
170
, input/output circuitry for transferring data to the external system and assorted supporting circuitry inherent in the necessary circuitry.
Individual memory elements can undergo three distinct operations, which are shown in 
FIGS. 3
, 
4
 and 
5
. The three operations are Program, shown in 
FIG. 3
, Erase, shown in 
FIG. 4
, and Read, shown in 
FIG. 5
, and are described below. This discussion will be limited to standard, n-channel, NAND-based non-volatile memory elements, although those ordinarily skilled in the art will appreciate that the basic operations described herein can be easily extended to at least NOR-based non-volatile memory elements and multi-level non-volatile memory elements in which more than two states can be programmed.
The structure of the memory element 
200
 is well known in the art: a p-type semiconductor well 
210
 is disposed within a n-type semiconductor well 
206
. The n-type semiconductor well 
206
 is contained within a p-type semiconductor substrate 
202
. A set of n-type semiconductor junctions comprising the source 
204
 and drain 
208
 are disposed within the p-type semiconductor well 
210
. The p-type semiconductor well 
210
 and n-type semiconductor well 
206
 are usually maintained at the same voltage during device operation to avoid current flow from one of the well regions to the other.
The memory element further includes a control gate 
212
 and a floating gate 
214
. The gates 
212
, 
214
 are conventionally formed from polysilicon deposited and patterned on the surface of the substrate, although the floating gate 
214
 may alternately be formed from an ONO layer. The gates 
212
, 
214
 are formed such that an oxide is formed on part of the substrate with the floating gate 
214
 formed above the oxide. The control gate 
212
 is formed above the floating gate 
214
 and isolated from the floating gate 
214
 by a second oxide. Control signals are applied to the control gate 
212
.
During the program operation, as shown in 
FIG. 3
, both the source 
204
 and the drain 
208
 of the memory element 
200
 are connected with Vss (usually ground). Prior to programming, the threshold voltage (or turn-on voltage) of the MOSFET is generally designed to be a negative voltage, so that a channel 
216
 of electrons 
218
 exists in the p-type semiconductor well 
210
 when the gate 
212
 is grounded. The channel 
216
 is disposed between the source 
204
 and drain 
208
 of the memory element 
200
. A large positive voltage is applied to the control gate 
212
, which causes electrons 
218
 to be trapped onto the floating gate 
214
 via Fowler-Nordheim tunneling. The threshold voltage of the transistor is increased if electrons are trapped on the floating gate 
214
. In this case, the threshold voltage of the programmed memory element changes from a negative voltage to a positive voltage.
During the erase operation, as shown in 
FIG. 4
, the source 
204
 and the drain 
208
 of the memory element 
200
 are left floating wh
Chung Michael
Hollmer Shane
Yano Masaru
Advanced Micro Devices , Inc.
Portka Gary
LandOfFree
Multi-set block erase does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Multi-set block erase, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Multi-set block erase will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3019568