Semiconductor device

This application is based upon and claims the benefit of priority from Japanese Patent Application No. 2010-175693, filed Aug. 4, 2010; the entire contents of which are incorporated herein by reference.

Embodiments described herein relate generally to a semiconductor device.

Nonvolatile semiconductor memories such as flash memories are used, as semiconductor storage, in various electronic apparatuses.

In any flash memory, a voltage of about 10 to 20 V is applied to the gate electrodes of the memory cell transistors to write data. To generate a high voltage of 10 V or more, a so-called “charge pump circuit” is arranged in a peripheral circuit of the flash memory.

The charge pump circuit comprises, for example, a plurality of MOS transistors and a plurality of capacitors. Each MOS transistor has its drain connected to the gate of another MOS transistor. The capacitors of the charge pump circuit are, for example, MOS capacitors. Each MOS capacitor is provided in one semiconductor region (element region) that is surrounded by an isolation region.

The charge pump circuit needs to have a plurality of MOS transistor groups and a plurality of MOS capacitor groups, in order raise the power-supply voltage (about 3 V) to about 20 V. The charge pump circuit therefore occupies a large area. Inevitably, the area the charge pump circuit occupies tends to increase.

As a result of this, the area for the memory cell array and the area for the peripheral circuit inevitably are limited. This makes it difficult to increase the storage capacity of the flash memory, to reduce the memory chip size and to lower the bit cost.

FIG. 1 is a block diagram showing the configuration of a flash memory;

FIG. 2 is a sectional view schematically showing a cross section of the flash memory;

FIG. 3 is a sectional view showing the structure of a memory cell array;

FIG. 4 is a plan view explaining the structure of a semiconductor device according to a first embodiment;

FIG. 5 is a sectional view explaining the structure of a semiconductor device according to a first embodiment;

FIG. 6 is a sectional view explaining the structure of a semiconductor device according to a first embodiment;

FIG. 7 is an equivalent circuit diagram showing an exemplary configuration of a charge pump circuit;

FIG. 8 is a waveform diagram showing control signals for the charge pump circuit;

FIG. 9 is a plan view showing the layout of the capacitors in the charge pump circuit;

FIG. 10 is a diagram explaining the structure of one of the capacitors in the charge pump circuit;

FIG. 11A and FIG. 11B are diagrams explaining the structure and operation of the capacitors in the charge pump circuit;

FIG. 12A and FIG. 12B are other diagrams explaining the structure and operation of the capacitors in the charge pump circuit;

FIG. 13A and FIG. 13B are sectional views explaining the structure of a semiconductor device according to a second embodiment;

FIG. 14 is an equivalent circuit diagram showing an exemplary configuration of a charge pump circuit in a third embodiment;

FIG. 15 is a waveform diagram showing control signals for the charge pump circuit;

FIG. 16 is a plan view showing the layout of the capacitors in the charge pump circuit;

FIG. 17A , FIG. 17B and FIG. 17C are diagrams explaining the structure of the capacitors in the charge pump circuit;

FIG. 18 is an equivalent circuit diagram showing an exemplary configuration of the charge pump circuit in a fourth embodiment;

FIG. 19 is a waveform diagram showing control signals for the charge pump circuit;

FIG. 20 is a plan view showing the layout of the capacitors in the charge pump circuit;

FIG. 21 is a plan view explaining a modification of the semiconductor devices according to the first to fourth embodiments; and

FIG. 22 is a sectional view explaining the modification of the semiconductor devices according to the first to fourth embodiments.

Embodiments will be described in detail, with reference to the accompanying drawings. In the following description, the components of any embodiment, which are identical in function and configuration to those of any other embodiment, will be designated by the same reference numbers and will be repeatedly described, only if necessary.

In general, according to one embodiment, a semiconductor device includes at least one semiconductor region provided in a semiconductor substrate; and a capacitor group including a plurality of capacitors provided in the semiconductor region, each capacitor including a capacitor insulating film provided on the semiconductor region, a capacitor electrode provided on the capacitor insulating film, and at least one diffusion layer provided in the semiconductor region adjacent to the capacitor electrode.

(1) First Embodiment

A semiconductor device according to a first embodiment will be described with reference to FIGS. 1 to 10 , FIGS. 11A and 11B , and FIGS. 12A and 12B . The semiconductor device according to this embodiment is, for example, a flash memory.

(a) Configuration of the Flash Memory

FIG. 1 is a block diagram showing the configuration of the flash memory, which is the semiconductor device according to the first embodiment.

As shown in FIG. 1 , the flash memory 1 has a memory cell array 2 , a row decoder 3 , a voltage generation circuit 4 , and a control circuit 5 .

The memory cell array 2 includes a plurality of memory cell transistors MT, each of which can hold data. Each memory cell transistor MT has stacked gate structure including a charge storage layer and a control gate. The charge storage layer can hold a charge. The control gate is formed on the charge storage layer. The control gate of each memory cell transistor MT is connected to a word line WL. The drain of the memory cell transistor MT is electrically connected to a bit line BL. The source of the memory cell transistor MT is electrically connected to a source line SL.

In order to write data, the word line WL associated with the write address is selected, a program voltage VPGM (e.g., 20 V) is applied to the word line WL selected, and an intermediate voltage VPASS or the like is applied to the word lines WL not selected. In order to read data, the word line WL associated with the read address is selected, a read voltage VCGR is applied to the word line WL selected, and a voltage VREAD is applied to the word lines not selected. The voltages VPASS and VREAD turn on the memory cells. To erase the data, 0 V is applied to the all word lines WL, and a high voltage (e.g., 20 V) is applied to the well region in which the memory cell array 2 is formed.

The internal structure of the memory cell array 2 will be described. The flash memory according to this embodiment is, for example, a NAND flash memory. The memory cell array 2 therefore includes a plurality of NAND cells NC as shown in FIG. 1 .

Each of the NAND cells NC includes 32 memory cell transistors MT 0 to MT 31 , two select transistors ST 1 and ST 2 . The number of memory cell transistors MT included in the NAND cell is not limited to 32. The NAND cell may include 8 memory cell transistors, 16 memory cell transistors, 64 transistors, 128 transistors, 256 transistors, or the like. Hereinafter, the memory cell transistors MT 0 to MT 31 may be called “memory cell transistors MT,” on some occasions, for the sake of simplicity.

Each memory cell transistors MT has a stacked gate structure. More precisely, the memory cell transistor MT has a first gate insulating film provided on a semiconductor substrate, a charge storage layer formed on the gate insulating film, an inter-gate insulating film provided on the charge storage layer, and a control gate electrode formed on the inter-gate insulting film. Any two adjacent memory cell transistors MT share the source and drain, and are arranged with their current paths connected in series between select transistors ST 1 and ST 2 . The drain region at one end of the unit constituted by the serially connected memory cell transistors MT is connected to the source region of select transistor ST 1 , and the source region at the other end of this unit is connected to the drain region of select transistor ST 2 .

The control gate electrodes of the memory cell transistors MT arranged in the same row are connected to one of the word lines WL 0 to WL 31 . Select transistors ST 1 and ST 2 arranged in the same row have their gates connected to select gate lines SGD and SGS, respectively. In the memory cell array, the drains of the select transistors ST 1 arranged in the same column are connected to one of the bit lines BL 0 to BLn (n is a natural number). Hereinafter, the word lines WL to WL 31 and the bit lines BL 0 to BLn may be called “word lines WL” and “bit lines BL” on some occasions, for the sake of simplicity. The source of select transistors ST 2 are connected to the source line SL.

In FIG. 1 , only one NAND cell is shown. Nonetheless, the memory cell array 2 has a plurality of NAND cells arranged in the form of an array. The NAND cells arranged in the same column are connected to the same bit line BL. Data can be written altogether at a time, into the memory cell transistors MT connected to the same word line. The data unit so written is called “one page.” The data can be erased altogether from the NAND cells. The data unit so erased is called “one block.”

The row decoder 3 selects any row of the memory cell array 2 , in accordance with the address input externally.

The voltage generation circuit 4 applies a voltage to the row decoder 3 . The voltage generation circuit 4 includes a plurality of charge pump circuits 6 . The charge pump circuits 6 are provided for the word lines WL and select gate lines SGD and SGS, in one-to-one relationship. The voltage generation circuit 4 applies the voltages generated by the charge pump circuits 6 , as the above-mentioned voltages VPGM and VPASS, to the row decoder 3 . Each charge pump circuit 6 has a plurality of capacitors C.

The control circuit 5 receives an address and/or a command from an external device, and controls the other circuits provided in the flash memory 1 in accordance with the address and/or command it has received. The control circuit 5 controls, for example, the voltage generation circuit 4 , instructing the same to generate a voltage required.

Of the components of the flash memory 1 , those other than the memory cell array 2 , such as voltage generation circuit 4 and control circuit 5 , will be called “peripheral circuits” in the present embodiment.

The positional relationship the memory cell array 2 has with the peripheral circuits, as seen in the cross section of the flash memory 1 , will be explained with reference to FIG. 2 . FIG. 2 is a sectional view of the flash memory 1 , showing one region of the memory cell array 2 and one region of the peripheral circuits. The sectional view of FIG. 2 is taken along a line extending along the gates of MOS transistors. It should be noted, however, that all elements need not be so arranged as illustrated in FIG. 2 .

As shown in FIG. 2 , a p-type well region 12 is formed in the surface region of a semiconductor substrate (e.g., p-type silicon substrate) 10 , in the memory cell array 2 . P-type well region 12 is doped with, for example, boron. The impurity concentration of well region 12 is set at, for example, from 10 14 cm −3 to 10 19 cm −3 . Two n-type well regions 50 are formed on both sides of well region 12 , respectively. On one side of well region 12 , an n-type well region 50 is formed. Beneath well region 12 , an n-type well region 11 is formed, contacting well regions 12 and 50 . N-type well regions 11 and 12 therefore surround p-type well region 12 . Well region 12 is thus electrically isolated from the semiconductor substrate 10 . Since well region 12 is electrically isolated from the semiconductor substrate 10 , the potential of well region 12 can be controlled, in dependently of the semiconductor substrate 10 , in accordance with the operating mode of the memory cell array 2 . This double-well structure can serve to decrease the load on the charge pump circuits 6 , ultimately reducing the power consumption of the flash memory 1 .

Memory cell transistors MT are provided on well region 12 . These memory cell transistors MT are isolated from one another along a word line, by isolation regions STI. The isolation regions STI have a depth of, for example, about 0.1 to about 0.5 μm. The isolation regions STI have been formed by burying, for example, silicon oxide film in the trenches made in the surface of the semiconductor substrate 10 .

A potential must be uniformly applied to the entire well region 12 , in order to accomplish uniform data erasure in the plurality of memory cell transistors MT. To achieve the uniform data erasure, the bottom of well region 12 , i.e., interface between well region 12 and well region 11 , is positioned far deeper than the bottoms of the isolation regions STI. The bottom of well region 12 lies at depth of, for example, 0.4 to 0.9 μm from the surface of the semiconductor substrate 10 .

The NAND flash memory 1 uses the tunnel current flowing in a tunnel insulating film 13 to inject a charge into charge storage layers 14 and to release the charges from the charge storage layers 14 .

Data is erased at a time at the memory cell transistors to increase the number of memory cells from which data can be erased per unit time. Thus, data can be erased from a plurality of memory cells at the same time. To erase the data, a positive voltage of 15 V or more is applied to well region 12 , with respect to the semiconductor substrate 10 . The charge is thereby transferred from the charge storage layers 14 into well region 12 .

On the other hand, the voltage applied to diffusion layers 17 is lowered in order to read or write data, while maintaining the voltage of well region 12 at 0 V. The power for electrically charging or discharging well region 12 can be decreased, and the operating speed of the flash memory 1 rises. Further, in the NAND flash memory 1 according to this embodiment, data is erased or written by selectively connecting the output nodes of the charge pump circuit 6 to well region 12 or the control gate 16 (i.e., word line WL).

The structure of each NAND cell, as viewed in a cross section, will be described in detail, with reference to FIG. 3 . FIG. 3 is a sectional view taken along the bit line.

As shown in FIG. 3 , an n-type well region 11 is formed in the surface region of the p-type semiconductor substrate 10 , and a p-type well region 12 is formed in the surface region of well region 11 . On well region 12 , a gate insulating film 13 is formed. The gate insulating film 13 has been formed by using a silicon oxide film or a silicon oxynitride film that has thickness ranging from 4 nm to 12 nm. The gate insulating film 13 functions as tunnel insulating film to the memory cell transistors.

On the gate insulating film 13 , the gate electrodes of the memory cell transistors MT and select transistors ST 1 and ST 2 are formed.

The gate electrodes of the memory cell transistors MT and select transistors ST 1 and ST 2 have a charge storage layer 14 formed on the gate insulating film 13 , an inter-gate insulating film 15 formed on the charge storage layer 14 , and a conductive layer 16 formed on the inter-gate insulating film 15 .

The charge storage layer 14 is, for example, either a polycrystalline silicon layer (i.e., conductive layer) or a silicon nitride layer.

The inter-gate insulating film 15 is, for example, a silicon oxide film, a silicon oxynitride film, an ON fill or NO film (each composed of silicon oxide film and silicon nitride film), or an ONO film, or a stacked gate structure including these. Alternatively, the inter-gate insulating film 15 may be an Al 2 O 3 film, HfSi x film, AlSi x film, HfO x film or TiO x film. Still alternatively, the inter-gate insulating film 15 may be a multi-layer film composed of at least one selected form the group consisting of Al 2 O 3 film, HfSi x film, AlSi x film, HfO x film and TiO x film, a silicon oxide film, silicon nitride film, and a silicon oxynitride film.

The conductive layers 16 are conductive silicon layers, silicide layers, or two-layer layers composed of a silicon layer and a silicide layer. Examples of the silicide layers are WSi, NISi, MoSi, TiSi and CoSi. Each conductive layer 16 connects any memory cell transistors that are adjacent in the direction at right angles to the bit lines BL, and functions as a word line WL.

In select transistors ST 1 and ST 2 , one through hole is made in each inter-gate insulating film 15 . The bottom of the conductive layer 16 contacts the upper surface of the charge storage layer 14 . In select transistors ST 1 and ST 2 , each conductive layer 16 connects any transistors that are adjacent along the word line extending direction. The conductive layer 16 functions as select gate line SGD or select gate line SGS.

In those parts of the surface region of the semiconductor substrate 10 , which lie between the gate electrodes adjacent to one another, n-type impurity diffusion layers 17 are formed. Impurity diffusion layers 17 are provided for adjacent transistors. Each impurity diffusion layer 17 is used as the source (S) or drain (D) of a transistor. The diffusion layer 17 used as source or drain will be hereinafter called “source/drain diffusion layer 17 .” The source/drain diffusion layer 17 contains, for example, phosphorus, arsenic or antimony. Its surface concentration is 10 17 cm −3 to 10 20 cm −3 . The region between the source and the drain functions as channel region, in which electrons move.

On the semiconductor substrate 10 , an inter-layer insulating film 19 is formed, covering the memory cell transistors MT and select transistors ST 1 and ST 2 . The inter-layer insulating film 19 is, for example, a silicon oxide film.

In the inter-layer insulating film 19 , a contact plug CP 1 is formed. Contact plug CP 1 reaches impurity diffusion layer 17 (source S) of the select transistor ST 1 on the source side. In the surface of the inter-layer insulating film 19 , a metal interconnect layer 20 is formed. Metal interconnect layer 20 is connected to contact plug CP 1 . Metal interconnect layer 20 functions as source line SL.

In the inter-layer insulating film 19 , a contact plug CP 2 is formed. Contact plug CP 2 reaches impurity diffusion layer 17 (drain D) of the select transistor ST 2 on the drain side. In the surface of the inter-layer insulating film 19 , a metal interconnect layer 21 is formed. Metal interconnect layer 21 is connected to contact plug CP 2 .

On the inter-layer insulating film 19 , an inter-layer insulating film 22 is formed, covering metal interconnect layer 21 . In the inter-layer insulating film 22 , a contact plug CP 3 is formed, reaching metal interconnect layer 21 . On the inter-layer insulating film 22 , a metal interconnect layer 23 is formed, which is connected to contact plug CP 3 . Metal interconnect layer 23 functions as it line BL.

As shown in FIG. 2 , the elements constituting the peripheral circuits are formed on the semiconductor substrate 10 , together with the memory cells, and are arranged adjacent to the memory cell array 2 .

A MOS transistor Q 1 having a breakdown voltage (high threshold voltage) higher than that of the memory cell transistors MT is arranged on the semiconductor substrate 10 , adjacent to the memory cell array 2 .

MOS transistor Q 1 is incorporated in, for example, the row decoder 3 or the voltage generation circuit 4 . As shown in FIG. 2 , MOS transistor Q 1 has a gate insulating film 42 , which is thicker than the gate insulating film (tunnel insulating film) 13 of each memory cell transistor MT.

MOS transistor Q 1 has a gate electrode 34 on the gate insulating film 42 . The gate electrode 34 has a stacked gate structure essentially identical to that of the memory cell transistors. The gate electrode 34 of MOS transistor Q 1 has a first conductive layer 43 formed on the gate insulating film 42 , an inter-gate insulating film 44 formed on the first conductive layer 43 , and a second conductive layer 45 formed on the inter-gate insulating film 44 . A source/drain diffusion layer 32 is formed in a well region 51 . The inter-gate insulating film 44 has a through hole. The bottom of the second conductive layer 45 contacts the top of the first conductive layer 43 . MOS transistor Q 1 has been formed, together with the memory cell transistors MT, practically in the same manufacturing step.

The flash memory 1 needs the voltage generation circuit 4 that generates a high voltage from a low voltage. The voltage generation circuit 4 is configured to generate a potential difference ranging from 10 to 15 V or more, with the potential of the semiconductor substrate 10 used as reference. This is why the voltage generation circuit 4 has voltage boosting circuits (e.g., charge pump circuits 6 ).

Each charge pump circuit 6 includes a plurality of capacitors C and a plurality of transistors. The capacitors C accumulate charges. The transistors control the charges accumulated in the respective capacitors. MOS transistors Q 3 are shown in FIG. 2 , which are incorporated in a charge pump circuit 6 . Like MOS transistor Q 1 , MOS transistors Q 3 have been formed, together with the memory cell transistors MT, practically in the same manufacturing step.

MOS transistors Q 3 are high-breakdown voltage transistors, each having a gate insulating film 42 thicker than the gate insulating film of any memory cell transistor MT, and can therefore withstand potential differences of 15 V or more.

The gate electrode 34 of each MOS transistor Q 3 has a first conductive layer 43 formed on the gate insulating film 42 , an inter-gate insulating film 44 formed on the first conductive layer 43 , and a second conductive layer 45 formed on the inter-gate insulating film 44 . A source/drain diffusion layer 32 is formed in a well region 51 .

In the inter-gate insulating film 44 , a through hole is made. The bottom of the second conductive layer 45 contacts the top of the first conductive layer 43 .

In addition to the high-breakdown voltage transistors Q 1 and Q 3 , MOS transistors (hereinafter called “low-breakdown voltage MOS transistors) having a breakdown voltage lower than that of MOS transistors Q 1 and Q 3 are provided on the substrate 10 , together with the memory cell transistors MT. The low-breakdown voltage MOS transistors have a stacked gate structure that is essentially identical to that of the high-breakdown voltage transistors Q 1 . The low-breakdown voltage MOS transistors have a gate insulating film thicker than the gate insulating film 13 of MOS transistors MT and thinner than the gate insulating film 42 of the high-breakdown voltage transistors Q 1 .

The capacitors C are formed on the isolation region (capacitor area) AA provided in the semiconductor substrate 10 . Each capacitor C has an insulating film 42 R having almost the same thickness as the gate insulating film 42 , an electrode 34 A having almost the same structure as the gate electrode 34 , and an impurity diffusion layer 32 A. The electrode 34 A of the capacitor C is provided above well region 51 formed in the surface of the semiconductor substrate 10 (more precisely, capacitor area AA), with the insulating film 42 A interposed between it and well region 51 . As shown in FIG. 2 , the capacitors C are provided p-type well region 51 . Instead, they may be provided in that part of the p-type semiconductor substrate 10 , in which no well regions are formed.

At the boundaries between the regions in which the memory cell array 2 , MOS transistor Q 1 , MOS transistor Q 3 and capacitors C are formed, respectively, isolation regions STI are formed. Beneath the isolation regions STI, well regions 36 having the same conductivity type (here, p type) as the semiconductor substrate 10 are formed as channel stoppers. Further, p-type regions 37 are formed to prevent a punch-through leakage current from flowing at the bottoms of the isolation regions STI. Well regions 36 serving as channel stoppers, and p-type regions 37 may not be formed in some cases.

An n-type well region (not shown) is formed, reaching the surface of the semiconductor substrate 10 . This n-type well region is used as a region in which n-channel MOS transistors may be formed.

The basic configuration the capacitors, which are incorporated in the semiconductor device (e.g., flash memory) according to this embodiment, will be described with reference to FIG. 4 , FIG. 5 and FIG. 6 .

FIG. 4 is a plan view showing capacitors Cm and Cn included in the semiconductor device according to this embodiment. FIG. 5 is a sectional view taken along line V-V in FIG. 4 , showing capacitors Cm and Cn. FIG. 6 is a sectional view taken along line VI-VI in FIG. 4 , showing capacitors Cm and Cn, too. In FIG. 5 , the broken lines indicate two members that exist in front and at the back of the plane of the cross section, respectively, in the depth direction of the drawing.

As shown in FIG. 4 to FIG. 6 , an isolation region STI is provided in the semiconductor substrate 10 . In the isolation regions STI, an isolation insulating film is buried. The isolation region STI defines one semiconductor region (active region) AAC in the semiconductor substrate 10 .

The semiconductor region AAC is a p- or n-type well region, or a semiconductor substrate (e.g., p-type silicon substrate) having no well regions.

In the semiconductor device according to this embodiment, at least two capacitors Cm and Cn are provided in one semiconductor region AAC. Capacitors Cm and Cn provided in the same semiconductor region AAC make one capacitor group GA. Hereinafter, the semiconductor region in which the capacitor group GA is provided may be called the “capacitor formation region (capacitor area),” on some occasions.

Capacitors Cm and Cn are MOS capacitors. Each of these capacitors Cm and Cn has a gate insulating film 42 A provided on the surface of the semiconductor region AAC, and gate electrodes 34 Am and 34 An provided on the gate insulating film 42 A. Hereinafter, the gate insulating film 42 A used as a part of capacitors Cm and Cn will be called “capacitor insulating film 42 A,” and the gate electrodes gate electrodes 34 Am and 34 An used as elements constituting capacitors Cm and Cn will be called “capacitor electrodes 34 Am and 34 An.” If the MOS capacitors Cm and Cn are formed on p-type well regions or on a p-type semiconductor substrate, channels are formed beneath capacitor electrodes 34 An and 34 Am when the MOS capacitors are driven (or when a positive voltage is applied to capacitor electrodes 34 An and 34 Am).

The capacitor insulating film 42 A of the capacitor C is thicker than the gate insulating film (tunnel insulating film) of each memory cell transistor MT so that the capacitor C may withstand a potential difference of 15 V or more, which it may have with respect to the potential of the semiconductor region ARC. The capacitor insulating film 42 A may be formed at the same as, for example, the gate insulating film 42 of MOS transistor Q 3 . In this case, the capacitor insulating film 42 A has the same thickness as the gate insulating film 42 of MOS transistor Q 3 . The material of the capacitor insulating film 42 A is identical to that of the gate insulating film 42 . Nonetheless, the capacitor insulating film 42 A may be formed in a step different from the step of forming the gate insulating film 42 . Hence, the capacitor insulating film 42 A may differ from the gate insulating film 42 in terms of thickness or material. Moreover, the capacitor insulating film 42 A may have a stacked gate structure.

Capacitor electrode 34 Am of capacitor Cm and capacitor electrode 34 An of capacitor Cn are formed, practically at the same time as, for example, the stacked gate electrodes of the memory cell transistors MT is formed. Therefore, capacitor electrodes 34 Am and 34 An of capacitors Cm and Cn, respectively, have a conductive layer 43 A, a conductive layer 45 A, and an inter-gate insulating film 44 A interposed between two conductive layers 43 A and 45 A. As in MOS transistor Q 3 , the inter-gate insulating film 44 A has, for example, a through hole, and the bottom of the second conductive layer 45 A contacts the top of the first conductive layer 43 a via the through hole.

Capacitor electrodes 34 Am and 34 An are arranged adjacent to each other in one semiconductor region (capacitor formation region) AAC.

On capacitor electrodes 34 Am and 34 An, contact plugs CPEm and CPEn are provided, respectively. It suffices to provide at least one contact plug CPEm to be connected to one capacitor electrode 34 Am, and at least one contact plug CPEn to be connected one capacitor electrode 34 An.

Contact plug CPEm connects a metal interconnect layer 29 m to capacitor electrode 34 Am, and contact plug CPEn connects a metal interconnect layer 29 n to capacitor electrode 34 An. Metal interconnect layers 29 m and 29 n are electrically isolated from each other with respect to each capacitor Cm and Cn, respectively. The potentials of capacitor electrodes 34 Am and 34 An of capacitors Cm and Cn are therefore controlled, independently of each other. While the circuit including the capacitor group GA is operating, potentials Vm and Vn of different values are applied from metal interconnect layers 29 m and 29 n to capacitor electrodes 34 Am and 34 An, respectively.

The capacitor group GA has, for example, the one diffusion layer 32 A surrounding the capacitor electrodes Cm and Cn Impurity diffusion layer 32 A is provided in the surface region of the semiconductor region AAC. Impurity diffusion layer 32 A is formed in the semiconductor region (capacitor formation region) AAC, surrounding, for example, capacitor electrodes 34 Am and 34 An. Impurity diffusion layer 32 A is continuous in the capacitor formation region AAC. Impurity diffusion layer 32 A is shared by capacitors Cm and Cn in the semiconductor region AAC. Impurity diffusion layer 32 A in the capacitor formation region AAC is provided to surround the capacitor electrodes Cm, Cn.

Impurity diffusion layer 32 A has been formed at the same time as, for example, the source/drain diffusion layers 32 of transistors MT, Q 1 and Q 3 . In this case, impurity diffusion layer 32 A provided in the semiconductor region AAC has the same impurity concentration as, for example, the source/drain diffusion layers 32 .

At least one contact plug CPA is provided on impurity diffusion layer 32 A provided in the semiconductor region AAC. Through contact plug CPA, potential Vaa is applied to impurity diffusion layer 32 A.

In this embodiment, potential Vaa applied to impurity diffusion layer 32 A controls the potential level at the one end (i.e., substrate or well) of any capacitor (Cm or Cn) provided in one semiconductor region AAC. That is, channels are formed beneath capacitor electrodes 34 An and 34 Am when the MOS capacitors are driven, and the potential level at one end of capacitor Cm and the potential level at one end of capacitor Cn are commonly controlled through the these channels. The potential level at the other end (electrode) of capacitor Cm and the potential level at the other end (electrode) of capacitor Cn are independently controlled, by potentials Vm and Vn applied to metal interconnect layers 29 A and 196 spaced apart from each other. To the semiconductor region AAC, ground potential (0 V) is applied in order to stabilize the operation of the semiconductor device.

Capacitors Cm and Cn, for example, are formed practically in the same manufacturing step, along with the memory cell transistors and the MOS transistors. Thus, capacitors Cm and On can be formed without increasing the number of manufacturing steps of the semiconductor device.

Capacitors Cm and Cn shown in FIG. 3 to FIG. 5 are used in, for example, the charge pump circuit 6 .

As described above, a plurality of capacitors Cm and Cn are provided in the semiconductor region (capacitor area) AAC in the semiconductor device according to this embodiment. In the present embodiment, these capacitors Cm and Cn provided in the semiconductor region AAC are commonly controlled in term of the potential at the diffusion layer 32 A that is used as one end of each capacitor.

The potential level of the diffusion layer 32 A used as one end of each capacitor may be controlled to different values for the respective capacitor groups GA. In this case, the capacitor groups GA must be electrically isolated from one another, in order to stabilize the operation of the semiconductor device. To stabilize the operation of the device, each capacitor group GA is provided in one semiconductor region surrounded by an isolation region.

Moreover, when a capacitor (MOS capacitor) is driven, the channel formed beneath the capacitor electrode extends, in some cases, to the adjacent semiconductor region formed for another capacitor via the bottom of the isolation insulating film. This extension of channel may possibly cause the adjacent capacitors to malfunction, depending upon the state in which the adjacent capacitors are driven. In order to suppress such malfunction, the isolation regions are elongated in the adjacent capacitor formation regions, in the direction parallel to the surface of the semiconductor substrate. Furthermore, channel stoppers (well regions 36 and 37 ) are formed at the bottom of the isolation insulating film provided in the isolation region. In order to provide margins for forming the channel stoppers, the isolation region (isolation insulating film) inevitably has a still large size.

In the semiconductor device according to this embodiment, a plurality of capacitors Cm and Cn are provided in the same semiconductor region AAC. Formed in the same semiconductor region AAC, one ends of capacitors Cm and Cn are controlled by the same potential levels when a potential is applied to the semiconductor region AAC.

While capacitors Cm and Cn provided in one semiconductor region AAC are being driven, or while channels (inversion layers) are being formed beneath capacitor electrodes 34 Am and 34 An, the diffusion layer 32 A, i.e., ends of capacitors Cm and Cn, are controlled through the channels at the same potential level. In the semiconductor region AAC, a channel is formed immediately below the capacitor insulating film 42 A. Beneath this channel, a depletion layer is formed. Nonetheless, the capacitance of the depletion layer is negligibly small as compared with the capacitance of the gate insulating film 42 A (i.e., capacitance of the capacitor composed of the capacitor electrodes 34 A, gate insulating film 42 A and the channel). As a result, the depletion will scarcely influence capacitors Cm and Cn, though capacitors Cm and Cn are arranged adjacent to each other, in one semiconductor region AAC.

Thus, neither capacitors Cm and Cn controlled by the same potential level at the one ends nor the capacitor group GA including these capacitors Cm and Cn is practically degraded though capacitors Cm and Cn are provided in the same semiconductor region AAC, in the semiconductor device according to this embodiment.

Note that capacitor electrodes of capacitors Cm and Cn formed in one semiconductor region RAC have their potentials controlled independently. Capacitor electrodes 34 Am and 34 An of capacitors Cm and Cn must, therefore, be electrically isolated. Nevertheless, the potential difference between capacitor electrodes 34 Am and 34 An can be adjusted to a relatively small value by adjusting the arrangement of capacitors Cm and Cn, though different potentials are applied to capacitor electrodes 34 Am and 34 An while capacitors Cm and Cn are being driven. In this embodiment, the potential difference between capacitor electrodes 34 Am and 34 An can be about 5 V at most if the distance (gap) between capacitor electrodes 34 Am and 34 An is set to, for example, about 0.2 to 1.0 μm. The distance between capacitor electrodes 34 Am and 34 An can indeed be rendered shorter than the width (e.g., about 2 to 5 μm) of the isolation region isolating adjacent two semiconductor regions AAC.

As specified already, a plurality of capacitors Cm and Cn can be provided in the same semiconductor region AAC, without degrading the operating characteristic of the semiconductor device and without the necessity of electrically isolating them by isolation regions, only if capacitors Cm and Cn can be commonly controlled in terms of potential level at one end (semiconductor region).

Hence, the area occupied by the isolation regions for electrically isolating the capacitors from one another can be reduced.

Thus, with the semiconductor device according to the first embodiment, it is possible to reduce the area for the regions in which capacitors are formed.

(b) Configuration of the Charge Pump Circuits in The Voltage Generation Circuit

Some cases where the capacitors provided in one semiconductor region ACC are used in any charge pump circuit 6 incorporated in the voltage generation circuit 4 will be explained with reference to FIG. 7 , FIG. 8 , FIG. 9 , FIG. 10 , FIG. 11A , FIG. 11B , FIG. 12A and FIG. 12B .

<Circuit Configuration>

First, the configuration of each charge pump circuit 6 incorporated in the voltage generation circuit 4 will be described with reference to FIG. 7 . FIG. 7 is an equivalent circuit diagram showing an exemplary configuration of the charge pump circuit 6 .

As FIG. 7 shows, the charge pump circuit 6 has nine n-channel MOS transistors TR 1 to TR 9 and eight capacitors C 1 to C 8 . The number of MOS transistors is not limited to nine. Nor is the number of capacitors limited to eight. The charge pump circuit 6 only needs to have at least three capacitors. In the following description, MOS transistors TR 1 to TR 9 and capacitors C 1 to C 8 will be called merely as “MOS transistors TR” and “capacitors C,” respectively, unless they need to be distinguished from one another.

MOS transistor TR 1 has its source or drain connected to its gate and also to the power-supply potential Vdd.

MOS transistor TR 2 has one of its source and drain connected to its gate (hereinafter, the resultant connection node will be referred to as “node N 1 ”). Further, the one of its source and drain is connected to the other of the source and drain of MOS transistor TR 1 .

MOS transistor TR 3 has one of its source and drain connected to its gate (hereinafter, the resultant connection node will be referred to as “node N 2 ”). Further, the one of its source and drain is connected to the other of the source and drain of MOS transistor TR 2 .

MOS transistor TR 4 has one of its source and drain connected to its gate (hereinafter, the resultant connection node will be referred to as “node N 3 ”). Further, the one of its source and drain is connected to the other of the source and drain of MOS transistor TR 3 .

MOS transistor TR 5 has one of its source and drain connected to its gate (hereinafter, the resultant connection node will be referred to as “node N 4 ”). Further, the one of its source and drain is connected to the other of the source and drain of MOS transistor TR 4 .

MOS transistor TR 6 has one of its source and drain connected to its gate (hereinafter, the resultant connection node will be referred to as “node N 5 ”). Further, the one of its source and drain is connected to the other of the source and drain of MOS transistor TR 5 .

MOS transistor TR 7 has one of its source and drain connected to its gate (hereinafter, the resultant connection node will be referred to as “node N 6 ”). Further, the one of its source and drain is connected to the other of the source and drain of MOS transistor TR 6 .

MOS transistor TR 8 has one of its source and drain connected to its gate (hereinafter, the resultant connection node will be referred to as “node N 7 ”). Further, the one of its source and drain is connected to the other of the source and drain of MOS transistor TR 7 .

MOS transistor TR 9 has one of its source and drain connected to its gate (hereinafter, the resultant connection node will be referred to as “node N 8 ”). Further, the one of its source and drain is connected to the other of the source and drain of MOS transistor TR 8 .

The other of the source and drain of MOS transistor TR 9 (hereinafter referred to as “node N 9 ”) serves as output node for output voltage Vout of the charge pump circuit 6 .

As specified above, each MOS transistor TR works as a rectification element, the anode of which is one of the source and drain of each MOS transistor TR and the cathode of which is the other of the source and drain. In other words, the MOS transistor TR incorporated in the charge pump circuit 6 is a diode-connected transistor.

Capacitors C 1 , C 3 , C 5 and C 7 are connected, at one electrode, to a signal line for a control (clock) signal Φ 2 . At the other electrode, capacitors C 1 , C 3 , C 5 and C 7 are connected to nodes N 1 , N 3 , N 5 and N 7 , respectively. On the other hand, capacitors C 2 , C 4 , C 6 and C 8 are connected, at one electrode, to a signal line for a control (clock) signal /Φ 2 . At the other electrode, capacitors C 2 , C 4 , C 6 and C 8 are connected to nodes N 2 , N 4 , N 6 and N 8 , respectively. As shown in FIG. 8 , clock signal /Φ 2 has been generated by inverting clock signal Φ 2 .

The charge pump circuit 6 includes a plurality of rectification elements connected in series. Clock signal Φ 2 is supplied to the input nodes of the even-numbered rectification elements (i.e., MOS transistors TR 2 , TR 4 , TR 6 , . . . ) through the capacitors Cj (j=1, 3, 5, . . . ). Clock signal /Φ 2 is supplied to the input nodes of the odd-numbered rectification elements, excluding MOS transistor TR 1 (i.e., MOS transistors TR 3 , TR 5 , . . . ) through the capacitors C(j+1).

As shown in FIG. 8 , clocks Φ 2 and /Φ 2 are clock signals that can assume voltage Vc and ground potential (GND). Clocks Φ 2 and /Φ 2 , used as control signals, are two phase clock signals that never overlap in terms of generation timing of voltage Vcc, and therefore provide some time, for which any adjacent MOS transistors TR 1 to TR 9 , serially connected can be disconnected from one another. This can prevent the charge from being applied backwards from node N 9 toward node N 1 .

A voltage of 3 to 10 V is used as voltage Vcc, whereby the output voltage Vout can be a high voltage of 10 V or more. The output voltage Vout is used as, for example, voltage VPGM or voltage PASS. MOS transistors Tri have threshold voltages Vth 1 , Vth 2 , . . . , Vthi (i is an integer ranging from 1 to 9), respectively. The power-supply voltage Vdd is, for example, 1 to 5 V. In order to decrease the power-supply impedance, and to suppress changes in the output voltage, it is desired that the power-supply voltage Vdd for the charge pump circuit should be equal to the power-supply voltage of the NAND flash memory 1 .

With regard to the charge pump circuit 6 shown in FIG. 7 , the capacitors Cj (j=1, 3, 5, . . . ) to which clock Φ 2 is input shall be called “odd-numbered capacitor group G odd ,” and the capacitors C(j+1) (j=1, 3, 5, . . . ) to which clock /Φ 2 is input shall be called “even-numbered capacitor group G even .”

The structure of the capacitors incorporated in the charge pump circuit 6 of FIG. 7 will be described with reference to FIG. 9 , FIG. 10 , FIG. 11A and FIG. 11B . FIG. 9 is a plan view showing the layout of the capacitor formation regions (capacitor areas). FIG. 10 is a sectional view taken along line X-X shown in FIG. 9 . FIG. 11A is a sectional view taken along line XIA-XIA shown in FIG. 9 , and FIG. 11B is a sectional view taken along line XIB-XIB shown in FIG. 9 . In FIG. 11A and FIG. 11B , the inter-layer insulating film is not illustrated. Further, neither the contact plugs nor the interconnect layers are shown in detail in FIG. 11A and FIG. 11B .

As shown in FIG. 9 , FIG. 10 , FIG. 11A and FIG. 11B , the odd-numbered capacitor group G odd and the even-numbered capacitor group G even are provided in, for example, semiconductor regions AAC-o and AAC-e of the semiconductor substrate (e.g., p-type silicon substrate) 10 , respectively. In semiconductor regions AAC-o and AAC-e, p-type well regions 51 p - e and 51 p - o , for example, are provided. P-type well regions 51 p - e and 51 p - o may not be formed, as shown in FIGS. 5 and 6 .

In semiconductor regions AAC-o and AAC-e, diffusion layers 32 A-A and 32 A-B are provided, surrounding capacitor electrodes 34 A 1 , 34 A 3 , 34 A 5 and 34 A 7 of the odd-numbered capacitor group G odd and capacitor electrodes 34 A 2 , 34 A 4 , 34 A 6 and 34 A 8 of the even-numbered capacitor group G even , respectively. Diffusion layers 32 A-A and 32 A-B provided in semiconductor regions AAC-o and AAC-e are respectively shared by capacitors C 1 , C 3 , C 5 and C 7 of capacitor group G odd and capacitors C 2 , C 4 , C 6 and C 8 of capacitor group G even . Diffusion layer 32 A-A is continuous in semiconductor region AAC-o. Similarly, diffusion layer 32 A-B is continuous in semiconductor region AAC-e.

Clocks Φ 2 and /Φ 2 are applied to the odd-numbered capacitor group G odd and the even-numbered capacitor group G even through contact plugs CPA-o and CPA-e provided in semiconductor regions AAC-o and AAC-e, respectively.

Clocks Φ 2 and /Φ 2 are control signals that rise to voltage Vcc at different times as shown in FIG. 8 . Thus, the two well regions 51 p - e and 51 p - o are controlled at different times in terms of potential level. This is why two diffusion layers 32 A-A and 32 A-B are electrically isolated from each other, by the isolation region STI.

The odd-numbered capacitor group G odd (i.e., capacitors C 1 , C 3 , C 5 and C 7 ) is arranged in p-type well region 51 p - o in which clock Φ 2 is applied to diffusion layer 32 A-A. The even-numbered capacitor group G even (i.e., capacitors C 2 , C 4 , C 6 and C 8 ) is arranged in p-type well region 51 p - e in which clock /Φ 2 is applied to diffusion layer 32 A-B. The odd-numbered capacitor group G odd and even-numbered capacitor group G ever , are thus provided in semiconductor regions AAC-o and AAC-e (well regions), and electrically isolated from each other, respectively. The adjacent semiconductor regions AAC-o and AAC-e are spaced apart by the isolation insulating film embedded in the semiconductor substrate 10 .

In the odd-numbered capacitor group G odd , channels are formed beneath capacitor electrodes 34 A 1 , 34 A 3 , 34 A 5 and 34 A 7 , respectively, when capacitors C 1 , C 3 , C 5 and C 7 are driven. In the even-numbered capacitor group G even , channels are formed beneath capacitor electrodes 34 A 2 , 34 A 4 , 34 A 6 and 34 A 8 , respectively, when capacitors C 2 , C 4 , C 6 and C 8 are driven.

As shown in the plan view of FIG. 9 , diffusion layer 32 A-A for capacitors C 1 , C 3 , C 5 and C 7 is continuous in the whole semiconductor region 51 p - o , surrounding capacitor electrodes 34 A 1 , 34 A 3 , 34 A 5 and 34 A 7 , respectively. Therefore, the diffusion layer 32 A-A has a common potential, the one ends of capacitors C 1 , C 3 , C 5 and 07 are commonly controlled to one another by diffusion layer 32 A-A even if the odd-numbered capacitor group G odd is not driven. Similarly, diffusion layer 32 A-B for capacitors C 2 , C 4 , C 6 and C 8 is continuous in the whole semiconductor region 51 p - e , surrounding capacitor electrodes 34 A 2 , 34 A 4 , 34 A 6 and 34 A 8 , respectively. The diffusion layer 32 A-B has a common potential, the one ends of capacitors C 2 , C 4 , C 6 and C 8 are commonly controlled to one another by diffusion layer 32 A-B even if the even-numbered capacitor group G even is not driven.

Diffusion layer 32 A-A for capacitors C 1 , C 3 , C 5 and C 7 and diffusion layer 32 A-B for capacitors C 2 , C 4 , C 6 and C 8 may not be continuous, respectively in semiconductor regions AAC-o and AAC-e. Instead, diffusion layer 32 A-A may be divided into segments by capacitor electrodes 34 A 1 , 34 A 3 , 34 A 5 and 34 A 7 , and diffusion layer 32 A-B may be divided into segments by capacitor electrodes 34 A 2 , 34 A 4 , 34 A 6 and 34 A 8 . This is because the segments of diffusion layer 32 A-A are electrically connected together by the formed channels when capacitors C 1 , C 3 , C 5 and C 7 of the odd-numbered capacitor group G odd are driven, and the segments of diffusion layer 32 A-B are electrically connected together by the formed channels when capacitors C 2 , C 4 , C 6 and C 8 of the even-numbered capacitor group G even are driven.

Contact plugs CPA-o and CPA-e are electrically connected to diffusion layers 32 A-A and 32 A- 2 , respectively, which are provided in p-type well regions 51 p - o and 51 p - e . Therefore, clock Φ 2 is applied to capacitors C 1 , C 3 , C 5 and C 7 of the odd-numbered capacitor group G odd , at one end, through contact plug CPA-o. Similarly, clock /Φ 2 is applied to capacitors C 2 , C 4 , C 6 and C 8 of the even-numbered capacitor group G even , at one end, through contact plug CPA-e.

As shown in FIG. 9 , two contact plugs CPA-o are provided in semiconductor region AAC-o, respectively at two ends thereof. Similarly, two contact plugs CPA-e are provided in semiconductor region AAC-e, respectively at two ends thereof. The number and layout of contact plugs CPA-o provided in semiconductor region AAC-o is not limited to two. Nor are the number and layout of contact plugs CPA-e provided in semiconductor region AAC-e.

On capacitor electrode 34 A 1 of capacitor C 1 , a contact plug CPE 1 is provided. Contact plug CPE 1 is connected to a metal interconnect layer 29 1 above diffusion layer 32 A-A. Capacitor electrode 34 A 1 is connected to node N 1 of transistor TR 2 by metal interconnect layer 29 1 .

On capacitor electrode 34 A 2 of capacitor C 2 , a contact plug CPE 2 is provided. Contact plug CPE 2 is connected to a metal interconnect layer 29 2 above diffusion layer 32 A-B. Capacitor electrode 34 A 2 is connected to node N 2 of transistor TR 3 by metal interconnect layer 29 2 .

On capacitor electrode 34 A 3 of capacitor C 3 , a contact plug CPE 3 is provided. Contact plug CPE 3 is connected to a metal interconnect layer 29 3 above diffusion layer 32 A-B. Capacitor electrode 34 A 3 is connected to node N 3 of transistor TR 4 by metal interconnect layer 29 3 .

On capacitor electrode 34 A 4 of capacitor C 4 , a contact plug CPE 4 is provided. Contact plug CPE 4 is connected to a metal interconnect layer 29 4 above diffusion layer 32 A-B. Capacitor electrode 34 A 4 is connected to node N 4 of transistor TR 5 by metal interconnect layer 29 4 .

On capacitor electrode 34 A 5 of capacitor C 5 , a contact plug CPE 5 is provided. Contact plug CPE 5 is connected to a metal interconnect layer 29 5 above diffusion layer 32 A-B. Capacitor electrode 34 A 5 is connected to node N 5 of transistor TR 6 by metal interconnect layer 29 5 .

On capacitor electrode 34 A 6 of capacitor C 6 , a contact plug CPE 6 is provided. Contact plug CPE 6 is connected to a metal interconnect layer 29 6 above diffusion layer 32 A-B. Capacitor electrode 34 A 6 is connected to node N 6 of transistor TR 7 by metal interconnect layer 29 6 .

On capacitor electrode 34 A 7 of capacitor C 7 , a contact plug CPE 7 is provided. Contact plug CPE 7 is connected to a metal interconnect layer 29 7 above diffusion layer 32 A-B. Capacitor electrode 34 A 7 is connected to node N 7 of transistor TR 8 by metal interconnect layer 29 7 .

On capacitor electrode 34 A 8 of capacitor C 8 , a contact plug CPE 8 is provided. Contact plug CPE 8 is connected to a metal line layer 29 8 above diffusion layer 32 A-B. Capacitor electrode 34 A 8 is connected to node N 8 of transistor TR 9 by metal interconnect layer 29 8 .

Capacitors C 1 to C 8 are so connected, at the other end, to nodes N 1 to N 8 of transistors TR 2 to TR 9 , respectively.

Metal interconnect layers 29 1 to 29 8 , which are connected to capacitor electrode electrodes 34 A 1 to 34 A 8 , are electrically isolated from one another. Capacitor electrode electrodes 34 A 1 to 34 A 8 , of capacitors C 1 to C 8 are connected to nodes N 1 to N 8 of different transistors TR 2 to TR 9 , respectively.

As will be described later, nodes N 1 to N 8 of transistors TR 2 to TR 9 have different potential levels, each level raised by the transistor immediately preceding the transistor having the node. That is, of any adjacent two nodes, the node nearer to the output side (i.e., transistor TR 9 ) has a higher voltage (potential) than the other node. Hence, different potentials V 1 to V 8 are applied to capacitor electrodes 34 A 1 to 34 A 8 through metal interconnect layers 29 1 to 29 8 that are electrically isolated from one another. Of potentials V 1 to V 8 applied to capacitor electrode 34 A 1 to 34 A 8 , each is higher than the one applied to the immediately preceding capacitor electrode.

While potentials V 1 to V g are applied to nodes N 1 to N 8 to which capacitors C 1 to C 8 are connected, channels 90 are formed in p-type well regions 51 p - o and 51 p - e that lie beneath capacitor electrodes 34 A 1 to 34 A 8 .

In the odd-numbered capacitor group G odd and even-numbered capacitor group G even , the interval (first distance) D 1 at which capacitor electrodes 34 A 1 to 34 A 8 are arranged is smaller than the width (second distance) D 2 of the isolation region STI.

At the bottom of the isolation region STI, channel stoppers 36 and 37 are provided. The channel stoppers 36 and 37 prevent a channel being formed between semiconductor regions AAC-o and AAC-e which are applied with clocks Φ 2 and /Φ 2 , respectively, through the bottom of the isolation region STI.

<Operation>

How the charge pump circuit 6 operates will be explained with reference to FIG. 7 , FIG. 8 , FIG. 9 , FIG. 10 , FIG. 11A , FIG. 11B , FIG. 12A and FIG. 12B . FIG. 11A , FIG. 11S , FIG. 12A and FIG. 12B are sectional view schematically showing the states that capacitors C 1 to C 8 take while the charge pump circuit 6 is operating.

Assume that clock Φ 2 is at ground potential GND (e.g., 0 V) and MOS transistor TR 2 is in non-conductive state. Capacitors C 1 to C 8 take the state shown in FIG. 11A and FIG. 11B . If MOS transistor TR 1 is rendered conductive in this state, node N 1 is charged to a voltage that the threshold voltage Vth 1 of MOS transistor TR 1 is reduced from the power-supply voltage Vdd, that is, (Vdd−Vth 1 ). At this point, the potential of node N 1 (Vdd−Vth 1 =V 1 ) is applied to capacitor electrode 34 A 1 , and 0 V is applied to diffusion layer 32 A-A. As a result, a channel 90 is formed as shown in FIG. 11A , beneath capacitor electrode 34 A 1 of capacitor C 1 at the first stage (odd-numbered stage).

Clock Φ 2 is raised from GND to Vcc (e.g., 4 V). That is, the potential level of well regions 51 p - o is changed from 0 to 4 V. Clock Φ 2 is a signal for controlling the potential level at one end of each of capacitors C 1 , C 3 , C 5 and C 7 of the odd-numbered capacitor group G odd . In the charge pump circuit of this embodiment, capacitors C 1 , C 3 , C 5 and C 7 of the odd-numbered capacitor group G odd share diffusion layer 32 A-A. Capacitors C 1 , C 3 , C 5 and C 7 are therefore applied, at one end, with voltage of 4 V.

The capacitive coupling resulting from the Vcc application raises potential V 1 of node N 1 by αVcc (up to about 4 V). As a result, potential V 1 of node N 1 increases to (Vdd−Vth 1 +αVcc) at most. The amplitude of the voltage at node N 1 decreases to, for example, αVcc because the capacitance ratio between capacitor C 1 and MOS transistors TR 1 and TR 2 . Note that α is a value equivalent to boost-up ratio and ranges, for example, from 0.7 to 1.

Also in this period, the channel 90 is formed beneath capacitor electrode 34 A 1 of capacitor C 1 of the odd-numbered capacitor group G odd . As the channel 90 is formed, a depletion layer is formed in well region (semiconductor region) 51 p - o , at a position below capacitor electrode 34 A 1 of capacitor C 1 (i.e., in the region beneath the channel 90 ). The serial capacitance defined by the capacitance of the depletion layer and the capacitance of the oxide film will not degrade the characteristic (e.g., electrostatic capacitance) of capacitor C 1 .

The odd-numbered capacitor group G odd and the even-numbered capacitor group G even are electrically isolated from each other. Therefore, the depletion layer of capacitor C 1 imposes no adverse influence on capacitors C 2 , C 4 , C 6 and 08 of the even-numbered capacitor group G even .

As clock Φ 2 rises from GND to Vcc, clock /Φ 2 does fall from Vcc to GND. The potential level of well region 51 p - o in which the even-numbered capacitor group G even is provided falls from 4 to 0 V. While potential V 2 of node N 1 is rising, 0 V is applied to one end of each of capacitors C 2 , C 4 , C 6 and C 8 of the even-numbered capacitor group G even .

While the charge pump circuit is being driven, the depletion layer of capacitor C 1 may impose adverse influence on capacitors C 2 , C 4 , C 6 and C 8 of the even-numbered capacitor group G ever , (or on the potential of well region 51 p - o ). Assume that the odd-numbered capacitor group G odd and the even-numbered capacitor group G even , are electrically isolated by an impurity region (e.g., high-concentration p-type impurity layer), without use of the isolation insulating film. Then, the ground potential is applied to the high-concentration p-type impurity layer. Inevitably, a leakage current will be generated due to punch-through caused by the potential difference between diffusion layer 32 A-A, 32 A-B and the impurity layer. Furthermore, the high-concentration p-type impurity layer suppresses the formation of channels in the MOS capacitors. Consequently, the charge pump circuit 6 may fail to perform the prescribed operation. In view of this, the electrical isolation of the well regions (i.e., diffusion layers 32 A-A and 32 A-B, in which the odd-numbered capacitor group G odd and even-numbered capacitor group G even are provided), should better be electrically isolated by the isolation region STI, not by an impurity region such as a high-concentration p-type impurity layer.

While clock Φ 2 remains at Vcc, MOS transistor TR 2 remains conductive. Since clock /Φ 2 is set to GND, 0 V is applied to one end of capacitor C 2 (at even-numbered stage), next to capacitor C 1 (at odd-numbered stage). In this case, capacitor C 1 at an odd-numbered stage is at high potential, whereas capacitor C 2 at an even-numbered stage is at low potential. A charge therefore moves between these capacitors C 1 and C 2 via MOS transistor TR 2 now in conductive state. As a result, potential V 1 of node N 1 is transferred to node N 2 until potential V 2 of node N 2 rises to {(potential V 1 of node N 1 )−Vth 2 }. Potential V 2 of node N 2 further rises to almost {{Vdd−Vth 1 +ΔVcc}−Vth 2 }. In capacitor C 2 connected to node N 2 , a potential {{Vdd−Vth 1 +ΔVcc}−Vth 2 } is applied to capacitor electrode 34 A 2 . Thus, potential V 2 of node N 2 forms a channel 90 beneath capacitor electrode 34 A 2 , as shown in FIG. 11B .

Next, as shown in FIG. 12A and FIG. 12B , clock Φ 2 is set to GND, rendering MOS transistor TR 2 non-conductive. Clock /Φ 2 rises from GND (0 V) to Vcc (4 V). At this point, diffusion layers 32 A-A of the odd-numbered capacitor group G odd have their potential level lowered to 0 V, whereas diffusion layers 32 A-B of the even-numbered capacitor group G even have their potential level raised from 0 to 4 V.

As node N 1 does, node N 2 has its potential V 2 raised by αVcc because of the capacitive coupling of capacitor C 2 . Potential V 2 of node N 2 therefore rises to almost {Vdd−Vth 1 +αVcc}−Vth 2 }+αVcc.

As described above, the isolation region STI electrically isolates the odd-numbered capacitor group G odd and the even-numbered capacitor group G even . Therefore, the depletion layer of capacitor Φ 2 will not impose adverse influence on capacitors C 1 , C 3 , C 5 and C 7 of the odd-numbered capacitor group G odd .

While clock /Φ 2 remains at 4 V, MOS transistor TR 3 becomes conductive, and clock Φ 2 remains at 0 V. Hence, 0 V is applied to capacitors C 1 , C 3 , C 5 and C 7 of the odd-numbered capacitor group G odd , at one end (i.e., well region 51 p - o ). As a charge moves between capacitors C 2 and C 3 via transistor TR 3 , potential V 2 of node N 2 is transferred to node N 3 until potential V 3 of node N 3 rises to {(potential of note N 2 )−Vth 3 }. In the capacitor C 3 connected to node N 3 , potential V 3 of node N 3 is applied to capacitor electrode 34 A 4 . Thus, a channel 90 is formed beneath capacitor electrode 34 A 3 , of capacitor C 3 , in accordance with the magnitude of potential V 3 of node N 3 .

The charge is repeatedly transferred in the same way as described above, alternately in the odd-numbered capacitor groups G odd and G even . The potential is thereby raised at any other nodes, up to node N 9 .

More precisely, while clock /Φ 2 is at Vcc, MOS transistors TR 1 , TR 3 , TR 5 , TR 7 and TR 9 at the odd-numbered stages become conductive, whereas MOS transistors TR 2 , TR 4 TR 6 and TR 8 at the even-numbered stages become non-conductive.

Currents therefore flow from the node at power-supply potential Vdd to node N 1 , from node N 2 to node N 3 , from node N 4 to node N 5 , from node N 6 to node N 7 , and from node N 8 to node N 9 . The charge is thereby transferred.

While clock Φ 2 is at Vcc, the even-numbered MOS transistors TR 2 , TR 4 , TR 6 and TR 8 become conductive, whereas the odd-numbered TR 1 , TR 3 TR 5 , TR 7 and TR 9 become non-conductive. In this case, currents flow from node N 1 to node N 2 , from node N 3 to node N 4 , from node N 5 to node N 6 , and from node N 7 to node N 8 . The charge is thereby transferred.

A capacitive load, such as a word line, may be connected to node N 9 . If so, the output voltage Vout is smoothed by the capacitance of the load.

As long as the charge pump circuit 6 is operating, each of nodes N 1 to N 8 remains at a higher potential than the immediately preceding node. Different potentials are therefore applied to capacitor electrodes 34 A 1 to 34 A 8 through metal interconnect layers 29 1 to 29 8 that are electrically isolated from one another.

The charge pump circuit 6 of FIG. 7 needs a plurality of capacitors C 1 to C 9 in order to raise the any voltage (e.g., Vdd) to the drive voltage (e.g., 20 V or more) for the memory cell array.

In the semiconductor device according to this embodiment, a plurality of capacitors (MOS capacitors) are provided in a common semiconductor region (well region), not electrically isolated by isolation regions, while semiconductor regions (well regions) 51 p - o and 51 p - e are commonly controlled in terms of potential level. Therefore, the area the isolation regions occupy are reduced, the area capacitors C 1 to C 8 occupy and the area the circuit 6 having these capacitors occupies can be smaller than in the case where such isolation regions are provided for every capacitor. Although capacitors C 1 to C 8 are provided in semiconductor regions 51 p - o 51 p - e , they are scarcely degraded in operating characteristic, because each of the capacitor groups G odd and G even is commonly controlled in terms of the potential level of semiconductor regions 51 p - o and 51 p - e as a common one ends of capacitors.

In the charge pump circuit, nodes N 1 to N 8 are applied with different voltages such that each node is applied with a lower voltage than the node in the immediately higher stage. The voltage difference between any adjacent two nodes is relatively small, nevertheless. In capacitors C 1 to C 8 arranged in a common semiconductor region, two capacitors having a small potential difference are adjacently arranged in the common semiconductor region. Thus suppresses the dielectric breakdown of the inter-layer insulating film between the capacitor electrodes or the mutual interference of the capacitor electrodes, without increasing the distance between the capacitor electrodes. This can reduce the area of the region in which the capacitors are formed.

Capacitors C 1 and C 3 , for example, are arranged adjacent to each other as shown in FIG. 9 . Capacitor electrodes 34 A 1 and 34 A 3 are therefore located adjacent to each other. Since capacitor electrodes 34 A 1 and 34 A 3 have a relatively small potential difference, the distance between them can be short.

Any two capacitors isolated by an isolation region should better be connected to adjacent two nodes. As a result, the potential difference between the two interconnects connected to the electrodes of either capacitor become small, ultimately decreasing the leakage between the interconnects.

As has been described, the area the capacitors occupy can be reduced in the semiconductor device according to the first embodiment.

(2) Second Embodiment

A semiconductor device (e.g., flash memory) according to a second embodiment will be described with reference to FIG. 13A and FIG. 13B . FIG. 13A and FIG. 13B schematically show the sectional structure of a semiconductor device according to the second embodiment.

In the semiconductor device according to the first embodiment, the semiconductor region in which a plurality of capacitors are provided is a p-type well region or a p-type semiconductor substrate.

In the second embodiment, a plurality of capacitors C 1 to C 8 may be provided in an n-type well region 51 n - o and 51 n - e.

Capacitors C 1 , C 3 , C 5 and C 7 are provided in n-type well region 51 n - o . The potential at one end (diffusion layer 32 A-A) of each of these capacitors C 1 , C 3 , C 5 and C 7 is controlled by a common potential level (control signal) Φ 2 .

Similarly, capacitors C 2 , C 4 , C 6 and C 8 are provided in n-type well region 51 n - e . The potential at one end (diffusion layer 32 A-B) of each of these capacitors C 2 , C 4 , C 6 and C 8 is controlled by a common potential level (control signal) /Φ 2 .

Thus, capacitors C 1 , C 3 , C 5 and C 76 provided in n-type well region 51 n - o are controlled, at one end, by a common potential level, and capacitors C 2 , C 4 , C 6 and C 8 provided in n-type well region 51 n - e are controlled, at one end, by a common potential level. Therefore, the area occupied by the regions in which the capacitors are provided can be reduced as in the first embodiment.

If capacitors C 1 to C 8 are provided on a p-type well region or a p-type substrate, the capacitance of the depletion layers will be formed as channels (inversion layers) are formed beneath the capacitor electrodes 34 , as has been described above.

In this embodiment, channels need not be formed beneath capacitor electrodes 34 A 1 , 34 A 3 , 34 A 5 and 34 A 7 , respectively, when the odd-numbered capacitors C 1 , C 3 , C 5 and C 7 are driven, as long as these capacitors, which are formed in n-type well region 51 n - o , are controlled by a common potential level. Similarly, channels need not be formed beneath capacitor electrodes 34 A 2 , 34 A 4 , 34 A 6 and 34 A 8 , respectively, when the even-numbered capacitors C 2 , C 4 , C 6 and C 8 are driven, as long as these capacitors, which are formed in n-type well region 51 n - e , are controlled by a common potential level. Capacitors C 1 to C 8 are formed by n-type well as a electrode, capacitor electrode 34 A, and capacitor insulating film 42 A.

Hence, the charge pump circuit can be prevented from being degraded in operating characteristic, because capacitors C 1 to C 8 are formed in n-type regions 51 n - o and 51 n - e , the capacitance of depletion layers is not formed at all.

If capacitors C 1 to C 8 are formed on n-type well regions 51 n - o and 51 n - e as described above, channels need not be formed beneath capacitor electrodes 34 A 1 to 34 A 8 . Therefore, diffusion layers 32 A-A and 32 A-B may not be formed in n-type well regions 51 n - o and 51 n - e to surround capacitor electrodes 34 A 1 to 34 A 8 . It is desired, however, that the diffusion layers should be provided at the junctions between contact plugs CPA-o and CPA-e and well regions 51 n - o and 51 n - e , in order to decrease the contact resistance between contact plugs CPA-o and CFA-e and well regions 51 n - o and 51 n - e.

As shown in FIG. 13A and FIG. 13B , both the odd-numbered capacitor groups G odd and even-numbered capacitor group G even are provided in the n-type well region. Nonetheless, one of the capacitor groups may be provided in the n-type well region, and the other capacitor group may be provided in a p-type well region.

As has been described, the area occupied by the region in which the capacitors are provided can be reduced in the semiconductor device according to the second embodiment, in the same manner as in the first embodiment. In addition, the capacitors and the circuit using them can be improved in their operating characteristic.

(3) Third Embodiment

A semiconductor device according to a third embodiment will be described with reference to FIG. 14 , FIG. 15 , FIG. 16 , and FIGS. 17A to 17C . In this embodiment, the clocks input to the charge pump circuit are three phase clocks. The features that distinguish this embodiment from the first and second embodiments will be described in the main.

FIG. 14 is an equivalent circuit diagram of the charge pump circuit of the third embodiment.

As FIG. 14 shows, the charge pump circuit 6 of this embodiment receives a clock Φ 3 - 1 at one end (semiconductor region) of each of three capacitors Cj connected to nodes Nj (j=1, 4 and 7). Further, a clock Φ 3 - 2 is input to one end (semiconductor region) of each of three capacitors C(j+1) connected to nodes Nj+1. Still further, a clock Φ 3 - 3 is input to one end (semiconductor region) of each of three capacitors C(j+2) connected to nodes Nj+2.

FIG. 15 shows the waveforms of the control signals (clocks) used in the charge pump circuit 6 . As can be seen from FIG. 15 , clocks Φ 3 - 1 , Φ 3 - 2 and Φ 3 - 3 have waveforms shifted in phase with respect to one another by, for example, 120°. Each clock has a leading edge having a smaller slope (through rate) than the trailing edge. In other words, clocks Φ 3 - 1 , Φ 3 - 2 and Φ 3 - 3 rise more slowly than they fall.

The charge pump circuit of the third embodiment has a plurality of capacitors C 1 to C 8 . The structure of capacitors C 1 to C 8 will be described, with reference to FIG. 16 and FIGS. 17A to 17C . FIG. 16 is a plan view showing the layout of capacitors C 1 to C 8 incorporated in the charge pump circuit 6 . FIG. 17A , FIG. 17B and FIG. 17C are sectional views of three capacitor formation regions AAC- 1 , AAC- 2 and AAC- 3 , respectively.

As shown in FIG. 16 , three semiconductor regions AAC- 1 , AAC- 2 and AAC- 3 are provided in the surface region of a semiconductor substrate 10 , in association with clocks Φ 3 - 1 , Φ 3 - 2 and Φ 3 - 3 that are shifted in phase with respect to one another. Semiconductor regions AAC- 1 , AAC- 2 and AAC- 3 are surrounded by isolation regions STI, respectively, and are electrically isolated from one another by the isolation regions STI. On semiconductor region AAC- 1 , a capacitor group G 1 is provided. On semiconductor region AAC- 2 , a capacitor group G 2 is provided. On semiconductor region AAC- 3 , a capacitor group G 3 is provided.

As FIG. 16 and FIGS. 17A to 17C show, clock Φ 3 - 1 is applied to a diffusion layer 32 A-A provided in semiconductor region AAC- 1 , through a contact plug CPA- 1 . Clock Φ 3 - 2 is applied to a diffusion layer 32 A-B provided in semiconductor region AAC- 2 , through a contact plug CPA- 2 . Similarly, clock Φ 3 - 3 is applied to a diffusion layer 32 A-C provided in semiconductor region AAC- 3 , through a contact plug CPA- 3 . Therefore, clocks (potentials) Φ 3 - 1 , Φ 3 - 2 and Φ 3 - 3 are applied, respectively to capacitor groups G 1 , G 2 and G 3 , respectively, more precisely to one end of each capacitor.

As can be seen from FIG. 16 and FIGS. 17A to 17C , capacitor group G 1 to which clock Φ 3 - 1 is input includes three capacitors C 1 , C 4 and C 7 . Capacitors C 1 , C 4 and C 7 of the group G 1 are provided in one semiconductor region AAC- 1 . The three capacitors C 1 , C 4 and C 7 therefore receive clock Φ 3 - 1 at one end (diffusion layer 32 A-A).

Capacitor group G 2 to which clock Φ 3 - 2 is input includes three capacitors C 2 , C 5 and C 8 . Capacitors C 2 , C 5 and C 8 of the group G 2 are provided in one semiconductor region AAC- 2 . These three capacitors C 2 , C 5 and C 8 therefore receive clock Φ 3 - 2 at one end (diffusion layer 32 A- 13 ).

Capacitor group G 3 to which clock Φ 3 - 3 is input includes two capacitors C 3 and C 6 . Capacitors C 3 and C 6 of the group G 3 are provided in one semiconductor region AAC- 3 . The two capacitors C 3 and C 6 therefore receive clock Φ 3 - 3 at one end (diffusion layer 32 A-C).

In this embodiment, capacitors C 1 to C 8 are divided into three groups, which are provided in three semiconductor regions AAC- 1 , AAC- 2 and AAC- 3 , respectively, and the capacitors of each group are provided on the same semiconductor region to which a common clock (control signal) is applied.

The layout of the three semiconductor regions AAC- 1 , AAC- 2 and AAC- 3 is not limited to the layout of FIG. 16 . Moreover, well regions 51 - 1 , 51 - 2 and 51 - 3 shown in FIG. 16 and FIGS. 17A to 17C , in which capacitor groups G 1 to G 3 are provided, respectively, may be p-type well regions or n-type well regions. Furthermore, capacitor groups G 1 to G 3 may be provided in a semiconductor region (e.g., p-type semiconductor region) in which no well regions are provided.

The charge pump circuit which receives three phase clocks operates in principle essentially identical to that of the charge pump circuit driven by the above-mentioned two phase clocks. That is, as the three clock signals Φ 3 - 1 to Φ 3 - 3 change in potential level in the charge pump circuit, the potential of each of nodes N 1 to N 8 rises above the value at the immediately preceding node because of the capacitive coupling of capacitors C 1 to C 8 . The voltages, each so raised, are sequentially transferred to the next stage nodes N 2 to N 9 through transistors TR 2 to TR 9 , respectively.

Since the three clocks are used as shown in FIG. 16 , the charge can be efficiently transferred even if MOS transistors of a large reverse current are used as rectification elements. This is because, as shown in FIG. 16 , clock Φ 3 - 1 is raised to Vcc at time t 1 when clock Φ 3 - 2 rises from GND to Vcc, and a current can therefore hardly flow in reverse direction through the junction. This is also because a current can easily flow from capacitor C 2 to capacitor C 3 , for example, since clock Φ 3 - 3 falls quickly to GND at time t 1 .

Time for recovering from the reversely biased state (hereinafter called “reverse recovery time” can be acquired at a ramp part where the through rate is low on the voltage rising side. Hence, a charge pump circuit can be provided, in which a smaller current flows back, from the high-potential side to the low-potential side, than in the case no reverse recovery time is acquired.

As described above, the first embodiment can be applied to the charge pump circuit according to this embodiment. Therefore, the third embodiment can achieve the same advantages as the first embodiment.

Hence, the area the regions in which the capacitors are provided can be reduced in the third embodiment, as in the first and second embodiments.

(4) Fourth Embodiment

A semiconductor device (e.g., flash memory) according to a fourth embodiment will be described with reference to FIG. 18 , FIG. 19 and FIG. 20 .

In this embodiment, the clocks input to the charge pump circuit are four phase clocks. The features that distinguish this embodiment from the first to third embodiments will be mainly described.

FIG. 18 is an equivalent circuit diagram of the charge pump circuit of this embodiment.

The charge pump circuit 6 shown in FIG. 18 differs in terms of configuration from the charge pump circuits shown in FIG. 7 and FIG. 14 , in the following respects.

In the charge pump circuit 6 of this embodiment, a clock Φ 4 - 1 is input to two capacitors Cj connected to nodes Nj 5 ), at the other electrode (semiconductor region) thereof. A clock Φ 4 - 2 is input to two capacitors C(j+1) connected to nodes N(j+1), at the other electrode (semiconductor region) thereof. A clock Φ 4 - 3 is input to two capacitors C(j+2) connected to nodes N(j+2), at the other electrode (semiconductor region) thereof. A clock Φ 4 - 4 is input to two capacitors C(j+3) connected to nodes N(j+3), at the other electrode (semiconductor region) thereof.

FIG. 19 shows the waveforms of the control signals (clocks) used in the charge pump circuit 6 . As can be seen from FIG. 19 , clocks Φ 4 - 1 , Φ 4 - 2 , Φ 4 - 3 and Φ 4 - 4 have waveforms shifted in phase with respect to one another by, for example, 90°. Unlike in the third embodiment, the leading edge and trailing edge of each clock have the same slope.

FIG. 20 shows the layout of capacitors C 1 to C 8 incorporated in the charge pump circuit 6 to which four phase clocks are input. This embodiment differs from the first to third embodiments, only in the number of capacitors provided in one semiconductor region. It is identical to the first to third embodiments in the sectional structure of the capacitors. Therefore, sectional structure of the capacitors will not described or shown in drawings.

As shown in FIG. 20 , four semiconductor regions AAC- 1 , AAC- 2 , ACC- 3 and AAC- 4 are provided in the surface region of a semiconductor substrate 10 , in association with the four clocks Φ 4 - 1 , Φ 4 - 2 , Φ 4 - 3 and Φ 4 - 4 , respectively.

The four semiconductor regions AAC- 1 , AAC- 2 , ACC- 3 and AAC- 4 are surrounded by isolation regions STI, respectively, and are therefore electrically isolated from one another. Semiconductor regions AAC- 1 , AAC- 2 , ACC- 3 and AAC- 4 may be p-type well regions or n-type well regions. Alternatively, they may be p-type or n-type semiconductor substrates in which no well regions are provided.

Capacitors C 1 to C 9 form four groups, which are provided in four semiconductor regions AAC- 1 , AAC- 2 , AAC- 3 and AAC- 4 , respectively, and to which the four clocks Φ 4 - 1 , Φ 4 - 2 , Φ 4 - 3 and Φ 4 - 4 are input, respectively.

Clock Φ 4 - 1 is supplied through a contact plug CPA- 1 to diffusion layer 32 A-A provided in semiconductor region AAC- 1 . In semiconductor region AAC- 1 , capacitors C 1 and C 5 are provided to receive clock Φ 4 - 1 .

Clock Φ 4 - 2 is supplied through a contact plug CPA- 2 to diffusion layer 32 A-B provided in semiconductor region AAC- 2 . In semiconductor region AAC- 2 , capacitors C 2 and C 6 are provided to receive clock Φ 4 - 2 .

Clock Φ 4 - 3 is supplied through a contact plug CPA- 3 to diffusion layer 32 A-C provided in semiconductor region AAC- 3 . In semiconductor region AAC- 3 , capacitors C 3 and C 7 are provided to receive clock Φ 4 - 3 .

Clock Φ 4 - 4 is supplied through a contact plug CPA- 4 to diffusion layer 32 A-D provided in semiconductor region AAC- 4 . In semiconductor region AAC- 4 , capacitors C 4 and C 8 are provided to receive clock Φ 4 - 4 .

Using four phase signals, the charge pump circuit 6 of FIG. 18 can efficiently transfer a charge, for reasons. As shown in FIG. 19 , a current can hardly flow though a junction in the reverse direction since clock Φ 4 - 1 is already at Vcc at time t 2 when clock Φ 4 - 2 rises from GND to Vcc. A current can easily flow, for example, from capacitor C 2 to capacitor C 3 since clock Φ 4 - 3 is at GND at time t 2 . Moreover, a reverse recovery time can be acquired at a ramp part where the clock rises, as in the third embodiment. Hence, a charge pump circuit can be provided, in which a smaller current flows back, from the high-potential side to the low-potential side.

Also in the charge pump circuit 6 according to this embodiment, which receives four phase clocks, four groups of capacitors, each composed of two or more capacitors driven by the same clock, are provided in the four semiconductor regions AAC- 1 to ACC- 4 , respectively. The flash memory according to this embodiment can therefore achieve the same advantage as the flash memory according to the first to third embodiments.

Hence, the area the regions in which the capacitors are provided can be reduced in the fourth embodiment, as in the first to third embodiments.

(5) Modification

A modification of the semiconductor devices according to the first to fourth embodiments will be described with reference to FIG. 21 and FIG. 22 . FIG. 21 is a plan view showing the layout of capacitors provided in the modification. FIG. 22 is a sectional view taken along line XXII-XXII shown in FIG. 21 .

In the modified semiconductor device, a ratio (hereinafter called “coverage ratio”) is set for the area occupies of an element/line pattern in a prescribed area provided any interconnect level. That is, a particular coverage ratio is set for the area capacitor electrodes (or gate electrodes) occupy in one semiconductor region (element region).

In the first to fourth embodiments, the electrodes of any capacitor are formed on semiconductor regions only. However, the layout of the capacitor electrodes in the semiconductor region is not limited to the above-described one, so it achieves a prescribed coverage ratio.

For example, as shown in FIG. 21 and FIG. 22 , capacitor electrodes 34 A 1 to 34 A 8 may not extend from semiconductor regions AAC-o and AAC-e to the isolation region (isolation insulating film) STI.

The diffusion layers 32 B provided in semiconductor regions AAC-o and AAC-e have been formed in self-alignment, using capacitor electrodes 34 A 1 to 34 A 8 as mask. Therefore, any one of capacitor electrodes 34 A 1 to 34 A 8 is interposed between two of diffusion layers 32 A 1 to 32 A 5 in the layout shown in FIG. 21 . That is, diffusion layers 32 A 1 to 32 A 5 of capacitors C 1 , C 3 , C 5 and C 7 are isolated from one another, not continuous to one another in semiconductor region AAC-o, in this modified semiconductor device. One capacitor electrode is, for example, sandwiched two diffusion layers. In the layout of FIG. 21 , five diffusion layers 32 A 1 , 32 A 2 , 32 A 3 , 32 A 4 and 32 A 5 are formed in one semiconductor region AAC-o. Thus, the capacitor group has two or more the diffusion layers sandwiching the each capacitor electrode.

Channels are therefore formed beneath capacitor electrodes 34 A 1 , 34 A 3 , 34 A 5 and 34 A 7 are when capacitors C 1 , C 3 , C 5 and C 7 are driven. Though diffusion layers 32 A 1 to 32 A 5 are not continuous to one another in semiconductor region AAC-o, through the channels thus formed, the potentials (clocks) Φ 2 and /Φ 2 applied are transferred to the adjacent capacitors C 1 , C 3 , C 5 and C 7 .

In brief, it suffices to form at least one contact plug CPA on one of diffusion layers 32 A 1 , 32 A 2 , 32 A 3 , 32 A 4 and 32 A 5 that are provided in the capacitor formation region AAC. As a result, the area the capacitors occupy can be decreased.

Moreover, contacts plugs CPE 1 to CPE 8 connected to capacitor electrodes 34 A 1 to 34 A 8 , respectively, may be arranged above the isolation film STI.

Also in the modified semiconductor device shown in FIG. 21 and FIG. 22 , the area the capacitors occupy can be reduced as in the first to fourth embodiments.

[Other Features]

The first to fourth embodiments described above are flash memories. In these embodiments, a two-dimensional memory cell array is formed in the surface region of a semiconductor substrate. The embodiments may, however, be applied to a flash memory having a three-dimensional memory cell array, such as a memory cell array of bit-cost scalable (BiCS) structure.

Further, the semiconductor devices according to the first to fourth embodiments are not limited to flash memories. They may be magnetoresistive random access memories (MRAMs), resistive random access memories (ReRAMs) or phase-change random access memories (PCRAMs).

Any one type of these memories may incorporate above-described capacitors provided in one semiconductor region and the above-described circuit (e.g., charge pump circuit) using these capacitors. Moreover, the semiconductor devices according to the embodiments are not limited to semiconductor memories. Rather, they may be embedded memories or logic LSIs.

In the first to fourth embodiments, a plurality of capacitors provided in one semiconductor region are used in a charge pump circuit. Nonetheless, capacitors provided in one semiconductor region may be used to form a circuit other than a charge pump circuit, in which the capacitors are controlled at one end (well region or substrate) by the same control signal (common potential level).

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