## Abstract

Parity-time (*𝒫𝒯*) symmetric structures have exhibited potential applications in developing various robust quantum devices. In an optical trimer with balanced loss and gain, we analytically study the *𝒫𝒯* symmetric phase transition by investigating the spontaneous symmetric breaking. We also illustrate the asymmetric photonic transmission behaviors in both of the *𝒫𝒯* symmetric and *𝒫𝒯* symmetry broken phases. We find (i) the non-periodical dynamics of photonic transmission in the *𝒫𝒯* symmetry broken phase instead of *𝒫𝒯* symmetric phase can be regarded as a signature of phase transition; and (ii) it shows asymmetric photonic transmission behavior in both of the phases but comes from different underlying physical mechanisms. The obtained results may be useful to implement the photonic devices based on coupled-cavity system.

© 2017 Optical Society of America

## 1. Introduction

Since Bender and Boettcher proposed the concept of parity-time (*𝒫𝒯*) symmetry [1, 2], it has attracted a lot of attentions due to its potential applications. Remarkably, the system with *𝒫𝒯* symmetry can undergo a phase transition when the parameter that controls the non-Hermiticity surpasses a critical value which is usually called exceptional point (EP) [2,3]. Below the EP, all of the eigen-values of the non-Hermit Hamiltonian are real, and some or all of the eigen values become complex beyond the EP.

The broken of *𝒫𝒯* symmetry will lead to a lot of interesting phenomena. For example, people have observed the non-reciprocal photonic transmission in the *𝒫𝒯* symmetric structure [4–12], and predicted the enhancement of nonlinear interaction due to field localization [13, 14]. Moreover, the unique property of the system with *𝒫𝒯* symmetry has potential applications in various fields, such as loss-induced or gain-induced transparency [15, 16], efficient photon or phonon lasing [17–21], ultralow-threshold optical chaos [22,23] as well as quantum metrology [24].

On the other hand, the coupled-cavity system is widely used to coherently control the photon transfer. In the coupled-cavity-array with infinite length, the defect can be introduced to construct a singe-photon switcher [25, 26], router [27] and frequency converter [28]. Moreover, within the capacity of current experiments, a lot of attentions have been paid on the optical dimer (also called optical molecule [29, 30], which is composed of two coupled cavities), such as the coherent polariton [31] and state transfer [32]. Furthermore, motivated by the simulation of photosynthesis harvest system in recent years, there are also studies about the photon [33] and thermal transport [34] in the trimer structure.

In this paper, we focus on an optical trimer with *𝒫𝒯* symmetry. Here, our scheme is composed of an active gain cavity and a passive loss cavity, which simultaneously couple with a third cavity without loss and gain, to form a coupled-cavity-array as shown in Fig. 1. With balanced gain and loss, which are described phenomenally in this paper, we show a *𝒫𝒯* symmetric phase transition in the non-Hermitian system. Different from previous work about the *𝒫𝒯* symmetry in optical trimer structure [35–38], we here focus on the property of the eigen-state of the non-Hermite Hamiltonian, and find that the field localization and spontaneous symmetry broken occurs along with the *𝒫𝒯* symmetry phase transition, in which a smaller critical coupling strength is needed than that in the dimer system [4,5,13,14].

To wit the *𝒫𝒯* phase transition, we study the dynamical evolution of the system in both of the *𝒫𝒯* symmetric phase (where all the eigen values of the Hamiltonian are real) and *𝒫𝒯* symmetry broken phase (where the complex eigen values emerge). Under the semi-classical approximation, we neglect the effect of quantum noise which contributes to the photonic correlation [39], and study the dynamical behavior of the system by solving the Schödinger equation analytically. As a result, we show the asymmetric photonic transmission in both of the phases. Actually, the authors have studied the asymmetric photonic transmission in an optical dimer [35–38], where some nonlinearities are evolved. On contrary, we here consider only the linear interaction, and give an intuitive explanation based on the symmetry of the system.

The rest of the paper is organized as follows. In Sec. 2, we present a *𝒫𝒯* symmetry model by an optical trimer with balanced loss and gain and discuss the *𝒫𝒯* symmetric phase transition. In Sec. 3, we illustrate the dynamical evolution of the system and show the asymmetric photonic transmission in both of the *𝒫𝒯* symmetric phase and *𝒫𝒯* symmetry broken phase. At last, we give a brief remark and conclusion in Sec. 4.

## 2. Model and *𝒫𝒯* phase transition

As shown in Fig. 1, our model consists of an array of three single-mode cavities [33], where a passive and an active cavity (labelled by “−1” and “1” respectively) simultaneously couple to the third cavity (labelled by “0”) without loss and gain. By describing the gain and loss in our scheme phenomenologically, the Hamiltonian is written as

*a*(

_{l}*l*= −1, 0, 1) is the annihilation operator for the

*l*th cavity with resonant frequency

*ω*.

_{l}*J*is the photon-tunneling strength between the two nearest cavities, which can be adjusted by changing the distance of them. In addition, we use

*γ*

_{−1}(> 0) to denote the decay of the passive cavity and

*γ*

_{1}(> 0) to denote the gain of active cavity. Hereafter, we consider the case that

*ω*

_{−1}=

*ω*

_{0}=

*ω*

_{1}=

*ω*and

*γ*

_{−1}=

*γ*

_{1}=

*γ*, therefore the Hamiltonian satisfies a

*𝒫𝒯*symmetry, that is [

*H*,

*𝒫𝒯*] = 0. Here

*𝒫*represents the mirror reflection 1 ↔ −1 and

*𝒯*denotes the time reversal

*i*↔ −

*i*[2].

To deeply investigate the *𝒫𝒯* phase transition in our system, we write the Hamiltonian in the form of

*ℋ*−

*EI*) = 0, where

*I*is a 3 × 3 identity matric, we will obtain the eigenvalues and the corresponding eigenstates of the Hamiltonian

*ℋ*, yielding

*𝒩*

_{±}are the normalized constants and we have defined

We note that the eigenvalue *E*_{0} is always real and is independent of *J* and *γ*, and the eigen-state satisfies the *𝒫𝒯* symmetry, that is *𝒫𝒯* |*E*_{0}〉 = *e ^{iπ}* |

*E*

_{0}〉. However, the other paired eigenvalues

*E*

_{±}are dependent not only on

*ω*, but also on

*γ*and

*J*.

To obtain a purely real spectrum, we need a strong inter-cavity coupling strength, i.e., $J>\gamma /\sqrt{2}$. It then satisfies *𝒩*_{+} = *𝒩*_{−} = 2 and the eigen-state |*E*_{±}〉 can be simplified as

*E*

_{±}〉 are transformation invariant under the

*𝒫𝒯*operation (except for a global phase). On the other hand, for the situation of $J<\gamma /\sqrt{2}$, the imaginary parts of

*E*

_{±}emerge and ${E}_{\pm}=\omega \pm i\sqrt{{\gamma}^{2}-2{J}^{2}}$. Meanwhile, we can not find a global phase

*ϕ*to satisfy

*𝒫𝒯*|

*E*

_{±}〉 =

*e*

^{±iϕ}|

*E*

_{±}〉. In other words, when the system undergoes a spontaneous symmetry breaking as the inter-cavity coupling crosses the EP $J=\gamma /\sqrt{2}$. In this sense, we name the phase in the regime $J>\gamma /\sqrt{2}$ as the

*𝒫𝒯*symmetric phase and that for $J<\gamma /\sqrt{2}$ as the

*𝒫𝒯*symmetry broken phase.

In Figs. 2(a) and 2(b), we give the real and imaginary parts of *E*_{±} respectively. In the *𝒫𝒯* symmetry broken phase ($J<\gamma /\sqrt{2}$), the small coupling strength protects the gained energy flowing from the active cavity to the passive one, and the long lifetime supermode *E*_{+} is localized at the active cavity as shown in Fig. 2(c), where we plot |*a*_{+}| as a function of the coupling strength *J*. On contrary, in the *𝒫𝒯* symmetric phase ($J>\gamma /\sqrt{2}$), the gained energy is transferred to the passive cavity quickly, and the photon yields an equal weight distribution in the passive and active cavities, that is |*a*_{±}| ≡ 1.

We point out that here, the similar *𝒫𝒯* symmetric phase transition also occurs in optical dimer which consists of two cavities with balanced loss and gain (see Refs. [4, 5] and the references therein). The differences stem in the following two aspects: On one hand, in our optical trimer system, there exist a single real energy level *E*_{0}, and the corresponding wave function is always invariant under the *𝒫𝒯* operation, independent of whether the phase transition occurs. On the other hand, the EP of optical trimer system is $J=\gamma /\sqrt{2}$ instead of *J* = *γ* in optical dimer [4,5,13,14], that is, a smaller coupling strength is needed.

## 3. Dynamical behavior

As a witness of the *𝒫𝒯* symmetry phase transition, we here consider the dynamical evolution of the system in both of the *𝒫𝒯* symmetric phase and *𝒫𝒯* symmetry broken phase. For the sake of simplicity and to grasp the major physics, we restrict ourselves in the subspace spanned by the basis {|1; 0; 0〉, |0; 1; 0〉, |0; 0; 1〉}. Here, |*m*; *n*; *q*〉 := |*m*〉_{−1} ⊗ |*n*〉_{0} ⊗ |*q*〉_{1}, and |*n*〉* _{i}* (

*i*= −1, 0, 1) represents that the

*i*th cavity is in the Fock state |

*n*〉. Then, the wave function at arbitrary time

*t*can be assumed as

Under the semi-classical approximation [40], we will focus on the photonic distribution in different cavities (that is |*α*(*t*)|^{2}, |*β*(*t*)|^{2} and |*ξ*(*t*)|^{2}). Therefore, we will neglect the quantum noise from the environment, which makes significant contributions to the photonic correlation, and therefore beyond our consideration in this paper. In this sense, the dynamics of the system is governed by the Schödinger equation *i∂ _{t}* |

*ψ*〉 =

*H*|

*ψ*〉, where the Hamiltonian

*H*is given in Eq. (1), and it yields

#### 3.1. Dynamics in *𝒫𝒯* symmetric phase

We now study the dynamical behavior of the system in its *𝒫𝒯* symmetric phase, that is $J>\gamma /\sqrt{2}$. Firstly, we consider the situation that the photon is excited in the passive cavity initially, that is *α*(0) = 1, *β*(0) = *ξ*(0) = 0, the probability amplitudes for finding the photon in the three cavities can be obtained explicitly as

*ϕ*

_{1}= arctan[Δ

*γ*/(

*J*

^{2}−

*γ*

^{2})],

*ϕ*

_{2}= 2 arctan(Δ/

*γ*). Secondly, when the photon is initially excited in the active cavity, that is

*α*(0) =

*β*(0) = 0,

*ξ*(0) = 1, the solution of Eqs. (3) are obtained as

In Figs. 3(a) and 3(b), we plot the corresponding probabilities |*h _{s}* (

*t*)|

^{2}and |

*h′*(

_{s}*t*)|

^{2}(

*h*=

*α*,

*β*,

*ξ*) as functions of evolution time

*t*. As shown in the figure, when the system is in the

*𝒫𝒯*symmetric phase, the dynamics shows regular periodical oscillations. If the photon is initially excited in the passive cavity, i.e.,

*α*(0) = 1, the decay makes |

*α*(

_{s}*t*)|

^{2}directly decrease to zero and then the revival occurs. However, if it is excited in the active cavity, i.e.,

*ξ*(0) = 1, with the assistance of the gain effect, |

*ξ′*(

_{s}*t*)|

^{2}firstly reaches its maximal value [2

*J*

^{2}/(2

*J*

^{2}−

*γ*

^{2})]

^{2}, which is obviously larger than 1 and then oscillates between the maximal value and zero. As for the central cavity, we can also observe that

*β*(

_{s}*t*) ≠

*β′*(

_{s}*t*). In this sense, our system exhibits an asymmetric photon transmission behavior even in the

*𝒫𝒯*symmetric phase. Furthermore, we find that the initial phases

*ϕ*

_{1}and

*ϕ*

_{2}are

*γ*dependent and

*ϕ*

_{1}(−

*γ*) = −

*ϕ*

_{1}(

*γ*),

*ϕ*

_{2}(−

*γ*) = −

*ϕ*

_{2}(

*γ*). As a result, by regarding the amplitudes as functions of

*t*and

*γ*, we will reach the relationship

*α*(

_{s}*t*, −

*γ*) =

*ξ′*(

_{s}*t*,

*γ*),

*β*(

_{s}*t*, −

*γ*) =

*β′*(

_{s}*t*,

*γ*) and

*ξ*(

_{s}*t*, −

*γ*) =

*α′*(

_{s}*t*,

*γ*). Meanwhile, it is obvious from Eq. (3) that

*ℋ*(

*i*↔ −

*i*) =

*ℋ*(

*γ*↔ −

*γ*), therefore, the asymmetric photonic transmission in the

*𝒫𝒯*symmetric phase comes from the breaking of time reversal symmetry, that is [

*𝒯*,

*ℋ*] ≠ 0. The breaking of time reversal symmetry can also be observed from the dynamical behavior of the system. For example, it is shown that |

*g*(

*t*)|

^{2}≠ |

*g*(−

*t*)|

^{2}and |

*g′*(

*t*)|

^{2}≠ |

*g*′(−

*t*)|

^{2}for

*g*=

*α*,

*β*.

#### 3.2. Dynamics in *𝒫𝒯* symmetry broken phase

In this subsection, we will continue to study the dynamics of the system in the *𝒫𝒯* symmetry broken phase, where $J<\gamma /\sqrt{2}$. On one hand, we consider that the photon is initially excited in the passive cavity, then the dynamics of system is obtained as

In Fig. 4, we depict the dynamics of the system when it is in the *𝒫𝒯* symmetry broken phase. Obviously, it shows a completely different behavior compared with the case in the *𝒫𝒯* symmetric phase. As shown in Fig. 4(a), when the photon is initially excited in the passive cavity, it will experience a loss firstly and then the gain in the active cavity will compensate the loss. As a result, the probability for finding the photon in the cavities will increase as the time elapse. As for the central cavity, the incident photon will hop to it, but the obtained photons will jump to the other two cavities due to the coherent coupling until the gained photon from the active cavity jumped back to it. Furthermore, at the time *t* = *arctanh*[*γδ*/(*γ*^{2} − *J*^{2})]/*δ*, the probability for finding the photon in the passive or the central cavities achieve their smallest values simultaneously. On the other hand, the probability for finding a photon in the active cavity will increase monotonously due to the combinational effect of the photonic hopping from the central cavity and the gain from the surrounding environment.

In Fig. 4(b), we show the results when the photon is initially excited in the active cavity. In such a situation, the gain effect will take action from the very beginning and the probabilities for finding photons in all of the cavities will undoubtedly increase as the time evolution.

Comparing the results in Figs. 4(a) and 4(b), we also observe the asymmetric photonic transmission in the *𝒫𝒯* symmetry broken phase. It can be explained from the following two aspects. Firstly, similar to that in the *𝒫𝒯* symmetric phase, the time reversal symmetry breaking of the Hamiltonian naturally results in the different transmission behaviors for the photons initially excited in the left and right sides of the system. However, the periodical triangle functions which describe the dynamics of the system in the *𝒫𝒯* symmetric phase is replaced by the monotonous hyperbola function in the broken phase. Therefore, the fixed phase difference between *α _{s}* (

*t*) and

*ξ′*(

_{s}*t*) does not hold any longer for

*α*(

_{b}*t*) and

*ξ′*(

_{b}*t*). Secondly, the asymmetric transmission also comes from the field localization in the

*𝒫𝒯*symmetry broken phase. As shown in Sec. 2, the long lifetime eigen-state |

*E*

_{+}〉 has a lager distribution weight in the 1th cavity, that is, the photon is localized in the active cavity when the system is in the

*𝒫𝒯*symmetry broken phase. As a result, for the photon excited in the active cavity, the overlap between the initial state and |

*E*

_{+}〉 is much larger than that excited in the passive cavity, and leading to a different transmission behavior.

In dealing with the dynamics of the system in this section, we have applied the semi-classical approximation, where the quantum noise is neglected. The effect of noise to the photonic correlation is deserved to be explored more deeply, especially in the *𝒫𝒯* symmetry broken phase, where the amplification induced by the active cavity becomes significant. However, it is beyond the discussions in this paper.

## 4. Remark and conclusion

In this paper, we have constructed an optical trimer system by coupling a passive and active cavities to a third one without gain and loss. Experimentally speaking, the active cavity can be realized by doping the ions inside the material of the cavity [4, 5]. On the other hand, the lossless cavity can be realized by coupling a decay cavity to an auxiliary one with a much larger decay rate, which is similar to the scheme proposed by Liu *et.al.* [31]. Following the similar approach to eliminate the mode in the auxiliary cavity, we can obtain the effective decay rate of the considered cavity, which can achieve zero by choosing the relative parameters properly. In other words, we are able to obtain an effectively lossless cavity.

In conclusion, we have studied the *𝒫𝒯* symmetric phase transition by demonstrating the spontaneous symmetry breaking in an optical trimer with balanced loss and gain. In the *𝒫𝒯* symmetric phase, all of the eigen values are real and the corresponding eigen states show a balanced distribution in the passive and active cavities. In the *𝒫𝒯* symmetry broken phase, the imaginary parts of the eigen values appear and the supermode with long lifetime is characterized by a strong field localization in the active cavity. Comparing with the optical molecule/dimer system, which was broadly studied recently, the critical coupling strength of the phase transition is much smaller in our trimer structure. As a signature of the *𝒫𝒯* symmetric phase transition, we subsequently find the dramatically different dynamical behaviors when the system is in the two phases. Our results show that the regular periodical oscillation is replaced by the non-periodical behavior as the system transfers from the *𝒫𝒯* symmetric phase to *𝒫𝒯* symmetry broken phase. Moreover, we find an asymmetric transmission phenomenon in both phases. We hope our study about the *𝒫𝒯* symmetry in optical trimer will be helpful for the designing of photonic device based on coupled-cavity system.

**Note added.** Recently, we have learned of recent related work about the *𝒫𝒯* symmetry trimer systems, which cares about the dynamical behavior just at the EP [41]

## Funding

National Natural Science Foundation of China (NSFC) (11404021 and 11504241); Jilin Province Science and Technology Development Plan (20170520132JH); Fundamental Research Funds for the Central Universities (2412016KJ015 and 2412016KJ004).

## References and links

**1. **C. M. Bender and S. Boettcher, “Real spectra in non-Hermitian Hamiltonians having PT symmetry,” Phys. Rev. Lett. **80**, 5243 (1998). [CrossRef]

**2. **C. M. Bender, “Making sense of non-Hermitian Hamiltonians,” Rep. Prog. Phys. **70**, 947 (2007). [CrossRef]

**3. **A. Mostafazadeh, “Pseudo-Hermiticity versus PT symmetry: the necessary condition for the reality of the spectrum of a non-Hermitian Hamiltonian,” J. Math. Phys. **43**, 205 (2002). [CrossRef]

**4. **B. Peng, S. K. Ödemir, F. Lei, F. Monifi, M. Gianfreda, G. L. Long, S. Fan, F. Nori, C. M. Bender, and L. Yang, “Parity-time-symmetric whispering-gallery microcavities,” Nat. Phys. **10**, 394 (2014). [CrossRef]

**5. **L. Chang, X. Jiang, S. Hua, C. Yang, J. Wen, L. Jiang, G. Li, G. Wang, and M. Xiao, “Parity-time symmetry and variable optical isolation in active-passive-coupled microresonators,” Nat. Photon. **8**, 524 (2014). [CrossRef]

**6. **J.-H. Wu, M. Artoni, and G. C. La Rocca, “Non-Hermitian degeneracies and unidirectional reflectionless atomic lattices,” Phys. Rev. Lett. **113**, 123004 (2014). [CrossRef] [PubMed]

**7. **C. Li, L. Jin, and Z. Song, “Non-Hermitian interferometer: Unidirectional amplification without distortion,” Phys. Rev. A **95**, 022125 (2017). [CrossRef]

**8. **H. Ramezani, T. Kottos, R. El-Ganainy, and D. N. Christodoulides, “Unidirectional nonlinear 𝒫𝒯-symmetric optical structures,” Phys. Rev. A **82**, 043803 (2010). [CrossRef]

**9. **Z. Lin, H. Ramezani, T. Eichelkraut, T. Kottos, H. Cao, and D. N. Christodoulides, “Unidirectional invisibility induced by 𝒫𝒯-symmetric periodic structures,” Phys. Rev. Lett. **106**, 213901 (2011). [CrossRef]

**10. **L. Feng, Y. Xu, W. S. Fegadolli, M. Lu, J. E. B. Oliveira, V. R. Almeida, Y. Chen, and A. Scherer, “Experimental demonstration of a unidirectional reflectionless parity-time metamaterial at optical frequencies,” Nat. Mater. **12**, 108 (2012). [CrossRef] [PubMed]

**11. **N. Bender, S. Factor, J. D. Bodyfelt, H. Ramezani, D. N. Christodoulides, F. M. Ellis, and T. Kottos, “Observation of asymmetric transport in structures with active nonlinearities,” Phys. Rev. Lett. **110**, 234101 (2013). [CrossRef] [PubMed]

**12. **J. D’Ambroise, P. G. Kevrekidis, and S. Lepri, “Asymmetric wave propagation through nonlinear PT-symmetric oligomers,” J. Phys. A: Math. Theor. **45**, 444012 (2012). [CrossRef]

**13. **J. Li, R. Yu, and Y. Wu, “Proposal for enhanced photon blockade in parity-time-symmetric coupled microcavities,” Phys. Rev. A **92**, 053837 (2015). [CrossRef]

**14. **J. Li, X. Zhan, C. Ding, D. Zhang, and Y. Wu, “Enhanced nonlinear optics in coupled optical microcavities with an unbroken and broken parity-time symmetry,” Phys. Rev. A **92**, 043830 (2015). [CrossRef]

**15. **A. Guo, G. J. Salamo, D. Duchesne, R. Morandotti, M. Volatier- Ravat, V. Aimez, G. A. Siviloglou, and D. N. Christodoulides, “Observation of 𝒫𝒯-symmetry breaking in complex optical potentials,” Phys. Rev. Lett. **103**, 093902 (2009). [CrossRef]

**16. **H. Jing, S. K. Ödemir, Z. Geng, J. Zhang, X.-Y. Lü, B. Peng, L. Yang, and F. Nori, “Optomechanically-induced transparency in partiy-time-symmetric microresonators,” Sci. Rep. **5**, 9663 (2015). [CrossRef]

**17. **B. Peng, S. K. Ödemir, S. Rotter, H. Yilmaz, M. Liertzer, F. Monifi, C. M. Bender, F. Nori, and L. Yang, “Loss-induced suppression and revival of lasing,” Science **346**, 328 (2014). [CrossRef] [PubMed]

**18. **L. Feng, Z. J. Wong, R.-M. Ma, Y. Wang, and X. Zhang, “Single-mode laser by parity-time symmetry breaking,” Science **346**, 972 (2014). [CrossRef] [PubMed]

**19. **H. Hodaei, M.-A. Miri, M. Heinrich, D. N. Christodoulides, and M. Khajavikhan, “Parity-time-symmetric microring lasers,” Science **346**, 975 (2014). [CrossRef] [PubMed]

**20. **H. Jing, S. K. Ödemir, X.-Y. Lü, J. Zhang, L. Yang, and F. Nori, “𝒫𝒯-symmetric phonon laser,” Phys. Rev. Lett. **113**, 053604 (2014). [CrossRef]

**21. **B. He, L. Yang, and M. Xiao, “Dynamical phonon laser in coupled active-passive microresonators,” Phys. Rev. A **94**, 031802 (2016). [CrossRef]

**22. **C. T. West, T. Kottos, and T. Prosen, “𝒫𝒯-symmetric wave chaos,” Phys. Rev. Lett. **104**, 054102 (2010). [CrossRef]

**23. **X.-Y. Lü, H. Jing, J.-Y. Ma, and Y. Wu, “𝒫𝒯-symmetry-breaking chaos in optomechanics,” Phys. Rev. Lett. **114**, 253601 (2015). [CrossRef]

**24. **Z.-P. Liu, J. Zhang, S. K. Ödemir, B. Peng, H. Jing, X.-Y. Lü, C.-W. Li, L. Yang, F. Nori, and Y.-x. Liu, “Metrology with PT-symmetric cavities: Enhanced sensitivity near the PT-phase transition,” Phys. Rev. Lett. **117**, 110802 (2016). [CrossRef] [PubMed]

**25. **L. Zhou, Z. R. Gong, Y.-x. Liu, C. P. Sun, and F. Nori, “Controllable scattering of a single photon inside a one-dimensional resonator waveguide,” Phys. Rev. Lett. **101**, 100501 (2008). [CrossRef] [PubMed]

**26. **Z. H. Wang, Y. Li, D. L. Zhou, C. P. Sun, and P. Zhang, “Single-photon scattering on a strongly dressed atom,” Phys. Rev. A **86**, 023824 (2012). [CrossRef]

**27. **L. Zhou, L.-P. Yang, Y. Li, and C. P. Sun, “Quantum routing of single photons with a cyclic three-level system,” Phys. Rev. Lett. **111**, 103604 (2013). [CrossRef]

**28. **Z. H. Wang, L. Zhou, Y. Li, and C. P. Sun, “Controllable single-photon frequency converter via a one-dimensional waveguide,” Phys. Rev. A **89**, 053813 (2014). [CrossRef]

**29. **M. Bayer, T. Gutbrod, J. P. Reithmaier, A. Forchel, T. L. Reinecke, P. A. Knipp, A. A. Dremin, and V. D. Kulakovskii, “Optical modes in photonic molecules,” Phys. Rev. Lett. **81**, 2582 (1998). [CrossRef]

**30. **Y. P. Rakovich and J. F. Donegan, “Photonic atoms and molecules,” Laser Photon. Rev. **4**, 179 (2010). [CrossRef]

**31. **Y.-C. Liu, X. Luan, H.-K. Li, Q. Gong, C. W. Wong, and Y.-F. Xiao, “Coherent polariton dynamics in coupled highly dissipative cavities,” Phys. Rev. Lett. **112**, 213602 (2014). [CrossRef]

**32. **C. D. Ogden, E. K. Irish, and M. S. Kim, “Dynamics in a coupled-cavity array,” Phys. Rev. A **78**, 063805 (2008). [CrossRef]

**33. **S. Felicetti, G. Romero, D. Rossini, R. Fazio, and E. Solano, “Photon transfer in ultrastrongly coupled three-cavity arrays,” Phys. Rev. A **89**, 013853 (2014). [CrossRef]

**34. **K. Joulain, J. Drevillon, Y. Ezzahri, and J. O.- Miranda, “Quantum thermal transistor,” Phys. Rev. Lett. **116**, 200601 (2016). [CrossRef] [PubMed]

**35. **M. Duanmu, K. Li, R. L. Horne, P. G. Kevrekidis, and N. Whitaker, “Linear and nonlinear parity-time-symmetric oligomers: a dynamical systems analysis,” Phil. Trans. R. Soc. A **371**, 20120171 (2013). [CrossRef] [PubMed]

**36. **S. V Suchkov, A. A Sukhorukov, J. Huang, S. V. Dmitriev, C. Lee, and Y. S Kivshar, “Nonlinear switching and solitons in PT-symmetric photonic systems,” Laser Photonics Rev. **10**, 177(2016). [CrossRef]

**37. **K. Li, P. G. Kevrekidis, D. J. Frantzeskakis, C. E. Ruter, and D. Kip, “Revisiting the PT -symmetric trimer: bifurcations, ghost states and associated dynamics,” J. Phys. A: Math. Theor. **46**, 375304 (2013). [CrossRef]

**38. **S. V Suchkov, F. F.- Ngaffo, A. K.- Jiotsa, A. D Tikeng, T. C Kofane, Y. S Kivshar, and A. A Sukhorukov, “Non-Hermitian trimers: PT-symmetry versus pseudo-Hermiticity,” New J. Phys. **18**, 065005 (2016). [CrossRef]

**39. **G. S. Agarwal and K. Qu, “Spontaneous generation of photons in transmission of quantum fields in PT-symmetric optical systems,” Phys. Rev. A **85**, 031802(R) (2012). [CrossRef]

**40. **Z. R. Gong, H. Ian, Lan Zhou, and C. P. Sun, “Controlling quasibound states in a one-dimensional continuum through an electromagnetically-induced-transparency mechanism,” Phys. Rev. A **78**, 053806 (2008). [CrossRef]

**41. **L. Jin, “𝒫𝒯-symmetric trimer systems,” arXiv: 1701.00316v1 (2017).