USRE37199E1 - Method of making nonvolatile semiconductor memory - Google Patents

Method of making nonvolatile semiconductor memory Download PDF

Info

Publication number
USRE37199E1
USRE37199E1 US09/340,682 US34068299A USRE37199E US RE37199 E1 USRE37199 E1 US RE37199E1 US 34068299 A US34068299 A US 34068299A US RE37199 E USRE37199 E US RE37199E
Authority
US
United States
Prior art keywords
insulating film
forming
diffusion layer
semiconductor substrate
tunnel
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Lifetime
Application number
US09/340,682
Other languages
English (en)
Inventor
Yoshimitsu Yamauchi
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sharp Corp
Original Assignee
Sharp Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from JP8164343A external-priority patent/JPH0992739A/ja
Priority to EP96304829A priority Critical patent/EP0752721B1/fr
Priority claimed from US08/672,666 external-priority patent/US5877054A/en
Application filed by Sharp Corp filed Critical Sharp Corp
Priority to US09/340,682 priority patent/USRE37199E1/en
Application granted granted Critical
Publication of USRE37199E1 publication Critical patent/USRE37199E1/en
Assigned to SHARP KABUSHIKI KAISHA reassignment SHARP KABUSHIKI KAISHA ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: YAMAUCHI, YOSHIMOTSI
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H10—SEMICONDUCTOR DEVICES; ELECTRIC SOLID-STATE DEVICES NOT OTHERWISE PROVIDED FOR
    • H10B—ELECTRONIC MEMORY DEVICES
    • H10B41/00—Electrically erasable-and-programmable ROM [EEPROM] devices comprising floating gates
    • H10B41/30—Electrically erasable-and-programmable ROM [EEPROM] devices comprising floating gates characterised by the memory core region
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C16/00—Erasable programmable read-only memories
    • G11C16/02—Erasable programmable read-only memories electrically programmable
    • G11C16/04—Erasable programmable read-only memories electrically programmable using variable threshold transistors, e.g. FAMOS
    • G11C16/0408—Erasable programmable read-only memories electrically programmable using variable threshold transistors, e.g. FAMOS comprising cells containing floating gate transistors
    • G11C16/0425—Erasable programmable read-only memories electrically programmable using variable threshold transistors, e.g. FAMOS comprising cells containing floating gate transistors comprising cells containing a merged floating gate and select transistor
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C16/00—Erasable programmable read-only memories
    • G11C16/02—Erasable programmable read-only memories electrically programmable
    • G11C16/04—Erasable programmable read-only memories electrically programmable using variable threshold transistors, e.g. FAMOS
    • G11C16/0491—Virtual ground arrays
    • H—ELECTRICITY
    • H10—SEMICONDUCTOR DEVICES; ELECTRIC SOLID-STATE DEVICES NOT OTHERWISE PROVIDED FOR
    • H10B—ELECTRONIC MEMORY DEVICES
    • H10B69/00—Erasable-and-programmable ROM [EPROM] devices not provided for in groups H10B41/00 - H10B63/00, e.g. ultraviolet erasable-and-programmable ROM [UVEPROM] devices

Definitions

  • the present invention relates to a nonvolatile semiconductor memory using a Fowler-Nordheim (FN) tunnel current for writing and erasing, and a driving method and fabrication method for such a nonvolatile semiconductor memory. More specifically, the present invention relates to a flash memory, and a driving method and fabrication method for the flash memory.
  • FN Fowler-Nordheim
  • FIG. 47 is a sectional view of memory cells of a conventional flash memory 500 of the type using hot electron injection.
  • FIG. 48 is an equivalent circuit of a memory cell array of the flash memory 500 .
  • the flash memory 500 includes a semiconductor substrate 151 , a tunnel insulating film 153 formed on the semiconductor substrate 151 , and floating gates 155 formed on the tunnel insulating film 153 .
  • An insulating film 156 made of ONO(SiO 2 /SiN/SiO 2 ) and the like is formed covering the floating gates 155 .
  • Control gates 157 are formed on the insulating film 156 .
  • memory cells memory cells (memory cells C 51 to C 53 and C 61 to C 63 are shown in FIG. 48) are arranged in a matrix.
  • the control gates 157 of the memory cells lined in an X direction shown in FIG. 47 are electrically connected to one another, forming a word line WL (word lines WL 1 and WL 2 are shown in FIG. 48 ).
  • an impurity diffusion layer (a source/drain diffusion layer) 161 is formed between every two memory cells adjacent in the X direction, and shared by the two memory cells as a source diffusion layer 159 for one memory cell and a drain diffusion layer 160 for the other memory cell.
  • Such source/drain diffusion layers are formed in a self-aligning manner by ion implantation using a film as a mask during the fabrication process. Upon completion of the fabrication process, the film will become the floating gates 155 formed in a stripe shape.
  • bit lines BL (bit lines BL 1 to BL 4 are shown in FIG. 48) extend in the Y direction, electrically connecting the diffusion layers 161 lined in the Y direction with one another.
  • Each bit line serves as a source wiring or a drain wiring depending on the selected memory cell.
  • Such a driving method where each bit line is not fixed as the source wiring or the drain wiring, but where the source wiring (ground line) and the drain wiring are appropriately switched is called a virtual ground method.
  • the impurity diffusion layer 161 is used both as the source diffusion layer 159 for one memory cell and the drain diffusion layer 160 for a memory cell adjacent in the X direction as described above, no isolation region is required between the source diffusion layer 159 for one memory cell and the drain diffusion layer 160 for the-adjacent memory cell in the X direction. Furthermore, since the bit lines BL are formed by connecting the impurity diffusion layers 161 lined in the Y direction via diffusion wirings, no contact regions are required for the connection of the bit lines BL with the memory cells. This makes it possible to realize high integration of memories.
  • Japanese Laid-Open Patent Publication No. 2-231772 discloses the configuration shown in FIG. 51, where two memory cells adjacent in the X direction are paired, sharing a source line SL but having individual bit lines BL. This configuration allows for parallel reading and parallel writing of data.
  • the write operation of a flash memory 500 is conducted in the following manner.
  • the memory cell C 52 is selected as a memory cell in which data is to be written (hereinafter, such a selected memory cell is referred to as a selected cell).
  • a high voltage is applied to the word line WL 1 connected with the selected cell C 52 .
  • a predetermined voltage is applied to one of the bit lines connected with the selected cell C 52 (e.g., the bit line BL 3 ) to define the drain side, while a ground voltage (0 V) is applied to the other bit line (e.g., the bit line BL 2 ) to define the source side.
  • non-selected cells certain voltages are applied to two bit lines connected with any one of the non-selected cells so that the two bit lines have the same potential. For example, for the non-selected cell C 51 , voltages are applied to the bit lines BL 3 and BL 4 so that the potentials of these bit lines are the same.
  • the erase operation of the flash memory 500 is conducted in the following manner.
  • a negative voltage is applied to the word lines WL, and simultaneously, a predetermined positive voltage is applied to all the bit lines BL (or all bit lines in a block when the memory cell array is divided into blocks).
  • This causes an FN tunnel current to flow, drawing out charges accumulated in the floating gates 155 and thus erasing data stored in all the memory cells (or all memory cells in the block) at one time.
  • the read operation of the flash memory 500 is conducted in the following manner.
  • a predetermined voltage is applied to a word line WL connected with a selected cell for reading.
  • a predetermined voltage is applied to one of two bit lines connected with the selected cell, while a ground voltage (0 V) is applied to the other bit line.
  • the amount of current flowing between the two bit lines is different depending on the amount of charge (i.e., data) stored in the floating gate 155 .
  • the data is thus read by detecting the amount of current.
  • voltages are applied to two bit lines connected with a non-selected cell where no reading is conducted so that the potential of the two bit lines is the same as in the write operation.
  • the write operation using the channel hot electron injection as described above has the following drawbacks.
  • the efficiency of electron injection is generally poor. Since the write current is large (about 1 mA), power consumption at writing is large.
  • a high-voltage power source (or a booster) is required to supply the comparatively large write current, which prevents the flash memory 500 from lowering the driving voltage and using a single power source.
  • FIG. 49 is a sectional view of memory cells of a conventional flash memory 600 of the type using an FN tunnel current for writing.
  • FIG. 50 is an equivalent circuit of a memory cell array of flash memory 600 . Similar components to those of flash memory 500 shown in FIGS. 46 and 47 are denoted by the same reference numerals.
  • the flash memory 600 includes a semiconductor substrate 151 , a tunnel insulating film 153 formed of a uniform oxide film on the semiconductor substrate 151 , and floating gates 155 formed on the tunnel insulating film 153 .
  • a source diffusion layer 159 and a drain diffusion layer 160 are formed on both ends of each of the floating gates 155 .
  • An element isolation film 162 is formed between every two adjacent memory cells, isolating the source diffusion layer 159 of one memory cell from the drain diffusion layer 160 of the other memory cell.
  • An insulating film 156 made of ONO (SiO 2 /SiN/SiO 2 ) is formed covering the floating gates 155 .
  • Control gates 157 are formed on the insulating film 156 .
  • memory cells (memory cells C 71 to C 73 and C 81 to C 83 are shown in FIG. 50) are arranged in a matrix.
  • the control gates 157 of the memory cells lined in the X direction shown in FIG. 50 are electrically connected with one another, forming a word line WL (word lines WL 1 and WL 2 are shown in FIG. 50 ).
  • Bit lines BL (bit lines BL 1 to BL 6 are shown in FIG. 50) extend in the Y direction, electrically connecting the corresponding source diffusion layers 159 or the drain diffusion layers 160 with one another.
  • the function of each bit line BL is fixed as the source line or the drain line.
  • the write operation of the flash memory 600 is conducted in the following manner.
  • a negative voltage or 0 V is applied to the word line WL connected with a selected cell, while a positive voltage is applied to the drain diffusion layer 160 of the connected cell so that electrons are drawn out of the floating gate 155 .
  • the erase operation is conducted in the following manner.
  • a positive high voltage is applied to one word line WL, while 0 V is applied to the source diffusion layers 159 and the drain diffusion layers 160 so that electrons are injected into the floating gates of all the memory cells connected with the word line WL.
  • FIG. 51 shows another example of the flash memory 700 using an FN tunnel current.
  • the drain diffusion layer and the source diffusion layer cell of the adjacent memory cells are isolated from each other.
  • the drain diffusion layers of respective memory cells are separated from one another and these individual drain diffusion layers lined in the Y direction are connected with one another to form a bit line.
  • each pair of memory cells adjacent in the X direction share the source diffusion layer to form a source line extending in the Y direction.
  • the conventional flash memory of the type using an FN tunnel current has the following drawbacks.
  • a tunnel current generated by the electric field applied to the tunnel insulating film 153 is used. Accordingly, if in the memory cells shown in FIG. 49, each pair of memory cells adjacent in the X direction share the diffusion layer as shown in FIG. 48, data will be written in a non-selected memory cell which shares the drain diffusion layer with a selected memory cell in data is to be written and is connected with the same word line as the selected memory cell. Accordingly, it is necessary to form the element isolation film 162 as shown in FIG. 49 and isolate at least the drain diffusion layers 160 of the adjacent memory cells in the X direction from each other as shown in FIGS. 50 and 51 . Thus, the conventional flash memory of the type using an FN tunnel current for writing cannot employ the virtual ground method. This prevents the flash memory of this type from being made smaller.
  • the tunnel insulating film 153 is comparatively thin. Accordingly, when a comparatively high voltage is applied to the drain diffusion layer 160 at writing, a current 10 6 times the write current may sometimes flow from the end of the drain diffusion layer 160 to the substrate 151 due to the interband tunnel phenomenon.
  • the tunnel insulating film 153 becomes thinner, the capacitance between the floating gate and the substrate is larger. This makes it difficult to increase the capacitance coupling ratio of the control gate. Also, as the tunnel insulating film 153 becomes thinner, the disturb margin against a high voltage to be applied to the control gate electrode at writing becomes smaller.
  • the nonvolatile semiconductor memory of this invention includes: a semiconductor substrate; a plurality of memory cells formed in a matrix on the semiconductor substrate, each of the memory cells including a first insulating film formed on the semiconductor substrate, a floating gate formed on the first insulating film, and a control gate formed on the floating gate via a second insulating film sandwiched therebetween, a source diffusion region, and a drain diffusion region; a diffusion layer formed in a portion of the semiconductor substrate located between two of the memory cells adjacent in a first direction, the diffusion layer including the drain diffusion region for one of the two memory cells and the source diffusion region for the other memory cell; a word line formed by connecting the control gates of the memory cells lined in the first direction; and a bit line formed by connecting the diffusion layers lined in a second direction substantially perpendicular to the first direction, wherein the memory cells have a structure in which a tunnel current flows between the drain diffusion region and the floating gate of one of the two adjacent memory cells via the first insulating film when a predetermined voltage is applied to the diffusion
  • the source diffusion region and the drain diffusion region of each of the memory cells have impurity densities different from each other.
  • the drain diffusion region has a double-layered diffusion layer structure where the impurity density is lower in an outer layer.
  • the first insulating film includes a first portion which does not constitute a tunnel region and a second portion which constitutes a tunnel region, and the drain diffusion region of each memory cell capacitively couples with the floating gate via the second portion, and the source diffusion region of each memory cell capacitively couples with the floating gate via the first portion.
  • the thickness of the first portion of the first insulating film which does not constitute a tunnel region is larger than that of the second portion which constitutes a tunnel region.
  • a method for fabricating a nonvolatile semiconductor memory includes the steps of: forming a tunnel insulating film on a semiconductor substrate of a first conductivity type; forming a first conductive layer pattern with a patterned portion having a predetermined width by patterning a first conductive layer formed on the semiconductor substrate; forming a first diffusion layer in the semiconductor substrate by implanting impurities of a second conductivity type using the first conductive layer pattern as a mask; forming a second diffusion layer in the semiconductor substrate in contact with the first diffusion layer by implanting impurities of the second conductivity type under implantation conditions different from those for forming the first diffusion layer, the impurity density of the second diffusion layer being higher than that of the first diffusion layer; forming an insulating film and a second conductive layer on the first conductive layer pattern in this order; and patterning the second conductive layer, the insulating film, and the first conductive layer, to form control gates, insulating layers, and floating gates, respectively.
  • the step of forming the second diffusion layer is performed using a mask having a second width, and the impurities of the second conductivity type are implanted in a predetermined region of the first diffusion layer.
  • the step of forming the first diffusion layer includes implanting the impurities of the second conductivity type at a predetermined angle directed to one side of an opening of the first conductive layer pattern used as the mask, and the step of forming the second diffusion layer includes implanting the impurities of the second conductivity type at a predetermined angle directed to the other side of the opening of the first conductive layer pattern used as the mask.
  • the method for fabricating a nonvolatile semiconductor memory of this invention includes: the steps of: forming a first insulating film on a semiconductor substrate of a first conductivity type; forming a drain diffusion layer by implanting impurities of a second conductivity type in the semiconductor substrate using a first resist pattern having a first opening as a mask, forming a gate insulating film using a second resist pattern having a second opening corresponding to a tunnel region as a mask and removing a portion of the first insulating film located below the second opening; forming a tunnel insulating film in the tunnel region where the first insulating film has been removed by forming a second insulating film on the semiconductor substrate after removal of the second resist pattern; forming a first conductive layer having a predetermined width by patterning to cover the tunnel region where the tunnel insulating film has been formed; forming a source diffusion layer in contact with the drain diffusion layer by implanting impurities of the second conductivity type in the semiconductor substrate using the first conductive layer as a mask
  • the method for fabricating a nonvolatile semiconductor memory of this invention includes the steps of: forming a first insulating film on a semiconductor substrate of a first conductive type; forming a drain diffusion layer by implanting impurities of a second conductivity type in the semiconductor substrate using a resist pattern having an opening corresponding to a tunnel region as a mask; removing a portion of the first insulating film using the resist pattern as a mask; forming a tunnel insulating film in the tunnel region where the first insulating film has been removed by forming a second insulating film on the semiconductor substrate after removal of the resist pattern; forming a first conductive layer having a predetermined width by patterning to cover the tunnel region where the tunnel insulating film has been formed; forming a source diffusion layer in contact with the drain diffusion layer by implanting impurities of the second conductivity type in the semiconductor substrate using the first conductive layer as a mask; forming a third insulating film; and forming a second conductive layer on the third insulating film
  • the method for fabricating a nonvolatile semiconductor memory of this invention includes the steps of: forming a first insulating film on a semiconductor substrate of a first conductivity type; forming a first nitride film having a first width on the first insulating film by patterning; covering the first nitride film with an oxide film and then forming spacers on both side walls of the first nitride film, the spacer on one of the side walls corresponding to a region defining a tunnel region; forming a diffusion layer by implanting impurities of a second conductivity type using the first nitride film and the spacers on the side walls as a mask; removing the spacer on the other side wall of the first nitride film; forming a second insulating film having a thickness larger than that of the first insulating film selectively using the first nitride film and the spacer as a mask; removing the spacer on the one of the both side walls of the first nitride
  • writing and erasing are conducted using a tunnel current
  • the memory including: a plurality of memory cells formed in a matrix on a semiconductor substrate and divided into a plurality of blocks; a word line formed by connecting control gates of the memory cells lined in a first direction; a first bit line formed by connecting diffusion layers formed between the memory cells adjacent in the first direction in a second direction perpendicular to the first direction; a second bit line provided to correspond to a predetermined number of the first bit lines; and a selective transistor provided to correspond to each of the first bit lines for connecting the first bit line to the corresponding second bit line, wherein the selective transistor electrically connects the first bit lines in a selected block to the corresponding second bit lines and puts the first bit lines in blocks other than the selected block in a floating state, in order to conduct an erase operation for every block.
  • the plurality of blocks are formed within one well region of the semiconductor substrate.
  • a method for driving a nonvolatile semiconductor memory includes: a plurality of memory cells formed in a matrix on a semiconductor substrate and divided into a plurality of blocks; a word line formed by connecting control gates of the memory cells lined in a first direction; a first bit line formed by connecting diffusion layers formed between the memory cells adjacent in the first direction in a second direction perpendicular to the first direction; a second bit line provided to correspond to a predetermined number of the first bit lines; and a selective transistor provided to correspond to each of the first bit lines for connecting the first bit line to the corresponding second bit line.
  • the method includes the steps of: electrically connecting the first bit lines in a selected block to the corresponding second bit lines, while putting the-first bit lines in blocks other than the selected block in a floating state by controlling the selective transistors; applying a predetermined negative voltage to the first bit lines and the semiconductor substrate; applying a predetermined positive voltage to the word lines in the selected block, while applying a ground voltage to the word lines in the blocks other than the selected block; and consequently erasing data in the memory cells in the selected block using a tunnel current.
  • the memory includes: a plurality of memory cells formed in a matrix on a semiconductor substrate; a word line formed by connecting control gates of the memory cells lined in a first direction; a first bit line formed by connecting diffusion layers formed between the memory cells adjacent in the first direction in a second direction perpendicular to the first direction; a second bit line provided to correspond to a predetermined number (N) of first bit lines; a selective transistor provided to correspond to each of the first bit lines for connecting the first bit line to the corresponding second bit line; and a signal line for controlling the N selective transistors separately.
  • the method includes the steps of: (a) selecting a word line; (b) supplying a predetermined potential to the second bit line connected with a drain of the memory cell in which data is to be written; (c) selecting one of the N first bit lines to be connected to the second bit line by controlling one of the signal lines to allow the corresponding selective transistor to be turned on, and writing data on the selected second bit line in the memory cell connected with the selected word line and the first bit line connected to the second bit line; and (d) repeating step (c) to write data in all the N memory cells connected with the selected word line and the N first bit lines corresponding to the selected second bit.
  • the memory includes: a plurality of memory cells formed in a matrix on a semiconductor substrate and divided into a plurality of blocks; a word line formed by connecting control gates of the memory cells lined in a first direction; a first bit line formed by connecting diffusion layers formed between the memory cells adjacent in the first direction in a second direction perpendicular to the first direction; a second bit line provided to correspond to a predetermined number of the first bit lines; and a selective transistor provided to correspond to each of the first bit lines for connecting the first bit line to the corresponding second bit line selectively.
  • the method includes the steps of: applying a predetermined voltage to the word line connected with the memory cell from which data is to be read; applying a predetermined voltage to one of the first bit lines connected with a source diffusion layer of the memory cell from which data is to be read, and applying the same predetermined voltage to the other first bit lines existing in parallel with the one of the first bit lines via the one of the first bit lines under a floating state.
  • a method for driving a nonvolatile semiconductor memory is provided.
  • the memory cells of the nonvolatile semiconductor device have a structure in which a tunnel current flows between the drain diffusion region and the floating gate of one of the two adjacent memory cells via the first insulating film when a predetermined voltage is applied to the diffusion layer and no tunnel current flows between the diffusion layer and the floating gate of the other memory cell.
  • the method comprises a step of applying a predetermined voltage to the memory cell from the source diffusion region so as to read data stored in the memory cell.
  • an asymmetric memory cell where the coupling capacitance of the floating gate is asymmetric between the source diffusion region side and the drain diffusion region side at writing is used. Accordingly, when voltages are applied to a word line and bit lines (drain diffusion layers) connected with a selected memory cell (selected cell), no data is written in a non-selected cell of which the source diffusion region is connected with the selected bit line.
  • Such an asymmetric memory cell can be realized by having the impurity density of the drain diffusion layer higher than that of the source diffusion layer. Since the floating gate is capacitively coupled with the high-density drain diffusion layer, a tunnel current easily flows between the floating gate and the drain diffusion layer via the tunnel insulating film.
  • the impurity density of the source diffusion layer capacitively coupled with the floating gate is low even if the same voltage as that applied to the drain diffusion layer is applied to the source diffusion layer.
  • the surface of this low-density source diffusion layer is thus depleted, lowering the electric field applied to the tunnel insulating film and thus preventing the tunnel phenomenon from occurring.
  • the above asymmetric memory cell can also be realized by forming a tunnel insulating film only between the drain diffusion layer and the floating gate.
  • a tunnel insulating film only between the drain diffusion layer and the floating gate.
  • the invention described herein makes possible the advantages of (1) providing a nonvolatile semiconductor memory using an FN tunnel current for writing and erasing and employing the virtual ground method for driving a memory cell array thereof, (2) providing a driving method of such a nonvolatile semiconductor memory, (3) providing a fabrication method of such a nonvolatile semiconductor memory, (4) providing a nonvolatile semiconductor memory with high write efficiency and reliability where a current flowing into a semiconductor substrate at writing is reduced, (5) providing a driving method of such a nonvolatile semiconductor memory, and (6) providing a fabrication method of such a nonvolatile semiconductor memory.
  • FIG. 1 is a plan view of a portion of a memory cell array of a nonvolatile semiconductor memory of Example 1 according to the present invention.
  • FIGS. 2A and 2B are sectional views of the memory cell array shown in FIG. 1 taken along lines A—A and B—B of FIG. 1, respectively.
  • FIG. 3 is an equivalent circuit diagram of the memory cell array of the nonvolatile semiconductor memory of Example 1.
  • FIG. 4 shows the relationship between the impurity dose amount of a source diffusion layer and the influence of the write operation to a non-selected cell.
  • FIG. 5 is a plan view illustrating a portion of a memory cell array of a nonvolatile semiconductor memory of Example 1 according to the present invention.
  • FIG. 6 is a sectional view illustrating a step of a fabrication method of memory cells of the nonvolatile semiconductor memory of Example 2 according to the present invention.
  • FIG. 7 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 2 according to the present invention.
  • FIG. 8 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 2 according to the present invention.
  • FIG. 9 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 2 according to the present invention.
  • FIG. 10 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 2 according to the present invention.
  • FIGS. 11A, 11 B, 11 C, and 11 D are sectional views illustrating steps of an alternative fabrication method of memory cells of the nonvolatile semiconductor memory of Example 2 according to the present invention.
  • FIGS. 12A and 12B are sectional views illustrating a tilt ion injection process of another alternative fabrication method of memory cells of the nonvolatile semiconductor memory of Example 2 according to the present invention.
  • FIG. 13A is a plan view of a portion of a memory cell array of an alternative nonvolatile semiconductor memory according to the present invention, where element isolation is conducted by a p-n joint
  • FIG. 13B is a sectional view of the memory cells shown in FIG. 13A taken along line D—D of FIG. 13 A.
  • FIG. 14 is a plan view of a portion of a memory cell array of a nonvolatile semiconductor memory of Example 3 according to the present invention.
  • FIGS. 15A and 15B are sectional views of the memory cell array shown in FIG. 14 taken along lines A—A and B—B of FIG. 14, respectively.
  • FIG. 16 is an equivalent circuit diagram of the memory cell array of the nonvolatile semiconductor memory of Example 3 according to the present invention.
  • FIG. 17 shows the relationship between the thickness of a gate insulating film and the write efficiency (leak current/tunnel current).
  • FIG. 18 is a plan view of a portion of a memory cell array of an alternative nonvolatile semiconductor memory according to the present invention.
  • FIG. 19 is a sectional view illustrating a step of a fabrication method of memory cells of a nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 20 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 21 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 22 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 23 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 24 is a sectional view illustrating a step of an alternative fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 25 is a sectional view illustrating a step of the alternative fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 26 is a sectional view illustrating a step of the alternative fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 27 is a sectional view illustrating a step of the alternative fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 28 is a sectional view illustrating a step of the alternative fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 4 according to the present invention.
  • FIG. 29A is a plan view of a portion of a memory cell array of an alternative nonvolatile semiconductor memory according to the present invention, where element isolation is conducted by a p-n joint
  • FIG. 29B is a sectional view of the memory cells shown in FIG. 29A taken along line D—D of FIG. 29 A.
  • FIG. 30 is a plan view of a portion of a memory cell array of a nonvolatile semiconductor memory of Example 5 according to the present invention.
  • FIGS. 31A and 31B are sectional views of the memory cell array illustrated in FIG. 30 taken along lines A—A and B—B of FIG. 30, respectively.
  • FIG. 32 is a sectional view illustrating a step of a fabrication method of memory cells of the nonvolatile semiconductor memory of Example 5 according to the present invention.
  • FIG. 33 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 5 according to the present invention.
  • FIG. 34 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 5 according to the present invention.
  • FIG. 35 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 5 according to the present invention.
  • FIG. 36 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 5 according to the present invention.
  • FIG. 37 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 5 according to the present invention.
  • FIG. 38 is a sectional view illustrating a step of the fabrication method of the memory cells of the nonvolatile semiconductor memory of Example 5 according to the present invention.
  • FIG. 39 is a plan view of a portion of a memory cell array of a nonvolatile semiconductor memory of Example 6 according to the present invention.
  • FIG. 40 shows a state of a memory cell in the read operation.
  • FIG. 41 shows the relationship between the read current (channel current) and the leak current in the read operation.
  • FIG. 42 is a timing chart schematically showing the applied voltages in the erase operation.
  • FIG. 43 shows the relationship between the voltage application time to a control gate and the threshold voltage of a memory cell when a negative voltage is applied to a source/drain diffusion layer and a semiconductor substrate.
  • FIG. 44 shows the relationship between the negative voltage applied to the semiconductor substrate and the life of the memory cell (reliability).
  • FIG. 45 is a timing chart schematically showing the applied voltages in the write operation.
  • FIG. 46 ia a plan view of an alternative example of the memory cell array of the nonvolatile semiconductor memory according to the present invention.
  • FIG. 47 is a sectional view of a portion of a memory cell array of a conventional nonvolatile semiconductor memory of the type of using hot electron injection for writing.
  • FIG. 48 is an equivalent circuit diagram of the memory cell array of the nonvolatile semiconductor memory illustrated in FIG. 47 .
  • FIG. 49 is a sectional view of a portion of a memory cell array of a conventional nonvolatile semiconductor memory of the type of using an FN tunnel current for writing.
  • FIG. 50 is an equivalent circuit diagram of the memory cell array of the nonvolatile semiconductor memory illustrated in FIG. 49 .
  • FIG. 51 is an equivalent circuit diagram of a portion of a memory cell array of another conventional semiconductor memory of the type of using an FN tunnel current for writing.
  • FIG. 1 is a plan view of a portion of a memory cell array of a nonvolatile semiconductor memory 100 of Example 1 according to the present invention.
  • FIGS. 2A and 2B are sectional views of the nonvolatile semiconductor memory 100 taken along lines A—A and B—B of FIG. 1, respectively,
  • the nonvolatile semiconductor memory 100 includes a plurality of memory cells C formed in a matrix on a semiconductor substrate 1 .
  • the respective memory cells C are denoted by individual codes such as C im .
  • Each of the memory cells C includes a tunnel insulating film 3 formed on the semiconductor substrate 1 and a floating gate 5 formed on the tunnel insulating film 3 .
  • a control gate 7 is formed on the floating gate 5 via an insulating film 6 made of ONO and the like. Such control gates 7 extend in a channel direction of the memory cell C (an X direction shown in FIG. 1 ), constituting a word line WL as shown in FIG. 3 for connecting the memory cells C lined in the X direction.
  • a diffusion layer 2 is formed-between every two memory cells adjacent each other in the X direction (e.g., memory cells C im and c in ).
  • the diffusion layer 2 includes a drain diffusion layer 2 a for one of the two adjacent memory cells (e.g., the memory cell C im ) and a source diffusion layer 2 b for the other memory cell (e.g., the memory cell C in ).
  • the diffusion layer 2 is thus shared by the two adjacent memory cells.
  • the impurity density of the drain diffusion layer 2 a (N + ) is made higher than that of the source diffusion layer 2 b (N ⁇ ).
  • the diffusion layer 2 has a double-layered structure where the drain diffusion layer 2 a is formed inside the diffusion region with a low impurity density constituting the source diffusion layer 2 b.
  • the floating gate 5 is capacitively coupled with the drain diffusion layer 2 a of one of the two diffusion layers 2 via the tunnel insulating film 3 (defining the drain side), while it is capacitively coupled with the source diffusion layer 2 b of the other diffusion layer 2 via the tunnel insulating film 3 (defining the source side).
  • the tunnel insulating film 3 may be composed of a silicon oxide film or a lamination of a silicon oxide film and a silicon nitride film.
  • the diffusion layers 2 lined in the Y direction are connected with one another via diffusion layer wirings, forming a so-called buried bit line.
  • the diffusion layers 2 may be electrically connected with one another with metal wirings. In the latter case, a contact region is required for each diffusion layer 2 .
  • a field oxide film (element isolation film) 12 is formed between two memory cells adjacent in the Y direction (e.g., the memory cells C in and C jn ).
  • the nonvolatile semiconductor memory 100 employs the virtual ground method, where bit lines BL (bit lines BL 1 to BL 4 are shown in FIG. 3) serve as source wirings or drain wirings depending on the selected cell.
  • Table 1 below shows the operation conditions when the memory cell C 12 shown in FIG. 3 is selected, for example.
  • a negative high voltage V H1 (e.g., ⁇ 8 V) is applied to the word line WL 1 connected with the control gate 7 of the memory cell C 12 , while 0 V is applied to the other word lines.
  • a predetermined positive voltage V cc (e.g., 4 V) is applied to the bit line BL 2 connected with the drain diffusion layer 2 a for the memory cell C 12 , while the other bit lines are put in a floating state.
  • V H1 e.g., ⁇ 8 V
  • V cc e.g., 4 V
  • the same write voltage is also applied to the control gate 7 of a non-selected cell which is connected with the selected word line WL and the source diffusion layer 2 b of the bit line BL of which source is selected, for example, the non-selected cell C 11 which is connected with the word line W 1 and the bit line BL 2 .
  • the impurity density of the source diffusion layer 2 b is low, no tunnel current flows between the source diffusion layer 2 b and the floating gate 5 of the non-selected cell C 11 . Accordingly, even though the diffusion layer 2 is shared with the selected cell, no data is written in the n on-selected cell.
  • 0 V is applied to all the bit lines BL.
  • a positive high voltage V H2 (e.g., 17 V) is applied to a desired word line WL.
  • V H2 e.g. 17 V
  • a high voltage V H2 is applied to the word line WL 1
  • date in the memory cells C 11 , C 12 , and C 13 are erased at one time.
  • a high voltage V H2 is applied to the word line WL 2
  • date in the memory cells C 21 , C 22 , and C 23 are erased at one time.
  • the data reading from the selected cell C 12 is conducted in a conventional manner.
  • a predetermined voltage V cc (e.g., 3 V) is applied to the word line WL 1 .
  • a predetermined voltage V L e.g., 1 V
  • 0 V is applied to the bit line BL 3 .
  • FIG. 4 shows a change in the cell voltage (threshold) of the non-selected cell C 11 when the impurity dose amount of the source diffusion layer 2 b is changed.
  • a curve 4 a represents the change in the threshold voltage when data is written in a memory cell of which drain diffusion layer 2 a has a impurity dose amount of 5 ⁇ 10 15 /cm 2
  • curves 4 b to 4 d represent changes in the threshold voltage when the same voltage as that applied to the drain diffusion layer in the write operation is applied to the source diffusion layer of memory cells of which source diffusion layer 2b has a impurity dose amount of 1 ⁇ 10 14 /cm 2 , 5 ⁇ 10 13 /cm 2 , and 1 ⁇ 10 13 /cm 2 , respectively.
  • the threshold voltage of the selected cell C 12 falls below 0 V.
  • the threshold voltage of the non-selected cell C 11 of which source diffusion layer is in contact with the drain diffusion layer of the selected cell C 12 hardly changes when the impurity dose amount of the source diffusion layer 2 b is 5 ⁇ 10 13 /cm 2 or less.
  • the threshold voltage of the non-selected cell C 11 is 4 V even when a high voltage is applied to the selected cell C 12 for 100 ms.
  • the threshold voltage of the non-selected cell C 11 hardly changes when a high voltage is applied to the selected cell C 12 for 1000 ms.
  • the nonselected cell adjacent to the selected cell is not influenced by the write operation for the selected cell even though the diffusion layer 2 is shared with the selected cell. This makes it possible to write data only in the selected cell.
  • FIG. 5 shows an example of an equivalent circuit of a nonvolatile semiconductor memory cell array 100 ′ where, in order to reduce the wiring resistance of the bit line BL composed of the diffusion layers shown in FIG. 3, a sub-bit line SBL is connected to a main bit line MBL made of metal via a selective transistor.
  • FIGS. 6 to 10 a first example of the fabrication method of a nonvolatile semiconductor memory according to the present invention will be described.
  • the fabrication method of the nonvolatile semiconductor memory 100 of Example 1 will be described.
  • FIGS. 6 to 10 are sectional views taken along line A—A of FIG. 1, showing steps of the fabrication method.
  • the field insulating films (element isolation films) 12 shown in FIG. 2B are formed on the semiconductor substrate (silicon substrate) 1 .
  • the tunnel insulating film 3 with a thickness of about 80 ⁇ is formed by thermal oxidation, and first polysilicon is deposited on the tunnel insulating film 3 to a thickness of about 1000 to 2000 ⁇ .
  • a first resist mask 8 a with a predetermined pattern is formed on the first polysilicon by photolithography, to form first polysilicon layers 5 ′ in a striped shape by patterning.
  • Each of the first polysilicon layers 5 ′ has the same width as the width in the channel direction of the floating gate 5 to be formed in a later step (FIG. 6 ).
  • phosphorous ions are implanted with an energy of 50 KeV and a dose amount of 3 ⁇ 10 13 /cm 2 .
  • the regions where the phosphorous ions are implanted will become the low-density n-type impurity diffusion layers (source diffusion layers) 2 b in a later step (FIG. 7 ).
  • a second resist mask 8 b with a predetermined pattern having openings corresponding to regions which will become the drain diffusion layers in a later step is formed.
  • arsenic ions are implanted with an energy of 70 KeV and a dose amount of 1 ⁇ 10 15 /cm 2 .
  • the regions where the arsenic ions are implanted will become the high-density N-type impurity diffusion layers (drain diffusion layers) 2 a in a later step (FIG. 8 ).
  • the resultant structure is heat-treated under a nitrogen atmosphere at 900° C. for 10 minutes, so that a DDD structure is formed in the drain diffusion layers 2 a, while an LDD structure is formed in the source diffusion layers 2 b.
  • the high-density impurity diffusion layer 2 a and the low-density impurity diffusion layer 2 b constitute one continuous diffusion layer 2
  • a series of the diffusion layers 2 in the Y direction constitute a one bit line.
  • An oxide film is formed by CVD and etched back to form oxide films 4 between the first polysilicon layers 5 ′.
  • the ONO film (oxide film/nitride film/oxide film) 6 is formed on the resultant structure (FIG. 9 ).
  • a second polysilicon film 7 ′ with a thickness of about 1000 ⁇ is formed on the ONO film 6 .
  • the control gates 7 , the insulating films 6 , and the floating gates 5 are formed (FIG. 10 ).
  • the control gates 7 of respective memory cells are formed consecutively in the X direction, forming one word line.
  • control gates 37 i.e., the word line WL
  • the control gates 37 may be of a double-layered structure where metal silicide with a high melting point such as tungsten silicide (WSi) is deposited to a thickness of about 1000 ⁇ on the second polysilicon layer with a thickness of about 1000 ⁇ and these layers are patterned.
  • metal silicide with a high melting point such as tungsten silicide (WSi)
  • FIGS. 11A to 11 D an alternative fabrication method of the nonvolatile semiconductor memory according to the present invention will be described with reference to FIGS. 11A to 11 D.
  • the first and second resist masks 8 a and 8 b are used.
  • the spaces between the adjacent floating gates 5 are made smaller. This makes it difficult to form the second resist mask 8 b.
  • both phosphorous and arsenic ions are implanted using the first resist mask 8 a.
  • the step until the first resist mask 8 a is formed on the semiconductor substrate 1 is the same as that in the above method (FIG. 6 ). Then, as shown in FIG. 11A, phosphorous ions are implanted to form the lowdensity impurity diffusion layer. At this time, by implanting phosphorous ions at a predetermined tilt angle (e.g., about 7° with respect to the normal of the semiconductor substrate 1 ), the phosphorous ions are implanted in the portions of the semiconductor substrate along one side (source side) of each of the openings of the first resist mask 8 a (corresponding to the bit lines). This implantation does not form the DDD structure, but is suitable for miniaturization.
  • a predetermined tilt angle e.g., about 7° with respect to the normal of the semiconductor substrate 1
  • Phosphorous ions may be implanted on only the source side as described above, or, as in the case shown in FIG. 7, may be implanted so that the DDD structure is formed.
  • arsenic ions are implanted at a predetermined tilt angle (e.g., about 7° inversely with respect to the normal of the semiconductor substrate 1 ).
  • a predetermined tilt angle e.g., about 7° inversely with respect to the normal of the semiconductor substrate 1 .
  • arsenic ions are implanted in the portions of the semiconductor substrate along the other side (drain side) of each of the openings of the first resist mask 8 a.
  • the first resist mask 8 a is then removed and the resultant structure is heat-treated, forming the diffusion layers 2 each including the highdensity N-type diffusion layer 2 a on the drain side and the low-density N-type diffusion layer 2 b on the source side consecutively (FIG. 11 C).
  • the ONO films 6 and the control gates 7 are then formed in a manner similar to the steps shown in FIGS. 9 and 10 (FIG. 11 D).
  • the step of forming the second resist mask 8 b can be omitted and an allowance for alignment is not required.
  • the cell size can be further reduced.
  • phosphorous and arsenic ions may be implanted after the removal of the first resist mask 8 a as shown in FIGS. 12A and 12B.
  • the tilt angle for ion implantation should be larger than the case using the resist mask 8 a.
  • FIGS. 13A and 13B show an alternative nonvolatile semiconductor memory according to the present invention.
  • a nonvolatile semiconductor memory 110 does not include the field oxide films 12 (see FIG. 2B) on the semiconductor substrate 1 , but includes p-type impurity diffusion layers 14 to isolate memory cells lined in the Y direction from one another.
  • the fabrication process of the nonvolatile semiconductor memory 110 is almost the same as those described above (see FIGS. 6 to 10 , 11 A to 11 D, and 12 A and 12 B).
  • the p-type impurity diffusion layers 14 for element isolation are formed by implanting phorone ions with an energy of 40 KeV and a dose amount of 1 ⁇ 10 13 /cm 2 using the control gates (word lines) as the mask after the formation of the control gates (word lines) (FIG. 13 B).
  • the distance between the memory cells adjacent in the Y direction can be made smaller to the limit of the resolution for photolithography.
  • FIG. 14 is a plan view of a portion of a memory cell array of a nonvolatile semiconductor memory 300 of Example 4 according to the present invention.
  • FIGS. 15A and 15B are sectional views of the nonvolatile semiconductor memory 300 taken along lines A—A and B—B of FIG. 14, respectively,
  • the nonvolatile semiconductor memory 300 includes a plurality of memory cells C′ formed in a matrix on a semiconductor substrate 31 .
  • the respective memory cells C′ are denoted by individual codes such as C′ im .
  • Each of the memory cells C′ includes a floating gate 35 formed on the semiconductor substrate 31 via an insulating film.
  • a control gate 37 is formed on the floating gate 35 via an insulating film 36 made of ONO and the like. Such control gates 37 extend in a channel direction of the memory cell C′ (an X direction shown in FIG. 14 ), constituting a word line WL as shown in FIG. 16 for connecting the memory cells C′ lined in the X direction.
  • a high-density n-type diffusion layer 32 is formed between two memory cells adjacent in the X direction (e.g., memory cells C′ im and C′ in ).
  • the diffusion layer 32 includes a drain diffusion layer 9 for one of the two adjacent memory cells (e.g.; the memory cell C′ im ) and a source diffusion layer 10 for the other memory cell (e.g., the memory cell C′ in ).
  • the diffusion layer 32 is thus shared by the two adjacent memory cells.
  • a tunnel insulating film 33 and a gate insulating film 34 are formed underneath each of the floating gate 35 .
  • the tunnel insulating film 33 is formed above the diffusion layer 32 and thinner than the gate insulating film 34 .
  • a tunnel current flows through the tunnel insulating film 33 , while it does not flow through the gate insulating film 34 .
  • the floating gate 35 is capacitively coupled with the drain diffusion layer 9 of one of the two diffusion layers 32 located on both sides of the memory cell C′ via the tunnel insulating film 33 on the drain side (defining a tunnel region 9 ′), while it is capacitively coupled with the source diffusion layer 10 of the other diffusion layer 32 via the gate insulating film 34 on the source side.
  • the transistor of the memory cell C′ in this example is an asymmetric transistor where the insulating film is formed asymmetrically between the drain side and the source side.
  • the tunnel insulating film 33 may be composed of a silicon oxide film or a lamination of a silicon oxide film and a silicon nitride film.
  • the diffusion layers 32 lined in the Y direction are connected with one another via diffusion layer wirings, forming a so-called buried bit line.
  • the diffusion layers 32 may be electrically connected via metal wirings. In the latter case, a contact region is required for each diffusion layer 32 .
  • a field oxide film (element isolation film) 42 is formed between the floating gates 35 of two adjacent memory cells in the Y direction (e.g., the memory cells C′ in and C′ jn ).
  • the nonvolatile semiconductor memory 300 employs the virtual ground method, where bit lines BL (bit lines BL 1 to BL 4 are shown in FIG. 16) serve as the source wirings or the drain wirings depending on the selected cell.
  • Table 2 below shows the operation conditions when the memory cell C′ 12 shown in FIG. 16 is selected, for example.
  • a negative high voltage V H1 (e.g., ⁇ 6 V) is applied to the word line WL 1 connected with the control gate 37 of the memory cell C′ 12 , while 0 V is applied to the other word lines.
  • a predetermined positive voltage V cc (e.g., 3 V) is applied to the bit line BL 2 connected with the drain of the memory cell C′ 12 , while the other bit lines are put in the floating state.
  • V H1 e.g., ⁇ 6 V
  • V cc e.g., 3 V
  • the same write voltage is also applied to the control gate 37 of the non-selected cell C′ 11 , which is connected with the word line WL 1 and the source diffusion layer 10 of the bit line BL 2 .
  • the comparatively thick gate insulating film 34 which does not cause the tunnel phenomenon is formed above the source diffusion layer 10 , no tunnel current flows between the source diffusion layer 10 and the floating gate 35 of the non-selected cell C′ 11 . Accordingly, though the non-selected cell shares the diffusion layer 32 with the selected cell, no data is written in the non-selected cell.
  • 0 V is applied to all the bit lines BL.
  • a positive high voltage V H2 (e.g., 12 V) is applied to a desired word line WL.
  • V H2 e.g. 12 V
  • a high voltage V H2 is applied to the word line WL 1
  • data in the memory cells C′ 11 , C′ 12 , and C′ 13 are erased at one time.
  • a high voltage V H2 is applied to the word line WL 2
  • data in the memory cells C′ 21 , C′ 22 , and C′ 23 are erased at one time.
  • the data reading from the selected cell C′ 12 is conducted in a conventional manner.
  • a predetermined voltage V cc (e.g., 3 V) is applied to the word line WL 1 .
  • a predetermined voltage V L e.g., 1 V
  • 0 V is applied to the bit line BL 2 .
  • Data is read by detecting a current flowing between these bit lines.
  • FIG. 17 is a graph plotting the ratio of the leakage current (i.e., the current flowing into the semiconductor substrate 31 ) to the write current (i.e., the tunnel current flowing between the drain diffusion layer 9 and the floating gate 35 ) with respect to the thickness of the gate insulating film 34 .
  • the thickness of the tunnel insulating film 33 is 11 nm (110 ⁇ ). As is observed from FIG. 17, when the thickness of the gate insulating film 34 is 17 nm or more, the leakage current flowing into the semiconductor substrate 31 is smaller than the tunnel current.
  • the write efficiency can be improved. Since the voltage applied in the write operation can be reduced by thinning the tunnel insulating film 33 , the gate insulating film 34 can be thinned in proportion to the thinning of the tunnel insulating film 33 . Also, the voltage can be lower than the case of the memory cells in Example 1.
  • the asymmetric memory cell can be realized by varying the thickness of the gate insulating film.
  • it can be realized by forming the tunnel insulating film of a different material from that of the other portion of the insulating film.
  • FIG. 18 shows an example of an equivalent circuit of a nonvolatile semiconductor memory cell array 420 where, in order to reduce the wiring resistance of the bit line BL composed of the diffusion layers shown in FIG. 16, a sub-bit line SBL is connected to a main bit line MBL made of metal via a selective transistor.
  • FIGS. 19 to 23 a second example of the fabrication method of a nonvolatile semiconductor memory according to the present invention will become described.
  • the fabrication method of the nonvolatile semiconductor memory 300 shown in FIG. 14 will be described.
  • the field insulating films (element isolation films) 42 shown in FIG. 15B are formed on the semiconductor substrate (silicon substrate) 31 .
  • the gate insulating film 34 with a thickness of about 200 ⁇ (20 nm) is formed by thermal oxidation.
  • a first resist mask 8 a with a predetermined pattern is formed on the gate insulating film 34 by photolithography.
  • the first resist mask 8 a has openings 8 ′ at positions corresponding to regions which will become drain diffusion layers 9 in a later step.
  • Arsenic ions are implanted using the first resist mask 8 a with an energy of 70 KeV and a dose amount of 5 ⁇ 10 15 /cm 2 (FIG. 19 ).
  • the resultant structure is heat-treated under a nitrogen atmosphere at 900° C. for 30 minutes so that the drain diffusion layers 9 which are high-density n-type diffusion layers are formed. Thereafter, a second resist mask 8 b with a predetermined pattern having openings 8 ′′ corresponding to regions which will become tunnel regions 9 ′ in a later step is formed by photolithography. Using the second resist mask 8 b, the portions of the gate insulating film 34 located above the tunnel regions 9 ′ are removed (FIG. 20 ).
  • the tunnel insulating films 33 with a thickness of about 80 ⁇ are formed in the tunnel regions 9 ′, and first polysilicon is deposited on the insulting films to a thickness of about 1500 ⁇ .
  • a third resist mask 8 c with a predetermined pattern is formed on the first polysilicon by photolithography. Using the third resist mask 8 c, the first polysilicon is patterned to form first polysilicon layers 35 ′ in a striped shape. Each of the first polysilicon layers 35 ′ has the same width as the width in the channel direction of the floating gates 35 to be formed in a later step (FIG. 21 ).
  • the first polysilicon layer 35 ′ As the mask, arsenic ions are implanted with an energy of 70 KeV and a dose amount of 1 ⁇ 10 15 /cm 2 (FIG. 22 ).
  • the resultant structure is heat-treated under a nitrogen atmosphere at 900° C. for 30 minutes, thereby to form the source diffusion layers 10 .
  • the source diffusion layer 10 of each memory cell is in contact with the drain diffusion layer 9 of an adjacent memory cell, jointly forming the diffusion layer (bit line) 32 .
  • the ONO film (oxide film/nitride film/oxide film) 36 is formed on the resultant structure by CVD, and then a second polysilicon layer 37 ′ with a thickness of about 1000 ⁇ is formed on the ONO film 36 .
  • the second polysilicon layer 37 ′, the ONO film 36 , and the first polysilicon layers 35 ′ are then patterned by photolithography (see FIG. 15 B), to form the control gates 37 , the insulating films 36 , and the floating gates 35 (FIG. 23 ).
  • the control gates 37 of the respective memory cells are formed consecutively, forming the word line.
  • control gates (word line) 37 may be of a double-layered structure where metal silicide with a high melting point such as tungsten silicide (WSi) is deposited to a thickness of about 1000 ⁇ on the second polysilicon layer with a thickness of about 1000 ⁇ and these layers are patterned.
  • metal silicide with a high melting point such as tungsten silicide (WSi) is deposited to a thickness of about 1000 ⁇ on the second polysilicon layer with a thickness of about 1000 ⁇ and these layers are patterned.
  • FIGS. 24 to 28 an alternative fabrication method of the nonvolatile semiconductor memory according to the present invention will be described with reference to FIGS. 24 to 28 .
  • two different masks 8 a and 8 b for ion implanting to form the drain diffusion layer and to define the tunnel region, respectively are used. Since allowances for aligning the respective masks are required at the formation of these masks, the size reduction of the memory cells is restricted.
  • impurity ions are implanted using the mask for defining the tunnel region, so as to form the diffusion layer in the tunnel region in a self-alignment manner.
  • the gate insulating film (first gate insulating film) 34 with a thickness of about 200 ⁇ (20 nm) is formed by thermal oxidation on the semiconductor substrate 31 having the field insulating films (element isolation films) 42 formed thereon.
  • a resist mask 8 d with a predetermined pattern having openings corresponding to regions which will become tunnel regions in a later step is formed on the gate insulating film by photolithography.
  • arsenic ions are implanted with an energy of 70 KeV and a dose amount of 5 ⁇ 10 15 /cm 2 (FIG. 24 ).
  • the portions of the gate insulating film 34 located above the tunnel regions are removed (FIG. 25 ).
  • the resultant structure is heat-treated under a nitrogen atmosphere at 900° C. for 10 minutes, forming the drain diffusion layers 9 which are high-density n-type diffusion layers.
  • the tunnel insulating films (second gate insulating films) 33 with a thickness of about 80 ⁇ are then formed by thermal oxidation at 900° C. for 10 minutes (FIG. 26 ).
  • a fist polysilicon film with a thickness of about 1000 to 2000 ⁇ is formed, which is then patterned by photolithography using a resist mask with a predetermined pattern to form the first polysilicon layers 35 ′ in a striped shape.
  • Each of the first polysilicon layers 35 ′ has the same width as the width in the channel direction of the floating gates 35 to be formed in a later step.
  • phosphorous ions are implanted at a tilt angle of about 45° with respect to the normal of the semiconductor substrate 31 (tilt ion implantation) with an energy of 30 KeV and a dose amount of 1 to 5 ⁇ 10 14 /cm 2 .
  • arsenic ions are implanted by tilt ion implantation with an energy of 70 KeV and a dose amount of 1 ⁇ 10 15 /cm 2 (FIG. 27 ). Only one-time implantation of either phosphorous ions or arsenic ions is acceptable. However, the contact with the diffusion layers 9 under the tunnel insulating films 33 can be more ensured by the two-time implantation of phosphorous ions and arsenic ions as described above. Thereafter, the source diffusion layers 10 are formed by predetermined heat-treatment. The source diffusion layer 10 of one memory cell is in contact with the drain diffusion layer 9 of an adjacent memory cell, jointly forming the diffusion layer (bit line) 32 .
  • the ONO film 36 is formed on the resultant structure by CVD, and then the second polysilicon layer 37 ′ with a thickness of about 1000 ⁇ is formed on the ONO film 36 .
  • the second polysilicon layer 37 ′, the ONO film 36 , and the first polysilicon layers 35 ′ are then patterned (see FIG. 15B) to form the control gates 37 , the insulating films 36 and the floating gates 35 (FIG. 28 ).
  • the control gates 37 of the respective memory cells are formed consecutively, forming the word line.
  • control gates (word line) 37 may have a double-layered structure where metal silicide with a high melting point such as tungsten silicide (WSi) is deposited to a thickness of 1000 ⁇ on the second polysilicon layer of a thickness of about 1000 ⁇ and these layers are patterned.
  • metal silicide with a high melting point such as tungsten silicide (WSi) is deposited to a thickness of 1000 ⁇ on the second polysilicon layer of a thickness of about 1000 ⁇ and these layers are patterned.
  • the memory cell structure is simplified and the cell size can be reduced. Since all the tunnel regions are formed above the diffusion layers (especially, the drain diffusion layers), a leakage current from the diffusion layers to the semiconductor substrate in the write operation can be reduced. This improves the write efficiency and lowers the power consumption. Since no electron/hole trapping due to the leakage current occurs, the reliability of the memory cells can be improved.
  • FIGS. 29A and 29B show an alternative example of the nonvolatile semiconductor memory according to the present invention.
  • a nonvolatile semiconductor memory- 310 does not include the field oxide films 42 (see FIG. 15B) on the semiconductor substrate 31 . Instead, it includes p-type impurity diffusion layers 39 to isolate memory cells lined in the Y direction from one another.
  • the fabrication process of the nonvolatile semiconductor memory 310 is almost the same as those described above (see FIGS. 19 to 23 , and FIGS. 24 to 28 ).
  • the p-type impurity diffusion layers 39 are formed by implanting phorone ions under the condition of an energy of 40 KeV and a dose amount of 1 ⁇ 10 13 /cm 2 using the control gates (word lines) 37 as the mask after the formation of the control gates (word lines) 37 (FIG. 29 B).
  • Example 5 a nonvolatile semiconductor memory 320 using asymmetric transistors where the insulating film is formed asymmetrically between the drain side and the source side, as the nonvolatile semiconductor memory 300 of Example 3, will be described.
  • FIG. 30 is a plan view of a portion of a memory cell array of the nonvolatile semiconductor memory 320 of this example according to the present invention.
  • FIGS. 31A and 31B are sectional views of the nonvolatile semiconductor memory 320 taken along lines A—A and B—B of FIG. 30, respectively,
  • the nonvolatile semiconductor memory 320 includes a plurality of memory cells C′′ formed in a matrix on a semiconductor substrate 41 .
  • a floating gate 51 is formed on the semiconductor substrate 41 via an insulating film 55 .
  • the insulating film 55 is not uniform in thickness, including a gate insulating film (first gate insulating film) 42 , a tunnel insulating film (second gate insulating film) 50 , and a bit line insulating film 48 of different thicknesses.
  • a control gate 53 is formed on the floating gate 51 via an insulating film 52 made of ONO and the like. Such control gates 53 extend in a channel direction of the memory cell C′′ (an X direction shown in FIG. 30 ), constituting a word line WL for connecting the memory cells C′′ lined in the X direction.
  • a high-density n-type diffusion layer 46 is formed between two memory cells adjacent in the X direction.
  • the diffusion layer 46 includes a drain diffusion. layer for one of the two adjacent memory cells and a source diffusion layer for the other memory cell.
  • the diffusion layer 46 is thus shared by the two adjacent memory cells.
  • the diffusion layers 46 lined in the Y direction are connected via diffusion layer wirings, forming a buried bit line.
  • the gate insulating film 42 and the tunnel insulating film 50 are formed underneath the floating gate 51 . All the tunnel insulating films 50 are formed above the diffusion layers 46 . Each of the diffusion layers 46 is capacitively coupled with the floating gate 51 of one of the adjacent memory cells via the tunnel insulating film 50 on the drain side (defining a tunnel region 50 ), while it is capacitively coupled with the floating gate 51 of the other memory cell via the gate insulating film 42 on the source side.
  • the tunnel insulating film 50 is thinner than the gate insulating film 42 .
  • the transistor of the memory cell C′′ in this example is an asymmetric transistor where the insulating film is formed asymmetrically between the drain side and the source side.
  • the tunnel insulating film 50 may be composed of a silicon oxide film or a lamination of a silicon oxide film and a silicon nitride film.
  • the bit line insulating films 48 with a comparatively large thickness are formed above the portions of the diffusion layers 46 other than the drain and source diffusion layers (i.e., in the regions located between adjacent memory cells).
  • the nonvolatile semiconductor memory 320 employs the virtual ground method, where the bit lines BL serve as the source wirings or the drain wirings depending on the selected cell.
  • the gate insulating film 42 with a thickness of about 40 nm, for example, is formed on the semiconductor substrate (silicon substrate) 41 .
  • a first silicon nitride film 43 of a thickness of about 100 nm, for example) is formed on the gate insulating film 42 and patterned into a predetermined stripe shape.
  • the first silicon nitride film 43 serves as a substantial mask for forming the buried bit lines (diffusion layers).
  • an oxide film 44 with a thickness of about 40 nm, for example, is formed over the semiconductor substrate 41 (FIG. 32 ).
  • a second silicon nitride film 45 is formed on the oxide film 44 and etched back to form nitride film spacers 45 on both side walls of each of the first silicon nitride films 43 covered with the oxide film 44 .
  • Each of the nitride film spacers 45 is composed of a nitride film spacer 45 a corresponding to the drain side (tunnel region) and a nitride film spacer 45 b corresponding to the source side.
  • arsenic ions are implanted, forming the high-density N-type impurity diffusion layers 46 by heat treatment (i.e., the bit lines BL) (FIG. 33 ).
  • a resist mask 47 with a predetermined pattern is formed to cover the nitride film spacers 45 a corresponding to the tunnel regions 50 ′ to be formed in a later step, and the nitride film spacers 45 b are removed by isotropic plasma etching using the resist mask 47 (FIG. 34 ).
  • the comparatively thick oxide films 48 are formed on the semiconductor substrate 41 , covering the diffusion layers 46 (i.e., the bit lines BL) (FIG. 35 ).
  • the nitride film spacers 45 a and then the oxide films 44 are removed by etching.
  • the portions of the oxide films 44 located above the tunnel regions 50 ′ are completely removed by removing the thermal oxide films ( 44 and 48 ) by the thickness of the oxide films 44 (about 40 nm) by etching.
  • the oxide films 48 are sufficiently thick, the portions of the semiconductor substrate other than the tunnel regions 50 ′ will not be exposed.
  • thin thermal oxide films 49 are formed on the tunnel regions 50 ′ (FIG. 36 ).
  • the first silicon nitride films 43 are completely removed by phosphorous boiling, for example.
  • the oxide films 49 are then removed by etching, to expose only the portion of the semiconductor substrate which is to be the tunnel regions 50 ′.
  • the tunnel insulating films 50 are formed by thermal oxidation.
  • a first polysilicon film with a thickness of about 1500 ⁇ , for example, is formed on the resultant structure, and patterned into a predetermined shape by photolithography, forming the floating gates 51 (FIG. 37 ).
  • the control gates 53 of the respective memory cells are formed consecutively, forming the word line.
  • the first polysilicon film may be formed in the shape of strips each having the same width as the width in the channel direction of the floating gates 51 and then patterned simultaneously with the second polysilicon film.
  • the control gates 53 may be the word line with the double-layered structure where metal silicide with a high melting point such as tungsten silicide (WSi) is deposited to a thickness of about 100 nm on the second polysilicon layer with a thickness of about 100 nm and these layers are patterned.
  • metal silicide with a high melting point such as tungsten silicide (WSi)
  • each transistor of this example is an asymmetric transistor having an insulating film including portions with different thicknesses (the gate insulating film 48 and the tunnel insulating film 50 ) underneath the floating gate 51 .
  • the tunnel region 50 ′ is defined by the nitride film spacer 45 a formed on the side walls of the first nitride film 43 . This makes it possible to reduce the regions where the tunnel insulating films 50 are formed.
  • the capacitance coupling ratio of the control gates increases. This eliminates the necessity of a high voltage transistor module and thus reduces processing costs.
  • the wiring portions of the diffusion layers 46 i.e., the portions of the diffusion layers 46 other than the source coupling regions and the drain coupling regions (tunnel regions 50 ′) coupled with the floating gates 51 are covered with the comparatively thick insulating films 48 . This reduces the parasitic capacitance between the control gates 53 and the bit lines.
  • FIG. 39 is a plan view showing the configuration of a nonvolatile semiconductor memory 400 of Example 6 according to the present invention.
  • a memory cell array of the nonvolatile semiconductor memory 400 is divided into a plurality of blocks 410 . This division of the memory cell array into a predetermined number of blocks contributes to reducing the parasitic capacitance and resistance of wirings and increasing the operation speed.
  • the memory cells C used in the nonvolatile semiconductor memory 100 in Example 1 are used.
  • the configuration of this example can also be applied to the nonvolatile semiconductor memory 300 of Example 3.
  • the operation of such a nonvolatile semiconductor memory obtained by dividing the memory cell array of the nonvolatile semiconductor memory 300 of Example 3 into a plurality of blocks is basically the same as that described hereinbelow. Voltages to be applied may be changed depending on the characteristics of the memory cells.
  • the diffusion layers 2 of the nonvolatile semiconductor memory 400 are coupled in the Y direction, forming sub-bit lines SBL (sub-bit lines SBL 1 to SBL 5 are shown in FIG. 39 ).
  • One main bit line MBL is arranged for every two sub-bit lines SBL (main bit lines MBL 1 to MBL 3 are shown in FIG. 39 ).
  • the main bit lines MBL are made of a metal layer and arranged in the same direction as the sub-bit lines SBL at a pitch of every two memory cells C.
  • the submit lines SBL are diffusion layer wirings
  • the main bit lines MBL are metal wirings so that the memory cells can be formed irrelevant of the pitch of the metal wirings, the pitch of the metal wirings can be every two memory cells, and thus the parasitic capacitance of the metal wirings can be reduced.
  • the sub-bit lines SBL are connected to the corresponding main bit lines MBL via selective transistors.
  • the sub-bit lines SBL 1 and SBL 3 are connected to the main bit line MBL 1 via selective transistors Q 1 and Q 2 , respectively, while the sub-bit lines SBL 2 and SBL 4 are connected to the main bit line MBL 2 via selective transistors Q 4 and Q 5 , respectively.
  • the selective transistors are controlled via corresponding selective lines ST (selective lines ST 11 to ST 41 , and ST 12 to ST 42 are shown in FIG. 39 ).
  • the right-side number of the subscript of the word line code (01, n1, etc.) represents the block number. The number is therefore 2 for the word lines in a second block 410 ′ shown in FIG. 39 (word lines WL 02 , WL n2 , etc.).
  • the configuration of the block 410 and the arrangement of the main bit lines MBL are not limited to those described above.
  • the selective transistors Q 1 to Q 5 are turned off to electrically isolate all the diffusion layers 2 (sub-bit lines SBL) in the block 410 from the corresponding main bit lines MBL (floating state). This prevents data in the non-selected block from being erased mistakenly even if the erase operation is repeated.
  • the sub-bit line closest to the boundary of each block is not connected with a memory cell included in the adjacent block.
  • Table 3 shows examples of applied voltages in the respective operation modes when a memory cell C 31 is selected.
  • a negative high voltage (e.g., ⁇ 8 V) is applied to the word line WL 01 , while the main bit line MBL 1 is put in the floating state, and a predetermined positive voltage (e.g., 4 V) is applied to the main bit line MBL 2 .
  • the selective line ST 21 is set at a low level (6 V) to turn on the selective transistor Q 1 and thus connect the sub-bit line SBL 1 to the main bit line MBL 1 .
  • the selective lines ST 11 and ST 41 are set at a low level (e.g., 0 V) to turn off the selective transistors Q 2 and Q 5 and put the sub-bit lines SBL 3 and SBL 4 in the floating state.
  • a negative high voltage is applied to the control gate 7 , 0 V is applied to the source diffusion layer 2 b, and a predetermined positive voltage is applied to the drain diffusion layer 2 a.
  • This causes a tunnel current to flow into the floating gate 5 (i.e., electrons are drawn out), and thus data to be written in the selected cell C 31 .
  • the drain disturb immunity of the memory cells connected with the same sub-bit line SBL 1 can be improved by applying a voltage lower than the voltage applied to the drain to the non-selected word lines WL in the same block.
  • the erase operation is conducted for every sector or every block.
  • the voltage of 0 V is applied to the main bit lines MBL 1 , MBL 2 , and MBL 3 and a positive high voltage (e.g., 17 V) is applied to a predetermined word lie e.g., the word line WL 01 , while 0 V is applied to the other word lines WL n1 .
  • Data of the memory cells connected with the word line WL to which the positive high voltage has been applied are erased, while data of the memory cells connected with the word lines WL to which 0 V is applied are not erased, by setting the selective lines ST 11 , ST 21 , ST 31 , and ST 41 3 at a high level (e.g., 3 V) and turning all the selective transistors in the block where the memory cells to be erased are included. Accordingly, when sector-unit erasing is conducted, a positive high voltage is applied to only one word line, while 0 V is applied to the other word lines.
  • a high level e.g. 3 V
  • the read operation for the selected cell C 31 is conducted in the following manner.
  • a predetermined positive voltage e.g., 1 V
  • a positive voltage e.g., 3 V
  • the selective lines ST 21 and ST 31 are set at a high level (e.g., 3 V) to turn on the selective transistors Q 1 and Q 4 . No “soft write” phenomenon occurs in the memory cells by reading data from the low-density source diffusion layer side.
  • FIG. 40 shows how the memory cells connect with the word line WL 01 . Referring to FIG. 40, in the read operation for a selected memory cell (A in FIG.
  • the drain diffusion layer (sub-bit line SBL 1 ) is connected to the main bit line MBL 1 (ground level) via the selective transistor Q 1
  • the source diffusion layer (sub-bit line SBL 2 ) is connected to the main bit line MBL 2 (1 V) via the selective transistor Q 4
  • the sub-bit line SBL 5 is connected to the main bit line MBL 3 via the selective transistor Q 3 .
  • the memory cell transistor is turned on or off depending on whether data stored in the memory cell C 31 is 1 or 0. When the memory cell transistor is off, correct reading of data may not be possible due to the lowering of the potential at the main bit line MBL 2 (sub-bit line SBL 2 ).
  • the same potential is applied to the main bit line MBL 3 (sub-bit line SBL 5 ).
  • the sub-bit lines SBL 3 and SBL 4 are in the floating state and the parasitic capacitance in the sub-bit lines is small compared with the main bit lines. The influence of the sub-bit lines is therefore small.
  • the leakage current flowing from the memory cell C 32 through the memory cell C 34 is negligible compared with the current flowing to the memory cell C 31 , even if the same voltage has been applied to the main bit line MBL 3 (sub-bit line SBL 5 ), causing no trouble for reading.
  • the capacitance of the main bit line is denoted as C MBL .
  • the leakage current flowing along the word line WL n1 from the sub-bit line SBL 2 to the sub-bit line SBL 1 adjacent each other in the Y direction is negligible compared with the channel current of the memory cell A in which the data has been written, as shown in FIG. 41 .
  • the threshold voltage of the memory cell is more than 0.5 V
  • the leakage current of the non-selected memory cell of which control gate is grounded is smaller than the read current by seven digits, which is therefore negligible.
  • a high voltage (17 V) is applied to the word line in the erase operation.
  • the high voltage applied to the word line can be reduced by applying a negative voltage to the diffusion layer 2 and the semiconductor substrate 1 in the erase operation.
  • the selective transistors in a nonselected block are turned off, to put the diffusion layers 2 (i.e., the sub-bit lines SBL) of the non-selected block in the floating state.
  • the selective transistors Q are turned on or off depending on the high level (0 V) or the low level (negative voltage, e.g., ⁇ 8 V) of the selective lines.
  • a negative voltage (e.g., ⁇ 8 V) is applied to the semiconductor substrate 1
  • a positive voltage e.g., 9 V
  • a negative voltage e.g., ⁇ 8 V
  • a negative voltage e.g., ⁇ 8 V
  • FIG. 43 shows the time required to erase data in a memory cell (erase time) when voltages of 8 V and 9 V are applied to the control gate 7 .
  • the threshold voltage of the memory cell in the erase state is 5 V
  • the control gate 7 data in the memory cell is erased in about 10 ⁇ 2 seconds.
  • the memory cell array is formed on a well which is formed in the semiconductor substrate.
  • a negative voltage is applied to the well, it is necessary to isolate the well for each block to prevent the negative voltage from being applied to other non-selected blocks formed on the same substrate. Isolating the well for every block results in increasing the size of the memory. In practice, however, no problem will occur in the aspect of reliability of the memory cells if the well is not isolated for every block, due to the following reasons.
  • FIG. 44 shows the relationship between the negative voltage ( ⁇ V sub ) applied to the semiconductor substrate and the reliability (disturb immunity) of the tunnel insulating film (thickness: about 80 ⁇ ). As is observed from FIG. 44, when the tunnel insulating film with a thickness of about 80 ⁇ is formed, the nonvolatile semiconductor memory has a life of 10 years or longer as long as the voltage applied to the substrate does not exceed ⁇ 20 V.
  • two sub-bit lines SBL are connected to one main bit line MBL via respective selective transistors.
  • the main bit line MBL is selected in the following manner so that data can be written in the corresponding memory cell.
  • the main bit line MBL 2 is selected as the drain wiring will be described with reference to Table 4 below and FIG. 45 .
  • a negative write voltage ( ⁇ 8 V) is applied to the selected word line WL o1 .
  • The, ground voltage (0 V) is retained for the non-selected word lines WL n1 (n ⁇ 0).
  • the selective lines ST 21 and ST 31 are set at a high level (6 V)
  • the sub-bit line SBL 1 is connected to the main bit line MBL 1 via the selective transistor Q 1
  • the sub-bit line SBL 2 is connected to the main bit line MBL 2 via the selective transistor Q 4
  • the sub-bit line SBL 3 is connected to the main bit line MBL 3 via the selective transistor Q 3 .
  • a write voltage is applied to the main bit line MBL depending on the data to be stored in the memory cell C in which the data is to be written. For example, 4 V is applied as the write voltage for data 1 to allow a tunnel current to flow, while the floating state is established for data 0 to inhibit writing.
  • data are written in the memory cells C 31 , C 32 ,C 35 (not shown), etc.
  • the selective lines for the other nonselected selective transistors are retained grounded.
  • the selective lines ST 21 , and ST 31 are set at a low level (0 V), and then the selective lines ST 11 and ST 41 are set at a high level (5 V).
  • the sub-bit line SBL 3 is connected to the main bit line MBL, via the selective transistor Q 2
  • the sub-bit line SBL 4 is connected to the main bit line MBL 2 via the selective transistor Q 5 .
  • a write voltage corresponding to the memory cell C to be selected next is applied to each main bit line MBL (4 V for data 1 , while the floating state is established for data 0 ).
  • data are written in the memory cells C 33 and C 34 .
  • two sub-bit lines SBL are connected to one main bit line MBL.
  • an arbitrary number (N) of sub-bit lines SBL may be connected to the main bit line MBL via corresponding selective transistors.
  • data can be written in all memory cells corresponding to the N sub-bit lines SBL connected to one main bit line MBL by N-time repetition of the write operation.
  • the sub-bit line SBL 1 for example, is connected to the main bit line MBL 1 via the selective transistor Q 1 .
  • the sub-bit lines SBL 1 , SBL 2 , . . . are electrically isolated from the main bit lines MBL 1 , MBL 2 , . . . and put in the floating state by turning off the selective transistors Q 1 , Q 2 , . . . , respectively.
  • nonvolatile semiconductor memory 430 data can be written in all the memory cells connected with one word line by one write operation. However, the main bit line is required for every sub-bit line. Since the memory cell size is determined by the pitch of the metal wirings, this configuration of the nonvolatile semiconductor memory 430 is not suitable for miniaturization compared with the device shown in FIG. 39 .
  • a tunnel current flows only between the drain diffusion layer and the floating gate by applying a predetermined voltage to the source/drain diffusion layer (bit line).
  • bit line the source/drain diffusion layer
  • the impurity density is made different between the source diffusion layer and the drain diffusion layer constituting the source/drain diffusion layer which is shared by two adjacent memory cells. This makes it possible to flow a tunnel current only in the region of the memory cell coupled with the drain diffusion layer.
  • the thicker portion (the gate insulating film) and the thinner portion (the tunnel insulating film) are formed in the insulating film formed underneath the floating gate. This makes it possible to flow a tunnel current only on the drain side of the memory cell capacitively coupled via the tunnel insulating film when a predetermined voltage is applied, even though the memory cell shares the source/drain diffusion layer with an adjacent memory cell.
  • All the comparatively thin tunnel insulating films are formed above the source/drain diffusion layers. This makes it possible to greatly reduce the generation of an interband tunnel current in the write operation and thus improve the write efficiency and the reliability of the memory cells.
  • the nitride film spacers are formed as the mask defining the arrangement of the source/drain diffusion layers, and the tunnel regions are defined by use of the nitride film spacers.
  • the tunnel regions can be made smaller. This makes it possible to increase the capacitance coupling ratio of the control gates, eliminating the necessity of a high voltage transistor module, and thus reducing processing costs.
  • the wiring portions of the source/drain diffusion layers i.e., the portions of the diffusion layers other than the source coupling regions and the drain coupling regions coupled with the floating gates (tunnel regions) are covered with the comparatively thick insulating films. This makes it possible to reduce the parasitic capacitance between the control gates and the bit lines.

Landscapes

  • Engineering & Computer Science (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Non-Volatile Memory (AREA)
  • Semiconductor Memories (AREA)
US09/340,682 1995-06-29 1999-06-29 Method of making nonvolatile semiconductor memory Expired - Lifetime USRE37199E1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP96304829A EP0752721B1 (fr) 1995-06-29 1996-06-28 Mémoire semi-conductrice non-volatile et méthode de commande et procédé de fabrication
US09/340,682 USRE37199E1 (en) 1995-06-29 1999-06-29 Method of making nonvolatile semiconductor memory

Applications Claiming Priority (8)

Application Number Priority Date Filing Date Title
JP7-163727 1995-06-29
JP16372795 1995-06-29
JP17728695 1995-07-13
JP7-177286 1995-07-13
JP8164343A JPH0992739A (ja) 1995-06-29 1996-06-25 不揮発性半導体メモリとその駆動方法及び製造方法
JP8-164343 1996-06-25
US08/672,666 US5877054A (en) 1995-06-29 1996-06-28 Method of making nonvolatile semiconductor memory
US09/340,682 USRE37199E1 (en) 1995-06-29 1999-06-29 Method of making nonvolatile semiconductor memory

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
US08/672,666 Reissue US5877054A (en) 1995-06-29 1996-06-28 Method of making nonvolatile semiconductor memory

Publications (1)

Publication Number Publication Date
USRE37199E1 true USRE37199E1 (en) 2001-05-29

Family

ID=27473884

Family Applications (1)

Application Number Title Priority Date Filing Date
US09/340,682 Expired - Lifetime USRE37199E1 (en) 1995-06-29 1999-06-29 Method of making nonvolatile semiconductor memory

Country Status (2)

Country Link
US (1) USRE37199E1 (fr)
EP (1) EP0752721B1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6621130B2 (en) * 2001-05-31 2003-09-16 Semiconductor Energy Laboratory Co., Ltd. Semiconductor device and an electronic device
US10276679B2 (en) * 2017-05-30 2019-04-30 Vanguard International Semiconductor Corporation Semiconductor device and method for manufacturing the same

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5837584A (en) * 1997-01-15 1998-11-17 Macronix International Co., Ltd. Virtual ground flash cell with asymmetrically placed source and drain and method of fabrication
EP0969507B1 (fr) * 1998-06-30 2006-11-15 STMicroelectronics S.r.l. Procédé de fabrication d'une cellule de mémoire EEPROM
JP3754600B2 (ja) * 2000-06-13 2006-03-15 シャープ株式会社 不揮発性半導体記憶装置およびそのテスト方法
JP3653449B2 (ja) 2000-06-15 2005-05-25 シャープ株式会社 不揮発性半導体記憶装置

Citations (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2073488A (en) * 1980-04-09 1981-10-14 Hughes Aircraft Co Electrically erasable programmable read only memory
EP0256993A1 (fr) * 1986-06-03 1988-02-24 STMicroelectronics S.r.l. Dispositif de mémoire morte altérable électriquement à grille flottante
JPS6399573A (ja) 1986-10-16 1988-04-30 Sony Corp メモリ装置
DE3844115A1 (de) * 1987-12-28 1989-07-13 Toshiba Kawasaki Kk Elektrisch loeschbarer, programmierbarer festwertspeicher mit nand-zellenstruktur
JPH022685A (ja) * 1988-06-15 1990-01-08 Fujitsu Ltd 半導体記憶装置の製造方法
JPH0287578A (ja) * 1988-09-26 1990-03-28 Toshiba Corp 半導体装置
JPH02174171A (ja) 1988-12-26 1990-07-05 Matsushita Electron Corp 半導体記憶装置
US4945068A (en) * 1988-10-25 1990-07-31 Matsushita Electronics Corporation Manufacturing method of semiconductor nonvolatile memory device
EP0419663A1 (fr) * 1988-10-21 1991-04-03 Kabushiki Kaisha Toshiba Memoire a semi-conducteur remanente et procede de production
JPH04130778A (ja) * 1990-09-21 1992-05-01 Mitsubishi Electric Corp 不揮発性半導体記憶装置
JPH04208573A (ja) * 1990-11-30 1992-07-30 Nec Corp 不揮発性半導体記憶装置の製造方法
EP0535694A2 (fr) * 1991-10-03 1993-04-07 Kabushiki Kaisha Toshiba Dispositif semi-conducteur de mémoire et méthode pour sa fabrication
US5273923A (en) * 1991-10-09 1993-12-28 Motorola, Inc. Process for fabricating an EEPROM cell having a tunnel opening which overlaps field isolation regions
US5283758A (en) * 1989-06-13 1994-02-01 Mitsubishi Denki Kabushiki Kaisha Non-volatile semiconductor memory device
US5379254A (en) * 1992-10-20 1995-01-03 National Semiconductor Corporation Asymmetrical alternate metal virtual ground EPROM array
JPH0794686A (ja) 1993-06-30 1995-04-07 Nippon Steel Corp 不揮発性半導体記憶装置及びその製造方法
JPH07183407A (ja) * 1993-12-22 1995-07-21 Nec Corp 不揮発性半導体記憶装置
JPH07254686A (ja) 1994-03-15 1995-10-03 Toshiba Corp 不揮発性半導体記憶装置
US5516713A (en) * 1994-09-06 1996-05-14 United Microelectronics Corporation Method of making high coupling ratio NAND type flash memory
US5521110A (en) * 1994-12-28 1996-05-28 National Semiconductor Corporation Method of making EEPROM devices with smaller cell size
US5592415A (en) 1992-07-06 1997-01-07 Hitachi, Ltd. Non-volatile semiconductor memory
US5677867A (en) 1991-06-12 1997-10-14 Hazani; Emanuel Memory with isolatable expandable bit lines
US5745417A (en) 1992-04-07 1998-04-28 Mitsubishi Denki Kabushiki Kaisha Electrically programmable and erasable nonvolatile semiconductor memory device and operating method therefor
US5747846A (en) 1993-11-25 1998-05-05 Nippondenso Co., Ltd. Programmable non-volatile memory cell

Patent Citations (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2073488A (en) * 1980-04-09 1981-10-14 Hughes Aircraft Co Electrically erasable programmable read only memory
EP0256993A1 (fr) * 1986-06-03 1988-02-24 STMicroelectronics S.r.l. Dispositif de mémoire morte altérable électriquement à grille flottante
JPS6399573A (ja) 1986-10-16 1988-04-30 Sony Corp メモリ装置
US4939690A (en) * 1987-12-28 1990-07-03 Kabushiki Kaisha Toshiba Electrically erasable programmable read-only memory with NAND cell structure that suppresses memory cell threshold voltage variation
DE3844115A1 (de) * 1987-12-28 1989-07-13 Toshiba Kawasaki Kk Elektrisch loeschbarer, programmierbarer festwertspeicher mit nand-zellenstruktur
JPH022685A (ja) * 1988-06-15 1990-01-08 Fujitsu Ltd 半導体記憶装置の製造方法
EP0364769A2 (fr) * 1988-09-26 1990-04-25 Kabushiki Kaisha Toshiba Dispositif semi-conducteur à électrode de commande comportant plusieurs couches
JPH0287578A (ja) * 1988-09-26 1990-03-28 Toshiba Corp 半導体装置
EP0419663A1 (fr) * 1988-10-21 1991-04-03 Kabushiki Kaisha Toshiba Memoire a semi-conducteur remanente et procede de production
US4945068A (en) * 1988-10-25 1990-07-31 Matsushita Electronics Corporation Manufacturing method of semiconductor nonvolatile memory device
JPH02174171A (ja) 1988-12-26 1990-07-05 Matsushita Electron Corp 半導体記憶装置
US5283758A (en) * 1989-06-13 1994-02-01 Mitsubishi Denki Kabushiki Kaisha Non-volatile semiconductor memory device
JPH04130778A (ja) * 1990-09-21 1992-05-01 Mitsubishi Electric Corp 不揮発性半導体記憶装置
JPH04208573A (ja) * 1990-11-30 1992-07-30 Nec Corp 不揮発性半導体記憶装置の製造方法
US5677867A (en) 1991-06-12 1997-10-14 Hazani; Emanuel Memory with isolatable expandable bit lines
EP0535694A2 (fr) * 1991-10-03 1993-04-07 Kabushiki Kaisha Toshiba Dispositif semi-conducteur de mémoire et méthode pour sa fabrication
US5273923A (en) * 1991-10-09 1993-12-28 Motorola, Inc. Process for fabricating an EEPROM cell having a tunnel opening which overlaps field isolation regions
US5745417A (en) 1992-04-07 1998-04-28 Mitsubishi Denki Kabushiki Kaisha Electrically programmable and erasable nonvolatile semiconductor memory device and operating method therefor
US5592415A (en) 1992-07-06 1997-01-07 Hitachi, Ltd. Non-volatile semiconductor memory
US5379254A (en) * 1992-10-20 1995-01-03 National Semiconductor Corporation Asymmetrical alternate metal virtual ground EPROM array
JPH0794686A (ja) 1993-06-30 1995-04-07 Nippon Steel Corp 不揮発性半導体記憶装置及びその製造方法
US5747846A (en) 1993-11-25 1998-05-05 Nippondenso Co., Ltd. Programmable non-volatile memory cell
JPH07183407A (ja) * 1993-12-22 1995-07-21 Nec Corp 不揮発性半導体記憶装置
JPH07254686A (ja) 1994-03-15 1995-10-03 Toshiba Corp 不揮発性半導体記憶装置
US5516713A (en) * 1994-09-06 1996-05-14 United Microelectronics Corporation Method of making high coupling ratio NAND type flash memory
US5521110A (en) * 1994-12-28 1996-05-28 National Semiconductor Corporation Method of making EEPROM devices with smaller cell size

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
Hisamune et al., "A High Capacitive-Coupling Ratio (HiCR) Cell for 3 V-Only 64 Mbit and future Flash Memories", IEDM 93, pp. 19-22, 1993.*
Kato et al., "A 0.4- m Self-Aligned Contactless Memory Cell Technology Suitable for 256-Mbit Flash Memories", IEDM 94, pp. 921-923, 1994. *
Ohi et al., "An Asymmetrical Offset Source/Drain Structure For Virtual Ground Array Flash Memory with Dinor Operation", Digest of Technical Papers of the Symposium on VLSI Technology, Kyoto, May 17-19, 1993, No. -, May 17, 1993, Institute of Electrical and.*
Yoshikawa et al., "An Asymmetrical Lightly Doped Source Cell For Virtual Ground High-Density Eprom's", IEEE Transactions on Electron Devices, vol. 37, No. 4, pp. 1046-1051, XP 000125218, Apr. 1, 1990.*

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6621130B2 (en) * 2001-05-31 2003-09-16 Semiconductor Energy Laboratory Co., Ltd. Semiconductor device and an electronic device
US20040036094A1 (en) * 2001-05-31 2004-02-26 Semiconductor Energy Laboratory Co., Ltd. Semiconductor device and an electonic device
US7268387B2 (en) 2001-05-31 2007-09-11 Semiconductor Energy Laboratory Co., Ltd. Semiconductor device and an electronic device
US10276679B2 (en) * 2017-05-30 2019-04-30 Vanguard International Semiconductor Corporation Semiconductor device and method for manufacturing the same

Also Published As

Publication number Publication date
EP0752721A3 (fr) 1997-06-11
EP0752721B1 (fr) 2009-04-29
EP0752721A2 (fr) 1997-01-08

Similar Documents

Publication Publication Date Title
US5877054A (en) Method of making nonvolatile semiconductor memory
US5399891A (en) Floating gate or flash EPROM transistor array having contactless source and drain diffusions
US5939749A (en) Split gate transistor array
US6212103B1 (en) Method for operating flash memory
US6531732B2 (en) Nonvolatile semiconductor memory device, process of manufacturing the same and method of operating the same
US5284785A (en) Diffusionless source/drain conductor electrically-erasable, electrically-programmable read-only memory and methods for making and using the same
US6157575A (en) Nonvolatile memory device and operating method thereof
US20060071265A1 (en) Nonvolatile memory devices and methods of forming the same
US6310374B1 (en) Nonvolatile semiconductor memory device having extracting electrode
KR100247228B1 (ko) 워드라인과 자기정렬된 부우스팅 라인을 가지는불휘발성 반도체 메모리
US7553725B2 (en) Nonvolatile memory devices and methods of fabricating the same
US6376295B1 (en) Method for manufacturing a semiconductor memory device with a fine structure
US5414286A (en) Nonvolatile memory, method of fabricating the same, and method of reading information from the same
JP2967346B2 (ja) 不揮発性メモリ装置の製造方法
US6144064A (en) Split-gate EEPROM device having floating gate with double polysilicon layer
US7187029B2 (en) Nonvolatile semiconductor memory device with floating gate and two control gates
USRE37199E1 (en) Method of making nonvolatile semiconductor memory
US5057446A (en) Method of making an EEPROM with improved capacitive coupling between control gate and floating gate
US5936889A (en) Array of nonvolatile memory device and method for fabricating the same
US6392927B2 (en) Cell array, operating method of the same and manufacturing method of the same
JP2901473B2 (ja) 不揮発性半導体集積回路装置
KR100267870B1 (ko) 비휘발성 반도체 메모리와 그의 구동방법 및 제조방법
JP3899601B2 (ja) 不揮発性半導体記憶装置及びその製造方法
US5151760A (en) Integrated circuit with improved capacitive coupling
EP0517353B1 (fr) Mémoire rémanente

Legal Events

Date Code Title Description
FPAY Fee payment

Year of fee payment: 4

FPAY Fee payment

Year of fee payment: 8

FPAY Fee payment

Year of fee payment: 12

AS Assignment

Owner name: SHARP KABUSHIKI KAISHA, JAPAN

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:YAMAUCHI, YOSHIMOTSI;REEL/FRAME:027528/0108

Effective date: 19960826