JP4836028B2 - Superconducting quantum arithmetic circuit - Google Patents

Description

  The present invention relates to a superconducting quantum arithmetic circuit, and more particularly to a superconducting quantum arithmetic circuit capable of completely turning off the interaction between qubits.

  With the development of technology in recent years, superconducting qubit devices for realizing quantum computation have been realized. For example, in a magnetic flux qubit device as described in Non-Patent Document 1, an external magnetic flux equivalent to 1/2 of the quantized magnetic flux is applied to a superconducting loop 201 having three Josephson junctions 202 as shown in FIG. When applied, the superposition state of the two states, the clockwise state and the counterclockwise state, of the permanent current that circulates the loop becomes the two eigenstates with the lowest energy, and the effective quantum two-level system, that is, the qubit Works as.

  Furthermore, in order to construct a qubit operation circuit by integrating qubit elements, a conditional operation gate between two qubits is required by introducing an interaction between qubits. For example, in the conventional qubit arithmetic circuit described in Non-Patent Document 2, a part of the loop of two magnetic flux qubits 301 and 302 is shared as shown in FIG. Is realized.

  In Non-Patent Document 3, the interaction is strengthened by increasing the inductance by providing a large Josephson junction 403 at the loop sharing portion of the two magnetic flux qubits 401 and 402 as shown in FIG. On the other hand, in Non-Patent Document 1, a superconducting loop including a superconducting magnetic flux quantum interference element (SQUID) 504 is provided as a transformer 503 between magnetic flux qubits as shown in FIG. Realization.

  Non-Patent Document 4 realizes magnetic coupling between two magnetic flux qubits 601 and 602 having different configurations by using a transformer 603 in which two superconducting loops including SQUID 604 are coupled symmetrically as shown in FIG. is doing.

  In Non-Patent Document 5, the SQUID 704 is arranged in parallel to the superconducting loop 703 instead of in series as shown in FIG. Further, in Non-Patent Document 6, as shown in FIG. 8, a current-biased SQUID 803 is provided between magnetic flux qubits to realize magnetic coupling between the qubits 801 and 802 via the SQUID.

  On the other hand, each qubit needs to maintain coherence for a long time in the quantum operation circuit. As shown in Non-Patent Document 7, it is known that coherence of qubits greatly depends on bias conditions, and a long coherence time can be obtained at an optimum operating point.

J.E. Mooij, T.P.Orlando, L.S. Levitov, L. Tian, C.H.van der Wal, and S. Lloyd, Science 285, 1036 (1999). J.B.Majer, F.G.Paauw, A.C.J.ter Haar, C.J.P.M.Harmans, and J.E. Mooij, Physical Review Letters 94, 090501 (2005). M. Grajcar, A. Izmalkov, SHW van der Ploeg, S. Linzen, E. Il'ichev, Th. Wagner, U. Hubner, H.-G. Meyer, A. Maassen van den Brink, S. Uchaikin, and AM Zagoskin, PHYSICAL REVIEW B 72, 020503 (R) 2005. T.V. Filippov, S.K.Tolpygo, J. Mannik, and J.E.Lukens, IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY 13, 1005 (2003). C. Cosmelli, M. G. Castellano, F. Chiarello, R. Leoni, D. Simeone, G. Torrioli, P. Carelli, arXiv: cond-mat / 0403690. BLT Plourde, J. Zhang, KB Whaley, FK Wilhelm, TL Robertson, T. Hime, S. Linzen, PA Reichardt, C.-E.Wu, and J. Clarke, PHYSICAL REVIEW B 70, 140501 (R) (2004 ). P. Bertet, I. Chiorescu, G. Burkard, K. Semba, C.J.P.M.Harmans, D.P.DiVincenzo, J.E. Mooij, arXiv: cond-mat / 0412485.

  However, these conventional techniques have some problems.

  The first problem is that the interaction between qubits is fixed and not variable. This problem corresponds to the examples of FIGS. This is because the coupling inductance between qubits is not variable. This is because part of the superconducting loop or Josephson junction 403 is used as the coupling inductance.

  The second problem is that the interaction between qubits cannot be completely turned off. This problem corresponds to the examples of FIGS. In the example of FIG. 5, this cause is due to the critical current of SQUID 504 becoming too small. The reason is that applying a magnetic flux that is half of the quantized magnetic flux to SQUID 504 to turn off the interaction between qubits 501 and 502 causes the critical current of SQUID 504 to become zero, and the Josephson junction switches to a finite voltage state. This is because dissipation occurs and coherent magnetic coupling cannot be achieved. In the example of FIG. 7, the cause is that the SQUID 704 is arranged in parallel to the transformer loop 703. That is, even if the magnetic flux inside the SQUID 704 is changed to change the inductance of the SQUID 704, the coupling inductance between the quantum bits 701 and 702 cannot be completely reduced to zero.

  The third problem is that it can be applied only to a specific type of magnetic flux qubit, as in the example shown in FIG. This is due to the geometry of the coupling transformer 604. This is because, in the example of FIG. 6, in order to turn off the interaction, the qubits 601 and 602 have a special structure in which Josephson elements are arranged at orthogonal positions, and each qubit 601 and 602 itself is symmetrical. It is because it has a structure with the property.

  The fourth problem is that the control current that flows when controlling the state of individual qubits or controlling the interaction between qubits is prevented from cross-coupling to other qubits or SQUIDs. It is not possible. This problem corresponds to the example of FIG. The reason for this is that the example of FIG. 8 is magnetic via SQUID803 sandwiched between direct magnetic coupling between adjacent qubits to completely turn off the interaction between qubits 801 and 802. This is due to the use of the coupling cancellation phenomenon. The reason is that in order to use the direct magnetic coupling between adjacent qubits 801 and 802, the distance between the qubits must be reduced, thereby suppressing crosstalk between control signals. It is because it is not possible.

  The fifth problem is that when the interaction between qubits is controlled, the qubit bias point changes, and the qubit operating point deviates from the optimum condition, so that the qubit coherence time is shortened. It means that. This problem corresponds to the example of FIG. This is due to the current flowing through SQUID 803 to control the interaction. This is because a change in current flowing through the SQUID 803 changes the magnetic flux applied to the qubits 801 and 802 on both sides.

  An object of the present invention is to provide a superconducting quantum arithmetic circuit capable of on / off controlling the interaction between qubits while maintaining a long coherence time of each qubit.

  Another object of the present invention is to provide a transformer element capable of temporarily turning on / off the interaction between qubits.

  Still another object of the present invention is to provide a semiconductor device including the superconducting quantum arithmetic circuit or the transformer element.

  Another object of the present invention is to provide a quantum computer including the superconducting quantum arithmetic circuit or the transformer element.

  The superconducting quantum arithmetic circuit of the present invention includes an integrated magnetic flux qubit, a bias current line for applying a bias magnetic flux to each magnetic flux qubit, and two transformers disposed between adjacent qubits, In other words, a transformer consisting of a superconducting loop shunted by SQUID, a transformer with the same twist but a twist in the loop, a current line for SQUID control associated with each transformer, and a DC magnetic field canceling DC associated with each transformer. It has a current line for magnetic flux adjustment and a SQUID for reading the state of the qubit.

  The first effect is that an interaction between adjacent magnetic flux qubits is performed via a superconducting transformer including SQUID, thereby providing a superconducting quantum arithmetic circuit in which the interaction between qubits is variable. .

  The second effect is to provide a superconducting quantum arithmetic circuit that can completely turn off the interaction between qubits in the off state by using two superconducting transformers that cancel each other out in pairs. it can.

  The third effect is that the bias point of the qubit does not change when the interaction between qubits is controlled due to the symmetry of the SQUID in the superconducting transformer, and the operating point of the qubit does not deviate from the optimum condition. Thus, it is possible to provide a superconducting quantum operation circuit capable of keeping the coherence time of each qubit long.

  The fourth effect is that the coupling between qubits does not use the direct magnetic coupling between adjacent qubits, but uses a long superconducting transformer (for example, 200 μm), so Since the distance can be sufficiently large, it is possible to provide a superconducting quantum arithmetic circuit that can suppress crosstalk between signals for qubit control.

  The fifth effect is to provide a superconducting quantum arithmetic circuit with few restrictions on the type and arrangement of qubits to be used because the symmetry of qubits itself is not used to turn off the interaction between qubits. Can do.

  Next, embodiments of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 1 (a), a superconducting quantum arithmetic circuit is shown as a perspective view as seen from the upper layer to the lower layer as an embodiment of the present invention. Other FIGS. 1B to 1D are exploded views of the multilayer structure for easy understanding, and the layers in each figure are separated by an insulating film. In FIG. 1B, a part where the lines intersect is a multilayer wiring, which is separated by an interlayer insulating film.

  Referring to FIG. 1 (b), there is shown a transformer element disposed below the magnetic flux qubits 101 and 102 shown in FIG. 1 (c), where two superconducting transformers 105 and 106 are used. It is configured. As shown in FIG. 1A, the illustrated superconducting transformers 105 and 106 are disposed between the flux qubits 101 and 102, adjacent to each other at the left end and the right end, and to the flux qubits 101 and 102, Superconducting loops that are provided adjacent to each other on the lower layer region and extend in the length direction are formed. In addition, the superconducting transformers 105 and 106 are formed by wirings close to each other at portions other than both ends.

  Further, a shunt SQUID 107 is connected to the superconducting loop of the superconducting transformer 105. The shunt SQUID 107 includes a loop including two Josephson elements indicated by a cross. In the illustrated example, both ends of the Josephson element of the shunt SQUID 107 and the two wires of the superconducting transformer 105 are connected by two lead wires. In this case, as shown in FIG. 1B, one lead-out wiring straddles one wiring of the superconducting transformer 105 and is connected to the other wiring of the superconducting transformer 105.

  On the other hand, a shunt SQUID 108 is connected to the superconducting loop of the superconducting transformer 106. The shunt SQUID 108 is also provided with a loop including two Josephson elements indicated by a cross, and both ends of the Josephson element of the shunt SQUID 108 are connected to the two wires of the superconducting transformer 106 by two lead wires. Has been. As is apparent from the figure, one of the two lead wires is connected to the upper wire in the superconducting loop of the superconducting transformer 106, and the other is connected to the lower wire in the superconducting loop of the superconducting transformer 106. It is connected. As a result, a twisted portion 109 is formed between the lead wiring of the shunt SQUID 108 and the superconducting loop. In FIG. 1B, other superconducting transformer pairs are also arranged on the left and right sides of the superconducting transformers 105 and 106 at predetermined intervals.

  Referring now to FIG. 1 (c), two flux qubits 101 and 102 are shown disposed on top of the superconducting transformers 105,106. In this example, a read SQUID 103 is provided along with the magnetic flux qubit 101. In the illustrated example, the read SQUID 103 includes two Josephson elements and is formed so as to surround the magnetic flux qubits 101 and 102. Here, the magnetic flux qubits 101 and 102 are superconducting loops having the same structure, each having three Josephson junctions. One of the three Josephson junctions in each flux qubit 101, 102 is slightly smaller than the other two, and the energy difference between the ground state and the excited state of the qubit is designed to be about 10 GHz. ing.

  Superconducting wiring as shown in FIG. 1D is provided on the upper layer of the magnetic flux qubits 101 and 102. That is, the qubit bias current line 104 is wired so as to cover the magnetic flux qubit 101, and the DC magnetic flux adjustment current line 111 is wired so as to cover the left end region of the superconducting transformers 105 and 106. Yes. Similarly, a qubit bias current line 104 is provided corresponding to the other magnetic flux qubit 102, and a DC magnetic flux adjustment current line 111 is provided so as to cover the right end region of the superconducting transformers 105 and 106. Yes. In FIG. 1C, a DC magnetic flux adjusting current line 111 provided corresponding to the adjacent superconducting transformer pair is also shown.

  Furthermore, SQUID control current lines 110 are provided at positions corresponding to the shunt SQUIDs 107 and 108 shown in FIG.

  Although only two magnetic flux qubits 101 and 102 are shown in FIGS. 1A and 1C, they can be easily expanded to n qubits. The illustrated circuit is formed on an insulator substrate with little high-frequency loss. In this case, examples of the substrate material include sapphire, magnesium oxide, silicon with a thermal oxide film, and the like.

  As described above, the wirings such as the qubit bias current line 104, the SQUID control current line 110, and the DC magnetic flux adjustment current line 111 are all superconducting wirings, and no heat is generated by passing a current. Here, examples of the superconductive material include aluminum, niobium, and niobium nitride.

  In the illustrated superconducting quantum arithmetic circuit, the qubit bias current line 104 shown in FIG. 1 (d) controls the bias magnetic flux of each of the magnetic flux qubits 101 and 102, and the optimum bias for maintaining coherence. A wiring for biasing the point and applying a microwave current pulse for 1-bit control is provided along with each of the qubits 101 and 102.

  In addition, two transformers 105 and 106 provided between adjacent qubits 101 and 102 (ie, a transformer 105 composed of a superconducting loop shunted by SQUID 107 and a transformer 106 having a twisted portion 109 in the loop, similar to the transformer 105). When the current flowing through the SQUID control current line 110 is zero, the mutual effects are canceled out, so that the magnetic coupling between the qubits 101 and 102 becomes zero. Therefore, the interaction between the qubits 101 and 102 can be completely turned off.

  The DC magnetic flux adjusting current lines 111 provided on the left and right ends of the superconducting transformers 105 and 106 have a current adjusted so that the DC magnetic flux in the loop of each superconducting transformer 105 and 106 becomes zero. Washed away. In other words, the direct current magnetic flux in each superconducting loop is adjusted to be zero by passing a current through the direct current flux adjusting current line 111. That is, the DC magnetic flux adjusting current line 111 is for compensating for magnetic field leakage due to the bias current of the qubits 101 and 102.

  Further, by passing the SQUID control current through the SQUID control current line 110 provided in the upper layer of the shunt SQUIDs 107 and 108, the cancellation between the two superconducting transformers 105 and 106 is lost. As a result, magnetic coupling occurs between the qubits 101 and 102, and the interaction can be temporarily turned on for a necessary time. This means that the interaction between the qubits 101 and 102 can be varied. In other words, the illustrated superconducting quantum computation circuit can perform conditional computation because the qubits 101 and 102 can be magnetically coupled as necessary.

  Since the shunts SQUID107 and 108 shown in the figure have symmetry, no current is induced in the superconducting transformer loops 105 and 106 even when a SQUID control current is passed, and the bias point of the qubits 101 and 102 is optimal. There is no deviation from the point. In addition, since the superconducting transformer loops 105 and 106 are elongated so that most of the opposing wirings are close to each other, the self-inductance of the loop can be reduced, and the interaction between qubits when turned on can be sufficiently increased. On the other hand, by ensuring the distance between the qubits 101 and 102, it is possible to prevent signal cross-coupling during control and readout.

  Next, the operation of the superconducting quantum arithmetic circuit shown in FIG. 1 will be described.

  Each of the integrated magnetic flux qubits 101 and 102 is biased to an optimum bias point for maintaining coherence by a current flowing through the associated qubit bias current line 104.

  Further, the control of the 1-bit state is executed by flowing a microwave current pulse that resonates with the magnetic flux qubits 101 and 102 through the qubit bias current line 104. Since the two superconducting transformers 105 and 106 arranged between the two magnetic flux qubits 101 and 102 cancel each other's effect when the current flowing through the SQUID control current line 110 is 0, the qubit The magnetic coupling between them becomes zero. Therefore, the interaction between qubits is completely off. The direct current magnetic flux in the superconducting transformer loops 105 and 106 is adjusted by passing a current through the direct current magnetic flux adjusting current line 110 so that the direct current magnetic flux becomes zero. By passing the current for controlling the SQUID, the cancellation between the two superconducting transformers 105 and 106 is lost, magnetic coupling occurs between the qubits 101 and 102, and the interaction is temporarily turned on for a necessary time. be able to.

  Thus, in the present invention, a variable transformer is provided between adjacent qubits 101 and 102, and a quantum operation circuit capable of controlling the interaction between magnetic flux qubits via the variable transformer is realized. Reading of the state after completion of the quantum operation can be performed using a reading SQUID associated with each qubit.

  As an application example of the present invention, there is a quantum arithmetic circuit used in a quantum computer or a quantum repeater.

(A) is a figure shown in the state which saw through the superconducting quantum arithmetic circuit concerning one embodiment of the present invention for convenience of explanation, and (b), (c), and (d) are the figures shown in Drawing 1 (a). It is a figure which decomposes | disassembles and shows. It is a figure which illustrates conceptually the magnetic flux qubit used in FIG. It is a figure explaining the coupling | bonding method of the magnetic flux qubit seen by the nonpatent literature 2. FIG. It is a figure explaining the coupling | bonding method of the magnetic flux qubit seen by the nonpatent literature 3. FIG. It is a figure explaining the coupling | bonding method of the magnetic flux qubit seen in the nonpatent literature 1. FIG. It is a figure explaining the coupling | bonding method of the magnetic flux qubit seen in the nonpatent literature 4. It is a figure explaining the coupling | bonding method of the magnetic flux qubit seen in the nonpatent literature 5. It is a figure explaining the coupling | bonding method of the magnetic flux qubit seen by the nonpatent literature 6. FIG.

Explanation of symbols

101: Magnetic flux qubit 1
102: Magnetic flux qubit 2
103: SQUID for reading
104: qubit bias current line
105: Superconducting transformer
106: Superconducting transformer
107: Shunt SQUID
108: Shunt SQUID
109: Twist of superconducting loop
110: Current line for SQUID control
111: DC magnetic flux adjustment current line
201: Superconducting loop
202: Josephson junction
301: Quantum bit 1
302: Quantum bit 2
401: Quantum bit 1
402: Quantum bit 2
403: Josephson junction for bonding
501: Quantum bit 1
502: Quantum bit 2
503: Superconducting transformer
504: SQUID
601: Quantum bit 1
602: Quantum bit 2
603: Superconducting transformer
604: SQUID
701: Quantum bit 1
702: qubit 2
703: Superconducting transformer
704: SQUID
801: Quantum bit 1
802: Quantum bit 2
803: SQUID for bonding

Claims (11)

  A plurality of magnetic flux qubits having the same structure as each other, a plurality of transformers arranged between adjacent magnetic flux qubits, each forming a superconducting loop, and a bias current for applying a bias magnetic flux to each magnetic flux qubit A superconducting quantum arithmetic circuit comprising: a line; and a DC magnetic flux adjusting current line for DC magnetic field cancellation associated with each transformer.   2. The plurality of transformers arranged between adjacent flux qubits according to claim 1, wherein the first transformer includes a superconducting loop shunted with a shunt SQUID, and the second transformer includes a shunt SQUID and a twist portion. And a SQUID control current line is provided corresponding to the SQUIDs of the first and second transformers.   3. The superconducting quantum arithmetic circuit according to claim 2, further comprising a read SQUID for reading a qubit adjacent to the magnetic flux qubit.   4. The first layer in which the plurality of transformers and the shunt SQUID are formed, the second layer in which the magnetic flux qubit and the read SQUID are formed, the bias current line, and the SQUID control current line. And a third layer provided with a DC magnetic flux adjusting current line, and an interlayer insulating film is provided between the first layer and the third layer. Quantum arithmetic circuit.   A quantum computer comprising the superconducting quantum arithmetic circuit according to claim 1.   A semiconductor device comprising the superconducting quantum arithmetic circuit according to claim 1.   A transformer forming step of forming two superconducting transformers adjacent to each other, each of which comprises two ends and constituting a superconducting loop extending in the longitudinal direction between the two ends; and two of the two superconducting transformers A method of manufacturing a superconducting quantum arithmetic circuit comprising: forming a magnetic flux qubit adjacent to one end; and providing a superconducting wiring necessary for the superconducting transformer and the magnetic flux qubit .   8. The transformer forming step according to claim 7, wherein the transformer forming step includes forming a shunt SQUID in the middle portion in the longitudinal direction of the two superconducting transformers, and forming a twist portion in one of the two superconducting transformers. A method of manufacturing a superconducting quantum arithmetic circuit, comprising the step of forming the superconducting loop connected to one of the shunt SQUIDs via the twisted portion. In a control method for controlling the state of two flux qubits that are arranged in a predetermined direction at intervals and have the same structure, a superconducting loop having the predetermined direction as a length direction between the two flux qubits The two superconducting transformers that form the
Biasing the two flux qubits to an optimal bias and turning off the interaction between the two flux qubits due to the canceling action of the two superconducting transformers,
A control method characterized in that the canceling action between the two superconducting transformers is canceled and the two magnetic flux qubits are temporarily magnetically coupled to perform a quantum operation.   10. The control method according to claim 9, wherein a read SQUID is provided in association with one of the two magnetic flux qubits, and a state obtained by the temporary coupling is read. In a transformer element used by being arranged between a plurality of magnetic flux qubits, a first transformer forming a superconducting loop extending in a longitudinal direction including a shunt SQUID, and a transformer adjacent to the first transformer are arranged. And a second transformer that includes a shunt SQUID and a twisted portion, and forms a superconducting loop extending in the length direction in the same manner as the first truss transformer.

Images