Fingerprints of Individual Supermassive Black Hole Binaries in Pulsar Timing Arrays

Abstract

With evidence for a nanohertz gravitational-wave background now established by Pulsar Timing Arrays, the search focuses on identifying individual supermassive black hole binaries. We show that these binaries produce a distinct spatial correlation pattern across the array, acting as a deterministic analogue to the stochastic Hellings and Downs curve. We derive a closed analytic expression for this single-source overlap reduction function, $Υ_{ab}$, factorizing the signal into a source-dependent amplitude and a purely geometric fingerprint. Using simulated datasets, we demonstrate that this fingerprint breaks the degeneracy between an individual binary and a stochastic background. Including these cross-correlations yields Bayes factors of $ 144$ favoring the continuous-wave model over a stochastic-background model and $\sim 80$ favoring the continuous-wave model over an uncorrelated red-noise model. Furthermore, these new cross-correlations improve sky localization by a factor of $11\times$ over an uncorrelated search. Finally, while coherent matched filtering offers higher theoretical sensitivity, we argue that a cross-correlation-based search for individual binaries provides a robust alternative that hedges against the possibility of overfitting to noise fluctuations by focusing on the evidence for the correlations. The geometric fingerprints we present here rely on stable spatial correlations rather than phase coherence to identify the first nanohertz gravitational-wave sources.

Figures (6)

• Figure 1: Computational frame used to define the single source overlap reduction function $\Upsilon_{ab}(\hat{\bm{\Omega}},\iota,\psi)$. Pulsar $a$ lies on the $+\hat{\mathbf{z}}$ axis and pulsar $b$ lies in the $x$–$z$ plane at an angular separation $\zeta$. The GW propagation direction $\hat{\boldsymbol{\Omega}}$ is specified by polar angle $\theta$ and azimuthal angle $\phi$ with respect to $+\hat{\mathbf{z}}$. The direction to the GW source is $-\hat{\boldsymbol{\Omega}}$. This fixed geometry determines the antenna pattern functions $F_a^{A}$ and $F_b^{A}$ and hence the spatial correlation $\Upsilon_{ab}$ used throughout the paper. Shown also are the GW polarization basis vectors $\hat{m}$ and $\hat{n}$, which may be subject to a rotation by angle $\psi$.
• Figure 2: Single source correlation fingerprints across the sky. Each colored point on the left marks a GW propagation direction in the computational frame $(\theta,\phi)$, and the curve with the same color on the right shows the corresponding ORF for a single orientation-marginalized GW source, $\Upsilon_{ab}(\hat{\bm{\Omega}})$, Eq. \ref{['eq:upsilon-final']}, for a representative pulsar pair as a function of their angular separation $\zeta$. Pulsar $a$ lies on the $+\hat{\bm{z}}$ axis and pulsar $b$ lies in the $x$-$z$ plane. Unlike the HD curve, these responses are not universal: their shapes depend on source position because different pulsar pairs probe different lobes and nodes of the quadrupolar antenna pattern, and they are periodic, rather than symmetric, about $\pi$ radians due to the innate anisotropy of the single GW source. Particular binary orientations will further distort the cross correlation in a $\psi$ and $\iota$-dependent way, following Eq. \ref{['eq:upsilon-final-psi']}. Each SMBHB therefore imprints a distinct "fingerprint" on PTA correlations that encodes its sky position and polarization, which can be used to identify individual binaries and distinguish their signals from, e.g. the GWB.
• Figure 3: Single source correlation fingerprints (Eq. \ref{['eq:upsilon-final']}) corresponding to CWs from ten randomly drawn sky locations, marginalized over $\iota$ and $\psi$. For comparison, the unnormalized Hellings and Downs curve is shown in black.
• Figure 4: Sub-block of the CW covariance matrix $\bm{\Phi}$ [Eq. \ref{['eq:Phi_matrix_CW']}] for 4 different choices of ORF, $\Gamma_{ab}$. These 4 ORFs correspond to (a) an elliptically polarized GW from a SMBHB (BHB; the $\Upsilon_{ab}$ fingerprint derived here) [Eq. \ref{['eq:gamma_psi']}], (b) an unpolarized point GW source [Eq. \ref{['eq:spike-pixel']}], (c) a Hellings-Downs correlated model [Eq. \ref{['eq:HD-curve']}], and (d) an uncorrelated CW-like model [Eq. \ref{['eq:Gamma-auto']}]. The total matrix is $2N_{\rm psr} \times 2N_{\rm psr}$, with the CW represented by two Fourier coefficients in each pulsar; these sub-blocks correspond to the first 20 pulsars of our simulated dataset. Matrices are constructed using the injected CW parameters (Table \ref{['tab:simulation']}) and visually partitioned into $N_{\rm psr} \times N_{\rm psr}$ blocks to distinguish each set of interpulsar correlations.
• Figure 5: Posterior parameter distributions for cross-correlated CW models. Each model recovers the injected CW parameters (dashed black lines) in an idealized dataset where the injection uses a deterministic, evolving SMBHB template with ${\rm S/N}=30$ (Table \ref{['tab:simulation']}). The physically-accurate cross-correlated CW models, Spike-Pixel (green; Schult2025) and BHB (blue), yield more informative posteriors than the misspecified auto-correlated (yellow) and HD-correlated (pink; anholm2009) models. Nonetheless, both the auto-correlated and HD-correlated models do recover a signal at nearly the correct frequency and strain amplitude. The inset skymap on the upper right shows that the cross-correlations of the Spike-Pixel and BHB models allows more precise and accurate source localization versus the purely auto-correlated model. By including $\psi$-dependence in the cross-correlations, the BHB model returns the most informative posteriors, recovering the edge-on binary orientation $\cos\iota = 0$. The BHB model also returns the most accurate $h_0$ posterior by resolving the impact of inclination angle $\iota$ on the signal amplitude. Nonetheless, frequency recovery and sky localization of the BHB model are mildly preferred to the Spike-Pixel model.