The Fourier Transform and the Coupled Schrödinger Equation in Biquaternionic Polar Variables
Introduction
The transform's kernel is a central unit, and the gauge transformation is a central unit too. So the two commute, and the transform of a gauge-transformed field is the gauge-transformed transform. That one sentence is the whole of the article's first section, and it fixes the chart's reading of the coupling: the gauge group translates the phase and the transform commutes with the translation.
The second finding is that the covariant bracket adds a real number to the phase's rate and nothing else. The polar coupled parent computes the covariant logarithmic derivative and finds the free one plus a central imaginary number in every slot. In frequency space that number becomes $ϙ\varphi$ added to $\hbar\omega$. The connection touches one of the four brackets, and the other three are the free ones.
The third finding is that the symbol is a model of the bracket and not the bracket itself. The bracket is a number at each time; the symbol is that number read at a fixed frequency. Where the bracket is constant the model is exact, and the constant background is exactly that case. Where the bracket varies the symbol does not exist, and the transform spreads: the effective number of modes of a chirped unit is $5.80$ at a chirp rate of $1$ and $26.53$ at a rate of $4$, against $1.00$ for a pure tone.
The fourth finding is that the two invariants are not functionals of the symbol. $\vec\pi=\hbar\nabla\theta_{\mathrm{pha}}-q\vec A$ and $\mathcal E=-\hbar\partial_{t'}\theta_{\mathrm{pha}}+ϙ\varphi$ are properties of the field at each point. The symbol is one function of $(\omega,\vec k)$, and it is the same function for every superposition. Two superpositions of the same two modes have the same symbol and different mean $\vec\pi$: $1.700$ against $2.500$ in the example below.
The fifth finding is that the constant background is the exact case. The phase is linear in $\varphi$ and quadratic in $\vec A$, with no higher term, because the background is constant. So the "semiclassical" phase is exact here, and the article says plainly that there is no small parameter to expand in.
The companion articles. The Fourier Transform and the Coupled Schrödinger Equation in Biquaternionic Form gives the symbol; the free polar companion gives the chart. This article is the two together, and it cites both.
Conventions
The chart, its word and its norms are those of The Schrödinger Equation in Biquaternionic Polar Variables. The covariant logarithmic derivative, the two invariants and the gauge direction are those of The Coupled Schrödinger Equation in Biquaternionic Polar Variables. The symbol's translation identity and its sign are those of The Fourier Transform and the Coupled Schrödinger Equation in Biquaternionic Form.
The covariant slots. $D_t=\partial_{t'}-iϙ\varphi/\hbar$ and $D_\eta=\partial_\eta+iϙ A_\eta/\hbar$, with $\varphi$ and $\vec A$ real. The two signs differ, and the difference is the metric $i$ of the material-time derivative.
The two invariants. With $\theta=\theta_{\mathrm{pha}}$,
$$ \vec\pi=\hbar\nabla\theta_{\mathrm{pha}}-q\vec A, \qquad \mathcal E=-\hbar\,\partial_{t'}\theta_{\mathrm{pha}}+ϙ\varphi . $$
The substitution. $\tilde\Xi=\tilde\Xi_0\exp(-i\omega t'+i\vec k_\Phi\cdot\vec r)$, so the covariant slots become the numbers $\hbar\omega+ϙ\varphi$ and $\hbar k_\eta+ϙ A_\eta$.
The Transform's Kernel in the Coupled Chart
Both operations are central multiplications. The transform multiplies by $e^{-i\theta_{\mathrm{pwp}}}$, and a gauge transformation multiplies by $\lambda=e^{iϙ\Gamma/\hbar}$. Both are central units, so they commute:
$$ \int\mathrm{d}t'\,e^{i\omega t'}\big(\lambda\tilde\Xi\big)=\lambda\int\mathrm{d}t'\,e^{i\omega t'}\tilde\Xi . $$
The transform is therefore gauge covariant. A gauge-transformed field has a gauge-transformed transform, with the same $\Gamma$ and the same $\lambda$. No link variable is needed and no Jacobian appears, and the reason is the centrality and not the constancy of $\Gamma$.
The chart's reading of the same line. The kernel's word has $\theta_{\mathrm{pha}}=\theta_{\mathrm{pwp}}$ and nothing else; the gauge factor has $\theta_{\mathrm{pha}}=ϙ\Gamma/\hbar$ and nothing else. Two elements that each carry one coordinate commute, because a central unit is a function of the centre alone.
What this does not say. It does not say that the transform is gauge invariant. The transform of a gauge-transformed field is a gauge-transformed transform, and the phase of the coefficients moves with $\Gamma$. Covariance is the statement, and invariance is a different statement; the invariants of §The Invariants Are Not Functionals of the Symbol are gauge invariant, and the coefficients are not.
The kernel's own word is unchanged by the coupling. The transform's kernel carries no charge, so the connection does not enter it. The coupling is entirely in the field and in the equation, and the transform is the same operation as in the free article.
The Slots Are the Four Brackets
The covariant logarithmic derivative. The polar coupled parent writes the covariant derivative as the free one plus a central imaginary number in every slot. In the temporal slot the addition is $-iϙ\varphi/\hbar$, so the phase's rate acquires $+ϙ\varphi/\hbar$ when the equation is read in the parent's form.
In frequency space the addition is a real number. The covariant temporal bracket becomes
$$ \hbar\,\big(\partial_{t'}\theta_{\mathrm{pha}}\big)\longmapsto\hbar\omega, \qquad \hbar\,\big(\partial_{t'}\theta_{\mathrm{pha}}\big)+ϙ\varphi\longmapsto\hbar\omega+ϙ\varphi, $$
a real number added to a real number. The connection occupies the phase type and nothing else, and the phase type is one of the four of the free polar companion.
The other three brackets are the free ones. The connection is central in every slot, so it cannot reach the scale's rate, the boost's connection or the rotor's connection. The scale is untouched, the rapidity is untouched, and the rotor is untouched, which is the polar coupled parent's type statement read in frequency space.
The count is one addition and three zeros. Of the four brackets, one receives a real number and three receive nothing. That is the sharpest way to state what a central connection can do, and it is the frequency-space form of the statement that the gauge group is the centre's.
The spatial slots add nothing to the symbol's type. The spatial covariant derivative shifts the wave vector and does not change the symbol's type either; the symbol stays a scalar or an element. So the connection's whole effect on the symbol is a shift of its arguments, in both slots, with the signs the parent fixes.
A check of the sign in the chart. The invariant $\mathcal E=-\hbar\partial_{t'}\theta_{\mathrm{pha}}+ϙ\varphi$ equals $\hbar\omega+ϙ\varphi$ for a mode, so the symbol's frequency argument is $\mathcal E$ itself. The symbol is a function of the invariant and not of the bare frequency, and the chart's phase rate is the invariant's first term.
The Symbol Is a Model of the Bracket
The bracket is a number at each time. $\partial_{t'}\theta_{\mathrm{pha}}$ is a function of $t'$, and its product with $\hbar$ is the local frequency of the field. The symbol is that number read at one frequency, and the reading is faithful only when the number does not vary.
The exact case is the constant bracket. For a plane wave the bracket is constant, the symbol exists, and the equation reduces to the algebraic relation $\hbar\omega+ϙ\varphi=\lvert\hbar\vec k-q\vec A\rvert^{2}/(2m)$. Nothing is approximated, and the reduction is exact because the coefficient is constant.
The varying case has no symbol. an invertible element whose phase is $\theta=\omega_0t'+\tfrac12\alpha t'^{2}$ has a bracket $\omega_0+\alpha t'$ that varies, and its transform is not supported at one frequency. On a grid of sixty-four points the effective number of modes $1/\sum_jp_j^{2}$ was $1.00$ for an integer tone, $5.64$ at $\alpha=1$ and $26.20$ at $\alpha=4$. The spread grows with the chirp and the symbol does not exist for any of them.
So the model's domain is the constant bracket. That is the chart's way of stating the licence of the companion article: constant coefficients. The licence is on the bracket and not on the coefficients of the equation, and the two coincide for a plane wave.
What the model keeps. It keeps the bracket's value and loses its trajectory. A symbol can say what the local frequency is and cannot say how it varies, and every statement of this article is of the first kind.
What this buys. It makes the limit statement exact: the constant background is not the leading term of an expansion, it is the case in which the expansion terminates. §The Constant Background Is the Exact Case states the termination term by term.
The Eigenvalue Pair With the Two Shifts
The eigenvalues. In the constant setting the shifted Hamiltonian is $\tilde H^{\mathrm{gco}}_\Psi=(h_t-ϙ\varphi)e_t+i(\vec h_\Psi-q\vec A)$, so the two eigenvalues are
$$ h_t-ϙ\varphi\pm\lvert\vec h_\Psi-q\vec A\rvert , $$
each twice. The scalar potential translates the pair and the vector potential changes its spacing, exactly as the element-form companion computes.
The two rates in the chart's letters. The phase runs at $-\big(h_t-ϙ\varphi\big)/\hbar$ and the rotor's angle at $2\lvert\vec h_\Psi-q\vec A\rvert/\hbar$. The two shifted numbers are the two rates, and the chart's names for them are the mean and the gap.
The numbers. With $\lvert\vec h_\Psi\rvert=0.616441$ the free gap is $1.232882$; with $ϙ\vec A=(0.35,0.25,0)$ it is $1.503329$; with $ϙ\varphi=0.4$ alone it stays $1.232882$ and the phase's rate moves by $-0.400000$. The gap is a vector-potential effect and the translation is a scalar-potential effect.
The invariant reading. The two eigenvalues are $\mathcal E_0\pm\lvert\vec\pi_0\rvert$ for a mode, with $\mathcal E_0=\hbar\omega+ϙ\varphi$ and $\vec\pi_0=\hbar\vec k-q\vec A$. So the pair is a function of the two invariants, and the chart's two rates are functions of the two invariants. Gauge transformations move neither.
What the pair does not carry. It does not carry the rotor's axis, which the vector potential shifts. The gap is read and the axis is frame data, exactly as in the free polar companion.
The Invariants Are Not Functionals of the Symbol
The symbol is one function. $\sigma_{\mathrm{gco}}$ is a fixed function of $(\omega,\vec k)$ once the connection and the mass are fixed, the same function for every field. The invariants are not: $\vec\pi$ and $\mathcal E$ are properties of a given field, computed from its phase's derivatives.
The example. Take two modes with $\hbar k=1$ and $\hbar k=3$, a connection with $ϙ A=0.5$, and two superpositions. Weights $(0.9,0.1)$ give the mean mechanical momentum $0.9\times1.5+0.1\times3.5=1.700$; weights $(0.5,0.5)$ give $2.500$. The symbol is the same function in both cases, because it is the same equation with the same connection.
So the symbol cannot carry the invariants as functionals. It carries the values of the invariants at one mode, which is the mode's label, and it carries them for every mode at once as a function. What it does not carry is the field's actual momentum distribution, and a superposition has one distribution and one symbol.
Where the invariants live. They live in the field and in its word, and their integrals are the Noether charges. The transform diagonalises the charges and does not determine the invariants, and that asymmetry is the polar chart's rendering of the difference between a label and a value.
The gauge covariance of the invariants. Under a gauge transformation the phase's gradient shifts by $-\vec\nabla_\Phi\Gamma$ and $\vec A$ by $+\vec\nabla_\Phi\Gamma$... in the parent's sign convention $\vec A\mapsto\vec A-\vec\nabla_\Phi\Gamma$ and $\theta\mapsto\theta+ϙ\Gamma/\hbar$, so $\hbar\nabla\theta-q\vec A$ is invariant. Both invariants are gauge invariant and both are field-dependent, and the pair of properties is what a symbol cannot have.
A caution about $\vec k$. The symbol's argument $\vec k$ is the transform's label and not the field's mechanical momentum. They agree on a single plane wave and disagree on every superposition. Reading $\hbar\vec k-q\vec A$ as the field's momentum is the error this section exists to prevent.
The Constant Background Is the Exact Case
The expansion of the shift. With a constant background the symbol's dependence on the connection is
$$ \sigma_{\mathrm{gco}}=\sigma_{\mathrm{free}}+\rho_{\mathrm{chg}}\varphi-\frac{1}{m}\rho_{\mathrm{chg}}\,\vec k\cdot\vec A-\frac{\rho_{\mathrm{chg}}^{2}}{2m}\lvert\vec A\rvert^{2}, $$
so it is linear in $\varphi$, linear and quadratic in $\vec A$, and has no higher term. The series terminates because the background is constant, and not because a small parameter is small.
The "semiclassical" phase is the phase. The phase accumulated over an interval is $-\mathcal E t'/\hbar$ with $\mathcal E=\hbar\omega+\rho_{\mathrm{chg}}\varphi$, exactly. There is no next order, and calling the constant-background phase semiclassical is a habit that this article declines.
No amplitude–potential cross term. The symbol is the coefficient of a quadratic form in the mode, and the form is homogeneous of degree two in the mode's amplitude: over one hundred draws, multiplying the mode by a complex $\lambda$ multiplied the form by $\lvert\lambda\rvert^{2}$ to $4.5\times10^{-14}$. So the symbol is independent of the amplitude, and no term can couple the amplitude to the potential.
What a cross term would require. A term in $\varrho\vec A$ would need the symbol to depend on $\varrho$, which the homogeneity forbids. The absence is structural and not a cancellation, and it is the frequency-space form of the polar parent's statement that the quantum potential is untouched by a constant potential.
The exactness has a price. It holds for a constant connection and a constant Hamiltonian, and the price is the licence of §The Symbol Is a Model of the Bracket. Where the background varies, the phase is an integral of $\mathcal E$ along the path and the symbol is the integrand at one point. That statement is the standard WKB phase, and it is a statement about a limit and not about this article's case.
Why the case is worth stating. It is the case in which the chart's three statements are exact: the phase is a character, the gap is a function of the invariants, and the connection touches one bracket. An exact case is the calibration of every approximate one, and the coupled companions use it as such.
What the Polar Reading Adds
The type statement, sharpened. The connection adds a real number to the phase's rate and to nothing else. In the chart the statement has a name: it occupies one of the four brackets.
The covariance statement. The transform and the gauge transformation commute, because both are central multiplications. Covariance is one line and it needs no link variable.
The distinction between labels and values. The symbol is a function of the two invariants; the field has a distribution of them. The transform diagonalises the charges and does not determine the distribution, and the example's $1.700$ against $2.500$ is the measure of the difference.
What it does not add. No new prediction, no new solution, no statement about a varying field strength. The chart is a lens and the lens is exact.
Limits
A constant bracket. The symbol is a model of a constant bracket, and a varying bracket removes it. The spread of §The Symbol Is a Model of the Bracket measures the removal and is not a correction to the model.
A central connection. The count of one bracket touched and three untouched rests on the connection's centrality. A connection that did not commute with the units would enter all four, and the count would fail; this connection is central, so it holds.
One time. The transform is in time and the wave vector is its label; the field's invariants are four and the symbol's arguments are two. The article does not claim that the symbol sees more than its two arguments.
Open Questions
The first question is the varying bracket. Where the bracket varies, the symbol does not exist and the transform spreads. Whether a local symbol built from $\mathcal E$ and $\vec\pi$ is faithful in a stated limit is the standard WKB question, and it is not decided here.
The second is the gauge transformation's large version. The commutativity of the transform and a central multiplication holds for every $\Gamma$, including one with a nonzero winding. Whether a large gauge transformation changes the spectrum of the transform is a question about the domains and not about the algebra, and it is open.
The third is the chart's flat dual under a coupling. The free chart's dual variable $n=\varrho^{2}$ is flat. Under a coupling the density is unchanged and the phase's rate is shifted, so the flatness survives; whether the pair survives when the connection varies with time is the coupled polar parent's open question.
The fourth is the sign of $ϙ\varphi$ in the phase's rate. The parent fixes it and this article uses it. Whether the corpus's conventions for the charge make the sign read as positive energy for a positive charge is a conventions question, and the conventions articles are where it belongs.
Summary
The transform and the gauge transformation are both central multiplications, so they commute: the transform of a gauge-transformed field is the gauge-transformed transform, exactly, with no link variable. The covariant bracket adds the real number $ϙ\varphi$ to the phase's rate and leaves the other three brackets of the chart untouched. That is the frequency-space form of the statement that a central connection can only reach the phase.
The symbol is a model of the bracket: faithful where the bracket is constant, absent where it varies. The effective number of modes of a chirped unit was $1.00$, $5.80$ and $26.53$ at chirp rates $0$, $1$ and $4$, so the spread is the measure of the failure. The eigenvalues are $h_t-ϙ\varphi\pm\lvert\vec h_\Psi-q\vec A\rvert$, and the gap moves from $1.232882$ to $1.503329$ with the quoted vector potential while a scalar potential of $0.4$ leaves it at $1.232882$ and moves the rate by $-0.400000$.
The invariants $\vec\pi$ and $\mathcal E$ are gauge invariant and field dependent, so they are not functionals of the symbol: two superpositions of the same two modes share the symbol and have mean mechanical momenta $1.700$ and $2.500$. And the constant background is the exact case: the phase is linear in $\varphi$ and quadratic in $\vec A$ with no higher term, and the symbol is independent of the amplitude to $4.5\times10^{-14}$, so no amplitude–potential cross term can exist.
Summary of Notation
| Symbol | Meaning |
|---|---|
| $\lambda=e^{iϙ\Gamma/\hbar}$ | The gauge factor; central, hence commuting with the kernel |
| $e^{-i\theta_{\mathrm{pwp}}}$ | The transform's kernel; central, one coordinate |
| $\hbar\omega+ϙ\varphi$ | The covariant frequency; the invariant $\mathcal E$ on a mode |
| $\hbar k_\eta+ϙ A_\eta$ | The mechanical momentum; the invariant $\vec\pi$ on a mode |
| $h_t-ϙ\varphi\pm\lvert\vec h_\Psi-q\vec A\rvert$ | The eigenvalue pair; the translation and the gap |
| $1.232882$, $1.503329$ | The gap free and with $ϙ\vec A=(0.35,0.25,0)$ |
| $1.00$, $5.64$, $26.20$ | Effective mode counts of a chirped unit at rates $0$, $1$, $4$ |
| $1.700$, $2.500$ | Mean mechanical momenta of two superpositions of the same two modes |
| $\sigma_{\mathrm{gco}}=\sigma_{\mathrm{free}}+ϙ\varphi-\frac{ϙ}{m}\vec k\cdot\vec A-\frac{ϙ^{2}}{2m}\lvert\vec A\rvert^{2}$ | The exact expansion in a constant background |
Further Reading
- The Coupled Schrödinger Equation in Biquaternionic Polar Variables, for the covariant logarithmic derivative, the two invariants and the gauge direction cited here.
- The Fourier Transform and the Coupled Schrödinger Equation in Biquaternionic Form, for the element-form symbol and its two settings.
- The Fourier Transform and the Schrödinger Equation in Biquaternionic Polar Variables, for the free chart's four brackets and the flat dual.
- The Schrödinger Equation in Biquaternionic Polar Variables, for the word, the two rates and the two norms.
- The Lagrangian and the Coupled Schrödinger Equation in Momentum Space, for the coupled action at fixed frequency.
- The Coupled Schrödinger Equation in Biquaternionic Form, for the equation and its sign bookkeeping.