General relativity (GR) is Einstein’s 1915 theory of gravitation. It replaces Newton’s force of gravity with the idea that massive objects curve the geometry of spacetime, and freely falling bodies (including light) follow the straightest possible paths (geodesics) through that curved geometry. GR has passed every experimental test to exquisite precision over a century. This post builds from the equivalence principle up to black hole thermodynamics.
1. The Equivalence Principle and Geodesics
Einstein’s Equivalence Principle (EEP) asserts that the effects of gravity are locally indistinguishable from the effects of acceleration. More precisely: in a sufficiently small region of spacetime, the laws of special relativity hold in a freely falling (inertial) frame. This implies that gravity must bend light (if a horizontal light beam is bent in an accelerating rocket, it must also bend in a gravitational field) and cause gravitational time dilation.
Geodesic Equation & Christoffel Symbols
Spacetime interval (general metric):
ds² = g_μν dxμ dxν
Signature: (−,+,+,+) [timelike ds²<0, spacelike >0, null =0]
Proper time: dτ² = −ds²/c²
Geodesic equation (free fall in curved spacetime):
d²xµ/dτ² + Γµ_(νρ) (dxν/dτ)(dxρ/dτ) = 0
Christoffel symbols (Levi-Civita connection):
Γµ_(νρ) = (1/2) gµσ (∂_ν g_σρ + ∂_ρ g_σν − ∂_σ g_νρ)
Symmetric in lower indices: Γµ_(νρ) = Γµ_(ρν)
Not a tensor (vanishes in locally inertial frame)
Covariant derivative:
∇_ν Vµ = ∂_ν Vµ + Γµ_(νρ) Vρ (contravariant vector)
∇_ν ω_µ = ∂_ν ω_µ − Γρ_(νµ) ω_ρ (covariant vector, 1-form)
Parallel transport:
DVµ/dλ = (dVµ/dλ) + Γµ_(νρ)(dxν/dλ)Vρ = 0
Geodesic = self-parallel curve: DUµ/dτ = 0 where Uµ = dxµ/dτ
2. Curvature: Riemann, Ricci, and the Einstein Tensor
The Riemann curvature tensor measures the failure of parallel transport around a closed loop — equivalently, the commutator of covariant derivatives. Its contractions, the Ricci tensor and Ricci scalar, appear directly in the Einstein field equations.
Riemann Tensor & Its Contractions
Riemann tensor (16&sup4; = 256 components; 20 independent by symmetry):
Rσ_(μνρ) = ∂_νΓσ_(ρµ) − ∂_ρΓσ_(νµ) + Γσ_(νλ)Γλ_(ρµ) − Γσ_(ρλ)Γλ_(νµ)
Measures curvature via commutator:
[∇_μ, ∇_ν]Vρ = Rρ_(σμν) Vσ
Symmetries:
R_μνρσ = −R_νμρσ = −R_μνσρ = R_ρσμν
Algebraic Bianchi: R_[μνρσ] = 0
Differential Bianchi: ∇_[λR|μν|ρσ] = 0
Ricci tensor (symmetric, 10 independent):
R_μν = Rρ_(µρν) (contraction on 1st and 3rd indices)
Ricci scalar:
R = gμνR_µν
Einstein tensor (symmetric, divergence-free):
G_μν = R_μν − (1/2)g_μνR
∇ν Gμν = 0 (contracted Bianchi identity → conservation of Tμν)
Weyl tensor C_μνρσ:
Traceless part of Riemann
Governs tidal forces in vacuum (R_μν=0 does not imply Rσ_µνρ=0)
3. Einstein Field Equations
The Einstein field equations (EFE) describe how the distribution of energy and momentum sources the curvature of spacetime. They are 10 coupled, nonlinear, partial differential equations for the metric tensor gμν.
Einstein Field Equations & Linearised Gravity
Einstein field equations:
G_μν + Λ g_μν = (8πG/c&sup4;) T_μν
G_μν = spacetime curvature (geometry)
T_μν = energy-momentum tensor (matter/energy content)
Λ ≈ 1.1 × 10−&sup5;² m−² = cosmological constant (dark energy)
Perfect fluid energy-momentum tensor:
Tµν = (ρ + p/c²) UµUν + p gµν
Dust: p=0 → Tµν = ρc² UµUν
∇_ν Tµν = 0 ↔ Euler equations + energy conservation
Trace-reversed form (useful for linearisation):
Rµν = (8πG/c&sup4;)(Tµν − (1/2)gµνT)
Vacuum: Tµν = 0 → Rµν = 0 (Ricci-flat ≠ flat)
Linearised gravity hµν = gµν − ηµν (|h| ≪ 1):
Trace-reversed perturbation: h̄µν = hµν − (1/2)ηµνh
Lorenz gauge condition: ∂νh̄µν = 0
□h̄µν = −(16πG/c&sup4;) Tµν (wave equation)
□ = ηµν∂µ∂ν = −∂&sub2;t/c² + ∇² (d'Alembertian)
Newtonian limit: g_00 ≈ −(1 + 2Φ/c²), Φ = gravitational potential
∇²Φ = 4πGρ (Poisson equation recovered)
4. The Schwarzschild Solution
Karl Schwarzschild found the first exact solution of the EFE in December 1915 — just weeks after Einstein published the final equations — while on the Eastern Front during World War I. The Schwarzschild metric describes the spacetime geometry outside a spherically symmetric, non-rotating, uncharged mass.
Schwarzschild Metric, Orbits & Classical Tests
Schwarzschild metric (spherical symmetry, vacuum, M, J=0, Q=0):
ds² = −(1−r_s/r)c²dt² + (1−r_s/r)−¹dr² + r²dΩ²
r_s = 2GM/c² = Schwarzschild radius
Examples:
Sun: r_s = 2.95 km (R_sun = 696 000 km → well outside horizon)
Earth: r_s = 8.9 mm (R_earth = 6371 km)
M87*: r_s = 2 × 10^10 km (M ≈ 6.5 × 10^9 M_sun)
Effective potential for massive particles:
V_eff(r) = (c²/2)[(1−r_s/r)(1 + L²/r²c²) − 1]
where L = specific angular momentum
Innermost stable circular orbit (ISCO): r_ISCO = 3r_s = 6GM/c²
Orbital precession of Mercury:
Δφ_prec = 6πGM / [ac²(1−e²)] per orbit
Mercury: a = 0.387 AU, e = 0.206 → Δφ = 43.0 arcsec/century ✓
Gravitational redshift:
ν_obs/ν_emit = √(g_00(r_emit) / g_00(r_obs))
z = (1 − r_s/r)^(−1/2) − 1 ≈ GM/(rc²) for r ≫ r_s
Pound-Rebka (1959): measured z = 2.46 × 10^−15 at h=22.5m ✓
Photon sphere: r_ps = (3/2)r_s = 3GM/c²
Unstable circular orbit for photons
Shadow radius in EHT images: r_shadow = 3√3 GM/c²
M87* EHT image (2019): shadow diameter 40±3 µas (predicted 39±1 µas) ✓
5. Gravitational Waves
Accelerating mass-energy emits gravitational waves — ripples in spacetime geometry propagating at the speed of light. The transverse-traceless (TT) gauge selects physical degrees of freedom: two polarisations, h+ and h×, which alternately squeeze and stretch perpendicular directions as a wave passes through.
Quadrupole Radiation & LIGO Detection
TT-gauge solution (far field, r ≫ λ_GW):
h_ij^TT(t,r) = (2G/rc&sup4;) Ï_ij^TT(t_ret)
where t_ret = t − r/c (retarded time)
I_ij = ∫ ρ x_i x_j d³x (mass quadrupole moment tensor)
Total radiated power (quadrupole formula):
P = −G/(5c&sup5;) 〈Q̈_ij Q̈^ij〉 where Q_ij = I_ij − (1/3)δ_ij I_kk
Binary system (two equal masses m, separation a, circular orbit f_orb):
P = (32/5) G^4 m^5 / (c^5 a^5) (Peters formula)
a(t) = a_0 (1 − t/t_merge)^(1/4)
t_merge = (12/19) c_0^4/(B F(e)) ≈ 12/85 (c²0 a^4_0)/(4G^3 m² M)
LIGO GW150914 (2015 Sept 14):
Two black holes: m_1 ≈ 36 M_sun, m_2 ≈ 29 M_sun
Peak strain: h ~ 10^−²¹ at 150 Hz
LIGO arm length: L = 4 km → ΔL = hL/2 ~ 10^−¹&sup8; m ≈ 1/1000 proton radius
Final BH: M_f ≈ 62 M_sun (3 M_sun → GW energy: ~ 5 × 10^47 J)
Pulsar Timing Arrays (PTAs):
Nanohertz GW background detected 2023 (NANOGrav, EPTA, PPTA, CPTA)
f ~1–100 nHz, T_obs ~ 12.5 yr baselines, ~67 millisecond pulsars
LISA planned (2034, ESA/NASA):
3 spacecraft, 2.5 Mkm arm, 0.1 mHz–1 Hz
Targets: SMBH mergers z<20, Galactic double white dwarfs, EMRI waveforms
6. Black Holes and Hawking Radiation
A black hole is a region of spacetime where the escape velocity exceeds c: no causal signal can escape from within the event horizon. The Kerr solution (1963) extends Schwarzschild to spinning black holes and predicts the ergosphere — a region where spacetime is dragged so violently that nothing can remain stationary. Stephen Hawking showed in 1974 that quantum field theory near the horizon leads to thermal radiation, giving black holes a finite temperature and lifetime.
Kerr Metric, Penrose Process & Hawking Radiation
Kerr metric (Boyer-Lindquist coordinates, M, J=Ma, Q=0):
ds² = −(1−r_s r/Σ)c²dt² − (2r_s ra sin²θ/Σ)c dt dφ
+ (Σ/Δ)dr² + Σdθ² + (r²+a²+r_s ra²sin²θ/Σ)sin²θ dφ²
Σ = r² + a²cos²θ, Δ = r² − r_s r + a²
a = J/(Mc) = specific angular momentum (0 ≤ a ≤ GM/c)
Outer event horizon (largest root of Δ=0):
r_+ = GM/c² + √((GM/c²)² − a²)
Extremal Kerr: a = GM/c² → r_+ = GM/c² (smallest possible BH radius)
Ergosphere (where g_tt = 0):
r_ergo = GM/c² + √((GM/c²)² − a²cos²θ)
Between ergosphere and horizon, angular momentum can be extracted (Penrose process)
Penrose process:
Particle splits inside ergosphere: one falls in with negative E (w.r.t. infinity)
Other escapes with E_esc > E_initial → BH loses rotation
Maximum extractable energy: ~21% of rest mass (for extremal Kerr)
Blandford-Znajek process: magnetic version, powers AGN jets
Hawking radiation (1974 — quantum effect):
T_H = ℏc³ / (8πGM k_B)
Solar-mass BH: T_H ≈ 60 nanokelvins (negligible vs CMB 2.7 K)
Asteroid-mass BH (M ~ 10^15 g): T_H ~ 10^11 K → evaporating now
Black hole entropy (Bekenstein-Hawking):
S_BH = k_B A / (4 l_Pl²) where l_Pl = √(ℏG/c³) = 1.6 × 10^−35 m
A = 4πr_+² (event horizon area; increases in classical processes: 2nd law)
Information paradox:
Hawking radiation is exactly thermal → no information escapes?
Violates unitarity of quantum mechanics (Page curve debate)
Resolutions: fuzzballs (string theory), firewall (AMPS), Island formula (holography)
GPS and GPS satellites: The GPS system provides a direct practical application of GR. Clocks in GPS satellites run faster by 45 μs/day due to weaker gravity (gravitational blueshift), but slower by 7 μs/day due to orbital velocity (special relativistic time dilation). The net +38 μs/day correction is essential — without it, GPS position errors would accumulate at ~11 km/day.
Try These Simulations
Gravitational Lensing
Light ray bending around a Schwarzschild mass: Einstein rings, arc formation, and the lensing cross-section as you vary impact parameter and mass.
Minkowski Diagram
Interactive spacetime diagrams (ct vs x): worldlines, light cones, simultaneity planes, and Lorentz boosts visualised geometrically.
Twin Paradox
Symmetric proper-time calculation for the travelling twin: acceleration phase, cruise, turnaround, and the asymmetry that resolves the “paradox”.
Gravitational N-Body
Newtonian many-body gravity with optional post-Newtonian corrections: stable orbits, binary inspiral, and chaotic three-body trajectories.