betaFeedback

WorksPhysics and Astronomy

Quantum Vortices in a Boundary Layer: New Results and Perspectives

Abstract

Here, we investigate the motion of a thin circular quantum vortex filament near the infinite planar surface. The fluid surrounding this surface moves with a non-zero velocity v\bf{v}, which is parallel to the surface. We study the specific features of this quantum system and show that they are quite suitable for the boundary layer theory. The developed model allows us to calculate the vortex energy spectrum, E=E(p)E = E({\bf p}), where p{\bf p} is the total momentum of a vortex ring. We have demonstrated that this function has complex non-trivial dependence on the velocity v\bf{v}. It is stated that the inverse effective mass of a vortex under consideration is of a tensorial nature. In certain quantum states, the system shows the possibility of both negative and positive effective mass existing. This study employs a novel quantization method for classical closed vortex filaments, developed by the author earlier.

Introduction

The boundary layer theory is a well-established concept in classical fluid dynamics. The early stages of this theory’s development, as well as its different results, can be seen, for example, in the book [1] (see also the book [2] where the behavior of the classical vortices near the surfaces was explored). The mathematical aspects of boundary layer theory were studied in detail in the book [3]. Since these books were edited, many issues in this field were investigated concerning classical fluid flows. We are not aiming to review these results here, because we will be considering the quantum case instead. For some time, it has been believed that there is no boundary layer in superfluids. But recent research [4] has shown that this statement needs to be revised.

In this study, we construct a mathematical model that describes the evolution of a quantum circular vortex filament near an infinite planar surface. We also explore how the findings apply to the process of boundary layer formation. The suggested model is based on the new approach developed earlier by the author for vortex quantization [5, 6]. In these and other author’s papers, the background and motivation for this non-traditional approach were discussed at length. Instead of repeating details, we highlight the essential and innovative aspects of our proposed method.

Firstly, we introduce quantum vortices into a theory in a special non-standard way. In contrast to the prevailing view, our approach proposes that vortices are quantized classical dynamical systems, rather than topological defects. Among other things, our findings differ from the standard in that the range of quantized circulations, Γ\Gamma, is wider than the well-known equally spaced levels described by the equation Γ=ℏn/μH\Gamma= \hbar n/\mu_H. (where n=1,2,…n = 1, 2, \ldots). The previous author’s papers thoroughly explored the reasoning behind this significant finding.

Secondly, we use a group-theoretical framework to determine the total energy of a vortex filament with very small but non-zero core radius aa. To implement this program, we introduced a set of independent Hamiltonian variables, enabling a vortex loop to be represented as a particle with internal degrees of the freedom. Here, we introduce the centrally-extended Galilean group G~3\widetilde{G}_3. This group serves as the space-time symmetry group of our theory. Lee algebra of the group G~3\widetilde{G}_3 has three Cazimir functions:

C^1=μ0I^,C^2=(M^i−∑k,j=x,y,zϵijkP^jB^k)2C^3=H^−12μ0∑i=x,y,zP^i2\hat{C}_1 = \mu_0 \hat{I}, \qquad\hat{C}_2 = \left(\hat{M}_i-\sum_{k,j=x,y,z}\epsilon_{ijk}\hat{P}_j\hat{B}_k\right)^2 \qquad\hat{C}_3 = \hat{H}-\frac{1}{2\mu_0}\sum_{i=x,y,z}\hat{P}_i^2

where I^\hat{I} is the unit operator, M^i\hat{M}_i, H^\hat{H}, P^i\hat{P}_i and B^i\hat{B}_i (i=x,y,zi = x, y, z) are the respective generators of rotations, time and space translations and Galilean boosts, value μ0\mu_0 is a central charge. As usual, the function C^3\hat{C}_3 is understood as the ”internal energy of the particle”. The central charge μ0\mu_0 can be seen as a sort of conditional ”bare mass”. This method allows us to bypass intricate and sometimes ambiguous hydrodynamic calculations involving vortex filaments with very thin cores. Let us recall here Donnelly’s insightful comment: ”. . . considering how small the vortex core in helium II is, i.e., of order an angstrom, it would seem that one either ought to know how it is constructed or one ought to find a way to ignore it. Unfortunately neither goal has been achieved” [7]. Thus, we propose the following formula as the fundamental equation for vortex energy:

Ecl=p22μ0+C^3(ϖ,χ,…).(1)E_{cl} = \frac{\mathbf{p}^2}{2\mu_0}+\hat{C}_3(\varpi,\chi,\ldots). \tag*{(1)}

Here, the vector p\mathbf{p} represents vortex momentum, while the variables ω\omega, χ\chi, . . . are a set of ”internal” variables. In section 3, we will elucidate the concept of vortex energy in detail. As a consequence, we will introduce the concept of the ”effective mass” for the circular vortex filament.

Thirdly, selecting independent Hamiltonian variables enables us to apply the many-body theory to describe interactions among vortices. The author has addressed this subject before, but it doesn’t align with the current paper’s primary focus. As regards the planned outcomes of our study, they are:

  • We intend to explain the concentration of vortices near the boundary plane;

  • We will calculate the vortex energy as a function of the vortex momentum. We will also calculate and analyze the inverse effective mass tensor for the circular vortex filament

  • Although our model does not explore the reasons for the vortex appearance1 near the surface, we will formulate criteria to determine when vortices will occur;

  • We will explore a simplest multi-vortex system near a plane, concentrating on how the boundary layer develops.

The physical system under study

Let’s define the physical system we’ll explore here. Generally speaking, we consider the circular-shaped, thin quantum vortex filaments that evolve near an infinite plane. As a classical object before quantization, filament’s behaviour is governed by well-known differential equation of Local Induction Approximation (LIA), augmented by term capturing flow within vortex core. Before writing it, we should make the important notes and outline our desired outcomes.

  1. Because the area of applicability of LIA is quite limited, we should consider the possibility of its generalization. This generalization has been made in the paper [5]. But to make our construction more visual and simpler, we consider only the LIA equation with the mentioned additional term;

  2. Of course, we can extend our theory by considering small perturbations to the ring2. But in our approach, the quantum filaments shaped like

circles can be seen as building blocks, or ”atoms,” for creating closed quantum filaments of any size and shape. Corresponding theory has been developed in the paper [8];

  1. The quantum states that will be constructed, admit the states with some flow in the vortex core as well as the states without it. The author recognizes diverse viewpoints on this matter. Maybe, the point of view given above in the quote by Donnelly justifies including such a flow in our consideration;

To build our model, we use the following constants: the surrounding fluid density ρ0\rho_0 and the speed of sound in this fluid v0v_0. Besides this, we also need to introduce the length constant RfR_f. This constant can be related to the size of a stable molecular cluster, intermolecular distance and so on. Here, we don’t specify it, but we assume that the constant Rf≫a>0R_f \gg a > 0 defines a certain minimal radius for the vortex ring. Unlike the constant RfR_f, the core radius aa remains undetermined here.

Further, to simplify the formulas, we will use the auxiliary constants such as t0=Rf/v0t_0 = R_f/v_0 and E0=μ0v02\mathcal{E}_0 = \mu_0v_0^2. The constants t0t_0 and E0\mathcal{E}_0 define the time and energy scales in constructed classical theory.

It is necessary remind that the standard LIA equation is deduced [2] for the ”evolution parameter” t∗=tΓ/4πt^* = t\Gamma/4\pi, where tt means the actual (physical) time and Γ\Gamma means the circulation. Exploring the dynamics of the vortex ring, it will be more convenient to use dimensionless parameters τ\tau and ξ\xi instead of the ”real time” tt and the natural parameter (arc length of the curve) ss:

τ=t∗/R2,ξ=s/R,ξ∈[0,2π],(2)\tau= t^*/R^2,\qquad\xi= s/R,\qquad\xi\in[0,2\pi], \tag*{(2)}

where value RR means the ring radius. If we introduce the projective vector r=r/R\mathbf r = \mathbf r/R instead of the radius-vector r\mathbf r, the convenient form of the LIA equation will be as follows:

∂τr(τ,ξ)=α(∂ξr(τ,ξ)×∂ξ2r(τ,ξ))+ω(2∂ξ3r(τ,ξ)+3∣∂ξ2r(τ,ξ)∣2∂ξr(τ,ξ)),(3)\partial_\tau\mathbf r(\tau,\xi) = \alpha\left(\partial_\xi\mathbf r(\tau,\xi) \times\partial_\xi^2 \mathbf r(\tau,\xi)\right) + \omega\left(2\partial_\xi^3 \mathbf r(\tau,\xi) + 3\left|\partial_\xi^2 \mathbf r(\tau,\xi)\right|^2\partial_\xi\mathbf r(\tau,\xi)\right), \tag*{(3)}

where the values α\alpha and ω\omega are finite dimensionless constants. In total, the parameter α∝ln⁡a+const\alpha\propto\ln a + const arises as a consequence of a certain regularization process when the LIA equation is derived. As regards the parameter ω\omega, it is determined by the velocity of the fluid flow Φω\Phi_\omega in the vortex core. The book [9] thoroughly derives this equation and explicitly expresses the parameter $\omega using conventional physical values which are connected with core flow. We will assume that both α\alpha and ω\omega take non-zero values in this study. We use the following exact solution to the Eq. (3):

r(τ,ξ)=(qx/R+cos⁡(ξ+ϕ),qy/R+sin⁡(ξ+ϕ),qz/R+ατ),(4)\mathbf{r}(\tau,\xi)=\left(q_x/R+\cos(\xi+\phi),\quad q_y/R+\sin(\xi+\phi),\quad q_z/R+\alpha\tau\right), \tag*{(4)}

where the variable ϕ≡ϕ(τ)=ϕ0+ωτ\phi\equiv\phi(\tau)=\phi_0+\omega\tau defines the rotation of the circular curve (4). This rotation models the flow into the core of the vortex. As regards the vector q=(qx,qy,qz)\mathbf{q}=(q_x,q_y,q_z), it determines the center of the circle.

Here and further, we will use the indices (x,y,z)(x,y,z) for the Cartesian coordinate system, which is naturally associated with solution (4). In this system, the x0yx0y plane coincides with a ring plane and the ZZ axis is directed along the ring axis. There is another natural Cartesian coordinate system that exists in our theory. In this system, X0YX0Y plane coincides with the planar surface under consideration. To avoid ambiguity, we will use the indices (1,2,3)(1,2,3) to denote coordinates in this system. As usual, the correspondence (1,2,3)↔(X,Y,Z)(1,2,3)\leftrightarrow(X,Y,Z) takes place for the coordinate axes.

Equation (3) describes the dynamics of a vortex filament as a purely geometric object. But because we are constructing a model to describe the vortex motion, we must take into account the motion of the surrounding fluid in some way. To do it, we fulfill two steps.

Firstly, we assume that the conventional formula for the canonical momentum p\mathbf{p} is fulfilled:

p=pv+ϱ02∫r×w(r) dV,(5)\mathbf{p}=\mathbf{p}_v+\frac{\varrho_0}{2}\int\mathbf{r}\times\mathbf{w}(\mathbf{r})\,\mathrm{d}V, \tag*{(5)}

where the vorticity w(r)\mathbf{w}(\mathbf{r}) is defined for the vortex filament by standard way [10]:

w(r)=Γ∫02πδ(r−r(ξ))∂ξr(ξ) dξ.\mathbf{w}(\mathbf{r})=\Gamma\int_{0}^{2\pi}\delta\left(\mathbf{r}-\mathbf{r}(\xi)\right)\partial_\xi\mathbf{r}(\xi)\,\mathrm{d}\xi.

The summand pv\mathbf{p}_v takes into account the momentum which arises when the surrounding fluid has a non-zero velocity v\mathbf{v} without vortex motion. The second summand in the Eq. (5) is the standard vortex canonical momentum p\mathbf{p} for zero boundary conditions (v=0\mathbf{v}=\mathbf{0}) at infinity [11]. If we substitute the solution (4) into Eq. (5) we deduce simple equation

p−pv=πϱ0R2Γ b.(6)\mathbf{p}-\mathbf{p}_v=\pi\varrho_0R^2\Gamma\,\mathbf{b}. \tag*{(6)}

The vector b\mathbf{b} here means the binormal unit vector of the considered circular filament; its coordinates (bx,by,bz)(b_x,b_y,b_z) are (0,0,1)(0,0,1). It is clear that this vector coincides with the direction of the vortex circle propagation.

Secondly, we include the circulation Γ\Gamma as a dynamic variable in our theory. Thus, the minimal set of variables needed to describe our dynamic system is as follows:

A={Γ,q,R,ϕ(τ),b}.\mathcal{A} = \{\Gamma,\mathbf{q},R,\phi(\tau),b\}.

Let us focus our attention on the summand pv\mathbf{p}_{\mathrm v} in the Eq. (5). In general, we assume that the following equality is fulfilled:

vi=∑j=1,2[Meff−1]ij(pv)j,i=1,2,(7)v_i = \sum_{j=1,2} [\mathbf{M}_{\mathrm{eff}}^{-1}]_{ij}(\mathbf{p}_{\mathrm v})_j,\qquad i=1,2, \tag*{(7)}

where the values vjv_j are components of the velocity v\mathbf{v} of the surrounding fluid without vortex motion. We assume also that the vector v\mathbf{v} is parallel to the considered plane surface, so that the fluid flow along the surface. The values [Meff−1]ij[\mathbf{M}_{\mathrm{eff}}^{-1}]_{ij} are components of the tensor of the ”inverse effective mass” for the vortex ring. To calculate its components, we need to determine the energy E=E(p,pv)E=E(\mathbf{p},\mathbf{p}_{\mathrm v}) of the quantum vortex ring in question. Indeed,

[Meff−1]ij=∂2E(p,pv)∂pi∂pj,i,j=1,2.(8)[\mathbf{M}_{\mathrm{eff}}^{-1}]_{ij}=\frac{\partial^2 E(\mathbf{p},\mathbf{p}_{\mathrm v})}{\partial p_i\partial p_j},\qquad i,j=1,2. \tag*{(8)}

More interesting cases for us are those when the vector p=psp\mathbf{p}=\mathbf{p}_{\mathrm{sp}} is an extreme or a stationary point of the function E(p,pv)E(\mathbf{p},\mathbf{p}_{\mathrm v}). We will return to this issue later.

Therefore, our subsequent goal is to quantize our circular-shaped vortex filament and determine its energy. To quantize our system in accordance with our approach, we need to redefine the set of dynamical variables A\mathcal{A}. Firstly, we fulfill replacement {Γ,b}→p\{\Gamma,b\}\to\mathbf{p} with help of the Eq. (6) where the vector pv\mathbf{p}_{\mathrm v} is considered as an external parameter. Secondly, we replace the variables RR and ϕ(τ)\phi(\tau) with new variables χ\chi and ϖ\varpi, as follows:

χ=ΔRRfcos⁡(ϕ0+ωτ),ϖ=ΔRRfsin⁡(ϕ0+ωτ),whereΔR=R2−Rf2.\chi=\frac{\Delta R}{R_f}\cos(\phi_0+\omega\tau),\qquad\varpi=\frac{\Delta R}{R_f}\sin(\phi_0+\omega\tau),\qquad\text{where}\quad\Delta R=\sqrt{R^2-R_f^2}.

Clearly, the behavior of the variables ϖ\varpi and χ\chi is similar to that of a harmonic oscillator. Finally, we postulate the set

A′={p,q;ϖ,χ}(9)\mathcal{A}'=\{\mathbf{p},\mathbf{q};\varpi,\chi\} \tag*{(9)}

as the set of fundamental independent variables for the considered dynamical system – circle-shaped vortex ring which evolves in accordance with Eq. (3). As a consequence of equality (6), the following formula holds:

(p−pv)2=π2θ02Rf4(1+ϖ2+χ2)2Γ2.(10)(\mathbf{p}-\mathbf{p}_{\mathrm v})^2=\pi^2\theta_0^2R_f^4\left(1+\varpi^2+\chi^2\right)^2\Gamma^2. \tag*{(10)}

This formula shows that the circulation Γ\Gamma is the function of new variables that parametrize the set (9).

Hamiltonian structure and quantization

The structure of the set A′\mathcal{A}' as well as the assumption about vortex energy (see Eq. (1)) naturally lead to the following Hamiltonian description of the model.

  • Phase space H=Hpq×Hb\mathcal{H}=\mathcal{H}_{pq}\times\mathcal{H}_b. The space Hpq\mathcal{H}_{pq} is the phase space for a 3D3D free structureless particle which is parametrized by the variables q\mathbf{q} and p\mathbf{p}. Note that the coordinate q3q_3 cannot equal zero due to the presence of an impermeable planar surface 33.3 The space Hb\mathcal{H}_b is a phase space for one-dimensional harmonic oscillator with (conditional) frequency ω/t0\omega/t_0.

  • Poisson structure:

{pi,qj}=δij(i,j=1,2,3) ;{ϖ,χ}=1ε0t0\{p_i,q_j\}=\delta_{ij}\quad(i,j=1,2,3)\ ;\qquad\{\varpi,\chi\}=\frac{1}{\varepsilon_0t_0}

the brackets that are not written are vanished;

  • Hamiltonian

H=p22μ0+ε0ω2(ω2+χ2).(11)H=\frac{\mathbf{p}^{2}}{2\mu_0}+\frac{\varepsilon_0\omega}{2}\left(\omega^2+\chi^2\right). \tag*{(11)}

This Hamiltonian structure naturally leads to a quantum description of our vortex filament using the following Hilbert space:

H1=Hpq⊗Hb.(12)\mathbf{H}_1=\mathbf{H}_{pq}\otimes\mathbf{H}_b. \tag*{(12)}

Symbol Hpq\mathbf{H}_{pq} denotes the Hilbert space of quantum states for a free, structureless, non-relativistic particle. In general, we assume that

Hpq=L2(R2)⊗L2(R1)≈L2(R3).\mathbf{H}_{pq}=L_2(\mathbb{R}_2)\otimes L_2(\mathbb{R}_1)\approx L_2(\mathbb{R}_3).

Here we limit ourselves to operators on Hilbert space Hpq\mathbf{H}_{pq} of the form

A=A2⊗I1+I2⊗A1.A=A_2\otimes I_1+I_2\otimes A_1.

where operators A2A_2, A1A_1 and unit operators I2I_2, I1I_1 act in the spaces L2(R2)L_2(\mathbb{R}_2), L2(R1)L_2(\mathbb{R}_1) correspondingly. To simplify formulas, we will use notations A2+A1A_2+A_1 instead A2⊗I1+I2⊗A1A_2\otimes I_1+I_2\otimes A_1. For our subsequent purposes, we will only consider the AA-operators with domains:

D(A)⊆L2(R2)×L2(R1)⊂L2(R2)⊗L2(R1).\mathop{\mathrm{D}}(A)\subseteq L_2(\mathbb{R}_2)\times L_2(\mathbb{R}_1)\subset L_2(\mathbb{R}_2)\otimes L_2(\mathbb{R}_1).

The symbol ×\times means the algebraic tensor product of the corresponding spaces: any functions ψ(q1,q2,q3)∈L2(R2)×L2(R1)\psi(q_1,q_2,q_3) \in L_2(\mathbb{R}_2) \times L_2(\mathbb{R}_1) are finite sequences of the functions ψa(q1,q2)ψb(q3)\psi_a(q_1,q_2)\psi_b(q_3), where ψa(q1,q2)∈L2(R2)\psi_a(q_1,q_2) \in L_2(\mathbb{R}_2) and ψb(q3)∈L2(R1)\psi_b(q_3) \in L_2(\mathbb{R}_1).4 In our case, the domain of any operator A1A_1 must takes into account the presence of an impenetrable (X0Y)(X_0Y) surface. Therefore, any operator A1A_1 which corresponds to a physical observable is defined in some domain D1(A1)⊆D1⊂L2(R1)\mathbb{D}_1(A_1) \subseteq\mathbb{D}_1 \subset L_2(\mathbb{R}_1), where domain D1\mathbb{D}_1 is defined as follows:

Ψ(q3)∈D1:Ψ(q3)∈L2(R1),∣Ψ(0)=0.(13)\Psi(q_3) \in\mathbb{D}_1:\quad\Psi(q_3) \in L_2(\mathbb{R}_1), \quad| \Psi(0) = 0. \tag*{(13)}

As it is known, such limitations frequently cause problems with self-adjointness of operators. We will return to the determination of domain D(H)\mathbb{D}(H) of the Hamiltonian HH later.

The symbol Hb\mathbf{H}_b represents the Hilbert space of the quantized harmonic oscillator, characterized by the classical variables χ\chi and ϖ\varpi. This space is formed by the vectors

∣n⟩=1n!(b^+)n∣0b⟩[b^,b^+]=I^b,b^∣0b⟩=0,|n\rangle= \frac{1}{\sqrt{n!}}(\hat{b}^{+})^n|0_b\rangle\qquad[\hat{b},\hat{b}^{+}] = \hat{I}_b,\qquad\hat{b}|0_b\rangle= 0,

where vector ∣0b⟩∈Hb|0_b\rangle\in\mathbf{H}_b is vacuum vector and symbols b^+\hat{b}^{+} and b^\hat{b} mean the operators which create and annihilate the oscillator levels. As regards to the classical variables ϖ\varpi and χ\chi, these become operators as follows:

χ+iϖ⟶2σphb^,whereσph=ℏE0t0.(14)\chi+i\varpi\longrightarrow\sqrt{2}\sigma_{ph}\hat{b},\qquad\text{where}\qquad\sigma_{ph}=\sqrt{\frac{\hbar}{\mathcal{E}_0t_0}}. \tag*{(14)}

The variables q\mathbf{q} and p\mathbf{p} are quantized using standard rules of quantum mechanics for free, non-relativistic particles in coordinate representation: q→q^=q\mathbf{q} \to\hat{\mathbf{q}}=\mathbf{q}, p→p^\mathbf{p}\to\hat{\mathbf{p}}. The features of the operator p^\hat{\mathbf{p}}, which arise due to the presence of an impenetrable surface will be discussed in the subsequent analysis of quantum classical equations.

The first significant finding is that the radius of a vortex ring is quantized. By examining the definitions of the variables χ\chi and ϖ\varpi, we can derive the formula for the quantized value R2R^2:

R2⟶R^2=Rf2[I^b+σph2(b^+b^+12)].R^2 \longrightarrow\hat{R}^{2}=R_f^2\left[\hat{I}_b+\sigma_{ph}^2\left(\hat{b}^{+}\hat{b}+\frac{1}{2}\right)\right].

It is easy to find the eigenvalues Rn2R_n^2 of this operator, and consequently, the values RnR_n of the vortex radius:

Rn=Rf1+σph2(n+1/2),n=0,1,… .(15)R_n=R_f\sqrt{1+\sigma_{ph}^2\left(n+1/2\right)},\qquad n=0,1,\ldots\,. \tag*{(15)}

We will consider the spectrum (15) for a restricted number of nn values. As the number nn grows, the system exhibits quasi-classical behavior5. Further, the quantization scheme results in two spectral problems.

  • The first spectral problem arises after quantizing the equation (10). This problem determines the circulation values Γ\Gamma:

[(p−pv)2−π2Ω02Γ2Rf4(I^+σph2(b^†b^+12))2]∣Ψ⟩=0.(16)\left[(\mathbf{p}-\mathbf{p}_{v})^{2}-\pi^{2}\Omega_{0}^{2}\Gamma^{2}R_{f}^{4}\left(\hat{I}+\sigma_{ph}^{2}\left(\hat{b}^{\dagger}\hat{b}+\frac{1}{2}\right)\right)^{2}\right]|\Psi\rangle=0. \tag*{(16)}
  • Second problem is determining the energy value E#E^{\#}, such that the corresponding Hamiltonian H^#\hat{H}^{\#} provides dynamics in accordance with conditional time t#=τt0t^{\#}=\tau t_{0}:

H^#∣ψ⟩=E#∣ψ⟩,H^#=−p^22μ0+ℏωt0(b^†b^+12).(17)\hat{H}^{\#}|\psi\rangle=E^{\#}|\psi\rangle,\qquad\hat{H}^{\#}=-\frac{\hat{p}^{2}}{2\mu_{0}}+\frac{\hbar\omega}{t_{0}}\left(\hat{b}^{\dagger}\hat{b}+\frac{1}{2}\right). \tag*{(17)}

Let us now discuss operators p^2\hat{\mathbf{p}}^{2} and (p−pv^)2(\widehat{\mathbf{p}-\mathbf{p}_{v}})^{2}. In the case where the impenetrable X0Y surface is absent, the quantization is trivial: p2→−ℏ2Δ3\mathbf{p}^{2}\to-\hbar^{2}\Delta_{3}, where notation Δn\Delta_{n} (n=2,3n=2,3) means the Laplace operator on space L2(Rn)L_{2}(\mathbb{R}_{n}) with domain W22(Rn)W_{2}^{2}(\mathbb{R}_{n}), where the space W22(Rn)W_{2}^{2}(\mathbb{R}_{n}) is corresponding Sobolev space. The quantization of the variable (p−pv)2(\mathbf{p}-\mathbf{p}_{v})^{2} is also trivial.

But the operator Δ3\Delta_{3}, with a domain that was restricted by Eq. (13), is not a self-adjoint operator! Indeed,

p2⟶p^2=−ℏ2Δ2+p^32,\mathbf{p}^{2}\longrightarrow\hat{\mathbf{p}}^{2}=-\hbar^{2}\Delta_{2}+\hat{p}_{3}^{2},

where the operator Δ2\Delta_{2} is self-adjoint operator with domain D(Δ2)=W22(R2)\mathop{\mathrm{D}}(\Delta_{2})=W_{2}^{2}(\mathbb{R}_{2}). Let us discuss the p^32\hat{p}_{3}^{2} operator. The operator p^32=−ℏ2∂2/∂q32\hat{p}_{3}^{2}=-\hbar^{2}\partial^{2}/\partial q_{3}^{2} is self-adjoint if it is defined on a domain W22(R1)W_{2}^{2}(\mathbb{R}_{1}). But in accordance with our assumption (13), we must consider the restriction

p^32⟶pˇ32=p^32∣D1∩W22(R1),\hat{p}_{3}^{2}\longrightarrow\check{p}_{3}^{2}=\left.\hat{p}_{3}^{2}\right|_{\mathop{\mathrm{D}}_{1}\cap W_{2}^{2}(\mathbb{R}_{1})},

that means the domain restriction

W22(R1)⟶D(pˇ32)=W22(R1)∩D1.W_{2}^{2}(\mathbb{R}_{1})\longrightarrow\mathop{\mathrm{D}}(\check{p}_{3}^{2})=W_{2}^{2}(\mathbb{R}_{1})\cap\mathop{\mathrm{D}}_{1}.

The operator pˇ32\check{p}_{3}^{2} is symmetric operator in the space L2(R1)L_{2}(\mathbb{R}_{1}), so that the inclusion D(pˇ32)⊂D((pˇ32)+)\mathop{\mathrm{D}}(\check{p}_{3}^{2})\subset\mathop{\mathrm{D}}((\check{p}_{3}^{2})^{+}) is fulfilled. Therefore, we must consider the self-adjoint extension p~32(λ)\tilde{p}_3^2(\lambda) (see, for example, [12]) of this operator. To avoid technical details that are not relevant to the main topic of this study, we will use the results from work [13]. In accordance with results [13], the all self-adjoint extensions p~32(λ)\tilde{p}_3^2(\lambda) of the operator p~32\tilde{p}_3^2 are marked by the real number λ\lambda. The corresponding domains D(p~32(λ))\mathbb{D}(\tilde{p}_3^2(\lambda)) includes the functions ψ(x)∈D((p~32)+)\psi(x) \in\mathbb{D}((\tilde{p}_3^2)^+) that meet specific conditions:

ψ(−0)−ψ(+0)=0,ψ′(−0)−ψ′(+0)=−2λℏ2ψ(0).\begin{aligned} \psi(-0)-\psi(+0)&=0,\\ \psi'(-0)-\psi'(+0)&=-\frac{2\lambda}{\hbar^2}\psi(0). \end{aligned}

Formally speaking, these functions can be considered as the solution of the differential equation

[−ℏ22μ0∂2∂q32+λδ(q3)]ψ(q3)=Eλψ(q3),Eλ<0,\left[-\frac{\hbar^2}{2\mu_0}\frac{\partial^2}{\partial q_3^2}+\lambda\delta(q_3)\right]\psi(q_3)=E_\lambda\psi(q_3),\qquad E_\lambda<0,

where λ=−ℏ−2Eλ/μ0\lambda=-\hbar\sqrt{-2E_\lambda/\mu_0}. For example, the function

Ψλ(q3)={exp⁡[−−2μ0Eλ q3/ℏ],q3>0,exp⁡[−2μ0Eλ q3/ℏ],q3<0,(18)\Psi_\lambda(q_3)= \begin{cases} \exp[-\sqrt{-2\mu_0E_\lambda}\,q_3/\hbar], & q_3>0,\\ \exp[\sqrt{-2\mu_0E_\lambda}\,q_3/\hbar], & q_3<0, \end{cases} \tag*{(18)}

solves this equation quite correctly: the product δ(q3)ψ(q3)\delta(q_3)\psi(q_3) is well-defined here. This fact gives some basis for interpreting of the extension procedure described above as a background for the Schrödinger equation with a δ\delta potential. In 3D3D case, this fact was investigated many years ago [14].

Of course, the differential expression [−(ℏ2/2μ0)(∂2/∂q32)+λδ(q3)][-(\hbar^2/2\mu_0)(\partial^2/\partial q_3^2)+\lambda\delta(q_3)] with the distribution-valued coefficient function does not define the operator in the space L2(R1)L_2(\mathbb{R}_1). As a consequence, treating this differential equation as the Schrödinger equation is problematic. To do this, we need to construct a theory in which the multiplication of distributions is a well-defined procedure. For example, corresponding theory was constructed in the work [15]. We do not go into detail, but we note that the solution (18) has also been deduced from the theory, suggested in the paper [15].

To simplify notations, we will use the summand 2μ0λδ(q3)2\mu_0\lambda\delta(q_3) formally, meaning that the procedure of the self-adjoint extension of the operator p~32\tilde{p}_3^2 was fulfilled.

Thus, Eq. (16) is transformed as follows:

[−ℏ2(∂∂q1−(kv)1)2−ℏ2(∂∂q2−(kv)2)2−ℏ2∂2∂q32+2μ0λδ(q3)−π2ϱ02Γ2Rf4(I^+σph2(b^†b^+1/2))2]∣Ψ⟩=0(19)\left[-\hbar^{2}\left(\frac{\partial}{\partial q_{1}}-(k_{v})_{1}\right)^{2}-\hbar^{2}\left(\frac{\partial}{\partial q_{2}}-(k_{v})_{2}\right)^{2}-\hbar^{2}\frac{\partial^{2}}{\partial q_{3}^{2}}+2\mu_{0}\lambda\delta(q_{3})-\pi^{2}\varrho_{0}^{2}\Gamma^{2}R_{f}^{4}\left(\hat{I}+\sigma_{ph}^{2}\left(\hat{b}^{\dagger}\hat{b}+1/2\right)\right)^{2}\right]|\Psi\rangle=0 \tag*{(19)}

where the values (kv)i(k_{v})_{i}, where i=1,2i=1,2, are the components of the wave vector pv/ℏ\mathbf{p}_{v}/\hbar. Spectral problem (19) has eigenvectors

∣Ψ⟩=∣Ψλ(p1,p2)⟩∣n⟩,(20)|\Psi\rangle=|\Psi_{\lambda}(p_{1},p_{2})\rangle|n\rangle, \tag*{(20)}

where the vectors ∣Ψλ(p1,p2)⟩∈Hpq′|\Psi_{\lambda}(p_{1},p_{2})\rangle\in\mathbf{H}'_{pq} in coordinate representation take form:

⟨q∣Ψλ(p1,p2)⟩≡Ψ(q1,q2,q3)=Ψλ(q3)exp⁡[−i(p1q1+p2q2)/ℏ].\langle q|\Psi_{\lambda}(p_{1},p_{2})\rangle\equiv\Psi(q_{1},q_{2},q_{3})=\Psi_{\lambda}(q_{3})\exp\left[-i(p_{1}q_{1}+p_{2}q_{2})/\hbar\right].

Symbol Hpq′\mathbf{H}'_{pq} means corresponding rigged (”equipped”) Hilbert space (see, for example, [16]).

Why is it a boundary layer?

Thus, we achieved one of our goals in this study: the vectors (20) visually demonstrate that the vortex in question is located near the Z=0Z=0 surface. The eigenvalues Γ2\Gamma^{2} of the spectral problem (19) determine the permitted circulation values in our theory. To write Γ\Gamma values in a convenient form, we will introduce some additional notations:

Mf=πϱ0Rf3,pℏ=ℏRf,κ=ppℏ,κv=pvpℏ=Rfkv.M_{f}=\pi\varrho_{0}R_{f}^{3},\qquad p_{\hbar}=\frac{\hbar}{R_{f}},\qquad\boldsymbol{\kappa}=\frac{\mathbf{p}}{p_{\hbar}},\qquad\boldsymbol{\kappa}_{v}=\frac{\mathbf{p}_{v}}{p_{\hbar}}=R_{f}\mathbf{k}_{v}.

The value pℏp_{\hbar} defines the momentum scale in our theory. As regards the mass parameter MfM_{f}, its physical meaning will be clarified later. Further, we will prefer dimensionless vectors, κ\boldsymbol{\kappa}, κv\boldsymbol{\kappa}_{v}, …, anywhere if it does not cause ambiguity.

As a consequence of the above studies, the values of circulation are written in terms of ”mass” MfM_{f} and dimensionless vectors κ\boldsymbol{\kappa}, κv\boldsymbol{\kappa}_{v} as follows:

Γλ(p,pv;n)=±ℏMfγλ(p,pv;n),(21)\Gamma_{\lambda}(\mathbf{p},\mathbf{p}_{v};n)=\pm\frac{\hbar}{M_{f}}\gamma_{\lambda}(\mathbf{p},\mathbf{p}_{v};n), \tag*{(21)}

where dimensionless function γλ(p,pv;n)\gamma_{\lambda}(\mathbf{p},\mathbf{p}_{v};n) is written as follows:

γλ(p,pv;n)≡γλ(κ,κv;n)=(κ1−(κv)1)2+(κ2−(κv)2)2−2βλ/σph2[1+σph2(n+1/2)].\gamma_{\lambda}(\mathbf{p},\mathbf{p}_{v};n)\equiv\gamma_{\lambda}(\boldsymbol{\kappa},\boldsymbol{\kappa}_{v};n)=\frac{\sqrt{\left(\kappa_{1}-(\kappa_{v})_{1}\right)^{2}+\left(\kappa_{2}-(\kappa_{v})_{2}\right)^{2}-2\beta_{\lambda}/\sigma_{ph}^{2}}}{\left[1+\sigma_{ph}^{2}(n+1/2)\right]}.

The notation βλ=Rf∣Eλ∣/v0ℏ\beta_{\lambda}=R_f|E_{\lambda}|/v_0\hbar was introduced here.

This formula requires further discussion. We won’t discuss here the no-table discrepancies between the found set of circulation values and the standard set Γ∝k\Gamma\propto k, where the number kk is natural number. The author’s previous works delve deeply into these disparities, examining specific cases and their corresponding arguments and motivations. The new feature of the system under discussion is the absence of a Γ\Gamma-value for some parameter values. Indeed, we can see that the function under the square root may be negative, but the circulation must always be a real number.

To better understand the physical implications of this observation, we will begin with a qualitative analysis. Let us suppose that the matrix Meff−1=([Meff−1]ij)\mathbf{M}_{\mathrm{eff}}^{-1}=([\mathbf{M}_{\mathrm{eff}}^{-1}]_{ij}) exists and the inverse matrix Meff\mathbf{M}_{\mathrm{eff}} also exists, such that each element [Meff]ij[\mathbf{M}_{\mathrm{eff}}]_{ij} is finite. Thus, we consider a ”easy vortices” here. In this case, the condition for the square root of Eq. (21) to take non-zero real values is as follows:

∑i,j=12(pi−[Meff]i,jvj)2>2μ0∣Eλ∣.(22)\sum_{i,j=1}^{2}\left(p_i-[\mathbf{M}_{\mathrm{eff}}]_{i,j}v_j\right)^2>2\mu_0|E_{\lambda}|. \tag*{(22)}

It is clear that this condition will break down in one of two ways. First one, when the values of v\mathbf{v} and p\mathbf{p} are sufficiently small. Second one, when the values of v\mathbf{v} and p\mathbf{p} can be arbitrarily large but the following condition holds:

∣pi−∑j=1,2[Meff]i,jvj∣≤2μ0∣Eλ∣,(23)\left|p_i-\sum_{j=1,2}[\mathbf{M}_{\mathrm{eff}}]_{i,j}v_j\right|\leq\sqrt{2\mu_0|E_{\lambda}|}, \tag*{(23)}

These facts means that there is no any vortex motion in this case. Thus, the prediction of our model agrees quite well with the paper [4] where the creation of vortices in a boundary layer is conditioned by irregularities in the surface. As it seems, the ”almost resting” fluid should not lead to vortex formation. As regards the condition (23), it predicts the disappearance of vortex motion in a certain range of fluid velocities. We’re still examining individual circular vortices. Of course, the turbulent media in the boundary layer consists of many vortices of various shapes and sizes. We’ll return to this issue in subsequent sections of this study.

The above qualitative analysis is quite conditional. Indeed, when the matrix Meff\mathbf{M}_{\mathrm{eff}} ceases to exist for certain parameter values, the criterion for the vortex appearance becomes more complex. In this case, we must consider the equations:

∣p−pv∣2>2μ0∣Eλ∣,(24)|\mathbf{p}-\mathbf{p}_v|^2>2\mu_0|E_{\lambda}|, \tag*{(24)}
vi=∑j=12∂2E(p,pv)∂pi∂pj(pv)j,i=1,2.(25)v_i=\sum_{j=1}^{2}\frac{\partial^2 E(\mathbf{p},\mathbf{p}_v)}{\partial p_i\partial p_j}(\mathbf{p}_v)_j,\qquad i=1,2. \tag*{(25)}

To deduce the condition of the vortices appearance, we need to find the values (pv)1(\mathbf{p}_v)_1 and (pv)2(\mathbf{p}_v)_2 from the system (25). After that, the result should be substituted into Eq. (24).

Thus, to continue our study, we need to return to the definition of the vortex energy E(p,pv)E(\mathbf{p},\mathbf{p}_v) in our theory. The spectral problem (17) has the eigenvalues

En#(p)=12μ0(p12+p22)+Eλ+ℏωt0(n+12)n=0,1,….(26)E_n^{\#}(\mathbf{p}) = \frac{1}{2\mu_0}\left(p_1^2+p_2^2\right) + E_\lambda+ \frac{\hbar\omega}{t_0}\left(n+\frac{1}{2}\right) \qquad n=0,1,\ldots. \tag*{(26)}

As an important point, we must remember that until now we have described LIA dynamics using the concept of ”conditional time” - a special evolution parameter, t∗=tΓ/4πt^{\ast}=t\Gamma/4\pi. After that, we introduced the dimensionless evolution parameter τ\tau in accordance with Eq. (2) for convenience. This fact means that the values (26) represent some ”conditional energy”, since this energy value is related to the dynamics with respect to ”conditional” time t#=t0τt^{\#}=t_0\tau. Therefore, our first goal is to return to the actual (physical) time tt as well as to return to actual energy En(p1,p2)E_n(p_1,p_2), which is connected to the time tt. To do it, let us note that the t#t^{\#} dependence of any vector ∣Ψ⟩∈H1|\Psi\rangle\in H_1 is described as follows:

∣Ψ⟩⟶∣Ψ(t#)⟩=exp⁡(iH^#t#ℏ)∣Ψ⟩==∑n∫dp1dp2 Cn(p1,p2)exp⁡(iEn#(p)t#ℏ)∣Ψλ(p1,p2)⟩∣n⟩,\begin{aligned} |\Psi\rangle\longrightarrow|\Psi(t^{\#})\rangle= \exp\left(\frac{i\hat{H}^{\#}t^{\#}}{\hbar}\right)|\Psi\rangle= \\ &= \sum_n \int dp_1dp_2\,C_n(p_1,p_2)\exp\left(\frac{iE_n^{\#}(\mathbf{p})t^{\#}}{\hbar}\right)|\Psi_\lambda(p_1,p_2)\rangle|n\rangle, \end{aligned}

where Cn(p1,p2)=⟨Ψλ(p1,p2)∣⟨n∣Ψ⟩C_n(p_1,p_2)=\langle\Psi_\lambda(p_1,p_2)|\langle n|\Psi\rangle. In the next step, we restore the real-time tt value in our formulas by using the equation

t#⟶t=4πR2t0∣Γ∣t#.(27)t^{\#}\longrightarrow t=\frac{4\pi R^2}{t_0|\Gamma|}t^{\#}. \tag*{(27)}

Our key assumption here is that the probabilities ∣⟨Φ∣Ψ(t#)⟩∣2|\langle\Phi|\Psi(t^{\#})\rangle|^2 of transition to any quantum state ∣Φ⟩|\Phi\rangle are observable values that do not depend on the time parametrization. To ensure this, we need to require the following equality for the real energy En(p1,p2)E_n(p_1,p_2):

En#(p1,p2)t#=En(p1,p2)t.E_n^{\#}(p_1,p_2)t^{\#}=E_n(p_1,p_2)t.

Therefore, the ”real - time” evolution of any vector ∣Ψ⟩∈H|\Psi\rangle\in H is written as

∣Ψ⟩⟶∣Ψ(t)⟩=∑n∫dp1dp2 Cn(p1,p2)exp⁡(iEn(p1,p2)tℏ)∣ψp;λ,n⟩.(28)|\Psi\rangle\longrightarrow|\Psi(t)\rangle= \sum_n\int dp_1dp_2\,C_n(p_1,p_2)\exp\left(\frac{iE_n(p_1,p_2)t}{\hbar}\right)|\psi_{p;\lambda,n}\rangle. \tag*{(28)}

where function En(p1,p2)≡En(p;pv)E_n(p_1,p_2) \equiv E_n(\mathbf{p};\mathbf{p}_v) means the actual vortex energy. This function takes the following form:

En(p;pv)=t0∣Γλ(p,pv;n)∣4πRn2En#(p)≡EℏEn(κ,κv),(29)E_n(\mathbf{p};\mathbf{p}_v)=\frac{t_0\left|\Gamma_\lambda(\mathbf{p},\mathbf{p}_v;n)\right|}{4\pi R_n^2}E_n^{\#}(\mathbf{p})\equiv E_{\hbar}\mathcal{E}_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v), \tag*{(29)}

where the constant

Eℏ=ℏ2MfRf2E_{\hbar}=\frac{\hbar^2}{M_fR_f^2}

defines the energy scale in our model. The dimensionless function En(κ,κv)\mathcal{E}_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v) is written as follows:

En(κ,κv)=(κ1−(κv)1)2+(κ2−(κv)2)2−2βλ/σph24π[1+σph2(n+1/2)]2××[ω(n+1/2)+σph22(κ12+κ22)−βλ].(30)\begin{aligned} \mathcal{E}_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v)=\frac{\sqrt{(\kappa_1-(\kappa_v)_1)^2+(\kappa_2-(\kappa_v)_2)^2-2\beta_\lambda/\sigma_{\mathrm{ph}}^2}}{4\pi\left[1+\sigma_{\mathrm{ph}}^2(n+1/2)\right]^2}\times\\ &\times\left[\omega(n+1/2)+\frac{\sigma_{\mathrm{ph}}^2}{2}(\kappa_1^2+\kappa_2^2)-\beta_\lambda\right]. \tag*{(30)} \end{aligned}

Let us discuss the breakdown of the condition (24) from the function En(κ,κv)E_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v) point of view. Indeed, the breakdown of this condition for some values (κ1,κ2)(\kappa_1,\kappa_2) means that this function takes the imaginary values. We will denote corresponding domain in the (κ1,κ2)(\kappa_1,\kappa_2) plane as D00D_{00}. Let the condition (24) be broken. By writing the corresponding appropriate square root value in the form

En(κ,κv)=iℏ/Tn(κ,κv),Tn(κ,κv)>0,E_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v)=i\hbar/T_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v),\qquad T_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v)>0,

we can transform some parts of a sub-integral expression in Equation (28) into a form:

exp⁡(iEn(p1,p2)t/ℏ)∣ψp;λ,n⟩⟶exp⁡(−t/Tn(κ,κv))∣ψp;λ,n⟩.\exp\left(iE_n(p_1,p_2)t/\hbar\right)|\psi_{\mathbf{p};\lambda,n}\rangle\longrightarrow\exp\left(-t/T_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v)\right)|\psi_{\mathbf{p};\lambda,n}\rangle.

Thus, the vortex modes corresponding to the given values of (κ1,κ2)(\kappa_1,\kappa_2) are unstable. Their contribution decreases exponentially over time for any given amplitude. In this context, the constant Tn(κ,κv)T_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v) can be linked to a ”vortex lifetime.” In a classical theory, this parameter has been studied in the work [17], from a different point of view.

The above outcome is consistent with the previous statement that vortices do not exist if condition (24) is broken. More specific results that justify the applicability of our model to a boundary layer description, require a detailed study of the function En(κ,κv)E_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v).

The energy structure of vortices and the conditions for their formation

The first obvious conclusion is that the tensor of the inverse effective mass (8) contains off-diagonal elements. To continue our study, we will investigate the structure of the 3D3D surface En=En(κ1,κ2;κv)\mathcal{E}_n = \mathcal{E}_n(\kappa_1,\kappa_2;\boldsymbol{\kappa}_v) in the Cartesian coordinate system (E,κ1,κ2)(\mathcal{E},\kappa_1,\kappa_2). One of the important issues that we intend to explore is the dependency of the set of the extremal and stationary points of the function En(κ1,κ2;κv)\mathcal{E}_n(\kappa_1,\kappa_2;\boldsymbol{\kappa}_v) on the parameter κv\boldsymbol{\kappa}_v. To avoid the imaging 3D3D graphics which will not sufficiently visually demonstrate the structure of the En(κ;κv)\mathcal{E}_n(\boldsymbol{\kappa};\boldsymbol{\kappa}_v) function, we consider the structure of the levels of this function. It turns out that the picture of these levels essentially depends on the vector pv=phκv\mathbf{p}_v = p_h\boldsymbol{\kappa}_v as well as other parameters of our model. We must consider two cases.

  • The case when ω(n+1/2)>βλ\omega(n+1/2)>\beta_\lambda. In this case, the square brackets in Eq. (30) do not vanish for any values of vector κ\boldsymbol{\kappa} and remain positive.

So, when the value of ∣κv∣|\kappa_\mathrm{v}| is sufficiently small, we get a simple level structure that is demonstrated in Figure 1. The domain D00D_{00} where the condition (24) is broken, is bounded in Figure 1 by the red dashed curve. At each point on this curve, the function En(κ,κv)\mathcal{E}_n(\kappa,\kappa_\mathrm{v}) reaches a local minimum which is equal to zero.

Energy levels

Figure 1. Energy levels for (κv)1=5(\boldsymbol{\kappa}_v)_1 = 5, (κv)2=3(\boldsymbol{\kappa}_v)_2 = 3.

What happens when En(κ;κv)>0\mathcal{E}_n(\kappa;\kappa_\mathrm{v}) > 0 in the case of small value of ∣κv∣|\kappa_\mathrm{v}|? To find out, let us view a simple formula

∂E∂t=∂E∂p1p˙1+∂E∂p2p˙2,p˙i≡dpidt.\frac{\partial\mathcal{E}}{\partial t} = \frac{\partial\mathcal{E}}{\partial p_1}\dot{p}_1 + \frac{\partial\mathcal{E}}{\partial p_2}\dot{p}_2,\qquad\dot{p}_i \equiv\frac{d p_i}{d t}.

Still, we consider the integrable Hamiltonian system (3), so that the equalities p˙i=0\dot{p}_i=0 hold. But really, in many vortex systems, we must consider the various interactions between vortices. In this scenario, p˙i\dot{p}_i generally does not equal zero. Therefore, the any stable (or quasi-stable) vortex state corresponds only to a stationary point, where the equalities ∂E/∂pi=0\partial\mathcal{E}/\partial p_i=0, $i=1,2 take place. The case under study the surface En=En(κ1,κ2;κv)\mathcal{E}_n=\mathcal{E}_n(\kappa_1,\kappa_2;\boldsymbol{\kappa}_v) has not any stationary points. Consequently, vortices do not form here and the flow of fluid in the boundary layer remains laminar.

When the value of ∣κv∣|\boldsymbol{\kappa}_v| increases, the level picture becomes more complex. Increasing the value of ∣κv∣|\boldsymbol{\kappa}_v|, the domain D00D_{00} also moves so that it is located in some neighborhood of the point ((κv)1,(κv)2)((\kappa_v)_1,(\kappa_v)_2). When the number ∣κv∣|\boldsymbol{\kappa}_v| reaches and exceeds some critical value κcr=pcr/ph\kappa_{cr}=p_{cr}/p_h, the surface E=En(κ,κv)\mathcal{E}=\mathcal{E}_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_v) gets two additional local extreme points – M+M_+ and M−M_- where the equalities ∂E/∂pi=0\partial\mathcal{E}/\partial p_i=0, i=1,2i=1,2, hold. The level structure for this case is demonstrated in Figure 2. The extreme points – M+M_+ and M−M_- have positive and negative curvature correspondingly. Critical point κv=κcr\boldsymbol{\kappa}_v=\boldsymbol{\kappa}_{cr} is characterized by the following condition being fulfilled:

Energy levels

Figure 2. Energy levels for (κv)1=250(\kappa_\mathrm{v})_1 = 250, (κv)2=70(\kappa_\mathrm{v})_2 = 70.

∃ p0=phκ0∂2En(κ,κcr)∂κi∂κj∣κ=κ0=0∀ i,j=1,2.(31)\exists\ \mathbf{p}_0=p_h\boldsymbol{\kappa}_0 \qquad\left.\frac{\partial^2\mathcal{E}_n(\boldsymbol{\kappa},\boldsymbol{\kappa}_{cr})}{\partial\kappa_i\partial\kappa_j}\right|_{\boldsymbol{\kappa}=\boldsymbol{\kappa}_0}=0 \qquad\forall\,i,j=1,2 . \tag*{(31)}

It is clear that a planar point occurs at a moment when the picture in Figure 1 transforms into the picture in Figure 2, by means of increasing of the value of κv\boldsymbol{\kappa}_v. Thus, we can only observe the vortices appearing near the plane Z=0Z=0 when pv=phκv>p0\mathbf{p}_v=p_h\boldsymbol{\kappa}_v>\mathbf{p}_0. Formally, taking into account Eqs. (7) and Eq. (31), we can see that this domain for the parameter pv\mathbf{p}_v corresponds to the domain v>0\mathbf{v}>0 for the fluid velocity v\mathbf{v}. It is clear that the vanishing of the matrix Meff−1\mathbf{M}_{\mathrm{eff}}^{-1} corresponds to the vortex with ”infinitely large mass”. We will explore this surprising phenomenon later in this section.

⋆ The case when ω(n+1/2)≤βλ\omega(n+1/2)\leq\beta\lambda. In this case, the square brackets in Eq. (30) take negative values for κ∈Dneg\boldsymbol{\kappa}\in D_{neg} and they are vanish for values of κ∈∂Dneg\boldsymbol{\kappa}\in\partial D_{neg}:

Dneg:{κ∈R2∣σph2(κ12+κ22)<2βλ−ω(2n+1)}.D_{neg}:\quad\{\boldsymbol{\kappa}\in\mathbb{R}^2\mid\sigma_{ph}^2(\kappa_1^2+\kappa_2^2)<2\beta\lambda-\omega(2n+1)\}.

The relative location of the domains DnegD_{neg} and D00D_{00} depends essentially on the vector κv\boldsymbol{\kappa}_v. If the value ∣κv∣|\boldsymbol{\kappa}_v| sufficiently small, the inclusion Dneg⊂D00D_{neg}\subset D_{00} takes place. This case demonstrated on the Figure 3. Thus, we have the same case as was visualized in Figure 1, because the energy becomes an imaginary value at all inner points in the domain D00D_{00}.

Boundaries of special domains $D_{neg}$ and $D_{00}$

Figure 3. Domain’s boundaries for (κv)1=3(\kappa_v)_1 = 3, (κv)2=1(\kappa_v)_2 = 1.

As ∣κv∣|\boldsymbol{\kappa}_v| grows, the center of the domain D00D_{00} shifts along with the vector κv\boldsymbol{\kappa}_v, while the domain DnegD_{neg} stays near the origin (0,0)(0,0). This case demonstrated on the Figure 4.

Boundaries of special domains $D_{neg}$ and $D_{00}$

Figure 4. Domain’s boundaries for (κv)1=20(\kappa_v)_1 = 20, (κv)2=10(\kappa_v)_2 = 10.

Thus, we have a local energy minimum that takes negative values in this case.

Our analysis demonstrates that the point p=p0\mathbf{p}=\mathbf{p}_0, where the ”inverse effective mass” tensor vanishes, exists. Nevertheless, the existence of the vortices with an infinite masses seems unphysical. Let us analyze additional restrictions imposed on our system by quantum mechanics.

  • Let us consider the case when the local extreme points M±M_{\pm} exist (case where ∣pv∣>∣pcr∣|\mathbf{p}_v| > |\mathbf{p}_{cr}|) but the equality ∣pv−pcr∣<ϵ|\mathbf{p}_v - \mathbf{p}_{cr}| < \epsilon holds for a small value of ϵ\epsilon. Let the extreme points M±M_{\pm} have coordinates p±\mathbf{p}_{\pm}. Then, the energy En(p,pv)E_n(\mathbf{p},\mathbf{p}_v) in small neighbourhoods of the points p±\mathbf{p}_{\pm} can be written in the form

En(p,pv)−En(p±,pv)∼∑i,j=12∂2En(p±,pv)∂pi∂pj(p−p±)i(p−p±)j.(32)E_n(\mathbf{p},\mathbf{p}_v) - E_n(\mathbf{p}_{\pm},\mathbf{p}_v) \sim\sum_{i,j=1}^{2} \frac{\partial^2 E_n(\mathbf{p}_{\pm},\mathbf{p}_v)}{\partial p_i \partial p_j}(\mathbf{p}-\mathbf{p}_{\pm})_i(\mathbf{p}-\mathbf{p}_{\pm})_j. \tag*{(32)}

Because our assumptions, the second derivatives in the (32) take small values.

  • When we study the domains of physical values of EE and p\mathbf{p} near certain fixed points, we must take into account the Heisenberg uncertainty principle. In actual quantum systems, the uncertainties ΔE\Delta E and Δp\Delta\mathbf{p} in measuring energy and momentum are never zero. So, the assumption ∣Δr∣min⁡≃Rf|\Delta\mathbf{r}|_{\min} \simeq R_f for the minimal uncertainty ∣Δr∣|\Delta\mathbf{r}| of coordinate r\mathbf{r} seems quite true. Therefore, the corresponding uncertainty ∣Δp∣1|\Delta\mathbf{p}|_1 that minimizes the value ∣Δr∣ ∣Δp∣|\Delta\mathbf{r}|\,|\Delta\mathbf{p}| will be roughly the order of pℏp_{\hbar}.

As regards of the the uncertainty ΔE\Delta E, we assume that ∣ΔE∣∼Eℏ|\Delta E| \sim E_{\hbar}. Indeed, in accordance with (29), we use a ”ruler” with divisions of EℏE_{\hbar} for the energy determination. Correspondingly, the precision of any measurement will not be higher than one division. Therefore, we can assume that in a small neighbourhood of the points p±\mathbf{p}_{\pm}, the estimating equalities hold:

∣En(p,pv)−En(p±,pv)∣≃∣ΔE∣≃Eℏ,∣p−p±∣≃∣Δp∣1≃pℏ\left| E_n(\mathbf{p},\mathbf{p}_v) - E_n(\mathbf{p}_{\pm},\mathbf{p}_v) \right| \simeq\left| \Delta E \right| \simeq E_{\hbar}, \qquad \left| \mathbf{p} - \mathbf{p}_{\pm} \right| \simeq\left| \Delta\mathbf{p} \right|_1 \simeq p_{\hbar}
  • Applying the Schwarz inequality to Equation (32), the estimates from above yield an inequality

∣En(p,pv)−En(p±,pv)∣≤∥Meff−1∥∣p−p±∣12\left| E_n(\mathbf{p},\mathbf{p}_v) - E_n(\mathbf{p}_{\pm},\mathbf{p}_v) \right| \leq\left\| \mathbf{M}_{\mathrm{eff}}^{-1} \right\| \left| \mathbf{p} - \mathbf{p}_{\pm} \right|_1^2

where symbol ∥Meff−1∥\left\| \mathbf{M}_{\mathrm{eff}}^{-1} \right\| means the norm of inverse effective matrix Meff−1\mathbf{M}_{\mathrm{eff}}^{-1}. Thus, quantum mechanics, together with the specific features of the dynamical system being studied, leads to the following restriction on the matrix Meff−1\mathbf{M}_{\mathrm{eff}}^{-1}:

∥Meff−1∥≥1Mf.\left\| \mathbf{M}_{\mathrm{eff}}^{-1} \right\| \geq\frac{1}{M_{\mathrm{f}}}.

This inequality clarifies the physical meaning of the mass constant MfM_{\mathrm{f}} which was introduced at the beginning of the section 4. This constant defines the upper limit of the effective mass of the quantum vortices under study:

∥Meff∥≤MfCond⁡(Meff),\left\| \mathbf{M}_{\mathrm{eff}} \right\| \leq M_{\mathrm{f}} \operatorname{Cond}(\mathbf{M}_{\mathrm{eff}}),

where the multiplier Cond⁡(Meff)\operatorname{Cond}(\mathbf{M}_{\mathrm{eff}}) is the condition number for the matrix Meff\mathbf{M}_{\mathrm{eff}}. Looking back at the Eq. (7), we are once again convinced that vortices do not form in a ”slowly flowing fluid”.

Concluding remarks: from a single quantum vortex to the boundary layer formation

Sinse we considered a single quantum vortex ring near the impenetrable plane. The proposed model demonstrates a higher probability of the location of such vortices near the surface. Discussing turbulent motion in the boundary layer, we need to consider the many-vortex system. The simplest system here consists of a number of non-interacting or incredibly weakly interacting circular vortices with meso- and micro- sizes. The state space of such a quantum system is as follows:

HK=H1⊗⋯⊗H1⏟K,\mathbf{H}_{K} = \underbrace{\mathbf{H}_{1} \otimes\cdots\otimes\mathbf{H}_{1}}_{K},

where Hilbert spaces H1\mathbf{H}_1 are defined as in Eq. (12) for each vortex individually. We assume the vortices follow Bose statistics, so we consider a symmetric tensor product.

The full Hamiltonian takes the form

H^v=∑k=1KI⊗I⊗⋯⊗I⏟k−1⊗H^(λk,ωk)⊗I⊗I⊗⋯⊗I⏟K−k,\widehat{H}_{v}=\sum_{k=1}^{K} \underbrace{I\otimes I\otimes\cdots\otimes I}_{k-1}\otimes\widehat{H}(\lambda_k,\omega_k)\otimes\underbrace{I\otimes I\otimes\cdots\otimes I}_{K-k},

where, by virtue of spectral theorem,

H^(λk,ωk)=∑n∫dp1dp2En(p;pv)∣ψp;λk,n⟩⟨ψp;λk,n∣.\widehat{H}(\lambda_k,\omega_k)=\sum_n\int dp_1dp_2 E_n(\mathbf{p};\mathbf{p}_v)\lvert\psi_{\mathbf{p};\lambda_k,n}\rangle\langle\psi_{\mathbf{p};\lambda_k,n}\rvert.

In accordance with our previous constructions, operators H^(λk,ωk)\widehat{H}(\lambda_k,\omega_k) are self-adjoint operators. This means that full Hamiltonian H^v\widehat{H}_v is self-adjoint operator. Parameters both λk<0\lambda_k<0 and ”dimensionless frequencies” ωk\omega_k may vary from vortex to vortex. The vortices in a turbulent flow are short-lived structures [17]. Because they are located near the different points of the plane Z=0Z=0, the set of the numbers λ∈[λminλmax]\lambda\in[\lambda_{min}\lambda_{max}] models the irregularities of the surface. The permitted interval [λminλmax][\lambda_{min}\lambda_{max}] as well as the distribution law of this parameter should depends on the type of surface. This issue is a separate study item. The vortex configurations that are more complex than rings can be considered using the method suggested by the author in the paper [8].

The above studies demonstrate the existence of stable minima in the vortex energy (point M+M_+ on Figure 2), which can take positive and negative values depending on the values of the parameters λk\lambda_k and ωk\omega_k. This means that a transition to a turbulent regime in the boundary layer does not imply a jump-like energy change in a quantum system. As regards the states which allow us to interpret our system as a quasi-particle with negative effective mass (point M−M_- on Figure 2), it could be valuable for simulating superfluidity in quantum turbulent media. The author hopes to return to this topic in future works.

Appendix

In order to facilitate the understanding of the graphs and their corresponding conclusions, we present in the table some of the notation used.

NotationWhat does this value mean?The page where the value was defined
p~h\tilde{p}_{h}Scale factor which has the dimension of momentumpage 11
p≡κp~h\mathbf{p} \equiv\kappa\tilde{p}_{h}Full vortex momentumpage 5
pv≡κvp~h\mathbf{p}_{v} \equiv\kappa_{v}\tilde{p}_{h}Additional vortex momentum caused by the movement of the surrounding fluidpage 5
pcr≡κcrp~h\mathbf{p}_{cr} \equiv\kappa_{cr}\tilde{p}_{h}Critical value of pv\mathbf{p}_{v} when the planar point on graph En(κ,κv)\mathcal{E}_{n}(\kappa,\kappa_{v}) appearspage 17
p0≡κ0p~h\mathbf{p}_{0} \equiv\kappa_{0}\tilde{p}_{h}Coordinates of the planar point, where components of ∂2En/∂pi∂pj\partial^{2}\mathcal{E}_{n}/\partial p_{i}\partial p_{j} vanishpage 17
p±≡κ±p~h\mathbf{p}_{\pm} \equiv\kappa_{\pm}\tilde{p}_{h}Coordinates of the local stationary points M±M_{\pm} which appear when ∣pv∣>∣pcr∣\lvert\mathbf{p}_{v}\rvert> \lvert\mathbf{p}_{cr}\rvertpage 18

Below, we provide the constants and quantum numbers employed in creating the graphs.

ConstantFigures 1, 2Figures 3, 4The page where the value was defined
σph2\sigma_{ph}^{2}10−610^{-6}10−610^{-6}page 8
βλ\beta_{\lambda}2⋅10−52\cdot10^{-5}3⋅10−53\cdot10^{-5}page 12
ω\omega10−410^{-4}14⋅10−614\cdot10^{-6}page 4
Quantum number nn1111page 8

The constants v0v_{0}, ϱ0\varrho_{0}, RfR_{f}, μ0\mu_{0} do not directly influence both the ”dimensionless circulation” γλ(p,pv;n)\gamma_{\lambda}(\mathbf{p},\mathbf{p}_{v};n) and the ”dimensionless energy” En(κ,κv)\mathcal{E}_{n}(\kappa,\kappa_{v}). Their effects are mediated solely through the values of the constants σph\sigma_{ph} and βλ\beta_{\lambda}. As noted earlier, the constants α\alpha, ω\omega and EλE_{\lambda}, can vary from one vortex to another.

References

  1. [1]H. Schlichting. Grenzschicht-Theorie. Verlad G. Braun, Karlsruhe (1951).
  2. [2]P.G. Saffman, Vortex dynamics. Cambrige Univ. Press, (1992).
  3. [3]S.A. Lomov, I.S. Lomov. Osnovy matematicheskoi teorii pogranichnogo sloya (The foundations of the mathematical theory of a boundary layer) Moscow: Moscow State University Press. (2011). In Russian, ISBN 978-5-211-05843-9.
  4. [4]G. W. Stagg, N. G. Parker, and C. F. Barenghi // Phys. Rev. Lett. 118, 135301 (2017).
  5. [5]S.V. Talalov // Chaos, Solitons and Fractals. 185. 115097. (2024).
  6. [6]S.V. Talalov // Chaos, Solitons and Fractals. 196. 116394. (2025).
  7. [7]R.J. Donnely. Quantum Vortices in Helium II Cambrige Univ. Press. (1991).
  8. [8]S.V. Talalov // Chaos, Solitons and Fractals. 202. 117582. (2026).
  9. [9]S.V. Alekseenko, P.A. Kuibin, V.L. Okulov, Theory of concentrated vortices. Springer-Verlag, Berlin Heidelberg (2007).DOI
  10. [10]S.K. Nemirovskii. Gydrodynamics of quantum fluids. Waves, vortices, turbulence. Part II. Quantum vortices, superfluid turbulence. Novosibirsk: Siberian Branch of the Russian Academy of Sciences (2016) In Russian, ISBN 978-5-7692-1482-0
  11. [11]G.K. Batchelor, An Introducton to Fluid Dynamics. Cambridge Univ. Press. (1970).
  12. [12]M. Reed, B. Simon. Methods of Modern Mathematical Physics. V.2. Fourier Analysis, Self-Adjointness. Academic Press: New York London (1975).
  13. [13]T.A. Filatova, A.I. Shafarevich // Theor. Math. Phys. 164. Issue 2. pp. 1081 - 1088 (2010).
  14. [14]F.A. Berezin, L.D. Faddeev // Sov. Sci. Doklady. 137, No 5, 1011 (in Russian, 1961).
  15. [15]Yu. M. Shirokov // Theor. Math. Phys. 41. pp. 1031 - 1038 (1979).
  16. [16]F.A. Berezin, M.A. Shubin, The Schrödinger Equation. Kluwer Academic Publisher, Netherlands. (1991).
  17. [17]H. Tennekes. in book: Handbook of Turbulence. ed. by W. Frost and T.H. Moulden. Plenum press, N.Y. (1977).

Paper details

vlnc:W_7cstfece
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