Library / Study
4 · The Schrödinger–Poisson system

From a classical scalar field to Schrödinger–Poisson

Fundamentally, fuzzy dark matter is a relativistic scalar field obeying the Klein–Gordon equation. But cosmic dark matter is non-relativistic and slowly-varying, and in that limit the field's equation collapses to the Schrödinger equation. This article makes that reduction, which is the bridge from particle physics to the wave dynamics we simulate.

The starting point: a scalar field

An ultralight boson is described by a real scalar field $\phi(\mathbf x,t)$ of mass $m$ obeying the Klein–Gordon equation in an expanding universe,

$$\ddot\phi+3H\dot\phi-\frac{\nabla^2}{a^2}\phi+\frac{m^2c^2}{\hbar^2}\phi=0 .$$

The field oscillates at the Compton frequency $\omega=mc^2/\hbar$ (Topic 2.4). For dark matter we care about the slow, gravitationally-driven modulation of that oscillation, not the fast carrier itself.

Factoring out the fast oscillation

Write the field as a fast carrier times a slowly-varying complex envelope $\psi$ (figure):

$$\phi(\mathbf x,t)=\frac{\hbar}{\sqrt{2m}}\Big(\psi\,e^{-imc^2t/\hbar}+\psi^*\,e^{+imc^2t/\hbar}\Big).$$

The key non-relativistic assumption is that $\psi$ changes slowly compared with the carrier: $|\dot\psi|\ll(mc^2/\hbar)|\psi|$, so we may drop $\ddot\psi$ relative to $mc^2\dot\psi/\hbar$. This separates the ~$10^{11}$-per-Hubble-time oscillation from the ~cosmological modulation.

The non-relativistic reduction: the relativistic field (red) oscillates at the fast Compton frequency $mc^2/\hbar$; factoring it out leaves the slowly-varying envelope $\psi$ (blue) that obeys the Schrödinger equation.

The result: Schrödinger's equation

Substituting and keeping leading order in $1/c^2$ collapses Klein–Gordon to the (comoving) Schrödinger equation,

$$i\hbar\,\dot\psi=-\frac{\hbar^2}{2ma^2}\nabla^2\psi+m\,\Phi\,\psi,$$

with $\Phi$ the Newtonian gravitational potential and $\rho=m|\psi|^2$ the dark-matter density. The relativistic field theory has become non-relativistic wave mechanics — the equation our codes actually integrate.

Worked example — is the non-relativistic limit safe?

The approximation needs $v\ll c$. Fuzzy dark matter in galaxies moves at $v\sim100\ \mathrm{km\,s^{-1}}$, so $v/c\sim3\times10^{-4}$ and $(v/c)^2\sim10^{-7}$: relativistic corrections are one part in ten million. The Schrödinger reduction is superbly accurate for structure formation — relativistic treatment is needed only near black holes or in the very early universe.

What survives, what is dropped

The reduction keeps gravity and the wave's kinetic (quantum-pressure) term — everything relevant to halos — and discards the fast Compton oscillation and $O(v^2/c^2)$ corrections. Adding self-gravity through Poisson's equation (next section) closes the system.

FDM in galaxies sits at $v/c\sim10^{-3}$ — deep in the non-relativistic regime where the Schrödinger reduction is exact.
In our research

This reduction is why the campaign simulates the Schrödinger equation rather than full field theory: at galactic $v/c\sim10^{-4}$ it is exact to one part in $10^{7}$. All three of JAXiON, GAMER, and the analytic tools inherit their governing equation from this step.

Key references
  • Hui, Ostriker, Tremaine & Witten (2017), Phys. Rev. D 95, 043541 (arXiv:1610.08297).
  • Marsh (2016), Axion Cosmology, Phys. Rep. 643, 1 (arXiv:1510.07633).
  • Ferreira (2021), Ultra-light dark matter, A&A Rev. 29, 7 (arXiv:2005.03254).