Frequency-domain reduced-order model of aligned-spin effective-one-body waveforms with higher-order modes
Frequency-domain reduced-order model of aligned-spin effective-one-body waveforms with higher-order modes
复制标题
DOI：
10.1103/physrevd.101.124040
复制
发表时间：
2020-03
影响因子：
5
通讯作者：
R. Cotesta;S. Marsat;M. Purrer
中科院分区：
文献类型：
--
作者：
R. Cotesta;S. Marsat;M. Purrer
我们提出了一种用于双黑洞 (BBH) SEOBNRv4HM 的对齐自旋有效单体 (EOB) 模型的频域降阶模型 (ROM)，其中包括超出主导 $(\ell, |m|) = (2,2)$ 模式的球谐模式 $(\ell, |m|) = (2,1),(3,3),(4,4),(5,5)$。这些更高的模式对于准确表示非对称 BBH 发出的波形至关重要。我们讨论了波形的分解，扩展了文献中的其他方法，使我们能够准确有效地捕获更高模式波形的形态。我们表明 ROM 非常准确，对于 $[2.8,100] M_\odot$ 中的总质量，针对 SEOBNRv4HM 的不忠实度中值（最大）低于 $0.001\% (0.03\%)$。对于 $M = 300 M_\odot$ 的总质量，不忠的中值（最大值）增加至 $0.004\% (0.17\%)$。与数值相对论模拟相比，这仍然比 SEOBNRv4HM 的估计精度低至少一个数量级。与 SEOBNRv4HM 相比，ROM 生成波形的速度快两个数量级。数据分析应用程序通常需要 $\mathcal{O}(10^6-10^8)$ 波形评估，而 SEOBNRv4HM 通常速度太慢。因此，ROM 对于允许在搜索和贝叶斯参数推断中使用 SEOBNRv4HM 波形至关重要。我们提出了一项有针对性的参数估计研究，该研究显示了使用包含更高模式的波形时测量二进制参数的改进，并与其他三种波形模型进行了比较。
We present a frequency domain reduced order model (ROM) for the aligned-spin effective-one-body (EOB) model for binary black holes (BBHs) SEOBNRv4HM that includes the spherical harmonics modes $(\ell, |m|) = (2,1),(3,3),(4,4),(5,5)$ beyond the dominant $(\ell, |m|) = (2,2)$ mode. These higher modes are crucial to accurately represent the waveform emitted from asymmetric BBHs. We discuss a decomposition of the waveform, extending other methods in the literature, that allows us to accurately and efficiently capture the morphology of higher mode waveforms. We show that the ROM is very accurate with median (maximum) values of the unfaithfulness against SEOBNRv4HM lower than $0.001\% (0.03\%)$ for total masses in $[2.8,100] M_\odot$. For a total mass of $M = 300 M_\odot$ the median (maximum) value of the unfaithfulness increases up to $0.004\% (0.17\%)$. This is still at least an order of magnitude lower than the estimated accuracy of SEOBNRv4HM compared to numerical relativity simulations. The ROM is two orders of magnitude faster in generating a waveform compared to SEOBNRv4HM. Data analysis applications typically require $\mathcal{O}(10^6-10^8)$ waveform evaluations for which SEOBNRv4HM is in general too slow. The ROM is therefore crucial to allow the SEOBNRv4HM waveform to be used in searches and Bayesian parameter inference. We present a targeted parameter estimation study that shows the improvements in measuring binary parameters when using waveforms that includes higher modes and compare against three other waveform models.