跳至正文

宇宙黎明与 21 cm 信号

目录

机制裁决第 173 篇 · 对称双向第 168 篇 · section B1 · 全库第 231 篇 · 2026-09-10 · 作者 opencode（GLM）

前景比信号亮三千倍，仪器与天空的简并空间里能画出无数种形状——2018 年那个「检测」，八年后仍以「被拒绝的最优剖面」的身份活着。这七年本身就是最好的承重测试。

本篇审的不是「21cm 宇宙学值不值得做」——它是第一批恒星与早期热历史的最有前途探针，这也不是问题。本篇审的是两句更大、传播更广的话：2018 年的「探测到宇宙第一缕星光/暗物质相互作用证据」与 2022 年的「宇宙黎明没戏了」。它接的是库内 JWST 早期星系篇（2026-08-10，审星系丰度争论——本题只在「高红移天体史有独立约束」处交叉，不重做）与热化篇的「极慢瞬态 vs 相」方法论。同族共用一把尺：把造尺人自己的论文、批评与新闻稿压在自己的叙事上承重。

先读三句红线：

  1. 本篇不裁决「第一颗恒星何时点亮」的最终答案，不评价任何在世研究者的动机——只审「宇宙黎明被看到了吗」在事件链层、理论层、实验层、名号层分别承不承重。
  2. 「EDGES 拟合剖面被拒绝」与「EDGES 数据有错」是两件不同的事；「信号未复现」与「宇宙黎明研究无进展」也是两件不同的事。本篇对称双向：既审把「3.8σ 异常」升格为「里程碑＋暗物质证据」的方向，也审把「非检测＋批评」读成「21cm 全是仪器把戏、领域没戏」的方向——两边都不许越界。
  3. 「宇宙黎明信号」不是一个问题，是四个问题共用一个词组：事件链层问「2018 那篇说了什么、被谁以什么理由质疑、回应在哪里」，理论层问「标准预测是多少、简并有多大」，实验层问「独立仪器看到了什么、自己仪器的偏差量化到哪」，名号层问「消费层相对论文原文升没升格、有没有实错」。四层答案不同，混用是全部越界的总开关。

结构胎记＝事件链跳 × 预测幅度跳 × 系统误差-天文信号跳 × 量词跳 ＋ 反向红跳。 共用动作＝把「受控的 3.8σ 异常」读成「第一颗恒星里程碑＋暗物质第二突破」（2018），或把「95.3% 拒绝最优剖面」读成「never existed」（2022）。事件链跳用 EDGES/Hills/Reply/SARAS 3 四件原文逐字称重；预测幅度跳用 Cohen 2017 与本篇复算称重；系统误差-天文信号跳用双方原文与第三方综述称重；量词跳用八份媒体原件逐字称重。

方法与纪律：全部承重事实一手取证落盘（原始论文 > 机构官方页面 > 媒体转述，分层标注 access 等级）；引语逐字保持原文语言（英文/中文各自照录）；卷期页码与 DOI 回 Crossref/Europe PMC 亲核（当场 curl；课题书凭记忆写的 2 个标识符当场证伪改正，入附录 A）；取不到的一律登记诚实空位（附录 B），不用二手填充。六捆取证原始件（理论层／EDGES 事件链／SARAS 与仪器／解释空间／名号谱系／复算锚）落盘于 evidence/2026-09-10-cosmic-dawn-21cm-edges-saras-stress-test/raw/，检索日志与通路受阻记录随件在案。关键数字（幅度比 2.39/2.08、分数吸收 0.03%、95.3%→1.99σ、z=17.2→180 Myr）全部用工具复算过，复算脚本与输出随件在案。

母裁决先行四句（每句在正文承重）：

  1. 事件链层：检测主张、批评与回应三方文本逐字都在案——EDGES（Nature 555, 67–70）逐字：”we report the detection of a flattened absorption profile … centred at a frequency of 78 megahertz and has a best-fitting full-width at half-maximum of 19 megahertz and an amplitude of 0.5 kelvin … however, the best-fitting amplitude of the profile is more than a factor of two greater than the largest predictions”；Barkana（Nature 555, 71–74）逐字 “at a confidence level of 3.8 standard deviations”；Hills 等（Nature 564, E32–E34）逐字 “their data implies unphysical parameters for the foreground emission and also that their solution is not unique … although our analysis does not prove that the feature … is absent from their data, we believe the issues … are such that the evidence for its presence falls well short of the level required to invoke new physics”；EDGES 回应（Nature 564, E35）逐字 “We broadly agree, but a general absorption profile remains the most justified a priori choice of signal model” [一手逐字·全文快照]。
  2. 理论层：预测幅度与测量难度都自有硬数字——标准预测谷深约 −170 mK（Cohen 等 2017，MNRAS 472, 1915 逐字 “∼ −170 mK in our standard case”）、193 组参数空间最大 −240 mK（逐字 “−240 mK < Tb,min < −25 mK”）；本篇复算：EDGES 500 mK 对最强标准预测（−209 mK，Fialkov–Barkana PRL 逐字 “the strongest possible absorption expected at z ∼ 17 is −209 mK”）为 2.39 倍，对 Cohen 最大为 2.08 倍——与 EDGES 摘要 “more than a factor of two” 自洽；带内前景 ~1500–1700 K（OVRO-LWA 2019）对 0.5 K 信号≈ 1:3000–3400 [一手逐字·复算]。
  3. 实验层：独立非检测与热历史旁证分线在案——SARAS 3（Nat. Astron. 6, 607–617）逐字 “their best-fitting profile is rejected with 95.3% confidence”（复算 ≈1.99σ）与 “Our non-detection bears out earlier concerns”；damped sinusoidal 系统误差被三方独立点名（Hills、Sims & Pober、Bevins，经 Shen 等 2025 逐字转述）；多波长联合（Dhandha 等 2025, MNRAS）给出全天信号谷深首个下界逐字 “−201 mK ≲ T21,min ≲ −68 mK at z_min ≈ 10–16″——EDGES 幅度在其参数化内被挤出；HERA 2023/MWA 2025 幂谱旁证（Nunhokee 2025 转述逐字 “heating of the IGM above the adiabatic cooling limit must have occurred by at least z = 10.4″）不依赖 EDGES 之争 [一手逐字·复算]。
  4. 名号层：四环升格＋一环实错＋一环降格，逐环可考——Guardian 副题逐字 “‘Revolutionary’ observations … may hold information on dark matter”＋”would mark a second major breakthrough”；科技文摘（文汇系）标题「人类首次！宇宙第一缕曙光被捕捉，或改写对暗物质认知」＋实错环逐字「”山谷”深度却比预期高两个数量级」（真实 2.08–2.39 倍——量词放大约 50 倍，复算在案）；澎湃逐字「氢原子气体大约比绝对零度高2度」vs Muñoz–Loeb 逐字 “baryons have a temperature of Tb ≈ 4 K … roughly half of its expected value”（「低一半」忠实、「2 度」与论文层不符）；2022 降格环：Sky & Telescope 逐字 “they also suggest it never existed to begin with”（高于 SARAS 95.3% 拒绝量词）、Quanta 自带边界句 “Whether there is any dip at all remains to be determined”；虚无化两组定向检索阴性（诚实空位 G4）[一手逐字·检索日志在案]。

反向红跳同样跑完：「宇宙黎明探测全无进展」不立（HERA 2023 z=10.4 加热约束、MWA 2025 最深幂谱上限、Dhandha 2025 首个谷深下界都是不依赖 EDGES 之争的硬数据）；「21cm 全是仪器把戏」不立（SARAS 3 非检测自身也量化了 2%–64% 的仪器偏差；REACH/MIST 的存在正是把系统误差显式建模的工程化回应）；「EDGES=里程碑铁案」不立（Hills 三桩＋damped sinusoidal 第三方共识＋SARAS 95.3% 拒绝）；「暗物质相互作用已被 21cm 证明」不立（Kovetz 微调约束逐字 “at least moderate fine tuning is required”；Muñoz–Loeb 只允许亚百分位微荷组分）。四句完整承重见第八章。

第一章 守真锚：21 cm 信号为什么是承重工具

1.1 物理机制的一句钉子

中性氢的超精细跃迁（静止频率 1420.405751768 MHz，波长 21.106 cm）通过碰撞与 Wouthuysen–Field 效应与气体温度耦合。红移 z 处的可观测频率为 1420.4/(1+z) MHz：EDGES 报告的 78 MHz 对应 z≈17.2（本篇复算：1420.4/18.2 = 78.0 MHz ✓，T_CMB(z=17.2)≈49.6 K）。差分亮度温度（Furlanetto–Oh–Briggs 2006 式 17–18，逐字转录）：

“δTb(ν) ≈ 27 xHI (1 + δ) (1 + z)^(1/2) [1 − Tγ/TS] mK” [一手逐字·转录]

第一代恒星的紫外辐射（Lyα）把 TS 耦合到气体温度（远低于 T_CMB）→ 氢气从 CMB 吸收 21cm 光子→全天吸收谷。Furlanetto 等 2006 参考模型的量词（逐字）：”this produces a relatively strong absorption signal (δTb ≈ −80 mK) over the range z ∼ 21–14 (or ν ∼ 70–95 MHz)”。

1.2 为什么难：两条硬数字

  • 对比度：带内前景（银河系同步辐射）在 70–78 MHz 处 ~620–1700 K（OVRO-LWA 2019 逐字转引 Mozdzen 2017 南天测量），信号 0.5 K——分数吸收 ≈0.03%（1:3000–3400，本篇复算）。
  • 简并度：Cohen 等 2017 用 193 组天体物理参数组合扫出全天信号（摘要逐字：”The early star-forming Universe is still poorly constrained … We show that the expected signal fills a large parameter space, but with a fixed general shape for the global 21-cm curve”）——预测本就充满简并空间。

第二章 事件链层：检测、批评与回应的三方文本

2.1 检测主张（EDGES, 2018-02-28）

Bowman, Rogers, Monsalve, Mozdzen & Mahesh（Nature 555, 67–70，Europe PMC PMID 29493587）。摘要逐字：

“After stars formed in the early Universe, their ultraviolet light is expected, eventually, to have penetrated the primordial hydrogen gas and altered the excitation state of its 21-centimetre hyperfine line. … Here we report the detection of a flattened absorption profile in the sky-averaged radio spectrum, which is centred at a frequency of 78 megahertz and has a best-fitting full-width at half-maximum of 19 megahertz and an amplitude of 0.5 kelvin. The profile is largely consistent with expectations for the 21-centimetre signal induced by early stars; however, the best-fitting amplitude of the profile is more than a factor of two greater than the largest predictions. This discrepancy suggests that either the primordial gas was much colder than expected or the background radiation temperature was hotter than expected. Astrophysical phenomena (such as radiation from stars and stellar remnants) are unlikely to account for this discrepancy; of the proposed extensions to the standard model of cosmology and particle physics, only cooling of the gas as a result of interactions between dark matter and baryons seems to explain the observed amplitude.” [一手逐字]

仪器（扩展数据图注逐字）：”The low-1 antenna with the 30 m × 30 m mesh ground plane. The darker inner square is the original 10 m × 10 m mesh.”——平天线＋大面积网格地平面；波束色差正是后续批评的焦点之一。

2.2 Barkana 的配套件（同刊姊妹）

Barkana（Nature 555, 71–74）摘要逐字：”The recent detection of the global 21-centimetre spectrum reveals a stronger absorption than the maximum predicted by existing models, at a confidence level of 3.8 standard deviations. Here we report that this absorption can be explained by the combination of radiation from the first stars and excess cooling of the cosmic gas induced by its interaction with dark matter. … the dark-matter particle is no heavier than several proton masses, well below the commonly predicted mass of weakly interacting massive particles. … These results indicate that 21-centimetre cosmology can be used as a dark-matter probe.”

注意两处自带分寸：①3.8σ 的口径是「相对现有模型最大预测的偏离」（Barkana 原文），不是「检测显著性」；②Barkana 在同月的 PRL 里自己把 flat-bottom 形状的显著性降格（捆 4.3：’the significance of the flattening is unclear given the level of systematic error’）——暗物质解释的焦点是幅度，形状分歧连提议者自己都挂起。

2.3 Hills et al 2018：三桩批评与自我设限的结论

Hills, Kulkarni, Meerburg & Puchwein（Nature 564, E32–E34，2018-12-19，Brief Communications Arising；全文快照落盘）。逐字：

“We have examined the modelling process they used and find that their data implies unphysical parameters for the foreground emission and also that their solution is not unique, in the sense that we found other simple formulations for the signal that are different in shape but that also fit their data. We argue that this calls into question the interpretation of these data as an unambiguous detection of the cosmological 21-cm absorption signature.” “We found that at least five free foreground parameters, in addition to the four absorption profile parameters, were always needed to obtain a good fit. The parameters found for the profile changed substantially when different formulations for the foreground were used.” “Adding higher-order foreground terms has simply moved the problem of unphysical parameter values from the ionosphere to the foreground.” “a good fit can also be achieved with two Gaussian emission features of equal height and width … a good fit was obtained with just a sine wave … the residuals take the form of undulations with a period of around 12.5 MHz” “Since the proposed 21-cm absorption profile does not match theoretical expectations in either shape or amplitude, it is not clear why it should be preferred to the other forms of signal explored here or to the many more that can be found in the degenerate space between signal and foreground model.” “Therefore, although our analysis does not prove that the feature identified by Bowman et al. is absent from their data, we believe the issues that we have raised are such that the evidence for its presence falls well short of the level required to invoke new physics for its explanation.” [一手逐字]

数据来源逐字：共用 EDGES Data Release（”Data for Figure 1 of Bowman et al. (2018)”，http://loco.lab.asu.edu/edges/edges-data-release/）。结论的自我设限是名号层称重的秤：批评的量词是「不足以支撑新物理」，不是「信号不存在」。

2.4 EDGES 回应：承认非唯一性＋子带/全带之辩

Bowman, Rogers, Monsalve, Mozdzen & Mahesh，”Reply to Hills et al.”（Nature 564, E35；全文快照落盘）。逐字：

“We broadly agree, but a general absorption profile remains the most justified a priori choice of signal model because we have disfavoured the instrument as the source of the structure and there is no known physical expectation for other shapes in either the foreground or 21-cm signal, whereas an absorption is expected. We have data that exclude some of the alternative signal models proposed by Hills et al. and plan to publish those results in the near future.” “In EDGES we do aim to measure a fully absolutely calibrated spectrum. Although in our Methods section we acknowledged potential residual calibration effects, we reported tests to show that any such effects are not consistent with the reported profile. We therefore concluded that the signal is astronomical.” “When using our polynomial foreground model over the full band (51-99 MHz), rather than over only the sub-band for which we used it (approximately 63-99 MHz), Hills et al. recovered best-fit profiles that are not consistent with our reported properties. We have shown using simulations that this outcome is consistent with the expected performance of that model. Therefore, their choice to use it over the full band was not justified.” [一手逐字]

事件链小结：双方共用同一组公开数据；批评的三桩（不物理参数/不唯一解/简并空间）在回应里未被逐一反驳，而是被「先验选择」论证接住；EDGES 承诺的排除性发表（’plan to publish those results in the near future’）在本篇检索范围内未定位到（诚实空位 B3）。

第三章 理论层：预测幅度与模型简并

3.1 标准预测的数字

Cohen, Fialkov, Barkana & Lotem 2017（MNRAS 472, 1915–1931，arXiv 全文落盘）逐字：

“the signal reaches its local minimum value, ∼ −170 mK in our standard case. … In our standard case the minimum point occurs at zmin = 18 with a depth of Tb,min = −170 mK, variation of the astrophysical parameters leads to a range in redshift 10.9 < zmin < 26.5 and temperatures −240 mK < Tb,min < −25 mK.” [一手逐字]

Fialkov & Barkana 2018 PRL（121, 011101，快照全文落盘）逐字：”In the standard cosmological scenario, the strongest possible absorption expected at z ∼ 17 is −209 mK, which corresponds to a gas temperature of T_gas ∼ 7 K.”——「标准宇宙学最强吸收」的另一套口径。

3.2 幅度对账（本篇复算）

  • EDGES 500 mK / Cohen 最大 −240 mK = 2.08；/ Fialkov–Barkana −209 mK = 2.39——两套「最大预测」口径都支持 EDGES 摘要的 “more than a factor of two”。
  • 换算：z=17.2 对应宇宙年龄 ~178 Myr（13.8 Gyr/(1+z)^1.5 粗口径，捆 6 复算）——与媒体通用的 “180 million years” 自洽。

3.3 简并的命门地位

Hills 的批评之所以有正典地位，因为它把 EDGES 的争点压到了模型简并上（2.3 逐字）。Cohen 2017 的参数空间扫描是同一枚硬币的正面：即使没有系统误差，全天信号的可预测形状本就充满简并（’fixed general shape’＋强简并的参数映射）。这正是 Bowmac Reply 用 ‘most justified a priori choice’（先验形状选择）回应 Hills 的原因——在简并空间里，结论取决于你相信哪种先验。

第四章 实验层（上）：SARAS 3 与独立非检测

4.1 SARAS 3（2022-02-28）

Singh 等（Nat. Astron. 6, 607–617）。摘要逐字：

“We report a radiometer measurement of the spectrum of the radio sky in the 55–85 MHz band, which shows that the profile found by Bowman et al. in data taken with the Experiment to Detect the Global Epoch of Reionization Signature (EDGES) low-band instrument is not of astrophysical origin; their best-fitting profile is rejected with 95.3% confidence. The profile was interpreted to be a signature of the cosmic dawn; however, its amplitude was substantially higher than that predicted by standard cosmological models. Our non-detection bears out earlier concerns and suggests that the profile found by Bowman et al. is not evidence for new astrophysics or non-standard cosmology.” [一手逐字]

复算：95.3% 双侧置信 ≈ 1.99σ（MIST 状态稿转述口径 ‘approximately 2σ’ 自洽——本篇复算在案）。仪器设计（Quanta 2022 逐字）：”The antenna — an aluminum cone propped on a Styrofoam raft — was floated in the middle of a placid lake … cutting their data collection short and returning to Bangalore hours before the first citywide COVID lockdown began.”——水体均匀性压地面分层系统误差；同篇 Parsons 引语逐字 “I don’t see a lot of wiggle room there” 与自带边界句 “Whether there is any dip at all remains to be determined; Parsons emphasizes that the SARAS team will need to do more work to understand the subtleties of their own measurement.”

4.2 SARAS 3 自查：非检测同样有仪器问题

Agrawal, Kavitha & Singh 2024（ApJ，快照逐字）：”We find a correlation between the nature of chromaticity in antenna properties and the bias in the recovered spectral profiles … The bias ranges from 2% to as high as 64%.”——非检测方自己的仪器偏差账在案。MIST 的无地平面设计自报土壤依赖（URSI 2025 状态稿逐字：”operates without a ground plane … the lack of a ground plane results in an antenna beam that is highly sensitive to the local electrical properties of the soil”）——所有全天实验都在与同一类系统误差搏斗。

4.3 第三方系统误差共识：damped sinusoidal

Shen 等 2025（MNRAS，快照逐字）：

“re-examinations (Hills et al. 2018; Singh & Subrahmanyan 2019) on their methods have raised concerns regarding the non-physical parameters, the non-uniqueness of their solution, and potentially unaccounted-for systematic structures in the data. Moreover, Hills et al. (2018), Sims & Pober (2020) and Bevins et al. (2021) have shown that a damped sinusoidal systematic is strongly preferred in the EDGES data.” [一手逐字·转述]

三件独立工作（Sims & Pober 2020、Bevins et al 2021 未取全文，登记空位）指向同一个系统误差形状——这是「系统误差侧」的第三方共识性陈述，比 Hills 单家批评重一档。

第四章 实验层（下）：解释空间与热历史旁证

4.1 两类解释的正典分类

Monsalve 等 2018（ApJ，快照逐字）：”Theoretical models have already been proposed to explain a deep absorption, which either invoke a stronger radiation background (e.g., Feng & Holder 2018; Ewall-Wice et al. 2018) or a colder intergalactic medium (e.g., … Barkana 2018; … Muñoz & Loeb 2018) during cosmic dawn.” OVRO-LWA 2019 同构分类（逐字见捆 2）。

4.2 重子-暗物质路线：从提出到被收紧

三步链逐字在案：

  1. 提出（Barkana 2018 Nature＋Muñoz–Loeb 2018 Nature 557, 684）：”if minicharged particles comprise a subpercent fraction of the dark matter, and have charges ∼ 10^-6 —in units of the electron charge—and masses mχ ∼ 1−60 MeV, they can significantly cool down the baryonic fluid”；”baryons have a temperature of Tb ≈ 4 K at z ≈ 17, roughly half of its expected value“。
  2. 约束（Kovetz 等 2018 PRD 98, 103529，快照逐字）：”only a fraction (mχ/MeV)×0.0115% ≲ f ≲ 0.4% of the entire DM content, and only for DM-particle masses between 0.5 MeV–35 MeV, can be charged at the level needed to marginally explain the anomaly”；”at least moderate fine tuning is required”。
  3. 涨落预言（Fialkov–Barkana PRL 逐字）：”consistent models predict a 21-cm fluctuation during cosmic dawn that is between 3 and 30 times larger than the largest previously expected”；”The expected power spectrum exhibits strong baryon acoustic oscillations … a conclusive evidence of the velocity-dependent scattering and could be used by interferometers to verify the dark matter explanation”——检验判据已被写下（幂谱 BAO 形状），未被执行。

4.3 2025：多波长联合的首个谷深下界

Dhandha 等 2025（MNRAS，快照逐字）：

“This includes high-redshift UV luminosity functions (UVLFs) from HST and JWST, cosmic X-ray and radio backgrounds (CXB and CRB), the SARAS 3 global 21-cm signal non-detection, and HERA 21-cm power spectrum upper limits. … We infer an absorption trough of depth −201 mK ≲ T21,min ≲ −68 mK at z_min ≈ 10–16, and a power spectrum of 8.7 mK² ≲ Δ²21(z = 15) ≲ 197 mK² at k = 0.35 hMpc⁻¹. … the results depend on our assumption of a redshift-independent X-ray and radio efficiency of galaxies, and the exclusion of a flexible model for Population III stars.” [一手逐字]

命门层的关键对账：EDGES 的 −500 mK 在该 95% C.I. 上端（−201 mK）之外——多波长联合在其参数化内把 EDGES 幅度挤出允许带；同时该联合给出了谷深的首个下界（”enable the first lower bound on the cosmic 21-cm signal”，逐字）——即信号应该存在但比 EDGES 浅。这就是「EDGES 事件」留给领域的净产出：一个被收紧的期望值区间。

4.4 与 EDGES 无关的热历史旁证

Nunhokee 等 2025（ApJ 989, 57，快照逐字）：

“These results represent the deepest upper limits achieved by the MWA to date and provide the first evidence of the heated intergalactic medium at redshifts z = 6.5 to 7.0.” “The best 21 cm upper limits … have been reported by HERA … The latest findings indicate that heating of the IGM above the adiabatic cooling limit must have occurred by at least z = 10.4, effectively ruling out a wide range of ‘cold-reionization’ models.” [一手逐字·转述]

分线声明：幂谱上界与全天谷深是不同统计量；热历史结论（早期已有加热）不依赖 EDGES 之争的胜负——这是「虚无化 21cm 方向」的直接反证。

第五章 命门：什么能终结这场分歧

命门不是「再争论拟合方法」，而是三条已被写下的判据：

  1. 独立复现或独立排除：REACH（双天线＋现场校准器＋贝叶斯联合拟合，Nat. Astron. 6, 984 摘要逐字：”The novel experimental approach focuses on detecting and jointly explaining these systematics together with the foregrounds and the cosmological signal using Bayesian statistics”）、MIST、PRIZM 等已部署在采（Dhandha 2025 谱系句逐字：”many already deployed and collecting data”）。REACH 出数据的速度与质量是下一个事件。
  2. EDGES 侧承诺的排除性数据：Reply 逐字 “We have data that exclude some of the alternative signal models proposed by Hills et al. and plan to publish those results in the near future“——本篇检索未定位到该发表（诚实空位 G2）；若发布，简并空间将被收紧。
  3. 幂谱 BAO 检验（Fialkov–Barkana PRL 逐字 “conclusive evidence”）——判据在案、判别实验未做；HERA/SKA 若在宇宙黎明波段测到速度依赖散射的 BAO 形状，暗物质路线获正典检验。

第六章 名号层：三条链逐环称重

6.1 链 A：2018 升格链

A0 论文锚：EDGES 摘要的受控表述（’more than a factor of two’＋’Astrophysical phenomena … unlikely to account’＋’only cooling … seems to explain’——限定在 proposed extensions 之内）。

A1 Guardian（Hannah Devlin）：标题逐字 “Cosmic dawn: astronomers detect signals from first stars in the universe”；副题逐字 “‘Revolutionary’ observations suggest the first stars appeared 180m years after the big bang – and may hold information on dark matter”；正文升格点逐字 “The faint imprint left by the glow of the earliest stars also appears to contain new and unexpected evidence about the existence and nature of dark matter which, if confirmed by independent observatories, would mark a second major breakthrough.”——「第二重大突破」框架＋if confirmed 条件句。

A2 BBC（Jonathan Amos）：Bowman 引语逐字 “This is the first time any team has been able to present evidence for the detection of this signal and hopefully it will go down as a milestone”；第三方谨慎引语在案（de Bernardis “If confirmed, it’s a breakthrough”；Glazebrook “astronomers worldwide would now be holding their breath until the result was confirmed by an independent experiment”）。

A3 ABC：数字升格候选逐字 “the signal was twice as strong as it should be. This indicated the hydrogen gas in the early universe was around -270 degrees C — much colder than expected.”——-270 °C≈3 K 与论文层口径（Muñoz–Loeb Tb≈4 K；Fialkov–Barkana T_gas<5.1 K）不符，温度换算粗糙，登记待核（B4）。

A4 科技文摘（文汇系）2018-03-02：标题逐字「人类首次！宇宙第一缕曙光被捕捉，或改写对暗物质认知」；实错环逐字「但是”山谷”深度却比预期高两个数量级」——EDGES 摘要量词是 “more than a factor of two”；本篇复算 500/240=2.08、500/209=2.39——「两个数量级」（100 倍）比真实差距夸大约 50 倍，是本链最重的实体错误。另一环：把 EDGES 的条件性表述读成「他们提出只有暗物质与重子的相互作用才能带走更多能量」——丢掉 ‘of the proposed extensions to the standard model’ 的限定域。同文自留「”山谷”出现的位置与理论模型符合得不错」。

A5 澎湃新闻 2018-03-01：主体转述忠实（”78兆赫对应大爆炸后约1.8亿年”＋Bowman 引语）；一处数字疑点逐字「氢原子气体大约比绝对零度高2度，比此前的理论预期要低一半」——「低一半」与 Muñoz–Loeb 逐字（Tb≈4 K，约为预期一半）一致；「2 度」与论文层 Tb≈4 K 不符（疑为换算/抄写误差），登记名号层一格。

A6 科学网 2018-03-02：转述克制（幅度「更大（吸收更深）」与暗物质质量句忠实）。

6.2 链 B：2022 降格环

B1 Sky & Telescope：逐字 “the results from SARAS 3 did, and they not only don’t confirm the signal, but they also suggest it never existed to begin with“——”never existed” 高于 SARAS 95.3%（≈2σ）拒绝量词一档以上；同篇转述 SARAS 团队把 EDGES 信号归因于 “unevenly distributed ground beneath the detector”（地面分层解释——登记为 SARAS 团队转述）。

B2 Quanta（Ben Brubaker）：标题 “In New Experiment, Astronomers See No Sign of ‘Cosmic Dawn'”——非检测方立场为主，但自带边界句（4.1 已逐字）；Parsons “I don’t see a lot of wiggle room there” 为强立场引语。

B3 MIST 状态稿（2025）：中性总结句逐字 “The EDGES detection was challenged in 2022 by the SARAS-3 team, which rejected the EDGES result with approximately 2σ confidence. The competing results motivates the need for independent teams to apply unique approaches to resolve the tension.”——第三方状态的准确表述。

6.3 虚无化方向：两组定向检索，消费层无靶

查询词与结果在捆 5 笔记（「EDGES artifact 假信号 中文」等→无消费层命中）。注意一个反直觉分寸：Hills 的 “falls well short” 是学术层的对称内容（受控表述），消费层若把它断章为「全是仪器把戏」将是反向升格；本轮未找到具体消费文本（阴性，空位 G4）。本题消费层失真集中于 2018 升格方向（含 50 倍量级放大实错）＋2022 的 “never existed” 一环。

第七章 对称红队：四句双向称重

7.1 升格方向

  • 「宇宙黎明已被探测/第一颗恒星的里程碑已立」——不立（作为已证命题）。 EDGES 自称 detection 但被独立 95.3% 拒绝＋damped sinusoidal 三方点名＋2025 下界挤出；「首次」是事件性描述，不是结论。
  • 「暗物质相互作用被发现了」——不立。 Barkana 的 3.8σ 是幅度-模型偏离口径；微荷路线被压到 0.4% 上限＋fine tuning 逐字；检验判据（幂谱 BAO）未执行；且 EDGES 前提本身未复现。
  • 「SARAS 已终结争论」——不立（在 95.3% 量词之外）。 SARAS 自身 bias 2%–64%；Parsons 的边界句；REACH/MIST 在采——争论的结构是「多方仪器正在与系统误差搏斗」。

7.2 虚无方向

  • 「21cm 全是仪器把戏/什么也没测出」——不立。 HERA 2023 的 z=10.4 加热约束、MWA 2025 最深幂谱限、Dhandha 2025 首个谷深下界——都是不依赖 EDGES 之争的热历史硬数据；SARAS 3 的非检测本身也是一次可复现的测量。
  • 「EDGES 数据就是错的」——超出一手文本。 Hills 自我设限（’does not prove absent’）；damped sinusoidal 是「强烈偏好」的统计陈述，不是仪器故障证明；EDGES 回应的仪器排除测试与子带批评（全带误用多项式）在案。
  • 「多波长下界说明 SARAS 才是对的」——未到量词。 Dhandha 2025 的下界依赖其模型参数化（自报 caveat：红移无关效率＋无灵活 Pop III）——它是收紧，不是判决。

7.3 红队净账

升格方向的最强真核：EDGES 0.5 K 谷是真实报告的测量（3.8σ 口径、双源可查）、暗物质解释曾给出可检验判据、领域投入了 REACH/MIST 等独立力量。虚无方向的最强真核：95.3% 拒绝、damped sinusoidal 三方点名、EDGES 幅度被 2025 下界挤出、暗物质路线 fine tuning。两边真核不冲突，因为它们落在不同的层。越界的从来不是某一层的结论，而是把某一层的结论搬进另一层：把「幅度异常」搬成「暗物质发现」，把「最优剖面被拒绝」搬成「信号从未存在」，把「两个数量级」式的换算错误留在转载链里。

第八章 结论及其适用边界

  1. 事件链裁决：EDGES 检测主张与其自带量词（0.5 K、>2×、3.8σ）逐字在案；Hills 三桩批评与 EDGES 回应的让步面（’broadly agree’＋先验选择承认）逐字在案。适用边界：两件 Brief Communications 共用公开数据，无新观测。
  2. 理论层裁决：标准预测谷深 −170/最大 −240 mK（Cohen）、−209 mK（Fialkov–Barkana）；EDGES 对应 2.08–2.39 倍（复算自洽）；信号在前景下 0.03% 量级；参数空间简并由 Cohen 193 模型自证。适用边界：半数值模拟的参数覆盖；「最大预测」两套口径。
  3. 实验层裁决：SARAS 3 独立非检测（95.3%≈1.99σ 拒绝最优剖面）；damped sinusoidal 系统误差三方点名；多波长联合首个下界 −201 mK（其参数化内排除 EDGES 幅度）；热历史旁证（HERA z=10.4、MWA 深限）独立于全局信号之争。适用边界：SARAS 自身 bias 2%–64% 在案；下界依赖模型参数化（自报）。
  4. 解释空间裁决：重子-暗物质路线被 CMB/BBN/恒星冷却压到 0.0115%×(mχ/MeV)–0.4% 且需 fine tuning；额外射电背景线登记未精读；检验判据（幂谱 BAO）在案未执行。适用边界：全部以 EDGES 幅度为真为前提。
  5. 名号层裁决：2018 升格（revolutionary/second breakthrough/人类首次）＋实错环（两个数量级）＋数字疑点（2 度）＋2022 降格环（never existed）逐环可考；虚无化消费层阴性。适用边界：只覆盖已落盘 10 份文本与检索日志。

第九章 仍不确定什么，什么证据会改变判断

仍不确定（按承重排序）：

  1. EDGES 结构的天文性。damped sinusoidal（Sims&Pober/Bevins）与仪器诊断未能被第三组复现检验。改变判断的证据：REACH/MIST 2026–2027 数据；或 EDGES 侧兑现 Reply 承诺发布备择模型排除数据（诚实空位 G5）。
  2. SARAS 3 的非检测强度。其自身 2%–64% 偏差账与湖泊环境的定向效应在案；若 SARAS 团队完成对自身测量的进一步理解（Quanta/Parsons 逐字），95.3% 的量词可能上移或下移。
  3. 若信号为真，新物理解释是否还有活路。微荷窗口 0.4% 上限＋fine tuning 逐字在案；2025 下界若被更强数据（JWST UVLF 更新）确认，EDGES 幅度空间进一步收窄；反之若排除灵活 Pop III 的假设被打破，窗口重开。
  4. 暗物质检验判据的执行。幂谱 BAO 形状检验（PRL 逐字 “conclusive evidence”）若被 HERA/SKA 在宇宙黎明波段执行，将同时裁决 EDGES 与暗物质解释。
  5. 名号层空白侧稳定性。虚无化消费层当前空位；若出现有影响力的「21cm 探测全是失败」类文本，第六章的不对称结论需复查。

第十章 关键来源与核验范围

理论层：Furlanetto–Oh–Briggs 2006（Phys. Rep. 433, 181–301，DOI 10.1016/j.physrep.2006.08.002，arXiv:astro-ph/0608032 全文落盘，A）；Cohen 等 2017（MNRAS 472, 1915–1931，DOI 10.1093/mnras/stx2065，arXiv:1609.02312 全文落盘，A）。

事件链：Bowman 等 2018（Nature 555, 67–70，DOI 10.1038/nature25792，摘要＋扩展数据图注逐字，◐——付费正文未取，摘要级承重）；Barkana 2018（Nature 555, 71–74，DOI 10.1038/nature25791，摘要 ◐）；Hills 等 2018（Nature 564, E32–E34，DOI 10.1038/s41586-018-0796-5，全文快照 A）；Bowman 等 2018 Reply（Nature 564, E35，DOI 10.1038/s41586-018-0797-4，全文快照 A——MIT Haystack 记 E32–E34 与 Crossref E35 两处口径不一，正文按 DOI 引）。

SARAS 与仪器：SARAS 3（Nat. Astron. 6, 607–617，DOI 10.1038/s41550-022-01610-5，摘要 ◐）；SARAS 2（ApJ 858, 54，DOI 10.3847/1538-4357/aabae1，条目亲核 B）；REACH 2022（Nat. Astron. 6, 984–998，DOI 10.1038/s41550-022-01709-9，快照 A）；SARAS 2024 定向效应（ApJ，DOI 10.3847/1538-4357/ad6d68，快照 ◐）；MIST 2025 状态稿（URSI，快照 ◐）；OVRO-LWA 2019（AJ，DOI 10.3847/1538-3881/ab2629，快照 ◐）；Nunhokee 2025（ApJ 989, 57，DOI 10.3847/1538-4357/adda45，快照 ◐）；Wilensky 2023（ApJ，DOI 10.3847/1538-4357/acffbd，登记）。

解释空间：Muñoz–Loeb 2018（Nature 557, 684–686，DOI 10.1038/s41586-018-0151-x，arXiv 快照 A）；Fialkov–Barkana 2018（PRL 121, 011101，DOI 10.1103/PhysRevLett.121.011101，快照 A）；Kovetz 等 2018（PRD 98, 103529，DOI 10.1103/PhysRevD.98.103529，快照 A）；Monsalve 2018（ApJ，DOI 10.3847/1538-4357/aace54，快照 ◐）；Dhandha 2025（MNRAS，DOI 10.1093/mnras/staf1736，快照 ◐）；Shen 2025（MNRAS，DOI 10.1093/mnras/staf678，快照 ◐）；Ewall-Wice 2018（arXiv:1803.01815）与 Feng & Holder 2018（arXiv:1802.07432）登记未读。

名号层（9 份原件落盘，逐字引于第六章）：Guardian 2018-02-28；BBC 2018-02-28；ABC 2018-03-01；Quanta 2022-02-28；Sky & Telescope 2022-02-28（快照级）；澎湃 2018-03-01；科技文摘 2018-03-02；科学网 2018-03-02；NatGeo 中文 2018-03-19（快照级）。

核验范围声明：全部书目经 Crossref/Europe PMC 逐条亲核（卷期页码以 Crossref 为准）；课题书 2 处凭记忆标识符当场证伪改正（SARAS 3 DOI 尾段、Cohen arXiv 号）；关键复算（幅度比、分数吸收、σ 换算、z→年龄）脚本与输出落盘；Nature 付费正文未取（EDGES/Barkana 摘要级承重）；Hills/Reply 全文经出版商页快照逐字；HERA 2023 DOI 未单独亲核（数字两处转述锚定）。

附录 A 课题书勘误登记（2 处）

  1. SARAS 3 DOI：课题书初稿猜 10.1038/s41550-022-01610-w——Europe PMC/nature.com 亲核为 s41550-022-01610-5。
  2. Cohen 2017 arXiv 号：初稿猜 1703.07398（实为 SU(2) Yang–Mills 论文）——经 Crossref 书目检索定位 arXiv:1609.02312。

附录 B 诚实空位登记

  • G1（EDGES 侧承诺件未定位）：Reply 逐字 “We have data that exclude some of the alternative signal models … plan to publish those results in the near future”——本篇检索未定位到该发表；若已发表需复查简并空间收紧情况。
  • G2（虚无化消费层空位）：「21cm 探测全是仪器把戏/宇宙黎明没戏」方向的消费层具体文本，两组定向检索未命中（捆 5 笔记）。
  • B3（Sims & Pober 2020、Bevins et al 2021 未取全文）：damped sinusoidal 主张经 Shen 2025 逐字转述锚定；不引未读件内容。
  • B4（ABC 温度换算疑点未核）：’-270 °C’ 与论文层口径（Muñoz–Loeb Tb≈4 K；Fialkov–Barkana T_gas<5.1 K）不符；原始换算路径未查，登记待核。
  • B5（EDGES/Barkana 付费正文未读）：摘要级承重；EDGES 的 Methods 细节（原始数据量、观测天数）未核对；引用全部以摘要与扩展数据图注为限。
  • B6（HERA 2023 DOI 未单独亲核）：z=10.4 数字经 Nunhokee 2025 与 Dhandha 2025 两处逐字转述锚定。
  • B7（Dhandha 2025 为快照级）：MNRAS 出版版全文未取；下界数字与其自报 caveat 以快照为限。
  • B8（复算的 z→年龄为量级锚）：13.8 Gyr/(1+z)^1.5 为 Ωm=1 粗口径（178 Myr vs 媒体 180 Myr 自洽）；不用于精确宇宙学断言。
  • B9（Morris 等后续 EDGES 重分析 2020–2024 线未穷尽）：EDGES 数据的再分析文献谱系（Tauscher 等）未逐件取证；以 Shen 2025 转述的三方点名承重。
  • B10（LOFAR 2025 幂谱件未单独核）：Dhandha 2025 引 LOFAR (Mertens et al 2025; Ghara et al 2025)；未取 DOI，热历史旁证以 HERA/MWA 两条亲核线承重。
标签: