Elsevier

Annals of Physics

Volume 418, July 2020, 168196
Annals of Physics

Multi-valued vortex solutions to the Schrödinger equation and radiation

https://doi.org/10.1016/j.aop.2020.168196Get rights and content

Highlights

  • A new explanation for why quantum mechanical wave functions are single-valued.

  • Transient multi-valued states might exist that emit radiation.

  • Aharonov–Bohm systems should emit radiation dependent on magnetic flux.

  • A new and large class of knotted vortex solutions are presented for study.

Abstract

This paper addresses the single-valued requirement for quantum wave functions when they are analytically continued in the spatial coordinates. This is particularly relevant for de Broglie–Bohm, hydrodynamic, or stochastic models of quantum mechanics where the physical basis for single-valuedness has been questioned. It first constructs a large class of multi-valued wave functions based on knotted vortex filaments familiar in fluid mechanics, and then it argues that for free particles, these systems will likely radiate electromagnetic radiation if they are charged or have multipolar moments, and only if they are single-valued will they definitely be radiationless. Thus, it is proposed that electromagnetic radiation is possibly the mechanism that causes quantum wave functions to relax to states of single-valuedness, and that multi-valued states might possibly exist in nature for transient periods of time. If true, this would be a modification to the standard quantum mechanical formalism. A prediction is made that electrons in vector Aharonov–Bohm experiments should radiate energy at a rate dependent on the solenoid’s magnetic flux.

Introduction

The question whether the Schrödinger equation must always be single-valued occupied a number of the founders of early quantum theory, starting with Schrödinger himself [1], then notably Pauli [2], and subsequently a number of others [3], [4], [5]. The question received renewed interest after the Aharonov–Bohm effect was discovered [6]. The general consensus reached was that even in the Aharonov–Bohm systems the wave function must still be single-valued [4], [5], [7], but there were some dissenters [8]. The subject received a new burst of interest from the insights of Wallstrom regarding the non-obvious motivation for the single-valued constraint in de Broglie–Bohm, hydrodynamic, and stochastic formulations of quantum mechanics [9], [10], [11], [12], [13]. Goldstein also independently commented on multi-valued wave functions in stochastic mechanics [14]. Derakhshani has recently suggested an interesting explanation for the single-valued constraint based on zitterbewegung [15], [16]. Smolin has also considered this issue [17]. Here I take a different approach to this question which is based on electromagnetic radiation.

The prototypical example of a multi-valued wave function is one of the form ψ(x)=f(x)eilzφ, where f(x) is a single-valued function and where φ is the azimuthal angle, and lz is a real constant. The wave function is single-valued only if lz=±n,nZ. The single-valued constraint leads to quantization effects. In this simple case, when lz is not an integer, we can consider the wavefunction to have multiple Riemann sheets that differ from one another by constant phase factors of the form eilz2πm for m integer. The fact that the different sheets differ by a constant phase factor which is independent of x, is important, and consequently I shall consider only multi-valued wave functions which have this property here.

Firstly I introduce a broad class of suitable multi-valued wave functions. I do this by borrowing results from the theory of vortex filaments in fluid mechanics. The fluid version of the Biot–Savart law allows one to construct a local potential function which is generated by a vortex filament taking an arbitrary stringlike shape, either closed or open . This includes arbitrary knots and links, and this function is generally multi-valued when analytically continued in x. From these known solutions, I construct multi-valued solutions of the Schrödinger equation. For the usual azimuthal example, the string would be the whole z axis, and be infinitely long. The analysis presented applies to such cases as well as to knots and links. The subject of quantum vortices applied to the Schrödinger equation is not new [18], [19], [20], [21]. Here we generalize these treatments to multi-valued wave functions.

Then I show, by using a nonlinear identity of the Schrödinger equation, that the multi-valued linear Schrödinger equation can be replaced by an equivalent single-valued but nonlinear equation. I then argue from this that if the particle is charged, then the nonlinear term will likely produce spontaneous bremsstrahlung even for a free particle, so that the multi-valued solution may not be stable to radiative decay. Consequently, I argue that the equilibrium state of the system that is approached must be single-valued, and since this could account for the observed fact of single-valuedness, at least for charged particles, that perhaps the single-valued condition of quantum mechanics might not be universally true, and in particular it might be violated for short time periods where, after a collapse of the wave function for example, a transient multi-valued wave function is produced which then subsequently decays into a single-valued one. Although this radiative relaxation to single-valuedness is easiest to understand for charged particles, it might be expected to hold for neutral particles too if they have some multipolar electromagnetic moments, since these too would radiate when accelerated. Neutrons or neutral atoms are examples. So we might expect that multi-valued wave functions if ever created would typically be short-lived for them as well. The exceptions might be neutrinos, or dark matter particles (if they exist), where the lack of any electromagnetic interaction would allow multi-valued wave functions to persist for longer periods of time. This transient multi-valuedness might yield experimentally detectable effects. We discuss one such test below in the vector Aharonov–Bohm discussion. Also, these results might also prove to be useful in the mathematical theory of knots.

When restricted by the single-valued constraint, the vortex solutions here are similar to the quantum vortex solutions in superconductors, and superfluids. Here they are considered as solutions to the single particle Schrödinger equation with or without a potential. I am not aware of the general knot solutions found here having been considered previously in the physics literature, although a number of special cases were elegantly analyzed in [19].

Section snippets

A multi-valued initial state based on a knotted vortex filament solution using the Biot–Savart law

It is well known that the single particle Schrödinger equation can be cast in the form of inviscid fluid mechanical equations by the Madelung transformation [22], [23]. The Euler equations take the following form in a conservative force field V with pressure p, and density ρ t+uu=1ρpVand with the Madelung ansatz for the quantum force, this becomes t+uu=Q+Vwhere Q=22mρρand where ρ is a conserved density. Of course, these equations are identical to those in Bohmian mechanics for

The Schrödinger equation in the de Broglie-Bohm-Madelung pilot wave formalism

We consider the single-particle Schrödinger equation 22m+Vψ=iψtand we write ψ(x,t)=R(x,t)eiS(x,t)where both R(x) and S(x) are real functions. The guidance equation is given by dX(t)dt=1mS(X(t))so we can equate the fluid velocity in a hydrodynamic picture with this. We wish to incorporate the potential from the vortex solution into the initial state of a wave function. We assume at an initial time t=0 that we have S(x,0)m=ϕf(x)+ϕw(x)where ϕf is the knot potential calculated above, and ex

Requirement that the filament must be a nodal curve for the wave function

It is typically assumed that if a Schrödinger wave function has a vortex filament, that it must be a nodal filament, so that the wave function vanishes along it. The reason for this is continuity of the wave function. In a neighborhood of a point on the vortex filament, the wave function’s phase takes on multiple values. If it does not vanish on the filament, then the different phases in a neighborhood would lead to a discontinuity. Since we typically assume that the quantum mechanical wave

Comparison with Dirac strings

When the quantization condition that ensures single-valuedness (20) is satisfied, the wave function with a vortex instantaneously looks similar to the wave function for a charged particle in the presence of a Dirac string [39], [40], [41], [42], except that there is no monopole at the end of our vortex string as it is either a closed curve, or one that goes off to infinity in both directions. A reasonable question to ask in this case is whether our vortex curves are essentially Dirac strings

An identity for the Schrödinger equation

The following two equations are equivalent at points of analyticity for the real functions R and S 22m+VReiS=itReiSis equivalent to ν22m+V+22mν21RRReiSν=iνtReiSν where ν is an arbitrary complex-valued constant. Since both sides of (28) are analytic functions of ν, if the equation is true for real ν then it is automatically true for complex ν by analytic continuations. Although the identity is true for complex ν, in this paper only real values of ν will be required. An

Bremsstrahlung for charged particles

For a classical charged particle undergoing acceleration, the instantaneous radiated power is given by Larmor’s formula PRadCl=23q2c3a2

For a quantum particle, with a wave function ψ which passes through a force field producing radiation, there are a number of calculations of Larmor’s formula for scalar particles [52], [53], [54], [55]. These calculations are based on rigorous second-quantized scalar electrodynamics. Since we are considering a non-plane wave packet here, I find it more suitable

Linear superposition

If we have two or more multi-valued solutions to Schrödinger equation, with different values of the vorticity constant, then their suitably normalized superposition is also a solution. However, the identity (28) we used for a single vorticity constant will not work anymore to produce a single-valued wave function in this case. Let ψj be a solution to (12) with a vorticity constant Γj. Let there be M such functions, and consider their normalized superposition ψ=1Nj=1Nψjwhere N is a

Some comments on the Aharonov–Bohm effect, and predicted energy loss due to radiation

Consider an ideal cylindrical solenoid whose centerline is the z axis with a static magnetic flux inside it. Let it be infinitely long in both directions. Next consider a Schrödinger wave packet of charged particles passing around this solenoid, and assume that the wave function vanishes at the solenoid’s surface and inside it. The space is no longer simply connected, and the phase factor is therefore not automatically single-valued as we analytically continue the wave function around the

Some comments on dissipation mechanisms for achieving single-valued equilibrium

It is not clear how to develop a theory that would take into account the energy loss due to radiation, and the influence that this would have on the wave function. We might try and invoke a perturbation expansion along the lines of conventional theory. It is not obvious that the vortex would be preserved. If it is, and if a single-valued limit is to be reached, then (44) must be satisfied. This means that either the mass or the vorticity of the wave function must change. Particle masses are

Topological considerations

If we ignore radiative energy loss, then the free particle Shrödinger equation is equivalent to an inviscid compressible Eulerian fluid described by the Madelung theory. This would be the case if the vortex is resulting in a single-valued linear wave function, so that it definitely does not radiate. Such fluids with vortex knots and links have been the subject of much research, and besides the circulation and the helicity, there are other topological invariants. For example, Liu and Ricca have

Discussion

A large class of multi-valued Schrödinger wave functions have been proposed and considered here, and it has been shown, using a nonlinear identity of Schrödinger’s equation, that they can be transformed into solutions of a single-valued but nonlinear differential equation. These were based on knotted vortex filament fluid methods. The Biot–Savart law commonly used in vortex analysis was supplemented by an additional factor along the filament to create wave functions which had the necessary

CRediT authorship contribution statement

Mark Davidson: Conceptualization, formal analysis, Methodology, Data curation, Writing - original draft, Writing - review and editing, Validation, Visualization, Investigation.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

I wish to acknowledge Fritz Bopp for helpful correspondence. I also wish to thank one of the reviewers for insightful suggestions.

References (87)

  • WallstromTimothy C.

    Phys. Lett. A

    (1994)
  • BliokhK.Y. et al.

    Phys. Rep.

    (2017)
  • KauffmanL.H.
  • PortmannFabian et al.

    J. Math. Pures Appl.

    (2018)
  • DavidsonMark

    Physica A

    (1979)
  • DavidsonMark P.

    Ann. Physics

    (2007)
  • ChoudhuryS. Rai et al.

    Phys. Lett. A

    (2019)
  • JingJian et al.

    Phys. Lett. B

    (2017)
  • MacdougallJames et al.

    Phys. Lett. A

    (2015)
  • SingletonDouglas et al.

    Phys. Lett. B

    (2013)
  • LiuXin et al.

    Procedia IUTAM

    (2013)
  • SchrödingerErwin

    Ann. Phys.

    (1938)
  • PauliWolfgang

    Helv. Phys. Acta

    (1939)
  • RiessJ.

    Helv. Phys. Acta

    (1973)
  • MerzbacherE.

    Amer. J. Phys.

    (1962)
  • YangC.N.
  • AharonovY. et al.

    Phys. Rev.

    (1959)
  • PeshkinMurray et al.
  • SivermanM.P.

    Lett. Nuovo Cimento (1971-1985)

    (1985)
  • WallstromTimothy C.

    Found. Phys. Lett.

    (1989)
  • WallstromTimothy C.

    Proc. Amer. Math. Soc.

    (1990)
  • WallstromTimothy C.

    Trans. Amer. Math. Soc.

    (1990)
  • WallstromTimothy C.

    Phys. Rev. A

    (1994)
  • GoldsteinSheldon

    J. Stat. Phys.

    (1987)
  • DerakhshaniMaaneli

    A suggested answer to Wallstrom’s criticism: Zitterbewegung stochastic mechanics I

    (2015)
  • DerakhshaniMaaneli

    A suggested answer to Wallstrom’s criticism: Zitterbewegung stochastic mechanics II

    (2016)
  • SmolinLee

    Could quantum mechanics be an approximation to another theory?

    (2006)
  • BerryM.V. et al.

    J. Phys. A: Math. Gen.

    (2001)
  • Bialynicki-BirulaIwo et al.

    Phys. Rev. A

    (2000)
  • LloydS.M. et al.

    Rev. Modern Phys.

    (2017)
  • MadelungE.

    Naturwissenschaften

    (1926)
  • MadelungE.

    Z. Phys.

    (1927)
  • BohmDavid

    Phys. Rev.

    (1952)
  • NelsonEdward

    Dynamical Theories of Brownian Motion

    (1967)
  • SaffmanP.G.

    Vortex Dynamics

    (1992)
  • KauffmanLouis H.

    On Knots

    (1987)
  • JacksonJohn David

    Classical Electrodynamics

    (1999)
  • FranklinJerrold

    Classical Electromagnetism: Second Edition

    (2017)
  • PortmannFabian et al.

    J. Geom. Anal.

    (2019)
  • GrossPaul W. et al.

    Electromagnetic Theory and Computation: A Topological Approach

    (2004)
  • KotiugaP.R.

    J. Appl. Phys.

    (1987)
  • BinyshJack et al.

    J. Phys. A

    (2018)
  • SeifertH.

    Math. Ann.

    (1935)
  • View full text