JP2005317787A - Switching element, and array type functional element using same switching element - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an array type functional element which can be more fined than conventional ones, by using a switching element as the selecting portion of the functional element. <P>SOLUTION: As each switching element, there is used each multilayer structure wherein a strongly correlated electronic material is sandwiched by two or more electrodes. Then, each multilayer structure is so connected in series with each variable-resistance type element of a functional element as to realize a random access memory by arranging the resultant intermediates in the form of an array. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to non-volatile memories used for information communication terminals, sensors for detecting changes in light, heat, stress, magnetism, etc., or electronic devices in general that require a random access function such as an image display. .

  Array-type functional elements are currently used in all functional fields and are important basic electronic components that support the current information society such as memories, sensors, and displays.

  In the conventional array type functional element, in order to enable random selection of each element, a two-terminal element having non-linear electrical characteristics and a three-terminal switching element such as a MOS transistor are attached on a one-to-one basis. Many devices use MOS transistors because of the requirements of high-speed and uniform characteristics.

FIG. 1 shows an example in which a MOS transistor is attached to a function unit as a switching element for selection. An array type function having random accessibility by arranging these in a two-dimensional or three-dimensional manner. It can be configured as an element.
Physics Letters A.Vol.54 Issue3, "Tunneling between containing films", Sep.8th 1975, pp.225-226

  In recent years, there has been an increasing demand for miniaturization and cost reduction of these functional elements due to the influence of the spread of portable information terminals and the deflation of electronic component prices. At present, non-volatile memories, which are expected to make great progress in portable information terminals, have been proposed that respond to miniaturization and cost reduction using new technologies as functional units.

  However, as shown in FIG. 2, in the element using the MOS structure, even if the functional part is sufficiently miniaturized, the fine limit of the element is limited by the MOS transistor in which the fine limit is pointed out. There is a problem that sufficient characteristics of the functional part cannot be exhibited.

  In particular, when using a magnetoresistive (MR) element, other variable resistance element, or a response sensor related to light or heat, which has attracted much attention in recent years for the functional part, it should be possible to make the device sufficiently finer than the MOS part. It has been known.

  The miniaturization limit of elements such as a MOS transistor and a PN diode forming the selection element occurs because a certain amount of carrier injection region is required for the semiconductor.

  Currently, semiconductor processes related to the manufacture of MOS transistors and the like continue to decrease year by year and progress to a level below F = 0.1 microns, but the prospect of further miniaturization is unclear.

In addition, the element region in the case where a switching element composed of a MOS transistor is configured is basically 8F ^{2 to} 12F ² (F is a design rule size for miniaturization, and the relative occupied area is determined by using F. However, it does not reach 4F ² which is the ideal when miniaturization is advanced.

  Some proposed functional units can be multi-layered or multi-valued, and a new proposal for a selecting unit is required for future breakthrough of array-type functional elements.

  The present invention addresses this conventional problem by using a specific strongly correlated electron material without using a conventional semiconductor, so that the bulk material itself has non-linear current-voltage characteristics, and current or voltage, or its It has been found out that the nonlinear conduction characteristic itself can be changed by applying both.

  By using the present invention as a switching element as a selection unit, it is possible to realize an array-type functional element that can be miniaturized as compared with the prior art, and by providing a preferred device configuration example, the present invention can be applied in a wide variety of ways. The purpose is to show.

  In order to solve such a problem, a first invention is a switching element in which a layer made of a strongly correlated electron material is sandwiched between at least two electrodes, and a non-linear conduction characteristic appears in the current-voltage characteristic. The fact that the nonlinear conduction characteristics change by applying a current or voltage in between is utilized.

  In addition, the second invention comprises a multi-stage structure formed by laminating at least two or more multi-layer structures each including a layer made of a strongly correlated electron material with at least two electrodes. In the switching element in which the nonlinear conduction characteristic appears in the current-voltage characteristic, the fact that the nonlinear conduction characteristic is changed by applying a current or voltage between the electrodes at both ends sandwiching the multistage structure is utilized.

  In the case of using this multistage structure, it is preferable that the nonlinear conduction characteristic is changed by applying a current or a voltage to at least one multilayer structure of the multistage structure.

  Further, it is preferable that the non-linear conduction characteristic change be kept non-volatile even when no current or voltage is applied.

  Note that application of current or voltage is preferably performed by pulses.

The strongly correlated electron material having such a function is represented by the formula: A _x B _y O _z (where A is an alkali metal (Group 1A element), alkaline earth metal (Group 2A element), rare earth element, scandium. , Or yttrium, lead, bismuth, thallium, B is a transition metal of any of group 1B elements to group 8 elements, or group 3A elements, group 4A elements, group 5A elements, group 6A elements, group 7A elements, O is preferably a substance having a crystal structure represented by oxygen).

  In the present invention, it is preferable that the switching elements to be configured and various functional elements are connected in series and arranged in an array, and the switching elements are used as elements for selecting random access to the functional unit. This is preferable for improving the random accessibility of the array type functional element.

  In addition, it is preferable to configure an array type functional element by utilizing a change in the non-volatile nonlinear conduction characteristic of the switching element.

  According to the present invention, at least two electrodes are sandwiched between a multilayer structure formed by sandwiching a layer made of a strongly correlated electron material, or a multi-stage structure formed by laminating at least two or more of these layers. The nonlinear conduction characteristic obtained between the electrodes at both ends can be changed by applying a current or voltage. Moreover, since they are held in a non-volatile manner, they can be used for various functional devices as switching elements using them. In particular, since the nonlinear conduction characteristic can be used as a switching element for selecting random access, an array type functional element arranged in an array can be realized.

BEST MODE FOR CARRYING OUT THE INVENTION

  A switching element of the present invention and an array type functional element using the same will be described with reference to the drawings.

(Embodiment 1)
In order to enable random selection of each element, a conventional array-type functional element has a one-to-one attachment of a two-terminal element having a non-linear electric conduction characteristic and a three-terminal switch element such as a MOS transistor.

  As shown in FIG. 3, in the case of an array-type element in which functional parts are connected at a cross point, avoiding wraparound from other than the functional part that is desired to be read has a significant meaning in functionality. Many devices use MOS transistors because of the requirements of high-speed and uniform characteristics.

  FIG. 1 shows an example in which a MOS transistor is selected as a selection element and is attached to a functional part. By arranging these in a two-dimensional or three-dimensional form, an array type functional element having random accessibility is shown. Can be configured.

  However, as shown in FIG. 2, in the use of a switching element using a MOS structure, even if the functional part is sufficiently miniaturized, the fine limit of the element is limited by the MOS transistor in which the fine limit is pointed out. Therefore, sufficient characteristics of the functional unit cannot be exhibited.

  In other words, since the width of the MOS transistor is larger than the width of the functional part, no matter how narrow the width of the functional part is, the width of the functional part is limited as long as it is difficult to reduce the width. Efforts to narrow down do not make sense.

  In FIG. 4 showing an example in which the elements shown in FIG. 2 are arranged in an array, it can be seen that the main miniaturization limit is determined by the switching element having the MOS structure as the selection unit. This indicates that even if an element having a new function is developed in the functional part, it is not possible to break through the miniaturization wall in the current semiconductor process technology.

  Therefore, in the present invention, a new switching element is provided as a switching element for selection without using a conventional Si-based or compound semiconductor. As shown in FIG. 5, the switching element of the present invention has two electrodes (for convenience of explanation, one of the two electrodes is simply referred to as “electrode” and the other is sometimes referred to as “COLUMN electrode”). It is a two-terminal switching element configured with a strongly correlated electron material interposed therebetween.

  By using the switching elements of the present invention in an array, it is possible to realize the miniaturization equivalent to the functional unit as shown in FIG. 6 as compared with the case of the MOS shown in FIG. This is because the non-linearity comes from the essential bulk characteristics of the strongly correlated electron material sandwiched between the electrodes, so that there is no limit on the size of the molecule to the molecular level. That is, since the switching element of the present invention can be made smaller than the MOS transistor, the width of the switching element of the present invention can be reduced as the functional part width is reduced. Is possible.

  A feature of the switching element of the present invention is that it is non-linear and asymmetric in the current-voltage characteristics, thereby ensuring element selectivity when connected to a functional unit, and thus random accessibility when an array type configuration is used. There is something you can do.

  As the configuration of the array type functional element, as shown in FIG. 6 (a) or FIG. 6 (b), a switching element and a functional unit connected in series are configured in an array, and the device as shown in FIG. 6 (c). Is realized. Also, it is preferable to use a multi-stage configuration as shown in FIG. 6 and use it in a configuration as shown in FIGS. 14A and 14B, or an array type as shown in FIG.

The strongly correlated electron materials be provided in the present invention is preferably an oxide material, the general formula: A _x is preferably B _y O _z represented by substances.

  In the formula, A is an alkali metal (Group 1A element), an alkaline earth metal (Group 2A element), a rare earth element, scandium, yttrium, lead, bismuth, or thallium. B is a transition metal of any one of group 1B elements to group 8 elements, or any of 3A group elements, 4A group elements, 5A group elements, 6A group elements, and 7A group elements. O represents oxygen. The substance has a crystal structure.

In general, a unit cell of a corresponding lattice structure includes a cell center molecule surrounded by a plurality of oxygen molecules each having a center molecule. Either type of central position can basically be occupied by compounds A and B.
The indices x, y, and z shown here have requirements that must be satisfied by the combination, and define several preferable categories.

When n = 0, 1, 2, 3, the combination of indices defined by x = n + 2, y = n + 1, z = 3n + 4 is, for example, (La, Ba) ₂ CuO ₄ or (La, Sr) ₂ CoO ₄ (xyz index (214)) or Sr ₃ Ru ₂ O ₇ or (La, Sr) ₃ Mn ₂ O ₇ (xyz index (327)), so-called Ruddlesden-Popper phase.

When n = 0, in the combination of defined indices, for example, the positions of A and B are reversed, that is, the B cation that was originally index y is at the position of index x, and the A cation is index y. An independent category with a spinel structure, such as Mg ₂ TiO ₄ (214), Cr ₂ MgO ₄ , Al ₂ MgO ₄ , or Fe ₂ CoO ₄ , Fe ₂ FeO ₄ (Fe ₃ O ₄ ) And the like, and substances whose x and y indices are only B cations (B ₂ BO ₄ ).

  When n = 1, 2, 3, 4, the combination of indices x, y, z defined by x = n + 1, y = n + 1, z = 3n + 5 is partially independent to give a substance with oxygen intercalation Also shows the category.

The combination of x = 1, y = 1, z = 1 and the index x, y, z defined by either index x or y is BeO, MgO, BaO, CaO,..., NiO, MnO Typical materials such as CoO, CuO, and ZnO are shown. Or when n = 1 or 2, the combination of x = n, y = n, z = n + 1 and the index x, y, z defined by either n = 1 and the index x or y is 0 is TiO ₂ , VO ₂ , MnO ₂ , GeO ₂ , CeO ₂ , PrO ₂ , SnO _{2 and} the like. When n = 2, substances such as Al ₂ O ₃ , V ₂ O ₃ , Ce ₂ O ₃ , Nd ₂ O ₃ , Ti ₂ O ₃ , Sc ₂ O ₃ , and La ₂ O ₃ are shown. Alternatively, a substance containing Li, Na, K, Cs, Ca, Sr, Ba, etc. in part A such as AM _x CoO ₂ (AE represents an alkali metal element or an alkaline earth metal element) is also included.

Or, when n = 2, x = n, y = n, z = 2n + 1, and the index x or y is 0, typical materials such as Nb ₂ O ₅ , V ₂ O ₅ , Ta ₂ O _5, etc. Indicates.

The combination of the indices x, y, z defined by n = 1, 2, or 3 and x = n, y = n, z = 3n is SrTiO ₃ , BaTiO ₃ , KNbO ₃ , LiNbO ₃ , n = 1 Independent substance categories, so-called perovskites, such as SrMnO ₃ , (Pr, Ca) MnO ₃ , BMnO ₃ and the like are shown. When n = 2, a substance having an index (226) such as Sr ₂ FeMoO ₆ can be obtained.

The combination of indices x, y, z defined by n = 1 or 2 and x = n + 1, y = n, z = 4n + 1 represents an independent substance category. When n = 1, a material having an index (215) such as Al ₂ TiO ₅ , Y ₂ MoO ₅ or the like is obtained. When n = 2, SrBi ₂ Ta ₂ O ₉ or the like is obtained.

  Substance categories such as these are modified by changing the composition of the substances such that at least one of the compounds Ax or By consists of one or several elements of the corresponding group of A, B be able to.

  Also, a superlattice formed by a combination of structural unit cells or small unit cells in which n is different and each of the structural unit cells and small unit cells is part of the corresponding homologous sequence obtained by oxygen intercalation is given. Can be changed. Furthermore, each n is different, and it is changed by giving a superlattice formed by a combination of structural unit cells and small unit cells of the Ruddlesden-Popper type structure where structural unit cells and small unit cells are part of the corresponding homologous series. Is added. In such a lattice modification, a lattice structure is formed in which one or more transition metal oxygen octahedral layers are separated by one or more block layers containing Compound A and oxygen.

  The switching element using the strongly correlated electron material of the present invention can further control and change non-linear and asymmetric characteristics appearing in the current-voltage characteristics by applying a current or voltage. 7 and 8 show this state. In particular, it is easily manifested by applying a voltage pulse or a current pulse between two electrodes.

  The characteristic shown in FIG. 7A shows that the characteristic of the switching element having the configuration as shown in FIG. 5 is shifted to the low voltage side by the application of the voltage pulse. In FIG. 7A, the voltage values of the pulses to be applied are (ai) initial state, (a-ii) 0.9V, (a-iii) 1.5V, and (a-iv) 10V, respectively. In either case, the pulse is a single pulse, and the width is 500 ns.

  Conversely, by reversing the polarity of the voltage pulse, it can be returned to the high voltage side. In FIG. 7 (b), the voltage values of the applied pulses are (bi) initial state (corresponding to (a-iv) state in FIG. 7 (a)), (b-ii) -1.2V, ( b-iii) −1.8V and (b-iv) −12V, both of which show the change in pulse width = single pulse at 500 ns.

  This operation can be obtained by adjusting the pulse width even at the same voltage. These behaviors are kept non-volatile even in the absence of pulse application.

  FIG. 8A shows a state in which the pulse width at an applied voltage of 15 V is (c-i) initial state, (c-ii) 30 ns, (c-iii) 100 ns, and (c-iv) 1 ms, respectively. ing. In FIG. 8B, the pulse widths in the case of the applied voltage of −15V are (di) initial state (state (c-iv) in FIG. 8A), (d-ii) 30 ns, (D-iii) 100 ns, (d-iv) 1 ms.

  In FIG. 7 and FIG. 8, the change in the response of the conduction characteristic with excellent controllability to the application of the constant voltage pulse can also be obtained by applying the constant current pulse.

  The non-linear and asymmetric conduction characteristic in the current-voltage characteristic found by the present invention can be controlled by applying a current or voltage pulse, and the state is maintained in a nonvolatile manner. This is a function unique to the strongly correlated electron material of the present invention, and is a unique one not seen in the past.

  In addition, it was confirmed that the oxide, which is a strongly correlated electron material, is stable in characteristics even at high temperatures. 7 and 8 show the characteristics at room temperature, but the characteristics of FIGS. 7 and 8 were stably reproduced even after a temperature change from 1K to 1123K. Further, in the preferable temperature range of 4K to 800K, the characteristics shown in FIGS. 7 and 8 could be reproduced with respect to the thermal history within at least 20 hours in any temperature range. That is, since it is resistant to thermal history, it does not impose restrictions on the subsequent thermal process, so it is a preferable function as an element disposed in a lower layer, such as a selection element, and is considered useful as a base process.

  Further, when the switching element having the non-linear and asymmetric electric conduction characteristic is used as a selection element for random access, the operation as shown in FIG. 11 to FIG. 13 is performed, and a desired operation can be performed.

  A selective read operation for such an array type functional element will be described.

  The operation shown in FIG. 11 and FIG. 12 can be realized without the non-volatile characteristics of the switching element of the present invention. However, when the switching element of the present invention is used, it can be used according to the specification purpose. It is preferable. First, FIG. 11 uses the rectifying property of the switching element, and selective reading of the element at the location (M, N) is performed between the RAW and COLUMN electrodes by controlling two voltage values.

  FIG. 11A shows a schematic diagram of an array-type functional element, where RAW electrodes are represented by indexes of M-1, M, and M + 1, and COLUMN electrodes are represented by indexes of N-1, N, and N + 1. Represents. In addition, the switching element in a figure is represented by the symbol of FIG.11 (d).

  FIG. 11B shows a serial connection of the switching element of the present invention and a functional element (here, one that operates as a memory by a nonvolatile resistance change. A write operation to such a functional unit will be described later). The state of the current-voltage characteristic at the time of being expressed is represented.

  In the operation under the binary voltage control, as shown in FIG. 11C, in the selection operation, the RAW electrode M is set to the high level for the portion (M, N) which is the selection place, and COLUMN Change electrode N to Low-level and apply + Vs bias. This operation is preferably performed using switching elements arranged around the array. Since the switching element at this time is located around the array, a conventional MOS transistor or the like may be used.

  At this time, at a non-selected place, a voltage is applied at least in a range where the switching element is below the breakdown voltage or breakdown voltage and below the voltage that can ensure rectification (the bias point in FIG. 11B is the O point). Or it is almost zero (the bias point in FIG. 11 (b) is V or U point), and the influence of the wraparound from other than the selected place (M, N) can be eliminated and the selected place can be read out. (The bias point is the point S in FIG. 11B).

  By comparing the current value at the voltage Vs at the point S with the current value flowing through the reference resistor, it is determined whether the information of the memory unit, which is a functional unit, is <1> or <0>. Can do. The bias point Vs may be at or above the voltage Vth at which nonlinearity rises.

  In FIG. 12, unlike the operation in FIG. 11, selective reading is realized by performing ternary voltage control instead of binary. In this method, the operating voltage range can be set lower than in the case of binary control, and more efficient reading is possible.

  FIG. 12A shows a schematic diagram of an array-type functional element, where the RAW electrode is represented by an index of M-1, M, M + 1, and the COLUMN electrode is represented by an index of N-1, N, N + 1. Represents.

  FIG. 12B shows the state of current-voltage characteristics when the switching element of the present invention and a functional element (in this case, one that operates as a memory by a nonvolatile resistance change) are connected in series.

  In the operation with ternary voltage control, as shown in FIG. 12C, in the selection operation, the RAW electrode M is set to the high level and the COLUMN electrode is used for reading the selected location (M, N). Change N to Low-level. This operation is preferably performed using switching elements arranged around the array. Since the switching element at this time is located around the array, a conventional MOS transistor or the like may be used.

  At this time, in the non-selected place, the bias point as shown in FIG. 12B is the U point or the O point, or it is almost zero (the bias point is the V point). It is possible to read out the selected place while eliminating the influence of the wraparound from (the bias point is the S point).

  By comparing the current value at the voltage Vs at the point S with the current value flowing through the reference resistor, it is determined whether the information of the memory unit, which is a functional unit, is <1> or <0>. Can do.

  By providing the middle level, it is possible to reduce the power consumption in the non-selected place, which is more preferable. Again, the bias point Vs may be equal to or higher than the voltage Vth at which nonlinearity rises.

  Most preferably, the operation shown in FIG. 13 is performed. FIG. 13 (a) shows a schematic diagram of an array-type functional element. The RAW electrode is represented by an index of M-1, M, M + 1, and the COLUMN electrode is represented by an index of N-1, N, N + 1. Represents.

  FIG. 13B shows the state of current-voltage characteristics when the switching element of the present invention and the functional element (in this case, one that operates as a memory by a nonvolatile resistance change) are connected in series. The operation when the switching characteristics described in FIGS. 7 and 8 are used by applying a pulse to (M, N) shown here is shown.

  In this case, in the selection operation, the characteristics of the switching element are changed by the same selection operation described with reference to FIG. 11 or 12 with respect to reading of the selected place (M, N).

  The applied voltage at this time is preferably applied on a pulse, and is applied by applying a voltage sufficiently higher than Vs.

  FIG. 13B shows a state in which the state is changed to a low resistance state by applying a pulse of Vb (> Vs). At this time, by comparing the current value at the voltage Vs at the S point with the current value flowing through the reference resistor, it is determined whether the information of the memory unit as the functional unit is <1> or <0>. Can be determined.

  Note that the non-selection unit can be selectively read because it is held in a sufficiently high resistance state with respect to the selection unit (X point at the voltage Vs).

  After the reading is completed, it is possible to return to the initial high resistance state (the X point at the voltage Vs) by applying a voltage pulse whose polarity is inverted as shown in FIG. 7 and FIG. it can.

  That is, in the selection operation, regarding the reading of the selected location (M, N), the switching element is turned on by applying a pulse, the information on the selected location is read using a reference resistance, and the switching device is applied by applying a pulse. A random access operation can be realized by performing a series of OFF operations on each selected location.

  This operation is preferably performed using switching elements arranged around the array. Since the switching element at this time is located around the array, a conventional MOS transistor or the like may be used.

  As described here, the voltage application is performed in a pulse form in performing the switching operation, but the operation by the pulse is preferable in terms of improving Joule heat loss and device efficiency. Switching operation by pulse width control was confirmed to respond at several hundred ps-DC (up to several tens of minutes). However, in order to improve Joule heat loss and device efficiency, it is still possible to pulse at several ns to several ms. It is preferable to operate by bias drive.

  The Low-Level shown here may be the GND-level.

  In addition, it is preferable to use one of the elements as a reference resistor used for reading, and it is preferable to detect by differential at that time.

  Reading using the reference resistor can be realized by a device configuration as shown in FIG. At the time of detection, the current flowing through the reference resistor array and the detection element can be detected as a voltage through an ohmic resistor, and compared with a comparator or the like, and the difference can be detected.

  At this time, using the principle of FIG. 21, it is preferable to detect the difference between the voltage V from the element resistance and the voltage Vref from the comparison resistance element through a differential amplifier as shown in FIG.

  By doing so, it is possible to reduce the contribution due to the series resistance of the pass transistor or the diode used for selecting the memory block as well as the wiring and the element portion, and the S / N of the output can be improved. Of course, the switching element of the present invention can also be used in this part.

  In the reading operation as described above, a preferable functional unit operates as a memory by a nonvolatile resistance change.

  FIG. 9 shows a memory based on the magnetoresistive effect as an example. A magnetoresistive memory element is configured by sandwiching a non-magnetic material between two ferromagnets, and a non-volatile resistance using a magnetoresistive difference generated by parallel and antiparallel magnetization of the two ferromagnets. It is a change element (nonpatent literature 1). The antiparallel / parallel control of the magnetization is performed using a magnetic field generated by passing a current through a wiring arranged near the magnetoresistive element, and the write operation as a memory corresponds to this (see FIG. 10A). ). In order to improve writing selectivity, writing by applying a biaxial magnetic field is preferable (see FIG. 10B).

  At this time, a Raw electrode or a Column electrode may be used, or other magnetic field generation wiring may be provided.

  In order to efficiently use the magnetoresistive difference, it is preferable that one magnetic body is a soft magnetic body and the other is a hard magnetic body to provide a holding force difference.

  Or the structure which provided the coercive force difference by contacting an antiferromagnetic body to one magnetic body is also preferable. Also, a laminated ferrimagnetic material having a multilayer structure of a magnetic material and a nonmagnetic material is preferably arranged as one of the ferromagnetic materials.

  Furthermore, it is preferable that a nonmagnetic material is arranged as a tunnel insulating film to constitute an element using a large magnetoresistance change that reflects the magnetic polarizability of the ferromagnetic material.

  In order to realize the switching elements as described above, preferable materials for electrodes (including RAW electrodes and COLUMN electrodes) include Cu, Al, Ag, Au, Pt, TiN, W, and a resistivity of 10 mΩ · cm. The following materials may be used.

Further, in view of the compatibility with the strongly correlated electron materials to be joined, as electrodes, resistivity of an oxide material of the following 10 m [Omega · cm, it may be a material represented by the general formula A _x B _y O _z .
In the formula, A is an alkali metal (Group 1A element), an alkaline earth metal (Group 2A element), a rare earth element, scandium, or yttrium. B is a transition metal of any one of group 1B elements to group 8 elements, or any of 3A group elements, 4A group elements, 5A group elements, 6A group elements, and 7A group elements. O represents oxygen. The substance has a crystal structure.

The electrode may also be magnetic, Fe, Co, Ni, FeCo alloy, NiFe alloy, CoNi alloy, NiFeCo alloy, or FeN, FeTiN, FeAlN, FeSiN, FeTaN, FeCoN, FeCoTiN, FeCo (Al, Si ) N, FeCoTaN and other nitrides, oxides, carbides, borides, fluoride magnetic materials represented by formula: TMA (where T is at least one selected from Fe, Co, Ni, M is At least one selected from Mg, Ca, Ti, Zr, Hf, V, Nb, Ta, Cr, Al, Si, Mg, Ge, Ga, and A is from N, B, O, F, C At least one selected),
Or the formula: (Co, Fe) M (wherein M is at least one selected from Ti, Zr, Hf, V, Nb, Ta, Cu, B),
Or FeCr, FeSiAl, FeSi, FeAl, FeCoSi, FeCoAl, FeCoSiAl, FeCoTi, Fe (Ni) (Co) Pt, Fe (Ni) (Co) Pd, Fe (Ni) (Co) Rh, Fe (Ni) (Co ) Formula represented by Ir, Fe (Ni) (Co) Ru, FePt, etc .: TL (wherein T is at least one selected from Fe, Co, Ni, L is Cu, Ag, Au, Pd, Pt, Rh, Ir, Ru, Os, Ru, Si, Ge, Al, Ga, Cr, Mo, W, V, Nb, Ta, Ti, Zr, Hf, La, Ce, Pr, Nd, Pm, Sm, A ferromagnetic material such as Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu)
Alternatively, Fe ₃ O ₄ or XMnSb (X is at least one selected from Ni, Cu, Pt), a half metal material represented by LaSrMnO, LaCaSrMnO, CrO ₂ ;
Or, the formula: QDA (wherein Q is at least one selected from Sc, Y, lanthanoid, Ti, Zr, Hf, Nb, Ta, Zn, A is selected from C, N, O, F, S) At least one selected from the group consisting of V, Cr, Mn, Fe, Co, and Ni),
Or GaMnN, AlMnN, GaAlMnN, AlBMnN, etc .: RDA (where R is one selected from B, Al, Ga, In, D is selected from V, Cr, Mn, Fe, Co, Ni) 1 type selected, A is a type selected from As, C, N, O, P, S)
Alternatively, a perovskite oxide, a spinel oxide such as ferrite, or a garnet oxide may be used as long as the resistivity satisfies 10 mΩ · cm or less.

  In addition, the material used for interlayer insulation between each electrode layer may be an insulator or a semiconductor, but particularly IIa to VIa, La, and Ce containing Mg, Ti, Zr, Hf, V, Nb, Ta, and Cr. Including lanthanoids, including Zn, B, Al, Ga, Si, IIb to IVb elements and at least elements selected from F, O, C, N, B, or polyimide or phthalocyanine organic molecules Preferably it is a material.

  The functional unit that can be used in the present invention needs to be unaffected by pulses applied to turn on and off the switching element of the present invention. An example of such a functional unit is a magnetoresistive element configured by sandwiching a non-magnetic material between a pair of ferromagnetic materials.

Ferromagnetic materials include Fe, Co, Ni, FeCo alloy, NiFe alloy, CoNi alloy, NiFeCo alloy, FeN, FeTiN, FeAlN, FeSiN, FeTaN, FeCoN, FeCoTiN, FeCo (Al, Si) N, FeCoTaN, etc. Formulas represented by nitrides, oxides, carbides, borides, and fluoride magnets: TMA
(Wherein, T is at least one selected from Fe, Co, Ni, M is Mg, Ca, Ti, Zr, Hf, V, Nb, Ta, Cr, Al, Si, Mg, Ge, Ga At least one selected from A, and A is at least one selected from N, B, O, F, C),
Or the formula: (Co, Fe) M (M is at least one selected from Ti, Zr, Hf, V, Nb, Ta, Cu, B),
Or FeCr, FeSiAl, FeSi, FeAl, FeCoSi, FeCoAl, FeCoSiAl, FeCoTi, Fe (Ni) (Co) Pt, Fe (Ni) (Co) Pd, Fe (Ni) (Co) Rh, Fe (Ni) (Co ) Formula represented by Ir, Fe (Ni) (Co) Ru, FePt, etc .: TL (wherein T is at least one selected from Fe, Co, Ni, L is Cu, Ag, Au, Pd, Pt, Rh, Ir, Ru, Os, Ru, Si, Ge, Al, Ga, Cr, Mo, W, V, Nb, Ta, Ti, Zr, Hf, La, Ce, Pr, Nd, Pm, Sm, A ferromagnetic material such as Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu)
Alternatively, Fe ₃ O ₄ or XMnSb (X is at least one selected from Ni, Cu, Pt), a half metal material represented by LaSrMnO, LaCaSrMnO, CrO ₂ ;
Or, the formula: QDA (wherein Q is at least one selected from Sc, Y, lanthanoid, Ti, Zr, Hf, Nb, Ta, Zn, A is selected from C, N, O, F, S) At least one selected from the group consisting of V, Cr, Mn, Fe, Co, and Ni),
Or GaMnN, AlMnN, GaAlMnN, AlBMnN, etc .: RDA (where R is one selected from B, Al, Ga, In, D is selected from V, Cr, Mn, Fe, Co, Ni) A is a magnetic semiconductor represented by A, one selected from As, C, N, O, P, S)
Alternatively, perovskite oxides, spinel oxides such as ferrite, and garnet oxides are preferable.

In particular, when a soft magnetic film is used, a material such as Co or Co—Fe, Ni—Fe, or Ni—Fe—Co alloy is excellent. When a Ni—Fe—Co film is used, Ni _x Fe _y Co _z Ni-rich soft magnetic film having an atomic composition ratio of 0.6 ≦ x ≦ 0.9, 0 ≦ y ≦ 0.3, and 0 ≦ z ≦ 0.4, or Ni _{x ′} Fe _{y ′} Co _{z It} is desirable to use a Co-rich film with 0 ≦ x ′ ≦ 0.4, 0 ≦ y ′ ≦ 0.5, and 0.2 ≦ z ′ ≦ 0.95.

  Further, as the hard magnetic material having high coercive force, a material having a coercive force of 100 Oe or more such as CoPt, FePt, CoCrPt, CoTaPt, FeTaPt, FeCrPt is preferable, and as the antiferromagnetic material, PtMn, PtPdMn, FeMn, IrMn NiMn, etc. are preferable, and as the laminated ferrimagnetic material having a multilayer structure of a magnetic material and a nonmagnetic material, the magnetic material is Co or Co-containing FeCo, CoFeNi, CoNi, CoZrTa, CoZrB, CoZrNb alloy, etc. In addition, the nonmagnetic material at this time is preferably composed of Cu, Ag, Au, Ru, Rh, Ir, Re, Os, or an alloy or oxide of these metals.

In particular, when the nonmagnetic layer is a tunnel insulating layer, for example, IIa to VIa containing Mg, Ti, Zr, Hf, V, Nb, Ta, Cr, lanthanoids containing La, Ce, Zn, B, Al, Ga A thin film precursor of an element or alloy or compound selected from IIb to IVb containing Si, Si is prepared, and this includes any element of F, O, C, N, B, molecule or ion, radical, etc. By reacting in an appropriate atmosphere, temperature, and time, it can be produced by almost completely fluoridating, oxidizing, carbonizing, nitriding, and boriding. In addition, as a thin film precursor, a non-stoichiometric compound containing F, O, C, N, B below the stoichiometric ratio is prepared, and this is an element, molecule or ion of any of F, O, C, N, B, You may make it suitable atmosphere, temperature, time, and reactivity containing a radical etc. For example, when Al ₂ O ₃ is produced as a tunnel insulating layer using a sputtering method, Al or AlO _x (X ≦ 1.5) is formed in an Ar atmosphere or an Ar + O ₂ atmosphere. This can be achieved by repeating the reaction in O ₂ or O ₂ + inert gas. For plasma and radical production, ordinary means such as ECR discharge, glow discharge, RF discharge, helicon or ICP can be used.

  The configuration of the present invention described above can be generally realized by using a normal thin film process and a microfabrication process. For the formation of each layer of strongly correlated electron system material layer, insulating layer, functional part, electrode, etc., pulse laser deposition (PLD), ion beam deposition (IBD), cluster ion beam or RF, DC, ECR, helicon Further, it can be produced by sputtering such as ICP or counter target, PVD such as MBE or ion plating, and other CVD, plating or sol-gel methods.

  Microfabrication includes semiconductor processes, physical or chemical etching methods such as ion milling, RIE, and FIB used in magnetic device manufacturing processes such as GMR and TMR magnetic heads and magnetic memories (MRAM), and fine patterns. This can be achieved by combining a photolithography technique using a stepper, EB method or the like for formation. It is also effective to use CMP or cluster ion beam etching for planarizing the surface of the electrode or the like.

  More specific embodiments of the present invention will be described below.

Example 1
Samples were prepared in the following manner using pulsed laser deposition (PLD) and magnetron sputtering.

From left to right,
Sample 1-1
MgO (100) substrate / YBa ₂ Cu ₃ O ₇ (100) / Pr _0.8 Ca _0.2 MnO ₃ (150) / Pt (200)
Sample 1-2
Nb: SrTiO ₃ (100) substrate / Pr _0.7 Ca _0.3 MnO ₃ (150) / Ag (200)
Sample 1-3
Si (SiO ₂ thermal oxidation) substrate / Pt (400) / Pr _0.8 Ca _0.2 MnO ₃ (150) / Au (200)
Sample 1-4
MgO (100) substrate / La _0.7 Sr _0.3 MnO ₃ (100) / Pr _0.8 Ca _0.2 MnO ₃ (150) / Pt (200)
Sample 1-5
MgO (100) substrate / Pt (400) / Pr _0.8 Ca _0.2 MnO ₃ (150) / ITO (200)
Sample 1-6
NdGaO ₃ (100) substrate / La _0.7 Sr _0.3 MnO ₃ (100) / Pr _0.8 Ca _0.2 MnO ₃ (150) / La _0.7 Sr _0.3 MnO ₃ (100)
Sample 1-7
SrTiO ₃ (100) substrate / La _0.1 Sr _0.9 RuO ₃ (100) / La _1.95 Ba _0.05 CuO ₄ (150) / La _0.1 Sr _0.9 RuO ₃ (100)
Sample 1-8
MgO (110) substrate / TiN (300) / Fe ₃ O ₄ (100) / TiN (100)
Sample 1-9
a-Al ₂ O ₃ (0001) substrate / V ₂ O ₃ (100) / TiO ₂ (150) / V ₂ O ₃ (100)
Sample 1-10
a-Al ₂ O ₃ (0001) substrate / V ₂ O ₃ (100) / TiO ₂ (10) / VO ₂ (50) / TiO ₂ (10) / V ₂ O ₃ (100)
Sample 1-11
MgO (100) substrate / YBa ₂ Cu ₃ O ₇ (100) / [Pr _0.8 Ca _0.2 MnO ₃ (100) / SrRuO ₃ (20)] _N / Pt (200)
(Unit in parenthesis is nm)
As a result, a laminate was prepared.

  The Y-Ba-Cu-O layer, Pr-Ca-Mn-O layer, and La-Sr-Mn-O layer each have a substrate temperature of about 600-900 degrees in PLD (typically 750-850 degrees). The ITO layer was prepared at a substrate temperature of about 400-800 degrees (typically 550 degrees), and in the PLD method, each layer was deposited at an oxygen atmosphere of 0.1 m to 500 mTorr (typically 1 m to 100 mTorr). . Each layer of Pt, Ag, Au, etc. was formed at the substrate temperature of room temperature by sputtering.

  In the process of PLD and sputtering, the sample was moved while maintaining a high vacuum by in-situ conveyance.

  The laminated body was processed by EB (electron beam) processing and a photolithographic technique to create a structure as shown in FIG.

In Sample 1-1, the Y-Ba-Cu-O layer was disposed as the lower electrode, the Pr-Ca-Mn-O layer as the layer using a strongly correlated electron material (SCEM), and Pt as the upper electrode. Hereinafter, in Sample 1-2, the Nb: SrTiO ₃ substrate is the lower electrode, the Pr-Ca-Mn-O layer is the SCEM layer, Ag is the upper electrode, and in Sample 1-3, the Pt layer is the lower electrode, Pr-Ca- Mn-O layer as SCEM layer, Au as upper electrode, sample 1-4 with La-Sr-Mn-O layer as lower electrode, Pr-Ca-Mn-O layer as SCEM layer, Pt as upper electrode In 1-5, the Pt layer is the lower electrode, the Pr-Ca-Mn-O layer is the SCEM layer, and the ITO layer is the upper electrode. In Sample 1-6, the La-Sr-Mn-O layer is the lower electrode, and the Pr-Ca- In Sample 1-7, the Mn-O layer is the SCEM layer, the La-Sr-Mn-O layer is the upper electrode, the La-Sr-Ru-O layer is the lower electrode in Sample 1-7, the La-Ba-Cu-O layer is the SCEM layer, La-Sr-Ru-O layer as the upper electrode, in sample 1-8, the TiN layer is the lower electrode, Fe-0 layer is the SCEM layer, Ti-N layer is the upper electrode, and in sample 1-9, the V-O layer Is the lower electrode, the Ti—O layer is the SCEM layer, the V—O layer is the upper electrode, and in Sample 1-10, the V—O layer is the lower electrode, / V-O / Ti-O multilayer as SCEM layer, VO layer as upper electrode, sample 1-11, Y-Ba-Cu-O layer as lower electrode, superlattice structure as SCEM layer, Pt as upper electrode Arranged.

  Further, Au, Ag, Pt, Cu, Al, etc. were used as wiring electrodes. In consideration of contact and processability, in this embodiment, an electrode having a multilayer structure such as Ta (5) / Cu (500) / Pt (10) was used.

In addition, in the Nb: SrTiO ₃ substrate of Sample 1-2, a substrate containing 0.5% Nb was used. As the ITO of Sample 1-5, tin-added indium oxide with 5% Sn added to In ₂ O ₃ was used.

  FIG. 20A shows a typical current I-voltage V characteristic between the upper and lower electrodes of Sample 1-1. A characteristic having a rectifying property is clearly observed. Next, it was confirmed that the behavior shown in FIG. 20B was changed by applying a single pulse of +5 V and a width of 50 ns to the sample 1-1. It was also confirmed that the behavior shown in FIG. 20A was restored by applying a single pulse of -5 V and 50 ns width. When viewed with a non-linearly strong low bias, the reverse behavior control can be realized by applying a positive pulse and switching from a high resistance state to a low resistance state by applying a negative bias. Since this behavior is observed at 400 K or less including at least room temperature, it can be used in a wide temperature range.

  Similar characteristics were observed for Samples 1-2, 1-3, 1-4, 1-5, 1-6, 1-7, 1-8, 1-9, 1-10. It has been shown that switching elements can be realized.

  It was found that the same kind of behavior in the range of at least 10 nm to 10 μm was confirmed in the thickness of the SCEM of the series of examples obtained here.

  For sample 1-11, a superlattice structure was provided in the SCEM layer, but the same type of behavior as sample 1 was confirmed when N was in the range of at least 1-30.

  In this example, a Pr-Ca-Mn-O layer (or La-Ba-Cu-O layer, Fe-O layer, Ti-O layer) was used as the SCEM layer, but Nd- Of the A-site elements, such as Ca-Mn-O and Sm-Ca-Mn-O, those obtained by changing rare earth elements, or Nd-Sr-Mn-O and Sm-Ba-Mn-O The same effect was observed even when the alkaline earth element was changed. In addition, it was observed even when B-site elements such as La-Sr-Ti-O and La- (Sr, Ba) -Zr-O were changed.

Furthermore, the following sample 1-12
YAlO ₃ (100) substrate / La _1.7 Sr _0.3 CuO ₄ (100) / Bi ₂ Sr ₂ CoO ₆ (100) / Ag (100)
Sample 1-13
SrTiO ₃ (100) substrate / La _1.7 Sr _0.3 CuO ₄ (100) / Bi ₂ Sr ₂ CuO ₆ (100) / Ag (100)
Sample 1-14
a-Al ₂ O ₃ (0001) substrate / IrO ₂ (100) / SrCoO ₃ (100) / Pt (100)
Sample 1-15
a-Al ₂ O ₃ (0001) substrate / IrO ₂ (100) / LiNbO ₃ (100) / Pt (100)
Sample 1-16
a-Al ₂ O ₃ (0001) substrate / IrO ₂ (50) / NiO (100) / Pt (100)
Sample 1-17
a-Al ₂ O ₃ (0001) substrate / IrO ₂ (50) / VO ₂ (100) / Pt (100)
Similar behavior was confirmed in the sample configuration.

  At this stage, the detailed mechanism of this change in nonlinearity is not clear, but the spatial inhomogeneity of the electronic state, which is believed to be inherently possessed by strongly correlated electron materials, is caused by the application of pulse bias. It is thought that it has changed and its conductivity has also changed.

In addition, as a switching element of this invention, the structure like FIG. 18 which arranged the electrode in the surface was also produced and evaluated. The SCEM layer used was Pr _0.7 Ca _0.3 MnO ₃ , La _1.95 Ba _0.05 CuO ₄ , Fe ₃ O ₄ , TiO _2, etc., and the electrode was Ag, Pt, or Au. In this configuration, the same characteristics as in FIG. 20 were observed, and it was shown that the desired operation was possible in this configuration.

(Example 2)
Samples were prepared in the following manner using magnetron sputtering and metal organic compound deposition (MOD).

From left to right,
Sample 2-1
Si (SiO ₂ thermal oxidation) substrate / Pt (400) / Pr _0.8 Ca _0.2 MnO ₃ (150) / Pt (20) / Ta (5) / Cu (100) / Pt _0.49 Mn _0.51 (15) / Co _0.5 Fe _0.5 (5) /Ru(0.9)/Co _0.5 Fe _0.5 (5) / Al ₂ O ₃ (1.2) / Co _0.5 Fe _0.5 (1) / Ni _0.81 Fe _0.19 (20) / Pt (20) Pr-Ca- The Mn-O layer was produced by the MOD method. The substrate temperature at this time was 350 to 700 degrees. All other layers were deposited at room temperature by magnetron sputtering. The sample was annealed in a magnetic field at 5 kOe and 280 ° C. for orientation of PtMn.

  This configuration is substrate / electrode / strongly correlated electron system material / electrode / pinned magnetic layer / nonmagnetic layer / free magnetic layer / electrode, and is basically the same as that shown in FIG.

  An array configuration as shown in FIG. 6B was produced using the produced sample. As shown in FIG. 16, the shape of the multilayer sample was regulated by using a photoresist mask so that the lower electrode is a column wiring and using argon ion milling or reactive ion etching (FIG. 16 ( b)).

  The same process is repeated to fabricate the element portion of the array (FIG. 16C), and the upper and lower wirings are separated by the interlayer insulating layer (FIG. 16D). Planarization is performed, a raw wiring electrode layer is deposited thereon, and a raw wiring shape is defined using argon ion milling or reactive ion etching using a photoresist mask similar to the above (FIG. 16 ( e)) After that, interlayer insulation for word wiring for magnetic field application (FIG. 16 (f)), deposition and shape definition of the word wiring electrode were performed, and a desired array type element was fabricated (FIG. 16 (g) )).

  Although an example of manufacturing is shown here, the present invention is not limited to this, and a lift-off process may be performed in the processes of FIGS. 16C to 16D, and any method that can manufacture a similar configuration may be used. Good. At this time, the raw line and the word line arranged on the upper part were mainly made of Cu.

  The produced element arrangement was 16 blocks of 16 × 16 elements, for a total of 8 blocks.

  Each element cross-sectional area of the sample is 0.2 μm × 0.3 μm, and the shape is as shown in FIG.

  By using a Word line and a Raw line and applying a magnetic field generated by passing a current, magnetization information can be selectively written to the free magnetic layer in each element. The magnetization reversal of each free magnetic layer was simultaneously written in accordance with the information on 8 elements of 8 blocks, and a signal of 8 bits was recorded. Next, the MOS gates created as pass transistors were turned on one by one for each block, and a current for reading was passed between the Column line and the Raw line. At this time, the voltages generated in the raw lines, elements, and pass transistors in each block were compared with dummy voltages by a comparator, and 8-bit information was simultaneously read from the output voltage of each element. Although a pass transistor made of MOS is used here, the switching element of the present invention can of course be used.

  At this time, the ratio of the major axis to the minor axis of the free magnetic layer was 1.5: 1 (minor axis was 0.2 μm), and an integrated memory having a shape changed to that shown in FIGS. The power consumption required for recording in these memories is about 3/5 to 1/2 of the shape of FIG. 17A in the shape of FIGS. 17B to 17E, which is preferable. .

  Further, a finer element shape was produced using an electron beam by using the configuration of Sample 2-1. The produced shape was as shown in FIG. 17C, and the ratio of the major axis to the minor axis was 1.5: 1 (minor axis was 0.07 μm). Also in this case, it was confirmed that the same behavior as in FIG. 20 was obtained, and reading of magnetoresistive memory information in the case where an array shape was produced was also possible.

  The switching element of the present invention and the array type functional element using the switching element include a multilayer structure constituted by sandwiching a layer made of a strongly correlated electron system material with at least two electrodes, or at least two or more of them in multiple stages In a multi-stage structure that is laminated, current and voltage are applied between the electrodes at both ends of the sandwiched structure to change the rectifying action that is manifested, and by using its non-volatility, it can be used as various functional devices can do.

  In particular, since the rectification type nonlinear characteristic can be used for random access selection, it is useful for realizing an array type functional element arranged in an array. For this reason, non-volatile random access memories used for conventional information communication terminals, reconfigurable logic devices using them, and light, heat, magnetism, electricity, stress change, acceleration change chemical reaction, etc. It is also possible to realize various sensors for detecting the image, an imaging device in which these are arranged in an array, a display using a light emitting element, and the like, and the characteristics such as array integration and access speed can be improved.

The figure which shows an example (conventional) of a functional element using a MOS transistor as a switching element The figure which shows an example (conventional) of a functional element using a MOS transistor as a switching element The figure which shows the mode of reading of the array type functional element when there is no switching element (conventional) Diagram showing an overview of the configuration of an array-type functional element (conventional) when MOS transistors are arranged as switching elements The figure which shows the structure outline figure of the switching element of this invention The figure which shows the structure outline figure of the array type functional element of this invention The figure which shows the current-voltage characteristic of the switching element of this invention The figure which shows the current-voltage characteristic of the switching element of this invention The figure which shows the structure schematic in the case of using the switching element and non-volatile memory of this invention as a functional element Operational explanatory diagram when using the switching element and nonvolatile memory of the present invention as a functional element Operation explanatory diagram of array type functional element of the present invention Operation explanatory diagram of array type functional element of the present invention Operation explanatory diagram of array type functional element of the present invention Schematic configuration of array type functional element comprising switching element and functional element of the present invention Schematic configuration of array type functional element comprising switching element and functional element of the present invention The figure which shows the preparation method of the array type functional element of this invention The figure which shows the element shape of the array type functional element of this invention Configuration schematic diagram of switching element of the present invention The figure which shows the array type functional element of this invention The figure which shows the current-voltage characteristic of the switching element of this invention Output detection operation explanatory diagram of the array type functional element of the present invention

Claims (8)

It is composed of at least two electrodes with a layer made of a strongly correlated electron system material. Non-linear characteristics appear in the current-voltage characteristics, and the non-linear characteristics change by applying current or voltage between the electrodes. A switching element. A multilayer structure composed of at least two electrodes sandwiching a layer of strongly correlated electron system material, and a multilayer structure composed of at least two or more layers, and its current-voltage characteristics are non-linear. And the nonlinear characteristic with respect to the current or the voltage is changed by applying a current or a voltage between the electrodes sandwiching the multistage structure. 3. The switching element according to claim 2, wherein at least one of the multi-stage structures has a nonlinear characteristic that is changed by applying a current or a voltage. 4. 4. The switching element according to claim 3, wherein in the change of the non-linear characteristic, the change is kept non-volatile when no current or voltage is applied. The strongly correlated electron material is represented by the formula: A _x B _y O _z (where A is an alkali metal (Group 1A element), alkaline earth metal (Group 2A element), rare earth element, scandium, yttrium, lead, bismuth , Thallium, and B are any of Group 1B elements to Group 8 elements, or 3A group elements, 4A group elements, 5A group elements, 6A group elements, 7A group element transition metals, and O is oxygen) 5. The switching element according to claim 1, wherein the switching element is a substance having a crystalline structure. The switching element according to claim 1, wherein the current or the voltage is applied in pulses. An array-type functional element comprising the switching elements according to claim 1 arranged in an array. 8. The array type functional element according to claim 7, wherein the switching element is used as a random access selection element.

Images