Linearized Gravity and Gravitational Waves
Writes spacetime as flat plus a small perturbation and turns Einstein's equations into a wave equation whose solutions travel at the speed of light with two polarizations. The unit derives what a passing wave does to a ring of free test masses and what the strain measures, shows why a spinning dumbbell radiates while a pulsating sphere does not, states the quadrupole formula, and follows the Hulse-Taylor pulsar, whose shrinking orbit proved forty years before any detector that gravitational waves carry away energy.
Units 10 to 12 solved the field equations where gravity is strong and static, around a star or a black hole. This unit asks what the same equations say where gravity is weak and changing: far from any mass, in a region of spacetime that is almost flat, through which a disturbance from some distant source is passing. The answer is that the disturbance obeys a wave equation, travels at the speed of light, stretches and squeezes everything it crosses by a tiny fraction, and carries energy away from whatever made it. Einstein found the waves on paper in 1916; two neutron stars in a tight orbit showed in the 1970s that the energy loss is real; and since 2015 the waves themselves have been measured on the Earth, which Units 14 and 15 describe.
1 Flat spacetime plus a ripple
Far from the source of any gravitational wave, spacetime is very nearly the flat spacetime of Unit 9, with the metric \(\eta_{\mu\nu} = \mathrm{diag}(-1, 1, 1, 1)\) and the signature \((-,+,+,+)\) that every unit from Unit 9 on uses. The idea of linearized gravity is to write the true metric as the flat one plus a small correction, \begin{equation} g_{\mu\nu} = \eta_{\mu\nu} + h_{\mu\nu} , \qquad |h_{\mu\nu}| \ll 1 , \label{eq:lin} \end{equation} and to keep only the terms of the field equations that are linear in \(h_{\mu\nu}\). The metric perturbation \(h_{\mu\nu}\) is symmetric, like the metric, so it has ten independent components. As in every Advanced unit, a repeated index, once up and once down, is summed over \(0, 1, 2, 3\), and indices on \(h_{\mu\nu}\) are raised and lowered with \(\eta_{\mu\nu}\), because the difference from using the full metric would be of second order in \(h\).
When is the approximation allowed? The weak-field metric of Unit 10 has \(h_{00} = -2\Phi/c^2\) with the Newtonian potential \(\Phi\), and \(|h_{00}|\) is \(1.4 \times 10^{-9}\) at the surface of the Earth, about one part in a thousand million, and \(4 \times 10^{-6}\) at the surface of the Sun. The whole solar system is therefore a region where Equation \eqref{eq:lin} holds with a very small \(h\). A gravitational wave arriving at the Earth from two merging black holes has \(|h_{\mu\nu}|\) of order \(10^{-21}\), which is smaller still. Where the approximation fails is near the source itself: within a few Schwarzschild radii of a black hole, \(h\) is of order one and only the full theory of Units 11 and 12 applies. The waves of this unit are made where gravity is strong and are studied where it is weak.
The perturbation is not unique, because a small change of coordinates, \(x^{\mu} \to x^{\mu} + \xi^{\mu}(x)\) with \(\xi^{\mu}\) as small as \(h_{\mu\nu}\), leaves spacetime exactly as it was but changes the components of the perturbation to \begin{equation} h_{\mu\nu} \to h_{\mu\nu} - \partial_{\mu}\xi_{\nu} - \partial_{\nu}\xi_{\mu} , \label{eq:gauge} \end{equation} where \(\partial_{\mu} = \partial/\partial x^{\mu}\) as in Unit 9. This freedom is called gauge freedom, by analogy with the potentials of electromagnetism, which can also be changed without changing the fields. Four functions \(\xi^{\mu}\) can be chosen at will, so four of the ten components of \(h_{\mu\nu}\) carry no physics at all. The choice of coordinates that makes the field equations simplest is stated most easily in terms of the trace-reversed perturbation \begin{equation} \bar{h}_{\mu\nu} = h_{\mu\nu} - \tfrac12\,\eta_{\mu\nu}\,h , \qquad h = \eta^{\mu\nu}h_{\mu\nu} , \label{eq:tracerev} \end{equation} whose trace is \(-h\), since \(\eta^{\mu\nu}\eta_{\mu\nu} = 4\), which is where the name comes from. The coordinates can always be chosen so that \begin{equation} \partial^{\mu}\bar{h}_{\mu\nu} = 0 , \label{eq:lorenz} \end{equation} which is called the Lorenz gauge, after the Danish physicist Ludvig Lorenz (1829-1891), who imposed the same condition on the electromagnetic potentials in 1867 (the Lorentz of the Lorentz transformation is a different person). In this gauge the field equations of Unit 10, \(G_{\mu\nu} = (8\pi G/c^4)\,T_{\mu\nu}\) with \(\Lambda = 0\), collapse into a single wave equation, as the derivation box shows.