2025 Volume 34 Issue 12
Article Contents

Wei Qin, Dong-Chen Zheng, Jia-Ying Lin, Yuan-Hong Chen, Renyuan Liao. Energy mechanism of the first-order superradiant phase transition in cavity-BEC system with double asymmetric pump beams[J]. Chinese Physics B, 2025, 34(12): 120507. doi: 10.1088/1674-1056/addeb7
Citation: Wei Qin, Dong-Chen Zheng, Jia-Ying Lin, Yuan-Hong Chen, Renyuan Liao. Energy mechanism of the first-order superradiant phase transition in cavity-BEC system with double asymmetric pump beams[J]. Chinese Physics B, 2025, 34(12): 120507. doi: 10.1088/1674-1056/addeb7

Energy mechanism of the first-order superradiant phase transition in cavity-BEC system with double asymmetric pump beams

  • Received Date: 21/03/2025
    Accepted Date: 22/05/2025
    Available Online: 01/12/2025
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Figures(5)

Article Metrics

Article views(569) PDF downloads(0) Cited by(0)

Access History

Energy mechanism of the first-order superradiant phase transition in cavity-BEC system with double asymmetric pump beams

Abstract: We consider a Bose–Einstein condensate loaded inside an optical cavity and exposed to two crossed coherent pump fields with same imbalance parameter γ. We identify different effects between pure standing wave fields (γ = 1) and the pump beams combining standing wave and running wave (γ ≠ 1). In particular, for γ = 1, the system only hosts a normal phase and a superradiant phase. In contrast, for γ ≠ 1, the system features three distinctive phases: the normal phase (NP), superradiant phase 1 (SR1), and superradiant phase 2 (SR2). Importantly, the superradiance is subdivided into different types characterized by the photon phase. Furthermore, we determine perturbatively the phase boundary separating the normal phase and the superradiant phases, and find that there exists a competitive relationship of energy minimum on the overlapping region between SR1 and SR2. Interestingly, the transition between the normal phase to SR1 or SR2 is identified to be a second-order phase transition, while the transition between SR1 and SR2 is a first-order transition. When the first-order phase transition occurs, the phase of the photons changes abruptly from 0 to π/2.

1.   Introduction
  • Loading quantum gases into a high-finesse optical cavity provides fascinating avenues for simulating novel quantum phenomena and exploring many-body physics.[1–14] In such light–atom hybrid systems, the ultracold dilute bosonic or fermionic gases inside an initially unoccupied cavity mode are illuminated by an external pump laser field. When the strength of the pump field exceeds a critical value, the long-range interaction between particles is generated, stemming from the scattering of photons between the transverse pump and the cavity mode via off-resonant Raman processes, and is essential for the realization of fascinating many-body phenomena.[7,15–20] Typically the incident pump consists of a pair of counterpropagating pumping lasers with equal strength, in which case different phases of the system can be distinguished by the arrangement of the atoms since the phase of intracavity photons is constrained to be in either 0 or π.[4,15,21,22] However, employing the unbalanced counterpropagating beams has discernible effects on the phase of intracavity photons, which can be used as a means to discriminate different superradiant phases of the system.[23,24]

    The system undergoes a superradiant phase transition from the normal phase into the self-ordered state on the condition that the gain in potential energy is able to overcome the loss in kinetic energy.[15] At the transition point the collective excitation mode softens at the coupled momentum, and regularly distributed atoms in emerging intracavity light field can further reduce their energy.[16,25–27] For red detuning, the atoms are in S orbitals due to the scattering terms being even spatial symmetry. In contrast, in the blue-detuned regime, the atoms are able to be scattered to higher orbital bands due to the odd spatial symmetry of the scattering terms.[28,29] In addition, the spatial arrangement of atoms leads to destructive interferences, which lowers the repulsive pump lattice potential and allows quantum phase transitions to occur in the repulsive potential.[23,24,28–31]

    Such studies utilizing unbalanced counterpropagating beams in the case of blue detuning are common. Accompanying the deep research of cavity-QED in the field of quantum simulation, it has become appealing to introduce multiple cavities or pump beams in the settings.[9,32–37] Our model integrates the configurations from Ref. [23] (a single unbalanced pump beam) and Ref. [33] (two single-mode cavities), investigating phase transitions when replacing two cavities with two pump beams under unbalanced conditions (see Fig. 1(a)). By constructing phase diagrams, we conclude that the system does not experience suppression of intracavity photons in the region of equal intensities compared to the two pairs of balanced beams, and an energy minimum occurs in this region only when the photon phase is π/2. We find that there is a first-order transition employing an asymmetric pump beam, and that the first-order phase transition occurs at the swap between a local energy minimum and the global energy minimum.

2.   Model
  • We consider that a spinless Bose–Einstein condensate (BEC) is trapped inside a high-finesse optical cavity, and are illuminated by two pairs of pump beams that are a mixture of standing and running waves. As shown in Fig. 1(a), the angle between the cavity axis and per pump beam is 60°. The wave vectors of the cavity light and the pump beams are given by kc=k0e^x and kj=(−1)jk0sin(60°)e^x+k0cos(60°)e^y with j = {1,2}. Here j is the identifier of the different pumps and k0 is the wave-vector magnitude of the pumping lasers and the cavity mode. The interaction between atoms and light can be understood from the scalar atomic polarizability, H^A−L=−ℏE^†⋅E^/Δa,[38–41] so that the many-body Hamiltonian in a frame rotating with the pump laser frequency reads

    where Δa = ωp − ωa and Δc = ωp − ωc are the atom–pump detuning and the cavity–pump detuning, respectively. We consider the pump beams that have the same frequency ωp and they are detuned far from the atomic resonance frequency ωa to ensure that the electronic excitation is sufficiently small and the upper atomic excited state can be eliminated adiabatically. We assume that the cavity frequency is closely detuned from the pump frequency. Ψ^†(Ψ^), a^†(a^), p^=−iℏ∇ denotes atomic creation (annihilation) operator with mass M, photon creation (annihilation) operator and atomic momentum operator, respectively. Here E^=E^c+∑jEpj is the total electric field operator and the quantized electric field of the cavity is given by E^c=E0cos(kc⋅r)a^e^z, where E0 is the electric field strength of a single cavity photon. It is worth emphasizing that the pump field originating from two unbalanced counterpropagating beams can be described as Epj=e^z(Ej+eikpj⋅r+Ej−e−ikpj⋅r)/2, where Ej± represents the strong (+) or weak (−1) amplitude of pump field j. Substituting the total electric field operator E^ into the Eq. (1), we find that the Hamiltonian of the atom–light interaction consists of the following six terms:

    where

    In the above expressions, Vj = Ej+Ej−/Δa are optical lattice depth for pump j and U0=E02/Δa is potential depth for the cavity. The first and second terms in Eq. (3) are the intracavity lattice and transverse pump lattice potentials, while the third term and the last term belongs to the coherence between the pump lights and the two-photon scattering process between the pump lights and the cavity field, respectively. Here ηj±=U0Vj(γ±1/γ)/2 are the amplitudes of the interference between the pump lights and the cavity field with γ = (Ej+/Ej−)1/2 being the imbalance parameter. In this study, we require two pairs of unbalanced counterpropagating pumps to have the same γ, choose the recoil energy Er=ℏ2k02/2M as the basic energy unit and set ℏ = 1 from now on for brevity.

    In general, the phase transition between the normal phase and the superradiant phase is judged by the presence or absence of intracavity photons. In the normal phase where 〈a^〉=0, the atoms feel an optical lattice consisting of two pairs of unbalanced pump fields and are localized at the position corresponding to the minimum of the lattice potential energy. Figure 1(b) shows the arrangement of these particles in real space, which enjoys the translational symmetry of 2π/(3k0) and 2π/k0 along the horizontal direction and the vertical direction, respectively. As Q^=(a^†+a^)/2 and P^=i(a^†−a^)/2 are orthogonal quadratures of the cavity field, by which the expectation value of the cavity field operator can be represented as a complex value [see Fig. 1(c)], 〈a^〉=α=|α|eiϕ. Except for |α|, which serves as an order parameter for distinguishing between the normal phase and the superradiant phase, the phase ϕ is capable of further subdividing the superradiant phase. For ϕ = 0, the coupling term which is proportional to P^ is always zero, such that the superradiant phase is coupled to the standing wave of pump via the term ∑jcos(kj · r). Conversely, when ϕ = π/2, the superradiant phase is coupled to the running wave of pump via the term ∑jsin(kj·r). Furthermore, for 0 ≪ ϕ ≪ π/2, it is evident that the two couplings can exist simultaneously. We refer to these three types of superradiant phases as superradiant phase 1 (SR1), superradiant phase 2 (SR2), and mixed phase (MP) based on the differences in phase ϕ. In particular, SR1 is coupled with the standing wave of the pumps while SR2 is coupled with the traveling wave of the pumps. Interestingly, subsequent studies does not reveal the existence of the mixed phase.

3.   Theory and method toward a steady solution
  • The dynamics of such a cavity–atom hybrid system is governed by the Heisenberg equations of motion for atomic and photonic field operators. To seek a steady-state solution with mean-field approximation, we formally replace the operators Ψ^ and a^ with the atomic mean-field Nψ0 and the coherent complex expectation value α with N being the number of atoms. As the dynamical evolution of the cavity field is very rapid compared to the center of mass motion of the atoms, its dynamical processes can be adiabatically eliminated, i.e., iℏ∂a^/∂t=0. Meanwhile, the evolution of the condensate wave function is governed by the Gross–Pitaevskii-type equation, including the mechanical effects of the radiation field inside the cavity.[25,42–44] Thus, the coupled mean-field equations that the system obeys can be written as

    Here, the cavity decay rate κ originates from the leakage of intracavity photons. The bunching parameter ℬ=∫dr〈Ψ^†(r)B^Ψ^(r)〉 with B^=cos2(kc⋅r). We have defined operators Θ^+ and Θ^− as Θ^+=∑j=1,2ηj+cos(kj⋅r)cos(kc⋅r) and Θ^j−=∑j=1,2ηj−sin(kj⋅r)cos(kc⋅r), respectively, similarly Θ+=∫dr〈Ψ^†(r)Θ^+Ψ^(r)〉 and Θ−=∫dr〈Ψ^†(r)Θ^−Ψ^(r)〉. μ0 is the chemical potential and H^A−L(α*,α) is the resultant Hamiltonian obtained by the replacement of a^(a^†) with α (α*) in Eq. (2).

    In Landau’s phase transition paradigm,[45,46] the ground-state energy of the system is a global minimum. We consider two ways to obtain the system’s ground state. The first strategy stems from the split-operator pseudospectral method,[25,47–49] where we propagate ψ0 in imaginary time according to Eq. (4b), with cavity field replaced by Eq. (4a). The second strategy for searching the lowest ground energy is performing the point-by-point scanning,[23] where the ground-state energy of many-body system is evaluated as Eg = Nμ0 − Δc|α|2 from Eq. (1) and can be plotted as a function of the cavity field quadratures P and Q, then we survey exactly the global minima of ground energy in relation to photon (see Figs. 5(a)–5(e)).

    The first strategy is fast but may suffer from errors. When the ground-state energy landscape possesses multiple minima, the problem is that it may search for a solution that is a local minima of the energy, depending on the choice of the initial value. Meanwhile, the second strategy is slow but the result is exact. Based on their strengths, the former is used for line drawings (see Figs. 2 and 5(f)–5(j)) and the latter is used for three-dimensional (3D) drawings (see Figs. 3, 4(c), 4(e) and 5(a)–5(e)).

    Inspection of Eq. (4a) indicates that both the order parameters Θ+ and Θ− affect the expectation value of cavity field operator, and α can be real of purely imaginary only if the two order parameters are simultaneously nonzero. The property of the complex number can be captured by the phase ϕ of the intracavity photons. In order to clear up the relationship between ϕ and Θ+, Θ−, we plot the behavior of the lattice potential in the cavity U0|α2| and Θ+, Θ− as a function of the pump lattice strength V1 (see Fig. 2).

    We find that Θ+ is not zero accompanied by ϕ = 0, Θ− is not zero accompanied by ϕ = π/2, and the two parameters cannot exist at the same time. This result is acceptable although it seems to contradict Eq. (4a). Cavity decay can be neglected, since the parameters used in our study, |Δc−U0ℬ|2≫κ2, it only leads to a minor shift of α involving κ. Thus, Θ+ and Θ− dominate real and imaginary parts of α, respectively. The two parameters cannot be finite at the same time, which demonstrates the subdivision of the superradiant phase does not include the presence of MP.

    We find that Θ+ and Θ− are negative, which leads to α is positive due to (Δc−U0ℬ)<0 (see Eq. (4a)). This suggests that the phase appearing in ϕ = π or ϕ = 3π/2 are rationality ignored (see the supplement of Ref. [23]). The parity of the scattering terms reflects the atoms scattering from the S band to higher bands or remaining in the S band.[50] In Eq. (3), the atoms couple to spatial patterns with different parities, which indicates that ηj+(α† + α) is more strongly coupled to the second (P) band compared to the first (S) band.[29] In contrast, ηj−(α† − α) coupled more strongly to the first (S) band compared to the second (P) band.

    We are curious about what different results would be obtained between two pump beams with pure standing wave and two pairs of pump beams with traveling and standing wave components, such two types of pump fields can be analog by setting γ to 1 or not. As shown in Fig. 3, the phase diagrams depicted by the intracavity lattice potentials U0|α|2 in the plane of the parameters V1 and V2, and draw the corresponding phase ϕ of the photons. Here we only consider the case where Δc ≪ 0, since the opposite case, Δc ≫ 0, leads to the existence of a transition from metastable regime to unstable regime.[28]

    Figures 3(a) and 3(c) depict the 3D distribution of the intracavity lattice depth U0|α|2 across the V1–V2 parameter plane for γ = 1 and γ ≠ 1 cases respectively, with Figs. 3(b) and 3(d) displaying their two-dimensional (2D) cross-sections. The colored regions in Figs. 3(b) and 3(d) indicate the presence of photons, where hue variations encode the photon phase differences. We find that as V2 gradually approaches 2Er from 0, the threshold pump strength V1 required for the emergence of intracavity lattice depth increases, indicating an increased difficulty for the system to enter the superradiant phase. For V2 ∈ [2Er, 7Er], the systems first transition from the superradiant phase to the normal phase and then re-enter the superradiant phase as V1 increases. The significant differences between the two cases are: for γ = 1, the intracavity photon is suppressed near V1 = V2, and the system exhibits only a ϕ = 0 superradiant phase; for γ ≠ 1, intracavity photon population is allowed near the V1 = V2 condition, and the superradiant phase admits additional phases (e.g., ϕ = π/2).

    Figures 3(a)–3(d) are symmetric along V1 = V2, which means it is equivalent to fix one of pump depths while varying the other. Meanwhile, it illustrates that the pump field of standing wave produces self-ordering phase with only real expectation value of intracavity photon, and the pump field of traveling wave produces a self-ordering phase with purely imaginary expectation value of intracavity photon. For the case of blue detuning Δa ≫ 0, the self-ordering phase around V1 = V2 is coupled to the traveling wave. The dashed arrows in Fig. 3(d) point out the most interesting difference between γ = 1 and γ ≠ 1.

    Figure 3(e) illustrates the effects of Δc and γ on the phase boundary along V1 = V2 at γ ≠ 1. For the larger of γ, the smaller the optical lattice depth of pumps are required for the transition of NP into SR2 at a certain cavity–pump detuning Δc on the one hand. On the other hand, for the larger of Δc, the contrary the optical lattice depth of pumps are required for this transition at a certain imbalance parameter γ.

4.   Energy mechanism of the first-order superradiant phase transition
  • The order parameters Θ+ and Θ− can be well used to subdivide the superradiant phase, here, Θ+ ≠ 0 but Θ− = 0, Θ+ = 0 but Θ− ≠ 0, and both Θ+ ≠ 0 and Θ− ≠ 0, which correspond to SR1, SR2, and MP, respectively. At the vicinity of the critical point, we can treat the terms in Eq. (1) involving a^(a^†) as a perturbation since the photon number is very small.

    At the zero-temperature limit, all atoms occupy the lowest-energy state, so that the atomic field operator can be expanded with a set of Bloch wave solutions of the unperturbed Hamiltonian H^0

    where

    Here |Ψ(m)(k) ⟩ denotes the Bolch state with quasi-momenta k and band index m, and |Ψ(1)(0)⟩ is the lowest-energy Bloch state. b^0,1 is the annihilation operator for state |Ψ(1)(0)⟩ and b^k,m is the annihilation operator for state |Ψ(m)(k)⟩ (k ≠ 0). Noteworthy, all possible two-photon processes coherently couple the zero-momentum state of the BEC |kx, ky ⟩ = |0,0⟩ to the twelve momentum-states, |kx, ky ⟩ = |±kc ± kj ⟩ and |kx, ky ⟩ = |±k1 ± k2 ⟩. We take into account the symmetry of the energy band structure to simplify the calculation, hence it is reasonable to use the momentum k1 = kc − k2 to replace all possible momentums in the second term of Eq. (5).

    Substituting the atomic field operator into Eq. (1) results in the effective Hamiltonian

    where Δ˜c=Δc−U0〈Ψ(1)(0)|B^|Ψ(1)(0)〉 is dispersive shift cavity detuning. E0,1 and Ek1,m are eigenvalues of H^0 corresponding to eigenvectors |Ψ(1)(0)⟩ and |Ψ(m)(k1)⟩, respectively. The scattering matrix elements are given by θm,+=〈Ψ(1)(0)|Θ^+|Ψ(m)(k1)〉, and the other as θm,−=〈Ψ(1)(0)|Θ^−|Ψ(m)(k1)〉. In the derivation, the particle number conservation b^0,1†b^0,1+∑mb^k1,m†b^k1,m=N should be satisfied, and the summation are carried over all band indices.

    We implement the mean-field theory by replacing the atomic operators and photon operator by their average values b^0,1=ψ0,1, b^k1,m=ψk1,m, and a^=α. We seek a steady state in which ∂tψk1,m = 0 with i∂tb^k1,m=[b^k1,m,H^eff] and obtain ψk1,m=−[(α*+α)θm,+*+i(α*−α)θm,−*]ψ0,1/(Ek1,m−E0,1). Together with the normalization condition |ψ0,1|2=N−∑m|ψk1,m|2. Substituting these expressions into Eq. (7) and obtaining the ground-state energy up to the fourth order in the photon field amplitudes α

    two susceptibilities are χ1 = ∑m|θm,+|2/(Ek1,m − E0,1) and χ2 = ∑m|θm,−|2/(Ek1,m − E0,1). We have introduce two symbols ν1 = ∑m|θm,j+|2/(Ek1,m − E0,1)2 and ν2 = ∑m|θm,j−|2/(Ek1,m − E0,1)2 for compact notations. The scattering matrix elements obey θm,+⁢θm,−∗+θm,+∗⁢θm,−=0. In the above discussion, equation (4a) informs us about the valuer of α in the steady condition, and we obtain α=(Θ++iΘ−)/Δ˜c by reasonably neglecting κ. So the ground-state energy becomes

    Some comments are in order. Firstly, as equation (7) is invariant under the simultaneous transformation a^→−a^ and b^k1,m→−b^k1,m, the system spontaneously breaks Z2 symmetry during the transition from the normal phase to superradiance phase.[51] Secondly, in Eq. (9), the coefficients of the fourth-order terms being positive ensures the stability of the system, while the coefficients of the second-order terms being positive or negative cause the minimization of the ground state energy to occur at α being zero or finite with the former case corresponding to the normal phase and the latter case to the superradiant phase. Accordingly, yielding the established superradiant phase condition

    here, equation (10a) corresponding to Θ+ is finite and equation (10b) corresponding to Θ− is finite.

    The key part of the phase transition condition is the two susceptibilities, which characterizes the tendency of inducing superradiance ϕ = 0 or ϕ = π/2. In Figs. 4(a) and 4(b), we present the numerical results for χ1 and χ2 at zero temperature with γ = 1.37. The difference between the two graphs lies in two aspects. One is the position of the maximum value in the plane: as V1 and V2 gradually increase, the maximum of χ1 moves closer to the center axis, but the maximum of χ2 is always close to both axes. The other is the value around V1 = V2: χ1 tends to zero, while χ2 is larger. Their similarity is not obvious but traceable. When the value of V2 is large, χ1 and χ2 follow the tendency of increasing firstly and then decreasing, and the descending rate is rapid in Fig. 4(a) but the ascending rate is quick in Fig. 4(b).

    Solving the phase transition condition Eqs. (10a) and (10b) with respect to Δc yields the critical cavity detuning

    where Δc,l is the pump–cavity detuning for the difference susceptibilities χl, with l = 1, 2.

    A larger χl indicate a greater critical magnitude of the cavity detuning Δc,l. We map out the phase diagrams spanned by V1 and Δc at V2 = 0. On one hand it is consistent with the results of Ref. [23], proving the correctness of methods. On the other hand, we find the connection between the analytical and numerical solutions through the phase diagrams. The boundary between the normal phase and superradiance is perfectly matched [see Figs. 4(c) and 4(d)]. However, the subdivision for the superradiant phase is not satisfactory due to the fact that the curve of critical cavity detuning corresponding to χ1 and the curve of critical cavity detuning corresponding to χ2 have both overlapping and independent regions [see Fig. 4(d): the non-overlapping dark-red is region SR1, light-blue is region SR2, and the gray indicates the overlapping region].

    Subdividing the independent regions is effortless, as they can be understood by what phases of intracavity photons can be allowed to occur inside the areas. At first glance, the gray regions imply that photons of the two phases ϕ = 0 and ϕ = π/2 are capable of coexisting, and the superradiance of the system is mixed phase in this region, where the atoms feel the coupling from both standing and traveling waves. This is not the case, however, the inset in Fig. 4(c) demonstrates that the gray region is completely occupied by SR1. In order to clarify the attribution of the overlapping region, we plot the phase diagrams at V2 = 3 for comparisons [see Figs. 4(e) and 4(f)]. Identically, the boundary between the normal phase and superradiant phase has the same tendency for both numerical and analytical solutions. However, the overlapping region is almost entirely occupied by the superradiant phase SR2. This implies that the gray region has a competitive relationship of ground-state energy minimum for intracavity photons of the phases ϕ = 0 and ϕ = π/2. In contrast, for the normal phase and independent regions [see Figs. 4(d) and 4(f)], white directly indicates that the minimum energy occurs at the photon number being zero, and the dark–red and light–blue indicate that the expectation value α is real and pure imaginary, respectively.

    The underlying algorithm for numerical simulation is to search for the global minimum of the ground-state energy of the system. To make clear of what is this energy trend on both sides of and at the point of phase boundaries. In other words, we want to sort out whether the transition between normal phase and superradiance is the first order or the second order, and similarly we want to sort out whether the transition between SR1 and SR2 is the first order or the second order. The zero-temperature energy landscapes for different strengths of lattice potentials V1 and V2 are depicted in Fig. 5. Specifically, figures 5(a)–5(e) show the energy landscape as a function of the cavity field quadratures Q=〈Q^〉 and P=〈P^〉. It is clearly seen in Fig. 5(a) that the energy minima appears at 〈a^〉=0 in the normal phase. Figures 5(b) and 5(e) indicate the system enters SR1 and SR2, respectively, and suggest that the energy minimum appears on the real axis in the SR1 and on the imaginary axis in the SR2. The transition between the normal phase and superradiant phase is a second-order phase transition, because there is only one energy minimum in the global plane and the transition occurs when α varies continuously across zero. Figures 5(c)–5(d) describe the case there are two different minima located at the quadratures Q and P axes. As small changes in tuning parameters occur, the energy can eventually run from the local minima to the global minima, and the accompanying exchange of the phase of the cavity light field reveals collective excitations from one superradiant phase to another. The distributions of the energy landscapes in Figs. 5(c) and 5(d) show that there is no competition for mixed phase on gray regions in Figs. 4(d) and 4(f).

    Figures 5(f) and 5(j) show the energy of ground state on the elliptic path connecting the real and imaginary minima, parameterized with the phase ϕ. Obviously, in SR1 region far away from the boundary of SR1 and SR2, there is no local minimum for the phase ϕ = π/2 of the cavity photons. Slowly moving towards the critical point, a local energy minimum develops and equilibrates with that of the phase ϕ = 0. Further increasing the pump strength, the energy swap from the local minimum to the global minimum, and eventually the local minimum for real α will vanish. Distinguishing superradiance SR1 from SR2 via the phase ϕ can be understood in terms of energy, the value equaling to 0 or π/2 causes the system to have a local and global minimal energies, and the first-order phase transition of the system occurs when local and global minima swap.

5.   Conclusion
  • We have considered a system of cavity–atom hybridization irradiated by two pump fields, and have thoroughly investigated the difference between the pumps composed of pure standing waves (γ = 1) and the pumps composed of standing and running waves (γ ≠ 1). By formulating and solving self-consistent mean-field equations, we obtained order parameters of the system and constructed ground-state phase diagrams spanned by the pump lattice depths V1/Er and V2/Er. Besides the normal phase, we found that the superradiant phase can be further subdivided into SR1 with ϕ = 0 and SR2 with ϕ = π/2. We identified that SR1 couples to the standing wave pumps via ∑j cos(kj · r) and SR2 couples to the running wave pumps via ∑j sin(kj · r). In addition, we calculated susceptibilities χ1 and χ2 to determine the tendency of the normal phase to enter into superradiance SR1 and SR2, respectively, and obtained the phase transition condition in terms of the cavity field quadratures acting as real and imaginary axes. We found the phase transition between the normal phase and superradiance is the second order, while the phase transition between SR1 and SR2 is the first order. In particular, approaching the critical point of the two superradiant phases, the ground-state energy will have a local minimum and a global minimum, and the two energy minima become equal at the critical point. When the local and global energy minima swaps, the system undergoes the first-order phase transition.

    Our approach is feasible for studying the phase transition in such a cavity-BEC system, but the study of the first-order phase transition cannot be quantitatively analyzed due to the difficulty of finding suitable formulas to simulate the simultaneous occurrence of two minima on the orthogonal axes. In this study, the phase boundaries also depend on the relative spatial phase Φ between different pump and cavity fields, and here we have assumed that Φ is zero. For nonzero Φ, the new Hamiltonian now depends on it.[47,51] Of course, altering the type of particles in the cavity is also interesting for discussing many-body phases.[52–56] Encouragingly, loading ultracold quantum gases into high-finesse cavities provides a versatile platform to explore structural phase transitions between different crystal configurations, since this leads to the cavity-mediated long-range interaction between atoms and different kinds of long-range interaction support different spatial structures in both light and atomic wave functions. We believe that our approach is favorable to the study of phase diagrams in cavity–atoms system and to the understanding of phase transitions from the energy point of view.[23]

Figure (5)  Reference (56)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return