The 1D Schrödinger equation with a spacetime white noise: the average wave function

Yu Gu

TL;DR

This paper explores how the 1D Schrödinger equation behaves when subjected to a highly oscillatory, random "spacetime white noise" potential. The author proves that as the noise becomes increasingly fine-grained, the average wave function converges to a deterministic Schrödinger equation featuring an "effective potential" that accounts for the cumulative scattering effects of the randomness.

What Problem Does This Paper Address?

In quantum mechanics, the Schrödinger equation typically describes a particle moving in a smooth, well-defined potential field. However, in complex environments—like disordered solids or turbulent plasmas—the potential can be highly irregular and random.

Mathematically, this is often modeled using "spacetime white noise," which is extremely "rough" and fluctuates at every point in time and space. Analyzing the Schrödinger equation in this regime is notoriously difficult because the dispersive nature of the wave function interacts with the noise in complex ways. Specifically, the randomness causes the wave function to scatter, leading to exponential attenuation of the original signal. The "gap" this paper fills is providing a rigorous mathematical proof for the limit of the average wave function (the first moment) under a specific diffusive scaling, a problem that had remained largely open.

Methods

The author employs a mix of stochastic analysis and partial differential equation (PDE) theory:

  • Mollification: Since raw spacetime white noise is too singular to handle directly, the author starts with a "mollified" (smoothed) version of the noise and then takes the limit as the smoothing disappears.
  • Probabilistic Representation: A key technique is the use of a Feynman-Kac-type formula. This allows the author to represent the solution of the Schrödinger equation using an auxiliary Brownian motion. This effectively transforms a difficult PDE problem into a problem of analyzing functionals of stochastic processes.
  • Fourier Domain Analysis: The proof relies heavily on looking at the wave function in the frequency domain ($\xi$).
  • Renormalization: Because the average wave function decays exponentially due to scattering, the author introduces a renormalization factor ($e^{z_1 t / \epsilon}$) to "boost" the signal so that a non-trivial limit can be observed.
  • Martingale Central Limit Theorem: To prove convergence as the scaling parameter $\epsilon \to 0$, the author uses tools from martingale theory to show that the random fluctuations of the Brownian functionals converge to a Gaussian limit.

Datasets

Not applicable. This is a theoretical paper in mathematical physics and does not use empirical datasets or benchmarks.

Key Insights and Ideas

  • The Ballistic Component: The most important insight is that the average wave function only captures the "ballistic" part of the wave—the part that manages to travel without being scattered away.
  • Energy Escape: The author notes that while the average wave function converges to a nice deterministic limit, this limit is "exponentially small" relative to the original system. Most of the wave's energy actually escapes to high frequencies through multiple scattering events, which would be captured by the second moment (the kinetic regime), not the average.
  • Effective Potential ($z_2$): The randomness doesn't just add "noise"; it fundamentally shifts the dynamics. The limit equation includes a complex constant $z_2$ which acts as an effective potential, representing how the medium "slows down" or dampens the average wave.
  • Renormalization as Compensation: The renormalization factor is physically interpreted as a compensation for the exponential attenuation caused by the medium.

Key Results

  • Theorem 1.1: The paper's central result proves that the expected value of the Fourier-transformed wave function $\mathbb{E}[\hat{\phi}_\epsilon]$ converges in distribution as $\epsilon \to 0$.
  • The Limit Equation: The average wave function $\bar{\phi}$ satisfies: $$i\partial_t\bar{\phi} + \frac{1}{2}\Delta\bar{\phi} - i z_2\bar{\phi} = 0$$ where $z_2$ is a constant derived from the covariance of the noise.

  • Explicit Constants: The paper provides explicit integral formulas for the renormalization constants $z_1$ and $z_2$ based on the choice of the mollifier (the smoothing function).

Limitations and Future Work

  • Physical Interpretation: The author acknowledges that the physical meaning of the renormalized average wave function is somewhat "unclear" because it focuses on a tiny fraction of the total energy that hasn't scattered.
  • 1D Constraint: While the author mentions that the convergence results hold in higher dimensions ($d \ge 1$), the specific scaling used to reach a spacetime white noise limit is unique to the 1D case.
  • Higher Moments: A natural next step is the study of the second moment (energy density). While this paper touches on kinetic equations (like the radiative transfer equation) in the remarks, a full treatment of the energy distribution in this specific spacetime noise regime remains a broader challenge.

Summary updated

Abstract

For the 1D Schrödinger equation with a mollified spacetime white noise, we show that the average wave function converges to the Schrödinger equation with an effective potential after an appropriate renormalization.
BibTeX
@inproceedings{gu2019d,
  author = {Yu Gu},
  title = {{The 1D Schrödinger equation with a spacetime white noise: the average wave function}},
  booktitle = {ESAIM},
  year = {2019},
  doi = {10.1051/ps/2019010},
}

Loading related content...

Figure