research31 min studyLesson 824 in Mathematics

Wavefront sets and microlocal regularity

Locate nonnegligible h-dependent oscillations and distinguish semiclassical phase-space localization from a fixed distribution's conic wavefront set.

On this page

What you’ll understand

  • Define absence from the finite-phase-space semiclassical wavefront set.
  • Distinguish classical smoothness, semiclassical residual size, and finite-momentum escape.
  • Analyze localized oscillatory families and an exponentially small variant.

A wave can be smooth at every fixed h and still contain a strong semiclassical oscillation. Ordinary smoothness does not decide where its mass lives in rescaled momentum. The semiclassical wavefront set detects phase-space locations that cannot be removed up to errors smaller than every power of h. Its quantifiers concern an entire family, and its finite-momentum version differs from the classical conic wavefront set of one fixed distribution.

Negligibility must hold to every order

Work with an h-tempered family, for example one whose local distribution norms grow at most polynomially in one over h. A finite phase-space point is absent from WF_h(u_h) if some fixed compact smooth cutoff a, elliptic at that point, satisfies ||a^w u_h||_(L²)=O(h^N) for every N. The cutoff may be chosen equal to one nearby. The constants may depend on N, but the neighbourhood and cutoff are fixed as h tends to zero. Such a family is called microlocally residual there.

An O(h) localized norm does not imply absence: it is only one power of smallness. Conversely an exponentially small family can be residual even when its normalized version has substantial oscillations. The wavefront set records the chosen normalization of the family. Local cutoffs in both position and momentum avoid confusing spatial support with oscillatory direction; a family can remain localized in position while its momentum moves outside every fixed compact set.

(x0,ξ0)∉WF⁡h(uh)⟺∃a∈Cc∞, a(x0,ξ0)≠0: ∥awuh∥L2=O(hN) for every N(x_0,\xi_0)\notin\operatorname{WF}_h(u_h)\quad\Longleftrightarrow\quad\exists a\in C_c^\infty,\ a(x_0,\xi_0)\ne0:\ \|a^wu_h\|_{L^2}=O(h^N)\ \text{for every }N
This lesson concerns finite phase-space points. The family is h-tempered and a is fixed and elliptic near the point.↔ Scroll wide equations horizontally.

Classical and semiclassical wavefront sets answer different questions

The classical wavefront set tests rapid Fourier decay in conic frequency regions for one distribution and excludes the zero covector. It is conic because only the direction at arbitrarily large ordinary frequency matters. The finite semiclassical wavefront set instead resolves an actual rescaled momentum ξ, can include ξ=0, and need not be conic. Multiplying ξ by two generally changes the point being tested. A fixed compact smooth function has no classical wavefront set but can have semiclassical wavefront over its support at zero momentum.

For χ(x)exp(iS(x)/h) with a smooth compact envelope and real smooth phase, nonstationary phase identifies the finite semiclassical momentum graph ξ=S′(x), where the envelope is nonzero. A cutoff away from that graph gives repeated integration by parts and residual size. On the graph, the oscillation survives. Zeros of the amplitude at isolated points do not automatically remove points from the closed wavefront set, so statements are most transparent on open sets where the amplitude does not vanish.

uh(x)=χ(x)eiS(x)/h⇝ξ=S′(x),S(x)=ξ0x⟹WF⁡h(uh)=supp⁡χ×{ξ0}u_h(x)=\chi(x)e^{iS(x)/h}\quad\leadsto\quad\xi=S'(x),\qquad S(x)=\xi_0x\Longrightarrow\operatorname{WF}_h(u_h)=\operatorname{supp}\chi\times\{\xi_0\}
The phase is real and smooth and χ is a fixed compact smooth function; the support includes closure points. The result is not a conic set in ξ.↔ Scroll wide equations horizontally.

Finite charts do not detect escape to infinity

A normalized family with phase exp(ix/h²) has rescaled momentum one over h. Every fixed compact momentum cutoff can miss it as h tends to zero, even though its L² norm is unchanged. Its finite-phase-space wavefront set can therefore be empty while information remains at fibre infinity. A compactified calculus can record that escape, but the finite definition above does not. An empty finite wavefront set is not a global residual conclusion without additional localization controlling the missing infinity.

This distinction is also why a finite-band drawing cannot prove a wavefront assertion. Any fixed finite sum of smooth sine waves is classically smooth, and inspecting a single value or a small plot does not establish an estimate for every power of h. The definition requires an asymptotic family and a neighbourhood test. Numerically visible oscillations may suggest a phase-space location, but the nonstationary-phase estimate or a microlocal operator argument supplies the actual regularity statement.

Smooth functions with a nonempty semiclassical wavefront

Let χ be nonzero and compactly supported smooth, and set u_h=χ(x)exp(2ix/h). Locate its finite semiclassical wavefront and compare with v_h=exp(−1/h)u_h.

  1. For each fixed positive h, u_h is smooth and compactly supported, so its classical distributional wavefront set is empty. Its h-tempered derivative bounds grow only as powers of one over h.

  2. The semiclassical phase has derivative two. Nonstationary phase removes any fixed momentum neighbourhood away from ξ=2, while a cutoff near (x,2) with χ(x) nonzero preserves a nonnegligible component. Taking closure gives support χ times {2}.

  3. The factor exp(−1/h) is smaller than h^N for every N. Since the original L² norm is fixed, v_h is globally O(h^∞) in L² and has empty finite semiclassical wavefront. Renormalizing v_h to unit norm would remove that exponential factor.

The result

Classical smoothness and semiclassical residual size are different properties. Normalization can change whether a whole family is microlocally negligible.

Explore the model

Interactive lab

Interactive lab · spectral

A pointwise zero is not microlocal absence

Pause and Reset, then set N = 1 and decay p = 0. At time zero and x=π/2, the sum is one. Reset and set both N = 3 and p = 0: the value there is 1+0−1=0, although nonzero modes remain in nearby positions. Predict that this single cancellation cannot establish a neighbourhood estimate. Every supported finite sum is smooth; the instrument has no h→0 family and cannot determine a semiclassical wavefront set.

A pointwise zero is not microlocal absenceA finite sum of sin(nx − t)/nᵖ. Playback changes t; increasing the cutoff includes finer spatial scales. This is not a proof about a continuum limit or renormalization. Current readings: Sum at x = π/2: 0.867; Smallest wavelength: 2π / 5; Highest-mode coefficient: 0.2. All parameters are adjustable with the labeled controls below.01.573.144.716.28−2.56−1.2801.282.56x · radiansFinite sine sum
Finite sumHighest included mode
Sum at x = π/2
0.867
Smallest wavelength
2π / 5
Highest-mode coefficient
0.2
t = 0 s

A finite sum of sin(nx − t)/nᵖ. Playback changes t; increasing the cutoff includes finer spatial scales. This is not a proof about a continuum limit or renormalization. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.

Look forWavefront absence is a uniform microlocal asymptotic statement, not a zero at one plotted point.

Check your understanding

Which statement correctly describes the finite semiclassical wavefront set?

Try it yourself

For a fixed nonzero compact smooth χ, analyze w_h=χ(x)exp(ix/h²). What happens under fixed compact semiclassical momentum cutoffs, and why does this not imply its L² norm tends to zero?

Take a moment to make a prediction or work it out on paper. Open a hint when you need one.

Give me a hint

Write the ordinary frequency as 1/h² and multiply by h to obtain the rescaled momentum. The modulus of the phase is one.

Compare with the solution

The semiclassical momentum is 1/h, which leaves every fixed compact momentum set. Nonstationary phase makes a cutoff with fixed compact phase-space support residual, so the finite semiclassical wavefront set is empty. Yet ||w_h||₂=||χ||₂ for every h. The mass escapes toward fibre infinity rather than disappearing. The family remains h-tempered, since each fixed derivative grows by only a finite inverse power of h. A global conclusion needs a compactified wavefront description or an additional frequency-tightness assumption, neither of which follows from the finite-cutoff tests alone.

Take it with you

  • Semiclassical absence requires O(h^N) for every N under a fixed microlocal cutoff.
  • Finite semiclassical wavefront sets differ from classical conic distributional wavefront sets.
  • Escape to momentum infinity can make all finite cutoffs negligible without losing global norm.

Read further

Original sources and deeper explanations to follow at your own pace.

Optional progress, saved in this browser when available. Revisit any lesson at any time.

Lesson not marked complete.