Introduction

Quantum computation1, which can implement the famous Shor’s algorithm2 for integer factorization and Grover/Long algorithm3,4 for unsorted database search, has attracted much attention in recent years. There are some interesting systems which have been used to realize quantum computation, such as photons5,6, quantum dots7,8, nuclear magnetic resonance9,10,11, diamond nitrogen-vacancy center12,13 and cavity quantum electrodynamics (QED)1. Achieving quantum computation, quantum state transfer14,15 and universal quantum gates have been studied a lot, especially the two-qubit controlled-phase (c-phase) gate or its equivalent (controlled-not gate) which can be used to construct a universal quantum computation assisted by single-qubit operations1. To construct the high-efficiency and high-fidelity quantum state transfer and the c-phase gate on fields or atoms, cavity QED, composed of a two-energy-level atom coupled to a single-mode filed, has been studied a lot.

Simulating cavity QED, circuit QED16,17,18,19,20,21,22,23,24,25,26,27, composed of a superconducting qubit coupled to a superconducting resonator, plays an important role in quantum computation because of its good ability for the large-scale integration28,29,30,31,32,33. By far, some important tasks of quantum computation based on the superconducting qubits have been realized in experiments. For example, DiCarlo et al. demonstrated a c-phase gate on two transmon qubits34 in 2009 and they prepared and measured the entanglement on three qubits in a superconducting circuit35 in 2010. In 2012, Lucero et al.36 computed the prime factors with a Josephson phase qubit quantum processor and Reed et al.37 constructed a controlled-controlled phase gate to realize a three-qubit quantum error correction with superconducting circuits. In 2014, Barends et al.38 realized the single-qubit gate and the c-phase gate on adjacent Xmon qubits with high fidelities of 99.94% and 99.4%, respectively.

Interestingly, the quality factor of a one-dimensional (1D) superconducting resonator39 has been enhanced to 106, which makes the resonator as a good carrier for quantum information processing40,41,42,43,44,45,46,47,48,49,50,51,52. For instance, Houck et al.53 generated single microwave photons in a circuit in 2007. In 2008, Hofheinz et al.54 generated the Fock states in a superconducting quantum circuit. In 2010, Johnson et al.55 realized the quantum non-demolition detection of single microwave photons in a resonator. In 2011, Wang et al.56 deterministically generated the entanglement of photons in two superconducting microwave resonators and Strauch et al.57 proposed a scheme to prepare the NOON state on two resonators. In 2013, Yang et al.58 presented two schemes for generating the entanglement between microwave photons and qubits. Recently, Hua et al.59 proposed some schemes to construct the universal c-phase and cc-phase gates on resonators.

There have been some theoretic studies on constructing the multi-resonator quantum entanglement and the universal quantum gate on local microwave-photon resonators in a processor composed of some resonators coupled to a superconducting qubit57,58,59,60,61. In this paper, we propose a quantum processor for quantum computation on distant resonators with the tunable coupling engineering62,63 between the superconducting resonator and the quantum bus. There is just one superconducting transmon qutrit q in our processor, which is coupled to a common resonator R (acts as a quantum bus). Different from the processors in previous works57,58,59,60,61, the resonators rj (j = 1, 2) (act as the information carriers) in our processor are coupled to the quantum bus R, not the qutrit, which makes it have the capability of integrating some distant resonators35 by coupling them to the bus in different positions. In contrast with the resonator-zero-qubit architecture by Galiautdinov et al.31, the resonators in our processor are used for quantum information processing, not the memory elements. It does not require more superconducting qubits. With our processor, we present an effective scheme for the quantum state transfer between r1 and r2 with the Fock states and and another for the c-phase gate on two resonators by using the resonance operations between R and rj and that between R and q. The fidelities of our quantum state transfer and c-phase gate are 99.97% and 99.66%, respectively. By catching and releasing the microwave photons from resonators64, our processor maybe play an important role in quantum communication.

Results

Quantum processor composed of resonators and a quantum bus

Our quantum processor is composed of some distant high-quality 1D superconducting resonators rj and a high-quality 1D superconducting resonator R, shown in Fig. 1. The common resonator R acts as a quantum bus for quantum information processing and it is capacitively coupled to a Ξ type three-energy-level superconducting transmon qutrit q whose frequency can be tuned by an external magnetic field. The qutrit is placed at the maximum of the voltage standing wave of R (not be drawn here). The simple superconducting quantum interferometer device (SQUID) with two Josephson junctions inserted between rj and R serves as the tunable-coupling function between them. The SQUID variables are not independent and introduce no new modes63. Here, the SQUIDs are not sensitive to the charge noise and can achieve a full tunability. Besides, the plasma frequencies of SQUIDs should be larger than the frequencies of the resonators. rj are laid far enough to each other to avoid their direct interaction generated by mutual capacitances and mutual inductive coupling. In the interaction picture, the Hamiltonian of the processor is (ħ = 1, under the rotating-wave approximation)

Figure 1
figure 1

Schematic diagram for the construction of the quantum state transfer between the two microwave-photon resonators rj (j = 1, 2) and the c-phase gate on rj assisted by a quantum bus (i.e., the common resonator R) which is coupled to only a superconducting transmon qutrit q.

Here, and Δj = ωj − ωR. ωR and ωj are the the first mode frequencies of R and rj, respectively. ωg,e(e,f) is the frequency of the transmon qutrit q with the transition in which , and are the ground, the first excited and the second excited states of the qutrit, respectively. and are the creation operators of R and rj, respectively. and are the creation operators of the two transitions of q, respectively. gg,e and ge,f are the coupling strengths between R and the two transitions of q, respectively. gj is the coupling strength between rj and R, which is contributed by their capacitive and inductive and can be tuned by the external flux through the SQUID63. By controlling the time dependence of the coupling, the cross-talk between resonators can be switched on and off.

The evolution of our processor can be described by the master equation65

Here, the operator D[L]ρ = (2LρL+ − L+ − ρL+L)/2 (L = a, b, , ). and . κ1, κ2 and κR are the decay rates of the resonators r1, r2 and R, respectively. γg,ee,f) is the energy relaxation rate of the qutrit with the transition . γϕ,eϕ,f) is the dephasing rate of the level of the qutrit. To achieve the resonance operations between R and rj, the transition frequencies of all the resonators are taken equal to each other.

Quantum state transfer between r1 and r2

Our quantum-state-transfer protocol between r1 and r2 can be completed with two resonance operations between the quantum bus R and the resonator rj. The interaction between R and rj can be described as

In our scheme, the states , and are required only. Here, the state keeps unchanged with the evolution . and are the Fock states of R and rj, respectively. and . For the resonance condition between R and rjj = 0) and if we take the initial state of the subsystem composed of R and rj to be , the state of the system composed of R and rj can be expressed as (further details can be found in the method)

Our scheme for the quantum state transfer between the two resonators r1 and r2 can be accomplished with two-step resonance operations described in detail as follows.

Initially, we assume the initial state of the processor is

which means r1 is in the state , R and r2 are all in the vacuum state and q is in the ground state. First, tuning the transition frequency of q to detune with R largely and turning off (on) the coupling strength between R and r2 (r1) by using the external flux through their SQUIDs, the state of the processor can evolve into

after a time of g1t = π/2.

Second, keeping the frequency of q detune with R largely, turning off g1 and turning on g2, the state of the processor can evolve from Eq. (6) to

within a time of g2t = π/2. Here, we complete the quantum state transfer as

If the operation time of the second step is taken as g2t = 3π/2, the final state after the information transfer is

This is just the result of the quantum state transfer between the two resonators r1 and r2 from the initial state .

Controlled-phase gate on r1 and r2

C-phase gate is an important universal two-qubit gate. In the basis of two resonators and , a matrix of the gate can be expressed as

which means a minus phase should be generated if and only if the two qubits are in the state . In our processor, the c-phase gate on the resonators r1 and r2 can be completed with five steps by combining the resonance operations between the quantum bus R and the resonator rj and those between R and q with the two transitions and

By taking the coupling strength between q and R much smaller than the anharmonicity of q (gg,e ωg,e − ωe,f), the interactions between R and q with the two transitions of and can be reduced into those of two individual two-energy-level qubits with R, whose Hamiltonians are

and

respectively. In the condition of resonance interactions between R and q with the transitions g,e = 0) and e,f = 0), the time-evolution operation of the system undergoing the Hamiltonians and are66

and

respectively.

Supposing the initial state of the processor is

Here, the amplitudes α1 = cosθ1 cosθ2, α2 = cosθ1 sinθ2, α3 = sinθ1 cosθ2 and α4 = sinθ1 sinθ2. The five steps for the construction of our c-phase gate on r1 and r2 can be described in detail as follows.

First, turning on the coupling strength between R and r1 with g1 = gg,e and turning off the interaction between R and r2, the state of the processor can evolve from to

with an operation time of 67.

Second, tuning the frequency of q to detune with R largely and turning off the coupling between R and r1, one can get the state of the processor as

after the time of g2t = π/2 when the coupling between R and r2 is turned on.

Third, resonating R and q with the transition of with a time of ge,ft = π and keeping R uncoupled to r1 and r2, the state of the the processor becomes

Fourth, repeating the second step, one can get the state of the processor as

Fifth, repeating the first step, we can get the state

This is just the result of our c-phase gate on r1 and r2 with the initial state .

Possible experimental implementation

Resonance operation between a superconducting qubit and a 1D superconducting resonator has been used to achieve some basic tasks in quantum information processing, such as generating Fock states in a superconducting quantum circuit54, realizing the NOON state entanglement on two superconducting microwave resonators56, constructing the resonant quantum gates on charge qubits in circuit QED68 or on resonators59 and completing a fast scheme to generate NOON state entanglement on two resonators69. To get a high-fidelity resonant operation between the qubit and the resonator, the magnetic flux with fast tunability is required.

To show the performance of our schemes for quantum state transfer and the c-phase gate, we simulate their fidelities by using the whole Hamiltonian in each step. In our simulations, the parameters are chosen as: g1/(2π) and g2/(2π) can be tuned from 0 MHz to 50 MHz, ωR/(2π) = 6.65 GHz63, δ = ωg,e/(2π) − ωe,f/(2π) = 0.72 GHz70, gg,e/(2π) = MHz,  μs and  μs. The transition frequency of a transmon qutrit can be tune with a range of about 2.5 GHz71, which is enough for us to make it detune with R largely. The maximal values of g1/(2π) and g2/(2π) taken by us are 50 MHz as the rotation-wave approximation can work well when the coupling strength is much smaller than the frequency of R and a theoretic predict of the coupling strength between two superconducting resonators can reach 1.2 GHz63.

The process for the generation of the initial states of and are not included in our simulations. To prepare the initial states, one should perform a proper single-qubit operation on q and send the information from q to rj by using the resonance operation, the same as the one in the first step for the construction of our c-phase gate. Here, the interactions which do not attend the resonance operation should be turned off. The single-qubit operation on a superconducting qubit has been realized in experiment with a quantum error smaller than 0.000638, which has little influence on our schemes. By taking the energy relaxation rate of the qutrit, the decay rates of resonators and gg,e and gj into account, the generation of the initial states just increases a little error value of the fidelities of the quantum state transfer and the c-phase gate.

Fidelity for our quantum state transfer

We numerically simulate the populations (vary with the operation time) of a microwave photon in r1, R and r2, shown in Fig. 2. The definition of the population is

Figure 2
figure 2

The populations of a microwave photon in r1, R and r2. P1, P2 and P3 with the red, green and blue solid lines represent the populations of the microwave photon in r1, R and r2, respectively.

The inset shows the populations varying with the decay rates of the resonators, in which the solid, the dot dash and the dotted lines represent those with the decay rates of the resonators κ−1 = ∞ μs, κ−1 = 50 μs and κ−1 = 10 μs, respectively.

Here m = 1, 2, 3. , and . ρ(t) is the realistic density operator of the processor for the quantum state transfer from the initial state . The parameters taken in the first step in our scheme are: ωg,e/(2π) = 5 GHz, g1/(2π) = 50 MHz, g2/(2π) = 0 MHz. In the second step, the parameters are: g1/(2π) = 0 MHz, g2/(2π) = 50 MHz and the other parameters are the same as the ones in the first step.

From the numerical simulation, the quantum state transfer between r1 and r2 with θ = π/4 can reach a fidelity of 99.97% within 10 ns by using the definition of the fidelity as with the initial state . In the inset in Fig. 2, we give the three conditions of the populations with different decay rates of r1, r2 and R.

Fidelity for our c-phase gate

We calculate the fidelity of our c-phase gate by using the average-gate-fidelity definition72,73

Here, is the final state of the processor by using the ideal c-phase gate operation on the initial state . ρ(t) is the realistic density operator after our c-phase gate operation on the initial state with the Hamiltonian H. The fidelity of our c-phase gate reaches 99.66% within 91.5 ns by using the parameters taken in each step as shown in Table 1. Here, if we take θ1 = θ2 = π/4 in Eq. (15) as an example, the density operators of and the real final state are shown in Fig. 3(a,b), respectively.

Table 1 Parameters for the construction of the c-phase gate on r1 and r2.
Figure 3
figure 3

(a) The density operator ρ0 of the initial state of our processor. (b) The realistic density operator of the final state after our c-phase gate operation is performed on the two microwave-photon resonators. The color bar indicates the phase information of the density matrix elements.

Actually, the fidelity of our c-phase gate is influenced by the decay rates κ of the resonators, the energy relaxation rate Γ of q and the anharmonicity δ of q, shown in Fig. 4. In Fig. 4(a), we show the fidelity of the gate varying with the decay rates and the energy relaxation rate of the resonators and q (κ = Γ). The fidelity of the gate is numerically simulated by using different optimal parameters corresponding to different Γ (keeping δ = 0.72 GHz unchanged) as the competition between the operation time (leads to the error from the coherence time of the qutrit) and the coupling strength between the qutrit and the bus R (leads to the error from the anharmonicity of the qutrit). Here, in order to choose Γ−1 = 10, 20, 30, 40 and 50 μs, we take , 19, 13, 13 and 13 MHz, respectively. The corresponding operation times are t = 58.1, 65.8, 91.5, 91.5 and 91.5 ns, respectively. By using κ = ωr/Qr is the frequency of the resonator)16, κ−1 = 50 μs corresponds to a quality factor Q ~ 2.08 × 106 of the resonators. In Fig. 4(b), the anharmonicity of the qutrit influences the fidelity with a small value as the coupling strength gg,e is much smaller than δ, which means that the transmon qutrit in our processor does not require a large anharmonicity.

Figure 4
figure 4

The fidelity of our c-phase gate on the two microwave-photon resonators r1 and r2 which varies with the parameters κ−1 = Γ−1 (a) and δ (b), respectively.

Conclusion

To show our processor can be used for an effective quantum computation based on resonators, we have given the scheme to achieve the quantum state transfer between two resonators and the one for the c-phase gate on them. These two schemes are just based on the Fock states and of the resonators rj. The fidelities of our quantum state transfer and c-phase gate reach 99.97% and 99.66% within 10 ns and 91.5 ns, respectively. In our processor, a single-qubit operation on the resonator rj can be achieved with the following steps: 1), one should transfer the information from rj to the qutrit with the resonance operation between them. 2), one can take the single-qubit gate on the qutrit. 3), one should transfer the information from the qutrit to rj. It is worth noticing that there are two steps with resonance operations in our scheme for the single-qubit operation on a microwave-photon resonator. Each resonance operation can generate a −phase for the state of the resonator rj or the state of the qutrit. The two steps with resonance operations can just eliminate this unwanted phase generated by each resonance operation as (−1)2 = 1. So, the single-qubit operation on the qutrit is convenient without considering the additional phase generated by the resonance operations. To readout the information of the photon states in rj, one can also transfer the information of the photon from rj (based on the Fock states and ) to the qutrit (based on the states and and then readout the state of the qutrit. To achieve the quantum non-demolition detection on the resonator rj, one can use a low-quality resonator coupled to the qutrit q to detect the information in the quantum bus R55 which comes from rj. By using the resonators which can catch and release the microwave photons64, our processor maybe play an important role in quantum communication.

In summary, we have proposed a quantum processor composed of some 1D superconducting resonators rj (quantum information carriers) which are coupled to a common 1D superconducting resonator R (the quantum bus), not the superconducting transmon qutrit, which makes it have the capability of integrating some distant resonators for quantum information processing on microwave photons assisted by circuit QED. With this processor, we have presented a scheme for the high-fidelity state transfer between two resonators. Also, we have given a scheme for the c-phase gate on two resonators with the resonance operations. With feasible parameters in experiment, the fidelities of our two schemes are 99.97% and 99.66%, respectively. Maybe this processor can play an important role in quantum communication in future.

Methods

Interaction between a resonator and a qubit

In the interaction picture, the Hamiltonian of a system composed of a two-energy-level qubit coupled to a resonator (Q-R system) can be written as (under the rotating-wave approximation):

Here, g is the coupling strength between the qubit and the resonator. and a+ are the create operators of the qubit and the resonator, respectively. Δ = ωq − ωr. ωqr) is the transition frequency of the qubit (resonator).

The state of the Q-R system can be solved with the equation of motion

in which is a linear combination of the states and , that is,

Here, ce,n(t) and cg,n(t) are the slowly varying probability amplitudes. is the Fock state of the resonator. Because the only transitions between and can be caused by the Hamiltonian HI, we just need to consider the evolutions of ce,n(t) and cg,n+1(t).

By combining Eqs. (24) and (25), one can get

A general solution for these amplitudes is

Here Ω2 = 4g2(n + 1) + Δ2.

Additional Information

How to cite this article: Hua, M. et al. Quantum state transfer and controlled-phase gate on one-dimensional superconducting resonators assisted by a quantum bus. Sci. Rep. 6, 22037; doi: 10.1038/srep22037 (2016).