Galactic Astronomy
Two-Body Relaxation Time: How Long Star Encounters Take to Erase a Cluster's Memory
Two stars gliding through a globular cluster almost never collide — the odds of a physical hit over the entire age of the universe are essentially nil. Yet after roughly a billion years of countless gravitational near-misses, a typical globular cluster completely forgets how its stars were originally moving. That timescale — the point at which the accumulated tiny velocity nudges from distant encounters add up to the star's own orbital speed — is the two-body relaxation time, one of the master clocks of stellar dynamics.
Formally, the two-body relaxation time t_relax is the interval over which the cumulative effect of weak, long-range gravitational deflections between individual stars randomizes a star's velocity by an amount comparable to its original velocity. It marks the boundary between a smooth, collisionless system (where each star feels only the mean gravitational field) and a collisional one (where star-star encounters drive evolution: energy equipartition, mass segregation, evaporation, and core collapse).
- TypeDynamical timescale (stellar dynamics)
- RegimeMarks collisional/collisionless boundary
- Foundational theoryChandrasekhar (1942), Spitzer (1987)
- Key relationt_relax ≈ (N / 8 ln N) × t_cross
- Typical scale~10^8–10^9 yr (globular clusters)
- Observed inGlobular & open clusters, galactic nuclei
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
What Two-Body Relaxation Actually Is
Imagine a single star threading its way through a swarm of a million others. It is pulled continuously by the smooth, averaged gravitational field of the whole cluster — this sets its overall orbit. But superimposed on that smooth pull are the individual tugs of every other star as it passes nearby. Each single tug is tiny, but they never quite cancel: after many passages the star's velocity vector has been randomly walked away from where it started.
The two-body relaxation time is the time for that random walk in velocity to accumulate a change comparable to the star's own speed — i.e., when (Δv)² summed over all encounters ≈ v². At that point the star has 'forgotten' its initial velocity; the system has partially thermalized.
- Collisionless (t_relax ≫ age): individual encounters are negligible — galaxies.
- Collisional (t_relax ≲ age): encounters drive evolution — star clusters.
Crucially, 'collision' here almost never means physical contact. It means a gravitational encounter — a deflection at a distance.
The Mechanism: Why Distant Encounters Dominate
Consider a field star of mass m passing a target star with impact parameter b and relative speed v. The perpendicular velocity kick is roughly δv ≈ 2Gm / (b v). A single close pass gives a big kick but is rare; distant passes give tiny kicks but are enormously more common because the number of encounters at impact parameter b scales as b db.
Summing (δv)² over all impact parameters gives an integral ∝ ∫ db/b, which evaluates to ln(b_max / b_min) = ln Λ — the famous Coulomb logarithm, borrowed by analogy from plasma physics. The logarithm is the fingerprint that many weak, distant encounters, not rare close ones, do most of the relaxing. Chandrasekhar's 1942 treatment cast this as a diffusion process in velocity space, the same framework that yields dynamical friction.
The upper cutoff b_max is the system size (or interparticle spacing debate aside), the lower cutoff b_min is where deflections become strong-angle (~Gm/v²). Because Λ ≈ γN, one typically writes ln Λ = ln(γN) with γ ≈ 0.4 for equal-mass stars (Spitzer 1987).
Key Quantities and a Worked Example
The compact scaling relation connects relaxation to the crossing time t_cross = R/σ (the time to traverse the system):
t_relax ≈ [ N / (8 ln N) ] × t_cross
A more precise, dimensional form is the Chandrasekhar/Spitzer expression:
t_relax ≈ 0.34 σ³ / ( G² m² n ln Λ )
where σ is the 1-D velocity dispersion, n the number density, m the stellar mass. Spitzer's widely-used half-mass relaxation time is:
t_rh ≈ 0.138 N^(1/2) r_h^(3/2) / ( m^(1/2) G^(1/2) ln(γN) )
- Globular cluster: N ≈ 10⁶, ln N ≈ 14, t_cross ≈ 10⁶ yr → t_relax ≈ (10⁶ / 112) × 10⁶ ≈ 10¹⁰ yr at half-mass, but only ~10⁸ yr in the dense core.
- Galaxy: N ≈ 10¹¹, t_cross ≈ 10⁸ yr → t_relax ≈ 10¹⁷ yr — roughly ten million times the age of the universe.
The N/ln N factor is why bigger systems are more collisionless: relaxation slows almost linearly with particle number.
How It's Observed and Where It Matters
You cannot 'see' a relaxation time directly, but its consequences are stamped all over dense stellar systems, and modern instruments measure them precisely:
- Mass segregation: heavier stars sink toward the center as the system moves toward energy equipartition. HST proper-motion studies of 47 Tucanae (t_rh ≈ 3 Gyr, but core t_relax ≈ 9×10⁷ yr) show blue stragglers and heavy remnants concentrated centrally.
- Core collapse: ~20% of Milky Way globulars have collapsed, cusp-like cores — a runaway relaxation-driven instability.
- Evaporation: relaxation kicks a few percent of stars above escape velocity per relaxation time; clusters lose ~1% of stars per t_relax, driving eventual dissolution and producing tidal tails (mapped by Gaia).
Relaxation times are inferred by combining star counts (density profiles from HST/Gaia photometry) with velocity dispersions from spectroscopy and proper motions. The comparison of t_relax to cluster age tells dynamicists whether a system is dynamically 'old' or 'young'.
Relaxation vs. Its Close Cousins
Two-body relaxation is the baseline, but several related processes can be faster or operate differently — and confusing them is a common error:
- Violent relaxation (Lynden-Bell 1967): relaxation via a rapidly changing collective potential during collapse/mergers, completing in a few crossing times — it is collisionless and independent of N. This is how galaxies reach equilibrium despite huge two-body relaxation times.
- Dynamical friction: the systematic drag (not random diffusion) a massive object feels moving through a field of lighter stars — the coherent, directed sibling of two-body relaxation, sharing the same ln Λ.
- Resonant relaxation: near a central black hole, coherent torques between orbits randomize angular momentum far faster than scalar two-body relaxation.
- Collisional (physical) relaxation: actual stellar collisions — negligible except in the densest cores.
The key distinction: two-body relaxation is a slow, incoherent, N-dependent diffusion; violent relaxation is a fast, coherent, N-independent rearrangement.
Significance, Famous Cases, and Open Questions
Two-body relaxation is arguably the single most important internal clock for dense stellar systems. It underwrites the entire theory of globular cluster evolution: equipartition, mass segregation, binary hardening, core collapse and post-collapse 'gravothermal oscillations', and eventual evaporation — the framework laid out by Spitzer and modeled numerically since the 1970s (Hénon, Aarseth, then Fokker-Planck and Monte Carlo codes).
- Core collapse in M15 and NGC 6397: textbook cases where the core relaxation time is short enough that the gravothermal catastrophe has run.
- Intermediate-mass black holes: whether short central relaxation times build IMBHs via runaway mergers is actively debated.
- Galactic Center: the relaxation time near Sgr A* (~10⁹–10^10 yr) governs the S-star cusp and tidal-disruption rates.
Open issues remain: the correct value of ln Λ (especially b_max for real, non-uniform systems), whether idealized 'equipartition' is ever fully reached (it isn't — the Spitzer instability), and how a spectrum of stellar masses, binaries, and stellar-mass black holes reshape the simple single-mass picture.
| System | N (stars) | Crossing time t_cross | Relaxation time t_relax | Regime |
|---|---|---|---|---|
| Open cluster (e.g. Pleiades) | ~10^3 | ~10^6 yr | ~10^7–10^8 yr | Collisional (dissolves) |
| Globular cluster (e.g. 47 Tuc) | ~10^6 | ~10^6 yr | ~10^8–10^9 yr | Collisional (relaxed) |
| Globular cluster core | — | ~10^5 yr | ~10^7–10^8 yr | Strongly collisional |
| Galactic nucleus / SMBH cusp | ~10^7 | ~10^4 yr | ~10^9–10^10 yr | Marginally collisional |
| Elliptical galaxy | ~10^11 | ~10^8 yr | ~10^17 yr | Collisionless |
| Galaxy cluster | ~10^3 galaxies | ~10^9 yr | ~10^11 yr | Effectively collisionless |
Frequently asked questions
What is the two-body relaxation time in simple terms?
It is the time for the cumulative gravitational tugs from many distant star-star encounters to change a star's velocity by an amount comparable to its own speed, effectively randomizing its motion. After one relaxation time, the system has 'forgotten' its initial velocity distribution and begins to thermalize. It separates systems evolving through encounters (star clusters) from those that don't (galaxies).
Why is the relaxation time so much longer than the crossing time?
Because each individual encounter produces only a tiny velocity kick. It takes a huge number of them for the random walk to accumulate a change equal to the star's speed. Mathematically t_relax ≈ (N / 8 ln N) × t_cross, so with a million stars the relaxation time is roughly ten thousand crossing times. The larger N is, the more the tiny kicks average out.
What is the Coulomb logarithm and why does it appear?
The Coulomb logarithm, ln Λ = ln(b_max/b_min) ≈ ln(γN), arises when you integrate the squared velocity kicks over all impact parameters — the integral ∫ db/b gives a logarithm. It quantifies that many weak, distant encounters dominate relaxation over rare close ones. For star clusters ln Λ ≈ 10, with γ ≈ 0.4 for equal-mass stars (Spitzer 1987).
Are galaxies collisionless because of relaxation time?
Yes. For a galaxy with ~10^11 stars, the two-body relaxation time is around 10^17 years — roughly ten million times the age of the universe. Individual star encounters are therefore utterly negligible, so galaxies are treated as collisionless systems governed only by their smooth mean gravitational field. They reached equilibrium instead through violent relaxation.
How does relaxation cause mass segregation and core collapse?
Relaxation drives the system toward energy equipartition, so heavier stars lose energy to lighter ones and sink toward the center (mass segregation). As the core contracts and heats, its negative heat capacity makes it contract further — a runaway 'gravothermal catastrophe' that ends in core collapse. About 20% of Milky Way globular clusters show collapsed cores from this process.
How is two-body relaxation different from violent relaxation?
Two-body relaxation is a slow, incoherent diffusion driven by individual star encounters, and it depends on N (number of stars). Violent relaxation, described by Lynden-Bell in 1967, is a fast, collisionless rearrangement driven by a rapidly changing collective potential during collapse or merger, completing in a few crossing times and independent of N. Galaxies relax violently, not through two-body encounters.