The Coupled Klein–Gordon Equation in the Non-Quantum Limit
Introduction
The limit in one sentence. The connection enters the parent's factorisation as one additive term inside the phase's gradient, so it changes the mass shell, the transport current and the criterion, and it changes no order: order zero becomes a mass shell in the mechanical four-momentum, and its gradient is the Lorentz force.
The parent proves the factorisation and prints the pair. The Coupled Klein–Gordon Equation in Biquaternionic Polar Variables shows that for the central field the covariant derivative factors exactly, $\tilde\nabla_{\mathrm{gco},\Phi,\mu}(\varrho\,e^{i\theta_{\mathrm{pha}}}e_{t})=e^{i\theta_{\mathrm{pha}}}\left(\partial_\mu \varrho+i\varrho\kappa^{\mathrm{kin}}_\mu\right)e_{t}$ with $\kappa^{\mathrm{kin}}_\mu=\partial_\mu\theta_{\mathrm{pha}}+\frac{ϙ}{\hbar}A_\mu$, and that the square of that parenthesis gives the two boxed equations. The coupling is one term inside one square, and that is the whole of its structure.
The consequence for the limit is one substitution. Every display of the free article survives with $S\to$ the mechanical action, that is with $\partial_\mu S$ replaced by $\partial_\mu S$ minus the potential's contribution. The orders are untouched because the substitution is inside a term and does not create one.
The article's sharpest result is the force, and it is order zero alone. Differentiating the mass shell in the mechanical four-momentum and dividing by the mechanical energy returns
$$ \frac{d\vec\pi}{dt'}=q\bigl(\vec E+\vec v\times\vec B\bigr), \qquad \vec v=\frac{c^{2}\vec\pi}{\mathcal E}, \qquad \mathcal E=E-q\varphi , $$
which is Newton's law with the magnetic force, with the relativistic relation between momentum and velocity. Two identities carry it: the mixed derivatives of the action commute, and the mechanical four-momentum is a gradient minus a potential.
The criterion must be written on the mechanical momentum, and here the argument is the free one. $\vec\pi=\vec\nabla_\Phi S-q\vec A$ is gauge invariant and $\vec\nabla_\Phi S$ is not, so a parameter defined on the canonical gradient changes value under a gauge transformation. A limit whose smallness depends on the gauge is not a statement about the field.
The second parameter is the free one. $\delta=1/(k_{m}L)$ does not involve the connection, because the Compton wavelength is the field's and not the potential's. The coupling changes neither parameter.
What the article establishes, in five sentences. The parent's pair is the free pair with two substitutions, and the quantum potential is untouched. Order zero is the mass shell in the mechanical four-momentum, $\mathcal E^{2}/c^{2}-\lvert\vec\pi\rvert^{2}=m^{2}c^{2}$, and its gradient is the Lorentz force with $\vec v=c^{2}\vec\pi/\mathcal E$. Order one is the transport equation with the mechanical current, which is gauge invariant and exact. The vorticity is the field's for the current velocity $\vec\pi/m$ and the field's plus an energy-gradient term for the kinematic velocity, and that split has no counterpart in the non-relativistic family because there the two velocities coincide.
What is not new. The covariant Madelung pair for a charged scalar field, the mass shell in mechanical variables, the Lorentz force from a Hamilton–Jacobi equation and the vorticity of a charged flow are all standard. Added here: the ordering of the parent's own factorisation against one parameter, the exact form of the vorticity split with the energy-gradient term named, and the statement that the coupling changes the value of every leading term and the accuracy of none.
The sign, stated once. The parent writes the coupling as $ϙ$ with $ϙ=-q$; its mass shell carries $+ϙ c a'_{t}$ in the mechanical energy and its transport carries $-q\vec A$ in the current. Every display here uses $q$, the electric charge, $\vec\pi=\vec\nabla_\Phi S-q\vec A$ and $\mathcal E=E-q\varphi$ with $E=-c\,\partial_{r'_{t}}S$.
Conventions. Those of the coupled polar parent: $\mathbb{B}=\mathbb{C}\otimes_{\mathbb{R}}\mathbb{H}$ with $e_{t}=1$ and $\vec e_\mu^{2}=-e_{t}$; $\mathrm{Sc}$ the coefficient of $e_{t}$; the word $\tilde\Xi=\varrho\,e^{i\theta_{\mathrm{pha}}}\varepsilon e_{t}$; the connection $\tilde A_\Phi=i(\varphi/c)e_{t}+\vec A$; $\tilde\nabla_{\mathrm{gco},\Phi,\mu}=\partial_\mu+iq\tilde A_\mu/\hbar$; $\Box_\Phi=\Delta_\Phi-\partial^{2}/\partial r_{t}^{\prime\,2}$ with $r'_{t}=ct'$; $\rho_{\mathrm{num}}=\varrho^{2}$; $k_{m}=mc/\hbar$; $\epsilon=1/(\pi_{\Phi}L)$ with $\pi_{\Phi}=\lvert\vec\pi\rvert/\hbar$.
The Factorisation and the Pair
The parent's exact identity.
$$ \tilde\nabla_{\mathrm{gco},\Phi,\mu}\bigl(\varrho\,e^{i\theta_{\mathrm{pha}}}e_{t}\bigr) =e^{i\theta_{\mathrm{pha}}}\Bigl(\partial_\mu \varrho+i\,\varrho\,\kappa^{\mathrm{kin}}_\mu\Bigr)e_{t}, \qquad \kappa^{\mathrm{kin}}_\mu=\partial_\mu\theta_{\mathrm{pha}}+\frac{ϙ}{\hbar}A_\mu, $$
and the parenthesis is central, so the derivation is a computation among scalars. The connection enters once, as an additive term in one scalar, and every later statement of this article is that term read at one order or another.
The pair, in the article's letters.
$$ \Bigl(\frac{\partial S}{\partial r'_{t}}+\frac{q\varphi}{c}\Bigr)^{2}-\lvert\vec\nabla_\Phi S-q\vec A\rvert^{2}=m^{2}c^{2}-\hbar^{2}\frac{\Box_\Phi \varrho}{\varrho}, \qquad \vec\nabla_\Phi\cdot\bigl(\rho_{\mathrm{num}}\vec\pi\bigr)+\frac{\partial}{\partial r'_{t}}\Bigl(\rho_{\mathrm{num}}\Bigl(\partial_{r'_{t}}S+\frac{q\varphi}{c}\Bigr)\Bigr)=0 . $$
The first is the amplitude equation and the second the transport equation, and both are the free ones with the two substitutions. The temporal bracket of the second is $-\mathcal E/c$, because $E=-c\,\partial_{r'_{t}}S$, so the two equations already speak in the mechanical variables and the sign is the one the free pair has.
The mass shell in mechanical variables.
$$ \frac{\mathcal E^{2}}{c^{2}}-\lvert\vec\pi\rvert^{2}=m^{2}c^{2}+2mQ_{\mathrm{qnt}}, \qquad Q_{\mathrm{qnt}}=-\frac{\hbar^{2}}{2m}\frac{\Box_\Phi \varrho}{\varrho}, \qquad \mathcal E=E-q\varphi . $$
The quantum potential is the free one, and the parent's reason is one line: it is built from the scale, the scale is inert under the gauge group, and no derivative of the potential reaches it. The coupling shifts the two leading terms and leaves the order-two term alone.
The Criterion on the Mechanical Momentum
The parameter. $\epsilon=1/(\pi_{\Phi}L)$ with $\pi_{\Phi}=\lvert\vec\pi\rvert/\hbar$ and $L^{-1}=\lvert\vec\nabla_\Phi\ln \varrho\rvert$. The covariant derivative differentiates the phase's gradient minus the potential's contribution, so that combination is what is large and what the scale's gradient is measured against.
Why the canonical gradient fails. Under $\theta_{\mathrm{pha}}\to\theta_{\mathrm{pha}}+q\Gamma/\hbar$ the canonical gradient shifts by $q\vec\nabla_\Phi\Gamma$, while $\vec A$ shifts by $\vec\nabla_\Phi\Gamma$ and the two cancel inside $\vec\pi$. A criterion on the canonical gradient would be gauge dependent, and a limit whose validity depends on the gauge is a statement about the chart and not about the field.
The order table is the free table with one substitution.
| object | order in $\epsilon$ | kept by the limit |
|---|---|---|
| $\mathcal E$, $\vec\pi$ | $0$ | yes |
| $\hbar\,\vec\nabla_\Phi\ln \varrho$ | $1$ | no |
| $\hbar^{2}\,\Box_\Phi \varrho/\varrho$ | $2$ | no |
Nothing in the table mentions the potential's strength. A large potential shifts the leading terms and creates no parameter, so the coupling costs no accuracy. The statement is the same as in the non-relativistic family and the reason is the same: the parameter counts derivatives of the scale.
Order Zero: the Mass Shell and the Lorentz Force
Delete the quantum potential.
$$ \frac{\mathcal E^{2}}{c^{2}}-\lvert\vec\pi\rvert^{2}=m^{2}c^{2}, $$
the mass shell of a charge $q$ in the potentials, written in the mechanical energy and momentum.
The two mechanical variables are the two gauge-invariant combinations. $\mathcal E$ is the temporal one and $\vec\pi$ the spatial one, and the free article's two invariants are their limits. The eikonal is again a normalisation and not an energy balance: it fixes the norm of the mechanical four-momentum, and the potential moves the norm's argument without moving the statement.
The flow. Differentiating the shell along the characteristics,
$$
\vec v=\frac{d\vec x}{dt'}=\frac{c^{2}\vec\pi}{\mathcal E},
\qquad
\lvert\vec v\rvert
so the velocity is the mechanical momentum over the mechanical energy, and the bound is the shell's own content.
The gradient of the shell is the Lorentz force, and the identity is exact. With $\vec E=-\vec\nabla_\Phi\varphi-\partial_{t'}\vec A$ and $\vec B=\vec\nabla_\Phi\times\vec A$,
$$ -\frac{c^{2}}{2\mathcal E}\,\partial_{i}\Bigl(\frac{\mathcal E^{2}}{c^{2}}-\lvert\vec\pi\rvert^{2}\Bigr)=\frac{d\pi_{i}}{dt'}-q\bigl(\vec E+\vec v\times\vec B\bigr)_{i}, $$
so on the shell the left side vanishes and
$$ \boxed{\ \frac{d\vec\pi}{dt'}=q\bigl(\vec E+\vec v\times\vec B\bigr),\qquad \vec v=\frac{c^{2}\vec\pi}{\mathcal E}\ } $$
Two cancellations carry it and both are identities. The mixed derivatives of $S$ commute, which produces $-\partial_{t'}\pi_{i}-q\partial_{t'}A_{i}$; and $\vec\pi$ is a gradient minus $q\vec A$, which produces the curl of $\vec A$ with the sign that makes it $\vec B$.
The force is the same bracket as the non-relativistic companion's and the kinematics is not. There $\vec\pi=m\vec v$ and the magnetic term is $q\vec v\times\vec B$ with the Newtonian velocity; here $\vec\pi=\gamma m\vec v$ and the same bracket is written in the same letters. A reader who compares the two families should compare $\vec\pi$ with $m\vec v$ and not the force with itself.
The magneto-static check. For $\vec E=0$ and a uniform $\vec B$, the shell's characters are helices with $p_{\parallel}$ constant and $p_{\perp}$ turning at $qB/\gamma m$, the relativistic Larmor rate, which reduces to the classical one as $\beta\to0$. The limit has produced the correct relativistic motion from order zero alone, with no spin term and no amplitude equation.
Order One: the Mechanical Current and the Two Vorticities
The transport equation.
$$ \frac{\partial\rho_{\mathrm{chg}}}{\partial r'_{t}}+\vec\nabla_\Phi\cdot\vec j_{\mathrm{mech}}=0, \qquad \vec j_{\mathrm{mech}}=\frac{\rho_{\mathrm{num}}}{m}\vec\pi, \qquad \rho_{\mathrm{chg}}=\frac{\rho_{\mathrm{num}}\,\mathcal E}{mc^{2}} , $$
with the mechanical energy in the charge density and the mechanical momentum in the current, the material length $r'_{t}=ct'$ supplying the one factor between the two forms. Both are gauge invariant and the free pair's were not: that is what the coupling buys.
It is exact and the limit does not touch it. No term is dropped at any order, and the equation is the free one with the substitutions. The current's carrier is the mechanical momentum, and the reason is the one the variational companion states: the conjugate momentum of the phase is $\rho\vec\pi$ and not $\rho\vec\nabla_\Phi S$.
The current velocity's vorticity is the field's, exactly. With $\vec V=\vec\pi/m$,
$$ \vec\nabla_\Phi\times\vec V=\frac{1}{m}\vec\nabla_\Phi\times\bigl(\vec\nabla_\Phi S-q\vec A\bigr)=-\frac{q}{m}\vec B , $$
because the curl of a gradient vanishes identically. The canonical flow is irrotational and the current flow is rotational exactly where the field is, and both statements are exact and inherited.
The kinematic velocity's vorticity is not, and the difference is a relativistic effect the corpus should name. With $\vec v=c^{2}\vec\pi/\mathcal E$ and $\Gamma=\mathcal E/c^{2}$ a scalar field,
$$ \vec\nabla_\Phi\times\vec v=\vec\nabla_\Phi\times\Bigl(\frac{\vec\pi}{\Gamma}\Bigr) =-\frac{q}{\Gamma}\vec B+\vec\nabla_\Phi\Bigl(\frac{1}{\Gamma}\Bigr)\times\vec\pi . $$
The second term is the energy's gradient and it is absent from the non-relativistic family, because there $\Gamma=m$ is constant and the two velocities coincide. A relativistic flow in a static electric field is rotational even where $\vec B$ vanishes, and the corpus's vorticity statement is about the current velocity and not about the trajectory's.
The circulation inherits the quantisation with the flux. For a loop bounding a surface that no node crosses,
$$ \oint\vec\pi\cdot\mathrm{d}\vec l=2\pi\hbar\,n-q\Phi_{B}, $$
the same statement as the non-relativistic companion's, and here it is the relativistic circulation of the mechanical momentum. The limit neither produces it nor modifies it.
Order Two: the Quantum Potential, Unchanged
The discarded term.
$$ Q_{\mathrm{qnt}}=-\frac{\hbar^{2}}{2m}\frac{\Box_\Phi \varrho}{\varrho} =-\frac{\hbar^{2}}{2m}\left[\frac{1}{2}\Box_\Phi\ln\rho_{\mathrm{num}}+\frac{1}{4}\left(\partial\ln\rho_{\mathrm{num}}\right)^{2}\right], $$
the free article's two forms, with the d'Alembertian and with the density alone.
Nothing about the coupling appears in either form. The scale is invariant under the gauge group, the logarithm of the density is invariant, and the second derivatives are ordinary. The quantum potential is the free one, term for term, and the parent states it as its first finding.
Its size is the free size in both denominators. Against the rest energy it is $\delta^{2}=\hbar^{2}/(m^{2}c^{2}L^{2})$; against the kinetic energy it is $\epsilon^{2}$. Neither involves $q$, so a strong potential does not make the approximation worse, and a rapidly varying one does — which is the hypothesis the next section states.
What the Coupling Costs and What It Does Not
It costs nothing in accuracy. The three orders are the free three, the deleted term is the free term, and both parameters are the free parameters. The coupling is leading in the value of every term and absent from the count of all of them.
It costs one hypothesis, and the hypothesis is stated and not met. $\vec\pi$ contains $\vec A$, so the criterion's meaning requires the potential to vary slowly on the length $L$: a potential varying as fast as the phase would make the leading term vary as fast as the scale, and the expansion would be in the wrong quantity. No article in the corpus states that hypothesis, and this one states it as a gap.
It gains three exact statements. The current is gauge invariant. The current velocity's vorticity is the field's. The circulation is quantised with the flux. None is semiclassical; all three are the exact pair's, and the limit's role is to leave them standing.
It changes one thing qualitatively, and it is the vorticities. In the non-relativistic family the kinematic and current velocities coincide, so there is one vorticity statement; here they differ by $\gamma$ and there are two, and the second carries the energy's gradient. That is a relativistic effect of the limit and not of the coupling, and it would be present with the field switched off.
Verification
The Lorentz-force identity was checked with no assumption on the potentials. With arbitrary $S$, $\varphi$ and $\vec A$, $\vec\pi=\vec\nabla_\Phi S-q\vec A$, $\mathcal E=-\partial_{t'}S-q\varphi$ and $\vec v=c^{2}\vec\pi/\mathcal E$, the three components of
$$ -\frac{c^{2}}{2\mathcal E}\partial_{i}\Bigl(\frac{\mathcal E^{2}}{c^{2}}-\lvert\vec\pi\rvert^{2}\Bigr)-\Bigl(\frac{d\pi_{i}}{dt'}-q(\vec E+\vec v\times\vec B)_{i}\Bigr) $$
vanish identically. The check was made with $c\neq1$ so that the factors of $c$ are exercised, and the sample point had $\lvert\vec v\rvert The vorticity was checked by direct differentiation. $\vec\nabla_\Phi\times(\vec\nabla_\Phi S-q\vec A)=-q\vec\nabla_\Phi\times\vec A$ because the curl of a gradient vanishes; the second term of the kinematic vorticity follows from $\vec\nabla_\Phi\times(f\vec X)=f\vec\nabla_\Phi\times\vec X+\vec\nabla_\Phi f\times\vec X$ with $f=1/\Gamma$. The quantum potential was checked for invariance. It contains $\varrho$ and $\Box_\Phi \varrho$; under the gauge transformation the scale is unchanged, so $Q_{\mathrm{qnt}}$ is unchanged term for term. The pair was checked against the free pair. Setting $\varphi=0$ and $\vec A=0$ returns the free article's two boxed lines term for term, so the coupling enters as a substitution and not as a modification. What was not checked. No numerical solution of the coupled time-dependent problem was run, no flux was computed, and no helices were integrated. The claims are about terms and identities, and both were verified as algebra. The parent's covariant derivative factors exactly on the central field, with the connection as one additive term in the phase's gradient. The limit is therefore one substitution in the free pair, and it changes no order. Order zero is the mass shell in the mechanical energy and momentum, and its gradient is the Lorentz force with $\vec v=c^{2}\vec\pi/\mathcal E$. The force is the same bracket as the non-relativistic family's and the kinematics is not: $\vec\pi=\gamma m\vec v$ here and $\vec\pi=m\vec v$ there. Order one is the transport equation in the mechanical current, gauge invariant and exact, and it carries two exact companions the limit inherits: the current velocity's vorticity is $-(q/m)\vec B$, and the circulation is $2\pi\hbar n-q\Phi_{B}$. The kinematic velocity's vorticity differs from the current velocity's by the energy's gradient, a relativistic effect that the non-relativistic family cannot have. Order two is the free quantum potential, with the free size in both denominators. The coupling changes the value of every leading term and the accuracy of none, and it costs one unstated hypothesis: the potential must be slowly varying on the scale $L$.Summary
Summary of Notation
Symbol
Meaning
$\tilde\Xi=\varrho\,e^{i\theta_{\mathrm{pha}}}\varepsilon e_{t}$
The parent's polar word on the central class
$\varrho>0$, $\rho_{\mathrm{num}}=\varrho^{2}$
The scale; the proper density
$S=\hbar\theta_{\mathrm{pha}}$, $E=-c\,\partial_{r'_{t}}S$
The action; the canonical energy
$\tilde A_\Phi=i(\varphi/c)e_{t}+\vec A$
The connection
$q$; the parent's $ϙ=-q$
The electric charge; the parent's coupling
$\vec\pi=\vec\nabla_\Phi S-q\vec A$
Mechanical momentum
$\mathcal E=E-q\varphi$
Mechanical energy
$\vec v=c^{2}\vec\pi/\mathcal E$
Kinematic velocity
$\vec V=\vec\pi/m$
Current velocity
$\rho_{\mathrm{chg}}=\rho_{\mathrm{num}}\mathcal E/(mc^{2})$
Charge density
$\epsilon=1/(\pi_{\Phi}L)$, $\delta=1/(k_{m}L)$
The semiclassical and one-particle parameters
$Q_{\mathrm{qnt}}=-\hbar^{2}\Box_\Phi \varrho/(2m\varrho)$
The quantum potential; the free one
$\Phi_{B}$
Magnetic flux through a surface the loop bounds
Further Reading