Model-based cross-correlation search for gravitational waves from Scorpius X-1: separating intrinsic amplitude from unknown spin inclination, and bounding the coherence time under spin wandering

The 2015 methods paper of Whelan, Sundaresan, Zhang and Peiris established the model-based (semicoherent, cross-correlation) search for continuous gravitational waves from the low-mass X-ray binary Scorpius X-1 and listed the issues that had to be solved before the method could deliver detection-level science: the maximum usable coherence time is limited by poorly constrained spin wandering, the intrinsic amplitude is entangled with the unknown spin inclination, computing cost must be tamed, the orbital ephemeris must be kept accurate, and exact false-alarm tails must be computable. The pipeline-deployment, computing-cost, and ephemeris items have been substantially resolved: the search has been executed on O1, O3, and O4a data, reaching upper limits below the standard torque-balance prediction for 50-200 Hz. What survives as a precise open core is (i) the implementation and validation of the paper's proposed second cross-correlation statistic, built from the circular-polarization weight, which would disentangle the intrinsic amplitude from the unknown inclination - never carried out, with all executed searches instead marginalizing over or optimizing the orientation; (ii) quantitative limits on the coherence time grounded in a realistic spin-wandering treatment for Scorpius X-1, whose actual behavior remains empirically unmeasured; and (iii) the detection itself - the overarching empirical goal these items serve rather than a methodological deliverable - still pending through the fourth observing run.

Background

A rotating non-axisymmetric neutron star emits continuous gravitational waves (CWs), a nearly monochromatic strain signal at twice the spin frequency (or at a comparable frequency for current modes of the stellar fluid, known as r-modes). Low-mass X-ray binaries (LMXBs) - neutron stars accreting from a low-mass companion - are prime CW targets because accretion can sustain a non-axisymmetric deformation. Scorpius X-1 (Sco X-1), the brightest LMXB in X-rays, is the loudest predicted source of this class. Under the torque-balance hypothesis, the spin-down torque of gravitational emission balances the accretion spin-up torque estimated from the observed X-ray flux, predicting a strain $h_0\approx3.4\times10^{-26}(\nu_s/300\,\mathrm{Hz})^{-1/2}$ for Sco X-1; whether the star is actually driven to this equilibrium is an open astrophysical question that only a detection or sufficiently deep upper limits can settle. The CW signal is described by an intrinsic amplitude $h_0$ and orientation parameters: the inclination $\iota$ of the neutron-star spin axis to the line of sight and a polarization angle, which enter the plus- and cross-polarization amplitude factors $\mathcal{A}_+=\tfrac{1}{2}(1+\cos^2\iota)$ and $\mathcal{A}_\times=\cos\iota$. The unknown $\cos\iota$ is a nuisance parameter that searches must marginalize or optimize over. The data analysis challenge is that a year of data is needed for detectable strains, while the unknown neutron-star spin frequency and orbital Doppler modulation demand a huge template bank. In semicoherent methods, the data are split into short Fourier transforms (SFTs, short-time discrete Fourier transforms of contiguous data stretches, typically minutes long), and pairs of SFTs separated by less than a maximum lag $T_{\max}$ are coherently cross-correlated using a signal model; $T_{\max}$ trades sensitivity (detectable $h_0$ scaling roughly as $T_{\max}^{-1/4}$) against a computing cost growing approximately as $f_0^4T_{\max}^4$. Whelan, Sundaresan, Zhang and Peiris (2015) gave the definitive formulation of this model-based cross-correlation search for Sco X-1: its parameter-space metric, the effect of filtering with amplitude-parameter-averaged templates (the search statistic observes a combination $h_0^{\mathrm{eff}}=h_0\sqrt{\tfrac{5}{2}\cdot\tfrac{\mathcal{A}_+^2+\mathcal{A}_\times^2}{2}}$ rather than $h_0$ itself), spectral leakage, exact single-template false-alarm probabilities in the extreme tails via Gil-Pelaez inversion (numerical inversion of the statistic's characteristic function, needed because trials factors (the false-alarm penalty from searching of order $10^8$ templates) push the relevant false-alarm probabilities to $10^{-10}$, where Gaussian tail approximations fail), the sensitivity projections, and the obstacles to a real search. Three obstacles stand out. First, the neutron star in Sco X-1 shows no coherent pulsations, so its spin frequency is unknown and its frequency stochastically wanders (spin wandering) as accretion torques fluctuate; the 2015 paper estimated significant signal-to-noise loss for $T_{\max}\gtrsim12$ hr under a simple drift model, but the actual $f(t)$, $\dot f(t)$ behavior is unmeasured. Prior work has modelled accretion-induced spin wandering as a stochastic torque process and inferred the spin-wandering statistics of Sco X-1 from X-ray timing (Mukherjee, Messenger and Riles 2018, Ref. 7). Second, because the standard statistic weights cross-correlations with the orientation-averaged factor $\Gamma_{KL}^{\mathrm{ave}}$, the search measures $h_0^{\mathrm{eff}}$, not $h_0$; the paper proposed adding a second statistic that uses the circular-polarization weight $\mathrm{i}\Gamma_{KL}^{\mathrm{circ}}$ and therefore observes $h_0^2\mathcal{A}_+\mathcal{A}_\times$, from which the pair ($h_0^{\mathrm{eff}}$, $h_0^2\mathcal{A}_+\mathcal{A}_\times$) would disentangle $h_0$ and $\cos\iota$, remarking that this prospect bears further investigation. Third, the search needs an accurate orbital ephemeris, in particular the ascending-node passage time $t_{\mathrm{asc}}$ and orbital period $P_{\mathrm{orb}}$, whose uncertainty grows as the reference epoch is propagated over many orbital cycles.

Problem Statement

The model-based cross-correlation search for continuous gravitational waves from Scorpius X-1, as formulated by Whelan et al. (2015), has been carried through to real interferometer data and to sub-torque-balance upper limits, but the methodological problems that the 2015 paper itself identified as unresolved survive in a precise core. Solve the following, jointly serving the single objective of making the search capable of detection-level science with correctly inferred intrinsic parameters. First, investigate the second cross-correlation statistic proposed by the 2015 paper - the one built from the circular-polarization weight $\mathrm{i}\Gamma_{KL}^{\mathrm{circ}}$ in place of the orientation-averaged $\Gamma_{KL}^{\mathrm{ave}}$, which observes the amplitude combination $h_0^2\mathcal{A}_+\mathcal{A}_\times$ - and determine whether, and under what conditions, the pair of statistics ($h_0^{\mathrm{eff}}$, $h_0^2\mathcal{A}_+\mathcal{A}_\times$) allows the intrinsic amplitude $h_0$ and the unknown spin inclination $\cos\iota$ to be disentangled for a signal from Sco X-1; the paper itself described this prospect as bearing further investigation, and it has never been carried out. Second, establish limits on the maximum coherent lag $T_{\max}$ grounded in a realistic spin-wandering treatment for Sco X-1 (since the actual $f(t)$, $\dot f(t)$ remain unmeasured), going beyond the idealized constant-drift bounds currently in use; the 2015 paper itself cautioned that the drift-model assumptions behind its illustrative 12-hour bound are "uncertain and somewhat arbitrary", so that number "should also not be viewed as an exact constraint on the method". The overarching goal these serve is the detection of continuous gravitational waves from Scorpius X-1 - or astrophysically decisive constraints on it - with the intrinsic amplitude $h_0$ (and hence the neutron-star deformation) inferred rather than marginalized over the unknown orientation.

References

1. John T. Whelan, Santosh Sundaresan, Yuanhao Zhang, Prabath Peiris, Model-Based Cross-Correlation Search for Gravitational Waves from Scorpius X-1, Phys. Rev. D 91, 102005 (2015), DOI: 10.1103/PhysRevD.91.102005, arXiv:1504.05890, https://arxiv.org/abs/1504.05890
2. B. P. Abbott et al. (LIGO Scientific Collaboration and Virgo Collaboration), Upper Limits on Gravitational Waves from Scorpius X-1 from a Model-Based Cross-Correlation Search in Advanced LIGO Data, Astrophys. J. 847, 47 (2017), DOI: 10.3847/1538-4357/aa86f0, arXiv:1706.03119, https://arxiv.org/abs/1706.03119
3. R. Abbott et al. (LIGO Scientific, Virgo, and KAGRA Collaborations), Model-based cross-correlation search for gravitational waves from the low-mass X-ray binary Scorpius X-1 in LIGO O3 data, Astrophys. J. Lett. 941, L30 (2022), DOI: 10.3847/2041-8213/aca1b0, arXiv:2209.02863, https://arxiv.org/abs/2209.02863
4. A. G. Abac et al. (LIGO Scientific, Virgo, and KAGRA Collaborations), Sub-Torque-Balance Upper Limits on Continuous Gravitational Waves from Scorpius X-1 (2026), arXiv:2607.07765, https://arxiv.org/abs/2607.07765
5. G. D. Meadors, B. Krishnan, M. A. Papa, J. T. Whelan, Y. Zhang, Resampling to accelerate cross-correlation searches for continuous gravitational waves from binary systems, Phys. Rev. D 97, 044017 (2018), DOI: 10.1103/PhysRevD.97.044017, arXiv:1712.06515, https://arxiv.org/abs/1712.06515
6. K. J. Wagner, J. T. Whelan, J. K. Wofford, K. Wette, Template lattices for a cross-correlation search for gravitational waves from Scorpius X-1, Class. Quantum Grav. 39, 075013 (2022), DOI: 10.1088/1361-6382/ac5012, arXiv:2106.16142, https://arxiv.org/abs/2106.16142
7. A. Mukherjee, C. Messenger, K. Riles, Accretion-induced spin-wandering effects on the neutron star in Scorpius X-1: Implications for continuous gravitational wave searches, Phys. Rev. D 97, 043016 (2018), DOI: 10.1103/PhysRevD.97.043016, arXiv:1710.06185, https://arxiv.org/abs/1710.06185
8. T. L. Killestein, M. Mould, D. Steeghs, J. Casares, D. K. Galloway, J. T. Whelan, Precision Ephemerides for Gravitational-wave Searches - IV: Corrected and refined ephemeris for Scorpius X-1, Mon. Not. R. Astron. Soc. 520, 5317 (2023), DOI: 10.1093/mnras/stad366, arXiv:2302.00018, https://arxiv.org/abs/2302.00018
9. J. T. Whelan, R. Tenorio, J. K. Wofford, J. A. Clark, E. J. Daw, E. Goetz, D. Keitel, A. Neunzert, A. M. Sintes, K. J. Wagner, G. Woan, T. L. Killestein, Search for Gravitational Waves from Scorpius X-1 in LIGO O3 Data with Corrected Orbital Ephemeris, Astrophys. J. 949, 117 (2023), DOI: 10.3847/1538-4357/acc8d7, arXiv:2302.10338, https://arxiv.org/abs/2302.10338
10. J. B. Carlin, A. Melatos, How much spin wandering can continuous gravitational wave search algorithms handle?, Phys. Rev. D 111, 083016 (2025), DOI: 10.1103/PhysRevD.111.083016, arXiv:2504.08163, https://arxiv.org/abs/2504.08163
11. J. Tau, J. T. Whelan, Investigating Hardware Injections in LIGO O3 Data: Simulated Signals from a Neutron Star in a Low-Mass X-ray Binary, J. Phys. Conf. Ser. 3177, 012072 (2026), DOI: 10.1088/1742-6596/3177/1/012072, arXiv:2509.00907, https://arxiv.org/abs/2509.00907
12. K. Wette, Searches for continuous gravitational waves from neutron stars: A twenty-year retrospective, Astropart. Phys. 153, 102880 (2023), DOI: 10.1016/j.astropartphys.2023.102880, arXiv:2305.07106, https://arxiv.org/abs/2305.07106
13. J. K. Wofford, A Cross-Correlation Search for Continuous Gravitational Waves from Scorpius X-1 in Advanced LIGO O3 data, PhD thesis (2023), INSPIRE-HEP record 2770156, https://inspirehep.net/literature/2770156