EXPLORING THE COSMIC REIONIZATION EPOCH IN FREQUENCY SPACE: AN IMPROVED APPROACH TO REMOVE THE FOREGROUND IN 21 cm TOMOGRAPHY
EXPLORING THE COSMIC REIONIZATION EPOCH IN FREQUENCY SPACE: AN IMPROVED APPROACH TO REMOVE THE FOREGROUND IN 21 cm TOMOGRAPHY
复制标题
DOI：
10.1088/0004-637x/763/2/90
复制
发表时间：
2012-11
期刊：
影响因子：
--
通讯作者：
Jingying Wang;Haiguang Xu;T. An;Junhua Gu;Xueying Guo;Weitian Li;Yu Wang;Chengze Liu;O. Martineau
中科院分区：
文献类型：
--
作者：
Jingying Wang;Haiguang Xu;T. An;Junhua Gu;Xueying Guo;Weitian Li;Yu Wang;Chengze Liu;O. Martineau
为了正确地恢复宇宙电离过程中中性氢发射的21厘米红移信号，我们重新审视了基于频率空间二次多项式拟合法的分离方法，以样本统计的形式定量评估恢复的21厘米信号的质量，以考察它们在复杂前景下是否能令人满意地工作。我们构建了前景模型来描述真实天空的空间子结构和频谱子结构，并用它来模拟观测到的射电频谱。通过对恢复的21厘米谱和相应的功率谱进行统计分析来比较不同的分离方法，以及它们对再电离过程的平均晕偏差b和平均电离分数Xe的约束，在z=8和60mK的噪声水平下，我们发现尽管复杂的前景可以很好地用二次多项式展开来逼近，但21 cm信号的MPC尺度分量的很大一部分(≳6 h−1 MPC尺度的75%和≳1 h−1 MPC尺度的34%)在应用单一窄段分离方法时容易被错误地识别为前景的一部分。用本文提出的三窄段拟合法可以获得21厘米信号的最佳恢复和对b和Xe的最紧密测定。在其他红移中也可以得到类似的结果。
With the intent of correctly restoring the redshifted 21 cm signals emitted by neutral hydrogen during the cosmic reionization processes, we re-examine the separation approaches based on the quadratic polynomial fitting technique in frequency space in order to investigate whether they work satisfactorily with complex foreground by quantitatively evaluating the quality of restored 21 cm signals in terms of sample statistics. We construct the foreground model to characterize both spatial and spectral substructures of the real sky, and use it to simulate the observed radio spectra. By comparing between different separation approaches through statistical analysis of restored 21 cm spectra and corresponding power spectra, as well as their constraints on the mean halo bias b and average ionization fraction xe of the reionization processes, at z = 8 and the noise level of 60 mK we find that although the complex foreground can be well approximated with quadratic polynomial expansion, a significant part of the Mpc-scale components of the 21 cm signals (75% for ≳ 6 h−1 Mpc scales and 34% for ≳ 1 h−1 Mpc scales) is lost because it tends to be misidentified as part of the foreground when the single-narrow-segment separation approach is applied. The best restoration of the 21 cm signals and the tightest determination of b and xe can be obtained with the three-narrow-segment fitting technique as proposed in this paper. Similar results can be obtained at other redshifts.