Gravitational Lensing

Gravitational lensing is the astrophysical and cosmological phenomenon in which the trajectory, flux, arrival time, and apparent shape of light (or any electromagnetic radiation) emitted by a distant background source are deflected and distorted as the radiation propagates through the curved spacetime geometry of an intervening mass distribution—such as a star, black hole, galaxy, or galaxy cluster—situated along the line of sight. Grounded in Albert Einstein's general relativity (1915–1936), mathematically generalized to extragalactic nebulae by Fritz Zwicky in 1937, and time-delay formulated by Sjur Refsdal in 1964, gravitational lensing was first observed at cosmological distances in 1979 with the discovery of the Twin Quasar (SBS 0957+561 A/B) by Dennis Walsh, Robert Carswell, and Ray Weymann.

Within observational cosmology and extragalactic astrophysics, gravitational lensing acts both as a natural cosmic telescope—magnifying faint high-redshift galaxies at z > 10 into the sensitivity window of space observatories—and as a direct, model-independent probe of total gravitating mass. Because spacetime curvature in general relativity responds identically to luminous baryonic matter and non-baryonic dark matter regardless of dynamical equilibrium or thermodynamic state, strong, weak, and microlensing observations reconstruct the spatial distribution of dark matter halos, measure the Hubble constant (H₀), and detect non-luminous compact objects and exoplanets.

Null Geodesic Deflection Angle (α̂ = 4*GM* / bc²)

In a weak, stationary gravitational field around a localized mass M (GM / bc² << 1), the spacetime metric tensor g_μν in isotropic coordinates is perturbed from flat Minkowski spacetime by the Newtonian potential Φ(r) = −*GM* / r:

Photons follow null geodesics (ds² = 0), propagating at a coordinate speed of light c′ ≈ c(1 + 2Φ/c²) that slows inside the gravitational potential well, equivalent to an effective Fermat refractive index n(r) = 1 − 2Φ(r) / c² > 1. Integrating the transverse gradient of n(r) along an unperturbed ray passing a point mass M at impact parameter b yields Einstein's gravitational deflection angle (α̂):

where r_s = 2*GM*/c² is the Schwarzschild radius (and the factor of 4—twice the 1804 Johann Georg von Soldner Newtonian corpuscular value—arises from equal contributions of temporal g₀₀ and spatial g_ij curvature). For a grazing ray at the solar limb (M = M_⊙, b = R_⊙), α̂ = 1.7505 arcseconds, confirmed by Arthur Eddington's 1919 solar eclipse observations.

Thin-Lens Geometry, the Lens Equation, and the Angular Einstein Radius (θ_E)

Because the physical thickness of a lensing galaxy or cluster (~kpc to Mpc) is tiny compared to the gigaparsec angular-diameter distances between the observer and lens (D_L), between the observer and background source (D_S), and between the lens and source (D_LS), deflection is modeled in the thin-lens approximation. Mapping the true unlensed angular position of the source β on the sky to its observed lensed image position θ via the reduced deflection angle α(θ) = (D_LS / D_S) α̂(D_Lθ) yields the fundamental gravitational lens equation:

For an axially symmetric point-mass lens M, substituting α̂ = 4*GM* / (D_L θ c²) gives the scalar quadratic equation β = θ − θ_E² / θ. When the background source lies directly behind the lens on the optical axis (β = 0), rotational symmetry deforms the source image into a closed circle of light—an Einstein ring—of angular radius θ_E (the Einstein radius):

For a galaxy or cluster of mass M ~ 10¹¹–10¹⁴ M_⊙ at cosmological distances (D ~ 1 Gpc), θ_E spans 1 to 30 arcseconds (resolvable by optical and radio telescopes); for a stellar-mass lens (M ~ 1 M_⊙) in the Milky Way Bulge or Magellanic Clouds (D ~ 8–50 kpc), θ_E is ~0.5 to 1 milliarcsecond.

Critical Surface Mass Density (Σ_crit), Convergence (κ), Shear (γ), and Magnification (μ)

For an extended mass distribution projected onto the lens plane with 2D surface mass density Σ(θ), multiple imaging occurs whenever Σ(θ) exceeds the critical surface mass density (Σ_crit):

The dimensionless ratio κ(θ) = Σ(θ) / Σ_crit is the convergence, which causes isotropic focusing (magnifying image size). Simultaneously, the tidal gravitational field of the surrounding mass distribution introduces anisotropic stretching quantified by the two-component trace-free complex shear γ = γ₁ + *i*γ₂. The 2 × 2 Jacobian distortion matrix of the lens mapping A(θ) = ∂β / ∂θ is:

Because gravitational lensing conserves surface brightness (by Liouville's theorem in phase space), the scalar flux magnification factor (μ) of an image is given by the inverse determinant of the Jacobian matrix:

Loci in the lens plane where det A(θ) = 0 (formally infinite magnification in geometric optics) define smooth closed critical curves (tangential and radial), which map via the lens equation onto cusped caustics in the source plane; whenever a background source crosses a caustic, a pair of lensed images is created or annihilated.

Strong Arcs, Weak Cosmic Shear, and Paczyński Microlensing Light Curves

Depending on impact parameter and surface density κ, lensing divides into three observational regimes:

1. Strong Gravitational Lensing (κ ≥ 1): Produces multiple resolved images of quasars (e.g., two-image doubles or four-image Einstein crosses plus a demagnified central fifth image by Burke's odd-number theorem), giant tangential arcs in galaxy clusters (such as Abell 1689 and SMACS 0723), and multiply imaged supernovae (SN Refsdal).
2. Weak Gravitational Lensing (κ << 1, ‖γ‖ << 1): In the outer halos of galaxy clusters and across large-scale cosmic web filaments, individual background galaxies are not multiply imaged, but their intrinsic ellipticities undergo a coherent 1% to 5% tangential stretch (cosmic shear). Averaging ellipticities over millions of background galaxies (in surveys such as Euclid, Vera C. Rubin Observatory LSST, and Nancy Grace Roman Space Telescope) reconstructs the projected total mass density κ(θ) via the Kaiser–Squires Fourier inversion.
3. Gravitational Microlensing: When a foreground star, brown dwarf, free-floating planet, or stellar-mass black hole in the Milky Way passes in front of a background bulge or Magellanic Cloud star with angular impact parameter u(t) = β(t) / θ_E, the two micro-images (separated by ~1 mas) are unresolved, but their combined brightness follows Bohdan Paczyński's (1986) achromatic magnification curve: A(u) = (u² + 2) / [ u (u² + 4)^1/2 ]. A planetary companion orbiting the lens star near its Einstein ring introduces a sharp caustic perturbation lasting hours to days, enabling microlensing surveys (OGLE, MOA, KMTNet, Roman) to detect cold terrestrial and jovian exoplanets near and beyond the snow line.

The Bullet Cluster (1E 0657-558) and Sjur Refsdal's Hubble Constant (H₀) Method

Two milestone applications highlight gravitational lensing's role in fundamental physics. First, in 2006 Douglas Clowe, Marusa Bradač, and colleagues combined weak and strong gravitational lensing maps of the Bullet Cluster (1E 0657−558)—a high-velocity merger of two galaxy clusters—with Chandra X-ray observations of the hot intracluster plasma. Whereas the collisional X-ray-emitting baryonic gas (representing ~85% of the normal matter) was slowed by ram pressure near the collision center, the weak-lensing convergence peaks (κ) traced the collisionless galaxies with an 8σ spatial offset from the baryonic gas, providing direct empirical evidence that the majority of gravitating mass in galaxy clusters consists of collisionless dark matter.

Second, as proposed by Sjur Refsdal (1964), the arrival-time difference Δ*t*_AB between two lensed images θ_A and θ_B of a variable quasar or supernova combines the geometric path-length difference with the relativistic Shapiro gravitational time delay through the Fermat potential Φ_F(θ, β):

Because the time-delay distance D_Δt = (1 + z_L) D_L D_S / D_LS is inversely proportional to the Hubble constant (H₀), measuring Δ*t*_AB alongside lens galaxy stellar kinematics yields an independent geometric measurement of H₀.

How does gravitational lensing magnify a distant galaxy if it conserves surface brightness?

By Liouville's theorem, gravitational deflection neither creates nor destroys photons along a bundle of rays in vacuum, so the surface brightness (flux per unit solid angle on the sky, I_ν) of a lensed source at fixed redshift remains strictly invariant. However, the lens Jacobian matrix stretches the apparent solid angle (*d*Ω) subtended by the background galaxy on the observer's sky by the factor μ = 1 / [(1 − κ)² − ‖γ‖²]. Because total observed flux is the product of surface brightness and solid angle (F = I_ν *d*Ω), enlarging the apparent area on the sky increases the total integrated photon flux received by the telescope.

Why is gravitational lensing achromatic (independent of light wavelength)?

Unlike optical glass lenses that bend light via wavelength-dependent dispersive refraction (n(λ)), gravitational lensing is a purely geometric effect of general relativity: photons of all energies travel along the exact same null geodesics (ds² = 0) of the curved spacetime metric g_μν. Consequently, radio waves, visible light, and X-rays emitted from the same spatial region of a background source experience identical deflection angles α̂ = 4*GM*/(bc²), a key signature used to distinguish true microlensing events from intrinsic stellar flares.

How did weak lensing of the Bullet Cluster provide direct evidence for dark matter?

In a quiescent galaxy cluster, both the hot X-ray-emitting intracluster gas (which accounts for roughly six times more baryonic mass than all the stars inside the cluster's galaxies) and the dark matter halo sit in the same central gravitational potential well. During the supersonic collision of the two subclusters in the Bullet Cluster (1E 0657−558), electromagnetic ram pressure decelerated the collisional X-ray plasma near the center, while the collisionless galaxies passed straight through. Weak gravitational lensing reconstructed the total gravitational potential and showed that the center of mass tracked the collisionless galaxies rather than the dominant baryonic gas mass at >8σ significance.

1. Einstein, A. (1936). "Lens-like action of a star by the deviation of light in the gravitational field." Science, 84(2188), 506–507. https://doi.org/10.1126/science.84.2188.506
2. Walsh, D., Carswell, R. F., & Weymann, R. J. (1979). "0957 + 561 A, B: twin quasistellar objects or gravitational lens?" Nature, 279(5712), 381–384. https://doi.org/10.1038/279381a0
3. Paczyński, B. (1986). "Gravitational microlensing by the galactic halo." The Astrophysical Journal, 304, 1–5. https://doi.org/10.1086/164140
4. Clowe, D., Bradač, M., Gonzalez, A. H., Markevitch, M., Randall, S. W., Jones, C., & Zaritsky, D. (2006). "A direct empirical proof of the existence of dark matter." The Astrophysical Journal Letters, 648(2), L109–L113. https://doi.org/10.1086/508162