1. 星系晕：宇宙中“失踪的重子”住在哪？
1. Galaxy halos: where do the “missing baryons” live?
把一个星系里所有看得见的恒星加起来，只占宇宙学家说的“普通物质（重子，baryons）”的一小部分。理论预计其余大部分气体弥散在星系周围几万到几十万光年的巨大晕里，温度高达百万度，称为周星系介质（circumgalactic medium, CGM）。这么热、这么稀薄的气体几乎只在 X 射线波段发光——所以要用 X 射线卫星去看。
Adding up all the visible stars of a galaxy accounts for only a small fraction of the ordinary matter (baryons) that cosmology predicts. Most of the rest is thought to lie in an enormous, million-degree, ultra-diffuse halo called the circumgalactic medium (CGM), extending tens to hundreds of thousands of light-years. Gas that hot and thin shines almost only in X-rays — hence X-ray satellites.
CGM 记录两件事：（a）星系形成时吹进晕里的金属与能量（feedback，反馈）；（b）把气体锁在位力阱里的引力（hydrostatic，静力学平衡）。温度和熵的径向分布正是区分两者的关键证据：如果气体主要被引力束缚且近似等温，熵应随半径上升；如果反馈把某方向加热，那里会局部偏热。
The CGM records two things: (a) the metals and energy that galaxy formation blew into the halo (feedback); (b) the gravity that confines the gas to the potential well (hydrostatic equilibrium). Radial temperature and entropy profiles are the key evidence separating them: a gravitationally confined, near-isothermal gas has entropy rising with radius; feedback heating a particular direction would make that direction locally hotter.
2. 为什么是 M104？
2. Why M104?
草帽星系（M104, Sombrero, NGC 4594）是一个侧向（近侧向）的 Sa 型大质量星系，距离约 9.55 Mpc（1′ ≈ 2.79 kpc），有很深的 XMM-Newton 数据（中心指向 0900170101 一次约 66 ks，加上偏置指向 0900170501/0601/0701 用于外区与背景），还有 174 ks 的 Chandra ACIS 观测（ObsID 9532+9533）。它近、亮、气体晕显著，是研究单个星系 CGM 的最佳目标之一。
The Sombrero Galaxy (M104, NGC 4594) is a massive, nearly edge-on Sa galaxy at about 9.55 Mpc (1′ ≈ 2.79 kpc), with deep XMM-Newton data (central pointing 0900170101, ~66 ks, plus offset pointings 0900170501/0601/0701 for the outer halo and background) and 174 ks of Chandra ACIS (ObsIDs 9532+9533). It is nearby, bright, and has a clear hot halo — one of the best single-galaxy CGM targets.
cognition-os/CLAUDE.md；1′↔kpc 换算来自 v18_complete_parameters.csv 的 r_mid_arcmin 与 r_mid_kpc 列（3.0′ → 9.861 kpc）
3. 这个项目具体测什么？
3. What exactly does this project measure?
项目把晕气体当作热等离子体，用 X 射线光谱测三个随半径变化的量：温度 kT（keV）、发射量归一 norm（换算成电子密度 n_e）、熵 K = kT·n_e^(−2/3)（keV cm²）。再回答：这个晕是不是统计上偏离等温？有没有方位角方向（盘侧 vs “风”轴）的温差？SZ 效应（宇宙微波背景光子被热气体逆康普顿散射）与 X 射线推出的气压是否一致？
The project treats the halo gas as a thermal plasma and measures three radial quantities from X-ray spectra: temperature kT (keV), the emission-measure normalization (converted to electron density n_e), and entropy K = kT·n_e^(−2/3) (keV cm²). It then asks: is the halo statistically distinguishable from isothermal? Are there azimuthal differences (disk side vs “wind” axis)? Do the Sunyaev–Zeldovich (SZ) effect and the X-ray-derived pressure agree?
当前最重要的结论（带标签）：
The most important current conclusions (with trust labels):
- 测得的径向温度数值是可信的数据事实：R2–4 kT = 0.715±0.025，R4–6 0.730±0.104，R6–8 0.761±0.069 keV（v18 扇区平均）。
- The measured radial temperatures are trusted data facts: R2–4 kT = 0.715±0.025, R4–6 0.730±0.104, R6–8 0.761±0.069 keV (v18 sector means).
- “内冷外热梯度”只是 Provisional：常数模型 p = 0.7097，加权斜率仅 1.37σ（D41, T133）。
- The “inner-cooler/outer-warmer gradient” is only Provisional: constant model p = 0.7097, weighted slope only 1.37σ (D41, T133).
- R4–6 方向差是唯一持续候选，但 P0 检验 NO-GO：full-nuisance 95.45% 区间 [−0.069, +0.343] keV 过零（D45）。
- The R4–6 directional contrast is the only persistent candidate, but the P0 test returned NO-GO: the full-nuisance 95.45% interval [−0.069, +0.343] keV crosses zero (D45).
cognition-os/insights/I061_temperature_profile_feedback_synthesis.md；cognition-os/04_DECISION_LOG.md D41/D45
4. 为什么值得一个本科生接手？
4. Why is it worth an undergraduate taking over?
这个项目已经把“大科学问题”（重子在哪、feedback 还是引力）翻译成了一系列可以一步步执行的测量任务：拟合一条谱、验证一个方法、画一张图。每一步都有明确的验收标准（例如温度区间两侧误差 ≤20%）、明确的停止规则（D45：不要重启 P1 反馈链）、和完整的决策日志。接手者不需要先懂天体物理前沿，只需要按本手册的“手动续做”页一步步来。
The project has already translated the “big science questions” (where are the baryons; feedback or gravity) into a sequence of concrete measurement tasks: fit one spectrum, validate one method, draw one map. Each step has a clear acceptance criterion (e.g. both sides of the temperature interval within 20%), clear stop rules (D45: do not restart the P1 feedback chain), and a complete decision log. A newcomer does not need frontier astrophysics first — just follow the “How to continue by hand” page step by step.