Reverberation Mapping

Reverberation mapping asks how a variable illuminating continuum is echoed by reprocessing matter. In an accreting black hole, energy-dependent delays can encode light-travel paths between a corona and disk, but the measured lag is not a ruler by itself. Intrinsic fluctuations propagate through the flow, both energy bands mix direct and reprocessed photons, phases wrap, detector response mixes energies, and the source can evolve within an exposure. The observable is a complex cross-spectrum; geometry enters only through a joint transfer model with declared sampling and calibration.

Required background. Waves, Green Functions, and Geometric Optics supplies retarded response and Fourier propagation used to define an impulse response and its phase.

Helpful background. Black-Hole Geometry and Mechanics supplies light-travel and redshift scales; Relativistic Matter and Compact Objects supplies radiative matter response; Black-Hole Astrophysics supplies accretion-state and inference context; and Computational Relativity supplies transfer-function and injection-recovery tests.

Cross-spectra and a signed lag

Use the transform pair consistent with the site’s $e^{-i\omega t}$ harmonic time dependence:

$$X(f)=\int_{-\infty}^{\infty}x(t)e^{+2\pi i f t}\,dt, \qquad x(t)=\int_{-\infty}^{\infty}X(f)e^{-2\pi i f t}\,df.$$

Let $x$ be a reference-band light curve and $y$ a comparison band. Define the cross-spectrum so a delayed $y(t)=x(t-\tau_0)$ has positive lag:

$$C_{yx}(f)=Y(f)X^*(f),\qquad \phi(f)=\arg\langle C_{yx}(f)\rangle,\qquad \tau(f)=\frac{\phi(f)}{2\pi f}.$$

Indeed, substituting $t'=t-\tau_0$ gives $Y=e^{+2\pi if\tau_0}X$, so $C_{yx}=|X|^2e^{+2\pi if\tau_0}$. The averaging brackets denote segments and, often, neighboring frequency bins. Many timing codes instead use a negative analysis exponent; conjugating both the transforms and the cross-spectrum order preserves this positive-delay definition. Translate the pair together. Sampling uncertainty depends on the number of averages and intrinsic coherence, while Poisson noise increases phase scatter. Uttley and collaborators derive the estimators and response relations in Uttley et al. 2014, §§ 2.1.2–2.1.3 and 2.2–2.4.

Transfer functions, propagation, and dilution

For a stationary linear response, a comparison band may be written

$$y(t)=a\,x(t)+b\int_0^\infty\psi(u)x(t-u)\,du+n(t),$$

where $a,b\geq0$ are the direct and reprocessed response weights, $\int_0^\infty\psi(u)\,du=1$, and $n$ is additional variability uncorrelated with the driver in this example. In Fourier space the transfer is $a+b\Psi(f)$. For a single delay, $\psi(u)=\delta(u-\tau_0)$, this becomes $a+be^{+2\pi if\tau_0}$. At small phase,

$$\tau_{\rm obs}\simeq\frac{b}{a+b}\tau_0, \qquad 2\pi f\tau_0\ll1.$$

Direct photons dilute the low-frequency lag. This is not an all-frequency inequality: interference and wrapping can change its magnitude or sign. The $x$ in this illustrative response is an uncontaminated driving continuum. If both observed bands contain echoes of a latent driver, $X=H_xS$ and $Y=H_yS$, then $C_{yx}=H_yH_x^*\langle|S|^2\rangle$. The measured lag is the difference of the two response phases, not the delay in either band alone.

Accretion-rate fluctuations introduce another causal chain. Slow fluctuations generated at large radius can propagate inward, modulating progressively harder emission and producing low-frequency hard lags. Reflection can add shorter soft or iron-band lags at higher frequency. Spectral pivoting, distant scattering, and an extended moving corona can yield related patterns. Ingram summarizes how mean spectra, power, cross-spectra, covariance, and reverberation models should be used together in Ingram 2023, §§ 2–3.

No sign is unique to a mechanism. When a true phase exceeds $\pi$, the reported principal phase wraps to the opposite sign. A lag–energy spectrum also depends on its reference band; including the channel of interest in that reference introduces correlated noise and additional dilution.

From lag structure to geometry

A compact disk-reverberation model predicts a coupled set of observables: a frequency-dependent response width, an iron-band lag–energy feature, a soft reflection lag, a covariance spectrum, and a mean reflection spectrum. General-relativistic propagation adds Shapiro delay, lensing, energy shifts, inclination, disk inner edge, and source geometry. A height inferred from lags must also be compatible with the illumination profile and reflection fraction.

The conversion to $GM$ requires an independently constrained mass. Geometry remains degenerate: an extended corona can broaden the transfer; radial propagation can change the phase baseline; dilution can make a large path appear short; and a distant reprocessor can contribute at low frequency. Reynolds discusses these timing constraints alongside reflection and spin systematics in Reynolds 2021, §§ 3.2.3–3.2.4. QPO phase lags and state dependence provide an independent timing axis but not an automatic geometric solution Ingram and Motta 2019, §§ 2.1–2.3 and 5.1–5.6.

Named event release and reduction boundary

The concrete data release is the 2008 XMM-Newton EPIC-pn long look at 1H 0707−495: ObsIDs 0511580101, 0511580201, 0511580301, and 0511580401. Miller and collaborators list the four O4 exposures, dates, durations, and count rates in Miller et al. 2010, § 2 and Table 1. Their historical reduction used SAS 8.0.0 and HEASoft 6.8, pn patterns 0–4, a 35-arcsec source region, tight background-flare screening, and 0.3–1, 1–4, and 4–7.5 keV bands. This page performs no new event reduction or likelihood fit.

A modern software reproduction must pin the XMM observation data files, SAS and current-calibration-file release, clock correction, event flags and patterns, extraction regions, response matrices, pile-up check, background threshold, good-time intervals, bin size, segment length, reference band, excluded channels, and gap treatment. It must recompute complex cross-spectra and covariance, inject a known response through the same sampling and energy redistribution, and test alternative reference bands and segment lengths. Statistical intervals require the sampling likelihood or simulations; calibration and nonstationarity remain separate systematic uncertainties.

Results and limitations through 2026-08-18

The existence of energy-dependent X-ray lags is durable. Their division into compact disk reverberation, inward propagation, spectral pivoting, and extended or distant scattering remains unresolved; credible competing positions depend on frequency range, source state, response, and transfer-function parameter regime.

Position or evidence	Result	Scope and limitation
Linear impulse-response analysis	Predicts coupled lag–frequency, lag–energy, covariance, and mean spectra	Dilution and phase wrapping can reverse or suppress the apparent lag
Compact disk reflection	Naturally links soft and iron-band lags to relativistically blurred reflection	Height and inner radius depend on mass, emissivity, continuum, disk density, and coronal extent
Inward propagation and pivoting	Explains broad low-frequency hard lags and state evolution	Must be fit jointly because it supplies the phase baseline beneath reverberation
1H 0707−495 O4 event release	Shows alternating lag signs across energy and frequency	Miller et al. demonstrate an extended scattering interpretation; one object does not select a universal geometry
Simulations and null recovery tests	Can measure bias from response, gaps, noise, reference bands, and model mismatch	Recovery is only as broad as the injected model family and calibration perturbations

This treatment should be reviewed after a new high-throughput timing data release changes the lag–energy structure, after an instrument clock or response calibration revision, after an independent transfer-function recovery succeeds across mean spectra and covariance, or after a model jointly fits propagation and reflection while decisively outperforming the competing extended-response explanations.

Common pitfalls

Multiplying lag by light speed immediately. The observed phase is diluted and mixed. A path length follows only after direct and reprocessed fractions and the full transfer function are modeled.

Ignoring phase wrapping. A negative principal lag can represent a positive delay whose phase has crossed $\pi$.

Exercise

At $f=3.0\times10^{-3}$ Hz, a transfer component has a true positive delay of $200$ s. Find its unwrapped phase, the principal phase in $(-\pi,\pi]$, and the apparent principal-value lag. What warning follows?

Solution

The unwrapped phase is

$$\phi=2\pi f\tau=2\pi(3.0\times10^{-3})(200)=3.77\ {\rm rad}.$$

Subtracting $2\pi$ gives the principal phase $\phi_{\rm p}=-2.51$ rad. If that wrapped phase is inserted without unwrapping,

$$\tau_{\rm p}=\frac{-2.51}{2\pi(3.0\times10^{-3})} =-133\ {\rm s}.$$

The same positive physical delay appears negative. A sign change in a lag spectrum is therefore not, by itself, evidence that the causal order reversed; phase continuity and a transfer model must be checked.

References

• Ingram, A. (2023). “Fourier Spectral-Timing Techniques for the Study of Accreting Black Holes.” Astronomische Nachrichten 344, e20230002. DOI; Open PDF.
• Ingram, A., and Motta, S. E. (2019). “A Review of Quasi-Periodic Oscillations from Black Hole X-Ray Binaries: Observation and Theory.” New Astronomy Reviews 85, 101524. DOI; Open PDF.
• Miller, L., Turner, T. J., Reeves, J. N., and Braito, V. (2010). “X-ray Reverberation in 1H 0707−495 Revisited.” Monthly Notices of the Royal Astronomical Society 408, 1928–1935. DOI; Open PDF.
• Reynolds, C. S. (2021). “Observational Constraints on Black Hole Spin.” Annual Review of Astronomy and Astrophysics 59, 117–154. DOI; Open PDF.
• Uttley, P., Cackett, E. M., Fabian, A. C., Kara, E., and Wilkins, D. R. (2014). “X-Ray Reverberation Around Accreting Black Holes.” Astronomy and Astrophysics Review 22, 72. DOI; Open PDF.