betaFeedback

WorksPhysics and Astronomy

Finite-temperature effects on the single-particle Green's function of a mobile impurity in a Luttinger liquid

Abstract

We consider the problem of a quadratically-dispersing quantum impurity interacting through a contact density-density term with a one-dimensional (1D) superfluid described by Luttinger liquid theory. By means of a field-theoretical approach based on the linked-cluster expansion, we quantitatively address the effects of a finite temperature of the 1D bath on the momentum-space single-particle Green's function. Through a careful identification of the resonant phonon-impurity scattering processes, we obtain asymptotic large-time estimates for the thermally-induced exponential damping of the zero-temperature signal both in the quasiparticle\textit{quasiparticle} regime, dominated by the presence of a finite lifetime for the host impurity, and in the infrared-dominated\textit{infrared-dominated} regime, where the impurity causes a Anderson's orthogonality catastrophe of the 1D bath. In both cases, we identify a temperature-induced coherence time that sets the characteristic timescale of the exponentially-decaying finite-temperature correction to the amplitude of the zero-temperature single-particle Green's function. Moreover, we draw connections between the aforesaid thermal timescale and the microscopic parameters of the Lieb-Liniger model, which admits an effective Luttinger-liquid description, highlighting the dependence of the inverse coherence time on the intra-bath interactions across the different dynamical regimes of the impurity.

Introduction

The problem of a mobile quantum impurity interacting with a one-dimensional many-body environment has attracted sustained interest across condensed matter and cold-atom physics[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]. In one dimension, the host medium is generically described as a Luttinger liquid[13, 14], where low-energy excitations are collective and the conventional quasiparticle picture breaks down[15, 16, 17, 18, 19]. As a consequence, even a single impurity exhibits highly nontrivial dynamics, shaped by its coupling to gapless bosonic modes.

A central object to characterize these effects is the impurity single-particle Green’s function, which encodes both dynamical and spectral properties. At zero temperature, its behavior is known to display nontrivial power-law decay in time and singular features in frequency within the so called infrared (IR) regime, reflecting the critical nature of the underlying liquid and the absence of well-defined quasiparticles[1, 2, 3, 4, 20, 21, 22, 23, 24]. These features are closely related to the physics captured by Luttinger liquid theory and to impurity-induced orthogonality effects[25, 26, 27, 28].

Despite this well-established zero-temperature picture, much less is understood about the role of finite temperature, which is unavoidable in experimental realizations[29, 30, 31]. Thermal fluctuations introduce an additional energy scale that competes with the intrinsic low-energy dynamics of the system, potentially modifying both the temporal decay and the spectral structure of the impurity Green’s function[32, 33, 34, 35]. In particular, finite temperature is expected to act as an infrared cutoff, smoothing zero-temperature singularities and inducing finite coherence times for the impurity[36].

In this work, we investigate finite-temperature corrections to the single-particle Green’s function of a mobile impurity in a Luttinger liquid. Focusing on the momentum-resolved and time-dependent behavior, we analyze how thermal effects modify the algebraic decay characteristic of the zero-temperature limit, giving rise to a crossover toward thermally dominated dynamics. We show that the interplay between impurity motion and collective excitations leads to a momentum-dependent decoherence and a nontrivial redistribution of spectral weight. Our results provide a unified framework to understand impurity dynamics beyond the zero-temperature regime and offer direct implications for ongoing experiments in one-dimensional quantum systems.

The manuscript is organized as follows: in Section II, we define the model Hamiltonian, the Keldysh field-theoretical framework that we employ, and the correlation function to be characterized quantitatively in the remainder of the manuscript. In Section III, we present the main analytical result of the present work, i.e., the linked-cluster approximation for the finite-temperature single-particle Green’s function, thereby considering distinct momentum regimes and the associated long-time properties of the aforementioned single-particle Green’s function. Then, in Section IV, the obtained results are discussed in their physical content through the definition of a characteristic time-scale for the temperature-induced degradation of the zero-temperature results, as well as through the connection with a microscopic model. Finally, in Section V, we summarize the main findings and discuss further theoretical setups which might be amenable to further investigation in future studies.

Model and methods

In this section, we introduce the physical model that we are going to consider, as well as the physical quantities that we are going to compute and the technical tools employed to this end. Specifically, we consider the model Hamiltonian H^=H^imp+H^TLL+H^int\hat{H}=\hat{H}_{imp}+\hat{H}_{TLL}+\hat{H}_{int}; the first contribution denotes the Hamiltonian of the bosonic impurity:

H^imp=∑kϵkd^k†d^k,(1)\hat{H}_{imp}=\sum_k\epsilon_k\hat{d}^{\dagger}_k\hat{d}_k, \tag*{(1)}

where the impurity dispersion relation satisfies ϵ−k=ϵk\epsilon_{-k}=\epsilon_k, the main focus of our work being the natural choice of a quadratic law of the form ϵk=tik2\epsilon_k=t_i k^2. The second term of the system’s Hamiltonian is the Hamiltonian of the bath, taken to be of the form of a Tomonaga-Luttinger liquid (TLL) [15] theory:

H^TLL=vs2π∫dx[K(∂xθ^)2+1K(∂xϕ^)2],(2)\hat{H}_{TLL}=\frac{v_s}{2\pi}\int dx\left[K(\partial_x\hat{\theta})^2+\frac{1}{K}(\partial_x\hat{\phi})^2\right], \tag*{(2)}

where the fields ϕ^(x)\hat{\phi}(x), θ^(x)\hat{\theta}(x) satisfy the commutation relation [ϕ^(x),∂x′θ^(x′)]=iπδ(x−x′)[\hat{\phi}(x),\partial_{x'}\hat{\theta}(x')]=i\pi\delta(x-x'), vsv_s is the sound velocity of the TLL and KK denotes the Luttinger parameter; the third contribution models instead the density-density interaction term between the TLL and the bosonic impurity:

H^int=−gπ∫dx (∂xϕ^)d^†(x)d^(x),(3)\hat{H}_{int}=-\frac{g}{\pi}\int dx\,(\partial_x\hat{\phi})\hat{d}^{\dagger}(x)\hat{d}(x), \tag*{(3)}

where we introduced the notation d^(x)=1L∑keikxd^k\hat{d}(x)=\frac{1}{\sqrt{L}}\sum_k e^{ikx}\hat{d}_k. We remark that the choice of the form of H^int\hat{H}_{int} relies on the assumption that the Luttinger parameter KK of the 1D bath satisfies K>1K>1, so that it is possible to retain only the long-wavelength part of the density fluctuations, while safely neglecting the Fourier modes at momenta that are multiples of 2πρ02\pi\rho_0 [1, 37], ρ0\rho_0 being the number density of the 1D bath.

In order to provide a quantitative description of the dynamical properties of a mobile impurity inside a generic 1D gapless quantum bath described by Luttinger-liquid theory, we initialize the system in the state:

ρ^0=e−β∑q≠0vs∣q∣b^q†b^qZTLL⊗∣0⟩⟨0∣,(4)\hat{\rho}_0=\frac{e^{-\beta\sum_{q\ne0}v_s|q|\hat{b}^{\dagger}_q\hat{b}_q}}{Z_{TLL}}\otimes|0\rangle\langle0|, \tag*{(4)}

where we introduced the notation ∣0⟩|0\rangle to denote the vacuum state for the impurity, and the bosonic operators b^q\hat{b}_q and b^q†\hat{b}^{\dagger}_q to indicate the bosonic modes that diagonalize H^TLL\hat{H}_{TLL}, while ZTLLZ_{TLL} is the partition function of the thermal density matrix of the 1D bath.

In the following, we evaluate the single-particle Green’s function:

G(y,t;x,0)=Tr⁡[eiH^td^(y)e−iH^td^†(x)ρ^0],(5)G(y,t;x,0)=\operatorname{Tr}\left[e^{i\hat{H}t}\hat{d}(y)e^{-i\hat{H}t}\hat{d}^{\dagger}(x)\hat{\rho}_0\right], \tag*{(5)}

in order to gain information on how the impurity’s dynamics is affected by the interaction with the host 1D quantum liquid.

The technical framework that we rely on to perform the calculation of time-dependent correlation functions and observables is the Keldysh formalism [38], where one introduces the Keldysh action S[ψ,ψ∗]S[\psi,\psi^*] in order to express the Keldysh partition function as Z=Tr⁡[ρ^(t)]=∫D[ψ,ψ∗]eiS[ψ,ψ∗]=1Z=\operatorname{Tr}[\hat{\rho}(t)]=\int\mathcal{D}[\psi,\psi^*]e^{iS[\psi,\psi^*]}=1 as a function of the retarded and advanced fields ψ=(φq,+,φq,−)\psi=(\varphi_{q,+},\varphi_{q,-}) associated with the degrees of freedom of the system, to be defined below. For our purposes, the action consists of 3 terms, identified as the impurity action SimpS_{imp}, the action of the TLL bath STLLS_{TLL} and the action SintS_{int} describing the interaction between the latter two. In the following, we introduce in the different terms of the action the so called classical and quantum component fields (ψcl,ψqu)(\psi_{cl},\psi_{qu}), respectively obtained from the retarded and advanced fields ψ±\psi_{\pm} with identical set of quantum numbers via the rotation ψcl=(ψ++ψ−)/2\psi_{cl}=(\psi_++\psi_-)/\sqrt{2}, ψqu=(ψ+−ψ−)/2\psi_{qu}=(\psi_+-\psi_-)/\sqrt{2}, where quantum number labels have been omitted for simplicity. The Keldysh action for the system of noninteracting bosonic modes modelling the bath reads:

STLL=∑q≠0∫dω2π∑α,β=cl,quφq,α∗(ω)Gq,α,β−1φq,β(ω),(6)S_{TLL}=\sum_{q\ne0}\int\frac{d\omega}{2\pi}\sum_{\alpha,\beta=cl,qu}\varphi^*_{q,\alpha}(\omega)\mathcal{G}^{-1}_{q,\alpha,\beta}\varphi_{q,\beta}(\omega), \tag*{(6)}
G−1=(0ω−ωq−iδω−ωq+iδ2iδF(ω)),(7)\mathcal{G}^{-1}= \begin{pmatrix} 0 & \omega-\omega_q-i\delta\\ \omega-\omega_q+i\delta& 2i\delta F(\omega) \end{pmatrix}, \tag*{(7)}

with F(ω)=1+2nB(ω)F(\omega)=1+2n_B(\omega) is simply related to the Bose-Einstein distribution nB(ω)n_B(\omega) and ωq=vs∣q∣\omega_q=v_s|q|. The Keldysh action for the bosonic impurity has a similar structure:

Simp=∑k∫dϵ2π∑α,β=cl,quφk,α∗(ϵ)Gk,α,β−1φk,β(ϵ),(8)S_{imp}=\sum_k\int\frac{d\epsilon}{2\pi}\sum_{\alpha,\beta=cl,qu}\varphi^*_{k,\alpha}(\epsilon)\mathcal{G}^{-1}_{k,\alpha,\beta}\varphi_{k,\beta}(\epsilon), \tag*{(8)}
G−1=(0ϵ−ϵk−iδϵ−ϵk+iδ2iδF(ϵ)),(9)\mathcal{G}^{-1}= \begin{pmatrix} 0 & \epsilon-\epsilon_k-i\delta\\ \epsilon-\epsilon_k+i\delta& 2i\delta F(\epsilon) \end{pmatrix}, \tag*{(9)}

with F(ϵ)=1F(\epsilon)=1 in the case of the initial density matrix being the vacuum state ∣0⟩⟨0∣|0\rangle\langle0| for the impurity modes. Finally, the interaction term takes the form:

Sint=∫dt[−Hint(φ+(t),φ+(t))+Hint(φ−(t),φ−(t))]=−gKL∑k∑p≠0V(p)∫−∞+∞dt×{φk,+∗(t)φk+p,+(t)[φp,+∗(t)+φ−p,+(t)]−φk,−∗(t)φk+p,−(t)[φp,−∗(t)+φ−p,−(t)]},(10)\begin{aligned} S_{int}&=\int dt\left[-H_{int}(\varphi_+(t),\varphi_+(t))+H_{int}(\varphi_-(t),\varphi_-(t))\right]\\ &=-\frac{g\sqrt{K}}{L}\sum_k\sum_{p\ne0}V(p)\int_{-\infty}^{+\infty}dt\\ &\quad\times\left\{\varphi^*_{k,+}(t)\varphi_{k+p,+}(t)\left[\varphi^*_{p,+}(t)+\varphi_{-p,+}(t)\right]\right.\\ &\qquad\left.-\varphi^*_{k,-}(t)\varphi_{k+p,-}(t)\left[\varphi^*_{p,-}(t)+\varphi_{-p,-}(t)\right]\right\}, \tag*{(10)} \end{aligned}

where V(p)=L∣p∣2πe−α∣p∣2V(p)=\sqrt{\frac{L|p|}{2\pi}}e^{-\alpha\frac{|p|}{2}}.

Finite-temperature Green's function

In order to compute the single-particle Green’s function (5), we reexpress it in the Keldysh formalism as:

G(y,t;x,0)=1L∑k1,k2ei(k2y−k1x)⟨φk2,+(t)φk1,+∗(0)⟩,(11)G(y,t;x,0)=\frac{1}{L}\sum_{k_{1},k_{2}}e^{i(k_{2}y-k_{1}x)}\langle\varphi_{k_{2},+}(t)\varphi_{k_{1},+}^{*}(0)\rangle, \tag*{(11)}

where the average is taken over the action obtained by summing (6), (8), (10), together with the vacuum condition F(ϵ)=1F(\epsilon)=1 in SimpS_{\mathrm{imp}}, while the bath is kept at finite temperature TT. A second-order perturbative expansion in the interaction SintS_{\mathrm{int}} followed by a reexponentiation of the result, in the spirit of the linked-cluster expansion[39], leads to the final expression (see App. § 1.2.1):

G(y,t;x,0)≈∫dk2πeik(y−x)−iϵkt−∫du[R(a)(u;k,T)+R(e)(u;k,T)]1−iut−e−iutu2,(12)G(y,t;x,0)\approx\int\frac{dk}{2\pi}e^{ik(y-x)-i\epsilon_{k}t-\int du[R^{(\mathrm{a})}(u;k,T)+R^{(\mathrm{e})}(u;k,T)]\frac{1-iut-e^{-iut}}{u^{2}}}, \tag*{(12)}

where we introduced the notation:

R(a)(u;k,T)=g2K(2π)2∫dp ∣p∣e−α∣p∣nB(ωp;T) δ(u−ϵk+p+ωp+ϵk),R^{(\mathrm{a})}(u;k,T)=\frac{g^{2}K}{(2\pi)^{2}}\int dp\,|p|e^{-\alpha|p|}n_{B}(\omega_{p};T)\,\delta(u-\epsilon_{k+p}+\omega_{p}+\epsilon_{k}),
R(e)(u;k,T)=g2K(2π)2∫dp ∣p∣e−α∣p∣(nB(ωp;T)+1)δ(u−ϵk+p−ωp+ϵk),(13)R^{(\mathrm{e})}(u;k,T)=\frac{g^{2}K}{(2\pi)^{2}}\int dp\,|p|e^{-\alpha|p|}\left(n_{B}(\omega_{p};T)+1\right)\delta(u-\epsilon_{k+p}-\omega_{p}+\epsilon_{k}), \tag*{(13)}

The interaction-induced correction term R(e)(u;k,T)R^{(\mathrm{e})}(u;k,T) accounts for the impurity’s phonon emission processes, and is present even at zero temperature, while R(a)(u;k,T)R^{(\mathrm{a})}(u;k,T) is a genuine finite-temperature effect, capturing the impurity’s phonon absorption processes.

In what follows, we discuss how the finite bath temperature corrects the zero-temperature behavior of the momentum-space Green’s function in the case of a quadratic dispersion relation ϵk=tik2\epsilon_{k}=t_{i}k^{2}; the latter is given by the expression:

G(k,t;β)=e−iϵkt−∫du[R(a)(u;k,T)+R(e)(u;k,T)]1−iut−e−iutu2,(14)G(k,t;\beta)=e^{-i\epsilon_{k}t-\int du[R^{(\mathrm{a})}(u;k,T)+R^{(\mathrm{e})}(u;k,T)]\frac{1-iut-e^{-iut}}{u^{2}}}, \tag*{(14)}

β=1/T\beta=1/T being the inverse temperature. At zero temperature, one has that R(a)(u;k,T)=0R^{(\mathrm{a})}(u;k,T)=0, so that G(k,t;β)G(k,t;\beta) satisfies:

G(k,t;β=+∞)=e−iϵkt−∫duR(e)(u;k,T=0)1−iut−e−iutu2,(15)G(k,t;\beta=+\infty)=e^{-i\epsilon_{k}t-\int duR^{(\mathrm{e})}(u;k,T=0)\frac{1-iut-e^{-iut}}{u^{2}}}, \tag*{(15)}

the latter quantity has been carefully examined in Ref. [2] and is known to display a transition in its time dependence at the critical momentum kc=vs/(2ti)k_{c}=v_{s}/(2t_{i}), separating a standard quasiparticle behavior for k>kck>k_{c}, associated with an exponential decay of ∣G(k,t;β=+∞)∣|G(k,t;\beta=+\infty)| as a function of time tt, from an infrared-dominated behavior for k<kck<k_{c}, signaled by a power-law decay of ∣G(k,t;β=+∞)∣|G(k,t;\beta=+\infty)| in time tt. We provide a detailed account of the aforesaid decays in App. § 1.2.2.

We study the finite-temperature corrections by evaluating the ratio:

G(k,t;β)G(k,t;+∞)=exp⁡{−g2K(2π)2∫dp∣p∣e−α∣p∣nB(ωp)[1−cos⁡[(ti(k+p)2−tik2−vs∣p∣)t](ti(k+p)2−tik2−vs∣p∣)2+1−cos⁡[(ti(k+p)2−tik2+vs∣p∣)t](ti(k+p)2−tik2+vs∣p∣)2]},(16)\frac{G(k,t;\beta)}{G(k,t;+\infty)} =\exp\left\{-\frac{g^{2}K}{(2\pi)^{2}}\int dp|p|e^{-\alpha|p|}n_{B}(\omega_{p})\left[\frac{1-\cos\left[(t_{i}(k+p)^{2}-t_{i}k^{2}-v_{s}|p|)t\right]}{(t_{i}(k+p)^{2}-t_{i}k^{2}-v_{s}|p|)^{2}}+\frac{1-\cos\left[(t_{i}(k+p)^{2}-t_{i}k^{2}+v_{s}|p|)t\right]}{(t_{i}(k+p)^{2}-t_{i}k^{2}+v_{s}|p|)^{2}}\right]\right\}, \tag*{(16)}

and estimating its large-time behavior (see App. § 1.2.3 for details). To this end, we observe that the dominant contributions come from the momentum values that fulfill the resonance condition ti(k+p)2−tik2−vs∣p∣=0t_{i}(k+p)^{2}-t_{i}k^{2}-v_{s}|p|=0 for the phonon absorption processes and the resonance condition ti(k+p)2−tik2+vs∣p∣=0t_{i}(k+p)^{2}-t_{i}k^{2}+v_{s}|p|=0 for the phonon emission processes. As detailed in App. § 1.2.3, we combine the knowledge of the poles of the aforementioned resonance conditions with the approximation of the delta function δ(x)=lim⁡t→+∞1−cos⁡(tx)πtx2\delta(x)=\lim_{t\to+\infty}\frac{1-\cos(tx)}{\pi tx^{2}}, to obtain large-time estimates of the behavior of the momentum-space Green’s function.

The IR regime k<vs/(2ti)k < v_s/(2t_i)

We start our analysis by characterizing the IR regime 0≤k<vs/(2ti)0 \leq k < v_s/(2t_i), thereby characterizing the large-time asymptotics of the finite-temperature correction to the single-particle Green’s function. The zeros of the resonance condition for the phonon absorption process occur now at p=0p = 0, p=(vs−2tik)/tip = (v_s - 2t_i k)/t_i and p=(vs+2tik)/tip = (v_s + 2t_i k)/t_i, while the phonon emission process is resonant only for p=0p = 0. The large-time asymptotic analysis, treated in detail in App. C, results ultimately in the following exponential suppression of the zero-temperature result:

∣G(k,t;β)G(k,t;+∞)∣≈exp⁡{−g2K(2π)2πt[1βvs(1vs−2tik+1vs+2tik)+e−α(vs−2tik)/titi(eβvs(vs−2tik)/ti−1)+e−α(vs+2tik)/titi(eβvs(vs+2tik)/ti−1)]}.(17)\left|\frac{G(k,t;\beta)}{G(k,t;+\infty)}\right| \approx\exp\left\{-\frac{g^2K}{(2\pi)^2}\pi t\left[\frac{1}{\beta v_s}\left(\frac{1}{v_s-2t_i k}+\frac{1}{v_s+2t_i k}\right)+\frac{e^{-\alpha(v_s-2t_i k)/t_i}}{t_i\left(e^{\beta v_s(v_s-2t_i k)/t_i}-1\right)}+\frac{e^{-\alpha(v_s+2t_i k)/t_i}}{t_i\left(e^{\beta v_s(v_s+2t_i k)/t_i}-1\right)}\right]\right\}. \tag*{(17)}

where the precise meaning of the asymptotic estimate shown in Eq. (17), as well as in the remainder of the main text, is discussed in App. D. As Eq. (17) manifestly shows, the effect of a finite-temperature consists in the appearance of a new set of zeros of the energy resonance conditions, namely the ones stemming from the first term in Eq. (16); the correction to the zero-temperature result displays moreover a linear-in-time behavior of the leading-order contribution to the integrals of Eq. (16) in the large-time limit. The net effect is an exponential suppression of the zero-temperature result in time, whose characteristic timescale decreases as temperature increases.

In Figure 1(a), we present some examples of the behavior of the ratio ∣G(k,t;β)∣/∣G(k,t;+∞)∣\lvert G(k,t;\beta)\rvert/\lvert G(k,t;+\infty)\rvert given by Eq. (16), showing that it displays a trend that is compatible with the predicted large-time exponential decay predicted by Eq. (17).

Ratio of finite-temperature and zero-temperature Green’s functions

Figure 1. (a) Ratio ∣G(k,t;β)∣/∣G(k,t;+∞)∣\lvert G(k,t;\beta)\rvert/\lvert G(k,t;+\infty)\rvert as a function of time tt for parameter values β=2.0\beta= 2.0, g=0.5g = 0.5, K=1.2K = 1.2, ti=0.25t_i = 0.25, vs=1.0v_s = 1.0, α=0.5\alpha= 0.5 with the choices of momentum k=0.5k = 0.5, 1.01.0, 1.51.5. (b) Ratio ∣G(k,t;β)∣/∣G(k,t;+∞)∣\lvert G(k,t;\beta)\rvert/\lvert G(k,t;+\infty)\rvert as a function of time tt for parameter values β=2.0\beta= 2.0, g=0.5g = 0.5, K=1.2K = 1.2, ti=0.25t_i = 0.25, vs=1.0v_s = 1.0, α=0.5\alpha= 0.5 with the choices of momentum k=2.5k = 2.5, 3.03.0, 3.53.5.

The quasiparticle regime k>vs/(2ti)k > v_s/(2t_i)

Moving on to the case where the momentum kk satisfies k>vs/(2ti)k > v_s/(2t_i), we are able to show that the resonance condition for the phonon absorption processes is satisfied for momenta p=0p = 0 and p=(vs+2tik)/tip = (v_s + 2t_i k)/t_i, while the phonon emission processes are resonant at p=0p = 0 and p=(2tik−vs)/tip = (2t_i k - v_s)/t_i. The final result for the ratio of the finite-temperature and zero-temperature Green’s functions is qualitatively analogous to the one found in the IR regime, despite the change in the distribution of the zeros of the resonance conditions. As shown in App. C, the explicit formula takes the form:

∣G(k,t;β)G(k,t;+∞)∣≈exp⁡{−g2K(2π)2πt[1βvs(12tik−vs+1vs+2tik)+e−α(2tik−vs)/titi(eβvs(2tik−vs)/ti−1)+e−α(vs+2tik)/titi(eβvs(vs+2tik)/ti−1)]}.(18)\left|\frac{G(k,t;\beta)}{G(k,t;+\infty)}\right| \approx\exp\left\{-\frac{g^2K}{(2\pi)^2}\pi t\left[\frac{1}{\beta v_s}\left(\frac{1}{2t_i k-v_s}+\frac{1}{v_s+2t_i k}\right)+\frac{e^{-\alpha(2t_i k-v_s)/t_i}}{t_i\left(e^{\beta v_s(2t_i k-v_s)/t_i}-1\right)}+\frac{e^{-\alpha(v_s+2t_i k)/t_i}}{t_i\left(e^{\beta v_s(v_s+2t_i k)/t_i}-1\right)}\right]\right\}. \tag*{(18)}

In the same spirit as for the IR regime, we present in Figure 1(b) a few instances of the behavior of the ratio ∣G(k,t;β)∣/∣G(k,t;+∞)∣\lvert G(k,t;\beta)\rvert/\lvert G(k,t;+\infty)\rvert given by Eq. (16), thereby providing evidence that it displays a trend that is compatible with the predicted large-time exponential decay predicted by Eq. (18).

The special momentum value k=vs/(2ti)k = v_s/(2t_i)

Finally, it is worth remarking that, for k=kck = k_c, the finite-temperature correction to the momentum-space Green’s function takes a different form from the ones already presented. Indeed, the result that we find for the finite-temperature correction to the momentum-space Green function takes the leading-order form:

∣G(k,t;β)∣∣G(k,t;+∞)∣≈exp⁡{−q2K(2π)2[23βvs2πtit3/2]}(19)\frac{\lvert G(k,t;\beta)\rvert}{\lvert G(k,t;+\infty)\rvert} \approx\exp\left\{-\frac{q^2K}{(2\pi)^2}\left[\frac{2}{3\beta v_s}\sqrt{\frac{2\pi}{t_i}}t^{3/2}\right]\right\} \tag*{(19)}

up to subleading linear-in-time corrections in the argument of the exponent (see App. § 1.2.3). Such an exceptional feature is responsible for the divergence of the finite-temperature coherence time τ\tau, to be introduced in Sec. IV, as k→kc±k \to k_c^\pm (see Figure 2).

Behavior of the inverse coherence time as a function of momentum

Figure 2. Behavior of the inverse coherence time 1/τ1/\tau as a function of momentum kk for parameter values β=2.0\beta= 2.0, g=0.5g = 0.5, K=1.2K = 1.2, ti=0.25t_i = 0.25, vs=1.0v_s = 1.0, α=0.5\alpha= 0.5. The dashed green line highlights the location of the momentum kc=vs/(2ti)k_c = v_s/(2t_i) on the xx-axis.

Discussion of the results

The leading behavior of the finite-temperature correction can be cast in the form:

∣G(k,t;β)G(k,t;+∞)∣≈exp⁡{−t/τ(k;g,vs,K,ti,T)}(20)\left\lvert\frac{G(k,t;\beta)}{G(k,t;+\infty)}\right\rvert\approx\exp\left\{-t/\tau(k;g,v_s,K,t_i,T)\right\} \tag*{(20)}

where the expression of the inverse lifetime can be read off from Eqs. (17)–(18). Thus, the effect of temperature can be interpreted through the introduction of a finite coherence time beyond which the zero-temperature properties of the impurity dynamics within the bath are fundamentally altered as a result of the presence of a finite temperature.

We start the discussion by plotting the inverse lifetime 1/τ1/\tau as a function of momentum kk in Figure 2 while fixing the other parameter values. We underline two features of the displayed curve, namely the divergence at k=kck = k_c and the emergent symmetry between the blue and orange curves in the neighborhood of kck_c. The divergence is readily explained in light of the non-analytic t3/2t^{3/2} time-dependence in the leading behavior of the finite-temperature correction in Eq. (16) when k=kck = k_c (see Eq. (19)), which formally causes τ−1\tau^{-1} to diverge as t→+∞t \to+\infty; on the other hand, the emergent symmetry of the behavior of 1/τ1/\tau in the left and right neighborhoods of kck_c traces back to the fact that the leading terms in the function 1/τ1/\tau as k→kc±k \to k_c^\pm depend on the distance ∣k−kc∣\lvert k-k_c\rvert, so that 1/τ1/\tau becomes to leading order a function of ∣k−kc∣\lvert k-k_c\rvert only as k→kc+k \to k_c^+ and k→kc−k \to k_c^-, respectively.

A further noteworthy observation is the effect of the impurity-bath interaction on the coherence time τ\tau: one has that τ∝g−2\tau\propto g^{-2}, as expected, since a larger value of gg is associated with a more effective hybridization of the impurity with the bath degrees of freedom.

In order to gain further physical insight into the dependence of τ−1\tau^{-1} on the microscopic parameters, we discuss the instance where the Luttinger liquid theory considered in the present work is an effective description of the low-energy excitations of the Lieb-Liniger model [40]. Since such model is Bethe-Ansatz solvable [41], we are able to relate the phenomenological parameters vsv_s and KK to the microscopic bath-bath interactions. Within such model, one defines the dimensionless parameter γ=mgint/ρ0\gamma= mg_{\mathrm{int}}/\rho_0, where mm is the bosons’ mass, gintg_{\mathrm{int}} the strength of their contact interaction and ρ0\rho_0 their number density. Then, the curve K(γ)K(\gamma) can be plotted and analytically estimated in the small and large γ\gamma limits, while Galilean invariance of the effective Luttinger liquid field theory allows to write the sound velocity as vs(γ)=πρ0/(mK(γ))v_s(\gamma)=\pi\rho_0/(mK(\gamma)).

In the strong-coupling regime, one has, to leading order, that K(γ)≈1K(\gamma)\approx1 and vs(γ)≈πρ0/m=:vFv_s(\gamma)\approx\pi\rho_0/m=:v_F (where the notation vFv_F underlines the fact that the combination πρ0/m\pi\rho_0/m equals the Fermi velocity of a free electron gas of density ρ0\rho_0 and mass mm) [40], so that the inverse coherence time becomes independent of the intra-bath interaction strength gintg_{\mathrm{int}} and converges to the value:

1τ=g24π[1βvF(1vF−2tik+1vF+2tik)+e−α(vF−2tik)/titi(eβvF(vF−2tik)/ti−1)+e−α(vF+2tik)/titi(eβvF(vF+2tik)/ti−1)](21)\frac{1}{\tau}=\frac{g^2}{4\pi}\left[\frac{1}{\beta v_F}\left(\frac{1}{v_F-2t_i k}+\frac{1}{v_F+2t_i k}\right)+\frac{e^{-\alpha(v_F-2t_i k)/t_i}}{t_i\left(e^{\beta v_F(v_F-2t_i k)/t_i}-1\right)}+\frac{e^{-\alpha(v_F+2t_i k)/t_i}}{t_i\left(e^{\beta v_F(v_F+2t_i k)/t_i}-1\right)}\right] \tag*{(21)}

for k<vs/(2ti)=πρ0/(2mti)k < v_s/(2t_i) = \pi\rho_0/(2mt_i), and to the value:

1τ=g24π[1βvF(12tik−vF+1vF+2tik)+e−α(2tik−vF)/titi(eβvF(2tik−vF)/ti−1)+e−α(vF+2tik)/titi(eβvF(vF+2tik)/ti−1)](22)\frac{1}{\tau} = \frac{g^2}{4\pi}\left[\frac{1}{\beta v_F}\left(\frac{1}{2t_i k-v_F}+\frac{1}{v_F+2t_i k}\right)+\frac{e^{-\alpha(2t_i k-v_F)/t_i}}{t_i\left(e^{\beta v_F(2t_i k-v_F)/t_i}-1\right)}+\frac{e^{-\alpha(v_F+2t_i k)/t_i}}{t_i\left(e^{\beta v_F(v_F+2t_i k)/t_i}-1\right)}\right] \tag*{(22)}

for the opposite scenario k>vs/(2ti)=πρ0/(2mti)k > v_s/(2t_i) = \pi\rho_0/(2mt_i).

In the weak-coupling regime, instead, one has that K(γ)≈π/γK(\gamma) \approx\pi/\sqrt{\gamma}, which results in vs(γ)≈ρ0γ/mv_s(\gamma) \approx\rho_0\sqrt{\gamma}/m [40]. In order to estimate the inverse coherence time, we consider separately the IR and quasiparticle regimes. When k<vs/(2ti)k < v_s/(2t_i), we provide an estimate of τ−1\tau^{-1} in the limit k≪vs(γ)/(2ti)=ρ0γ/(2mti)≪1k \ll v_s(\gamma)/(2t_i) = \rho_0\sqrt{\gamma}/(2mt_i) \ll1 by setting k=0k=0 to leading order in Eq. (17) and considering the leading term in the limit of small γ\gamma, so that the inverse coherence time reads in this limit:

1τ≈g2Km2πβρ02γ(23)\frac{1}{\tau} \approx\frac{g^2Km^2}{\pi\beta\rho_0^2\gamma} \tag*{(23)}

Thus, one observes that the temperature-induced inverse coherence time decreases, to a first approximation, as γ−1\gamma^{-1} with the intrabath interaction strength γ\gamma. On the other hand, when k>vs(γ)/(2ti)=ρ0γ/(2mti)k > v_s(\gamma)/(2t_i) = \rho_0\sqrt{\gamma}/(2mt_i), the leading contribution to Eq. (18) in the limit of small values of γ\gamma gives the result:

1τ≈g2Km(1+e−2αk)4πβtikρ0γ(24)\frac{1}{\tau} \approx\frac{g^2Km(1+e^{-2\alpha k})}{4\pi\beta t_i k\rho_0\sqrt{\gamma}} \tag*{(24)}

In such regime, we observe instead a γ−1/2\gamma^{-1/2} dependence of the inverse coherence time on the intrabath interaction strength γ\gamma, differently from what predicted in the IR regime. Nevertheless, in both cases, in the limit γ≪1\gamma\ll1, the decrease in the strength of the interactions within the chosen Lieb-Liniger bath causes the temperature-induced coherence time to shrink.

Conclusion and outlook

In this work, we have investigated the effect of a finite temperature in the Luttinger liquid bath on the dynamical signatures of a mobile impurity that interacts with it through a density-density contact term. Within a linked-cluster approximation in the framework of a Keldysh field-theoretical representation of the system’s dynamics, we have derived an analytical expression for the one-particle Green’s function of the impurity in presence of a finite-temperature 1D bath. Focusing on the momentum-space Green’s function, we unveiled that, regardless of the chosen parameter regime, the finite value of the temperature TT introduces a temperature- and momentum-dependent coherence time that sets the time-scale for coherent zero-temperature propagation of the impurity, and beyond which the aforesaid coherence is exponentially degraded.

The present results could be tested in experimental setups based on ultracold atoms, where the observation of impurity dynamics in 1D many-body systems is already within reach in state-of-the-art facilities [29, 30, 42, 43]. Moreover, while quantum gas microscopes naturally allow for direct access to the atoms’ density profiles [44, 45, 46, 47, 48, 49], there exist concrete theoretical proposals aiming at the measurement of single-particle Green’s functions in strongly-correlated quantum gases [50, 51, 52, 53]. We leave the discussion of the interplay of such techniques with our results as a promising direction of future research.

On the theoretical side, this work invites further investigations of the complex dynamics of mobile impurities in one-dimensional many-body environments. More specifically, it would be timely to discuss the role of the bath-impurity interaction range in the impurity’s dynamical properties, by replacing the standard contact interaction with a structured finite-range or, potentially, long-range one. Moreover, 1D quantum many-body systems may realize several instabilities of the Luttinger-liquid theory; the exploration of the features of impurity dynamics within different 1D phases of matter for the bath is therefore a potential avenue for future detailed investigations.

Acknowledgements: This work was supported by the Swiss National Science Foundation under grant number 200020-219400.

Derivation of the real-space Green's function

We construct an estimate of the single-particle Green’s function of the impurity in real space:

⟨0∣d^(y)e−iH^td^†(x)∣0⟩=1L∑k1,k2ei(k2y−k1x)⟨0∣d^k2(t)d^k1†(0)∣0⟩=1L∑k1,k2ei(k2y−k1x)⟨φk2,+(t)φk1,+∗(0)⟩(25)\langle0|\hat{d}(y)e^{-i\hat{H}t}\hat{d}^{\dagger}(x)|0\rangle= \frac{1}{L}\sum_{k_1,k_2}e^{i(k_2y-k_1x)}\langle0|\hat{d}_{k_2}(t)\hat{d}^{\dagger}_{k_1}(0)|0\rangle= \frac{1}{L}\sum_{k_1,k_2}e^{i(k_2y-k_1x)}\langle\varphi_{k_2,+}(t)\varphi^{*}_{k_1,+}(0)\rangle \tag*{(25)}

where the average is taken over the action obtained by summing:

Sbath=∑q≠0∫0+∞dt∫0+∞dt′∑α,β=cl,quφq,α∗(t)Gα,β−1φq,β(t′)(26)S_{\mathrm{bath}}=\sum_{q\ne0}\int_{0}^{+\infty}dt\int_{0}^{+\infty}dt'\sum_{\alpha,\beta=\mathrm{cl},\mathrm{qu}}\varphi^{*}_{q,\alpha}(t)\mathcal{G}^{-1}_{\alpha,\beta}\varphi_{q,\beta}(t') \tag*{(26)}
G−1=(0GA,bath−1(t,t′)GR,bath−1(t,t′)i1+wq1−wqδ(t)δ(t′))(27)\mathcal{G}^{-1}=\begin{pmatrix} 0 & G^{-1}_{A,\mathrm{bath}}(t,t') \\ G^{-1}_{R,\mathrm{bath}}(t,t') & i\frac{1+w_q}{1-w_q}\delta(t)\delta(t') \end{pmatrix} \tag*{(27)}
Simp=∑k∫0+∞dt∫0+∞dt′∑α,β=cl,quφk,α∗(t)Gα,β−1φk,β(t′)(28)S_{\mathrm{imp}}=\sum_k\int_{0}^{+\infty}dt\int_{0}^{+\infty}dt'\sum_{\alpha,\beta=\mathrm{cl},\mathrm{qu}}\varphi^{*}_{k,\alpha}(t)\mathcal{G}^{-1}_{\alpha,\beta}\varphi_{k,\beta}(t') \tag*{(28)}
G−1=(0GA,imp−1(t,t′)GR,imp−1(t,t′)i1+uk1−ukδ(t)δ(t′))(29)\mathcal{G}^{-1}=\begin{pmatrix} 0 & G^{-1}_{A,\mathrm{imp}}(t,t') \\ G^{-1}_{R,\mathrm{imp}}(t,t') & i\frac{1+u_k}{1-u_k}\delta(t)\delta(t') \end{pmatrix} \tag*{(29)}
Sint=∫dt[−Hint(φ+(t),φ+(t))+Hint(φ−(t),φ−(t))]=(30)S_{\mathrm{int}}=\int dt\left[-H_{\mathrm{int}}(\varphi_{+}(t),\varphi_{+}(t))+H_{\mathrm{int}}(\varphi_{-}(t),\varphi_{-}(t))\right]= \tag*{(30)}
=−gKL∑k∑p≠0V(p)∫−∞+∞dt{φk,+∗(t)φk+p,+(t)[φp,+∗(t)+φ−p,+(t)]−φk,−∗(t)φk+p,−(t)[φp,−∗(t)+φ−p,−(t)]}(31)=-\frac{g\sqrt{K}}{L}\sum_k\sum_{p\ne0}V(p)\int_{-\infty}^{+\infty}dt\left\{\varphi^{*}_{k,+}(t)\varphi_{k+p,+}(t)\left[\varphi^{*}_{p,+}(t)+\varphi_{-p,+}(t)\right]-\varphi^{*}_{k,-}(t)\varphi_{k+p,-}(t)\left[\varphi^{*}_{p,-}(t)+\varphi_{-p,-}(t)\right]\right\} \tag*{(31)}

where V(p)=L∣p∣2πe−α∣p∣2V(p)=\sqrt{\frac{L|p|}{2\pi}}e^{-\alpha\frac{|p|}{2}}, together with the vacuum conditions uk=0 ∀ku_k=0\ \forall k and wq=e−βωq ∀qw_q=e^{-\beta\omega_q}\ \forall q. The free correlators read:

⟨φk,±(t)φk′,∓∗(t′)⟩S0=δk,k′e−iϵk(t−t′)F(ϵk)∓12(32)\langle\varphi_{k,\pm}(t)\varphi^{*}_{k',\mp}(t')\rangle_{S_0}=\delta_{k,k'}e^{-i\epsilon_k(t-t')}\frac{F(\epsilon_k)\mp1}{2} \tag*{(32)}
⟨φk,±(t)φk′,±∗(t′)⟩S0=θ(t−t′)δk,k′e−iϵk(t−t′)F(ϵk)±12+θ(t′−t)δk,k′e−iϵk(t−t′)F(ϵk)∓12(33)\langle\varphi_{k,\pm}(t)\varphi^{*}_{k',\pm}(t')\rangle_{S_0}=\theta(t-t')\delta_{k,k'}e^{-i\epsilon_k(t-t')}\frac{F(\epsilon_k)\pm1}{2}+\theta(t'-t)\delta_{k,k'}e^{-i\epsilon_k(t-t')}\frac{F(\epsilon_k)\mp1}{2} \tag*{(33)}
⟨ϕq,±(t)ϕq′,∓∗(t′)⟩S0=δq,q′e−iωq(t−t′)F(ωq)∓12(34)\langle\phi_{q,\pm}(t)\phi^{*}_{q',\mp}(t')\rangle_{S_0}=\delta_{q,q'}e^{-i\omega_q(t-t')}\frac{F(\omega_q)\mp1}{2} \tag*{(34)}
⟨ϕq,±(t)ϕq′,±∗(t′)⟩S0=θ(t−t′)δq,q′e−iωq(t−t′)F(ωq)±12+θ(t′−t)δq,q′e−iωq(t−t′)F(ωq)∓12(35)\langle\phi_{q,\pm}(t)\phi^{*}_{q',\pm}(t')\rangle_{S_0}=\theta(t-t')\delta_{q,q'}e^{-i\omega_q(t-t')}\frac{F(\omega_q)\pm1}{2}+\theta(t'-t)\delta_{q,q'}e^{-i\omega_q(t-t')}\frac{F(\omega_q)\mp1}{2} \tag*{(35)}

where F(ϵ)=1+2nB(ϵ)F(\epsilon)=1+2n_B(\epsilon) and nB(ϵ)=(eβϵ−1)−1n_B(\epsilon)=(e^{\beta\epsilon}-1)^{-1} is the Bose-Einstein distribution function. The finite-temperature condition is applied solely to the bath degrees of freedom.

An expansion to second order yields the result:

⟨0∣d^(y)e−iH^td^†(x)∣0⟩=1L∑k1,k2ei(k2y−k1x)δk1,k2e−iϵk1t[1−∫du(R(a)(u;k1,T)+R(e)(u;k1,T))1−iut−e−iutu2]≈≈∫dk2πeik(y−x)−iϵkt−∫du[R(a)(u;k,T)+R(e)(u;k,T)]1−iut−e−iutu2(36)\begin{aligned} \langle0|\hat{d}(y)e^{-i\hat{H}t}\hat{d}^{\dagger}(x)|0\rangle=\frac{1}{L}\sum_{k_1,k_2}e^{i(k_2y-k_1x)}\delta_{k_1,k_2}e^{-i\epsilon_{k_1}t}\left[1-\int du\left(R^{(a)}(u;k_1,T)+R^{(e)}(u;k_1,T)\right)\frac{1-iut-e^{-iut}}{u^2}\right]\approx\\ &\approx\int\frac{dk}{2\pi}e^{ik(y-x)-i\epsilon_k t-\int du\left[R^{(a)}(u;k,T)+R^{(e)}(u;k,T)\right]\frac{1-iut-e^{-iut}}{u^2}} \tag*{(36)} \end{aligned}

where R(a/e)(u;k,T)R^{(a/e)}(u;k,T) is defined as:

R(a)(u;k,T)=g2K(2π)2∫dp ∣p∣e−α∣p∣nB(ωp;T)δ(u−ϵk+p+ωp+ϵk)(37)R^{(a)}(u;k,T)=\frac{g^2K}{(2\pi)^2}\int dp\,|p|e^{-\alpha|p|}n_B(\omega_p;T)\delta(u-\epsilon_{k+p}+\omega_p+\epsilon_k) \tag*{(37)}
R(e)(u;k,T)=g2K(2π)2∫dp ∣p∣e−α∣p∣(nB(ωp;T)+1)δ(u−ϵk+p−ωp+ϵk)(38)R^{(e)}(u;k,T)=\frac{g^2K}{(2\pi)^2}\int dp\,|p|e^{-\alpha|p|}\left(n_B(\omega_p;T)+1\right)\delta(u-\epsilon_{k+p}-\omega_p+\epsilon_k) \tag*{(38)}

and in the second equality we have reexponentiated the term in square brackets.

Review of the zero-temperature results for the momentum-space single-particle Green's function

Here, we present the large-time expressions of the single-particle Green’s function in the distinct dynamical regimes, in order to provide the zero-temperature expressions that finite-temperature corrections systematically alter. We start by realizing that the T=0T=0 momentum-space single-particle Green’s function can be read off from Eq. (36) as:

G(k,t;β=+∞)=e−iϵkt−∫duR(e)(u;k,T=0)1−iut−e−iutu2,(39)G(k,t;\beta=+\infty)=e^{-i\epsilon_k t-\int duR^{(e)}(u;k,T=0)\frac{1-iu t-e^{-iu t}}{u^2}}, \tag*{(39)}

whose modulus takes the form:

∣G(k,t;β=+∞)∣=exp⁡{−g2K(2π)2∫dp ∣p∣e−α∣p∣1−cos⁡[(ϵk+p+ωp−ϵk)t](ϵk+p+ωp−ϵk)2},(40)\left|G(k,t;\beta=+\infty)\right|=\exp\left\{-\frac{g^2K}{(2\pi)^2}\int dp\,|p|e^{-\alpha|p|}\frac{1-\cos\left[(\epsilon_{k+p}+\omega_p-\epsilon_k)t\right]}{(\epsilon_{k+p}+\omega_p-\epsilon_k)^2}\right\}, \tag*{(40)}

where we assume a quadratic impurity dispersion relation, i.e., ϵk=tik2\epsilon_k=t_i k^2, together with the linear dispersion relation ωq=vs∣q∣\omega_q=v_s|q| of the sound modes of the Luttinger liquid theory. The large-time limit of Eq. (40) is presented here below in the distinct dynamical regimes 0≤k<vs/(2ti)0\leq k<v_s/(2t_i) and k>vs/(2ti)k>v_s/(2t_i).

In the first case, i.e., 0≤k<vs/(2ti)0\leq k<v_s/(2t_i), one obtains the following power-law decay:

∣G(k,t;β=+∞)∣≈(2tik+vsαt)−g2K(2π)2(2tik+vs)2(vs−2tikαt)−g2K(2π)2(vs−2tik)2,(41)\left|G(k,t;\beta=+\infty)\right|\approx\left(\frac{2t_i k+v_s}{\alpha}t\right)^{-\frac{g^2K}{(2\pi)^2(2t_i k+v_s)^2}}\left(\frac{v_s-2t_i k}{\alpha}t\right)^{-\frac{g^2K}{(2\pi)^2(v_s-2t_i k)^2}}, \tag*{(41)}

associated with an instance of the Anderson’s orthogonality catastrophe in the one-dimensional bath. In the second case, namely k>vs/(2ti)k>v_s/(2t_i), the large-time estimate gives instead a leading linear-in-time term in the argument of the exponential of the expression of the zero-temperature Green’s function, thus resulting in the exponential decay:

∣G(k,t;β=+∞)∣≈exp⁡{−g2K(2π)2πe−α2tik−vstitit},(42)\left|G(k,t;\beta=+\infty)\right|\approx\exp\left\{-\frac{g^2K}{(2\pi)^2}\frac{\pi e^{-\alpha\frac{2t_i k-v_s}{t_i}}}{t_i}t\right\}, \tag*{(42)}

that is associated with the onset of quasi-particle behavior.

Asymptotic large-time behavior of the momentum-space Green's function at finite temperature

From Eq. (36), one can read off the correction of the modulus of the momentum-space Green’s function due to a finite temperature, which takes the form:

∣G(k,t;β)∣∣G(k,t;+∞)∣=exp⁡{−g2K(2π)2∫dp ∣p∣e−α∣p∣nB(ωp;T)[1−cos⁡[(ϵk+p−ϵk−ωp)t](ϵk+p−ϵk−ωp)2+1−cos⁡[(ϵk+p−ϵk+ωp)t](ϵk+p−ϵk+ωp)2]}.(43)\frac{|G(k,t;\beta)|}{|G(k,t;+\infty)|}=\exp\left\{-\frac{g^2K}{(2\pi)^2}\int dp\,|p|e^{-\alpha|p|}n_B(\omega_p;T)\left[\frac{1-\cos\left[(\epsilon_{k+p}-\epsilon_k-\omega_p)t\right]}{(\epsilon_{k+p}-\epsilon_k-\omega_p)^2}+\frac{1-\cos\left[(\epsilon_{k+p}-\epsilon_k+\omega_p)t\right]}{(\epsilon_{k+p}-\epsilon_k+\omega_p)^2}\right]\right\}. \tag*{(43)}

We analyze the large-time behavior of the above expression in the case of a quadratic dispersion relation, defined by the relation ϵk=tik2\epsilon_k=t_i k^2.

The case of a quadratic dispersion relation

If the dispersion relation is quadratic, Eq. (43) reads:

∣G(k,t;β)∣∣G(k,t;+∞)∣=exp⁡{−g2K(2π)2∫dp ∣p∣e−α∣p∣nB(ωp;T)[1−cos⁡[(ti(k+p)2−tik2−ωp)t](ti(k+p)2−tik2−ωp)2+1−cos⁡[(ti(k+p)2−tik2+ωp)t](ti(k+p)2−tik2+ωp)2]}=exp⁡{−g2K(2π)2(I1+I2)}.(44)\begin{aligned} \frac{|G(k,t;\beta)|}{|G(k,t;+\infty)|} &=\exp\left\{-\frac{g^2K}{(2\pi)^2}\int dp\,|p|e^{-\alpha|p|}n_B(\omega_p;T)\left[\frac{1-\cos\left[(t_i(k+p)^2-t_i k^2-\omega_p)t\right]}{(t_i(k+p)^2-t_i k^2-\omega_p)^2}+\frac{1-\cos\left[(t_i(k+p)^2-t_i k^2+\omega_p)t\right]}{(t_i(k+p)^2-t_i k^2+\omega_p)^2}\right]\right\}\\ &=\exp\left\{-\frac{g^2K}{(2\pi)^2}(I_1+I_2)\right\}. \tag*{(44)} \end{aligned}

The above expression is extremely challenging, yet, it is possible to make analytical progress by noticing that the integrals above will be dominated by the behavior of the function close to the poles of the arguments of the cosine function. Let us discuss separately the two integrals. The first one can be manipulated as:

I1=∫dp ∣p∣e−α∣p∣nB(ωp;T)1−cos⁡[(ti(k+p)2−tik2−ωp)t][ti(k+p)2−tik2−ωp]2=∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[(ti(k+p)2−tik2−vsp)t][ti(k+p)2−tik2−vsp]2+∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[(ti(k−p)2−tik2−vsp)t][ti(k−p)2−tik2−vsp]2=∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[(2tik−vs+tip)pt][(2tik−vs+tip)p]2+∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[(tip−2tik−vs)pt][(tip−2tik−vs)p]2.(45)\begin{aligned} I_1 &= \int dp\, |p|e^{-\alpha|p|}n_B(\omega_p;T)\frac{1-\cos\left[\left(t_i(k+p)^2-t_i k^2-\omega_p\right)t\right]}{\left[t_i(k+p)^2-t_i k^2-\omega_p\right]^2}\\ &= \int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[\left(t_i(k+p)^2-t_i k^2-v_s p\right)t\right]}{\left[t_i(k+p)^2-t_i k^2-v_s p\right]^2}\\ &\quad+\int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[\left(t_i(k-p)^2-t_i k^2-v_s p\right)t\right]}{\left[t_i(k-p)^2-t_i k^2-v_s p\right]^2}\\ &= \int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[\left(2t_i k-v_s+t_i p\right)pt\right]}{\left[\left(2t_i k-v_s+t_i p\right)p\right]^2}\\ &\quad+\int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[\left(t_i p-2t_i k-v_s\right)pt\right]}{\left[\left(t_i p-2t_i k-v_s\right)p\right]^2}. \tag*{(45)} \end{aligned}

Analogously, the second integral can be rewritten as:

I2=∫dp ∣p∣e−α∣p∣nB(ωp;T)1−cos⁡[(ti(k+p)2−tik2+ωp)t][ti(k+p)2−tik2+ωp]2=∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[(ti(k+p)2−tik2+vsp)t][ti(k+p)2−tik2+vsp]2+∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[(ti(k−p)2−tik2+vsp)t][ti(k−p)2−tik2+vsp]2=∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[(2tik+vs+tip)pt][(2tik+vs+tip)p]2+∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[(tip−2tik+vs)pt][(tip−2tik+vs)p]2.(46)\begin{aligned} I_2 &= \int dp\, |p|e^{-\alpha|p|}n_B(\omega_p;T)\frac{1-\cos\left[\left(t_i(k+p)^2-t_i k^2+\omega_p\right)t\right]}{\left[t_i(k+p)^2-t_i k^2+\omega_p\right]^2}\\ &= \int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[\left(t_i(k+p)^2-t_i k^2+v_s p\right)t\right]}{\left[t_i(k+p)^2-t_i k^2+v_s p\right]^2}\\ &\quad+\int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[\left(t_i(k-p)^2-t_i k^2+v_s p\right)t\right]}{\left[t_i(k-p)^2-t_i k^2+v_s p\right]^2}\\ &= \int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[\left(2t_i k+v_s+t_i p\right)pt\right]}{\left[\left(2t_i k+v_s+t_i p\right)p\right]^2}\\ &\quad+\int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[\left(t_i p-2t_i k+v_s\right)pt\right]}{\left[\left(t_i p-2t_i k+v_s\right)p\right]^2}. \tag*{(46)} \end{aligned}

To this end, it is now necessary to distinguish between two cases, namely, 0≤k<vs/(2ti)0 \le k < v_s/(2t_i) or k>vs/(2ti)k > v_s/(2t_i), as well as present a quick discussion of the special point k=vs/(2ti)k = v_s/(2t_i).

The case 0≤k<vs/(2ti)0 \le k < v_s/(2t_i)

In the present case, it is possible to notice that the first integrand in the final expression of I1I_1 is dominated by the poles of the argument of the cosine, which are given by p1∗=0p_1^* = 0 and p2∗=(vs−2tik)/tip_2^* = (v_s-2t_i k)/t_i, whereas the same consideration for the second integrand leads to the identification of the poles p3∗=p1∗=0p_3^* = p_1^* = 0 and p4∗=(2tik+vs)/tip_4^* = (2t_i k+v_s)/t_i. Notice how these poles fall either inside the domain of integration or at its left boundary. In a similar fashion, in the case of I2I_2, it is possible to notice that the only relevant poles of the cosine arguments of the integrands are the ones located at p5∗=p6∗=0p_5^* = p_6^* = 0.

To proceed with the large-time estimate, we approximate the full integrals in I1I_1 and I2I_2 by restricting the integration domain to the neighborhoods of the poles, by linearizing the argument of the cosine around the poles and then, after a change of variable, by using the delta-function representation:

lim⁡t→+∞1−cos⁡(tx)πtx2=δ(x).(47)\lim_{t\to+\infty}\frac{1-\cos(tx)}{\pi t x^2}=\delta(x). \tag*{(47)}

The estimate of I1I_1 proceeds as follows:

I1≈∫p≈0+dp pe−αpnB(ωp;T)1−cos⁡[(2tik−vs)pt][(2tik−vs)p]2+∫p≈(vs−2tik)/tidp pe−αpnB(ωp;T)1−cos⁡[(vs−2tik)(p−vs−2tikti)t][(vs−2tik)(p−vs−2tikti)]2+∫p≈0+dp pe−αpnB(ωp;T)1−cos⁡[(2tik+vs)pt][(2tik+vs)p]2+∫p≈(vs+2tik)/tidp pe−αpnB(ωp;T)1−cos⁡[(vs+2tik)(p−vs+2tikti)t][(vs+2tik)(p−vs+2tikti)]2≈πt2βvs(1vs−2tik+1vs+2tik)+πttie−αvs−2tiktieβvsvs−2tikti−1+πttie−αvs+2tiktieβvsvs+2tikti−1.(48)\begin{aligned} I_1 &\approx\int_{p\approx0^+}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[(2t_i k-v_s)pt\right]}{\left[(2t_i k-v_s)p\right]^2}\\ &\quad+\int_{p\approx(v_s-2t_i k)/t_i}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[(v_s-2t_i k)\left(p-\frac{v_s-2t_i k}{t_i}\right)t\right]}{\left[(v_s-2t_i k)\left(p-\frac{v_s-2t_i k}{t_i}\right)\right]^2}\\ &\quad+\int_{p\approx0^+}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[(2t_i k+v_s)pt\right]}{\left[(2t_i k+v_s)p\right]^2}\\ &\quad+\int_{p\approx(v_s+2t_i k)/t_i}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[(v_s+2t_i k)\left(p-\frac{v_s+2t_i k}{t_i}\right)t\right]}{\left[(v_s+2t_i k)\left(p-\frac{v_s+2t_i k}{t_i}\right)\right]^2}\\ &\approx\frac{\pi t}{2\beta v_s}\left(\frac{1}{v_s-2t_i k}+\frac{1}{v_s+2t_i k}\right)+\frac{\pi t}{t_i}\frac{e^{-\alpha\frac{v_s-2t_i k}{t_i}}}{e^{\beta v_s\frac{v_s-2t_i k}{t_i}}-1}+\frac{\pi t}{t_i}\frac{e^{-\alpha\frac{v_s+2t_i k}{t_i}}}{e^{\beta v_s\frac{v_s+2t_i k}{t_i}}-1}. \tag*{(48)} \end{aligned}

Analogously, one may estimate I2I_2 as:

I2≈∫p≈0+dp pe−αpnB(ωp;T)1−cos⁡[(2tik−vs)pt][(2tik−vs)p]2+∫p≈0+dp pe−αpnB(ωp;T)1−cos⁡[(2tik+vs)pt][(2tik+vs)p]2≈πt2βvs(1vs−2tik+1vs+2tik).(49)\begin{aligned} I_2 &\approx\int_{p\approx0^+}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[(2t_i k-v_s)pt\right]}{\left[(2t_i k-v_s)p\right]^2}+\int_{p\approx0^+}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[(2t_i k+v_s)pt\right]}{\left[(2t_i k+v_s)p\right]^2}\\ &\approx\frac{\pi t}{2\beta v_s}\left(\frac{1}{v_s-2t_i k}+\frac{1}{v_s+2t_i k}\right). \tag*{(49)} \end{aligned}

Plugging the obtained asymptotic estimates for I1I_1 and I2I_2 into Eq. (44), one can write:

∣G(k,t;β)∣∣G(k,t;+∞)∣≈exp⁡{−g2K(2π)2[πtβvs(1vs−2tik+1vs+2tik)+πttie−αvs−2tiktieβvsvs−2tikti−1+πttie−αvs+2tiktieβvsvs+2tikti−1]}.(50)\frac{\lvert G(k,t;\beta)\rvert}{\lvert G(k,t;+\infty)\rvert}\approx\exp\left\{-\frac{g^2K}{(2\pi)^2}\left[\frac{\pi t}{\beta v_s}\left(\frac{1}{v_s-2t_i k}+\frac{1}{v_s+2t_i k}\right)+\frac{\pi t}{t_i}\frac{e^{-\alpha\frac{v_s-2t_i k}{t_i}}}{e^{\beta v_s\frac{v_s-2t_i k}{t_i}}-1}+\frac{\pi t}{t_i}\frac{e^{-\alpha\frac{v_s+2t_i k}{t_i}}}{e^{\beta v_s\frac{v_s+2t_i k}{t_i}}-1}\right]\right\}. \tag*{(50)}
The case k>vs/(2ti)k>v_s/(2t_i)

The case k>vs/(2ti)k>v_s/(2t_i) can be treated in the same way as the opposed scenario by identifying the poles of the argument of the cosine functions appearing in the expressions of I1I_1 and I2I_2 such that they fall within the domain of integration. In this way, it is immediate to realize that I1I_1 is dominated by the behavior of the first integrand at the pole p1∗=0p_1^*=0 and of the second integrand at the poles p2∗=0p_2^*=0 and p3∗=(2tik+vs)/tip_3^*=(2t_i k+v_s)/t_i, while I2I_2 is dominated by the behavior of the first integrand at the pole p4∗=0p_4^*=0 and of the second integrand at the poles p5∗=0p_5^*=0 and p6∗=(2tik−vs)/tip_6^*=(2t_i k-v_s)/t_i.

The estimate of I1I_1 can be carried out as follows:

I1≈∫p≈0+dp pe−αpnB(ωp;T)1−cos⁡[(2tik−vs)pt][(2tik−vs)p]2+∫p≈0+dp pe−αpnB(ωp;T)1−cos⁡[(2tik+vs)pt][(2tik+vs)p]2+∫p≈vs+2tiktidp pe−αpnB(ωp;T)1−cos⁡[(vs+2tik)(p−vs+2tikti)t][(vs+2tik)(p−vs+2tikti)]2≈πt2βvs(12tik−vs+1vs+2tik)+πttie−αvs+2tiktieβvsvs+2tikti−1.(51)\begin{aligned} I_1\approx{}&\int_{p\approx0^+}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos[(2t_i k-v_s)pt]}{[(2t_i k-v_s)p]^2}\\ &+\int_{p\approx0^+}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos[(2t_i k+v_s)pt]}{[(2t_i k+v_s)p]^2}\\ &+\int_{p\approx\frac{v_s+2t_i k}{t_i}}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[(v_s+2t_i k)\left(p-\frac{v_s+2t_i k}{t_i}\right)t\right]}{\left[(v_s+2t_i k)\left(p-\frac{v_s+2t_i k}{t_i}\right)\right]^2}\\ &\approx\frac{\pi t}{2\beta v_s}\left(\frac{1}{2t_i k-v_s}+\frac{1}{v_s+2t_i k}\right)+\frac{\pi t}{t_i}\frac{e^{-\alpha\frac{v_s+2t_i k}{t_i}}}{e^{\beta v_s\frac{v_s+2t_i k}{t_i}}-1}. \tag*{(51)} \end{aligned}

Similarly, the estimate of I2I_2 is performed as:

I2≈∫p≈0+dp pe−αpnB(ωp;T)1−cos⁡[(2tik−vs)pt][(2tik−vs)p]2+∫p≈0+dp pe−αpnB(ωp;T)1−cos⁡[(2tik+vs)pt][(2tik+vs)p]2+∫p≈2tik−vstidp pe−αpnB(ωp;T)1−cos⁡[(2tik−vs)(p−2tik−vsti)t][(2tik−vs)(p−2tik−vsti)]2≈πt2βvs(12tik−vs+1vs+2tik)+πttie−α2tik−vstieβvs2tik−vsti−1.(52)\begin{aligned} I_2\approx{}&\int_{p\approx0^+}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos[(2t_i k-v_s)pt]}{[(2t_i k-v_s)p]^2}\\ &+\int_{p\approx0^+}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos[(2t_i k+v_s)pt]}{[(2t_i k+v_s)p]^2}\\ &+\int_{p\approx\frac{2t_i k-v_s}{t_i}}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos\left[(2t_i k-v_s)\left(p-\frac{2t_i k-v_s}{t_i}\right)t\right]}{\left[(2t_i k-v_s)\left(p-\frac{2t_i k-v_s}{t_i}\right)\right]^2}\\ &\approx\frac{\pi t}{2\beta v_s}\left(\frac{1}{2t_i k-v_s}+\frac{1}{v_s+2t_i k}\right)+\frac{\pi t}{t_i}\frac{e^{-\alpha\frac{2t_i k-v_s}{t_i}}}{e^{\beta v_s\frac{2t_i k-v_s}{t_i}}-1}. \tag*{(52)} \end{aligned}

The end result for the ratio ∣G(k,t;β)∣/∣G(k,t;+∞)∣\lvert G(k,t;\beta)\rvert/\lvert G(k,t;+\infty)\rvert reads therefore:

∣G(k,t;β)∣∣G(k,t;+∞)∣≈exp⁡{−g2K(2π)2[πtβvs(12tik−vs+1vs+2tik)+πttie−α2tik−vstieβvs2tik−vsti−1+πttie−αvs+2tiktieβvsvs+2tikti−1]}.(53)\frac{\lvert G(k,t;\beta)\rvert}{\lvert G(k,t;+\infty)\rvert}\approx\exp\left\{-\frac{g^2K}{(2\pi)^2}\left[\frac{\pi t}{\beta v_s}\left(\frac{1}{2t_i k-v_s}+\frac{1}{v_s+2t_i k}\right)+\frac{\pi t}{t_i}\frac{e^{-\alpha\frac{2t_i k-v_s}{t_i}}}{e^{\beta v_s\frac{2t_i k-v_s}{t_i}}-1}+\frac{\pi t}{t_i}\frac{e^{-\alpha\frac{v_s+2t_i k}{t_i}}}{e^{\beta v_s\frac{v_s+2t_i k}{t_i}}-1}\right]\right\}. \tag*{(53)}
The case k=vs/(2ti)k=v_s/(2t_i)

Here, we discuss the exceptional behavior occurring for k=vs/(2ti)k=v_s/(2t_i). To this end, we observe that:

lim⁡t→+∞32ti2πt31−cos⁡(tip2t)(tip2)2=δ(p).(54)\lim_{t\to+\infty}\frac{3}{2}\sqrt{\frac{t_i}{2\pi t^3}}\frac{1-\cos(t_i p^2t)}{(t_i p^2)^2}=\delta(p). \tag*{(54)}

By means of the identity presented in Eq. (54), we manipulate the first term in the last expression of Eq. (45) in the large-time limit as follows:

∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[(2tik−vs+tip)pt][(2tik−vs+tip)p]2∣k=kc=∫0+∞dp pe−αpnB(ωp;T)1−cos⁡[tip2t](tip2)2≈232πt3ti12βvs.(55)\left.\int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos[(2t_i k-v_s+t_i p)pt]}{[(2t_i k-v_s+t_i p)p]^2}\right|_{k=k_c} =\int_0^{+\infty}dp\,p e^{-\alpha p}n_B(\omega_p;T)\frac{1-\cos[t_i p^2t]}{(t_i p^2)^2}\approx\frac{2}{3}\sqrt{\frac{2\pi t^3}{t_i}}\frac{1}{2\beta v_s}. \tag*{(55)}

where we considered the large-time limit in the last step of the calculation. An analogous manipulation of the second term in the last expression of Eq. (46) leads to the same result as the one presented in Eq. (55). Thus, after treating the remaining terms as done in the other parameter regimes, the leading finite-temperature correction to the single-particle Green’s function reads in this case:

∣G(k,t;β)∣∣G(k,t;+∞)∣≈exp⁡{−g2K(2π)2[23βvs2πtit3/2+πt2βvs2+πtie−2αvst/tie2βvs2/ti−1]}.(56)\frac{\lvert G(k,t;\beta)\rvert}{\lvert G(k,t;+\infty)\rvert}\approx\exp\left\{-\frac{g^2K}{(2\pi)^2}\left[\frac{2}{3\beta v_s}\sqrt{\frac{2\pi}{t_i}}t^{3/2}+\frac{\pi t}{2\beta v_s^2}+\frac{\pi}{t_i}\frac{e^{-2\alpha v_s t/t_i}}{e^{2\beta v_s^2/t_i}-1}\right]\right\}. \tag*{(56)}

Eq. (56) thus displays an unexpected, novel leading-order term in the argument of the exponent, that scales non-analytically as t3/2t^{3/2}, rather than linearly with time tt. This feature, associated with the exceptional critical momentum value kck_c, is responsible for the divergence of the temperature-induced coherence time τ\tau while k→kc±k\to k_c^{\pm}.

Convergence of the asymptotic formulas

In the present section, we present additional details on the convergence of the l.h.s. of Eqs. (50)–(53), derived with a linked-cluster expansion and displayed explicitly in Eq. (43), to the analytical large-time approximations that are shown on the r.h.s. of the very same equations. Indeed, by denoting the full expression of Eq. (43) as ILCEI_{LCE} and the r.h.s. of Eqs. (50)–(53) as ILCE,estI_{LCE,est}, namely its asymptotic long-time estimate, we observe numerically in Figure 3 that the ratio log⁡ILCE/log⁡ILCE,est\log I_{LCE}/\log I_{LCE,est} displays a behavior that is consistent with its convergence to 1 for all explored parameter values. This highlights the precise meaning of the asymptotic approximations displayed in the main text of the manuscript, which consist indeed in the evaluation of the large-time leading term in the argument of the exponent of Eq. (43).

Ratio of logarithms of the linked-cluster expression and its asymptotic estimate as a function of time, shown for two parameter sets

Figure 3. Ratio log⁡ILCE/log⁡ILCE,est\log I_{LCE}/\log I_{LCE,est} as a function of time tt for parameter values k=1.0k=1.0, g=0.5g=0.5, K=1.2K=1.2, ti=0.3t_i=0.3, vs=1.0v_s=1.0, α=0.5\alpha=0.5, β=2.0\beta=2.0 (blue curve), k=0.5k=0.5, g=0.75g=0.75, K=1.1K=1.1, ti=0.25t_i=0.25, vs=1.1v_s=1.1, α=0.5\alpha=0.5, β=3.0\beta=3.0 (orange curve) and k=0.25k=0.25, g=0.25g=0.25, K=1.3K=1.3, ti=0.5t_i=0.5, vs=1.25v_s=1.25, α=0.5\alpha=0.5, β=1.5\beta=1.5 (green curve), respectively, compared to the value 1 (continuous red curve).

References

  1. [1]M. B. Zvonarev, V. V. Cheianov, and T. Giamarchi, Spin dynamics in a one-dimensional ferromagnetic bose gas, Phys. Rev. Lett. 99, 240404 (2007).DOI
  2. [2]A. Kantian, U. Schollwöck, and T. Giamarchi, Competing regimes of motion of 1d mobile impurities, Phys. Rev. Lett. 113, 070601 (2014).pubmed.ncbi.nlm.nih.gov/25170695
  3. [3]N. A. Kamar, A. Kantian, and T. Giamarchi, Mobile impurity in a two-leg bosonic ladder, Phys. Rev. Res. 7, 033136 (2025).DOI
  4. [4]L. Gotta and T. Giamarchi, Dirac impurity in a luttinger liquid, Phys. Rev. B 110, 165116 (2024).arxiv.org/abs/2404.16214
  5. [5]A. Lamacraft, Dispersion relation and spectral function of an impurity in a one-dimensional quantum liquid, Phys. Rev. B 79, 241105(R) (2009).DOI
  6. [6]M. Schecter, A. Kamenev, D. M. Gangardt, and A. Lamacraft, Critical velocity of a mobile impurity in one-dimensional quantum liquids, Phys. Rev. Lett. 108, 207001 (2012).
  7. [7]E. Burovski, V. Cheianov, O. Gamayun, and O. Lychkovskiy, Momentum relaxation of a mobile impurity in a one-dimensional quantum gas, Phys. Rev. A 89, 041601(R) (2014).
  8. [8]J.-T. Shen and S. Fan, Strongly correlated multiparticle transport in one dimension through a quantum impurity, Phys. Rev. A 76, 062709 (2007).DOI
  9. [9]T. Enss, V. Meden, S. Andergassen, X. Barnabé-Thériault, W. Metzner, and K. Schönhammer, Impurity and correlation effects on transport in one-dimensional quantum wires, Phys. Rev. B 71, 155401 (2005).DOI
  10. [10]K. Sacha and E. Timmarmans, Self-localized impurities embedded in a one-dimensional bose-einstein condensate and their quantum fluctuations, Phys. Rev. A 73, 063604 (2006).DOI
  11. [11]F. Lesage and H. Saleur, Correlations in one-dimensional quantum impurity problems with an external field, Nuclear physics B 490, 543 (1997).DOI
  12. [12]J. Bonart and L. F. Cugliandolo, From nonequilibrium quantum brownian motion to impurity dynamics in one-dimensional quantum liquids, Phys. Rev. A 86, 023636 (2012).DOI
  13. [13]F. D. M. Haldane, ’luttinger liquid theory’ of one-dimensional quantum fluids. i. properties of the luttinger model and their extension to the general 1d interacting spinless fermi gas, Journal of Physics C: Solid State Physics 14, 2585 (1981).
  14. [14]T. Giamarchi, Quantum physics in one dimension, Vol. 121 (Clarendon press, 2003).
  15. [15]T. Giamarchi, Quantum Physics in One Dimension, International series of monographs on physics, Vol. 121 (Oxford University Press, Oxford, 2004).
  16. [16]I. Bouchoule, R. Citro, T. Duty, T. Giamarchi, R. G. Hulet, M. K. Jałowski, E. Orignac, and B. Weber, Platforms for the realization and characterization of tomonaga–luttinger liquids, Nature Reviews Physics 7, 565 (2025).DOI
  17. [17]K. Schönhammer, Luttinger liquids: the basic concepts, Strong interactions in low dimensions , 93 (2004).
  18. [18]M. P. Fisher and L. I. Glazman, Transport in a one-dimensional luttinger liquid, in Mesoscopic electron transport (Springer, 1997) pp. 331–373.arxiv.org/abs/cond-mat/9610037
  19. [19]K.-V. Pham, M. Gabay, and P. Lederer, Fractional excitations in the luttinger liquid, Phys. Rev. B 61, 16397 (2000).DOI
  20. [20]A. H. Castro Neto and M. P. A. Fisher, Dynamics of a heavy particle in a luttinger liquid, Phys. Rev. B 53, 9713 (1996).pubmed.ncbi.nlm.nih.gov/9982527
  21. [21]V. V. Cheianov and M. B. Zvonarev, Nonunitary spin-charge separation in a one-dimensional fermion gas, Phys. Rev. Lett. 92, 176401 (2004).pubmed.ncbi.nlm.nih.gov/15169174
  22. [22]V. V. Cheianov and M. Zvonarev, Zero temperature correlation functions for the impenetrable fermion gas, Journal of Physics A: Mathematical and General 37, 2261 (2004).DOI
  23. [23]F. Massel, A. Kantian, A. J. Daley, T. Giamarchi, and P. Törmä, Dynamics of an impurity in a one-dimensional lattice, New Journal of Physics 15, 045018 (2013).
  24. [24]N. A. Kamar, A. Kantian, and T. Giamarchi, Dynamics of a mobile impurity in a two-leg bosonic ladder, Phys. Rev. A 100, 023614 (2019).DOI
  25. [25]P. W. Anderson, Infrared catastrophe in fermi gases with local scattering potentials, Phys. Rev. Lett. 18, 1049 (1967).
  26. [26]S. Qin, M. Fabrizio, and L. Yu, Impurity in a luttinger liquid: A numerical study of the finite-size energy spectrum and of the orthogonality catastrophe exponent, Phys. Rev. B 54, R9643(R) (1996).
  27. [27]A. Furusaki, Local perturbation in a tomonaga–luttinger liquid at g = 1/2 : orthogonality catastrophe, fermi-edge singularity, and local density of states, Phys. Rev. B 56, 9352 (1997).DOI
  28. [28]M. A. Caracanhas and R. G. Pereira, Mobile impurities and orthogonality catastrophe in two-dimensional vortex lattices, Phys. Rev. A 91, 013608 (2015).DOI
  29. [29]T. Fukuhara, A. Kantian, M. Endres, M. Cheneau, P. Schauß, S. Hild, P. Bellem, U. Schollwöck, T. Giamarchi, C. Gross, et al., Quantum dynamics of a mobile spin impurity, Nature Physics 9, 235 (2013).DOI
  30. [30]J. Catani, G. Lamporesi, D. Naik, M. Gring, M. Inguscio, F. Minardi, A. Kantian, and T. Giamarchi, Quantum dynamics of impurities in a one-dimensional bose gas, Phys. Rev. A 85, 023623 (2012).DOI
  31. [31]D. Cavazos-Cavazos, R. Senaratne, A. Kafle, and R. G. Hulet, Thermal disruption of a luttinger liquid, Nature Communications 14, 3154 (2023).
  32. [32]C. L. Kane and M. P. A. Fisher, Transport in a one-channel luttinger liquid, Phys. Rev. Lett. 68, 1220 (1992).pubmed.ncbi.nlm.nih.gov/10046110
  33. [33]A. E. Mattsson, S. Eggert, and H. Johannesson, Properties of a luttinger liquid with boundaries at finite temperature and size, Phys. Rev. B 56, 15615 (1997).DOI
  34. [34]G. Jüttner, A. Klümper, and J. Suzuki, Exact thermodynamics and luttinger liquid properties of the integrable tj model, Nuclear Physics B 487, 650 (1997).
  35. [35]G. Jüttner, A. Klümper, and J. Suzuki, The hubbard chain at finite temperatures: ab initio calculations of tomonaga-luttinger liquid properties, Nuclear Physics B 522, 471 (1998).
  36. [36]K. Le Hur, Electron lifetime in luttinger liquids, Phys. Rev. B 74, 165104 (2006).DOI
  37. [37]B. Horovitz, T. Giamarchi, and P. Le Doussal, Phase transitions for a collective coordinate coupled to luttinger liquids, Phys. Rev. Lett. 111, 115302 (2013).pubmed.ncbi.nlm.nih.gov/24074101
  38. [38]A. Kamenev, Field theory of non-equilibrium systems (Cambridge University Press, 2011).
  39. [39]G. D. Mahan, Many Particle Physics (Plenum, New York, 1981).
  40. [40]M. A. Cazalilla, R. Citro, T. Giamarchi, E. Orignac, and M. Rigol, One dimensional bosons: From condensed matter systems to ultracold gases, Rev. Mod. Phys. 83, 1405 (2011).DOI
  41. [41]E. H. Lieb and W. Liniger, Exact analysis of an interacting bose gas. i. the general solution and the ground state, Phys. Rev. 130, 1605 (1963).
  42. [42]F. Meinert, M. Knap, E. Kirilov, K. Jag-Lauber, M. B. Zvonarev, E. Demler, and H.-C. Nägerl, Bloch oscillations in the absence of a lattice, Science 356, 945 (2017).DOI
  43. [43]M. Horvath, S. Dhar, E. Wybo, D. Trypogeorgos, Y. Guo, M. Zvonarev, M. Knap, M. Landini, and H.-C. Nägerl, Observing dissipationless flow of an impurity in a strongly repulsive quantum fluid (2026), arXiv:2602.12320 [cond-mat.quant-gas].arxiv.org/abs/2602.12320
  44. [44]W. S. Bakr, J. I. Gillen, A. Peng, S. Fölling, and M. Greiner, A quantum gas microscope for detecting single atoms in a hubbard-regime optical lattice, Nature 462, 74 (2009).DOI
  45. [45]W. S. Bakr, A. Peng, M. E. Tai, R. Ma, J. Simon, J. I. Gillen, S. Foelling, L. Pollet, and M. Greiner, Probing the superfluid-to-mott insulator transition at the single-atom level, Science 329, 547 (2010).
  46. [46]J. F. Sherson, C. Weitenberg, M. Endres, M. Cheneau, I. Bloch, and S. Kuhr, Single-atom-resolved fluorescence imaging of an atomic mott insulator, Nature 467, 68 (2010).DOI
  47. [47]E. Haller, J. Hudson, A. Kelly, D. A. Cotta, B. Peaudecerf, G. D. Bruce, and S. Kuhr, Single-atom imaging of fermions in a quantum-gas microscope, Nature Physics 11, 738 (2015).DOI
  48. [48]L. W. Cheuk, M. A. Nichols, M. Okan, T. Gersdorf, V. V. Ramasesh, W. S. Bakr, T. Lompe, and M. W. Zwierlein, Quantum-gas microscope for fermionic atoms, Phys. Rev. Lett. 114, 193001 (2015).
  49. [49]C. Gross and W. S. Bakr, Quantum gas microscopy for single atom and spin detection, Nature Physics 17, 1316 (2021).DOI
  50. [50]M. Knap, A. Kantian, T. Giamarchi, I. Bloch, M. D. Lukin, and E. Demler, Probing real-space and time-resolved correlation functions with many-body ramsey interferometry, Phys. Rev. Lett. 111, 147205 (2013).pubmed.ncbi.nlm.nih.gov/24138270
  51. [51]H. H. Jen and D.-W. Wang, Theory of electromagnetically induced transparency in strongly correlated quantum gases, Phys. Rev. A 87, 061802(R) (2013).DOI
  52. [52]H. H. Jen and D.-W. Wang, Extracting dynamical green’s function of ultracold quantum gases via electromagnetically induced transparency, Journal of the Optical Society of America B 31, 2931 (2014).DOI
  53. [53]C. Weitenberg, L. Asteria, O. Carlsson, A. Bohrdt, and F. Grusdt, Protocols for a many-body phase microscope: From coherences and d-wave superconductivity to green’s functions, PRX Quantum 7, 033068 (2026).DOI

Paper details

vlnc:W_t2p4wqft
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents