8. Phase Spiral 9 pts

Astrophysics · Galactic dynamics, Gravitational potential, Phase mixing

Infer the Milky Way's vertical gravitational potential and dark matter density from the phase-space winding spiral of nearby stars (after Antoja et al. 2018, Guo et al. 2024).

Solution by Taavet Kalda.

Part i (1 point)

Gravitational acceleration obeys Gauss’ law, i.e., the number of field lines passing through a closed surface is proportional to the enclosed mass. We can see from the example of a point mass MM that gdA=4πGM\int g\,dA = 4\pi GM. Applied for the case of an infinite plane with constant density ρ0\rho_0 with a cuboid of area AA and half-thickness zz, we get 2azA=4πG2Azρ0-2a_z A = 4\pi G\cdot 2Az\rho_0, so

az=4πGρ0z.a_z = -4\pi G\rho_0 z.

Grading (preliminary)

  • Idea of using Gauss’ law: 0.3 pts
  • Formula relating the mass inside with gravitational flux: 0.3 pts
  • Application of Gauss’ law on a cuboid: 0.2 pts
  • Final result: 0.2 pts

Alternative solution:

  • Finding the acceleration of a thin disk by integrating over its surface: 0.5 pts
    • Writing down the integral: 0.3 pts
    • Correct evaluation, including finding that the acceleration is independent of the displacement from the surface: 0.2 pts
  • Inferring that only the surface density inside a<z<a-a < z < a contributes to the final acceleration: 0.3 pts
  • Final result: 0.2 pts

Part ii (0.5 points)

The acceleration is proportional to displacement and therefore corresponds to a harmonic oscillator. The period of oscillation is thus

T=2π4πGρ0=πGρ0.T = \frac{2\pi}{\sqrt{4\pi G\rho_0}} = \sqrt{\frac{\pi}{G\rho_0}}.

Grading (preliminary)

  • Noticing that the movement is that of a harmonic oscillator: 0.3 pts
  • Expression for the oscillation period: 0.2 pts

Part iii (2.5 points)

If we follow the trajectory of a single star that lies on the spiral, we would find it oscillating around the mid-plane with some period T(z)T(z) that decreases with increasing zz. Over the course of an orbit, the energy per unit mass is conserved and is given by E=vz2/2+Φ(z)E = v_z^2/2 + \Phi(z). We know that near the mid-plane, the gravitational potential is minimal and equal to zero, so vzv_z is maximal and the kinetic energy equals the total energy. Hence, we know the total energy of stars at the seven intersection points of the spiral with z=0z = 0. Similarly, when vz=0v_z = 0, the kinetic energy is zero, so the total energy equals Φ(z)\Phi(z) at those points.

As one traces the various intersection points with vz=0v_z = 0 and z=0z = 0 along the spiral, the maximal extent of the orbit keeps increasing. To the first order, assuming the maximal extent increases linearly with each crossing, we can find the potential energy of the crossings of vz=0v_z = 0 as the average between the kinetic energies of the previous and subsequent crossing over z=0z = 0. This allows us to determine the potential energy at all the crossings with vz=0v_z = 0, as tabulated below.

iiziz_i (kpc)Φ(zi)\Phi(z_i) (km²/s²)
10.27180
20.39330
30.54530
40.72800
50.971200
61.341800

Figure: Reconstructed gravitational potential \Phi(z) (in km²/s²) versus vertical distance z (in kpc), plotted from the six tabulated v_z = 0 crossings. \Phi rises from \sim 180\,\text{km}^2/\text{s}^2 at z \approx 0.27\,\text{kpc} to \sim 1800\,\text{km}^2/\text{s}^2 at z \approx 1.34\,\text{kpc}, showing roughly quadratic growth near the mid-plane and a more linear regime at larger z.

Grading (preliminary)

  • Making use of the total energy at z=0z = 0 intersection points being known (either explicitly or implicitly): 0.7 pts
  • Interpolating the values at vz=0v_z = 0 from neighbouring z=0z = 0 crossovers (should be explicitly mentioned): 0.8 pts
  • Tabulating the potential:
    • Using six points: 0.7 pts
    • Using four to five points: 0.4 pts
    • Using one to three points: 0.1 pts
  • Φ(z)\Phi(z) vs zz correctly plotted: 0.3 pts

Part iv (1 point)

For a harmonic oscillator, the potential energy would grow as z2z^2. Based on the plot of potential energy we obtained, it seems to grow roughly quadratically in the beginning, and then transition into a more linear regime, implying that smaller values of zz have more uniform ρ\rho. Taking the first value of Φ(z1)=180km2s2\Phi(z_1) = 180\,\text{km}^2\text{s}^{-2} and using Φ(z)=azdz=2πGρ0z2\Phi(z) = \int a_z\,dz = 2\pi G\rho_0 z^2:

ρ0=Φ(z1)2πGz12=6.1×1021kgm3=0.090Mpc3.\rho_0 = \frac{\Phi(z_1)}{2\pi G z_1^2} = 6.1\times 10^{-21}\,\text{kg\,m}^{-3} = 0.090\,M_\odot\text{pc}^{-3}.

Grading (preliminary)

  • Connecting Φ(z1)\Phi(z_1) with ρ0\rho_0 by assuming constant mass density: 0.8 pts
  • If the first data point is not used: −0.2 pts
  • Final expression for ρ0\rho_0: 0.1 pts
  • Numerical value within 10%: 0.1 pts

Part v (2 points)

We can compute the enclosed surface density Σ(z)\Sigma(z) between 0<z0 < z by using the previous harmonic oscillator estimate. With the constant density approximation:

Σ(z)=ρ0z=Φ(z)2πGz.\Sigma(z) = \rho_0 z = \frac{\Phi(z)}{2\pi G z}.

(Here ρ0\rho_0 is a placeholder variable while using the constant profile approximation to simplify the calculus. The final result is expected to deviate from the true value by a numerical factor close to unity.)

Assuming dark matter density dominates far away, we can use the difference between the farthest two data points at z5z_5 and z6z_6 to estimate the dark matter density:

Σ(z6)Σ(z5)=ρDM(z6z5)=12πG(Φ(z6)z6Φ(z5)z5).\Sigma(z_6) - \Sigma(z_5) = \rho_\text{DM}(z_6 - z_5) = \frac{1}{2\pi G}\left(\frac{\Phi(z_6)}{z_6} - \frac{\Phi(z_5)}{z_5}\right).

Thus,

ρDM12πG(z6z5)(Φ(z6)z6Φ(z5)z5)=7.7×1022kgm3=0.011Mpc3.\rho_\text{DM} \approx \frac{1}{2\pi G(z_6 - z_5)}\left(\frac{\Phi(z_6)}{z_6} - \frac{\Phi(z_5)}{z_5}\right) = 7.7\times 10^{-22}\,\text{kg\,m}^{-3} = 0.011\,M_\odot\text{pc}^{-3}.

Dark matter therefore makes up around ρDM/ρ0=13%\rho_\text{DM}/\rho_0 = 13\% of the total local matter budget.

Grading (preliminary)

  • Obtaining an expression for the total mass contained within zz: 0.7 pts
  • Taking the difference between the total mass within z6z_6 and z5z_5 for calculating the dark matter content: 0.9 pts
  • Final expression for ρDM\rho_\text{DM} based on z6z_6 and z5z_5: 0.3 pts
  • Numerical value within 10%: 0.1 pts

Alternative scheme:

  • Using the previous expression for ρ\rho based on Φ(z)\Phi(z) to express (z6z5)ρDM=z6ρ(z6)z5ρ(z5)(z_6 - z_5)\rho_\text{DM} = z_6\rho(z_6) - z_5\rho(z_5): 1.6 pts
  • Final expression for ρDM\rho_\text{DM} based on z6z_6 and z5z_5: 0.3 pts
  • Numerical value within 10%: 0.1 pts

Part vi (2 points)

We can estimate the time of the perturbation by using the winding rate between two points on the spiral and how many full turns around the origin they have made relative to each other. Using the harmonic estimate, ρ0=Φ(z)/(2πGz2)\rho_0 = \Phi(z)/(2\pi Gz^2) and the angular frequency is

ω(z)=4πGρ0=2Φ(z)/z2.\omega(z) = \sqrt{4\pi G\rho_0} = \sqrt{2\Phi(z)/z^2}.

Picking points z1z_1 and z6z_6 and seeing that they have 2.5 full turns between them, we can express how long ago the perturbation happened:

T0=2.52πω(z1)ω(z6)=5π(2Φ(z1)z122Φ(z6)z62)1T_0 = 2.5\cdot\frac{2\pi}{\omega(z_1) - \omega(z_6)} = 5\pi\left(\sqrt{\frac{2\Phi(z_1)}{z_1^2}} - \sqrt{\frac{2\Phi(z_6)}{z_6^2}}\right)^{-1} =1.9×1016s=620Myr.= 1.9\times 10^{16}\,\text{s} = 620\,\text{Myr}.

The timescale is relatively long, but compared to the lifespan of the Milky Way, which is around 13.6 billion years, it’s relatively recent.

Grading (preliminary)

  • Idea of using differences in the winding rate between two points on the spiral: 1.0 pts
  • Expression for angular frequency ω\omega in terms of Φ(z)\Phi(z) by assuming a harmonic oscillator: 0.5 pts
  • Picking two points and connecting the age of the spiral, ω\omega and the winding amount: 0.3 pts
  • Numerical value within 10% of the solution value: 0.2 pts