Anomalous Doppler effect in superfluid and supersolid atomic gases

Abstract

We investigate the Doppler effect at zero temperature in superfluids with broken Galilean invariance and hosting permanent currents, with special focus on atomic gas platforms. We consider the case when Galilean invariance is broken explicitly (by an external periodic potential) or spontaneously, as it happens in a supersolid. In the first case, the presence of a stationary current affects the propagation of sound (fourth sound) via an anomalous Doppler term proportional to the density derivative of the superfluid fraction. In supersolids, where, according to Goldstone theorem, distinct sounds of hybrid superfluid and crystal nature can propagate, the Doppler effect can be very different for each sound, including the possibility of being negative for the lower phonon branch. We obtain analytical predictions within the hydrodynamic theories for superfluids and supersolids, which are compared with the numerical results of time-dependent simulations for weakly interacting atomic Bose-Einstein condensates.

In a classical fluid, moving at a certain velocity $v_{f}$, the sound speed $c_{0}$ is modified by the kinematic Doppler shift as $c=c_{0}\pm v_{f}$, depending on whether the sound propagates parallel ($+$) or antiparallel ($-$) to the velocity $v_{f}$.

The Doppler effect has been predicted to behave quite differently in a superfluid as a consequence of the fact that the normal and the superfluid components can move at different velocities. The problem was first addressed a long time ago by Khalatnikov [1] for superfluid helium at finite temperature, where the motion of the liquid is described by Landau’s two-fluid hydrodynamics. Stimulated by early experimental results on the Doppler effect in the propagation of fourth sound in Helium [2], further theoretical studies have pointed out the occurrence of non-trivial, anomalous Doppler shifts in ⁴He [3, 4, 5], ⁴He$-^{3}$He mixture [6], as well as in superfluid ³He [7]. To our knowledge, however, a clear experimental confirmation of the novel theoretical predictions for the anomalous Doppler effect in liquid helium configurations is still missing. The availability of alternative platforms to investigate the effects of superfluidity, based on ultracold atomic gases [8, 9, 10], including the recent experimental observation of supersolidity [11, 12, 13, 14, 15] (see also [16] and references therein), is opening new perspectives in the field.

In this Letter, we focus on density modulated configurations, where the superfluid density can differ significantly from the total average density even at zero temperature. Density modulated phases in ultracold atomic gases can result from an external periodic potential (optical lattices [17]) or from the spontaneous breaking of translational symmetry, yielding supersolidity. The propagation of sound and the occurrence of collective oscillations, in the absence of steady currents, have already been the object of experimental investigations in both cases (see [18, 19] and [20, 21, 22], respectively). Furthermore, the Doppler effect was observed in a uniform Bose-Einstein condensed gas confined in a ring and used to measure the quanta of circulation characterizing the velocity of the superfluid flow [23].

In the presence of an external periodic potential the normal component of the fluid is locked, only the superfluid component being able to move. In this case sound is also called fourth sound [24, 25]. In supersolid configurations both the superfluid and the normal (crystal) components can instead move. As predicted in the seminal paper by Andreev and Lifshtz [26], Goldstone modes, of crystalline and superfluid nature, are expected to occur in a supersolid (see [27] for the counting of the modes in the supersolid phase). The different sounds are expected to react differently to the presence of relative motion between the two components. The purpose of our work is to determine the Doppler effect in both cases, comparing the predictions of hydrodynamic theory with the results of dynamic numerical simulations currently available in ultracold atomic gases.

For the sake of simplicity, we consider a linear tube geometry of length $L$, with periodic boundary conditions, where we can use one-dimensional hydrodynamic theories. In cold atomic gases, quasi-one-dimensional configurations hosting a permanent current can be experimentally realized using ring-shaped trapping potentials with sufficiently large radii.

Doppler shift of fourth sound.

Fourth sound characterizes the propagation of sound in a superfluid subject to an external potential locking the motion of the normal component. Experiments on fourth sound are available both in liquid helium, where the superfluid moves within a porous medium [25] and in ultracold atomic gases [19] where the pinning of the normal component is provided by an optical lattice.

Doppler effect in fourth sound was first observed in superfluid Helium [2, 28]. According to Landau’s theory the fluid velocity $v_{f}$ – defined as the current divided by the mass density $\rho$ – is $v_{f}=v_{s}\rho_{s}/\rho$, with $\rho_{s}$ and $v_{s}$ the superfluid density and the superfluid velocity, respectively. In the pioneering work [2] it was suggested that fourth sound is Doppler shifted by this amount. However, a more detailed analysis based on finite temperature two-fluid hydrodynamics has shown that this prediction is in general not correct [5].

Here we derive a straightforward result for the Doppler effect of fourth sound at zero temperature. Assuming that the stationary superfluid velocity is small, the energy density of the system, in the frame where the normal component is at rest, is given by the expansion $\epsilon(\rho,v_{s})=\epsilon_{0}(\rho)+(1/2)\rho_{s}v^{2}_{s}$. The superfluid velocity is irrotational and can be written as $v_{s}=\hbar\partial_{x}\phi/m$, where the macroscopic phase $\phi$ is canonically conjugated to the density . The Hamiltonian equations of motion for $\rho$ and $v_{s}$ then take the form of the following (collisionless) hydrodynamic (HD) equations:

$$\partial_{t}\rho+\partial_{x}(\rho_{s}v_{s}) = 0$$ (1)

$$\partial_{t}v_{s}+\partial_{x}\left({\mu_{0}}+\frac{1}{2}\frac{\partial\rho_{s}}{\partial\rho}v_{s}^{2}\right) = 0$$ (2)

where $\mu_{0}(\rho)=\partial\epsilon_{0}/\partial\rho$ is the chemical potential calculated at $v_{s}=0$. The two HD equations correspond to the atom number conservation and to the presence of the superfluid long-range order. It is worth noticing that the quantities entering the above equations are coarse grained and correspond to averages over distances larger than the period of the periodic potential, but, of course, smaller than the phonon wavelength.

Sound propagation – corresponding to the superfluid Goldstone mode – is described by linearizing the above equations around the uniform equilibrium values $\rho_{0}$ and $v_{s}^{0}$ as $\rho(x,t)=\rho_{0}+\delta\rho(x,t)$ and $v_{s}(x,t)=v_{s}^{0}+\delta v_{s}(x,t)$. In this way we find that the excitations obey the phononic dispersion relation $\omega(q)=c|q|$, with $q$ the wave vector of the sound wave and $c$ the speed of sound. The latter is given by $c^{\pm}=c_{0}\pm\Delta c$, where $c_{0}=\sqrt{f_{s}\kappa^{-1}}$ is the value of the sound velocity in absence of current, with $f_{s}=\rho_{s}/\rho$ the superfluid fraction and $\kappa^{-1}=\rho\partial_{\rho}\mu_{0}$ the inverse compressibility. Finally, the Doppler shift reads

$$\Delta c=\frac{\partial\rho_{s}}{\partial\rho}v_{s}^{0}=\left(f_{s}+\rho\frac{\partial f_{s}}{\partial\rho}\right)v_{s}^{0}$$ (3)

The expressions given by Eqs. (1-3) hold for both Bose and Fermi superfluids at zero temperature, the Doppler shift (Eq.(3)) being in general different from the kinematic expectation $f_{s}v_{s}^{0}$ as a consequence of the density dependence of the superfluid fraction. Similar results were obtained in [3] in the case of superleak helium configurations, with special focus on finite size effects, and in [30, 31] in the case of atomic Bose-Einstein condensates in optical lattices, after linearization of the time-dependent Gross-Pitaevskii equation (GPE). The superfluid fraction in these works has to be interpreted in terms of effective tunneling or effective mass, respectively.

The result Eq.(3) can be easily generalized to the case of an optical lattice moving with velocity $v_{n}^{0}$, by a simple Galilean transformation, i.e., $\Delta c=v_{n}^{0}+(\partial\rho_{s}/\partial\rho)(v_{s}^{0}-v_{n}^{0})=(\partial\rho_{s}/\partial\rho)v_{s}^{0}+(\partial\rho_{n}/\partial\rho)v_{n}^{0}$, with $\rho_{n}=\rho-\rho_{s}$.

Figure 1: a) An example of transversely integrated fluid density held in an optical lattice of strength $s=3$. b) Doppler shift relative to $v_{s}$ of the fourth sound in an optical lattice as a function of the strength parameter $s$. Superfluid fraction $f_{s}$ is estimated with Leggett’s upper bound [32] while the derivative $\partial\rho_{s}/\partial\rho$ is calculated numerically.

We verify the validity of the hydrodynamic result Eq.(3), when applied to a Bose-Einstein condensed gas, by numerically solving the time dependent Gross-Pitaevskii equation in presence of an optical lattice. We consider realistic parameters for a gas of $N=2\cdot 10^{5}$ ⁸⁷Rb atoms confined in a tube of length $L=96\mu\text{m}$ along $x$, subject to periodic boundary conditions, and radially confined by a harmonic potential of the form $1/2m\omega_{\perp}^{2}r^{2}$ with $r^{2}=y^{2}+z^{2}$, $\omega_{\perp}=2\pi\cdot 150$Hz, with $\sqrt{\hbar/m\omega_{\perp}}\ll L$. The gas is subject to an external periodic potential along $x$ of the form $V_{ext}=sE_{R}\cos(qx)$, where $E_{R}=\hbar^{2}q^{2}/2m$ is the recoil energy, and the dimensionless parameter $s$ gives the strength of the external potential. The gas at zero temperature forms a Bose-Einstein condensate, characterized by the healing length $\xi=\hbar/(m\sqrt{2\mu_{0}})$, which represents the length scale on which the condensate can react to an external potential. We choose the period of the lattice $d=2\pi/q=4\mu$m to be much larger than the healing length: $d\approx 16.7\xi$. In the opposite regime, $d\ll\xi$, the superfluid fraction becomes density independent and the kinematic Doppler shift is recovered [33]. An example of the transversely integrated GP ground state density ${\rho_{GP}}(x)$ is reported in Fig.(1a). A permanent current providing the superfluid velocity $v_{s}^{0}=47.8\mu\text{m/s}$, satisfying the proper boundary condition, is applied to the system. Following a previously proposed method for probing excitations [23, 34], the system is prepared in a stationary state in presence of a small additional potential $\lambda\cos(kx)$, which is suddenly removed. Since we are interested in the lowest energy modes we take the smallest possible value $k=2\pi/L$, corresponding to a wavelength much larger than the transverse size of the gas. Within linear response theory, the time evolution of the average $\langle\cos(kx)\rangle$ contains frequencies of the modes excited by the perturbation. In the absence of current, the signal contains a single frequency, $\omega$ from which the speed of sound is determined as $c=\omega/k$. The current induces a mode splitting, ($\omega^{+}\neq\omega^{-}$), from which the Doppler shift $2\Delta c=(\omega^{+}-\omega^{-})/k$ can be extracted.

In order to compare the numerical result with Eq.(3) we need the derivative of the superfluid density with respect to the average density $\rho$. We obtain it by determining the ground state of the system for different atom number and extracting the superfluid density by means of the Leggett estimate [32, 35], which provides a very accurate estimate in the present configuration (see, e.g., [36]). In Fig.(1) we report the numerically obtained Doppler shift (dots) along with the HD prediction (solid line) and the kinematic Doppler shift (dashed line). The HD result perfectly accounts for the Doppler shift, while the kinematic expression significantly underestimates it. We emphasize that in our configuration the density $\rho$ corresponds to the transverse integrated coarse grained density and the inverse compressibility is the corresponding one dimensional value.

It is worth mentioning that an experimental verification of the hydrodynamic relation between $c_{0}$ and $f_{s}$ has been very recently obtained in [19]. The measurement of the shift $\Delta c$ will be an important proof of the existence of a zero temperature anomalous Doppler effect. In cold gases one could easily change the ratio $d/\xi$ in order to explore such an effect.

Doppler effect in a supersolid.

Supersolids are systems which present both superfluid and crystalline long-range order, which, consequently, show two classes of phononic (Nambu-Goldstone) modes [27], due to the simultaneous spontaneous breaking of $U(1)$ symmetry and Galilean invariance. With respect to the case of fourth sound, the HD theory of supersolid includes additional equations for the current (following from momentum conservation) and for the lattice displacement $u_{x}\coloneqq\partial_{x}u$, where $u$ is the lattice dispacement (following from crystalline long range order), see, e.g., [37, 38, 39, 40]:

$$\partial_{t}\rho+\partial_{x}j = 0$$ (4)

$$\partial_{t}j+\partial_{x}\left(p+\rho_{n}v_{n}^{2}+\rho_{s}v_{s}^{2}\right) = 0$$ (5)

$$\partial_{t}v_{s}+\partial_{x}(v_{n}v_{s}+\mu) = 0$$ (6)

$$\partial_{t}u_{x}+\partial_{x}(v_{n}u_{x}-v_{n}) = 0,$$ (7)

where we remind that $\rho_{n}$, $v_{n}$ ($\rho_{s}$, $v_{s}$) are the normal (superfluid) density and velocity in a two fluid model, $\rho=\rho_{n}+\rho_{s}$ is the total density and $j=\rho_{n}v_{n}+\rho_{s}v_{s}$ is the total current. The quantities $p$ and $\mu$ are the pressure and the chemical potential in the presence of a stationary flow. Since we are interested in the linear expansion of the HD equation the constitutive relations can be written as [42, 40]:

$$\delta\mu = {1\over{\rho\kappa}}\delta\rho+\gamma u_{x}+{\rho_{s}^{\prime}\over 2}\delta w^{2}-{1\over 2}\delta v_{n}^{2},$$ (8)

$$\delta p = \left({1\over\kappa}-\gamma\right)\delta\rho+\left(\gamma\rho-\lambda\right)u_{x}+\frac{\rho_{n}-\rho\rho_{n}^{\prime}}{2}\delta w^{2},$$ (9)

$$\delta\rho_{n} = \rho_{n}^{\prime}\delta\rho+\rho_{n}^{u}u_{x},$$ (10)

where we introduce the relative velocity $w\equiv v_{n}-v_{s}$, $\gamma$ the strain-density constant, $\lambda$ the elastic constant. For the sake of notation we define $\rho_{i}^{\prime}\coloneqq\partial\rho_{i}/\partial\rho$, with $i=n,s$ and $\rho_{n}^{u}\coloneqq\partial\rho_{n}/\partial u_{x}$. The dependence of $\rho_{n}$ on the velocity difference $w$ is quadratic and can be consequently neglected at the linear order.

In order to calculate the speeds of sound we linearize Eqs. (4-7) around their stationary values $\{n^{0},v_{s}^{0},v_{n}^{0},u_{x}^{0}=0\}$ and determine the dispersion relation to the lowest order in the stationary velocities. We find two phononic modes $\omega_{1,2}=c_{1,2}|q|$, where the speeds of sound can be written as $c_{1,2}^{\pm}=c^{0}_{1,2}\pm(v_{s}^{0}+w\delta_{1,2}$). The expressions for two speeds of sound at rest $c^{0}_{1,2}$ in the supersolid phase have been already derived in a number of previous works [43, 40, 44, 45] (see below).

When $w=0$, i.e., the superfluid and normal components move together as a single fluid. Both sounds are simply shifted by the Doppler kinematic effect, fixed by the fluid velocity $v_{s}^{0}=v_{n}^{0}$. When $w\neq 0$ we instead find the additional contribution $w\delta_{1,2}$, with

$$\delta_{1,2}=\frac{\frac{\rho_{n}^{u}}{2\rho_{n}}\left[\left(c^{0}_{1,2}\right)^{2}-c_{\kappa}^{2}+\gamma\right]+2(c_{1,2}^{0})^{2}-\left(1+f_{s}\right)c_{\kappa}^{2}-\rho_{n}^{\prime}\beta+\gamma\left[1+\frac{\rho^{\prime}_{n}}{\rho_{n}}(\rho-2\rho_{n})\right]}{2\left(c^{0}_{1,2}\right)^{2}-c_{\kappa}^{2}-\beta},$$ (11)

where $c_{\kappa}=\kappa^{-1/2}$ and $\beta=\lambda/\rho_{n}$ and the speeds of sound at rest are $(c_{1,2}^{0})^{2}=(A\pm\sqrt{A^{2}-4B})/2$ with $A=c_{\kappa}^{2}+\beta-2\gamma$, $B=f_{s}(\beta c_{\kappa}-\gamma^{2})$. In the limit of an incompressible lattice, i.e. as $\beta\to\infty$, the Doppler shift approaches the value $\rho_{s}^{\prime}v_{s}^{0}+\rho^{\prime}_{n}v_{n}^{0}$, which coincides with the result obtained above for fourth sound. Moreover, approaching the crystal phase, i.e., $\rho_{n}\rightarrow\rho$ ($\rho_{n}^{u}\rightarrow 0,\;\rho^{\prime}_{n}\rightarrow 1$), one finds $\delta_{1,2}\rightarrow 1$, confirming the intuitive result that in the crystal phase the sound speed simply feels a kinematic Doppler shift, given by the normal velocity $v_{n}$.

In general Eq.(11) predicts that the Doppler effect affects differently the two sound speeds. The result corresponds to the zero temperature analogous of the predictions obtained for the first and second sound velocities in superfluid Helium-4 [1, 4].

We explore the HD prediction for the Doppler effect in the case of a Bose-Einstein condensed gas of atoms interacting with dipolar forces, where successful experimental achievements of supersolid configurations have been obtained in the last few years (see, e.g., [16] and references therein). Theoretically, the supersolid phase in such systems is stabilized by the inclusion of the so-called Lee-Huang-Yang correction to the mean-field approach, yielding an extended version of the Gross-Pitaevskii equation [46], which has been already employed to predict the sound velocities in the absence of persistent currents. Within such a formalism, the speeds of sound can be extracted following the protocol previously described for the fourth sound of a BEC in an optical lattice. In the absence of a persistent current, we already successfully applied the protocol to the supersolid phase of ultra-cold dipolar gases in [34].

In particular, we consider $N=1.6\cdot 10^{5}$ ¹⁶⁴Dy atoms confined, as before, in a tube of $L=96\mu$m length with periodic boundary conditions and with a transverse potential given by $m\omega_{\perp}^{2}r^{2}/2$, where $\omega_{\perp}=2\pi\cdot 100\text{Hz}$. Dipoles are aligned by an external magnetic field along the $z$ direction. The system can be in a homogeneous superfluid phase, in a supersolid or in the independent cluster phase – in which the superfluid long-range order along $x$ is lost – depending on the parameter $\epsilon_{dd}$, corresponding to the ratio between the dipolar and the contact 2-body interaction strength (see, e.g., [47, 13, 14, 15]).

An example of transversely integrated density $\rho_{GP}(x)$ is shown in Fig.(2a). The density modulations look very similar to the ones obtained in an optical lattice (see Fig.(1a)), but their origin and dynamics are profoundly different. We prepare the initial state of the system by evolving it in imaginary time in the moving frame with velocity $v_{n}$, by adding the term $-v_{n}P_{x}$ to the extended Gross-Pitaevski equation for the order parameter, with $P_{x}$ the momentum operator along $x$. The initial stationary state is characterised by a momentum $\langle P_{x}\rangle=m(f_{n}v_{n}+f_{s}v_{s})$, with $f_{n}=\rho_{n}/\rho$ the normal fraction. The superfluid velocity is quantized in the tube and exhibits a first jump to a non-zero value at $v_{n}^{\text{cr}}=\hbar\pi/(mL)=12.7\mu\text{m/s}$ (see also [48, 49] for the current generation in a ring) as shown in Fig.(4a).

The supersolid phase at rest exhibits two phononic modes, each of the two split into two in the presence of a current. As a consequence, four different frequencies with linear dispersion $\omega=c|q|=c\cdot 2\pi/L$ are expected to occur, with respect to two frequencies occurring in the superfluid and in the crystal phase. An example of Doppler splitting of two sound modes in the supersolid phase is presented in Fig.(2b). Both sounds experience an abrupt change of their speed at $v_{n}/v_{n}^{\text{cr}}=1$, i.e., when the phase of the order parameter hosts a quantum of circulation.

Figure 2: a) Transversely integrated density of a supersolid at $\epsilon_{dd}=1.398$. b) Splitting of the first (upper) and second sound (lower panel) into four phonon modes as a function of normal velocity $v_{n}$ for the system presented in a). Within the shaded area the Doppler shift of the second sound becomes negative.

In order to make a quantitative comparison between the GP simulation and the predictions of the linearized hydrodynamic equations we need to extract the parameters entering in Eq.(11). The superfluid velocity is extracted from the relation between $\langle P_{x}\rangle$ and the normal velocity, where the superfluid fraction $f_{s}$ is estimated via the Leggett formula [35], while $\rho^{\prime}_{n}$ and $\rho^{\prime}_{s}$ are determined numerically by marginally changing the total atom number $N$. The strain-density coupling has been shown to be very small for the supersolid phase of a dipolar gas [45] and will be neglected in our analysis. Finally, the term $\rho_{n}^{u}$ is not easily estimated within the eGPE framework and will be used here as a free parameter. Finally, the most relevant parameter $\delta_{i}$, accounting for the anomalous Doppler effect, is extracted by fitting the Doppler shift $\frac{1}{2}\left(c_{i}^{+}-c_{i}^{-}\right)$, calculated in the dynamic simulation.

In Fig.(3) we show the comparisons between the HD prediction (lines) and the numerical results (symbols) within eGPE. The solid lines represents the HD prediction for $\rho_{n}^{u}=0$. The agreement is already reasonable. Furthermore Eq.(11) suggests that $\delta_{2}$ is more sensitive than $\delta_{1}$ to the value of $\rho_{n}^{u}$, since, close to the transition to the superfluid phase, $c_{1}^{0}$ approaches $c_{\kappa}$ [44, 34, 50]. We therefore extract $\rho_{n}^{u}$ by fitting the quantity $\delta_{2}$ using Eq.(11). By inserting the extracted values for $\rho_{n}^{u}$ in Eq.(11) we find an almost perfect match also for $\delta_{1}$ (see the figure). In the inset we show the extracted value of $\rho_{n}^{u}=\partial_{u_{x}}\rho_{n}$, where for completeness we report also the value extracted by fitting $\delta_{1}$. We can therefore conclude that the such quantity is not negligible for a dipolar supersolid and can be estimated by inverting Eq.(11) using the numerically extracted values for the Doppler shifts. Our analysis also confirms that $\gamma$ is not important for describing the supersolid phase of dipolar gases [45].

Figure 3: Anomalous Doppler shift of first and second sound in the supersolid phase. Results of simulations (markers) are compared with hydrodynamic model (solid lines) where the unknown derivative $\rho_{n}^{u}$ and $\gamma$ are set to 0. Dotted lines present the result of the hydrodynamic model with $\gamma=0$ and $\rho_{n}^{u}$ extracted and fitted from $\delta_{1}$. The gray vertical line marks the superfluid-supersolid transition point. The inset presents values of $\rho_{n}^{u}$ calculated from GPE simulation and estimated with Eq.(11).

Already for Helium-4 it was predicted that within a certain temperature range such a phenomenon can occur[4, 6, 5, 1]. According to Eq.(11), the negative shift is possible for $v_{n}/v_{s}<1-(\delta_{i}+\rho^{\prime}_{n})^{-1}$. As shown in Fig.(4b), for the lower branch ($c_{2}$) there exists a parameter range for which the negative Doppler effect can be observed. In particular, since both currents have the same direction, a negative Doppler shift is possible only for values of $v_{n}>v_{n}^{cr}$, corresponding to the presence of a quantized vortex in the case of a gas confined in a ring.

Figure 4: a) Normal and superfluid velocities and their difference $w$ in a stationary state. Shapes of all three parameters are independent from $\epsilon_{dd}$. b) The region where the Doppler shift of $c_{2}$ is negative (see text). The shaded area in a) and the point on horizontal axis in b) represent the interaction parameters as used in Fig.(2a). The parameter space closer to the superfluid phase is numerically inaccessible using the dynamical method employed in this work.

Conclusions.

We have provided analytic and numerical predictions for the zero-temperature Doppler effect exhibited by ultracold atomic gases in the presence of density modulations caused by external periodic potentials, as well as by the spontaneous breaking of translational symmetry (supersolids). Analytic results are derived by linearizing the hydrodynamic theory of superfluids and supersolids and compared with dynamic simulations based on the numerical solution of the Gross-Pitaevskii equation. We have shown that in atomic gases anomalous Doppler effects are possible even at zero temperature provided the normal part is finite, which is possible thanks to the breaking of Galilean invariance. In the presence of an optical lattice only fourth sound can propagate and the Doppler effect takes a particularly simple form (see Eq.(3)), fixed by the density derivative of the superfluid density. For supersolids, where there exist sounds of hybrid superfluid and crystalline nature, the Doppler shifts exhibit highly non-trivial features and for certain values of the parameters the shift can have a negative sign. We have considered the case of very elongated configurations which could be experimentally implemented confining the atomic gas in a ring geometry and where one-dimensional hydrodynamic theory can be safely applied. The study of the precession phenomena caused by the Doppler effect and the identification of protocols to generate permanent currents in superfluids and supersolids will be the object of a future work.