JP6094582B2 - Semiconductor device and programming method - Google Patents

Description

  The present invention relates to a semiconductor device and a programming method thereof, and more particularly to a semiconductor device having a variable resistance nonvolatile element (hereinafter referred to as a “resistance change switch”) and a programming method thereof.

  Silicon integrated circuits (LSIs: Large Scale Integration) have been miniaturized and integrated according to scaling rules, and high density and high performance of semiconductor elements have been realized. On the other hand, development costs have soared with miniaturization and initial investment has become high. The development cost includes an original plate (reticle) cost for manufacturing an LSI, a design cost, a manufacturing cost, a re-spin cost when there is a bug, and the like. Since general-purpose products such as processors and memories are produced in large quantities, it is easy to recover initial costs. However, it is not easy to recover the initial cost in a custom LSI that is designed according to customer specifications and does not have many shipments.

  As one method for reducing the initial investment cost of a small quantity, high variety of custom LSI, an FPGA (Field-Programmable Gate Array) may be used. The FPGA is equipped with a logic circuit that can be rewritten after manufacture, and becomes an LSI capable of realizing a function that satisfies customer requirements by programming. A combination of SRAM (Static Random Access Memory) and a transistor is used for switching the circuit, and a highly integrated and highly reliable switch is realized without using a special manufacturing process. On the other hand, there are problems such as a minute current flowing when information is held, volatility, and a large switch area. For this reason, the chip area is larger than that of the custom LSI, and the cost of the chip is high as a price for reducing the initial investment cost. In addition, the operating speed decreases and the power consumption increases.

  As a method for solving these problems, a nonvolatile variable resistance change switch may be used as proposed by the inventors. According to Non-Patent Document 1, it has been demonstrated that the area and power consumption of an FPGA can be reduced to one third by using a resistance change switch.

  FIG. 1A shows current characteristics of the resistance change switch. The resistance change switch 10 basically has a structure including the first electrode 101, the resistance change layer 103 (insulator), and the second electrode 102 (FIG. 1B). The resistance change switch 10 can transition between a low resistance state (ON state) and a high resistance state (OFF state) by applying a voltage. In FIG. 1A, when a positive voltage is applied to the first electrode 101, the resistance change switch 10 is set from the OFF state to the ON state using a desired set voltage as a threshold voltage. Here, applying a voltage in the direction of setting to the ON state is referred to as applying a voltage in the forward direction. Subsequently, in the resistance change switch 10 in the ON state, when a positive voltage is applied to the first electrode 101 again, a linear current / voltage characteristic as seen in the electric conduction characteristic of the first electrode 101 is obtained. On the other hand, when a negative voltage is applied to the first electrode 101, the voltage is reset from the ON state to the OFF state using a certain voltage as a threshold voltage. Here, applying a voltage in the direction of resetting to the OFF state is referred to as applying a voltage in the reverse direction.

  The operation for setting from the OFF state to the ON state is called a set operation, the operation for resetting from the ON state to the OFF state is called a reset operation, and voltages at which each transition occurs are called a set voltage and a reset voltage. The resistance change switch 10 is set from the OFF state to the ON state only when a positive voltage is applied to the first electrode 101, and from the ON state to the OFF state only when a negative voltage is applied to the first electrode 101. A reset operation occurs. Thus, the resistance change switch that transitions between the ON and OFF states depending on the polarity of the applied voltage is called a bipolar resistance change switch. The first electrode 101 is an electrode that is reset to an ON or OFF state by applying a positive or negative voltage, and the second electrode 102 is an electrode facing the first electrode 101.

  An example of the bipolar resistance change switch is disclosed in Non-Patent Document 1. Non-Patent Document 1 discloses a resistance change switch using metal ion movement and an electrochemical reaction in an ionic conductor (a solid in which ions can move freely by application of an electric field or the like). The switch disclosed in Non-Patent Document 1 is configured to include an ion conductive layer and two electrodes provided to face each other with the ion conductive layer interposed therebetween. Among these, one electrode plays a role for supplying metal ions to the ion conductive layer, and is called an “active electrode”. Metal ions are not supplied from one electrode to the ion conductive layer, and are called “inert electrodes”. The active electrode corresponds to the first electrode 101 and the inactive electrode corresponds to the second electrode 102 from the equivalence of the current-voltage characteristics of the switch.

  The operation of the switch using this electrochemical reaction will be briefly described. When the inactive electrode (corresponding to the second electrode 102) is grounded and a positive voltage is applied to the active electrode (corresponding to the first electrode 101), the metal of the active electrode becomes a metal ion and the ion conductive layer (resistance change layer 103). Equivalent). And the metal ion in an ion conductive layer turns into a metal in an ion conductive layer, precipitates, and the metal bridge | crosslinking which connects an inactive electrode and an active electrode with the deposited metal is formed. When both electrodes are electrically connected by metal bridge, the switch is turned on. On the other hand, when the inactive electrode is grounded and a negative voltage is applied to the active electrode in the ON state, a part of the metal bridge is cut. Thereby, the electrical connection between both electrodes is cut off, and the switch is turned off. It should be noted that the electrical characteristics change from the stage before the electrical connection is completely broken, such as the resistance between the two electrodes increases or the capacitance between the electrodes changes, and the electrical connection is finally broken. In order to change from the OFF state to the ON state, the inactive electrode is grounded again and a positive voltage is applied to the active electrode.

  According to Non-Patent Document 1, the resistance change type switch is non-volatile that is smaller in size than a combination of an SRAM and a transistor, which is a switch used in an FPGA, and does not require power to maintain the ON and OFF states. Have Therefore, the resistance change type switch is considered to be promising for application to FPGA. In addition, application as a non-volatile memory element is conceivable by utilizing the non-volatility of this switch.

  In the FPGA, a crossbar switch having a resistance change switch as an element is used. The crossbar switch is a maximum disposed at the intersection of each wiring to switch connection / disconnection of N first wiring groups and M second wiring groups, one set of first wiring and second wiring. It consists of N × M switches.

  In the following, as shown in FIG. 2, for simplification, three first wirings 21A, 21B, 21C and three second wiring groups 22A, 22B, 22C, 3 × 3 resistance change switches 20A˜ Consider a crossbar switch composed of 20I. Here, taking the resistance change switch 20G as an example, the first node 23 connected to the first electrode is connected to the first wiring 21A, and the second node 24 connected to the second electrode is connected to the second wiring 22C. It is connected. In the 3 × 3 crossbar switch, in order to transmit signals from the first wiring group to the second wiring group, a maximum of three switches are set to the ON state. Of the three switches connected to one first wiring or second wiring, two or more switches are not set. By setting the voltage applied to the first wiring and the second wiring to any one of VP, 1 / 2VP, and 0V, one switch can be set or reset.

  In FIG. 2, a method for programming a 3 × 3 crossbar switch will be described. Assume that all the switches in the 3 × 3 crossbar switch are in the OFF state. 0V and VP are applied to the first wiring 21B and the second wiring 22B connected to the resistance change switch 20E of the target that performs the set operation, respectively. A voltage half the VP is applied to the other first wirings 21A and 21C and the second wirings 22A and 22C.

  As a result, VP is applied to the first node 23 and 0 V is applied to the second node 24 of the resistance change switch 20E selected as the target. Further, in the resistance change switches 20B and 20H that are connected to the first wiring 21B and are not selected, a voltage of 1/2 VP is applied to the potential of the first electrode with respect to the second electrode. In addition, in the resistance change switches 20D and 20F that are connected to the second wiring 22B and are not selected, a voltage of ½ VP is applied to the potential of the first electrode with respect to the second electrode. Further, in the non-selected resistance change switches 20A, 20G, 20C, and 20I other than the above, the voltage between the second electrode and the first electrode is 0V. If the threshold voltage when the resistance change switch is programmed is between VP and 1 / 2VP, only the resistance change switch 20E selected as the target is set.

  In order to erase the resistance change switch 20E and reset it to the OFF state on the assumption that the resistance change switch 20E is in the ON state, the first wiring 21B and the second wiring connected to the resistance change switch 20E of the target that performs the reset operation VE and 0 are applied to 22B, respectively. A voltage half the VE is applied to the other first wirings 21A and 21C and the second wirings 22A and 22C. As a result, 0 is applied to the first electrode of the resistance change switch 20E selected as the target, and VE is applied to the second electrode. Further, in the resistance change switches 20B and 20H connected to the first wiring 21B and not selected, a voltage of −1 / 2VE is applied to the potential of the first electrode with respect to the second electrode. . In addition, in the resistance change switches 20C and 20I connected to the second wiring 22B and the non-selected resistance change switches 20C and 20I, a voltage of −1 / 2VE is applied to the potential of the first electrode with respect to the second electrode. . Further, in the non-selected resistance change switches 20A, 20G, 20C, and 20I other than the above, the voltage between the second electrode and the first electrode is 0V. If the set voltage to which the resistance change switch is programmed is between VE and 1 / 2VE, only the selected resistance change switch E of the target is reset.

  This programming method is called a half-selection method because a non-selected resistance change switch is applied with half of the voltage applied to the selected resistance change switch or 0V. Non-Patent Document 1 shows that a 32 × 32 crossbar switch can be programmed by this method.

  When the above-described two-terminal resistance change switch is used in an FPGA, the guaranteed years that are held in the ON and OFF states are required to be 10 years or more, which is the same as that of a nonvolatile memory such as a flash memory. As disturbance to each state, a voltage applied to the signal line is applied to the switch in the ON state, while a voltage applied to the signal line is applied to the switch in the OFF state. The problem that the state transitions for each disturb is called disturb failure.

  As one method of avoiding a disturb failure in the OFF state, as disclosed in Non-Patent Document 2, there is a method in which one set of two bipolar resistance change switches connected in series is used as one switch. (Figure 3). The second electrodes 321A and 322A (FIG. 3A) in the two resistance change switches 301A and 302A (FIG. 3A) or the first electrodes 311B and 312B (FIG. 3B) are electrically connected to form a common node 33A or 33B. A terminal-type switch. Hereinafter, this three-terminal type switch is referred to as a three-terminal resistance change switch. As shown in FIG. 3, the three-terminal resistance change switch includes three electrodes, that is, a first node 31A or 31B, a second node 32A or 32B, and a common node 33A or 33B. The common node is used for setting or resetting each resistance change switch, and is floating except during the set / reset. The ON and OFF states of the three-terminal resistance change switch are distinguished by the resistance state between the first node and the second node. In the OFF state, one of the two resistance change switches is in the OFF state, and in the ON state, both the two resistance change switches are in the ON state.

  At the time of OFF, when the voltage applied between the first node and the second node is VD, the applied voltage is distributed to two first resistance change switches and second resistance change switches connected in series. . Therefore, the voltage applied to each resistance change switch is smaller than that of one resistance change switch. Furthermore, since the first resistance change switch and the second resistance change switch are bipolar type, any one resistance change with respect to a positive or negative voltage applied between the first node and the second node. Since the voltage is applied in the direction in which the switch is reset, the OFF state is maintained. For example, in FIG. 3A, consider a case where the first resistance change switch and the second resistance change switch are in the OFF state, and a positive voltage VD is applied to the first node and 0 V is applied to the second node. In this case, in the second resistance change switch, the potential of the first electrode is lower than that of the second electrode, and the OFF state is maintained according to FIG. 1A.

Makoto Miyamura, et al., "Programmable cell array using rewritable solid-electrolyte switch integrated in 90nmCMOS", Solid-State Circuits Conference Digest of Technical Papers (ISSCC), 2011 IEEE International, pp. 228--229, February2011. DOI: 10.1109 / ISSCC.2011.5746296. Munehiro Tada, et al., "Highly reliable, complementary atom switch (CAS) with low programming voltage embedded in Cu BEOL for Nonvolatile Programmable Logic", IEEE International Electron Devices Meeting 2011 IEDM Technical Digest, pp. 30.2.1--30.2. DOI: 10.1109 / IEDM.2011.6131642.

  The resistance change switch is arranged between two different wirings, and switches between connecting or not connecting the two wirings by changing the low resistance state and the high resistance state. In order to efficiently change the connection between a large number of wirings, a crossbar switch is used. However, the crossbar switch composed of a two-terminal resistance change switch has a problem in disturb resistance when it is OFF.

  An object of the present invention is to provide a semiconductor device capable of high reliability.

According to the present invention,
A first resistance change switch including a resistance change layer having a first terminal and a second terminal, the resistance state of which changes when the applied voltage exceeds a reference value;
A resistance change layer having a third terminal and a fourth terminal, wherein the third terminal is connected to the second terminal to form a common node, and the resistance state changes when the applied voltage exceeds a reference value A second resistance change switch including:
A first wiring connected to the first terminal;
A second wiring connected to the fourth terminal and extending in a direction intersecting with the first wiring in a plan view;
A current control switch element having one end connected to the common node and the other end connected to the first terminal ;
I have a,
A semiconductor device is provided in which no element other than the second resistance change switch is connected between the common node and the fourth terminal .

According to the present invention,
A first resistance change switch including a resistance change layer having a first terminal and a second terminal, the resistance state of which changes when the applied voltage exceeds a reference value;
A resistance change layer having a third terminal and a fourth terminal, wherein the third terminal is connected to the second terminal to form a common node, and the resistance state changes when the applied voltage exceeds a reference value A second resistance change switch including:
A first wiring connected to the first terminal;
A second wiring connected to the fourth terminal and extending in a direction intersecting with the first wiring in a plan view;
A current control switch element having one end connected to the common node and the other end connected to the first terminal;
Have
A programming method using a semiconductor device in which no element other than the second resistance change switch is connected between the common node and the fourth terminal,
A voltage exceeding the reference value is applied between one of the first wiring and the second wiring and a terminal not connected to the common node of the current control switch element; By making the current control switch element conductive, the resistance state of either one of the first resistance change switch or the second resistance change switch is changed first,
Thereafter, a voltage exceeding the reference value is applied between the other wiring and the terminal not connected to the common node of the current control switch element, and the current control switch element is made conductive. Thus, a programming method is provided in which the resistance state of the other resistance change switch is changed to program the three-terminal resistance change switch composed of the first resistance change switch and the second resistance change switch. .

  According to the present invention, high reliability of a semiconductor device can be achieved.

  The above-described object and other objects, features, and advantages will become more apparent from the preferred embodiments described below and the accompanying drawings.

It is a figure which shows the characteristic and structure of a resistance change switch. It is a figure which shows the structure of the crossbar switch which has 3x3 resistance change switches. It is a figure which shows the structure of 3 terminal resistance change switch. 1 is a diagram showing a configuration of a semiconductor device according to a first embodiment of the present invention. It is a figure which shows the structure of the semiconductor device which concerns on the 2nd Embodiment of this invention. It is a figure for demonstrating operation | movement of the semiconductor device which concerns on the 2nd Embodiment of this invention. It is a figure for demonstrating operation | movement of the semiconductor device which concerns on the 2nd Embodiment of this invention. It is a figure which shows the structure of the semiconductor device which concerns on the 3rd Embodiment of this invention. It is a figure which shows the conductivity distribution of each switch at the time of making 32 switches in the diagonal component of the 32 * 32 crossbar switch using a 3 terminal resistance change switch make an ON state transition.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In all the drawings, the same reference numerals are given to the same components, and the description will be omitted as appropriate.

(First embodiment)
A first embodiment of the present invention will be described with reference to FIG.

  As illustrated in FIG. 4A, the semiconductor device according to the present embodiment includes a three-terminal resistance change switch 400 including a first resistance change switch 401 and a second resistance change switch 402, a first wiring 411, and a second wiring. 421 and a current control switch element 404. The first resistance change switch 401 is in contact with the resistance change layer whose resistance state changes according to the polarity of the applied voltage, and the first terminal 4011 and the second terminal that are in contact with the resistance change layer and are not electrically connected to each other. A terminal 4012 is provided. The second resistance change switch 402 is in contact with the resistance change layer whose resistance state changes according to the polarity of the applied voltage, and the third terminal 4021 and the fourth terminal that are in contact with the resistance change layer and are not electrically connected to each other. A terminal 4022 is included. The second terminal 4012 of the first resistance change switch 401 and the third terminal 4021 of the second resistance change switch 402 are connected to each other to form a common node 403. One terminal of the current control switch element 404 is connected to the common node 403. The current control switch element 404 controls the current flowing between the first terminal 4011 and the common node 403 and between the fourth terminal 4022 and the common node 403 by the voltage applied to the gate terminal. The first wiring 411 transmits a signal for controlling the voltage applied to the first terminal 4011. The second wiring 421 transmits a signal for controlling the voltage applied to the fourth terminal 4022. The second wiring 421 extends in a direction intersecting with the first wiring 411 in plan view. For example, the second wiring 421 extends so as to intersect with the first wiring 411 at a right angle in plan view, and is connected to the first wiring 411 via the three-terminal resistance change switch 400.

  Here, the first resistance change switch 401 and the second resistance change switch 402 are preferably of a bipolar type as shown in the present embodiment. Further, as shown in this embodiment, the common node 403 preferably connects terminals having the same polarity.

  Further, the semiconductor device according to the present embodiment may have a configuration as shown in FIG. 4B. In the semiconductor device shown in FIG. 4B, the first resistance change switch 401 is connected to the first wiring 411 via the first current control switch element 405. The first current control switch element 405 controls the current flowing between the first terminal 4011 and the common node 403 according to the voltage applied to the gate terminal. The second resistance change switch 402 is connected to the second wiring 421 via the second current control switch element 406. The second current control switch element 406 controls the current flowing between the fourth terminal 4022 and the common node 403 according to the voltage applied to the gate terminal. A third wiring 431 is connected to the common node 403. The third wiring 431 transmits a signal for controlling the voltage applied to the common node 403.

  With respect to the three-terminal resistance change switch 400 in FIG. 4A, a method of transitioning from the OFF state to the ON state will be described. The three-terminal resistance change switch being in the OFF state means that both the first resistance change switch 401 and the second resistance change switch 402 are in the OFF state. On the other hand, that the three-terminal resistance change switch is in the ON state means that both the first resistance change switch 401 and the second resistance change switch 402 are in the ON state.

  In the present specification, terms of high voltage and low voltage are used, but these terms do not specify a voltage value. In the resistance change switch, the resistance state changes depending on the voltage difference between the first electrode and the second electrode and the polarity of the voltage applied to each electrode. Therefore, the voltage value of the other electrode is higher than one electrode. Whether or not is a requirement. Of course, since the resistance state depends on whether the voltage value is larger than the set voltage or reset voltage, if it is necessary to specify such a voltage value, such as "high positive voltage lower than the set voltage" Describe. The low voltage includes a ground potential.

  In FIG. 4A, the second terminal 421 of the first resistance change switch 401 and the third terminal 4021 of the second resistance change switch 402 are connected to form a common node 403. Further, a current control switch element 404 is connected to the common node 403. The current control switch element 404 is used to control the ON resistance of each resistance change switch by limiting the current flowing through the first resistance change switch 401 and the second resistance change switch 402 during the set operation. Here, it is known that there is a proportional relationship between the reciprocal of the ON resistance after the set operation (that is, the ON conductivity) and the flowing current.

  First, a high voltage is applied to the second wiring 421. In the current control switch element 404, a low voltage is applied to the terminal that is not connected to the common node 403. Further, a voltage is applied to the gate terminal of the current control switch element 404 to turn on the current control switch element 404. Here, by controlling the potential difference of the second wiring 421 with respect to the common node 403 to be larger than the set voltage, the second resistance change switch 402 transitions to the ON state. At this time, by adjusting the gate voltage of the current control switch element 404, the current flowing through the second resistance change switch 402 during the set operation is controlled to about 1 milliA, for example. As a result, the ON resistance of the second resistance change switch 402 is about 500 ohms.

  Next, a high voltage is applied to the first wiring 411. In the current control switch element 404, a low voltage is applied to the terminal that is not connected to the common node 403. Further, a voltage is applied to the gate terminal of the current control switch element 404 to turn on the current control switch element 404. Here, by controlling the potential difference of the first wiring 411 with respect to the common node 403 to be larger than the set voltage, the first resistance change switch 401 transitions to the ON state. At this time, by adjusting the gate voltage of the current control switch element 404, the current flowing through the first resistance change switch 401 during the set operation is controlled to about 500 μA, for example. As a result, the ON resistance of the first resistance change switch 401 is about 1 kilohm.

  After the second resistance change switch 402 was changed to the ON state, the first resistance change switch 401 was changed to the ON state. Here, when the first resistance change switch 401 is changed to the ON state, the potential of the common node 403 changes, so that a slight current may flow through the second resistance change switch 402 in the ON state. It is conceivable that the resistance state of the second resistance change switch 402 changes from the ON state to the OFF state due to this slight current. Here, it is assumed that the resistance state of each resistance change switch transitions when a current of the same level as the current at the time of setting or resetting flows. Therefore, it is preferable that the ON resistance of the second resistance change switch 402 that is initially set to the ON state is lower than the first resistance change switch 401 that is set to the ON state later. Thereby, even if an electric current flows into the 2nd resistance change switch 402 made to change to ON state previously, the resistance state of the 2nd resistance change switch 402 does not change.

  As described above, the difference between the second resistance change switch 402 and the first resistance change switch 401 is about the same as that of the first resistance change switch 401, so that the three-terminal resistance change switch 400 can It was possible to make a transition to the ON state without fail. Note that the order of changing the resistance state is not limited to the above, and the first resistance change switch 401 may be switched to the ON state first.

  Next, a method of transitioning from the OFF state to the ON state with respect to the three-terminal resistance change switch 400 in FIG. 4B will be described.

  First, a high voltage is applied to the third wiring 431. Further, a low voltage is applied to the first wiring 411. Further, the first current control switch element 405 is turned on. Here, by controlling the potential difference of the first wiring 411 with respect to the common node 403 to be larger than the set voltage, the first resistance change switch 401 transitions to the ON state. At this time, by adjusting the gate voltage of the first current control switch element 405, the current flowing through the first resistance change switch 401 during the set operation is controlled to about 1 milliA, for example. As a result, the ON resistance of the first resistance change switch 401 is about 500 ohms.

  Next, a high voltage is applied to the third wiring 431. Further, a low voltage is applied to the second wiring 421. Further, the second current control switch element 406 is turned on. By controlling the potential difference of the second wiring 421 with respect to the common node 403 to be larger than the set voltage, the second resistance change switch 402 transitions to the ON state. At this time, by adjusting the gate voltage of the second current control switch element 406, the current flowing through the second resistance change switch 402 during the set operation is controlled to about 500 microA, for example. As a result, the ON resistance of the second resistance change switch 402 is about 1 kilohm.

  After the first resistance change switch 401 was changed to the ON state, the second resistance change switch 402 was changed to the ON state. Here, when the second resistance change switch 402 is changed to the ON state, the potential of the common node 403 changes, so that a slight current may flow through the first resistance change switch 401 in the ON state. It is conceivable that the resistance state of the first resistance change switch 401 changes from the ON state to the OFF state due to this slight current. Here, it is assumed that the resistance state of each resistance change switch transitions when a current of the same level as the current at the time of setting or resetting flows. Therefore, it is preferable that the ON resistance of the first resistance change switch 401 that is first set to the ON state is lower than the second resistance change switch 402 that is set to the ON state later. As a result, even if a current flows through the first resistance change switch 401 that has been previously changed to the ON state, the resistance state of the first resistance change switch 401 does not change.

  As described above, the difference between the first resistance change switch 401 and the second resistance change switch 402 is approximately the same as that of the first resistance change switch 401 and the second resistance change switch 402. It was possible to make a transition to the ON state without fail. Note that the order of changing the resistance state is not limited to the above, and the second resistance change switch 402 may be switched to the ON state first.

  As described above, according to the present embodiment, the resistance state of the three-terminal resistance change switch 400 is determined based on the resistance states of both the first resistance change switch 401 and the second resistance change switch 402. With this configuration, even if the resistance state of one resistance change switch changes due to a malfunction, the resistance state of the entire three-terminal resistance change switch 400 can be maintained, so that high reliability of the semiconductor device can be realized.

  In this embodiment, the common terminal 403 is formed by connecting the second terminal 4012 and the third terminal 4021 having the same polarity to each other. With this configuration, when a voltage is applied to the three-terminal resistance change switch 400, voltages in different directions are applied to the first resistance change switch 401 and the second resistance change switch 402. Thus, malfunction of the three-terminal resistance change switch 400 can be suppressed, and the reliability of the semiconductor device can be further improved.

  In the present embodiment, a configuration in which a voltage is applied between the first wiring 411 and the second wiring 422 and the common node 403 and the resistance states of two resistance change switches connected in series are individually changed. Take. With this configuration, the set voltage can be suppressed lower than in the case where the resistance states of two resistance change switches connected in series are collectively changed, and power saving of the semiconductor device can be realized.

  In the present embodiment, the current is controlled using the current control switch element 404 when the resistance state of each resistance change switch is changed. With this configuration, the ON resistance of the resistance change switch whose resistance state has been changed first can be controlled to be lower than the ON resistance of the resistance change switch whose resistance state has been changed later. As a result, when the resistance state of the subsequent resistance change switch is changed, the resistance state can be prevented from changing even if a current flows through the resistance change switch whose resistance state has been changed first. Therefore, the reliability of the semiconductor device can be further increased.

(Second Embodiment)
A second embodiment of the present invention will be described with reference to FIGS. FIG. 5 is a configuration diagram of a semiconductor device using the three-terminal resistance change switch 500.

  The semiconductor device includes a three-terminal resistance change switch 500, a first wiring 511, and a second wiring 521. In the present embodiment, at least one of the first wiring 511 and the second wiring 521 is provided in plural. In FIG. 5, the first wiring 511 and the second wiring 521 intersect each other, and the shortest distance between the first wiring 511 and the second wiring 521 or a position in the vicinity thereof (hereinafter referred to as an intersection). A three-terminal resistance change switch 50 is provided. One three-terminal resistance change switch 500 is provided at each intersection of the first wiring 511 and the second wiring 521.

  A terminal of the first resistance change switch 501 connected to the first wiring 511 is a first terminal, and a terminal of the second resistance change switch 502 connected to the second wiring 521 is a fourth terminal. Then, the second terminal of the first resistance change switch 501 and the third terminal of the second resistance change switch 502 are connected to form a common node 503.

  Thereby, the conduction state between the first wiring 511 and the second wiring 521 can be controlled in accordance with the ON / OFF state of the three-terminal resistance change switch 500.

  Furthermore, a current control switch element 505 is connected to the three-terminal resistance change switch 500 in order to change the ON / OFF state. A source terminal and a drain terminal of the current control switch element 505 are connected to the first terminal and the common node 503. Thereby, the current flowing through the first resistance change switch 501 can be bypassed. ON / OFF of the current control switch element 505 is controlled by a voltage applied to the control node 504. When the current control switch element 505 is ON, a current can flow through the current control switch element 505.

  Next, the operation of the semiconductor device will be described. At this time, it is assumed that the first resistance change switch 501 and the second resistance change switch 502 constituting the three-terminal resistance change switch 500 are both set to the OFF state. (1) When a positive high voltage lower than the set voltage is applied to the first wiring 511 and a low voltage is applied to the second wiring 521, (2) a low voltage is applied to the first wiring 511. The behavior of the three-terminal resistance change switch 500 with respect to noise or the like in each case where a positive high voltage lower than the set voltage is applied to the second wiring 521 will be described.

(1) When a positive high voltage lower than the set voltage is applied to the first wiring 511 and a low voltage is applied to the second wiring 521, the bias condition is that the voltage (VL1) of the first wiring 511 is the first This means that the voltage (VL1> VL2) is higher than the voltage (VL2) of the two wirings 521. In the second resistance change switch 502, the potential of the fourth terminal is lower than that of the third terminal, and a voltage in the direction of resetting from the ON state to the OFF state is applied as shown in FIG. 1A. On the other hand, in the first resistance change switch 501, the potential of the first terminal is higher than that of the second terminal, and as shown in FIG. 1, a voltage in the direction of setting from the OFF state to the ON state is applied. For this reason, the first resistance change switch 501 may malfunction and shift to the ON state.

  However, since the voltage in the direction to reset from the ON state to the OFF state is applied to the second resistance change switch 502, the second resistance change switch 502 maintains the OFF state. Accordingly, since at least the second resistance change switch 502 maintains the OFF state, the three-terminal resistance change switch 500 maintains the OFF state.

(2) When a low voltage is applied to the first wiring 511 and a positive high voltage lower than the set voltage is applied to the second wiring 521, the bias condition is that the voltage (VL1) of the first wiring 511 is the first It means that it is lower than the voltage (VL2) of the two wirings 521 (VL1 <VL2). In the first resistance change switch 501, the potential of the first terminal is lower than that of the second terminal, and a voltage in a direction to reset from the ON state to the OFF state is applied as shown in FIG. 1A. On the other hand, in the second resistance change switch 502, the potential of the third terminal is higher than that of the fourth terminal, and as shown in FIG. 1, a voltage in the direction of setting from the OFF state to the ON state is applied. Therefore, there is a possibility that the second resistance change switch 502 malfunctions and transitions to the ON state.

  However, since the voltage in the direction of resetting from the ON state to the OFF state is applied to the first resistance change switch 501, the first resistance change switch 501 maintains the OFF state. Therefore, since at least the first resistance change switch 501 maintains the OFF state, the three-terminal resistance change switch 500 maintains the OFF state.

  The three-terminal resistance change switch 500 is formed by connecting a first resistance change switch 501 and a second resistance change switch 502 in series. Therefore, the voltage applied to each of the first resistance change switch 501 and the second resistance change switch 502 is the same as the voltage applied to the three-terminal resistance change switch 500 of the first resistance change switch 501 and the second resistance change switch 502. The voltage is divided according to the resistance value. Therefore, even if the voltage difference applied to the three-terminal resistance change switch 500 is in the vicinity of the set voltage, the three-terminal resistance change switch 500 in the OFF state is unlikely to malfunction.

  In addition, the reliability of the semiconductor device is improved by using a three-terminal resistance change switch that is unlikely to cause a malfunction to be set to the ON state when turned OFF in the semiconductor device. When an FPGA is assumed as the semiconductor device, a highly reliable FPGA can be manufactured by using a three-terminal resistance change switch for the FPGA.

  Next, a method (programming method) for changing the ON / OFF state of the three-terminal resistance change switch will be described.

  First, a method for setting the three-terminal resistance change switch in the OFF state to the ON state will be described with reference to FIGS. 6A and 6B.

  A positive high voltage (VH) equal to or higher than the set voltage is applied to the second wiring 521, and a low voltage (VL) is applied to the first wiring 511. Then, a voltage is applied to the control node 504 so that the current control switch element 505 is turned on.

  When the current control switch element 505 is turned on, the voltage of the common node 503 becomes a low voltage substantially the same as the voltage of the first wiring 511. Therefore, a voltage equal to or higher than the set voltage is applied to the second resistance change switch 502, and the second resistance change switch 502 transitions to the ON state (FIG. 6A). At this time, as in the first embodiment, the second resistance change switch 502 is turned on by setting the current flowing through the current control switch element 505 to about 1 milliA by the gate voltage of the current control switch element 505. It is preferable to control the resistance to around 500 ohms.

  By the operation of FIG. 6A, the second resistance change switch 502 is turned on. On the other hand, since the current control switch element 505 is ON, the first resistance change switch 501 to which a high voltage is not applied holds the OFF state.

  Next, a low voltage is applied to the second wiring 521 and a high voltage is applied to the first wiring 511. Since the second resistance change switch 502 is in the ON state, the common node 503 has substantially the same low voltage as the second wiring 521. When the current control switch element 505 is turned off, a high voltage is applied to the first terminal side of the first resistance change switch 501. Accordingly, a voltage equal to or higher than the set voltage is applied to the first resistance change switch 501, and the first resistance change switch 501 transitions to the ON state (FIG. 6B). At this time, as in the first embodiment, for example, another current control switch element is connected between the fourth terminal and the second wiring 521 to control the flowing current to about 500 microA. By doing so, it is preferable to control the ON resistance of the first resistance change switch 501 to about 1 kilohm.

  Since the current is applied to the second resistance change switch 502 when the first resistance change switch 501 is changed to the ON state, the resistance state of the second resistance change switch 502 may be changed from the ON state to the OFF state. There is. Here, when a current comparable to the current at the time of setting (that is, 1 milliA) flows through the second resistance change switch 502, the resistance state of the second resistance change switch 502 transitions. For this reason, the current that flows when the first resistance change switch 501 is set needs to be smaller than 1 milliA. Here, since it is about 500 micro A, it can prevent that the resistance state of the 2nd resistance change switch 502 changes.

  As described above, since both the first resistance change switch 501 and the second resistance change switch 502 are set to the ON state, the three-terminal resistance change switch is turned on. Accordingly, signal transmission is possible between the first wiring 511 and the second wiring 521. Note that the order of changing the resistance state of each resistance change switch is not limited to the above, and the current control switch element 505 is installed between the fourth terminal and the common node 503, and the first resistance change switch 501 is first. You may make a transition to the ON state.

  On the other hand, a method of resetting the three-terminal resistance change switch 500 in the ON state to the OFF state will be described with reference to FIGS. 7A and 7B. In the three-terminal resistance change switch 500 in the ON state, it is desirable that the ON resistance of the first resistance change switch 501 is higher than the ON resistance of the second resistance change switch 502. For example, it is desirable to set it to about 1 kilohm and 500 ohm, respectively.

  First, the first resistance change switch 501 having a high ON resistance is reset to the OFF state.

  A high positive voltage higher than the set voltage is applied to the second wiring 521, and a low voltage is applied to the first wiring 511 (FIG. 7A). Then, a voltage is applied to the control node 504 so that the current control switch element 505 is turned off. Here, current flows through the second resistance change switch 502 and the first resistance change switch 501 in the ON state. Since the forward bias voltage is applied to the second resistance change switch 502 in the ON state, the resistance state does not change. On the other hand, since a reverse bias voltage is applied to the first resistance change switch 501, the first resistance change switch 501 is reset to the OFF state.

  By the operation of FIG. 7A, the first resistance change switch 501 was turned off. On the other hand, the second resistance change switch 502 to which the forward voltage is applied maintains the ON state.

  Next, a low voltage is applied to the second wiring 521 and a high voltage is applied to the first wiring 511. Further, a voltage is applied to the control node 504 so that the current control switch element 505 is turned on (FIG. 7B). Since the current control switch element 505 is in the ON state, the potential of the common node 503 becomes the same high voltage as that of the first wiring 511, and a reverse bias voltage is applied to the second resistance change switch 502. Accordingly, the second resistance change switch 502 is reset to the OFF state.

  As described above, since both the first resistance change switch 501 and the second resistance change switch 501 are reset to the OFF state, the three-terminal resistance change switch 500 is turned OFF. Accordingly, signal transmission is blocked between the first wiring 511 and the second wiring 521.

  As described above, also in this embodiment, the same effect as that of the first embodiment can be obtained. According to the present embodiment, since the signal line for controlling the voltage of the common node 503 is not necessary, the configuration of the semiconductor device can be simplified.

(Third embodiment)
In the present embodiment, a case where the state of one three-terminal resistance change switch in the crossbar switch is changed will be described. FIG. 8 illustrates a semiconductor device having nine three-terminal resistance change switches 80A to 80I, three first wires 81A, 81B, 81C, and second wires 82A, 82B, 82C.

  A case where the three-terminal resistance change switch 80E in the OFF state is transitioned to the ON state will be described. At this time, the first wiring 81B and the second wiring 82B are used. First, a low voltage is applied to the first wiring 81B. Further, a high voltage is applied to the second wiring 82B, and control is performed so that the potential difference between the first wiring 81B and the second wiring 82B is equal to or higher than the set voltage. When the current control switch element 804 is turned on, a voltage equal to or higher than the set voltage is applied to the second resistance change switch 803 of the three-terminal resistance change switch 80E, and the second resistance change of the three-terminal resistance change switch 80E. The switch transits to the ON state (the same operation as in FIG. 6A). Next, a high voltage is applied to the first wiring 81B. Further, a low voltage is applied to the second wiring 82B, and the potential difference between the first wiring 81B and the second wiring 82B is controlled to be equal to or higher than the set voltage. Then, by setting the current control switch element 804 to the OFF state, a voltage equal to or higher than the set voltage is applied to the first resistance change switch 802 of the three-terminal resistance change switch 80E, and the first resistance change of the three-terminal resistance change switch 80E. The switch 801 transitions to the ON state (the same operation as in FIG. 6B). With the above operation, the three-terminal resistance change switch 80E changes to the ON state.

  During the above operation, the high voltage applied to the first wiring 81B and the second wiring 82B are applied to the first wiring 81A, 81C and the second wiring 82A, 82C that are not connected to the selected three-terminal resistance switch 80E. It is preferable to apply a voltage intermediate to the applied low voltage. By applying the intermediate voltage, the three-terminal resistance change switches 80B, 80H, 80D, and 80F are in a half-selected state, and do not exceed the set voltage for transitioning to the ON state.

  Next, a case where the three-terminal resistance change switch 80E in the ON state is reset to the OFF state will be described. At this time, the first wiring 81B and the second wiring 82B are used. First, a low voltage is applied to the first wiring 81B. Further, a high voltage is applied to the second wiring 82B, and control is performed so that the potential difference between the first wiring 81B and the second wiring 82B is equal to or higher than the reset voltage. Then, by setting the current control switch element 804 to the OFF state, a voltage equal to or higher than the reset voltage is applied to the second resistance change switch 802 of the three-terminal resistance change switch 80E, and the second resistance change of the three-terminal resistance change switch 80E. The switch 802 is reset to the OFF state (the same operation as in FIG. 7A). Further, a high potential is applied to the first wiring 81B. Further, a low voltage is applied to the second wiring 82B, and control is performed so that the potential difference between the first wiring 81B and the second wiring 82B is equal to or higher than the reset voltage. Then, by turning on the current control switch element 804, a voltage equal to or higher than the reset voltage is applied to the first resistance change switch 801 of the three-terminal resistance change switch 80E, and the first resistance change of the three-terminal resistance change switch 80E. The switch 801 is reset to the OFF state (the same operation as in FIG. 7B). With the above operation, the three-terminal resistance change switch 80E is reset to the OFF state.

  During the above operation, the first wiring 81A, 81C and the second wiring 82A, 82C not connected to the selected three-terminal resistance switch are applied to the high voltage applied to the first wiring 81B and the second wiring 82B. It is preferable to apply a voltage intermediate to the low voltage. By applying the intermediate voltage, the three-terminal resistance change switches 80B, 80H, 80D, and 80F are in a half-selected state, and do not exceed the reset voltage for transitioning to OFF.

  FIG. 9 shows an example using the semiconductor device in this embodiment. FIG. 9 shows the conductivity distribution of each switch when 32 switches in the diagonal component of a 32 × 32 crossbar switch using a three-terminal resistance change switch are shifted to the ON state. The shade of the color represents the conductivity, the dark color represents that the conductivity is small, and the light color represents that the conductivity is large. FIG. 9 shows that only a desired three-terminal resistance change switch can be correctly programmed in the 32 × 32 crossbar switch.

  As described above, also in this embodiment, the same effects as those of the first embodiment and the second embodiment can be obtained.

  As mentioned above, although embodiment of this invention was described with reference to drawings, these are the illustrations of this invention, Various structures other than the above are also employable.

In addition, according to embodiment mentioned above, the following invention is disclosed.
(Appendix 1)
A first resistance change switch including a resistance change layer having a first terminal and a second terminal, the resistance state of which changes when the applied voltage exceeds a reference value;
A resistance change layer having a third terminal and a fourth terminal, wherein the third terminal is connected to the second terminal to form a common node, and the resistance state changes when the applied voltage exceeds a reference value A second resistance change switch including:
A first wiring connected to the first terminal;
A second wiring connected to the fourth terminal and extending in a direction intersecting with the first wiring in a plan view;
A current controlling switch element having one end connected to the common node;
A semiconductor device.
(Appendix 2)
A first resistance change switch including a resistance change layer having a first terminal and a second terminal, the resistance state of which changes when the applied voltage exceeds a reference value;
A resistance change layer having a third terminal and a fourth terminal, wherein the third terminal is connected to the second terminal to form a common node, and the resistance state changes when the applied voltage exceeds a reference value A second resistance change switch including:
A first wiring connected to the first terminal via a first current control switch element;
A second wiring connected to the fourth terminal via a second current control switch element and extending in a direction crossing the first wiring in plan view;
A third wiring connected to the common node;
A semiconductor device.
(Appendix 3)
In the semiconductor device according to attachment 1,
The current control switch element is:
A semiconductor device in which the other end is connected to one of the first terminal and the fourth terminal.
(Appendix 4)
In the semiconductor device according to any one of appendices 1 to 3,
The first resistance change switch and the second resistance change switch are bipolar semiconductor devices.
(Appendix 5)
In the semiconductor device according to attachment 4,
The second terminal and the third terminal are semiconductor devices having the same polarity.
(Appendix 6)
In the semiconductor device according to any one of appendices 1 to 5,
At least one of the first wiring and the second wiring is provided in plural,
A semiconductor device in which a three-terminal resistance switch composed of the first resistance change switch and the second resistance change switch is connected to each intersection of the first wiring and the second wiring.
(Appendix 7)
A first resistance change switch including a resistance change layer having a first terminal and a second terminal, the resistance state of which changes when the applied voltage exceeds a reference value;
A resistance change layer having a third terminal and a fourth terminal, wherein the third terminal is connected to the second terminal to form a common node, and the resistance state changes when the applied voltage exceeds a reference value A second resistance change switch including:
A first wiring connected to the first terminal;
A second wiring connected to the fourth terminal and extending in a direction intersecting with the first wiring in a plan view;
A current control switch element connected to the common node;
A semiconductor device having
A voltage exceeding the reference value is applied between one of the first wiring and the second wiring and a terminal not connected to the common node of the current control switch element; By making the current control switch element conductive, the resistance state of either one of the first resistance change switch or the second resistance change switch is changed first,
Thereafter, a voltage exceeding the reference value is applied between the other wiring and the terminal not connected to the common node of the current control switch element, and the current control switch element is made conductive. Thus, a programming method of changing the resistance state of the other resistance change switch and programming the three-terminal resistance change switch constituted by the first resistance change switch or the second resistance change switch.
(Appendix 8)
A first resistance change switch including a resistance change layer having a first terminal and a second terminal, the resistance state of which changes when the applied voltage exceeds a reference value;
A resistance change layer having a third terminal and a fourth terminal, wherein the third terminal is connected to the second terminal to form a common node, and the resistance state changes when the applied voltage exceeds a reference value A second resistance change switch including:
A first wiring connected to the first terminal via a first current control switch element;
A second wiring connected to the fourth terminal via a second current control switch element and extending in a direction crossing the first wiring in plan view;
A third wiring connected to the common node;
A semiconductor device having
A voltage exceeding the reference value is applied between one of the first wiring and the second wiring and the third wiring, and the first current control switch element and the second current control Of the switch elements, one of the first resistance change switch and the second resistance change switch is turned on by turning on the current control switch element connected to the wiring to which the voltage is applied. Change the resistance state of the resistance change switch first,
Thereafter, a voltage exceeding the reference value is applied between the other wiring and the third wiring, and the resistance state of the other resistance change switch is changed by making the other current control switch element conductive. A programming method for programming a three-terminal resistance change switch constituted by the first resistance change switch or the second resistance change switch.
(Appendix 9)
In the programming method described in appendix 7,
The current control switch element is:
The other end is connected to one of the first terminal and the fourth terminal,
By turning on the current control switch element, either one of the first resistance change switch and the second resistance change switch is bypassed, and the first resistance change switch and the second resistance change switch A programming method for performing individual programs.
(Appendix 10)
In the programming method according to any one of appendices 7 to 9,
The first resistance change switch and the second resistance change switch are a bipolar type programming method.
(Appendix 11)
In the programming method according to attachment 10,
A programming method in which the second terminal and the third terminal have the same polarity.
(Appendix 12)
In the programming method according to any one of appendices 7 to 11,
At least one of the first wiring and the second wiring is provided in plural,
A three-terminal resistance switch composed of the first resistance change switch and the second resistance change switch is connected to each intersection of the first wiring and the second wiring,
A programming method for selecting a unique three-terminal resistance change switch to be programmed based on a signal from the first wiring and a signal from the second wiring.

  This application claims the priority on the basis of Japanese application Japanese Patent Application No. 2012-139054 for which it applied on June 20, 2012, and takes in those the indications of all here.

Claims (7)

A first resistance change switch including a resistance change layer having a first terminal and a second terminal, the resistance state of which changes when the applied voltage exceeds a reference value;
A resistance change layer having a third terminal and a fourth terminal, wherein the third terminal is connected to the second terminal to form a common node, and the resistance state changes when the applied voltage exceeds a reference value A second resistance change switch including:
A first wiring connected to the first terminal;
A second wiring connected to the fourth terminal and extending in a direction intersecting with the first wiring in a plan view;
A current control switch element having one end connected to the common node and the other end connected to the first terminal ;
I have a,
A semiconductor device in which no element other than the second resistance change switch is connected between the common node and the fourth terminal . The semiconductor device according to claim 1 ,
The first resistance change switch and the second resistance change switch are bipolar semiconductor devices. The semiconductor device according to claim 2 ,
The second terminal and the third terminal are semiconductor devices having the same polarity. The semiconductor device according to any one of claims 1 to 3 ,
At least one of the first wiring and the second wiring is provided in plural,
A semiconductor device in which a three-terminal resistance switch composed of the first resistance change switch and the second resistance change switch is connected to each intersection of the first wiring and the second wiring. The semiconductor device according to any one of claims 1 to 4,
A semiconductor device, wherein an ON resistance value of the first resistance change switch and an ON resistance value of the second resistance change switch are different from each other. A programming method using the semiconductor device according to claim 1,
A voltage exceeding the reference value is applied between one of the first wiring and the second wiring and a terminal not connected to the common node of the current control switch element; By making the current control switch element conductive, the resistance state of either one of the first resistance change switch or the second resistance change switch is changed first,
Thereafter, a voltage exceeding the reference value is applied between the other wiring and the terminal not connected to the common node of the current control switch element, and the current control switch element is made conductive. Thus, a programming method of changing the resistance state of the other resistance change switch and programming the three-terminal resistance change switch including the first resistance change switch and the second resistance change switch. The programming method according to claim 6, wherein
A programming method in which the on-resistance value of the first resistance change switch and the second resistance change switch that is set to the on state first is smaller than the other on-resistance value.

Images