2025 Volume 34 Issue 10
Article Contents

Ke-Ji Chen, Fan Wu. Dynamically generating superflow in a bosonic ring via phase imprinting[J]. Chinese Physics B, 2025, 34(10): 106701. doi: 10.1088/1674-1056/add67b
Citation: Ke-Ji Chen, Fan Wu. Dynamically generating superflow in a bosonic ring via phase imprinting[J]. Chinese Physics B, 2025, 34(10): 106701. doi: 10.1088/1674-1056/add67b

Dynamically generating superflow in a bosonic ring via phase imprinting

  • Received Date: 31/03/2025
    Accepted Date: 06/05/2025
    Available Online: 01/10/2025
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(3)

Article Metrics

Article views(395) PDF downloads(0) Cited by(0)

Access History

Dynamically generating superflow in a bosonic ring via phase imprinting

Abstract: Phase imprinting enables the dynamic generation of superflow in bosonic atoms, effectively overcoming traditional limitations such as vortex number constraints and heating effects. However, the mechanisms underlying superflow formation remain insufficiently understood. In this work, we reveal these mechanisms by studying the time evolution of the transferred total angular momentum and the quantized current throughout the phase imprinting process, achieved through numerically solving the time-dependent Schrödinger and Gross–Pitaevskii equations. We demonstrate that the Bose gas dynamically acquires angular momentum through the density depletion induced by the phase imprinting potential, whereas quantized currents emerge from azimuthal phase slips accompanied by complete density depletions. Regarding the impact of system parameters, such as interactions, we find that interactions hinder superflow formation, as the azimuthal density distribution becomes less susceptible to the phase imprinting potential. Our findings offer microscopic insights into the dynamic development of superflow during the phase imprinting process and provide valuable guidance for ongoing experimental efforts.

1.   Introduction
  • Superflow (or persistent current) in ring geometries, induced by vector potentials in superconductors,[1] or by synthetic gauge fields in neutral cold atoms,[2,3] has significant application potentials due to the vanishing energy dissipation and the robustness of magnetic-flux quantization. For instance, atomic superflow, realized in weakly interacting Bose–Einstein condensates (BECs) confined in toroidal traps, can be useful for quantum sensing, offering an atomtronic analogue to the celebrated superconducting quantum interference devices (SQUIDs).[4–8] An annular or ring-shaped superfluid symbolizes the minimal realization of a matter-wave circuit, serving as a fundamental building block for future atomtronic devices.[9,10]

    The achievements of atomic superflows in BECs have spanned over a decade[11–13] and generally rely on two schemes. One is rotation, which has proven efficient in producing superflows with well-defined winding numbers,[14] but it is limited to preparing relatively small winding numbers of superflows and requires a long preparation time. The other is utilizing Laguerre–Gauss beams, which carry angular momentum with helicoidal phases, to transfer angular momentum to the condensates through two-photon Raman processes,[12,15–17] and give rise to various intriguing phases.[18–30] This scheme is efficient for the preparation of a specific superflow, determined by the order of the angular momentum associated with the Laguerre–Gauss mode. However, the Raman process relies on the hyperfine structure of the atoms and introduces inevitable heating effects,[29,31] which limits its application.

    Phase imprinting, as a fundamental experimental technique, has been proposed and utilized to induce superflow in bosonic atoms.[32] More recently, superflow in fermionic rings has also been demonstrated either by stirring strongly interacting Fermi gases[33] or through a phase imprinting process.[34,35] In Ref. [34], Del Pace et al. introduce an angular-dependent phase imprinting potential and observe persistent current by measuring the winding number through an interferometric technique.[36–38] This work has attracted considerable interest in exploring the generation and stability of superflow in fermionic rings.[39–41] Building on the seminal experiment by Del Pace et al.,[34] several fundamental questions naturally arise. Can superflow emerge if we introduce such a phase imprinting potential into a Bose gas? Whereas a conventional atomic gas is irrotational, superflow is distinguished by a finite circulation. How does the system evolve from a non-circulatory state to one with quantized circulation? Furthermore, how do system parameters influence the formation of superflow?

    To address these questions and concentrate our discussions, we consider a Bose gas confined by a ring-shaped potential and illuminated by a phase imprinting potential, as realized in Ref. [34]. We focus on the time evolution of the transferred total angular momentum and quantized current throughout the entire phase imprinting process. Many intriguing dynamical features, such as density depletions, phase slips of wave functions, and dynamical phase transitions, emerge. We demonstrate that the density depletion, induced by the phase imprinting potential, leads to an increase in the total angular momentum. Conversely, the quantized currents arise from azimuthal phase slips, accompanied by complete density depletions, driving dynamical phase transitions from a zero to a nonzero circulation state through the phase imprinting process.

    This paper is organized as follows. Section 2 describes our system and analyzes its angular momentum dynamics. Section 3 reveals the mechanism of superflow generation by analyzing the features of the wavefunction. In Section 4, we investigate the impact of interactions and show that they hinder superflow formation. Finally, Section 5 summarizes our main results.

2.   Model
  • We consider a Bose gas with mass M confined by an annular potential

    as illustrated in Fig. 1(a) in the x–y plane, and by V(z)=Mωz2z2/2 in the z direction. Here, V0 denotes the trapping strength, and R1 (R2) represents the inner (outer) radius of V(r,θ). The parameter d is a length scale, and ωz is the trapping frequency along the z axis. To capture the essential physics and simplify the calculation, we model the Bose gas confined in a ring-shaped potential. This configuration can be realized by ensuring ħωz exceeds any other energy scale, and by requiring d ≪ Rj = 1, 2 with R1 ≈ R2. Under these constraints, atomic degrees of freedom along the axial and radial directions are suppressed, resulting in a one-dimensional gas confined in a ring with a radius R, where R ≡ (R1 + R2)/2.

    Phase imprinting is achieved via an angular-dependent potential U(θ), as shown in the inset of Fig. 1(b), which is given by

    with U0 the angular potential depth and Δθ ≪ 2π. As [U(θ),L^z]≠0 with L^z≡−iℏ∂/∂θ, U(θ) plays a crucial role in introducing angular momentum into the system. Here, we consider a specific scenario, in which U(θ) is switched on (off) for t ⩽ τ (t > τ), with τ the interval time. Therefore, a noninteracting Bose gas can be described as H0(t)=∫dθψ†(θ,t)ℋs(θ,t)ψ(θ,t), where ψ(θ,t) is the field operator, satisfying the time-dependent Schrödinger equation (TDSE)

    with

    where ℋs(θ)=−ℏ2/(2MR2)∂2/∂θ2 and ϑ(x) is the Heaviside step function. For the Bose gas, we choose the ground state of H0=∫dθψ†(θ)ℋs(θ)ψ(θ) as the initial state, and we calculate the time evolution of lztot(t), where

    Before delving into the detailed calculation of the time evolution of lztot(t), it is helpful to analyze its equation of motion. According to Eqs. (3) and (5), the equation of motion for lztot(t) reduces to

    where

    with n(θ,t) = |ψ(θ,t)|2 representing the density profile. Here, n¯L(t) and n¯R(t) denote the average density for θ ∈ [0,2π − Δθ] and θ ∈ (2π − Δθ, 2π), respectively. As illustrated in Eq. (6), for t ⩽ τ, an inequality between n¯L(t) and n¯R(t) breaks the conservation of lztot(t) and introduces angular momentum into the system. This imbalance naturally arises from the form of U(θ), which inevitably introduces a density depletion for θ ≈ 2π due to its sharp increase for θ ∈ (2π − Δθ, 2π).

    We validate the above analysis by solving the TDSE. Specifically, due to the ring geometry, we expand ψ(θ,t) = ∑mcm(t)Θm(θ), where Θm(θ)=eimθ/2π with m the magnetic qunatum number, and cm(t) is the time-dependent coefficient. This expansion transforms the TDSE into a matrix equation for cm(t), given by

    where ℋm,m′(t)=Km,m′+fm,m′ϑ(τ−t) with Km,m′=m2ℏ2/(2MR2)δm,m′ and

    Equation (9) is then directly numerically solved using the fourth-order Runge–Kutta algorithm, and consequently ψ(θ,t) is determined as well.

    Figure 1(b) shows the density profile n(θ,t), where a density depletion emerges for θ ≈ 2π as expected. Regarding the equation of motion for lztot(t), as depicted in Fig. 1(c), we observe that lztot(t) increases continuously from zero as our anticipation. The continuity of lztot(t) can be attributed to the angular dependence of both the modulus and phase of ψ(θ, t).

    The above analysis and calculations elucidate the origin of the injected angular momentum, yet the question of whether superflow can emerge remains ambiguous. To address this, we define the angular momentum with respect to the phase of ψ(θ,t) as

    which is the quantized component of the current. Here, φ(θ,t) represents the phase of ψ(θ,t), where ψ(θ,t) = |ψ(θ,t)|eiφ(θ,t) with |ψ(θ,t) the modulus of ψ(θ,t). The quantization of lzφ(t) arises from the single-valued nature of ψ(θ,t), which requires ψ(0,t) = ψ(2π,t) since θ = 0 and θ = 2π represent the same point on the ring. This leads to ∫ dθ∂φ(θ,t)/∂θ = φ(2π,t) − φ(0,t) = 2πκ with κ the winding number. In realistic calculation, ∂φ(θ, t)/∂θ in Eq. (11) can be extracted from the current density j(θ,t), defined as j(θ,t) = [ψ*(θ,t)∂ψ(θ,t)/∂θ − h.c.]/(2i) and ∂ φ(θ,t)/∂ θ = j(θ,t)/n(θ,t). Figure 1(c) shows the time evolution of lzφ(t). Three dynamical features of lzφ(t) emerge. First, lzφ(t) becomes quantized as expected. Second, jumps of lzφ(t) occur during the phase imprinting process. Third, lzφ(t) becomes stable at a long time, indicating that the persistent current state is robust. From these unique features, we conclude that superflow with different winding numbers emerge.

3.   Mechanism of superflow generation
  • To further elucidate the mechanism of superflow formation, we calculate the phase evolution of ψ(θ,t), which is given by

    As illustrated in Fig. 2(a), we show the time evolution of Δφ(θ,t) at different times as θ ranges from 0 to 2π. We observe that Δφ(2π,t) = 0 for tE0/ħ = 0.5; however, for tE0/ħ = 1.5, Δφ(2π,t) = 2π, clearly indicating the occurrence of a phase slip between these two time points.

    We then consider the time evolution of |ψ(θ,t)|min, the minimal value of |ψ(θ,t)|. As depicted in Fig. 2(b), we observe that the jump time of lzφ(t) [see the inset of Fig. 2(b)] and the time when |ψ(θ,t)|min = 0 are identical, indicating that the phase slip occurs when the wave function (or density) becomes completely depleted. We further confirm this observation by plotting the modulus profile of ψ(θ,t) at different times as illustrated in Fig. 2(c). We find that when tE0/ħ = 0.98, a node point of |ψ(θ,t)| emerges at θ ≈ 2π, with |ψ(θ, t)| being completely depleted, before or after this time, such as tE0/ħ = 0.5 and tE0/ħ = 1.5, no node point of |ψ(θ, t)| emerges. Therefore, the appearance of a node point of |ψ(θ,t)| can be considered as a criterion for the phase slip. In Fig. 2(d), we present the relation between the stable lztot(t)(lzφ(t)) and τ at long times, meaning t ≫ τ. Figure 2(d) indicates that a specific quantized current state can be obtained by tuning the duration τ.

    From the detailed analysis above, we conclude that the generation of superflow arises from phase slips, accompanied by the depletions of the wave functions, driving phase transitions from a zero-circulation state to a state with nonzero circulation during the phase imprinting process. By tuning the duration τ, quantized current states with different phase windings can be obtained.

4.   Impacts of interactions
  • Regarding realistic systems, interactions are inevitable. Here, we consider an s wave interaction, and the full Hamiltonian is given by H(t) = H0(t) + Hint(t), where

    with g being the interaction strength. Under the mean-field approximation, the Hamiltonian reduces to

    where n0(θ,t) = |ϕ0(θ,t)|2 is the condensate density of the interacting Bose gases with ϕ0(θ,t) being the condensate wave function. Here, ϕ0(θ,t) satisfies the time-dependent Gross–Pitaevskii equation, which is given by

    Similar to solving the TDSE, we expand ϕ0(θ,t)=∑mcm(t)Θm(θ) and substitute this expansion into Eq. (15), yielding

    where

    Equation (16) can be solved using the fourth-order Runge–Kutta method, allowing us to obtain ϕ0(θ,t) as well.

    Analogous to the noninteraction case, as illustrated in Eqs. (5) and (11), we define the total angular momentum lztot(t) and the quantized current lzφ(t) in the interaction case by replacing the field operator ψ(θ, t) with the condensate wave function ϕ0(θ, t). Here, we choose the ground state of an interacting Bose gas without phase imprinting potential as the initial state, and calculate the time evolution of lztot(t) and lzφ(t). As shown in Fig. 3(a), by comparing the results with and without interactions, we find that the presence of interactions suppress both lztot(t) and lzφ(t). This can be understood as follows: interactions favor a homogeneous density, which reduces the density depletion and, in turn, suppresses the increase of lztot(t). This understanding is further confirmed by numerical calculating the density difference between n¯L(t) and n¯R(t), as illustrated in Fig. 3(b). Here, we present the time evolution of n¯L(t)−n¯R(t) under different interaction strengths and demonstrate that the stronger interaction strength leads to a smaller difference between n¯L(t) and n¯R(t). According to Eq. (6), this reduced difference results in a slower increase in lztot(t). Ultimately, the smaller lztot(t) gives rise to a diminished lzφ(t). Figure 3(c) presents |ϕ0(θ, t)|/f0 under different interaction strengths at the same time. When contrasted with the results in the absence of interactions, we observe that the interactions reduce the oscillation of the wave function, which confirms our analysis that interactions favor a homogeneous density, making the system less susceptible to the phase imprinting potential, and ultimately suppressing both lztot(t) and lzφ(t).

5.   Conclusions
  • We investigated the generation of superflow and dynamic transitions induced by angular phase imprinting in a Bose gas. We demonstrated that dynamic transitions occur between the zero-circulation state and finite-circulation current states through the phase imprinting technique. Our microscopic approach reveals that the transferred angular momentum arises from the density depletion, induced by the phase imprinting potential, whereas transitions between states with different quantized persistent currents are driven by phase slips and wave function (or density) depletions in the angular direction. We further investigated the impact of interaction strength in the phase imprinting process and found interactions suppress the formation of superflow. Our results provide a microscopic understanding for the dynamic generation of superflow, several distinctive features that can be probed through time-of-flight imaging combined with interferometric techniques. These insights may contribute to refining current experimental protocols for superflow preparation and control.

Figure (3)  Reference (41)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return