SEMICONDUCTOR MEMORY DEVICE AND DRIVING METHOD OF SEMICONDUCTOR MEMORY DEVICE

Abstract

A novel semiconductor memory device whose power consumption is low is provided. A source of a writing transistor WTr_n_m, a gate of a reading transistor RTr_n_m, and one electrode of a capacitor CS_n_m are connected to each other. A gate and a drain of the writing transistor WTr_n_m are connected to a writing word line WWL_n and a writing bit line WBL_m, respectively. The other electrode of the capacitor CS_n_m is connected to a reading word line RWL_n. A drain of the reading transistor RTr_n_m is connected to a reading bit line RBL_m. Here, the potential of the reading bit line RBL_m is input to an inverting amplifier circuit such as a flip-flop circuit FF_m to be inverted by the inverting amplifier circuit. This inverted potential is output to the writing bit line WBL_m.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a divisional of U.S. application Ser. No. 13/288,089, filed Nov. 3, 2011, now allowed, which claims the benefit of a foreign priority application filed in Japan as Serial No. 2010-249435 on Nov. 8, 2010, both of which are incorporated by reference.

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to a memory device including a semiconductor.

2. Description of the Related Art

As a memory device including a semiconductor, which is used in various electronic products and electronics products, a dynamic random access memory (DRAM), a static random access memory (SRAM), and the like can be given.

In a DRAM, data is stored by holding charge in a capacitor which is provided in a memory cell. However, even when a transistor used for switching is in an off state, a slight amount of leakage current is generated between a source and a drain; thus, the data is lost within a relatively short time (several tens of seconds at the longest). Therefore, the data needs to be rewritten (refreshed) on a regular cycle (generally once every several tens of milliseconds), and power consumption is high even in a standby period.

While miniaturization of a circuit has been attempted, a deep hole (a trench) or a chimney-like projection (a stack) is formed to function as a capacitor because the capacitance of the capacitor needs to be kept constant (generally, 10 ff or higher). With the miniaturization, the aspect ratio thereof (the ratio of height or depth to base) has become 50 or more. A special technique for forming such a structure has been needed (see Non-Patent Document 1 and Non-Patent Document 2).

In an SRAM, data is held by utilizing a bistable state of a flip-flop circuit. When a CMOS inverter (a complementary inverter) is used in a flip-flop circuit of an SRAM, the amount of power consumption in a standby period is significantly smaller than that of a DRAM (see Patent Document 1). Therefore, an SRAM is used instead of a DRAM for applications, e.g., a cellular phone, in which the frequency of data writing and data reading is not so high and a standby period is much longer than a period during which data writing and data reading are performed. However, since six transistors are used in one memory cell, the degree of integration is lower than that of a DRAM and the unit cost per bit is ten times or more as high as that of the DRAM.

In recent years, a transistor in which the amount of leakage current between a source and a drain in an off state is extremely small and which has excellent charge holding characteristics has been devised, and a memory cell using it has been proposed (see Patent Document 2). In the case where a transistor of this structure is used, two transistors are needed for one memory cell; however, a capacitor having large capacitance is not needed unlike in a DRAM. In addition, data can be held for an extremely long period without refresh operation.

REFERENCE

Patent Document

• [Patent Document 1] U.S. Pat. No. 5,744,844
• [Patent Document 2] United States Patent Application Publication No. 2011/0101334

Non-Patent Document

• [Non-Patent Document 1] K. Kim, “Technology for sub-50 nm DRAM and NAND flash manufacturing”, Technical Digest of International Electron Devices Meeting, pp. 333-336, 2005.
• [Non-Patent Document 2] W. Mueller et al., “Challenges for the DRAM cell scaling to 40 nm”, Technical Digest of International Electron Devices Meeting, pp. 347-350, 2005.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide a novel semiconductor device (particularly, a semiconductor memory device). It is another object to provide a driving method of a novel semiconductor device (particularly, a driving method of a semiconductor memory device). Further, it is another object to provide a manufacturing method of a novel semiconductor device (particularly, a manufacturing method of a semiconductor memory device).

According to the present invention, a semiconductor memory device whose power consumption per bit in a standby period is lower than that of a DRAM and whose degree of integration is higher than that of an SRAM, a memory cell used in the semiconductor memory device, driving methods thereof, and manufacturing methods thereof are provided.

In addition, according to the present invention, a memory cell in which three or less transistors are used and which consumes a current of 1×10⁻²⁰ A or less in a standby period, and a semiconductor device including such a memory cell are provided. According to the present invention, at least one of the above objects is achieved.

The present invention will be described below; terms used in this specification are briefly described. First, when one of a source and a drain of a transistor is called a drain, the other is called a source in this specification. That is, they are not distinguished depending on the potential level. Therefore, a portion called a source in this specification can be alternatively referred to as a drain.

Even when the expression “to be connected” is used in this specification, there is a case in which no physical connecting portion is formed and a wiring is only extended in an actual circuit. For example, in a circuit including a field-effect transistor (FET), one wiring functions as gates of a plurality of FETs in some cases. In that case, one wiring having a plurality of branches may be illustrated in a circuit diagram. Even in such a case, the expression “a wiring is connected to a gate” may be used in this specification.

Note that in this specification, in referring to a specific row, a specific column, or a specific position in a matrix, a reference sign is accompanied in some cases by a sign denoting a coordinate as follows, for example: “a writing transistor WTr_n_m”, “a bit line BL_m”, and “a writing word line WWL_n”. In the case where a row, a column, or a position is not specified, the case where elements are collectively referred to, or the case where the position is obvious, the following expressions may be used: “a writing transistor WTr”, “a bit line BL”, and “a writing word line WWL” or simply “a writing transistor”, “a bit line”, and “a writing word line”.

The expression “the potential of a word line is set to H” (or “the potential of a word line is set to L”) means that the potential of the word line is set to a potential at which a transistor whose gate is connected to the word line is turned on (or turned off).

In one embodiment of the present invention, one memory cell includes a transistor as a writing transistor, in which leakage current between a source and a drain in an off state is small, another transistor (a reading transistor), and a capacitor. Further, as wirings connected to these, four kinds of wirings, that is, a writing word line, a writing bit line, a reading word line, and a reading bit line, are prepared.

The source of the writing transistor is connected to a gate of the reading transistor and one electrode of the capacitor. In a portion where they are connected to each other, charge can be transferred only through the writing transistor; when the writing transistor is off, the portion is insulated from its periphery and charge is confined therein. Therefore, this portion is referred to as a floating node, and a portion of the gate of the reading transistor is particularly referred to as a floating gate.

In addition, a gate of the writing transistor is connected to the writing word line. The drain of the writing transistor is connected to the writing bit line. A drain of the reading transistor is connected to the reading bit line. The other electrode of the capacitor is connected to the reading word line.

Note that a source of the reading transistor is supplied with an appropriate potential by another wiring. Depending on the reading method, a fluctuation in this potential can be small. For example, a driving method by which a constant potential can be held for 1 second or longer can be employed. Therefore, the resistance of the wiring connected to the source of the reading transistor does not necessarily need to be low. For example, silicon which is doped with impurities, or silicon which is doped with impurities and has a surface where silicide is formed may be used.

As the writing transistor, a transistor in which current flowing between a source and a drain can be 1×10⁻²¹ A or less, preferably 1×10⁻²⁴ A or less at ambient temperature when the transistor is in use (e.g., 25° C.), or can be 1×10⁻²⁰ A or less, preferably 1×10⁻²³ A or less at 85° C. by adjusting the potential of a gate is preferably used. Under such conditions, the capacitance of the capacitor can be much smaller than that of a conventional DRAM. In addition, the interval between refresh operations, which are necessary in a conventional DRAM, can be significantly long, and the refresh operation can be substantially unnecessary.

For example, even when the capacitance of the capacitor is 0.01 ff, which is 1/1000 or less of that of a conventional DRAM, the time constant is 1×10⁷ seconds (115 days) in the case where current flowing between the source and the drain is 1×10⁻²⁴ A. Thus, data can be held for a long period, which cannot be assumed in a conventional DRAM. That is, in using a usual personal computer, the refresh operation may be regarded as unnecessary, or the refresh operation may be performed at least once every ten days.

In other words, refresh operation (rewriting of data for the purpose of compensating a reduction in charge accumulated in a capacitor) which needs to be performed ten or more times per second in a conventional DRAM becomes unnecessary in a usual usage.

In the case of a general silicon semiconductor, it is difficult to realize leakage current having such a small value; however, such a value can be achieved in a transistor in which a semiconductor whose band gap is 2.8 electron volts (eV) or more (i.e., a wide bandgap semiconductor), such as an oxide semiconductor, is processed under a preferable condition. Therefore, a wide bandgap semiconductor is preferably used as a material for the writing transistor. Needless to say, in the present invention, a silicon semiconductor is not excluded from examples of a semiconductor used for the writing transistor.

Although a variety of known materials can be used as the oxide semiconductor, a material with a bandgap greater than or equal to 3 eV and less than 3.6 eV is desirable. In addition, it is desirable to use a material with an electron affinity greater than or equal to 4 eV, preferably a material with an electron affinity greater than or equal to 4 eV and less than 4.9 eV. In particular, an oxide including gallium and indium is preferable for the object of the present invention. Among such materials, a material whose carrier concentration derived from a donor or an acceptor is less than 1×10⁻¹⁴ cm⁻³, preferably less than 1×10⁻¹¹ cm⁻³ is desirable.

As for the reading transistor, although there is no particular limitation on the leakage current between the source and the drain in an off state, smaller leakage current is preferable because power consumption can be reduced. Further, a transistor which operates at high speed is desirable in order to increase the reading speed. Specifically, it is preferable to use a transistor with a switching speed of 10 nanoseconds or less.

Further, in both the writing transistor and the reading transistor, gate leakage current (leakage current between the gate and the source or between the gate and the drain) needs to be extremely small. Also in the capacitor, internal leakage current (leakage current between the electrodes) needs to be small. Each leakage current may be 1×10⁻²¹ A or less, preferably 1×10⁻²⁴ A or less at ambient temperature when the transistor or the capacitor is in use (e.g., 25° C.).

The potential of the gate (floating node) of the reading transistor is changed according to the potential of the reading word line. As a result, the gate capacitance of the reading transistor is changed. That is, the gate capacitance of the reading transistor in the case where the reading transistor is in an on state is larger than that in the case where the reading transistor is in an off state. When change in the gate capacitance of the reading transistor is larger than the capacitance of the capacitor, a problem is caused in operation of the memory cell in some cases.

Therefore, the capacitance of the capacitor is preferably larger than or equal to the gate capacitance of the reading transistor, further preferably larger than or equal to twice as large as the gate capacitance of the reading transistor. For this, the permittivity of a dielectric of the capacitor is preferably larger than that of a gate insulator of the reading transistor. Note that in the case where the dielectric of the capacitor and a gate insulator of the writing transistor are formed of the same material, there is an advantageous effect of improving the current driving capability of the writing transistor with the use of such a material having high permittivity.

To the reading word line, many capacitors are connected in parallel in this manner; thus, the capacitance of the reading word line is increased, which does not lead to a problem in many cases. The reason is as follows. Since the capacitance connected to the reading word line corresponds to the capacitance of the capacitor and the gate capacitance of the reading transistor which are connected in series, the combined capacitance is always smaller than the smaller one (under the above condition, the gate capacitance of the reading transistor).

Note that in a miniaturized semiconductor circuit, in order to form a capacitor without manufacturing a special structure having an extremely large aspect ratio, the capacitance of the capacitor is preferably 1 fF or less, more preferably 0.1 fF or less if possible. However, the capacitance may be 1 fF or more because the capacitance is preferably large in order to reduce the probability of data fluctuation due to a soft error.

Note that in the above structure, a fluctuation in charge of the capacitor due to a soft error is caused by the writing transistor. However, when a semiconductor layer of the writing transistor has a small thickness of 50 nm or less, the probability of charge fluctuation due to a soft error can be negligible even when the capacitance of the capacitor is 0.1 fF or less. Therefore, when the semiconductor layer used in the writing transistor has a thickness of 50 nm or less, reliability can be maintained even in the case where the capacitance of the capacitor is 0.1 fF or less.

Note that in order to suppress a short-channel effect of the writing transistor, the semiconductor layer is preferably thin. When the channel length of the writing transistor, the thickness of the gate insulator, the permittivity of the gate insulator, the thickness of the semiconductor layer, and the permittivity of the semiconductor layer are expressed as L, t₁, ∈₁, t₂, and ∈₂, respectively, L/5>(∈₂t₁/∈₁+t₂) is preferably satisfied. For example, in the case where L=100 nm, t₁=10 nm, and ∈₁=∈₂ are satisfied, t₂ is preferably less than 10 nm. When the semiconductor layer is thin in such a manner, a soft error described above can be prevented.

The writing word line, the writing bit line, the reading word line, and the reading bit line are arranged in a matrix. In order to perform matrix driving, it is preferable that the writing word line and the writing bit line be orthogonal, the writing word line and the reading word line be in parallel, and the writing bit line and the reading bit line be in parallel.

That is, one writing word line and one reading word line are needed per row in a matrix, and one writing bit line and one reading bit line are needed per column in a matrix. Therefore, for a matrix with N rows and M columns (N and M are natural numbers of 2 or more) of a memory device, at least (2N+2M) wirings are necessary. In addition, a wiring connected to the source of the reading transistor RTr is needed.

Some of these wiring are formed so as to have a three-dimensional structure, whereby the area occupied by the wirings can be reduced. For example, the wiring connected to the source of the reading transistor RTr is formed so as to overlap with the writing word line or the reading word line, or formed between the writing word line and the reading word line, whereby the memory cell can be formed without changing the practical area of the memory cell.

The reading transistor and the writing transistor may be formed in different layers. Note that a structure in which a writing word line for one memory cell functions also as a reading word line for another memory cell, or a structure in which a writing bit line for one memory cell functions also as a reading bit line for another memory cell is employed, whereby the number of necessary wirings can be reduced.

In such a memory cell, data writing is performed in the following manner: the potential of the writing word line is set to H so that the writing transistor is turned on, and charge corresponding to the potential of the writing bit line is supplied to the capacitor of the memory cell.

Writing transistors of a large number of memory cells are connected to the writing word line. In some cases, data needs to be written to some memory cells, and does not need to be written to the other memory cells. When the potential of the writing word line is set to H, all the writing transistors connected to the same writing word line are turned on so that there is a possibility that data of the memory cell for which data writing is unnecessary is rewritten to false one.

In order to prevent this, data reading operation is performed before data writing. Data is read to be output to the reading bit line, and that data has a phase opposite to that of held data. That is, in the case where data “1” is held, data to be output to the reading bit line corresponds to data “0”.

Output from the reading bit line is inverted by an inverting amplifier circuit such as an inverter circuit or a flip-flop circuit. That is, in the case where data output from the reading bit line corresponds to data “0”, output from the inverter circuit or output from another input terminal of the flip-flop circuit (these are collectively referred to as output from the inverting amplifier circuit) corresponds to data “1”.

If data of the memory cell does not need to be rewritten, output from the inverting amplifier circuit is output to the writing bit line. As described above, output from the inverting amplifier circuit has the same phase as data which is initially held.

In this state, when the potential of the writing word line is set to H so that the writing transistor is turned on, the potential of the source of the writing transistor has the same phase as the potential of the writing bit line. That is, data which is equivalent to initially stored one is written. As a result, “data is not rewritten”.

Note that in the case where data of the memory cell needs to be rewritten, the data to be rewritten is output to the writing bit line and the writing transistor is turned on by setting the potential of the writing word line to H.

One embodiment of the present invention is a semiconductor memory device including one or more writing bit lines; one or more writing word lines; one or more reading bit lines; one or more reading word lines; one or more memory cells; and a mechanism in which a potential of each one of the reading bit lines is inverted and amplified to be supplied to corresponding one of the writing bit lines. Each of the memory cells includes a writing transistor, a reading transistor, and a capacitor. A source of the writing transistor, a gate of the reading transistor, and one electrode of the capacitor are connected to each other. A drain of the writing transistor is connected to one of the writing bit lines. A gate of the writing transistor is connected to one of the writing word lines. A drain of the reading transistor is connected to one of the reading bit lines. The other electrode of the capacitor is connected to one of the reading word lines.

Another embodiment of the present invention is a semiconductor memory device including two or more bit lines; two or more word lines; one or more memory cells; and a mechanism in which a potential of one of the bit lines is inverted and amplified to be supplied to another one of the bit lines. Each of the memory cells includes a writing transistor, a reading transistor, and a capacitor. A source of the writing transistor, a gate of the reading transistor, and one electrode of the capacitor are connected to each other. A drain of the writing transistor is connected to one of the bit lines. A gate of the writing transistor is connected to one of the word lines. A drain of the reading transistor is connected to another one of the bit lines. The other electrode of the capacitor is connected to another one of the word lines.

Another embodiment of the present invention is a driving method of the above semiconductor memory device, including the steps of charging the writing bit line and the reading bit line to different potentials; changing a potential of the reading word line; and outputting a potential whose phase is opposite to a phase of a potential of the reading bit line to the writing bit line with an inverting amplifier circuit.

In the above memory cell, the resistance of the writing transistor is extremely high in an off state; thus, charge accumulated in the capacitor is held for a sufficiently long period, so that frequent refresh operation is unnecessary unlike in a conventional DRAM. For example, when current flowing between the source and the drain in the writing transistor in an off state is 1×10⁻²⁶ A and the capacitance of the capacitor is 0.01 ff, charge can be held for 10 years or more.

When the potential of the reading bit line and that of the source of the reading transistor in a standby period are the same, power consumption of this portion is ideally 0 W. Further, as described above, leakage current through the capacitor is sufficiently low. Therefore, current consumed by one memory cell in a standby period can be 1×10⁻²⁰ A or less.

Further, as apparent from the above description, the number of transistors used in one memory cell is three or less, typically two. These are provided in different layers, whereby the area occupied by the memory cells can be reduced. Furthermore, as described above, the wirings are arranged so as to have a three-dimensional structure or the wiring is shared by different elements, whereby the number of wirings can be reduced. Consequently, the degree of integration can be further increased.

Note that even in the case where current flowing between the source and the drain in the writing transistor in an off state is not extremely low as described above, a semiconductor memory device whose degree of integration is sufficiently high can be manufactured as described in the following embodiments. In this semiconductor memory device, a capacitor having large capacitance is unnecessary unlike in a DRAM and a soft error does not easily occur.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGS. 1A to 1E illustrate an example of a semiconductor memory device according to the present invention and an example of a driving method thereof.

FIG. 2 illustrates an example of a semiconductor memory device according to the present invention.

FIGS. 3A to 3D illustrate an example of a driving method of a semiconductor memory device according to the present invention.

FIG. 4 illustrates an example of a semiconductor memory device according to the present invention.

FIG. 5 illustrates an example of a semiconductor memory device according to the present invention.

FIGS. 6A to 6D illustrate an example of a manufacturing process of a semiconductor memory device according to the present invention.

FIGS. 7A to 7C illustrate an example of a manufacturing process of a semiconductor memory device according to the present invention.

FIGS. 8A to 8F illustrate an example of a semiconductor memory device according to the present invention.

FIGS. 9A to 9E illustrate an example of a driving method of a semiconductor memory device according to the present invention.

FIG. 10 illustrates an example of a semiconductor memory device according to the present invention.

DETAILED DESCRIPTION OF THE INVENTION

Hereinafter, embodiments will be described with reference to the drawings. Note that the embodiments can be implemented in various modes, and it is easily understood by those skilled in the art that modes and details can be changed in various ways without departing from the spirit and scope of the present invention. Therefore, the present invention is not construed as being limited to the description of the embodiments below.

Note that specific values of potentials are given below for the purpose of aid for understanding a technical idea of the present invention. Needless to say, such values are changed depending on various characteristics of a transistor, a capacitor, or the like, or for convenience of the practitioner. Further, in the semiconductor memory device described in the embodiments, data can be written or read using a method other than a method described below.

For the purpose of aid for understanding, in some circuit diagrams, a cross mark on a transistor indicates that the transistor is in an off state, and a circle on a transistor indicates that the transistor is in an on state.

Embodiment 1

FIG. 1A illustrates a memory cell of this embodiment. Here, n and m are natural numbers of 1 or more. In FIG. 1A, a memory cell including a writing transistor WTr_n_m, a reading transistor RTr_n_m, and a capacitor CS_n_m is illustrated. Here, a source of the writing transistor WTr_n_m is connected to a gate of the reading transistor RTr_n_m and one electrode of the capacitor CS_n_m.

In the memory cell illustrated in FIG. 1A, the writing transistor WTr_n_m and the reading transistor RTr_n_m are each an n-channel transistor; however, without limitation to this, the following structures can be employed, for example: the writing transistor WTr_n_m and the reading transistor RTr_n_m are each a p-channel transistor; the writing transistor WTr_n_m and the reading transistor RTr_n_m are an n-channel transistor and a p-channel transistor, respectively; and the writing transistor WTr_n_m and the reading transistor RTr_n_m are a p-channel transistor and an n-channel transistor, respectively. Note that when the conductivity type of the transistor is changed, the potentials of a gate, a source, and a drain need to be changed accordingly.

A writing word line WWL_n and a reading word line RWL_n are in parallel, and a writing bit line WBL_m and a reading bit line RBL_m are in parallel. The writing word line WWL_n and the writing bit line WBL_m intersect with each other to form a matrix.

A gate of the writing transistor WTr_n_m is connected to the writing word line WWL_n, a drain of the writing transistor WTr_n_m is connected to the writing bit line WBL_m, a drain of the reading transistor RTr_n_m is connected to the reading bit line RBL_m, and the other electrode of the capacitor CS_n_m is connected to the reading word line RWL_n.

The potential of a source of the reading transistor RTr_n_m is held at a fixed potential (here, 0 V). Further, the potential of the writing bit line WBL_m and that of the reading bit line RBL_m are each 0 V or more. Note that here, the threshold value of the writing transistor WTr_n_m is assumed to be +1 V, and that of the reading transistor RTr_n_m is assumed to be +0.5 V.

In the memory cell illustrated in FIG. 1A, the potential of the writing word line WWL_n is set to H, whereby the writing transistor WTr_n_m is turned on. By the potential of the writing bit line WBL_m at this time, charge is injected into the capacitor CS_n_m. The amount of charge injected at this time is determined depending on the potential of the writing bit line WBL_m, the gate capacitance of the reading transistor RTr_n_m, the capacitance of the capacitor CS_n_m, and the like and the result is thus always almost the same in the case where the conditions are the same, and dispersion is small. In this manner, data is written.

Then, the potential of the writing word line WWL_n is set to L, whereby the writing transistor WTr_n_m is turned off. At this time, current flowing between the source and the drain of the writing transistor WTr_n_m is set to 1×10⁻²¹ A or less, preferably 1×10⁻²⁴ A or less, whereby charge in the capacitor CS_n_m can be held for an extremely long period.

At the time of data reading, an appropriate potential is applied to the reading word line RWL_n, and the state of the reading transistor RTr_n_m is monitored; thus, the written data can be found. Hereinafter, a specific example of data writing and data reading will be described with reference to FIGS. 1B to 1E.

In the following example, the gate capacitance of the reading transistor RTr is treated as much smaller than the capacitance of the capacitor CS. Therefore, in the case where the writing transistor WTr_n_m is off, when the potential of the reading word line RWL_n is decreased by 1 V, the potential of the gate of the reading transistor RTr_n_m is decreased by 1 V regardless of the state of the reading transistor RTr_n_m.

First, an example of a writing method will be described. The potential of the writing word line WWL_n is assumed to be +2 V, and the potential of the reading word line RWL_n is assumed to be 0 V. In the case where data “1” is written, the potential of the writing bit line WBL_m is set to +1 V. In the case where data “0” is written, the potential of the writing bit line WBL_m is set to 0 V. With this operation, the writing transistor WTr_n_m is turned on, and charge is accumulated in the capacitor CS_n_m (see FIG. 1B).

Note that at this time, when the potential of the reading bit line RBL_m is kept at 0 V, current does not flow between the source and the drain of the reading transistor RTr_n_m regardless of data to be written, which is effective for reducing power consumption. In a similar manner, when the potential of the reading bit line RBL_m is set to a potential whose phase is opposite to that of the potential of the writing bit line WBL_m (i.e., the potential of the reading bit line RBL_m is set to 0 V when the potential of the writing bit line WBL_m is +1 V, and the potential of the reading bit line RBL_m is set to +1 V when the potential of the writing bit line WBL_m is 0 V), current does not flow between the source and the drain of the reading transistor RTr_n_m.

After that, the potential of the writing word line WWL_n is set to −1 V, and in addition, the potential of the reading word line RWL_n is set to −1 V. With this operation, the writing transistor WTr_n_m is turned off, and charge in the capacitor CS_n_m is held. The potential of the gate of the reading transistor RTr_n_m (which is also the potential of the capacitor CS_n_m or the potential of the floating node) is decreased by 1 V from the written potential to 0 V or −1 V; thus, the reading transistor RTr_n_m is off regardless of written data.

Note that the writing bit line WBL_m is supplied with data which is to be written to a memory cell in another row, so that the potential of the writing bit line WBL_m fluctuates between 0 V and +1 V (see FIG. 1C). In a similar manner, the potential of the reading bit line RBL_m fluctuates between 0 V and +1 V in some cases.

Next, an example of a reading method will be described. First, the reading bit line RBL_m is charged to +1 V (see FIG. 1D). Charging a wiring before one operation in this manner is referred to as pre-charge. Then, the potential of the reading word line RWL_n is set to 0 V (which is equal to that at the time of data writing). Then, according to written data, the potential of the gate of the reading transistor RTr_n_m becomes +1 V (when data “1” is written) or 0 V (when data “0” is written), or becomes a value close thereto. In the former case, the reading transistor RTr_n_m is turned on, and in the latter case, the reading transistor RTr_n_m remains off.

When the reading transistor RTr_n_m is turned on, charge in the reading bit line RBL_m is given to the source (having a potential of 0 V) of the reading transistor RTr_n_m; thus, the potential of the reading bit line RBL_m becomes 0 V. On the other hand, when the reading transistor RTr_n_m is off, the potential of the reading bit line RBL_m is kept at +1 V. Therefore, by measuring the potential of the reading bit line RBL_m, held data can be judged (see FIG. 1E).

Here, the phase of the potential of the reading bit line RBL_m is opposite to that of the writing bit line WBL_m at the time of data writing. That is, when data “0” (data “1”) is written, the potential of the writing bit line WBL_m is set to 0 V (+1 V), and the potential of the reading bit line RBL_m at the time of reading data “0” (data “1”) is +1 V (0 V). Note that since the writing transistor WTr_n_m is kept off during the above reading operation, charge accumulated in the capacitor CS_n_m is held.

FIG. 2 illustrates an example of a circuit for driving a memory cell array including a plurality of memory cells a structure of which is the same as that of the memory cell illustrated in FIG. 1A. In this circuit, a drain of a second pre-charge transistor CTr2_—m is connected to a writing bit line WBL_m, and a drain of a first pre-charge transistor CTr1_—m is connected to a reading bit line RBL_m. The potential of a source of the first pre-charge transistor CTr1_—m is kept at +1 V, and a gate thereof is connected to a first pre-charge control line CL1. The potential of a source of the second pre-charge transistor CTr2_—m is kept at +0.5 V, and a gate thereof is connected to the first pre-charge control line CL1.

That is, the potential of the first pre-charge control line CL1 is set to H, whereby the potential of the writing bit line WBL_m can be set to +0.5 V and the potential of the reading bit line RBL_m can be set to +1 V.

The reading bit line RBL_m is also connected to a drain of a selection transistor STr_m. A source of the selection transistor STr_m is connected to one input/output terminal of a flip-flop circuit FF_m, and a gate thereof is connected to a data selection line SL0_—m. The potential of the data selection line SL0_—m is set to H, whereby the selection transistor STr_m is turned on, and the potential of the reading bit line RBL_m can be input to the flip-flop circuit FF_m.

The other input/output terminal of the flip-flop circuit FF_m is connected to the writing bit line WBL_m. Note that the high power supply potential of the flip-flop circuit FF_m is assumed to be +1 V and the low power supply potential thereof is assumed to be 0 V. The writing bit line WBL_m is also connected to a data input/output terminal DATA_m. In data reading, the potential of the data input/output terminal DATA_m is measured. As described above, although the phase of the potential of the reading bit line RBL_m is opposite to that of written data, a potential inverted by the flip-flop circuit FF_m (the phase of that inverted potential is the same as that of written data) is output to the writing bit line WBL_m and the data input/output terminal DATA_m.

In data writing, the potential of the data input/output terminal DATA_m is set according to the written data. Note that in a column where data is rewritten, the potential of the data input/output terminal DATA_m is preferably changed in the state where the selection transistor STr_m is turned off by setting the potential of the data selection line SL0_—m to L.

For example, the case where data “1” is stored in the memory cell and is rewritten to data “0” is described. In that case, the potential of the data input/output terminal DATA_m is set to 0 V in the state where the selection transistor STr_m is off. In the state where data “1” is stored in the memory cell, a reading transistor RTr_n_m is on. Therefore, the potential of the reading bit line RBL_m is 0 V.

Note that when the potential of the data input/output terminal DATA_m is set to 0 V in the state where the selection transistor STr_m is off, the potential of the reading bit line RBL_m is kept at 0 V and current does not flow between a source and a drain of the reading transistor RTr_n_m.

An example of a driving method in the case where such a driver circuit is used will be described with reference to FIGS. 3A to 3D. As described above, writing transistors WTr of a large number of memory cells are connected to a writing word line WWL. In some cases, data needs to be written to some of the memory cells, and does not need to be written to the others. When the potential of the writing word line WWL is set to H, all the writing transistors WTr connected to the writing word line WWL are turned on, and data of the memory cell for which data writing is unnecessary might be rewritten to false one.

In a semiconductor device having the circuit illustrated in FIG. 2, to a memory cell for which data writing is unnecessary, data which is equivalent to initially stored one is written. In that case, a process in which stored data is read is needed before data writing. Here, the case where data “1” is initially stored in the memory cell in the n-th row and the m-th column is described.

First, the writing bit line WBL_m and the reading bit line RBL_m are pre-charged to +0.5 V and +1 V, respectively (see FIG. 3A). For this, in the state where the selection transistor STr_m in FIG. 2 is off, the potential of the first pre-charge control line CL1 is set to H, and the first pre-charge transistor CTr1_—m and the second pre-charge transistor CTr2_—m are turned on.

Next, the potential of the reading word line RWL_n is set to 0 V. As a result, the potential of the gate of the reading transistor RTr_n_m becomes +1 V, and the reading transistor RTr_n_m is turned on. The potential of the reading bit line RBL_m is decreased from +1 V to 0 V (see FIG. 3B).

After that, the potential of the data selection line SL0_—m is set to H, and the selection transistor STr_m is turned on. Since the selection transistor STr_m is turned on, the potential of the reading bit line RBL_m is input to the flip-flop circuit FF_m. Here, since the potential of the reading bit line RBL_m (0 V) is lower than that of the writing bit line WBL_m (+0.5 V), the potential of the reading bit line RBL_m and that of the writing bit line WBL_m become 0 V and +1 V, respectively, by the effect of the flip-flop circuit FF_m. The potential of the data input/output terminal DATA_m connected to the writing bit line WBL_m becomes +1 V (see FIG. 3C).

In this state, when the potential of the writing word line WWL_n is set to +2 V, the writing transistor WTr_n_m is turned on and the capacitor CS_n_m is charged to +1 V. That is, data which is equivalent to initially stored one is written.

Note that in this process, although the reading transistor RTr_n_m is on, current does not flow between the source and the drain because the potential of the source and that of the drain are equal to each other (0 V).

The case where data “1” is initially stored is described above; however, also in the case where data “0” is initially stored, the writing bit line WBL_m has a potential corresponding to data which is initially stored (i.e., 0 V) (see FIG. 3D).

Then, when the potential of the writing word line WWL_n is set to +2 V, the writing transistor WTr_n_m is turned on and the capacitor CS_n_m is charged to 0 V. That is, data which is equivalent to initially stored one is written.

At this time, current does not flow between the source and the drain of the reading transistor RTr_n_m. The reason is that, in this case, the potential of the reading bit line RBL_m is kept at the pre-charged potential, i.e., +1 V and there is a potential difference between the source and the drain of the reading transistor RTr_n_m, but the reading transistor RTr_n_m is off because the potential of the gate of the reading transistor RTr_n_m is 0 V.

The case where data is not rewritten is described above, and in the case where data is rewritten, the following process may be performed. First, as illustrated in FIG. 3A, the writing bit line WBL_m and the reading bit line RBL_m are pre-charged to +0.5 V and +1 V, respectively. Note that data does not need to be read; thus, in the case where pre-charge can be controlled per column, pre-charge is not performed in a column where data is rewritten, which leads to a reduction in power consumption.

After that, in the state where the selection transistor is kept off, the potential of the data input/output terminal DATA_m is set to one corresponding to data to be written. The potential of the writing bit line WBL_m is also set to one corresponding to data to be written. In this state, when the potential of the writing word line WWL_n is set to +2 V, the writing transistor WTr_n_m is turned on and the capacitor CS_n_m is charged to a potential corresponding to written data.

Embodiment 2

FIG. 4 is a circuit diagram illustrating part of a memory cell array of a semiconductor memory device in this embodiment, and FIG. 5 is a circuit diagram illustrating part of a driver circuit of a semiconductor memory device in this embodiment. FIGS. 9A to 9E illustrate an example of a driving method of a semiconductor memory device in this embodiment.

In the semiconductor memory device in this embodiment, as illustrated in FIG. 4, a plurality of memory cells including a memory cell in the (n−1)-th row and the m-th column, a memory cell in the n-th row and the (m−1)-th column, and the like are arranged in a matrix. Here, n and m are each an even number of 2 or more.

In the memory cell in the n-th row and the (m−1)-th column, a source of a writing transistor WTr_n_m−1, a gate of a reading transistor RTr_n_m−1, and one electrode of a capacitor CS_n_m−1 are connected to each other. In the memory cell in the (n−1)-th row and the m-th column, a source of a writing transistor WTr_n−1_m, a gate of a reading transistor RTr_n−1_m, and one electrode of a capacitor CS_n−1_m are connected to each other.

Here, as the writing transistor WTr, a transistor whose characteristics are similar to those of the writing transistor WTr in Embodiment 1 may be used. As the reading transistor RTr, a transistor having the opposite conductivity type (here, a p-channel transistor) to the writing transistor WTr is used.

A gate of the writing transistor WTr_n_m−1 and the other electrode of the capacitor CS_n−1_m are connected to a word line WL_n in the n-th row. A gate of the writing transistor WTr_n−1_m and the other electrode of the capacitor CS_n_m−1 are connected to a word line WL_n−1 in the (n−1)-th row. A drain of the writing transistor WTr_n_m−1 and a drain of the reading transistor RTr_n−1_m are connected to a bit line BL_m−1 in the (m−1)-th column. A drain of the writing transistor WTr_n−1_m and a drain of the reading transistor RTr_n_m−1 are connected to a bit line BL_m in the m-th column.

The word line WL in FIG. 4 functions as the writing word line WWL and the reading word line RWL in FIGS. 1A to 1E. The bit line BL in FIG. 4 functions as the writing bit line WBL and the reading bit line RBL in FIGS. 1A to 1E. Therefore, the number of wirings can be reduced, and the degree of integration can be increased.

Specifically, for the memory cell in the n-th row and the (m−1)-th column, the word line WL_n, the word line WL_n−1, the bit line BL_m−1, and the bit line BL_m correspond to the writing word line WWL_n, the reading word line RWL_n, the writing bit line WBL_m, and the reading bit line RBL_m in FIGS. 1A to 1E, respectively.

FIG. 5 illustrates part of a circuit for driving the memory cell array in FIG. 4. A drain of a first pre-charge transistor CTr1_—m−1 and a drain of a second pre-charge transistor CTr2_—m−1 are connected to a bit line BL_m−1. A drain of a first pre-charge transistor CTr1_—m and a drain of a second pre-charge transistor CTr2_—m are connected to a bit line BL_m.

Similarly, a drain of a first pre-charge transistor CTr1_—m+1 and a drain of a second pre-charge transistor CTr2_—m+1 are connected to a bit line BL_m+1, and a drain of a first pre-charge transistor CTr1_—m+2 and a drain of a second pre-charge transistor CTr2_—m+2 are connected to a bit line BL_m+2.

A gate of the first pre-charge transistor CTr1_—m−1 and a gate of the second pre-charge transistor CTr2_—m are connected to a first pre-charge control line CL1, and operate in conjunction with each other. Similarly, a gate of the second pre-charge transistor CTr2_—m−1 and a gate of the first pre-charge transistor CTr1_—m are connected to a second pre-charge control line CL2.

Similarly, a gate of the first pre-charge transistor CTr1_—m+1 and a gate of the second pre-charge transistor CTr2_—m+2 are connected to the first pre-charge control line CL1, and operate in conjunction with each other. Similarly, a gate of the second pre-charge transistor CTr2_—m+1 and a gate of the first pre-charge transistor CTr1_—m+2 are connected to the second pre-charge control line CL2.

Note that the potentials of sources of the first pre-charge transistors CTr1_—m−1, CTr1_—m, CTr1_—m+1, and CTr1_—m+2 are kept at 0 V, and the potentials of sources of the second pre-charge transistors CTr2_—m−1, CTr2_—m, CTr2_—m+1, and CTr2_—m+2 are kept at +0.5 V.

Therefore, when the potential of the first pre-charge control line CL1 is set to H, the potentials of the bit lines BL_m−1 and BL_m+1 become 0 V and the potentials of the bit lines BL_m and BL_m+2 become +0.5 V. When the potential of the second pre-charge control line CL2 is set to H, the potentials of the bit lines BL_m−1 and BL_m+1 become +0.5 V and the potentials of the bit lines BL_m and BL_m+2 become 0 V.

Drains of selection transistors STr_m−1 and STr_m are connected to the bit lines BL_m−1 and BL_m, respectively. Sources of the selection transistors STr_m−1 and STr_m are connected to two input terminals of a flip-flop circuit FF_m, respectively. The sources of the selection transistors Str_m−1 and STr_m are also connected to data input/output terminals DATA_m−1 and DATA_m, respectively.

Similarly, drains of selection transistors STr_m+1 and STr_m+2 are connected to the bit lines BL_m+1 and BL_m+2, respectively. Sources of the selection transistors STr_m+1 and STr_m+2 are connected to two input terminals of a flip-flop circuit FF_m+2, respectively. The sources of the selection transistors Str_m+1 and STr_m+2 are also connected to data input/output terminals DATA_m+1 and DATA_m+2, respectively.

Gates of the selection transistors STr_m−1, STr_m, STr_m+1, and STr_m+2 are connected to a first data selection line SL1; thus, when the potential of the first data selection line SL1 is set to H, the selection transistors STr_m−1, STr_m, STr_m+1, and STr_m+2 can be turned on. In such a manner, the bit line BL and the flip-flop circuit FF can be connected to each other. The high power supply potential of the flip-flop circuit FF is assumed to be +1 V and the low power supply potential of the flip-flop circuit FF is assumed to be 0 V.

An operation example of such a circuit will be described with reference to FIGS. 9A to 9E. Here, an example of operation of the memory cell in the (n−1)-th row and the m-th column and the memory cell in the n-th row and the (m−1)-th column will be described. In the operation described below, the potential of the bit line BL is higher than or equal to 0 V. The threshold voltage of the writing transistor WTr is assumed to be +1 V, and the threshold voltage of the reading transistor RTr is assumed to be −0.5 V. Note that the potential of the source of the reading transistor RTr is kept at a constant potential (here, +1 V) in the operation described below.

First, reading operation will be described. The case where data “1” is initially stored in the memory cell in the (n−1)-th row and the m-th column and data “0” is initially stored in the memory cell in the n-th row and the (m−1)-th column is described. In a holding state, as illustrated in FIG. 9A, the potential of the word line WL_n−1 and that of the word line WL_n are assumed to be −1 V. As described below, since the potential of the word line WL connected to the capacitor CS at the time of data writing is −2 V, the potential of the gate of the reading transistor RTr in the memory cell where data “1” has been stored is +2 V, and that in the memory cell where data “0” has been stored is +1 V.

Therefore, as illustrated in FIG. 9A, the potential of the gate of the reading transistor RTr_n_m−1 is +1 V, the potential of the gate of the reading transistor RTr_n−1_m is +2 V, and both of the reading transistors are off. Further, the writing transistors WTr_n−1_m and WTr_n_m−1 are also off.

Before data in the memory cell in the n-th row and the (m−1)-th column is read, as illustrated in FIG. 9A, the bit line BL_m−1 and the bit line BL_m are pre-charged to +0.5 V and 0 V, respectively. For this, the potential of the second pre-charge control line CL2 in FIG. 5 may be set to H.

Next, the potential of the word line WL_n−1 is set to −2 V. As a result, the potential of the gate of the reading transistor RTr_n_m−1 becomes 0 V, and the reading transistor RTr_n_m−1 is turned on. Then, charge is supplied from the source of the reading transistor RTr_n_m−1 to the bit line BL_m, so that the potential of the bit line BL_m is increased from 0 V to +1 V (see FIG. 9B).

Here, the potential of the first data selection line SL1 in FIG. 5 is set to H and the flip-flop circuit FF_m is connected to the bit lines BL_m−1 and BL_m. In that case, the potential of the input terminal of the flip-flop circuit FF_m, which is connected to the bit line BL_m having a higher potential, becomes a higher potential (+1 V). Further, the potential of the input terminal of the flip-flop circuit FF_m, which is connected to the bit line BL_m−1 having a lower potential, becomes a lower potential (0 V). As a result, the data input/output terminal DATA_m−1 has a potential corresponding to data in the memory cell in the n-th row and the (m−1)-th column (i.e., 0 V) (see FIG. 9C).

The case where data “0” has been stored in the memory cell in the n-th row and the (m−1)-th column is described above. Further, also in the case where data “1” has been stored, the data input/output terminal DATA_m−1 has a potential corresponding to data in the memory cell (i.e., +1 V). That is to say, in that case, in the above process, the reading transistor RTr_n_m−1 (the potential of the gate is +1 V) remains off, and the potential of the bit line BL_m is kept at 0 V and is lower than that of the bit line BL_m−1 (+0.5 V); therefore, owing to the operation of the flip-flop circuit FF_m, the potential of the bit line BL_m becomes 0 V and the potential of the bit line BL_m−1 (i.e., the potential of the data input/output terminal DATA_m−1) becomes +1 V.

The reading operation is thus completed. Next, writing operation will be described. As in Embodiment 1, data reading is performed before data writing. The process thereof is as described above.

If data in the memory cell in the n-th row and the (m−1)-th column does not need to be rewritten, the potential of the word line WL_n is set to +2 V. Then, the writing transistor WTr_n_m−1 is turned on, and the potential of the capacitor CS_n_m−1 becomes 0 V as illustrated in FIG. 9D. At this time, although the reading transistor RTr_n_m−1 is on, current does not flow between the source and the drain because both the potential of the source and that of the drain are +1 V.

In the case where data “1” has been stored in the memory cell in the n-th row and the (m−1)-th column, although the potential of the drain of the reading transistor RTr_n_m−1 (the potential of the bit line BL_m) is 0 V and is different from that of the source (+1 V), current does not flow between the source and the drain because the reading transistor RTr_n_m−1 is off (see FIG. 9E).

If data in the memory cell in the n-th row and the (m−1)-th column is rewritten, in the state where the potential of the word line WL_n is set to +2 V and the writing transistor WTr_n_m−1 is turned on, the data input/output terminal DATA_m−1 in FIG. 5 may have a potential corresponding to rewritten data. At this time, when the data input/output terminal DATA_m has a potential corresponding to data which is inverted from data to be written (that is, it has a potential of +1 V in the case where data “0” is to be written, and it has a potential of 0 V in the case where data “1” is to be written), data writing can be performed stably.

Embodiment 3

In this embodiment, examples of a layout and a manufacturing method of the semiconductor memory device described in Embodiment 2 will be described with reference to FIGS. 6A to 6D, FIGS. 7A to 7C, and FIGS. 8A to 8F. FIGS. 6A to 6D and FIGS. 7A to 7C are cross-sectional views illustrating a manufacturing process, and FIGS. 8A to 8F illustrate a layout of components such as main wirings and the like in main layers. Note that line A-B denotes the same position through FIGS. 8A to 8F. Further, dotted lines in FIGS. 8A to 8F denote coordinates, and can be used as reference when a positional relationship between structures in different layers is referred to.

FIG. 8A illustrates shapes of element isolation insulators 102 provided over a semiconductor substrate. The element isolation insulators 102 have a C-like shape. A continuous region 101a is formed in a direction intersecting with line A-B in the drawing, and this region is to be an impurity region 104a functioning as a wiring. The unit memory cell of the semiconductor memory device described in this embodiment occupies a region denoted by dashed-dotted line in FIG. 8A.

FIG. 8B illustrates a layout of floating gates 103 and first contact holes 106. The first contact holes 106 are each provided in the center of the C-like shape of the element isolation insulator 102. Each of the floating gates 103 is provided so as to overlap with two element isolation insulators.

FIG. 8C illustrates a layout of interlayer wirings 107 provided in contact with the floating gates 103 and the first contact holes 106. FIG. 8D illustrates a layout of oxide semiconductor layers 109 provided in contact with the interlayer wirings 107. FIG. 8E illustrates a layout of word lines 111 and second contact holes 113. FIG. 8F illustrates a layout of bit lines 114. The second contact holes 113 may be provided in the substantially same positions as the respective first contact holes 106.

A manufacturing process of a semiconductor memory device having the layout structure illustrated in FIGS. 8A to 8F will be described below with reference to FIGS. 6A to 6D and FIGS. 7A to 7C. Note that FIGS. 6A to 6D and FIGS. 7A to 7C are cross-sectional views taken along line A-B in FIGS. 8A to 8F.

<FIG. 6A>

By a known semiconductor processing technique, the element isolation insulator 102 is formed on one surface of a substrate 101 of a single crystal semiconductor such as silicon or gallium arsenide. As described above, the region 101a denoted by dotted line in FIG. 6A is to be the impurity region 104a functioning as a wiring.

<FIG. 6B>

By a known semiconductor processing technique, the floating gate 103 and a p-type impurity region 104 are formed. In addition, a first interlayer insulator 105 is formed. Note that part of the p-type impurity region 104 (denoted by dotted line in FIG. 6B) is the impurity region 104a functioning as a wiring. The impurity region 104a functioning as a wiring extends in a direction intersecting with line A-B (that is, a direction of the word line 111).

<FIG. 6C>

The first interlayer insulator 105 is planarized by a means such as chemical mechanical polishing (CMP). This planarization may be stopped in the state where the floating gate 103 is exposed. A planarized first interlayer insulator 105a is obtained in this manner. Further, the planarized first interlayer insulator 105a is etched, whereby the first contact hole 106 is formed.

<FIG. 6D>

The interlayer wiring 107 and an embedded insulator 108 are formed. A forming method of the first interlayer insulator may be referred to for a forming method of the embedded insulator 108, and planarization treatment may be performed so as to expose a surface of the interlayer wiring 107. The embedded insulator 108 is preferably formed using silicon oxide. The thickness of the embedded insulator 108 is 100 nm to 500 nm, and in a region having a thickness of 100 nm from a surface of the embedded insulator 108, the hydrogen concentration may be lower than 1×10¹⁸ cm⁻³, preferably lower than 1×10⁻¹⁷ cm⁻³.

<FIG. 7A>

The oxide semiconductor layer 109 and a gate insulator 110 covering the oxide semiconductor layer 109 are formed. For the oxide semiconductor layer 109, an oxide semiconductor in which indium accounts for 20 at % or more of all metal elements is preferably used. The thickness thereof is 1 nm to 20 nm, preferably 1 nm to 10 nm. As for dispersion in thickness, the root-mean-square (RMS) may be set to 0.01 nm to 1 nm.

At the time of formation of the oxide semiconductor layer 109, attention needs to be paid to prevent mixture of hydrogen, and deposition of the oxide semiconductor is preferably performed by a sputtering method in which hydrogen and water in an atmosphere and a target are sufficiently reduced. The hydrogen concentration in the oxide semiconductor layer 109 may be lower than 1×10¹⁸ cm⁻³, preferably lower than 1×10⁻¹⁷ cm⁻³. Note that Patent Document 2 can be referred to for the oxide semiconductor layer 109 and a forming method thereof.

As a material for the gate insulator 110, silicon oxide, silicon oxynitride, aluminum oxide, hafnium oxide, zirconium oxide, or the like can be used. The thickness thereof is 6 nm to 20 nm, preferably 10 nm to 16 nm. As for dispersion in thickness, the root-mean-square (RMS) may be set to 0.01 nm to 1 nm. The hydrogen concentration in the gate insulator 110 may be lower than 1×10¹⁸ cm⁻³, preferably lower than 1×10⁻¹⁷ cm⁻³.

<FIG. 7B>

The word line 111 is formed. As a material for the word line 111, a material whose work function is larger than the electron affinity of the oxide semiconductor, such as tungsten, tungsten nitride, platinum, palladium, nickel, or indium nitride, is preferably used. Alternatively, only part of the word line 111, which is in contact with the gate insulator 110, may be formed using such a material.

Further, the second interlayer insulator 112 is formed, and the second interlayer insulator 112, the gate insulator 110, and the oxide semiconductor layer 109 are etched. Thus, the second contact hole 113 is formed.

<FIG. 7C>

The bit line 114 is formed. Thus, a writing transistor 115, a reading transistor 116, and a capacitor 117 can be formed. As illustrated in FIG. 7C, in the memory cell described in this embodiment, most of a portion where the oxide semiconductor layer 109 and the word line 111 overlap with each other (a channel portion of the writing transistor 115) is formed over the element isolation insulator 102.

If the channel portion of the writing transistor 115 is formed over the impurity region 104, the writing transistor 115 is turned on or substantially turned on due to a change in potential of the impurity region 104; however, such a problem does not occur by employing the layout described in this embodiment. That is, even when the planarized first interlayer insulator 105a and embedded insulator 108 each have a small thickness of 200 nm or less, charge holding operation of the memory cell is not affected.

When F is used to express the minimum feature size, the area of one memory cell in the semiconductor memory device disclosed in this embodiment can be expressed as 8F², which means that this semiconductor memory device has the same degree of integration as a DRAM. In addition, a capacitor having large capacitance is unnecessary unlike in a DRAM. With the use of a writing transistor in which the amount of current flowing between a source and a drain in an off state is extremely small as described above, the interval between refresh operations can be sufficiently long, or the refresh operation can be substantially unnecessary.

Although an oxide semiconductor is employed as the semiconductor used in the writing transistor in the above example, another kind of semiconductor may be alternatively used. For example, a polycrystalline silicon film which is formed by laser crystallization or a single crystal silicon film may be used.

Embodiment 4

In Embodiment 1 and Embodiment 2, the writing bit line WBL_m or the bit line BL_m−1 is pre-charged to +0.5 V at the time of data reading; in this embodiment, an example of a driving method in which such pre-charge is unnecessary and an example of a driver circuit therefor will be described. Since pre-charge operation is unnecessary, power consumption can be reduced.

FIG. 10 illustrates part of a driver circuit of a semiconductor memory device used in this embodiment. Here, m is an even number of 2 or more. The memory cell array illustrated in FIG. 4 is used. To each bit line BL, a drain of a first pre-charge transistor CTr1 for pre-charging the bit line BL to 0 V is connected. Sources of the first pre-charge transistors CTr1 are kept at +1 V. Gates of the first pre-charge transistors CTr1 in odd-numbered columns are connected to a first pre-charge control line CL1, and gates thereof in even-numbered columns are connected to a second pre-charge control line CL2.

That is, when the potential of the first pre-charge control line CL1 is set to H, the bit lines in the odd-numbered columns are pre-charged to 0 V; when the potential of the second pre-charge control line CL2 is set to H, the bit lines in the even-numbered columns are pre-charged to 0 V.

The bit line BL_m−1 is connected to a drain of a selection transistor STr_m−1 and an output terminal of an inverter INV_m, and a source of the selection transistor STr_m−1 is connected to an input terminal of an inverter INV_m−1.

On the other hand, the bit line BL_m is connected to a drain of a selection transistor STr_m and an output terminal of the inverter INV_m−1, and a source of the selection transistor STr_m is connected to an input terminal of the inverter INV_m. Note that the high power supply potential of the inverter is assumed to be +1 V, and the low power supply potential thereof is assumed to be 0 V.

A gate of the selection transistor STr_m−1 is connected to a first data selection line SL1, and a gate of the selection transistor STr_m is connected to a second data selection line SL2. Therefore, when the potential of the first data selection line SL1 is set to H, the selection transistor STr_m−1 is turned on; when the potential of the second selection line SL2 is set to H, the selection transistor STr_m is turned on.

In a similar manner, the bit line BL_m+1, the bit line BL_m+2, a selection transistor STr_m+1, a selection transistor STr_m+2, the first data selection line SL1, the second data selection line SL2, an inverter INV_m+1, and an inverter_m+2 also form the above connection relations. Note that each bit line BL is connected to a data input/output terminal DATA.

Data reading is performed as follows. For example, in reading data of a memory cell in the n-th row and the (m−1)-th column, the m-th column is pre-charged to 0 V first. This operation can be performed as described above by setting the potential of the second pre-charge control line CL2 to H to turn on the first pre-charge transistor.

Next, the potential of the word line WL_n−1 is set to −2 V as described in Embodiment 2, whereby the state of a reading transistor RT_n_m−1 is changed. When data “1” is stored in the memory cell, the potential of the bit line BL_m is not changed because the reading transistor RT_n_m−1 is off; however, when data “0” is stored in the memory cell, the potential of the bit line BL_m is increased from 0 V to +1 V because the reading transistor RT_n_m−1 is turned on.

Then, the potential of the second data selection line SL2 in FIG. 10 is set to H, whereby the selection transistor STr_m is turned on. As a result, the potential of the bit line BL_m is input to the inverter INV_m. From the inverter INV_m, a potential whose phase is inverted from that of the potential of the bit line BL_m is output to the bit line BL_m−1. That is, the potential of the bit line BL_m−1 becomes +1 V when the potential of the bit line BL_m is 0 V, and the potential of the bit line BL_m−1 becomes 0 V when the potential of the bit line BL_m is +1 V.

In data reading, the potential of the data input/output terminal DATA_m−1 at this time may be read. Further, in the case where data writing is performed and data in the memory cell does not need to be rewritten, a writing transistor WTr_n_m−1 may be turned on by setting the potential of a word line WL_n to +2 V.

On the other hand, in the case where data writing is performed and data in the memory cell is rewritten, after the writing transistor WTr_n_m−1 is turned on, the potential of the data input/output terminal DATA_m−1 in FIG. 10 may be set to one corresponding to rewritten data, or may be set to one corresponding to data which is inverted from data to be written (that is, it may be set to a potential of +1 V in the case where data “0” is to be written, and it may be set to a potential of 0 V in the case where data “1” is to be written). This application is based on Japanese Patent Application serial no. 2010-249435 filed with Japan Patent Office on Nov. 8, 2010, the entire contents of which are hereby incorporated by reference.

Claims

1. (canceled)

2. A semiconductor memory device comprising:

a writing bit line;

a writing word line;

a reading bit line;

a reading word line;

a memory cell;

an inverting amplifier circuit configured to supply an inverted and amplified potential of the reading bit line to the writing bit line; and

a switch provided between the reading bit line and an input terminal of the inverting amplifier circuit,

wherein the memory cell comprises a writing transistor, a reading transistor, and a capacitor comprising a first electrode and a second electrode,

wherein the switch is configured to control the connection between the reading bit line and the input terminal of the inverting amplifier circuit, depending on whether the memory cell is rewritten with new data or not,

wherein one of a source and a drain of the writing transistor, a gate of the reading transistor, and the first electrode of the capacitor are connected to each other,

wherein the other of the source and the drain of the writing transistor is connected to the writing bit line,

wherein a gate of the writing transistor is connected to the writing word line,

wherein one of a source and a drain of the reading transistor is connected to the reading bit line, and

wherein the second electrode of the capacitor is connected to the reading word line.

3. The semiconductor memory device according to claim 2, wherein the writing transistor and the reading transistor are provided in different layers.

4. The semiconductor memory device according to claim 2, wherein a kind of semiconductor used in the writing transistor and a kind of semiconductor used in the reading transistor are different from each other.

5. The semiconductor memory device according to claim 2, wherein the inverting amplifier circuit is a flip-flop circuit.

6. The semiconductor memory device according to claim 2, wherein the inverting amplifier circuit is an inverter.

7. A driving method of the semiconductor memory device according to claim 2, comprising:

pre-charging the other of the source and the drain of the writing transistor and the other of the source and the drain of the reading transistor to different potentials;

changing a potential of the reading word line; and

outputting a potential whose phase is opposite to a phase of a potential of the other of the source and the drain of the reading transistor to the other of the source and the drain of the writing transistor with the inverting amplifier circuit.

8. A semiconductor memory device comprising:

a writing bit line;

a writing word line;

a reading bit line;

a memory cell;

an inverting amplifier circuit configured to supply an inverted and amplified potential of the reading bit line to the writing bit line,

a switch provided between the reading bit line and an input terminal of the inverting amplifier circuit;

wherein the memory cell comprises a writing transistor and a reading transistor,

wherein the switch is configured to control the connection between the reading bit line and the input terminal of the inverting amplifier circuit, depending on whether the memory cell is rewritten with new data or not,

wherein one of a source and a drain of the writing transistor and a gate of the reading transistor are connected to each other,

wherein the other of the source and the drain of the writing transistor is connected to the writing bit line,

wherein a gate of the writing transistor is connected to the writing word line,

wherein one of a source and a drain of the reading transistor is connected to the reading bit line.

9. The semiconductor memory device according to claim 8, wherein the writing transistor and the reading transistor are provided in different layers.

10. The semiconductor memory device according to claim 8, wherein a kind of semiconductor used in the writing transistor and a kind of semiconductor used in the reading transistor are different from each other.

11. The semiconductor memory device according to claim 8, wherein the inverting amplifier circuit is a flip-flop circuit.

12. The semiconductor memory device according to claim 8, wherein the inverting amplifier circuit is an inverter.

13. A driving method of the semiconductor memory device according to claim 8, comprising:

pre-charging the reading bit line and the write line to different potentials;

turning on the switch after the pre-charging; and

outputting a potential whose phase is opposite to a phase of a potential of the reading bit line to the write bit line with the inverting amplifier circuit after the turning on the switch.