Strong-field gravitational lensing by black holes
Jake O. Shipley School of Mathematics and Statistics University of Sheffield
Applied Mathematics Colloquium University of Sheffield 5 December 2018 Outline
1. General relativity and gravitational lensing
2. Black holes
3. Strong-field tests of general relativity
4. Gravitational lensing by black holes
5. Conclusions 1. General relativity and gravitational lensing General relativity
I Geometric theory of space, time and gravitation, published by A. Einstein in 1915.
I Gravitational field is coded in a metric gab of Lorentzian signature on the curved 4D spacetime manifold M.
I GR relates curvature of spacetime to energy and momentum of matter in spacetime through Einstein’s field equations,
1 8πG R − R g = T . (1) ab 2 ab c4 ab
I Light follows null geodesics on curved spacetime.
I Key consequence of GR: light rays are deflected by gravitational fields. Gravitational deflection of light Newtonian prediction
I Traces of the idea of light deflection date back to I. Newton:
“Do not Bodies act upon Light at a distance, and by their action bend its Rays, and is not this action (cæteris paribus) strongest at the least distance?”
— Isaac Newton, Query 1, Opticks, 1704 Gravitational deflection of light Newtonian prediction
I First (unpublished) calculation using corpuscular model of light and Newtonian theory by H. Cavendish (1784).
I First published calculation by J. G. von Soldner (1804):
2GM δθ = . (2) N c2b
I M = mass of body, b = impact parameter, G = gravitational constant, c = speed of light. Gravitational deflection of light Einstein’s predictions
I Einstein’s first prediction (1911) using only the equivalence principle (special relativity) was equal to those of Cavendish and Soldner.
I For the Sun, δθN ' 0.87 arcsec.
I Attempted observation of light deflection by the Sun during 1914 solar eclipse on Crimean Peninsula by E. Freundlich.
I Unfortunately, Freundlich and colleagues were imprisoned. They were eventually released, but missed the eclipse!
I However, Einstein’s 1911 calculation was incorrect.
I In 1916, Einstein found that the deflection angle predicted by GR is twice the Newtonian prediction:
4GM δθ = 2δθ = . (3) GR N c2b Gravitational deflection of light Experimental verification
I For a ray which grazes the Sun’s surface, the value would be
4GM δθ = 2 ' 1.74 arcsec. (4) c R
I In 1919, this was determined to within 20% during A. Eddington’s solar eclipse experiment.
I More recently, this result has been verified Figure: Solar Eclipse, 1919. experimentally to within 0.0002%. [Credit: F. W. Dyson et al., Philos. Trans. Royal Soc. A 220 (1920).] Weak-field lensing
I Gravitational fields affect light in a similar way to an optical lens.
I Light deflection by gravitational fields = gravitational lensing.
I Weak-field gravitational lensing involves:
GM 2 weak gravitational fields, |ΦN| = r c ; and small deflection angles δθ.
I Theory of weak-field gravitational lensing is well understood. [S. Dodelson, Gravitational Lensing (CUP, 2017).]
I Consequences of weak-field lensing: Einstein rings; multiple images; image distortion, . . . Weak-field lensing phenomena Multiple images and Einstein rings
I Multiple rays reach observer from a single source ⇒ multiple images of the source along the tangent to the ray.
I Perfect alignment of source, lens and observer results in an Einstein ring.
I Einstein radius is
r 4GM θE ' 2 . c DL Weak-field lensing phenomena Multiple images and Einstein rings
Figure: Einstein ring observed by Sloan Digital Sky Survey & Hubble Space Telescope. [Credit: ESA/Hubble & NASA.] Testing general relativity
I Three classical tests of GR proposed by Einstein in 1916:
1. perihelion precession of Mercury’s orbit;
2. deflection of light by the Sun;
3. gravitational redshift of light in the solar system.
I All established GR as our best theory of gravity.
I These are all weak-field tests of GR.
I GR also leads to objects with very strong gravitational fields, including black holes and neutron stars. 2. Black holes Schwarzschild black hole
a I Spacetime geometry specified in coordinates x by line element 2 a b ds = gab dx dx , where gab is a 4D Lorentzian metric.
I Solution found by K. Schwarzschild in 1916:
2M 2M −1 ds2 = − 1 − dt2 + 1 − dr2 + r2(dθ2 + sin2 θ dφ2). r r
I Describes a spherically symmetric black hole (BH) of mass M.
I The BH is characterised by:
I an event horizon (“point of no return”) at r = 2M;
I a curvature singularity at r = 0. Kerr(–Newman) black hole
I Schwarzschild solution generalised by R. Kerr (1963) to include rotation.
I Generalised by E. Newman (1965) to include electric charge.
I Kerr–Newman solution describes spacetime of a charged, rotating BH of mass M, angular momentum J = aM, and electric charge Q.
I Kerr–Newman solution is given by the line element
2Mr Σ ds2 = − 1 − dt2 + dr2 + Σ dθ2 Σ ∆ 2Ma2r 4Mra sin2 θ + r2 + a2 + sin2 θ sin2 θ dφ − dt dφ, Σ Σ
where Σ(r, θ) = r2 + a2 cos2 θ, ∆(r) = r2 − 2Mr + a2 + Q2.
I Not spherically symmetric due to rotation. No-hair conjecture
Conjecture Stationary isolated BHs in GR are uniquely characterised by three parameters: mass M, angular momentum J, and electric charge Q.
I All other parameters are “hair”.
I Stationary isolated BHs in GR are therefore described by the Kerr–Newman solution.
I What about astrophysical BHs in the Universe? Kerr’s solution and astrophysical black holes
I Q thought to be negligible for astrophysical BHs.
I Astrophysical BHs described by Kerr’s solution:
“The most shattering experience has been the realization that [Kerr’s] solution of Einstein’s equations of general relativity provides the abso- lutely exact representation of untold numbers of massive black holes that populate the universe.” — Subrahmanyan Chandrasekhar, 1975
I Can we test the no-hair conjecture, and other aspects of strong gravity, by observing astrophysical BHs? 3. Strong-field tests of general relativity Gravitational waves
I Gravitational waves: ripples in the fabric of spacetime caused by accelerating masses.
I GW150914: GW signal from the merger of two BHs detected by LIGO.
I Confirmed existence of BHs in Nature.
I Nobel Prize (2017): Awarded to R. Weiss, B. Barish and K. Thorne “for decisive contributions to the LIGO detector and the observation of gravitational waves”.
I Four new detections announced this week (03/12/18)! Multimessenger astronomy
I GW170817: “gravity and light”.
I LIGO–Virgo: GWs from neutron star merger.
I Telescopes observed gamma-ray burst (and afterglow) from EM counterpart.
I Birth of multimessenger astronomy. Before we could only see the universe, now we can “hear” it. The Event Horizon Telescope Overview
I Aims:
I Image nearby galactic centres.
I Take the first picture of a BH at exquisite angular resolution at event horizon scales.
wavelength angular resolution ∼ . I dish size
I Employs mm-wavelength very-long-baseline interferometry.
Figure: The Event Horizon I Global array of radio telescopes. Telescope. [Credit: EHT.]
I Earth-sized virtual telescope. The Event Horizon Telescope Main aims
I EHT’s targets:
1. Sgr A∗ 6 4 M ∼ 4 × 10 M , D ∼ 8 × 10 pc, θ ∼ 53 µas
2. M87 9 6 M ∼ 6 × 10 M , D ∼ 16 × 10 pc, θ ∼ 22 µas
I View the BH shadow: “silhouette” of the BH.
I Test GR: deviations from Kerr; no-hair conjecture.
I Understand accretion and jets. The Event Horizon Telescope Results
I So what does a BH actually look like?
I Results from the EHT expected in early 2019!
I However, we can answer this question theoretically. 4. Gravitational lensing by black holes Black hole shadow
I What would an observer see if they were to look at a BH?
I Light from distant stars does not refract off the BH – we can’t actually see it . . .
I . . . but light is deflected by the BH’s gravitational field.
I Some light is scattered, some is absorbed ⇒ BH will cast a shadow (or “silhouette”).
I Shadow = region of the observer’s sky which lacks radiation, due to the blockage of the BH. Ray-tracing
Undeflected ray
Black hole I We simulate BH shadows using Background backwards ray-tracing. Deflected rays Camera plane
I Shadow associated with the set of photons which reach the BH event horizon, when traced back in time from the observer.
I See Simulating eXtreme Spacetimes for simulations: https://www.black-holes.org/. Lensing by a Schwarzschild black hole
I Schwarzschild spacetime is spherically symmetric: consider motion in π equatorial plane (θ = 2 ).
I Curvature singularity at r = 0; event horizon at r = 2M. √ I Unstable photon sphere (or light-ring) at rcrit = 3M, bcrit = 3 3M.
I Determines the boundary of the BH shadow.
10
5 5
0 0
-5 -5
-10 -10 -5 0 5 10 -5 0 5 Lensing by a Kerr black hole Null geodesics
10
5
0
-5
-10 -10 -5 0 5 10
I Kerr BH rotates with angular momentum J = aM (0 < a < 1).
I Prograde (+) and retrograde (−) circular photon orbits:
2 a r = 2M 1 + cos arccos ∓ , b = ±3pMr − a. (5) ± 3 M ± ± Lensing by a Kerr black hole Black hole shadow
10 8 θ = 90 6
5 4
2
0 0
-2 -5
-4
-6 -10 -10 -5 0 5 10 -8 -8 -6 -4 -2 0 2 4 6 8
I Geodesic motion is integrable (due to existence of Carter constant).
I Equation for shadow boundary known in closed form.
I Shadow deformed due to lack of spherical symmetry. Binary black holes
I Pairs of spinning BHs which orbit one another – binary BHs – have recently been discovered by LIGO.
I Verification that BHs occur in binary pairs.
I LIGO–Virgo since confirmed an abundant binary BH population; 10 GW detections so far. Lensing by binary black holes
I What would a binary BH merger look like?
I Problem: No known solution to EFEs describing a dynamical binary.
I Solution: Evolve null geodesic equations on a fully numerical simulation to EFEs. (This is computationally very expensive!)
I Binary shadow is not superposition of two singleton BH shadows. Figure: Binary shadow with eyebrows. [Credit: SXS Lensing I Binary BH shadows look like they have Group, www.black-holes.org.] interesting features, e.g. eyebrows.
I Can we understand qualitative features of lensing by binary BHs by considering a simple imitative model? Toy model: Majumdar–Papapetrou static binary black hole
I Majumdar–Papapetrou (MP) binary BH solution to Einstein–Maxwell equations given in cylindrical coord’s {t, ρ, z, φ} by
2 −2 2 2 2 2 2 2 −1 ds = −U dt + U dρ + dz + ρ dφ ,Aa = [U , 0]. (6)
with
M+ M− ±dM∓ U(ρ, z) = 1 + + , z± = . p 2 2 p 2 2 ρ + (z − z+) ρ + (z − z−) M+ + M−
I Describes two fixed BHs of mass M±(= Q±) in static equilibrium, separated by coordinate distance d.
I Focus on equal-mass case M± = M = 1.
I To analyse lensing effects, consider geodesic motion, governed by Hamiltonian 1 1 1 H = gabp p = p 2 + p 2 − (ρU 2 − p )(ρU 2 + p ) = 0. (7) 2 a b 2 ρ z 2ρ2 φ φ Fate of rays for the static binary black hole
I Planar rays: pφ = 0.
10 I Rays can: 1. fall into the upper BH; 2. fall into the lower BH; 5 3. escape to spatial infinity.
I Open Hamiltonian system with three 0 exits.
I Fate of rays can be visualised using an exit 0 5 10 basin diagram.
I This is actually a 1D binary BH shadow. BH1
I Aim: understand set of non-escaping BH2 orbits – the repellor ΩR. (Corresponds to 0 2 4 6 8 10 boundary of BH shadow.) Non-escaping orbits and the repellor
I There are thee types of unstable, periodic “fundamental orbits”. 10
5 I Fundamental orbits are “dynamically connected”. A ray can transition between asymptotic neighbourhoods of fundamental 0 orbits. -5
There exists . . . -10 I a countable infinity of periodic orbits. -10 -5 0 5 10
I an uncountable infinity of aperiodic orbits.
2 2 2 2
1 1 1 1
0 0 0 0
-1 -1 -1 -1
-2 -2 -2 -2
-2 -1 0 1 2 -2 -1 0 1 2 -2 -1 0 1 2 -2 -1 0 1 2 Chaotic scattering
Eckhardt’s definition of chaotic scattering (1988):
I Scattering in a Hamiltonian system is irregular (or chaotic) if there exists on some manifold of initial data an infinity of distinct “scattering singularities” of zero (Lebesgue) measure, typically arranged into a fractal set.
I Scattering singularity: a value of the initial data for which the scattering process is undefined, and some physical quantity (e.g. time-delay, deflection angle) becomes singular.
For the Schwarzschild BH, there is a unique scattering singularity, I √ bcrit = 3 3M, r = 3M.
I For the Kerr BH, there are two distinct scattering singularities, b = b±,
r = r±.
I Neither system exhibits chaotic scattering. Binary black holes and chaotic scattering
In our work† ...
I We find that static binary BHs exhibit chaotic scattering.
I Each distinct perpetual orbit generates a scattering singularity in the initial data (e.g. impact parameter b).
I Any open interval in b contains either zero or infinitely many scattering singularities.
I Set of scattering singularities = strange repellor.
The scattering singularities. . .
I have a Cantor-like distribution in initial data (1D self-similar fractal);
I have zero (Lebesgue) measure;
I can be understood using symbolic dynamics.
† J. O. Shipley & S. R. Dolan, Classical and Quantum Gravity 33, 175001 (2016); arXiv:1603.04469 [gr-qc]. Self-similarity of the one-dimensional shadow
BH1
10
BH2 5.75 6.10 6.45 6.80 7.15 7.50 5
BH1 0
BH2 6.94 6.97 7.00 7.02 7.05 7.08 0 5 10
BH1 BH1
BH2 BH2 0 2 4 6 8 10 7.040 7.042 7.044 7.047 7.049 7.051 Static binary black hole shadows 5. Conclusions Conclusions
I Weak-field gravitational lensing phenomena: multiple images, Einstein rings, . . .
I GR predicts BHs – as confirmed by LIGO–Virgo detections.
I Will soon observe BHs directly with the EHT (early 2019). Images formed will help us test GR in the strong-field regime.
I BHs exhibit interesting (strong) lensing effects associated with light rings and event horizons: black hole shadows.
I Geodesic motion on Kerr spacetime is integrable → shadow is “boring”!
I Can understand lensing by binary BHs using a toy model of two fixed BHs.
I Geodesic motion on Majumdar–Papapetrou is non-integrable → chaotic scattering → fractal BH shadows with a Cantor-like structure. Selected references
1. Simulating eXtreme Spacetimes: www.black-holes.org. 2. S. Dodelson, Gravitational Lensing (CUP, 2017). 3. D. Psaltis, arXiv:1806.09740 [astro-ph.HE] (2018). 4. B. Eckhardt, Journal of Physics A 20, 5971 (1987). 5. G. Contopoulos, Proc. Roy. Soc. Lond. A 431, 183 (1990). 6. A. Bohn et al., Class. Quant. Grav. 32, 065002 (2015). 7. J. O. Shipley & S. R. Dolan, Class. Quant. Grav. 33, 175001 (2016). 8. S. R. Dolan & J. O. Shipley, Phys. Rev. D 94, 044038 (2016). 9. A.´ Daza, J. O. Shipley, S. R.Dolan & M. A. F. Sanju´an, Phys. Rev. D 98, 084050 (2018). Appendix Gravitational lensing at home
I Can mimic gravitational lensing using a logarithmic lens (e.g. the foot of a wine glass). Geodesic motion
a I A geodesic is a spacetime path q (λ) which extremises the action
Z dqa S[qa(λ)] = L(qa, q˙a) dλ, q˙a = , λ = affine parameter . (8) dλ
1 a b I Lagrangian: L = 2 gab q˙ q˙ . ∂L Conjugate momenta: p = . I a ∂q˙a
1 ab I Hamiltonian: H = 2 g papb.
a ∂H ∂H I Null geodesics from Hamilton’s equations: q˙ = , p˙a = − a . ∂pa ∂q Symmetries
I Null geodesics ⇒ L = 0 = H.
I Static ⇒ pt = constant; set pt = −1 w.l.o.g.
I Axisymmetric ⇒ pφ = constant. Geodesic motion
I In the MP di-hole spacetime, geodesic motion is governed by a 2D
Hamiltonian with a conserved parameter (pφ) for rays: 1 1 H = gabp p = p 2 + p 2 + V = 0, (9) 2 a b 2 ρ z with geodesic potential 1 V (ρ, z) = − (ρU 2 − p )(ρU 2 + p ). (10) 2ρ2 φ φ
I Motion in (ρ, z) is not separable.
I The system is non-integrable.
I We should anticipate rich dynamics. Static binary black hole shadows Viewing angles
0° 10° 20° 30° 40°
50° 60° 70° 80° 90°