JP2002280627A - Superconductive switching element - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a superconductive four-terminal element capable of performing an ultra-high speed switching operation. SOLUTION: The four-terminal element is constructed by using an S-N-S junction capable of controlling the flow of electrons without using a dielectric barrier to remove parasitic capacitance (electronic polarization), and thicknesses of superconductive layers and conductive layers are optimized so that a spin exchange correlation operation is maximized. In this way, the superconductive layer operates as if an insulator ware provided, Bloch resonance is generated in the case of a semiconductor superlattice, and a resonant state is varied by a current or voltage applied to a gate layer to perform a switching operation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION In the field of electronic information, a central processing unit (Central Processing) is used.
(Unit; CPU) and a high-speed switching element used for a storage device and a transmission device. It is capable of operating at a high frequency and is used as an element for converting an electric signal to a high-energy photon and an element for converting a photon to an electric signal.

[0002]

2. Description of the Related Art Due to an increase in the amount of information processing in the field of computers, a central processing unit (Central Processing Uni) at the heart of the computer.
t; CPU) on the order level. Up to now, the speeding-up of the CPU has been delegated to a principle of action called a "reduction rule", and has been performed by integration of elements. However, speeding up by this "reduction rule" is beginning to see its limits. To overcome this problem, devices using superconductivity have been devised, but have not been put to practical use. This is considered to be due to the presence of the dielectric material used for the element to partition the current. if,
If the current can be partitioned without using dielectric materials,
It seems that the problem of ultra-high-speed switching operation using a superconducting material is solved. In this application, the applicant proposes a four-terminal switching element that does not use a dielectric barrier. W. B. Shockley, J.M. Barde
en, W.C. H. Since Brattain et al. Discovered the transistor effect in 1948, the semiconductor devices used in computers have created a number of world-changing concepts. The speed of processing the data has been required to be the fastest in keeping with the changing flow of the concept. Also in the field of electronic devices, components have been integrated based on the "reduction rule" in order to follow this "flow". The “reduction rule” refers to a field effect transistor (Field Effect Tran) which is a basic element of an integrated circuit.
The concept is to enable high-speed switching by reducing the product of the resistance value and the capacitance of the entire circuit including the FET (sistor; FET). However, we are beginning to realize that the demands of modern society cannot be overcome by this "reduction rule." Because
Since the switching operation of the FET is performed by controlling the carrier near the gate of the FET with a bias voltage, the switching operation depends on the mobility of the carrier. Recently, Y. Nishi has projected a long-term increase in CPU operating frequency by 2010. According to that, even in 2010, the operating frequency is 1.1
¹ be up to x10 ⁹ Hz are prophetic. That is, the information processing capability of the CPU only increases at most several times. This means that an enormous amount of information processing performed by huge software depends on carrier mobility. On the other hand, B. D. Josephs
The superconducting element proposed by on operates at high speed,
² which is known to be a low power consumption. But Jo
Sephson elements have not been put to practical use in the high-number wavenumber region because noise due to chaotic signal fluctuation becomes dominant at a switching frequency of 7 × 10 ⁸ Hz or more. In addition, it has the properties of reduced transmittance due to the presence of the tunnel barrier, signal delay, and mechanical vulnerability to thermal strain. K. A.
Muller, J .; G. FIG. The oxide superconductor ³ discovered by Bednorz et al. Was expected to be applied to an element operating at a high temperature, but has not yet been put to practical use as a switching element. The fact that the oxide superconductor itself did not come into practical use despite numerous trials was due to the properties of the oxide superconductor itself. This application shows that the use of a metal superconductor superlattice can solve all the problems for practical use.

[0003]

The oxide superconductor is made of Jo.
The reason why it was not put into practical use as a sephson element is as follows.
There are two. First, the phase of the wave function fluctuates due to the existence of an interface having low consistency such as a crystal grain boundary. Second, at such poorly matched interfaces, the signal of the wave function is extremely attenuated during integration because the transmittance of the wave function is very small. The coherence length of oxide superconductors is designed to be extremely short (〜0.1 nm) to increase the superconducting phase transition temperature (T _c ) ⁴ . However, for that purpose, the coherent region must be partitioned by an ionic crystal having a high dielectric constant. As a result, the dielectric constant of the crystal itself increases, and a delay in the wave function between the conductive layers is inevitable (FIG. 1a). Also, at interfaces with low consistency,
This delay is not uniform (FIG. 1b). This Not dynamic impedance from uniformity element ends up rising ^5,6 such ^as, transmittance will be lowered. That is, the problem of the oxide superconductor can also be attributed to the existence of the dielectric constant.

[0004]

[Means for Solving the Problems]

[About the method of partitioning the coherent region without using a dielectric material] The characteristics of the superconductivity phenomenon are described by the coherence of the wave function of the Cooper pair, and the coherence is realized by spin exchange correlation. This means that if the spin exchange correlation can be controlled, the electron flow can be manipulated. Therefore, as an example of artificially controlling the spin exchange correlation, a giant magnetoresistance effect (Giant Magnene) is used.
To-Resistence (GMR) expression mechanism will be considered. It has been reported that the giant magnetoresistance effect appears in a metal system having a mesoscopic structure, but the application of external magnetism reduces the electrical resistance by as much as 50% ⁷ . The giant magnetoresistance effect occurs because the transition of itinerant electrons is restricted by the Kondo effect ⁸ when the wave function described up to spin differs between adjacent layers. In a system showing a giant magnetoresistance effect, although there are reports ⁹ regarding mesoscopic magnetic order, according to which the double magnetic order of the superlattice period is present it is confirmed by neutron diffraction. What should be emphasized here is that the domain structure on a mesoscopic scale appears by controlling the period of the superlattice. This means that, from the standpoint of spin exchange correlation, the transition of itinerant electrons is controlled without disposing a dielectric by changing the lamination period of the superconducting layer that is strongly correlated. With this suggestion, it is possible to create a superlattice that operates as a switch without using a dielectric.

[Possibility of switching element using metal superconducting superlattice] Superconducting metal-normal metal-superconducting metal (Super
conducting metal-Normal m
etal-Superconductingmeta
l; SNS) In the superlattice, a dimensional crossover phenomenon has been observed. I. K. Group of Schuller et al., Plays a pioneering role in the investigation of dimensional crossover phenomenon in the super lattice of Nb / Cu system ^10. According to the report, the coherence length of the Nb / Cu superlattice has anisotropy only in a mesoscopic system, and the correlation between the superconducting layers is maximized by optimizing the superlattice period. The applicant has long described spins using the correlation between layers. It has been proposed to partition the wave function. Recently, the applicants reported that an Nb / Cu superlattice with a mesoscopic superlattice period has antiferromagnetic magnetic order per superconducting layer ¹¹ . When the electrical resistance of the Nb / Cu-based superconducting superlattice having Nb, Cu, and the layer thickness of each layer fixed at 16.8 nm and 14.7 nm was measured, abnormal resistance was observed with good reproducibility. Considering that this resistance anomaly appears only when the correlation between the layers becomes maximal, the antiferromagnetic order of the electronic band is realized on a mesoscopic scale. (Fig. 2). if,
Assuming that antiferromagnetic magnetic order is formed in a superconducting system, the analogy between the superconductivity phenomenon and GMR can be concluded. As described above, in a system exhibiting GMR, the flow of electrons can be partitioned without using a dielectric barrier. Therefore, it is considered that even in the SNS junction, the flow of the Cooper pair can be partitioned without using the dielectric barrier.

[0005]

FIG. 3 shows an outline of a four-terminal switching element. The electron transition between layers becomes dominant only when the wave functions including the spin description coincide, and the transition to the next layer is restricted by the antiferromagnetic spin arrangement ( Figure 3a). Based on this philosophy, a four-terminal device having a source layer, a gate layer, and a drain layer has been proposed by one of the applicants (K. Tsukui) (FIG. 3b). The operating principle of the switching operation is to modulate the Broch resonances ¹² , ¹³ and Br
Due to the nonlinearity of electron transport at och resonance. As depicted in FIG. 4, the occupied orbitals having different spins in the superconducting layer are separated by a normal conducting layer, similar to a semiconductor superlattice. Cooper Pa inside the superlattice
ir is in a resonance state, and the resonance state is modulated by a gate operation or a single event introduced from the outside. The modulation of the Broch resonance results in the source-
The voltage across the drain is multi-valued by the nonlinear transport of the superlattice ¹⁴ . The phenomenon is the same as the semiconductor superlattice, but the handling of the tunnel barrier in the superlattice is different. In the case of the SNS superlattice, itinerant electrons propagate through the paramagnetic metal. However, the transition at low temperature from one superconducting layer to the next superconducting layer is limited by the Kondo effect. The value of the electrical resistance due to itinerant electrons (ρ _↑ ) is determined by the following equation ⁸ . The value of the electrical resistance is a function of the temperature and the electron density (N _Para (ε _F )) at the Fermi level of the normal metal. As a result, the transition probability of itinerant electrons below 100K is significantly reduced, and the inserted paramagnetic metal acts like an insulator barrier. This function is similar to the band gap of the semiconductor superlattice. This "tunnel barrier effect"
^15, and electrons in a strong layer of spin exchange-correlation, as a phenomenon which occurs when between the angular momentum quantum number of electron paramagnetic in metal inserted in different, which has been recognized. In conclusion, since the four-terminal device operates while maintaining the superconducting current due to the Broch resonance, the upper limit of the operating frequency is determined by the gap frequency that determines the operating frequency of the conventional RC-type Josephson device. And different. Further, since the unit cells of the element are stacked in the order of source layer (S) / gate layer (G) / drain layer (D) / G / S / G / D / G. The correlation from the source layer to the drain layer approaches 1 in geometric progression. Josephso with dielectric barrier
The n-junction suffered from chaotic noise, fluctuations in the wave function, low transmittance, and mechanical brittleness. This means that all problems for the transition have been resolved. In addition, this element can operate at an extremely high frequency without energy loss like an element that operates by breaking a superconducting state. The upper limit of the operating frequency is expected to reach 10 ¹⁸ Hz, considering the plasmon loss of metals (on the order of 10 ³ eV for Cu). If operation at such high frequencies is possible, converting electrical signals into high-energy photons;
It is also considered that the reverse conversion and the detection of high energy particles are possible.

[0006]

DESCRIPTION OF THE PREFERRED EMBODIMENTS To date, there has been no example in which a switching element has been produced using the spin exchange correlation action. However, as described above, it is considered possible to manufacture a switching element capable of operating at high speed without a parasitic capacitance by using a superconductor. The method of making this device is described below. (1) A method for producing a layered structure by a mechanical alloying method. (2) A method of producing a layered structure by a processing method based on mechanical deformation (including a forging method, a rolling method, and the like). (3) Layered structure production method by controlling solidification structure (including rapid solidification method). (4) Thin film preparation method (physical preparation method, chemical preparation in gas phase, liquid phase and solid phase) Layered structure manufacturing method (including a synthetic method). (5) A method for producing a layered structure using a chemical technique. (6) A method of forming a layered structure using an electrochemical method. (7) A method for producing a layered structure using a biochemical technique. (8) A method for producing a layered structure utilizing the periodicity of a structure existing in the natural world. (9) A method for producing a layered structure using a driving force existing in the natural world. (10) A method for producing a layered structure using a physical method (such as a particle accelerator). It should be noted that a method combining a plurality of these is included.

[0007]

In conclusion, this device can perform a switching operation at a very high frequency without energy loss like a device that operates by breaking a superconducting state. The upper limit of the operating frequency is the plasmon loss (C
Considering 10 ³ eV for u), 10
Expected to reach ¹⁸ Hz. If it is possible to operate at such a high frequency, it would be possible to convert an electrical signal into high-energy photons, vice versa, and detect high-energy particles.

[0008]

[Acknowledgment] The applicant expresses his deep appreciation to Professors Hideo Shingu and Keiichi Ishihara of Kyoto University. The point that GMR and superconductivity are very similar has matured in discussions with the teachers. Special thanks to Dr. Masanori Yada of National Institute for Metals Technology for guidance on thin film preparation. The applicant also expresses his deep appreciation for the fruitful discussions with the Professors at the National Institute of Metals and Materials, the National Institute of Advanced Industrial Science and Technology, the Institute of Mechanical Engineering, the Institute of Industrial Science and Technology, Waseda University, and Kyoto University. You. We would like to thank Professor Yasunori Otsuka of Otsuka International Patent Office for much advice on patenting. Finally, one of the applicants (K. Tsukui) would like to thank Dr. Taiki Kocha of Sakuragaoka Memorial Hospital for his kind guidance on his overall research life.

[Brief description of the drawings]

FIG. 1 shows the effect of electron polarization inside an oxide superconductor on propagation of a wave function between conductive sheets (a). In the case of a Josephson device using an oxide superconductor, 10
⁴ in Hz or more switching frequencies, thermal noise becomes dominant ^6. K. K. According to Likharev's report ⁵ , such thermal noise is due to an increase in dynamic impedance in a high-frequency region. That is, a crystal having a perovskite structure has large electron polarization. Wave functions are delayed during propagation due to the presence of their electronic polarization. FIG.
FIG. 2 shows a diagram illustrating that the delay of the wave function fluctuates near the interface having low matching. Due to the delay of the wave function, the order parameter near the interface becomes smaller, and the wave function necessarily becomes weakly coupled.

FIG. 2 shows the antiferromagnetic spin order per layer in a superconductor having a mesoscopic structure obtained as a result of the resistance measurement experiment ¹¹ . The electronic state in each layer was calculated by the DV-Xα method ¹⁶ , which is an ab initio molecular orbital method. The "thermal stability mechanism of the Cooper pair" ¹¹ proposed at the same time allows spin fluctuations to be considered independently of the normal electronic state. Therefore, it becomes possible to describe the mesoscopic spin order based on the normal electronic state. The spin direction of the next layer is reversed to reduce the total energy represented by the function of the exchange correlation parameter. If such an antiferromagnetic magnetic order exists in a superconductor, the flow of electrons can be separated without using a dielectric.

FIG. 3 is a schematic diagram of a four-terminal switching element and its operation principle. Transitions between layers having the same wave function including spin are prioritized (a). As a result of the Kondo effect ⁸ , the transition to the next layer is restricted. The unit cell of the four-terminal element includes a source layer, a gate layer, and a drain layer (b). The OFF operation of the switch is performed by modulating the Broch resonance existing in the element by the electronic state of the gate layer or a single event from the external environment. Considering that the superconductivity phenomenon and the GMR are analogous, the flow of electrons can be separated without using a dielectric. That is, it can be seen that this layer can be manufactured by stacking layers having a strong spin exchange correlation action.

FIG. 4 is a schematic diagram of an energy band inside a unit cell of a proposed four-terminal device. As described above, the spin fluctuation can be considered independently of the normal electronic state by the “thermal stability mechanism of the Cooper pair” ¹¹ . Because of that, Spin order can be described based on the normal electronic state. The electronic states inside the superconducting layer are arranged antiferromagnetically on a mesoscopic scale and are separated by a normal metal. As a result of the Kondo effect or "tunnel barrier effect", Br
An och resonance has been established.

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 4M113 AA04 AA14 AA29 AB01 AB11 AB15 CA13

Claims (1)

[Claims] (1) A switch element in which superconductors (layer A) and paramagnetic normal conductors (layer B) are alternately laminated. (2) A switch element having terminals provided only in the A layer. (3) A switch element using a superconductor instead of the paramagnetic normal conductor of the B layer. (4) As shown in FIG. 3, the A layer is named a source layer, a gate layer, and a drain layer in order, and applies a Broch resonance between the source layer and the drain layer to the gate layer. Element that performs a switching operation. (5) An element obtained by stacking the device of item 4 a plurality of times. (6) An element in which another metal is inserted between the A layer and the B layer.