betaFeedback

WorksPhysics and Astronomy

Eccentricity greatly enhances the detectability of line-of-sight acceleration in gravitational wave signals

Abstract

Compact object binaries formed dynamically are characterized by a tail of eccentric sources and by the potential presence of environmental perturbations. Here we show that the presence of eccentricity in a gravitational-wave (GW) signal strongly enhances the sensitivity to environmental effects, focusing in particular on line-of-sight acceleration (LSA). We demonstrate that this enhancement is recovered in Fisher-analysis forecasts and, crucially, persists in full Bayesian parameter estimation using state-of-the-art eccentric waveform models. The key insight is that the signal should be analyzed from the moment the first detectable eccentric harmonic enters the detector band, rather than from the customary entry of the n=2n=2 harmonic: this allows to probe the early inspiral and thus a longer signal, over which the GW dephasing introduced by LSA becomes substantially larger. We show that with this approach constraints on LSA can increase by more than an order of magnitude between eccentricities at 20~Hz e20=0e_{20}=0 and e20∼0.1e_{20}\sim0.1, for signals with sufficient signal-to-noise-ratio (SNR), while we predict a factor 55 amplification even for low-SNR ones. Fully exploiting this effect necessitates the continued improvement of efficient eccentric waveform models and likelihood evaluations, enabling the efficient analysis of signals lasting thousands of seconds and containing a large number of harmonics.

Introduction

The first candidate eccentric gravitational-wave (GW) signals have already been identified in the LIGO–Virgo–KAGRA (LVK) catalogs [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]. The current picture that binary systems of stellar-mass compact objects originate from a mixture of isolated and dynamical formation channels [13, 14, 15, 16, 17, 18] suggests that a fraction of detected signals must retain some amount of eccentricity in the detector band [2, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43]. It is therefore likely that more will emerge in future observing runs, in particular as detector sensitivities improve toward lower frequencies. Beyond their astrophysical significance, eccentric GW signals present some additional opportunities for parameter inference. The presence of eccentric harmonics extends the observable inspiral to earlier times and also provides multiple, independently measurable phase evolutions, improving constraints on both intrinsic and extrinsic binary parameters [44, 45, 46, 47, 48, 49]. Recently, it was shown that these same properties can substantially enhance the measurement of environmental signatures encoded in the GW phase [50, 51]. Collectively known as “environmental effects”, GW perturbations caused by acceleration, drag and tidal forces can provide direct information about the surroundings of the binary [52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90]. Line-of-sight acceleration (LSA) offers one of the cleanest examples [59, 78, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104], producing a characteristic negative post-Newtonian correction to the GW phase. Such accelerations can arise naturally in formation channels involving close encounters and prompt GW captures, as well as in active galactic nuclei disks [90, 105]. In the case of eccentric binaries, preliminary work has suggested that the resulting dephasing is significantly amplified by the higher-harmonic structure of the signal [50, 51], suggesting a substantial increase in detectability compared to circular binaries.

In this paper, we delineate the properties of eccentric signals that lead to an enhancement of the detectability of line-of-sight acceleration effects. We then show that this enhancement persists in both Fisher-matrix forecasts and, crucially, in full Bayesian parameter estimation (PE) using state-of-the-art eccentric waveform models and sources representative of the LVK population. We propose the following data-analysis strategy aimed at exploiting this enhancement in eccentric signals more effectively: The key requirement is to analyze the signal from the moment the first detectable harmonic enters the detector band, rather than the customary entry of the n=2n = 2 harmonic. In the time domain, this has the following simple interpretation. For an eccentric binary the GW frequency is not monotonic, but peaks at each periastron passage, where the instantaneous frequency exceeds the orbit-averaged one, typically used by waveform models to define the starting frequency. Therefore, the periastron frequency can reach the lower edge of the detector band much earlier than the orbit-averaged frequency for eccentric systems, allowing us to probe the early inspiral of the binary. This point, and its impact on PE, was also made in Refs. [48, 49]. Here we also show how not adopting the above requirement explains why recent studies jointly inferring eccentricity and line-of-sight acceleration have not recovered the expected enhancement [106, 107, 108].

The paper is organized as follows: in § 2 we explain the enhancement of LSA dephasing in eccentric signals, and delineate the analysis methods employed for both the full Bayesian inference and the Fisher matrix analysis. We present the results for the analysis of a set of injections, and for the real event GW200105_162426 (GW200105) in Sec. III, and we summarize our conclusions in Sec. IV. We also provide additional derivations and results for further tests in the appendix.

Methods

As shown in Refs. [50, 51, 89], the enhancement of dephasing signatures in eccentric binaries originates from a simple argument. For circular orbits, an environmental dephasing typically scales as a negative power of the GW frequency ff,

ΔΨcircn=2∝fα,α<0.(1)\Delta\Psi_{\mathrm{circ}}^{n=2} \propto f^{\alpha}, \qquad\alpha< 0 . \tag*{(1)}

In an eccentric binary, the nn-th eccentric GW harmonic observed at a frequency ff is emitted when the binary orbital frequency is F=f/nF=f/n. Consequently, the corresponding phase shift scales as:

ΔΨn(f)∝nΔΨcircn=2(f)n−α[tinsp(e,F)tinsp(e=0,F)]2,(2)\Delta\Psi_n(f) \propto n\Delta\Psi_{\mathrm{circ}}^{n=2}(f)n^{-\alpha}\left[\frac{t_{\mathrm{insp}}(e,F)}{t_{\mathrm{insp}}(e=0,F)}\right]^2 , \tag*{(2)}

where tinspt_{\mathrm{insp}} is the inspiral time at orbital frequency FF, and the additional factor of nn comes from the increased phase accumulation in higher harmonics.1 Physically, this result reflects the fact that higher harmonics probe progressively earlier stages of the inspiral, where environmental effects are intrinsically stronger. For LSA, α=−13/3\alpha=-13/3, therefore the contribution from the nn-th harmonic to the environmental signal-to-noise ratio, defined using the difference strain due to LSA, SNRLSA∝Δh≡hLSA−hnoLSA\mathrm{SNR}^{\mathrm{LSA}} \propto\Delta h \equiv h_{\mathrm{LSA}}-h_{\mathrm{noLSA}}, is:

SNRnLSA∼n16/3SNRn,(3)\mathrm{SNR}^{\mathrm{LSA}}_n \sim n^{16/3}\mathrm{SNR}_n , \tag*{(3)}

leading to an expected enhancement in SNRLSA_{\mathrm{LSA}} of:

SNRLSA∼nmax⁡16/3SNRnmax⁡,(4)\mathrm{SNR}^{\mathrm{LSA}} \sim n_{\max}^{16/3}\mathrm{SNR}_{n_{\max}} , \tag*{(4)}

where nmax⁡n_{\max} denotes the highest harmonic that contributes appreciable power to the observed signal. Crucially, the large power power of 16/316/3 means that even

including few additional eccentric harmonics will significantly enhance the detectability of the dephasing induced by LSA.

Phase shift from line-of-sight acceleration

We parametrize the LSA effects by Γ≡a∥/c\Gamma\equiv a_{\parallel}/c, i.e., the acceleration of the center-of-mass of the binary along the line of sight divided by the speed of light (in units of s−1^{-1}), taken to be positive when the acceleration points away from the observer. To first order in Γ\Gamma, the effect of the LSA can be included in frequency domain waveforms through the phase shift [e.g. 50, 78, 92, 95, 106]:

Δψj(f)=πΓtj(f)2,(5)\Delta\psi_j(f)=\pi\Gamma t_j(f)^2 , \tag*{(5)}

where tj(f)t_j(f) is the stationary time, measured from coalescence, at which the harmonic jj has frequency ff. The label j=(l,m,n)j=(l,m,n) denotes the spherical-harmonic multipole (l,m)(l,m) and the eccentric harmonic nn of the mean anomaly.

One important observation to make here is that, for a fixed GW frequency, the dependence of Eq. (5) on the harmonic is entirely encoded in the stationary time. Harmonics with large nn reach a given frequency at increasingly earlier times in the inspiral. Since the LSA-induced dephasing grows as t2t^2, it is much larger for modes with large nn. This is the origin of the enhanced sensitivity of eccentric harmonics to environmental effects. Additional corrections to the GW amplitude are suppressed by a factor of (tjf)−1(t_j f)^{-1} (see Appendix A), hence they are not included in this study.

Bayesian inference using pyEFPEHM

In the full Bayesian PE analysis, the parameters θ⃗\vec{\theta} of a system emitting GWs are inferred through Bayes theorem:

p(θ⃗∣d,H)=p(d∣θ⃗,H)p(θ⃗∣H)p(d∣H).(6)p(\vec{\theta}|d,\mathcal{H})=\frac{p(d|\vec{\theta},\mathcal{H})p(\vec{\theta}|\mathcal{H})}{p(d|\mathcal{H})}. \tag*{(6)}

In Eq. (6), p(θ⃗∣d,H)p(\vec{\theta}|d,\mathcal{H}) denotes the posterior probability of a set of parameters θ⃗\vec{\theta} given the data dd and a model (or hypothesis) H\mathcal{H} to describe the data, p(d∣θ⃗,H)p(d|\vec{\theta},\mathcal{H}) is the likelihood of the data, and p(θ⃗∣H)p(\vec{\theta}|\mathcal{H}) the prior probability of the parameters θ⃗\vec{\theta} given H\mathcal{H}, which encodes our a-priori knowledge about the system. In PE analyses, the evidence p(d∣H)p(d|\mathcal{H}) at the denominator serves just as a normalization constant.

To describe the data, we employ the waveform model pyEFPEHM [110], an inspiral-only approximant including eccentricity, precession, and higher-order modes. It extends pyEFPE [109] within the efficient fully precessing eccentric (EFPE) framework [111, 112, 113]. The orbital motion is described with the quasi-Keplerian parametrization and the spin precession with a multiple-scale analysis, so that only the precession-averaged radiation-reaction equations need to be integrated numerically. These combine eccentric post-Newtonian (PN) results (3PN for aligned spins) with quasi-circular corrections up to 4.5PN order and horizon absorption [114]. All (l,m)(l,m) multipoles entering the GW amplitude up to 1PN order are included, and each (l,m)(l,m) multiple is analytically decomposed into harmonics nn of the mean anomaly, with amplitudes valid at arbitrary eccentricity [115]. The frequency-domain waveform is computed harmonic by harmonic with the stationary phase approximation, using shifted uniform asymptotics for the precession-induced modulations [116]. The model thus provides the stationary time tj(f)t_j(f) of every harmonic j=(l,m,n)j=(l,m,n), measured from coalescence, which allows us to apply Eq. (5) mode by mode at a negligible additional cost.

For a few selected values of eccentricity, we perform PE with Bilby pipe [117, 118] and the nested sampler dynesty [119], using 2048 live points. To make the analysis computationally feasible, we speed-up the likelihood evaluation by using the multibanding approximation [120], which exploits the chirping nature of GW signals: the data strain is divided into frequency bands, and the likelihood is evaluated on custom frequency grids, determined by the specific time to merger of each band. We use multibanding with highest-mode 6 and an accuracy factor of 10. More details about the reliability of multibanding in the analysis of eccentric signals are provided in Appendix C.

Analysis setup

We simulate signals in a detector network comprising two Advanced LIGO in Hanford and Livingston [121], and Advanced Virgo [122]. To focus on the effects under investigation and avoid potential biases due to noise fluctuations, we run in a zero-noise configuration. We analyze the signals in the frequency range [20, 100] Hz, where the upper limit is determined by the fact that the waveform needs to be terminated before the innermost stable circular orbit (ISCO). As discussed above, the effects we are interested in are expected at low frequencies, therefore the upper cutoff does not risk discarding any information. Crucially, even if the analysis is started at 20 Hz (given the detector sensitivity limits), the waveform generation, both for the injection and during the sampling process, is initiated at the point where the frequency of the 22-mode n=2n=2 harmonic is finif_{\mathrm{ini}}. This ensures that harmonics up to nmax⁡=(40 Hz)/finin_{\max}=(40\ \mathrm{Hz})/f_{\mathrm{ini}} start below 20 Hz, meaning that all the relevant information

from these harmonics is included in the model. For both Fisher and PE analyses, we consider a non-spinning source with equal masses m1=m2=20M⊙m_1=m_2=20M_\odot. This choice of relatively large masses helps reduce the computational cost of the PE analysis, while still ensuring that the signal has enough cycles in band. For the PE injections, we place the source at a fixed random sky location, and at a luminosity distance of 100 Mpc, resulting in an SNR ∼110\sim110–120120 depending on the system’s properties. For all the tests described below, we set Γ=0\Gamma=0.

In Appendix B, we investigate the robustness of the analysis in case of different systems, such as precessing sources, as well as unequal-masses, and Γ\Gamma different from zero.

We employ priors uniform in component masses, with chirp mass sampled in the range [17.0, 18.0] M⊙M_\odot and mass ratio in [0.05, 1.0]. Eccentricity at 10 Hz and mean anomaly are sampled uniformly in [0.0, 0.4] and [0, 2π2\pi], respectively, while for Γ\Gamma we choose a uniform prior in [−0.01, 0.01]. The luminosity distance prior is uniform in source frame, in the range [10, 1000] Mpc, and for all the other parameters we use the standard uninformative priors implemented in BILBY and described in Ref. [118].

Fisher analysis

In order to be able to scan a large portion of the parameter space and understand the general trends in the measurement uncertainty of the LSA in an efficient way, we also study eccentric systems using Fisher analysis [123, 124]. For this approach we also use pyEFPEHM together with the same setup as described in § 2 C and Δψn(f)\Delta\psi_n(f) calculated according to Eq. (5).

The Fisher matrix is defined as:

Fϑϑ′=(∂ϑh∣∂ϑ′h),(7)\mathcal{F}_{\vartheta\vartheta'}=(\partial_{\vartheta}h|\partial_{\vartheta'}h), \tag*{(7)}

where ϑ\vartheta denotes the waveform parameters and (⋅∣⋅)(\cdot|\cdot) is the usual noise-weighted inner product. The derivatives with respect to the various parameters are calculated numerically, without any analytic shortcuts that might not hold with higher modes, eccentricity or precession. Each one is computed in three independent runs of central finite differences (one-sided at the boundary of the domain of validity), using symmetric steps and Richardson extrapolation up to fourth order. The initial step is chosen by a pilot search such that the change in the waveform is linear. The three runs are then combined through taking the median in each frequency bin. To ensure the waveform is smooth in the parameters, the adaptive harmonic selection of pyEFPEHM is frozen at the injection. This is especially important when calculating the derivatives at low eccentricities, where the adaptive harmonic selection can make these derivatives non-convergent.

We then calculate the measurement uncertainty of Γ\Gamma marginalized over the remaining parameters using the inverted Fisher matrix:

σΓ=(F−1)ΓΓ.(8)\sigma_{\Gamma} = \sqrt{(\mathcal{F}^{-1})_{\Gamma\Gamma}} . \tag*{(8)}

Results

We investigate how the constraints on Γ\Gamma change with the system eccentricity and with the starting frequency at which the waveform for the analysis is generated. We look at a few selected configurations with full PE, and then investigate the trend systematically with Fisher matrices.

Results from Bayesian inference

Varying initial waveform frequency

As discussed in § 2, the phase-shift contribution of LSA scales with a large negative power of frequency, hence most information about this effect is imprinted in the early evolution of the system. In eccentric systems, we can access this early stage thanks to higher eccentric harmonics that enter the band earlier than the dominant quasi-circular mode. In Bayesian inference, at each sampling step the data is compared to the GW signal that would be generated by the sampled parameters with the chosen waveform model. The initial frequency finif_{\mathrm{ini}} at which we start generating the waveform plays a crucial role: the lower the frequency, the more information from higher harmonics, produced at earlier times, appears in band. Therefore, a sufficiently low finif_{\mathrm{ini}} is required to fully exploit the enhancement. As an example, if fini=20f_{\mathrm{ini}} = 20 Hz, then only the full information from the n=1,2n = 1, 2 harmonics is recovered in the detector frequency band. Only a fraction of the n≥3n \ge3 would be included in band, since (i) their contribution to the signal would start at a GW frequency fGW=30,40,…f_{\mathrm{GW}} = 30, 40, \ldots Hz, while (ii) the early evolution of the binary where these harmonics would have a much stronger contribution is completely ignored. Instead, with fini=2f_{\mathrm{ini}} = 2 Hz, all the measurable information from up to the nmax⁡=20n_{\max} = 20 harmonics is fully included in the waveform.

Figure 1 shows the posterior probability distributions resulting from our Bayesian inference on signals simulated as described in § 2 C, with eccentricity at 20 Hz set to e20=0.1e_{20} = 0.1. The figure compares results from analyses with different waveform finif_{\mathrm{ini}}, showing how the constraints on eccentricity and especially on Γ\Gamma become tighter with lower finif_{\mathrm{ini}}. Notably, the constraint remains unchanged when finif_{\mathrm{ini}} goes from 4 Hz to 2 Hz: for the chosen eccentricity value e20=0.1e_{20} = 0.1, a waveform generated from 4 Hz already contains all the relevant eccentric harmonics information. All further harmonics would have SNRn≪1\mathrm{SNR}_{n} \ll1, therefore lowering finif_{\mathrm{ini}} yields no significant improvement (see discussion in Sec. III B). Larger values of eccentricity will require lower values of finif_{\mathrm{ini}} to encode all the harmonic content, and consequently the information we can obtain from them also saturates at lower finif_{\mathrm{ini}} values, as shown later in Figure 4.

Posterior probability densities for chirp mass, mass ratio, effective inspiral spin, eccentricity, and line-of-sight acceleration

Figure 1. Posterior probability densities for chirp mass Mc\mathcal{M}_c, mass ratio qq, effective inspiral spin χeff\chi_{\mathrm{eff}}, eccentricity value at 20 Hz e20e_{20}, and line-of-sight acceleration parameter Γ\Gamma for analyses where the model is generated starting form different initial frequencies finif_{\mathrm{ini}}. The inset zooms on the posterior distribution of Γ\Gamma. The black vertical lines indicate the injected values.

Figure 4. The measurement error of Γ\Gamma as a function of the eccentricity for different values of finif_{\mathrm{ini}}. Results from the Fisher analysis without the SNR cut are depicted by the dashed lines, while the ones including the SNR cut are shown by the solid lines. Results from full PE for selected parameters are shown by markers. The zero-eccentricity PE results for fini=20f_{\mathrm{ini}}=20 Hz and 4 Hz are denoted by arrows.

Varying eccentricity

Figure 2 shows the posteriors obtained for analyses of systems with different eccentricities, while everything else is kept fixed as described in § 2 C. Here we use fini=4f_{\mathrm{ini}} = 4 Hz for all the analyses. As expected, more eccentric systems lead to overall tighter constraints also on Γ\Gamma. However, this trend inverts after e=0.125e = 0.125. For increasing eccentricity, more and more higher harmonics start contributing to the signal, which then carry information about the early evolution of the binary, providing more stringent constraints on Γ\Gamma. However, if the value of finif_{\mathrm{ini}} is fixed, the bound on the length of the signal cuts off information carried by higher harmonics from the very early inspiral, and hence constraints on Γ\Gamma saturate. At very high eccentricities, the inspiral time and thus the signal duration start decreasing, naturally giving worse constraints. Overall, the two competing effects from a richer harmonic content and a shorter overall duration generate an optimal eccentricity for which Γ\Gamma can be measured with the smallest uncertainty.

Posterior probability densities for chirp mass, mass ratio, effective inspiral spin, eccentricity, and line-of-sight acceleration

Figure 2. Posterior probability densities for chirp mass Mc\mathcal{M}_c, mass ratio qq, effective inspiral spin χeff\chi_{\mathrm{eff}}, eccentricity value at 20 Hz e20e_{20}, and line-of-sight acceleration parameter Γ\Gamma for signals simulated from systems with different eccentricity values. The inset zooms on the posterior distribution of Γ\Gamma. The black vertical lines indicate the injected values.

Note that the measurement accuracy of Γ\Gamma is improved by slightly more than an order of magnitude between systems with e20=0e_{20} = 0 and e20=0.125e_{20} = 0.125, which can have profound implications on the observability of LSA and of environmental effects in general in mildly-eccentric systems.

Results from Fisher analysis

Fisher information by harmonic

In order to first gain insight about the contribution of various harmonics to the enhancement in measurement precision of Γ\Gamma, it is helpful to look at the contribution to the Fisher-matrix element FΓΓ\mathcal{F}_{\Gamma\Gamma} from the different harmonics, at various eccentricities. This matrix element still misses information about parameter correlations, but can point us to the main factors contributing to the enhancement. Figure 3 shows, for fini=2f_{\mathrm{ini}} = 2 Hz, the ratio FΓΓn/FΓΓ\mathcal{F}^{n}_{\Gamma\Gamma}/\mathcal{F}_{\Gamma\Gamma}, where FΓΓ\mathcal{F}_{\Gamma\Gamma} is computed from the full waveform and

Relative contribution to the Fisher matrix element $\mathcal{F}_{\Gamma\Gamma}$ from different harmonics as a function of orbital eccentricity $e_{20}$

Figure 3. Relative contribution to the Fisher matrix element FΓΓ\mathcal{F}_{\Gamma\Gamma} from different harmonics, as a function of orbital eccentricity e20e_{20}. The solid line denotes the harmonic with the dominant contribution at each eccentricity value. The black dashed line show the boundary above which harmonics are neglected after applying the SNR cut for harmonics with (SNRn)2<0.5(\mathrm{SNR}^{n})^{2}<0.5, while the red dashed line corresponds to the dominant harmonics after this cut. The waveforms were generated with fini=2f_{\mathrm{ini}}=2 Hz.

FΓΓn=∑l,m(∂Γhlm|∂Γhlm),(9)\mathcal{F}^{n}_{\Gamma\Gamma} = \sum_{l,m} \left(\partial_{\Gamma}h_{lm} \middle| \partial_{\Gamma}h_{lm}\right), \tag*{(9)}

summing over all the (l,m)(l,m) multipoles with the same nn. The cross terms between harmonics are thus neglected in FΓΓn\mathcal{F}^{n}_{\Gamma\Gamma}, and ∑nFΓΓn\sum_n \mathcal{F}^{n}_{\Gamma\Gamma} only approximates FΓΓ\mathcal{F}_{\Gamma\Gamma}. Nevertheless, these cross terms are small, since harmonics with different nn reach a given frequency at different times in the inspiral, and for harmonics with the same nn but different mm, periastron advance makes their relative phase wind many times across the band, so in both cases they average out. Only multipoles with the same (m,n)(m,n) add coherently, but the higher-ll ones are much weaker.

The solid red line shows the harmonic with the dominant contribution to the Fisher-matrix element. For e20≲10−2e_{20} \lesssim10^{-2}, even though contributing to roughly half of the Fisher-matrix element, most of the information comes not from the n=2n=2 harmonic, as one would naively expect, but from an n=4n=4 harmonic. This is because even though in our equal-mass system higher-order modes are suppressed, the fact that the dominant harmonic from the (l,m)=(4,4)(l,m)=(4,4) mode carries information from an earlier epoch, greatly amplifies its importance. The n=4n=4 harmonic of the (2,2)(2,2) mode, instead, scales as e2e^2 and is negligible at these eccentricities. For eccentricities as low as e20∼10−2e_{20} \sim10^{-2} a few higher harmonics start giving relevant contribution. At eccentricity e20∼4×10−2e_{20} \sim4 \times10^{-2}, when the eccentricity is high enough such that in the early phases of the evolution a significant number of higher harmonics would appear with a substantial contribution, the dominant harmonic suddenly jumps up to the maximum harmonic included in the waveform, then climbs up to reach the n=20n=20 harmonic, only going further up for e20≳0.1e_{20} \gtrsim0.1. This saturation behaviour at n=20n=20 is exactly due to the somewhat artificial choice of initializing the waveform at fini=2f_{\mathrm{ini}}=2 Hz, which means that the highest harmonic that has its full observable contribution in the waveform is nmax⁡=2×(20 Hz/2 Hz)=20n_{\max}=2 \times(20\ \mathrm{Hz}/2\ \mathrm{Hz})=20. Higher harmonics might potentially have more information content but this information is stored in a part of the waveform corresponding to fini<2f_{\mathrm{ini}} < 2 Hz.

The sharp jump of the dominant harmonic at e20∼4×10−2e_{20} \sim4 \times10^{-2}, which would reach even higher harmonics provided that they were included in the waveform is due to a limitation of the Fisher-matrix approach. Since the Fisher matrix approximates waveform differences from local derivatives, it does not take into account the saturation effect, i.e., the fact that the difference in strain for phase shifts Δψ≳1\Delta\psi\gtrsim1 do not grow linearly but are bounded form above. Due to this, the contribution of harmonics with a low SNR but substantial phase shift at early times are amplified beyond their real effect on the measurement uncertainty.

We solve this problem heuristically by dropping harmonics with an SNR below a threshold (SNRcutn)2=0.5(\mathrm{SNR}_{\mathrm{cut}}^n)^2 = 0.5. This choice of value reflects the fact that the contribution of a harmonic with (SNRcutn)2≤0.5(\mathrm{SNR}_{\mathrm{cut}}^n)^2 \leq0.5 to the log-likelihood is ≤1\leq1, even with a maximal dephasing. Thus, these harmonics, even with a maximal dephasing, cannot exclude a 1σ1\sigma change of Γ\Gamma. Note that such a sharp cut in harmonics with low SNR can create discontinuous features in the measurement uncertainties of Γ\Gamma, since the contributions of harmonics with SNR growing above this threshold would change from zero to a finite, non-negligible value. In a full PE setup, however, the contribution from harmonics below this threshold would not be neglected completely but would be bounded, leading to a smooth dependency on the eccentricity. In addition, we emphasize that the exact choice of the SNR threshold remains heuristic. Therefore, developing more robust approaches to handling low-SNR harmonics in Fisher analyses remains an important direction for the future.

The black dashed line in Figure 3 shows the boundary above which harmonics are dropped due to the SNR cut discussed above. This suppresses unphysical contributions from higher harmonics between e20∼0.04−0.1e_{20}\sim0.04-0.1, leading to a smooth increase of the dominant harmonic in this range (shown by the red dashed line). The saturation of the dominant harmonic above this range indicates that even tighter LSA constraints could be gained at high eccentricities by choosing an even lower value of finif_{\mathrm{ini}}.

Forecast on marginalized uncertainties

The measurement error of Γ\Gamma marginalized over the other parameters, σΓ\sigma_{\Gamma}, is calculated according to Eq. (8). Figure 4 shows this error as a function of eccentricity e20e_{20} for different values of finif_{\mathrm{ini}} shown by lines with different colors, while markers represent the errors obtained from full PE. Dashed lines correspond to Fisher results without the SNR cut, while the solid lines represent results with this cut included.

The two types of trends discussed before in varying eccentricity and finif_{\mathrm{ini}} are both visible in Figure 4, i.e. that (i) choosing a lower value for finif_{\mathrm{ini}} provides more in-band information from the early inspiral carried by higher harmonics, thus giving tighter constraints, which for low eccentricities already saturates for fini∼4−10f_{\mathrm{ini}}\sim4-10 Hz; and that (ii) choosing a sufficiently low value for finif_{\mathrm{ini}} and increasing the eccentricity enhances constraints on Γ\Gamma substantially, which then start degrading at high values of eccentricity. Note that there is some improvement in σΓ\sigma_{\Gamma}

even at zero eccentricity due to the higher-order spherical modes accounted for in pyEFPEHM.

The Fisher analysis, being an approximation, naturally underestimates the measurement uncertainties, however, the general trends are well-recovered. The main differences between the PE and the Fisher results are the difference in improvement at e20∼0e_{20}\sim0, and at eccentricities in the range e20∼0.04−0.1e_{20}\sim0.04-0.1 between full PE and Fisher results without the SNR cut. The second is due to the effect of the Fisher analysis overestimating the contribution from low-SNR harmonics, and is in fact corrected once the SNR cut is introduced. The first is also due to a similar reason: the SNR of the (l,m)=(4,4)(l,m)=(4,4) mode contribution does not fall below the threshold and thus its contribution, which would be reduced in a full PE setup, is fully accounted for in the Fisher analysis. Note that this large difference between PE and Fisher is especially true for lower eccentricities, where including or not including a harmonic in addition to the other 1−21-2 harmonics can make a huge difference in the results, while this effect at higher eccentricities with ∼10\sim10 harmonics already included is more subtle. Notably, σΓ\sigma_{\Gamma} matches the PE results quite well at e20≳0.1e_{20}\gtrsim0.1 and fini≤4f_{\mathrm{ini}}\leq4 Hz, and for these finif_{\mathrm{ini}} values we also recover the location of the optimal eccentricity relatively accurately. As discussed previously, while results from the Fisher analysis without the SNR cut show a smooth dependency on the eccentricity, the introducing the sharp SNR cut creates discontinuous breaks in this trend.

The increase of σΓ\sigma_{\Gamma} at high eccentricities is recovered in the Fisher results as well and is due to the reasons discussed in Sec. III A 2, i.e. that such high eccentricities substantially reduce the inspiral time of the binary, thus reducing the phase shift as well.

As mentioned before, the errors for Γ\Gamma shown here were marginalized in order to take into account correlations with other parameters, as they can increase the value of σΓ\sigma_{\Gamma} by more than an order of magnitude; in particular, we find that the effect of these correlations is the strongest for e20≳0.1e_{20} \gtrsim0.1.

Comparison with recent results

Multiple recent studies investigated constraints on LSA using eccentric waveforms [106, 107, 108, 125, 126]. Notably, Refs. [106, 107, 108] found that the value of Γ\Gamma is recovered with a larger uncertainty for binaries with eccentricity e20∼0.2−0.4e_{20} \sim0.2 - 0.4 compared to the circular case. In contrast, we find that introducing eccentricity to the binary system can greatly enhance the measurability of Γ\Gamma. For fini=20f_{\mathrm{ini}} = 20 Hz, however, even though σΓ\sigma_{\Gamma} slightly decreases between e20=0e_{20} = 0 and e20=0.1e_{20} = 0.1, it increases significantly above e20∼0.1e_{20} \sim0.1. While for eccentricities e20≳0.5e_{20} \gtrsim0.5 we also predict the accuracy of LSA-recovery to degrade for all values of finif_{\mathrm{ini}}, the worse recovery of Γ\Gamma for eccentric binaries reported in Refs. [106, 107] is almost certainly linked to the reason discussed before, i.e. to initializing the waveform at a higher reference frequency and thus ignoring the early evolution of the binary, where most of the information from the LSA carried by higher harmonics would come from. In fact, it is directly implied in Ref. [107] that all waveforms were initialized at fini=20f_{\mathrm{ini}} = 20 Hz, while a fixed signal duration of 256 s is mentioned in Ref. [108], once again implying signal generation from 20 Hz for a GW170817-like system.

Dependence on signal-to-noise ratio

Since the effect of the harmonic cut depends on the SNR of the signal, we study its effect on signals with different total SNR. The results of this analysis are shown in Figure 5, where the relative change in σΓ\sigma_{\Gamma} with respect to the zero-eccentricity case is shown for signals with different total SNRs, marked by the different line styles. Note that regarding these relative changes, reducing the total SNR of the signal by a given factor is equivalent to increasing the SNR cut threshold by the same factor. As expected, the amplification in the measurability of Γ\Gamma for eccentric binaries is reduced when the total SNR is 3 or 10 times less, however, we can still achieve a maximal improvement of a factor of ∼20\sim20 or ∼5\sim5, respectively, by generating the signal with an appropriately low value of finif_{\mathrm{ini}}. On the other hand, with a total SNR 3 times larger than the system we considered in our previous analyses, it is possible to reduce σΓ\sigma_{\Gamma} by a factor of ∼100\sim100 by increasing the eccentricity to e20∼0.1e_{20} \sim0.1 and choosing fini=1f_{\mathrm{ini}} = 1 Hz. This means that increasing the SNR has two additive effect on the measurability of LSA in eccentric systems: it both reduces σΓ\sigma_{\Gamma} at all eccentricities due to the higher signal power, while it also amplifies the improvement at higher eccentricities by making additional higher harmonics become relevant, thus providing information from even earlier times of the inspiral.

Figure 5. The relative change in the measurement error of Γ\Gamma, σΓ\sigma_{\Gamma}, with respect to the zero-eccentricity value with fini=20f_{\mathrm{ini}} = 20 Hz, as a function of the eccentricity, including the SNR cut, dropping the contribution of harmonics with (SNRℓm)2<0.5(\mathrm{SNR}^{\ell m})^2 < 0.5. Different values of finif_{\mathrm{ini}} are shown by different colors, while different line styles mark GW signals with different total SNR.

Finally, we note that with a total SNR reduced by a factor of 10, the (l,m)=(4,4)(l,m) = (4,4) mode is completely discarded from the Fisher analysis at low eccentricities, creating a similar trend for fini=4f_{\mathrm{ini}} = 4 Hz to what is seen in Figure 4 for the Fisher results, verifying our previous conclusion.

Analysis of GW200105

We analyze the real event GW200105 [127]2 for which previous works showed evidence of eccentricity (see, e.g., Ref. [7, 8, 9, 11, 12, 130]), to investigate whether we can infer constraints on LSA. The publicly available cleaned data strain lasts 64 s, therefore we can safely analyze the signal with waveforms generated from 16 Hz or larger. Considering lower values of finif_{\mathrm{ini}} would require analyzing longer segments of data, which are currently not cleaned. For the analysis we employ the waveform model pyEFPEHM and the multibanding approach for the likelihood evaluation, as for the injection study described above.

The top panel of Figure 6 shows the posteriors inferred for the cosine of inclination cos⁡(θJN)\cos(\theta_{JN}), the eccentricity e20e_{20} and the LSA parameter Γ\Gamma, comparing the case in which the model is generated starting from 20 or 16 Hz. Results are consistent with the ones in Ref. [107] and Ref. [106], where different waveform approximants were employed in the analysis. Also in our case, the Γ\Gamma posterior peaks at non-zero values: this is due to the strong degeneracy with eccentricity, whose posterior is considerably wider than the one recovered from analyses that do not take into account LSA effects.

Figure 6. Top: Posterior probability distributions for the inclination θJN\theta_{JN}, eccentricity e20e_{20} and the LSA parameter Γ\Gamma recovered from the analysis of GW200105 with the model generated at fini=20 Hzf_{\mathrm{ini}} = 20\,\mathrm{Hz} and fini=16 Hzf_{\mathrm{ini}} = 16\,\mathrm{Hz}. Bottom: Relative change in σΓ\sigma_{\Gamma} with respect to the value at e20=5×10−3e_{20} = 5 \times10^{-3} and fini=20 Hzf_{\mathrm{ini}} = 20\,\mathrm{Hz}, as a function of the eccentricity, using the Fisher-method for GW200105-like systems, with different values of finif_{\mathrm{ini}}.

The analysis from 16 Hz does not significantly tighten the constraints – yielding a relative improvement of roughly 10% –, which is expected: changing finif_{\mathrm{ini}} to 16 Hz we do not even extract the full information from the n=3n = 3 harmonics. The broad eccentricity posterior, peaking at an eccentricity as low as e20∼5×10−2e_{20} \sim5 \times10^{-2}, hinders the possibility to get a significant increase in the LSA constraint. Looking at Figure 5 as a general reference, the potential improvement at such low eccentricities, which is already quite small for a system with such low SNR as GW200105 – corresponding roughly to the dashed SNR/10 lines in the figure –, would derive mostly from the GW higher-order modes contributions. However, as shown in the posteriors in the top panel of Figure 6, the source emitting GW200105 was most likely a face-on system, for which a large part of higher-order modes contributions is suppressed.

In order to confirm this hypothesis, we calculated the uncertainties for Γ\Gamma for a GW200105-like system using our Fisher-framework. The source masses in this case were set to m1=8.9 M⊙m_1 = 8.9\,M_{\odot} and m2=1.9 M⊙m_2 = 1.9\,M_{\odot}, with dL=280 Mpcd_L = 280\,\mathrm{Mpc} and θJN=0\theta_{JN} = 0 set as the luminosity distance and inclination. The bottom panel of Figure 6 shows the relative change in σΓ\sigma_{\Gamma} as a function of eccentricity for such systems for different values of finif_{\mathrm{ini}}. This predicts that using fini=16 Hzf_{\mathrm{ini}} = 16\,\mathrm{Hz} instead of 20 Hz shrinks σΓ\sigma_{\Gamma} by ∼20%\sim20\% at e20∼5×10−3e_{20} \sim5 \times10^{-3}, while the maximal gain around e20∼0.1e_{20} \sim0.1 is ∼40%\sim40\%. Considering that the Fisher-method with the previously introduced SNR cut overestimates this improvement for low eccentricities (as shown in Figure 4), we find the results from the full Bayesian analysis of GW200105 to agree with the Fisher predictions reasonably well. This Fisher analysis also predicts a further marginal improvement when using fini=10 Hzf_{\mathrm{ini}} = 10\,\mathrm{Hz}, and, based on the results shown in Figure 5, we expect no further improvement beyond this for fini=4 Hzf_{\mathrm{ini}} = 4\,\mathrm{Hz}; however, analyzing the signal from these lower initial frequencies values would require using ∼200 s\sim200\,\mathrm{s} or ∼2000 s\sim2000\,\mathrm{s} of data, respectively.

In general, reanalyzing events in search for observable LSA from lower values of finif_{\mathrm{ini}} would require cleaning longer strains and re-evaluating the PSDs. A more systematic reanalysis of real events with waveform models generated from such lower frequencies is therefore left for future work.

Discussion

Using a state-of-the-art eccentric GW waveform to investigate detectability of LSA in eccentric sources through full Bayesian inference for selected cases, as well as more broadly through Fisher analysis, in this paper we have shown that:

  1. The detectability of the LSA-induced phase shift can be significantly enhanced for moderately eccentric binaries with e20∼0.1e_{20} \sim0.1, and this enhancement – more than an order of magnitude for the systems we investigated – persists in the full PE setup. This effect is due to higher harmonics entering the band at a much earlier time than the n=2n = 2 harmonic.

  2. In order to fully capture the information from higher eccentric harmonics, one has to initialize the waveform when the first detectable eccentric harmonic enters the detector band, while the frequency of the n=2n = 2 harmonic might still be below the sensitive band of GW detectors. This equivalently means

that a significantly longer time segment of the waveform should be analyzed. We have shown how ignoring this condition can lead to worse constraints on Γ\Gamma, similar to the results of previous studies [106, 107, 108].

(iii) One can expect an improvement in LSA constraints up to a factor ∼5\sim5 or more for systems like the ones investigated here with an SNR of ∼10\sim10, while this increase can be as large as a factor of ∼50\sim50 for systems with an SNR of ∼100\sim100.

These results have several important implications. Firstly, note that an increase in SNR doubly strengthens the measurability of LSA in eccentric systems. In addition to simply increasing the precision at all eccentricities, information from further observable harmonics also becomes accessible, amplifying the gain at non-zero eccentricities. Thus an order of magnitude increase in SNR for a system with e20∼0.1e_{20} \sim0.1 can increase LSA measurement precision by two orders of magnitude instead of just one. This suggests very promising perspectives for LSA measurement with third-generation detectors, like Einstein Telescope (ET) [131]. For such high-SNR signals, one might have to include even further harmonics in eccentric waveform models to fully exploit all the available information from the early evolution of the binary.

Secondly, such a strong amplification in LSA measurability is achieved despite strong parameter correlations at eccentricities e20≳0.1e_{20} \gtrsim0.1. In case these correlations can be broken, e.g. through the inclusion of higher-order LSA effects in the waveform, a further increase of a factor of few could be achieved.

Thirdly, the measurability of LSA has a very direct correspondence with the length of the analyzed signal containing measurable information, due to the phase shift given by Eq. (5). This means that if one was to measure LSA of the order 10−8 s−110^{-8}\,\mathrm{s}^{-1}, corresponding to a typical acceleration in a nuclear cluster, then the analysis of a waveform with ∼1000\sim1000 seconds would become inevitable. This necessitates the development of waveform generation methods even more efficient than the ones currently available for eccentric binaries, as well as fast and accurate likelihood-estimation techniques.

Finally, LSA is only one example of a broad class of astrophysical environmental effects that imprint characteristic dephasing on the GW signals. Since most environmental effects are strongest during the early inspiral, and therefore exhibit steep negative frequency scalings, they should similarly benefit from the harmonic enhancement associated with eccentric binaries. This may substantially improve constraints on the properties of the source environment, including surrounding gas, external tidal fields, or nearby massive companions. More broadly, our results demonstrate that eccentricity and environmental effects should be analyzed jointly. Once eccentricity has been established in a candidate signal, it should be re-analyzed with waveform models including environmental effects, as the harmonic structure can enhance their detectability by orders of magnitude.

Data availability

The data from the analyses in this work are available in Zenodo at 10.5281/zenodo.24087524 [132].

Acknowledgments

We thank Lorenzo Pompili for his helpful comments and suggestions as internal LIGO reviewer. J.T. acknowledges support from the Alexander von Humboldt Foundation under the project no. 1240213 - HFST-P. A.P. acknowledges funding from the EU Horizon under ERC Starting Grant, no. SMArt-101076369. Views and opinions expressed are those of the authors only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them. G.M.’s research is supported in part by the European Research Council (ERC) Horizon Synergy Grant “Making Sense of the Unexpected in the Gravitational-Wave Sky” grant agreement no. GWSky–101167314. L.Z. is supported by the European Union’s Horizon 2024 research and innovation program under the Marie Skłodowska-Curie grant agreement No. 101208914.

This research has made use of data or software obtained from the Gravitational Wave Open Science Center (gwosc.org), a service of the LIGO Scientific Collaboration, the Virgo Collaboration, and KAGRA. This material is based upon work supported by NSF’s LIGO Laboratory which is a major facility fully funded by the National Science Foundation, as well as the Science and Technology Facilities Council (STFC) of the United Kingdom, the Max-Planck-Society (MPS), and the State of Niedersachsen/Germany for support of the construction of Advanced LIGO and construction and operation of the GEO600 detector. Additional support for Advanced LIGO was provided by the Australian Research Council. Virgo is funded, through the European Gravitational Observatory (EGO), by the French Centre National de Recherche Scientifique (CNRS), the Italian Istituto Nazionale di Fisica Nucleare (INFN) and Dutch Nikhef, with contributions by institutions from Belgium, Germany, Greece, Hungary, Ireland, Japan, Monaco, Poland, Portugal, Spain. KAGRA is supported by Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan Society for the Promotion of Science (JSPS) in Japan; National Research Foundation (NRF) and Ministry of Science and ICT (MSIT) in Korea; Academia Sinica (AS) and National Science and Technology Council (NSTC) in Taiwan.

Derivation of line-of-sight acceleration phase shift

Perturbations in the waveform due to LSA are purely an observational effect, originating from the delay between the GW being emitted in the source frame and observed in the detector frame. The relation between the source and detector time and frequency, which correspond to the same GW phase ϕsrc(tsrc)=ϕdet(tdet)\phi_{\mathrm{src}}(t_{\mathrm{src}})=\phi_{\mathrm{det}}(t_{\mathrm{det}}), reads:

fdet=fsrc1+Γtsrc(10)f_{\mathrm{det}}=\frac{f_{\mathrm{src}}}{1+\Gamma t_{\mathrm{src}}} \tag*{(10)}
tdet=tsrc+Γ2tsrc2(11)t_{\mathrm{det}}=t_{\mathrm{src}}+\frac{\Gamma}{2}t_{\mathrm{src}}^{2} \tag*{(11)}

where the times are synchronized such that tsrc=0t_{\mathrm{src}}=0 corresponds to tdet=0t_{\mathrm{det}}=0, and we dropped the term corresponding to an initial non-zero velocity, since it can be absorbed through a shift in the chirp mass. The time domain phase shift due to acceleration becomes:

Δϕ(t)=ϕdet(t)−ϕsrc(t)≈ϕdet(tdet)+dϕdetdtΔt−ϕsrc(t)≈−πfΓt2,(12)\begin{aligned} \Delta\phi(t)&=\phi_{\mathrm{det}}(t)-\phi_{\mathrm{src}}(t)\\ &\approx\phi_{\mathrm{det}}(t_{\mathrm{det}})+\frac{\mathrm{d}\phi_{\mathrm{det}}}{\mathrm{d}t}\Delta t-\phi_{\mathrm{src}}(t)\\ &\approx-\pi f\Gamma t^{2}, \tag*{(12)} \end{aligned}

where t≡tsrct\equiv t_{\mathrm{src}} and f≡fsrcf\equiv f_{\mathrm{src}}. Then, using the first-order accurate relation Δψ(f)=−Δϕ(t(f))\Delta\psi(f)=-\Delta\phi(t(f)) [50], we get

Δψ(f)=πΓft(f)2.(13)\Delta\psi(f)=\pi\Gamma f t(f)^{2}. \tag*{(13)}

Note that the above relation, although leading order in Γ\Gamma, is accurate up to arbitrary PN order, provided the appropriate PN-accurate formula is used for t(f)t(f). This insight can substantially simplify otherwise lengthy derivations of Δψ\Delta\psi for higher PN orders (c.f. equations in Refs. [78, 101]).

One can easily see that the above argument generalizes to waveforms with a complex structure of eccentric harmonics and higher-order modes, leading us to the following formula:

Δψj(f)=πΓftj(f)2.(14)\Delta\psi_j(f)=\pi\Gamma f t_j(f)^{2}. \tag*{(14)}

where jj represents the indices (l,m,n)(l,m,n) of the spherical and eccentric harmonic decompositions.

Unlike some previous studies [e.g. 106], here we neglect amplitude corrections. Due to the scope of the present paper, here we do not deliver a detailed derivation but provide a basis for our choice based on the general structure of the correction. After expanding to first order in Γ\Gamma, the frequency-domain amplitude of the waveform with LSA takes the following form:

ALSA(f)≈Avac(f)[1+αΓt(f)+βΓff˙],(15)A_{\mathrm{LSA}}(f)\approx A_{\mathrm{vac}}(f)\left[1+\alpha\Gamma t(f)+\beta\Gamma\frac{f}{\dot{f}}\right], \tag*{(15)}

where AvacA_{\mathrm{vac}} is the “vacuum” amplitude without LSA, and α,β\alpha,\beta are scalar factors of O(1)O(1). At leading order f/f˙=−8/3 t(f)f/\dot{f}=-8/3\,t(f). Thus, the amplitude correction is proportional to Γt(f)\Gamma t(f), which is suppressed by a factor of [t(f)f]−1[t(f)f]^{-1} compared to the first-order accurate phase correction, which, at leading order, is:

∣[t(f)f]−1∣=256π5(πfM)5/3,(16)\left|[t(f)f]^{-1}\right|=\frac{256\pi}{5}(\pi f\mathcal{M})^{5/3}, \tag*{(16)}

meaning that the amplitude correction might only become comparable to the phase correction near the merger frequency, where the phase shift itself is the smallest. During earlier phases of the inspiral, when contributions from the phase shift can become significant, amplitude corrections are substantially suppressed.

Lastly, let us mention that the above Newtonian approach only remains valid when Γt≪1\Gamma t\ll1, which physically corresponds to the binary velocity remaining in the non-relativistic regime.

Additional tests

The choice of the parameters of the systems simulated throughout this work was oriented towards simplicity. In this section, we show the robustness of the framework against different, often more realistic, choices. All the analyses described below are performed with fini=4 Hzf_{\mathrm{ini}}=4\,\mathrm{Hz}, and with an eccentricity at 20 Hz of e20=0.1e_{20}=0.1; the specific injected parameters are listed in Table I.

Figure 7 shows the results obtained for a spinning system: all the parameters are well recovered, and spin does not bias the recovery of the other parameters, including eccentricity and Γ\Gamma, whose posteriors reproduce the ones obtained in the non-spinning case.

Parameter posteriors for a spinning, precessing system

Figure 7. Parameter posteriors recovered for a spinning, precessing system, compared to the reference one. The vertical lines indicate the injected values.

m1m_1, m2m_2 [M⊙M_\odot]a1a_1, a2a_2θ1\theta_1, θ2\theta_2, ϕ12\phi_{12}, ϕJL\phi_{JL}Γ\GammaFigure
[20, 20][0.0, 0.0][0., 0., 0., 0.]0.0Reference
[20, 20][0.4, 0.2][1.05, 0.52, 0.1, 0.2]0.0Figure 7
[41.83, 10.46][0.4, 0.2][1.05, 0.52, 0.1, 0.2]0.0Figure 8
[10, 10][0.0, 0.0][0., 0., 0., 0.]0.0Figure 9
[20, 20][0.0, 0.0][0., 0., 0., 0.]0.001Figure 10

Table 1. Injection parameters for the additional tests described in Appendix B, compared to the reference system used throughout this work, whose parameters are listed in Sec. II C. a1,2a_{1,2} are the components’ dimensionless spin magnitudes, with θ1,2\theta_{1,2} being the respective zenith angles with respect to the orbital angular momentum; ϕ12\phi_{12} denotes the difference between the azimuthal angles of the individual spin vectors, and ϕjl\phi_{jl} the angle between total and orbital angular momentum. The last column points to the figure with the results for the specific injection.

In Figure 8, we simulate the signal emitted by a spinning and unequal-mass system, with the same chirp mass as our reference configuration: in this case, the additional information from GW higher-order-modes yields an improvement in the constraints of all the parameters, as shown also in Ref. [106].

Parameter posteriors for a precessing system with mass ratio q = 0.25

Figure 8. Parameter posteriors obtained for a precessing system with mass ratio q=0.25q = 0.25, compared to the reference one. The vertical lines indicate the injected values.

The mass values for the systems simulated throughout this paper were chosen in order to reduce the computational cost of the analysis. However, we expect systems with lower mass to carry more information since their signals spend more time in band. This is shown in Figure 9, where we compare the results for a system with m1=m2=20M⊙m_1 = m_2 = 20 M_\odot, i.e., the reference system in this

Parameter posteriors for a simulated system with component masses m1 = m2 = 10 M⊙

Figure 9. Parameter posteriors recovered for a simulated system with component masses m1=m2=10M⊙m_1 = m_2 = 10 M_\odot, compared to the reference one. The vertical lines indicate the injected values.

work, with a system with m1=m2=10M⊙m_1 = m_2 = 10 M_\odot, keeping all the other parameters the same.

So far, we always injected signals with Γ=0\Gamma= 0. However, if this environmental effect is present, Γ\Gamma will have a non-zero value. To investigate how the analysis would behave in this case, we simulate a signal with e=0.1e = 0.1, and Γ=0.001\Gamma= 0.001, and the other parameters as in the reference configuration described in Sec. II C. This specific value of Γ\Gamma lies outside the 90% confidence interval obtained from the reference analysis with e20=0.1e_{20} = 0.1 and fini=4f_{\mathrm{ini}} = 4 Hz – as shown in Figure 2–Figure 1, and would therefore indicate whether we can confidently recover non-zero values of Γ\Gamma. Figure 10 shows not only that non-zero values of Γ\Gamma do not bias the recovery of the other parameters of the system, but also that we can correctly recover the injected value of Γ\Gamma.

Parameter posterior plots for a simulated system with $\Gamma=0.001$ compared to the reference system

Figure 10. Parameter posteriors recovered for a simulated system with Γ=0.001\Gamma= 0.001, compared to the reference one. The vertical lines indicate the injected values.

Finally, we investigate the improvements achievable with next-generation detectors, and in particular with ET [131, 133, 134]. Thanks to the technological and design improvements, ET is expected to gain sensitivity well below the one of current detectors, down to a few Hz. This will allow us to access all the information from the early evolution of the systems of interest, leading to massively tighter constraints on their parameters, including Γ\Gamma. This is evident from the posterior shown in Figure 11, for which we analyzed the detector data from flow=5f_{\mathrm{low}} = 5 Hz, while the starting frequency for waveform generation is kept to fini=4f_{\mathrm{ini}} = 4 Hz.

Parameter posterior plots for the reference system simulated in an LVK and ET network

Figure 11. Parameter posteriors recovered for our reference system simulated in an LVK and ET network. The vertical lines indicate the injected values.

Testing the multibanding likelihood for eccentric signals

In order to verify that employing the multibanding approach to likelihood computation for eccentric signals does not introduce biases, we perform an analysis on a simulated non-spinning system with m1=m2=10 M⊙m_1 = m_2 = 10\,M_\odot, e20=0.2e_{20} = 0.2, and Γ=0\Gamma= 0, using both the multibanding likelihood and the “full”, non-approximated one (i.e., the GWTransient likelihood in Bilby). Results are shown in Figure 12, where we see that the analysis performed with the multibanding likelihood can accurately reproduce the results obtained with the full one.

Comparison of parameter posteriors obtained with the full likelihood and the multibanding approach

Figure 12. Comparison between the parameter posteriors obtain when analyzing the same simulated signal with the full likelihood, in gray, and with the multibanding approach, in red. Black, vertical lines show the injected values.

Moreover, we test different settings for the multibanding approach, to ensure that the likelihood estimation is sufficiently close to the non-approximated one. In particular, we focus on the highest-mode setting: although it refers to the highest GW higher-order mode considered in the construction of the adaptive frequency grid, it indirectly affects the number of eccentric harmonics that we can capture. This might produce relevant effects especially when fini=2f_{\mathrm{ini}} = 2 Hz, given the larger number of harmonics included in the signal. Figure 13 shows the difference between the multibanding likelihood and the full one computed for a subset of 5000 samples among the ones collected for the analysis of our reference system when generating the waveform from fini=2f_{\mathrm{ini}} = 2 Hz. Overall, we find a very small difference in likelihood also when setting highest-mode to 6, which confirms the goodness of the multibanding approximation. However, we see how the error reduces when going to larger values of highest-mode, which, in turn, we expect to slightly influence the constraints on the parameters, including Γ\Gamma. Therefore, in the analyses with fini=2f_{\mathrm{ini}} = 2 Hz, we set highest-mode to 18.

Difference between the full likelihood and the multibanding likelihood for different highest-mode settings

Figure 13. Difference between the full likelihood and the multibanding one, computed with different values for the highest-mode setting and with fini=2f_{\mathrm{ini}} = 2 Hz for a subset of the samples obtained in our reference run, as function of the full likelihood normalized to the maximum likelihood value in the samples. The horizontal lines indicate the average likelihood difference for each value of highest-mode.

References

References

  1. [1]I. M. Romero-Shaw, P. D. Lasky, E. Thrane, and J. C. Bustillo, GW190521: orbital eccentricity and signatures of dynamical formation in a binary black hole merger signal, Astrophys. J. Lett. 903, L5 (2020), arXiv:2009.04771 [astro-ph.HE].
  2. [2]I. M. Romero-Shaw, P. D. Lasky, and E. Thrane, Signs of Eccentricity in Two Gravitational-wave Signals May Indicate a Subpopulation of Dynamically Assembled Binary Black Holes, Astrophys. J. Lett. 921, L31 (2021), arXiv:2108.01284 [astro-ph.HE].
  3. [3]I. M. Romero-Shaw, P. D. Lasky, and E. Thrane, Four Eccentric Mergers Increase the Evidence that LIGO–Virgo–KAGRA’s Binary Black Holes Form Dynamically, Astrophys. J. 940, 171 (2022), arXiv:2206.14695 [astro-ph-HE].
  4. [4]H. L. Iglesias et al., Eccentricity Estimation for Five Binary Black Hole Mergers with Higher-order Gravitational-wave Modes, Astrophys. J. 972, 65 (2024), arXiv:2208.01766 [gr-qc].arxiv.org/abs/2208.01766
  5. [5]I. Romero-Shaw, J. Stegmann, H. Tagawa, D. Gerosa, J. Samsing, N. Gupta, and S. R. Green, GW200208_222617 as an eccentric black-hole binary merger: Properties and astrophysical implications, Phys. Rev. D 112, 063052 (2025), arXiv:2506.17105 [astro-ph.HE].
  6. [6]N. Gupte et al., Evidence for eccentricity in the population of binary black holes observed by LIGO-Virgo-KAGRA, Phys. Rev. D 112, 104045 (2025), arXiv:2404.14286 [gr-qc].DOI
  7. [7]G. Morras, G. Pratten, and P. Schmidt, Orbital Eccentricity in a Neutron Star–Black Hole Merger, Astrophys. J. Lett. 1000, L2 (2026), arXiv:2503.15393 [astro-ph.HE].DOI
  8. [8]A. Jan, B.-J. Tsao, R. O’Shaughnessy, D. Shoemaker, and P. Laguna, GW200105: A detailed study of eccentricity in the neutron star-black hole binary, Phys. Rev. D 113, 024018 (2026), arXiv:2508.12460 [gr-qc].
  9. [9]K. Kacanja, K. Soni, and A. Nitz, Eccentricity signatures in LIGO-Virgo-KAGRA’s binary neutron star and neutron-star black holes, Phys. Rev. D 112, 122007 (2025), arXiv:2508.00179 [gr-qc].
  10. [10]P. McMillan, K. J. Wagner, G. Ficarra, C. O. Lousto, and R. O’Shaughnessy, Parameter estimation with targeted eccentric numerical-relativity simulations for GW200208_22 and GW190620, Phys. Rev. D 114, 024083 (2026), arXiv:2507.22862 [gr-qc].
  11. [11]M. d. L. Planas, S. Husa, A. Ramos-Buades, and J. Valencia, First Eccentric Inspiral–Merger–Ringdown Analysis of Neutron Star–Black Hole Mergers, Astrophys. J. 995, 47 (2025), arXiv:2506.01760 [astro-ph.HE].
  12. [12]A. Tiwari, S. A. Bhat, M. A. Shaikh, and S. J. Kapadia, Testing the Nature of GW200105 by Probing the Frequency Evolution of Eccentricity, Astrophys. J. 995, 48 (2025), arXiv:2509.26152 [astro-ph.HE].DOI
  13. [13]I. Mandel and A. Farmer, Merging stellar-mass binary black holes, Phys. Rept. 955, 1 (2022), arXiv:1806.05820 [astro-ph-HE].
  14. [14]M. Mapelli, Binary Black Hole Mergers: Formation and Populations, Front. Astron. Space Sci. 7, 38 (2020), arXiv:2105.12455 [astro-ph.HE].DOI
  15. [15]I. Mandel and F. S. Broekgaarden, Rates of compact object coalescences, Living Rev. Rel. 25, 1 (2022), arXiv:2107.14239 [astro-ph.HE].DOI
  16. [16]M. Zevin, I. M. Romero-Shaw, K. Kremer, E. Thrane, and P. D. Lasky, Implications of Eccentric Observations on Binary Black Hole Formation Channels, Astro- phys. J. Lett. 921, L43 (2021), arXiv:2106.09042 [astro-ph.HE].
  17. [17]J. Stegmann, A. Vigna-Gómez, A. Rantala, T. Wagg, L. Zwick, M. Renzo, L. A. C. van Son, S. E. de Mink, and S. D. M. White, Close Encounters of Wide Binaries Induced by the Galactic Tide: Implications for Stellar Mergers and Gravitational-wave Sources, ApJ 972, L19 (2024), arXiv:2405.02912 [astro-ph.GA].
  18. [18]J. Stegmann and J. Klencki, Orbital Eccentricity and Spin–Orbit Misalignment Are Evidence that Neutron Star–Black Hole Mergers Form through Triple Star Evolution, ApJ 991, L54 (2025).
  19. [19]K. Gültekin, M. C. Miller, and D. P. Hamilton, Three-Body Dynamics with Gravitational Wave Emission, Astrophys. J. 640, 156 (2006).DOI
  20. [20]J. Samsing and E. Ramirez-Ruiz, On the Assembly Rate of Highly Eccentric Binary Black Hole Mergers, ApJ 840, L14 (2017), arXiv:1703.09703 [astro-ph.HE].
  21. [21]K. Silsbee and S. Tremaine, Lidov-Kozai Cycles with Gravitational Radiation: Merging Black Holes in Isolated Triple Systems, Astrophys. J. 836, 39 (2017), arXiv:1608.07642 [astro-ph.HE].
  22. [22]F. Antonini, S. Toonen, and A. S. Hamers, Binary Black Hole Mergers from Field Triples: Properties, Rates, and the Impact of Stellar Evolution, Astrophys. J. 841, 77 (2017), arXiv:1703.06614 [astro-ph.GA].
  23. [23]J. Samsing and T. Ilan, Topology of black hole binary-single interactions, MNRAS 476, 1548 (2018), arXiv:1706.04672 [astro-ph.HE].DOI
  24. [24]J. Samsing, A. Askar, and M. Giersz, MOCCA-SURVEY Database. I. Eccentric Black Hole Mergers during Binary-Single Interactions in Globular Clusters, Astrophys. J. 855, 124 (2018), arXiv:1712.06186 [astro-ph.HE].
  25. [25]J. Samsing and D. J. D’Orazio, Black Hole Mergers From Globular Clusters Observable by LISA I: Eccentric Sources Originating From Relativistic N-body Dynamics, MNRAS 10.1093/mnras/sty2334 (2018), arXiv:1804.06519 [astro-ph.HE].
  26. [26]C. L. Rodriguez, P. Amaro-Seoane, S. Chatterjee, K. Kremer, F. A. Rasio, J. Samsing, C. S. Ye, and M. Zevin, Post-Newtonian dynamics in dense star clusters: Formation, masses, and merger rates of highly-eccentric black hole binaries, Phys. Rev. D 98, 123005 (2018), arXiv:1811.04926 [astro-ph.HE].
  27. [27]B. Liu, D. Lai, and Y.-H. Wang, Black Hole and Neutron Star Binary Mergers in Triple Systems. II. Merger Eccentricity and Spin-Orbit Misalignment, Astrophys. J. 881, 41 (2019), arXiv:1905.00427 [astro-ph.HE].
  28. [28]J. Samsing, A. S. Hamers, and J. G. Tyles, Effect of distant encounters on black hole binaries in globular clusters: Systematic increase of in-cluster mergers in the LISA band, Phys. Rev. D 100, 043010 (2019), arXiv:1906.07189 [astro-ph.HE].
  29. [29]J. Samsing, D. J. D’Orazio, K. Kremer, C. L. Rodriguez, and A. Askar, Single-single gravitational-wave captures in globular clusters: Eccentric deci-Hertz sources observable by DECIGO and Tian-Qin, Phys. Rev. D 101, 123010 (2020), arXiv:1907.11231 [astro-ph.HE].
  30. [30]K. Kremer, C. L. Rodriguez, P. Amaro-Seoane, K. Breivik, S. Chatterjee, M. L. Katz, S. L. Larson, F. A. Rasio, J. Samsing, C. S. Ye, and M. Zevin, Post-Newtonian dynamics in dense star clusters: Binary black holes in the LISA band, Phys. Rev. D 99, 063003 (2019), arXiv:1811.1812 [astro-ph.HE].
  31. [31]R. M. O’Leary, B. Kocsis, and A. Loeb, Gravitational waves from scattering of stellar-mass black holes in galactic nuclei, Mon. Not. Roy. Astron. Soc. 395, 2127 (2009), arXiv:0807.2638 [astro-ph].
  32. [32]F. Antonini and H. B. Perets, Secular evolution of compact binaries near massive black holes: Gravitational wave sources and other exotica, Astrophys. J. 757, 27 (2012), arXiv:1203.2938 [astro-ph.GA].
  33. [33]J. Samsing, M. MacLeod, and E. Ramirez-Ruiz, The Formation of Eccentric Compact Binary Inspirals and the Role of Gravitational Wave Emission in Binary-Single Stellar Encounters, Astrophys. J. 784, 71 (2014), arXiv:1308.2964 [astro-ph.HE].
  34. [34]L. Gondán, B. Kocsis, P. Raffai, and Z. Frei, Eccentric Black Hole Gravitational-Wave Capture Sources in Galactic Nuclei: Distribution of Binary Parameters, Astrophys. J. 860, 5 (2018), arXiv:1711.09989 [astro-ph.HE].
  35. [35]J. Samsing, Eccentric Black Hole Mergers Forming in Globular Clusters, Phys. Rev. D 97, 103014 (2018), arXiv:1711.07452 [astro-ph.HE].DOI
  36. [36]J. Takátsy, B. Bécsy, and P. Raffai, Eccentricity distribution of eccentric binary black holes in galactic nuclei, Mon. Not. Roy. Astron. Soc. 486, 570 (2019), arXiv:1812.04012 [astro-ph.HE].
  37. [37]M. Zevin, J. Samsing, C. Rodriguez, C.-J. Haster, and E. Ramirez-Ruiz, Eccentric Black Hole Mergers in Dense Star Clusters: The Role of Binary–Binary Encounters, Astrophys. J. 871, 91 (2019), arXiv:1810.00901 [astro-ph.HE].
  38. [38]A. Rasskazov and B. Kocsis, The rate of stellar mass black hole scattering in galactic nuclei, Astrophys. J. 881, 20 (2019), arXiv:1902.03242 [astro-ph.HE].
  39. [39]H. Tagawa, Z. Haiman, and B. Kocsis, Formation and Evolution of Compact Object Binaries in AGN Disks, Astrophys. J. 898, 25 (2020), arXiv:1912.08218 [astro-ph.GA].
  40. [40]G. Fragione, N. Leigh, and R. Perna, Black hole and neutron star mergers in Galactic Nuclei: the role of triples, Mon. Not. Roy. Astron. Soc. 488, 2825 (2019), arXiv:1903.09160 [astro-ph.GA].
  41. [41]H. Tagawa, B. Kocsis, Z. Haiman, I. Bartos, K. Omukai, and J. Samsing, Eccentric Black Hole Mergers in Active Galactic Nuclei, Astrophys. J. Lett. 907, L20 (2021), arXiv:2010.10526 [astro-ph.HE].DOI
  42. [42]J. Samsing, I. Bartos, D. J. D’Orazio, Z. Haiman, B. Kocsis, N. W. C. Leigh, B. Liu, M. E. Pessah, and H. Tagawa, AGN as potential factories for eccentric black hole mergers, Nature 603, 237 (2022), arXiv:2010.09765 [astro-ph.HE].
  43. [43]M. Dall’Amico, M. Mapelli, S. Torniamenti, and M. A. Sedda, Eccentric black hole mergers via three-body interactions in young, globular, and nuclear star clusters, Astron. Astrophys. 683, A186 (2024), arXiv:2303.07421 [astro-ph.HE].
  44. [44]N. Yunes, K. G. Arun, E. Berti, and C. M. Will, Post-Circular Expansion of Eccentric Binary Inspirals: Fourier-Domain Waveforms in the Stationary Phase Approximation, Phys. Rev. D 80, 084001 (2009), [Erratum: Phys.Rev.D 89, 109901 (2014)], arXiv:0906.0313 [gr-qc].
  45. [45]B. Mikóczi, B. Kocsis, P. Gopakumar, and M. Vasuth, Parameter estimation for inspiraling eccentric compact binaries including pericenter precession, Phys. Rev. D 86, 104027 (2012), arXiv:1206.5786 [gr-qc].
  46. [46]B. Sun, Z. Cao, Y. Wang, and H.-C. Yeh, Parameter estimation of eccentric inspiraling compact binaries using an enhanced post circular model for ground-based detectors, Phys. Rev. D 92, 044034 (2015).DOI
  47. [47]B. Moore and N. Yunes, Data Analysis Implications of Moderately Eccentric Gravitational Waves, Class. Quant. Grav. 37, 225015 (2020), arXiv:1910.01680 [gr-qc].DOI
  48. [48]P. Sinha, R. Prasad, M. K. Singh, P. Kumar, A. Maurya, and K. Paul, Early warning from eccentric compact binaries: Template initialization and subdominant mode effects, Phys. Rev. D 113, 123060 (2026), arXiv:2507.07021 [gr-qc].DOI
  49. [49]L. Pompili, A. Gamboa, and A. Buonanno, Eccentric and unbound compact binaries in the LIGO-Virgo-KAGRA catalog: Parameter estimation and waveform systematics with seobnrv4hem, Phys. Rev. D 114, 104009 (2026), arXiv:2605.28716 [gr-qc].
  50. [50]J. Takátsy, L. Zwick, K. Hendriks, P. Saini, G. Fabj, and J. Samsing, The construction and use of dephasing prescriptions for environmental effects in gravitational wave astronomy, Class. Quant. Grav. 42, 215006 (2025), arXiv:2505.09513 [astro-ph.HE].DOI
  51. [51]L. Zwick, K. Hendriks, P. Saini, J. Takátsy, C. Rowan, J. Samsing, and J. Stegmann, Environmental Effects in Stellar-mass Gravitational-wave Sources. II. Enhanced Detectability of Phase Shifts in Eccentric Subpopulations, Astrophys. J. 1006, 96 (2026), arXiv:2511.04540 [astro-ph.HE].arxiv.org/abs/2511.04540
  52. [52]S. K. Chakrabarti, Binary Black Holes in Stationary Orbits and a Test of the Active Galactic Nucleus Paradigm, Astrophys. J. 411, 610 (1993).DOI
  53. [53]F. D. Ryan, Gravitational waves from the inspiral of a compact object into a massive, axisymmetric body with arbitrary multipole moments, Phys. Rev. D 52, 5707 (1995).pubmed.ncbi.nlm.nih.gov/10019103
  54. [54]E. Barausse and L. Rezzolla, Influence of the hydrodynamic drag from an accretion torus on extreme mass-ratio inspirals, Phys. Rev. D 77, 104027 (2008), arXiv:0711.4558 [gr-qc].DOI
  55. [55]Y. Levin, Starbursts near supermassive black holes: young stars in the Galactic Centre, and gravitational waves in LISA band, MNRAS 374, 515 (2007), arXiv:astro-ph/0603583 [astro-ph].DOI
  56. [56]B. Kocsis, N. Yunes, and A. A. Loeb, Observable signatures of extreme mass-ratio inspiral black hole binaries embedded in thin accretion disks, Phys. Rev. D 84, 024032 (2011).DOI
  57. [57]E. Barausse, V. Cardoso, and P. Pani, Can environmental effects spoil precision gravitational-wave astrophysics?, Phys. Rev. D 89, 104059 (2014), arXiv:1404.7149 [gr-qc].DOI
  58. [58]K. Inayoshi, N. Tamanini, C. Caprini, and Z. Haiman, Probing stellar binary black hole formation in galactic nuclei via the imprint of their center of mass acceleration on their gravitational wave signal, Phys. Rev. D 96, 063014 (2017), arXiv:1702.06529 [astro-ph.HE].DOI
  59. [59]Y. Meiron, B. Kocsis, and A. A. Loeb, Detecting Triple Systems with Gravitational Wave Observations, Astrophys. J. 834, 200 (2017), arXiv:1604.02148 [astro-ph.HE].DOI
  60. [60]M. Bonetti, E. Barausse, G. Faye, F. Haardt, and A. Sesana, About gravitational-wave generation by a three-body system, Classical and Quantum Gravity 34, 215004 (2017), arXiv:1707.04902 [gr-qc].DOI
  61. [61]A. Torres- Orjuela, X. Chen, Z. Cao, P. Amaro-Seoane, and P. Peng, Detecting the beaming effect of gravitational waves, Phys. Rev. D 100, 063012 (2019), arXiv:1806.09857 [astro-ph.HE].
  62. [62]L. Randall and Z.-Z. Xianyu, Observing Eccentricity Oscillations of Binary Black Holes in LISA, arXiv e-prints , arXiv:1902.08604 (2019), arXiv:1902.08604 [astro-ph.HE].arxiv.org/abs/1902.08604
  63. [63]V. Cardoso and A. Maselli, Constraints on the astrophysical environment of binaries with gravitational-wave observations, A&A 644, A147 (2020), arXiv:1909.05870 [astro-ph.HE].DOI
  64. [64]D. J. D’Orazio and A. Loeb, Repeated gravitational lensing of gravitational waves in hierarchical black hole triples, Phys. Rev. D 101, 083031 (2020), arXiv:1910.02966 [astro-ph.HE].DOI
  65. [65]B. Liu, D. J. D’Orazio, A. Vigna-Gómez, and J. Samsing, Uncovering a hidden black hole binary from secular eccentricity variations of a tertiary star, Phys. Rev. D 106, 123010 (2022), arXiv:2207.10091 [astro-ph.HE].DOI
  66. [66]Z. Xuan, S. Naoz, and X. Chen, Detecting Accelerating Eccentric Binaries in the LISA Band, arXiv e-prints , arXiv:2210.03129 (2022), arXiv:2210.03129 [astro-ph.HE].DOI
  67. [67]M. Garg, A. Derdzinski, L. Zwick, P. R. Capelo, and L. Mayer, The imprint of gas on gravitational waves from LISA intermediate-mass black hole binaries, MNRAS 517, 1339 (2022), arXiv:2206.05292 [astro-ph.GA].DOI
  68. [68]P. S. Cole, A. Coogan, B. J. Kavanagh, and G. Bertone, Measuring dark matter spikes around primordial black holes with Einstein Telescope and Cosmic Explorer, arXiv e-prints , arXiv:2207.07576 (2022), arXiv:2207.07576 [astro-ph.CO].DOI
  69. [69]R. S. Chandramouli and N. Yunes, Ready-to-use analytic model for gravitational waves from a hierarchical triple with Kozai-Lidov oscillations, Phys. Rev. D 105, 064009 (2022), arXiv:2107.00741 [gr-qc].DOI
  70. [70]L. Sberna, S. Babak, S. Marsat, A. Caputo, G. Cusin, A. Toubiana, E. Barausse, C. Caprini, T. Dal Canton, A. Sesana, and N. Tamanini, Observing GW190521-like binary black holes and their environment with LISA, Phys. Rev. D 106, 064056 (2022), arXiv:2205.08550 [gr-qc].DOI
  71. [71]L. Zwick, P. R. Capelo, and L. Mayer, Priorities in gravitational waveforms for future space-borne detectors: vacuum accuracy or environment?, MNRAS 521, 4645 (2023), arXiv:2209.04060 [gr-qc].DOI
  72. [72]C. Tiede, D. J. D’Orazio, L. Zwick, and P. C. Duffell, Disk-induced Binary Precession: Implications for Dynamics and Multimessenger Observations of Black Hole Binaries, Astrophys. J. 964, 46 (2024), arXiv:2312.01805 [astro-ph.HE].DOI
  73. [73]A. Torres-Orjuela, X. Chen, and P. Amaro Seoane, Excitation of gravitational wave modes by a center-of-mass velocity of the source, Phys. Rev. D 104, 123025 (2021), arXiv:2010.15856 [astro-ph.CO].
  74. [74]K. Destounis, A. Kulathingal, K. D. Kokkotas, and G. O. Papadopoulos, Gravitational-wave imprints of compact and galactic-scale environments in extreme-mass-ratio binaries, arXiv e-prints , arXiv:2210.09357 (2022), arXiv:2210.09357 [gr-qc].DOI
  75. [75]V. Cardoso, K. Destounis, F. Duque, R. P. Macedo, and A. Maselli, Gravitational Waves from Extreme-Mass-Ratio Systems in Astrophysical Environments, Phys. Rev. Lett. 129, 241103 (2022), arXiv:2210.01133 [gr-qc].DOI
  76. [76]A. Caputo, L. Sberna, A. Toubiana, S. Babak, E. Barausse, S. Marsat, and P. Pani, Gravitational-wave Detection and Parameter Estimation for Accreting Black-hole Binaries and Their Electromagnetic Counterpart, Astrophys. J. 892, 90 (2020), arXiv:2001.03620 [astro-ph.HE].DOI
  77. [77]L. Zwick, A. Derdzinski, M. Garg, P. R. Capelo, and L. Mayer, Dirty waveforms: multiband harmonic content of gas-embedded gravitational wave sources, MNRAS 511, 6143 (2022), arXiv:2110.09097 [astro-ph.HE].DOI
  78. [78]A. Vijaykumar, A. Tiwari, S. J. Kapadia, K. G. Arun, and P. Ajith, Waltzing binaries: Probing the line-of-sight acceleration of merging compact objects with gravitational waves, Astrophys. J. 954, 105 (2023).DOI
  79. [79]L. Zwick, C. Tiede, A. A. Trani, A. Derdzinski, Z. Haiman, D. J. D’Orazio, and J. Samsing, Novel category of environmental effects on gravitational waves from binaries perturbed by periodic forces, Phys. Rev. D 110, 103005 (2024), arXiv:2405.05698 [gr-qc].arxiv.org/abs/2405.05698
  80. [80]L. Zwick, K. Hendriks, D. O’Neill, J. Takátsy, P. Kirkeberg, C. Tiede, J. Stegmann, J. Samsing, and D. J. D’Orazio, Dissecting environmental effects with eccentric gravitational wave sources, Phys. Rev. D 112, 063005 (2025), arXiv:2506.09140 [astro-ph.HE].arxiv.org/abs/2506.09140
  81. [81]A. Derdzinski, D. D’Orazio, P. Duffell, Z. Haiman, and A. MacFadyen, Evolution of gas disc-embedded intermediate mass ratio inspirals in the LISA band, MNRAS 501, 3540 (2021), arXiv:2005.11333 [astro-ph.HE].DOI
  82. [82]P. Basu, S. Chatterjee, and S. Mondal, Eccentric orbits in disc-embedded EMRIs : orbital evolution and observability trend in LISA, MNRAS 531, 1506 (2024).DOI
  83. [83]G. Caneva Santoro, S. Roy, R. Vicente, M. Haney, O. J. Piccinni, W. Del Pozzo, and M. Martinez, First Constraints on Compact Binary Environments from LIGO-Virgo Data, Phys. Rev. Lett. 132, 251401 (2024), arXiv:2309.05061 [gr-qc].arxiv.org/abs/2309.05061
  84. [84]R. Vicente, T. K. Karydas, and G. Bertone, A fully relativistic treatment of EMRIs in collisionless environments, arXiv e-prints , arXiv:2505.09715 (2025), arXiv:2505.09715 [gr-qc].
  85. [85]C. Dyson, T. F. M. Spieksma, R. Brito, M. van de Meent, and S. Dolan, Environmental effects in extreme mass ratio inspirals: perturbations to the environment in Kerr, arXiv e-prints , arXiv:2501.09806 (2025), arXiv:2501.09806 [gr-qc].arxiv.org/abs/2501.09806
  86. [86]K. Destounis and P. G. S. Fernandes, Chaos and Carter: Extreme-mass-ratio systems of relativistic rotating black holes in astrophysical environments, arXiv e-prints , arXiv:2508.20191 (2025), arXiv:2508.20191 [gr-qc].arxiv.org/abs/2508.20191
  87. [87]L. Copparoni, L. Speri, L. Sberna, A. Derdzinski, and E. Barausse, The implications of stochastic gas torques for asymmetric binaries in the LISA band, arXiv e-prints , arXiv:2502.10087 (2025), arXiv:2502.10087 [gr-qc].DOI
  88. [88]A. Torres-Orjuela, V. Vázquez-Aceves, and T.-X. Wang, Detection of Intermediate-mass Ratio Inspirals in Globular Clusters: Revealing the Brownian Motion with Gravitational Waves, Astrophys. J. 986, 155 (2025), arXiv:2501.13466 [astro-ph.HE].
  89. [89]L. Zwick, J. Takátsy, P. Saini, K. Hendriks, J. Samsing, C. Tiede, C. Rowan, and A. A. Trani, Environmental Effects in Stellar Mass Gravitational-wave Sources. I. Expected Fraction of Signals with Significant Dephasing in the Dynamical and Active Galactic Nucleus Channels, Astrophys. J. 991, 131 (2025), arXiv:2503.24084 [astro-ph.HE].
  90. [90]H. Tagawa, C. Rowan, J. Takátsy, L. Zwick, K. Hendriks, W.-B. Han, and J. Samsing, Gravitational Wave Phase Shifts of Black Hole Mergers in AGN Disks, Astrophys. J. 998, 244 (2026), arXiv:2511.15193 [astro-ph.HE].arxiv.org/abs/2511.15193
  91. [91]N. Yunes, M. Coleman Miller, and J. Thornburg, The Effect of Massive Perturbers on Extreme Mass-Ratio Inspiral Waveforms, Phys. Rev. D 83, 044030 (2011), arXiv:1010.1721 [astro-ph.GA].DOI
  92. [92]C. Bonvin, C. Caprini, R. Sturani, and N. Tamanini, Effect of matter structure on the gravitational waveform, Phys. Rev. D 95, 044029 (2017), arXiv:1609.08093 [astro-ph.CO].DOI
  93. [93]T. Robson, N. J. Cornish, and C. Liu, The construction and use of LISA sensitivity curves, Classical and Quantum Gravity 36, 105011 (2019), arXiv:1803.01944 [astro-ph.HE].DOI
  94. [94]N. Tamanini, A. Klein, C. Bonvin, E. Barausse, and C. Caprini, Peculiar acceleration of stellar-origin black hole binaries: Measurement and biases with LISA, Phys. Rev. D 101, 063002 (2020), arXiv:1907.02018 [astro-ph.IM].DOI
  95. [95]J. Samsing, K. Hendriks, L. Zwick, D. J. D’Orazio, and B. Liu, Gravitational-wave Phase Shifts in Eccentric Black Hole Mergers as a Probe of Dynamical Formation Environments, Astrophys. J. 990, 211 (2025), arXiv:2403.05625 [astro-ph.HE].arxiv.org/abs/2403.05625
  96. [96]K. Hendriks, D. Atallah, M. Martinez, M. Zevin, L. Zwick, A. A. Trani, P. Saini, J. Takátsy, and J. Samsing, Large gravitational wave phase shifts from strong three-body interactions in dense stellar clusters, Phys. Rev. D 114, 043071 (2026), arXiv:2411.08572 [astro-ph.HE].
  97. [97]K. Hendriks, L. Zwick, P. Saini, J. Takátsy, and J. Samsing, Towards gravitational wave parameter inference for binaries with an eccentric companion, arXiv e-prints (2026), arXiv:2601.14918 [astro-ph.HE].arxiv.org/abs/2601.14918
  98. [98]A. Tiwari, A. Vijaykumar, S. J. Kapadia, G. Fragione, and S. Chatterjee, Accelerated binary black holes in globular clusters: forecasts and detectability in the era of space-based gravitational-wave detectors, Mon. Not. Roy. Astron. Soc. 527, 8586 (2023), arXiv:2307.00930 [astro-ph.HE].DOI
  99. [99]A. Tiwari, A. Vijaykumar, S. J. Kapadia, S. Chatterjee, and G. Fragione, Profiling stellar environments for gravitational wave sources, Phys. Rev. D 112, 084034 (2025), arXiv:2407.15117 [astro-ph.HE].DOI
  100. [100]M. Lazarow, N. Leslie, and L. Dai, Gravitational waveform model for detecting accelerating inspiraling binaries, Phys. Rev. D 110, 083008 (2024), arXiv:2401.04175 [gr-qc].arxiv.org/abs/2401.04175
  101. [101]A. Tiwari, A. Vijaykumar, S. J. Kapadia, S. Ghosh, and A. B. Nielsen, Pipeline to search for signatures of line-of-sight acceleration in gravitational wave signals produced by compact binary coalescences, Phys. Rev. D 113, 104072 (2026), arXiv:2506.22272 [astro-ph.HE].DOI
  102. [102]S. Gera and P. Dutta Roy, Impact of neglecting center-of-mass acceleration in parameter estimation of stellar-mass black holes, arXiv e-prints (2025), arXiv:2512.21979 [gr-qc].
  103. [103]L. Roy and A. H. Nitz, Constraints on Line-of-Sight Acceleration from O1-O4, arXiv e-prints (2026), arXiv:2606.25304 [astro-ph.HE].arxiv.org/abs/2606.25304
  104. [104]A. G. Abac et al. (LIGO Scientific, VIRGO, KAGRA), GWTС-4.0: Tests of General Relativity. II. Parameterized Tests, arXiv e-prints (2026), arXiv:2603.19020 [gr-qc].
  105. [105]J. Samsing, L. Zwick, P. Saini, and J. Takátsy, Probing the Formation Environment of Strongly Lensed Black Hole Mergers: Implications for the AGN-disk Channel, arXiv e-prints (2025), arXiv:2511.19407 [astro-ph.HE].arxiv.org/abs/2511.19407
  106. [106]S. Roy and J. Janquart, Line-of-sight acceleration in compact binaries with higher harmonics and eccentricity, Phys. Rev. D 114, 044075 (2026), arXiv:2606.08838 [gr-qc].DOI
  107. [107]L. Pompili, A. Gamboa, and A. Buonanno, Joint inference of line-of-sight acceleration and orbital eccentricity in neutron-star–black-hole binaries, arXiv e-prints (2026), arXiv:2606.28156 [gr-qc].arxiv.org/abs/2606.28156
  108. [108]M. A. Ansari, A. Tiwari, S. A. Bhat, S. Nair, and S. J. Kapadia, Line-of-sight Acceleration Driven Effects in Eccentric Compact Binary Coalescences, arXiv e-prints (2026), arXiv:2610.08459 [gr-qc].arxiv.org/abs/2610.08459
  109. [109]G. Morras, G. Pratten, and P. Schmidt, Improved post-Newtonian waveform model for inspiralling precessing-eccentric compact binaries, Phys. Rev. D 111, 084052 (2025), arXiv:2502.03929 [gr-qc].DOI
  110. [110]G. Morras, G. Pratten, P. Schmidt, and A. Buonanno, Post-Newtonian inspiral waveform model for eccentric precessing binaries with higher-order modes and matter effects, Phys. Rev. D 114, 044032 (2026), arXiv:2604.11903 [gr-qc].DOI
  111. [111]A. Klein, Y. Boetzel, A. Gopakumar, P. Jetzer, and L. de Vittori, Fourier domain gravitational waveforms for precessing eccentric binaries, Phys. Rev. D 98, 104043 (2018), arXiv:1801.08542 [gr-qc].DOI
  112. [112]A. Klein, EFPE: Efficient fully precessing eccentric gravitational waveforms for binaries with long inspirals, (2021), arXiv:2106.10291 [gr-qc].arxiv.org/abs/2106.10291
  113. [113]J. N. Arredondo, A. Klein, and N. Yunes, Efficient gravitational-wave model for fully-precessing and moderately eccentric, compact binary inspirals, Phys. Rev. D 110, 044044 (2024), arXiv:2402.06804 [gr-qc].DOI
  114. [114]A. Álvaro-Díaz and G. Morras, Horizon absorption in eccentric precessing binary black hole inspirals and its importance for gravitational wave data analysis, Phys. Rev. D 114, 044033 (2026), arXiv:2606.11705 [gr-qc].
  115. [115]G. Morras, Modeling gravitational wave modes from the inspiral of binaries with arbitrary eccentricity, Phys. Rev. D 112, 084015 (2025), arXiv:2507.00169 [gr-qc].DOI
  116. [116]A. Klein, N. Cornish, and N. Yunes, Fast Frequency-domain Waveforms for Spin-Precessing Binary Inspirals, Phys. Rev. D 90, 124029 (2014), arXiv:1408.5158 [gr-qc].DOI
  117. [117]G. Ashton et al., BILBY: A user-friendly Bayesian inference library for gravitational-wave astronomy, Astrophys. J. Suppl. 241, 27 (2019), arXiv:1811.02042 [astro-ph.IM].DOI
  118. [118]I. M. Romero-Shaw et al., Bayesian inference for compact binary coalescences with bilby: validation and application to the first LIGO–Virgo gravitational-wave transient catalogue, Mon. Not. Roy. Astron. Soc. 499, 3295 (2020), arXiv:2006.00714 [astro-ph.IM].
  119. [119]J. S. Speagle, DYNESTY: a dynamic nested sampling package for estimating Bayesian posteriors and evidences, MNRAS 493, 3132 (2020), arXiv:1904.02180 [astro-ph.IM].DOI
  120. [120]S. Morisaki, Accelerating parameter estimation of gravitational waves from compact binary coalescence using adaptive frequency resolutions, Phys. Rev. D 104, 044062 (2021), arXiv:2104.07813 [gr-qc].DOI
  121. [121]J. Aasi et al. (LIGO Scientific), Advanced LIGO, Class. Quant. Grav. 32, 074001 (2015), arXiv:1411.4547 [gr-qc].
  122. [122]F. Acernese et al. (VIRGO), Advanced Virgo: a second-generation interferometric gravitational wave detector, Class. Quant. Grav. 32, 024001 (2015), arXiv:1408.3978 [gr-qc].
  123. [123]L. S. Finn, Detection, measurement and gravitational radiation, Phys. Rev. D 46, 5236 (1992), arXiv:gr-qc/9209010.DOI
  124. [124]C. Cutler and É. E. Flanagan, Gravitational waves from merging compact binaries: How accurately can one extract the binary’s parameters from the inspiral waveform?, Phys. Rev. D 49, 2658 (1994), arXiv:gr-qc/9402014 [gr-qc].DOI
  125. [125]M. A. Ansari, A. Tiwari, S. A. Bhat, A. Vijaykumar, and S. J. Kapadia, Line-of-Sight Acceleration-driven waveform corrections for unequal-mass compact binary coalescences, arXiv e-prints (2026), arXiv:2610.08492 [gr-qc].arxiv.org/abs/2610.08492
  126. [126]L. Pathak, H. P. Phurailatpam, and A. Gopakumar, On the Presence of a Tertiary Compact Object in GW190814, Astrophys. J. 1009, 214 (2026), arXiv:2605.21955 [gr-qc].arxiv.org/abs/2605.21955
  127. [127]R. Abbott et al. (LIGO Scientific, KAGRA, VIRGO), Observation of Gravitational Waves from Two Neutron Star–Black Hole Coalescences, Astrophys. J. Lett. 915, L5 (2021), arXiv:2106.15163 [astro-ph.HE].DOI
  128. [128]https://doi.org/10.7935/wc8b-rz86 (), GWOSC data release GW200105_162426.
  129. [129]https://dcc.ligo.org/LIGO-P2100143/public (), Parameter estimation sample release for GW200105_162426.
  130. [130]K. S. Phukon, P. Schmidt, G. Morras, and G. Pratten, Detection of GW200105 with a targeted eccentric search, Phys. Rev. D 113, 103023 (2026), arXiv:2512.10803 [gr-qc].DOI
  131. [131]A. Abac et al. (ET), The Science of the Einstein Telescope, JCAP 03, 081, arXiv:2503.12263 [gr-qc].DOI
  132. [132]J. Takátsy, A. Puecher, G. Morras, L. Zwick, Eccentricity greatly enhances the detectability of line-of-sight acceleration in gravitational wave signals - data release [Dataset] (2026).
  133. [133]M. Branchesi et al., Science with the Einstein Telescope: a comparison of different designs, JCAP 07, 068, arXiv:2303.15923 [gr-qc].DOI
  134. [134]S. Danilishin and T. Zhang, Einstein Telescope sensitivity curves used for CoBA Science study, Tech. Rep. ET-0304A-22 (Einstein Telescope Collaboration, 2022).

Paper details

vlnc:W_pjbp5a72
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY-SA 4.0
Usage

views · 12

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents