20 分钟研讨会提纲：M104 的热晕——压扁的气体与测不出的温差（10 张幻灯片，面向大一学生）
A 20-minute seminar outline: the M104 hot halo — flattened gas and an unmeasurable temperature contrast (10 slides, for a first-year student)
cognition-os/07_REVISION_LOG.md R107–R138 与 Synthesis 段；04_DECISION_LOG.md D148–D157；REDTEAM_QUEUE.md RT-102…RT-115；.agent/jobs/2026-10-20-p47b-morphology-corrected/REPORT.sol.md；.agent/jobs/2026-10-20-p57-dust-lane-absorption/REPORT.sol.md；.agent/jobs/2026-10-20-p58-gas-only-morphology/REPORT.sol.md；.agent/jobs/2026-10-20-p59-chandra-inner-morphology/REPORT.sol.md；.agent/jobs/2026-10-20-p60-flattening-models/REPORT.sol.md；.agent/jobs/2026-10-20-p34-b031-rebuild/REPORT.sol.md；.agent/jobs/2026-10-20-p43-hardband-2t-constraint/REPORT.sol.md；.agent/jobs/2026-10-20-p52-soft-instrument-forecast/REPORT.sol.md；.agent/jobs/2026-10-20-p50-projection-vs-2t/REPORT.sol.md；.agent/jobs/2026-10-20-p53-sz-xray-joint-2t/REPORT.sol.md；.agent/jobs/2026-10-20-p54-paper-results-draft/out/sol/results.tex；.agent/jobs/2026-10-20-p36b-forecast-redesign/REPORT.sol.md；cognition-os/references/literature_m104_hot_halo_2026-10-16.md
使用说明。每张幻灯片 = 一个观点 + 一张图或一张表 + 讲者备注。所有数字都有来源（备注里给出）；标 Provisional 的结论不要说成定论。总时长 20 分钟 ≈ 每张 2 分钟。
How to use this page. Each slide = one idea + one figure or table + speaker notes. Every number carries a source (given in the notes); do not state a Provisional conclusion as established. 20 minutes total ≈ 2 minutes per slide.
幻灯片 1 · 问题：草帽星系的百万度气体晕从哪里来？
Slide 1 · The question: where does the Sombrero's million-degree halo come from?
一个观点：M104 的热晕可能有两种命运——被星系引力束缚的气体，或者被星系风吹走的气体。要区分它们，我们测量两件事：晕的形状（沿盘还是沿极向拉长）和晕的温度走向（内冷外热还是平的）。
One idea: M104's hot halo could live two lives — gas bound to the galaxy's gravity, or gas blown outward by a wind. To tell them apart we measure two things: the halo's shape (elongated along the disk or the poles) and its temperature trend (cool inside, hot outside, or flat).
讲者备注：一张表，两列：束缚气体 vs 风的预言（形状、温度走向、判据）。文献锚点：Li+2011 的深 Chandra、Igarashi+2014 / Tsuchiya+2013 的"平 kT 分不开束缚与慢风"。
Speaker notes: a table with two columns: bound gas vs wind predictions (shape, temperature trend, test). Literature anchors: Li+2011's deep Chandra; Igarashi+2014 / Tsuchiya+2013's "a flat kT profile cannot separate bound gas from a slow wind".
幻灯片 2 · 数据：XMM-Newton 四次指向 + Chandra 深场
Slide 2 · The data: four XMM-Newton pointings + a deep Chandra field
一个观点：XMM-Newton 的 MOS1/MOS2/pn 在四个指向（0900170101/0501/0601/0701，约 390 ks 等效）给出 0.375–9.0 keV 的源谱；Chandra 1586/9532/9533（约 188.5 ks）分辨率更高、用于独立复核。中心 2′ 的点源太密，两种填洞法都在 9 个能段的独立注入检验中失败——中心被隔离，不画值。
One idea: XMM-Newton's MOS1/MOS2/pn over four pointings (0900170101/0501/0601/0701, ~390 ks effective) give 0.375–9.0 keV source spectra; Chandra 1586/9532/9533 (~188.5 ks) has higher resolution and serves as an independent check. The central 2′ is too source-crowded — both hole-filling methods fail independent injection tests in all 9 bands — so the centre is quarantined and no value is plotted.
讲者备注：图：四次指向的足迹图（或 P35 的九能段图之一）。来源：R023（拟合能段）、R107/R115/R131（中心隔离链）。
Speaker notes: figure: the four pointings' footprint (or one of P35's nine-band maps). Sources: R023 (bands actually fitted), R107/R115/R131 (the central quarantine chain).
幻灯片 3 · 背景：三样东西叠在晕信号上
Slide 3 · The background: three things sit on top of the halo signal
一个观点：每条 X 射线谱 = 晕 + 银河系软晕前景 + 未解恒星（LMXB/CV/AB）+ 宇宙背景 + 仪器粒子本底。前景约定本身把 ΔkT 从 +0.094 摆到 +0.198 keV（2.4 倍于记账的系统误差）；外部更冷的测量（eROSITA 0.1646、HaloSat 0.179 keV）会把对比度推到 ≈0 或负。
One idea: every spectrum = halo + Milky-Way soft foreground + unresolved stars (LMXB/CV/AB) + cosmic background + instrumental particle background. The foreground convention alone swings ΔkT from +0.094 to +0.198 keV (2.4× the booked systematic); cooler external measurements (eROSITA 0.1646, HaloSat 0.179 keV) push the contrast to ≈0 or negative.
讲者备注：图：背景模型的分解谱（v18 拟合会话图）。来源：R085/D128、RT-84/R088/D130、P31 重放 R091。
Speaker notes: figure: a decomposed background-model spectrum (a v18 fit-session plot). Sources: R085/D128, RT-84/R088/D130, the P31 replay R091.
幻灯片 4 · 形态学 I：内晕沿盘压扁（本工作的头条）
Slide 4 · Morphology I: the inner halo is flattened along the disk (this work's headline)
一个观点：在 2–3.5′，B0（0.4–1.25 keV）合并小轴/大轴 = 0.582±0.066——比圆（1）低 4.75/6.31σ，与 K 光基准 0.484 相差 +1.49/+1.56。背景单算的比值 ≈1，所以减背景只放大、不制造压扁。
One idea: at 2–3.5′ the combined B0 (0.4–1.25 keV) minor/major ratio is 0.582±0.066 — 4.75/6.31σ below round (1), +1.49/+1.56 from the K-light benchmark 0.484. The background-only ratio is ≈1, so background subtraction amplifies but does not create the flattening.
讲者备注：图：assets/fig/p47b_conditional_morphology.png（P47b 订正后的条件形态学图）。来源：R134、P47b REPORT.sol.md；红队 RT-112 的质疑全部被吸收。
Speaker notes: figure: assets/fig/p47b_conditional_morphology.png (P47b's corrected conditional-morphology figure). Sources: R134, the P47b report; all of red team RT-112's objections were absorbed.
幻灯片 5 · 形态学 II：这不是尘埃、不是恒星、不是仪器
Slide 5 · Morphology II: not dust, not stars, not the instrument
一个观点：三个独立检验逐个排除替代解释——尘埃改正把比值抬高 ×1.16–1.17（内禀 ≈0.50，压扁更深，P57）；扣掉恒星（小轴 8–16 %、大轴 9–13 %）后气体本身仍 0.567 [0.497,0.642]（P58）；Chandra 独立测 0.585±0.042、Chandra−XMM 差 +0.003±0.079（P59，仪器无关）。
One idea: three independent checks eliminate the alternatives one by one — the dust correction raises the ratio by ×1.16–1.17 (intrinsic ≈0.50, even more flattened, P57); after removing stars (8–16 % minor, 9–13 % major) the gas itself stays at 0.567 [0.497,0.642] (P58); Chandra independently measures 0.585±0.042 with a Chandra−XMM difference of +0.003±0.079 (P59, instrument-independent).
讲者备注：表：三个检验一行一个（改正方向、改正后的值、含义）。来源：R134（P57）、R136（P58）、R137（P59）。
Speaker notes: table: one row per check (direction of the correction, corrected value, meaning). Sources: R134 (P57), R136 (P58), R137 (P59).
幻灯片 6 · 形态学 III：物理模型——旋转/恒星注入可行，静力学与双锥风被排除
Slide 6 · Morphology III: physical models — rotating/stellar-injected gas viable, hydrostatic and bicone-wind excluded
一个观点：把模型形态对到气体 R=0.567 上：非旋转 BDH+NFW 静力学气体给 0.85–0.96（95% 外，排除）；纯小轴双锥风给 ≥1.49（95% 外，排除）；旋转束缚气体（f=0.75–1、kT 0.6–0.8 keV）与恒星质量损失注入（0.475/0.438）都能进入 68% 区间。形态学支持旋转/恒星注入的气体。
One idea: mapping model shapes onto the gas R = 0.567: non-rotating BDH+NFW hydrostatic gas gives 0.85–0.96 (outside 95 %, excluded); a pure minor-axis bicone wind gives ≥1.49 (outside 95 %, excluded); rotating bound gas (f = 0.75–1, kT 0.6–0.8 keV) and stellar-mass-loss injection (0.475/0.438) both enter the 68 % intervals. Morphology favours rotating/stellar-injected gas.
讲者备注：表：模型 → 预言 R → 是否进入 68%/95% 区间。来源：R138、P60 REPORT.sol.md；条件性（M/L、势、旋转律）要口头点名。
Speaker notes: table: model → predicted R → inside the 68 %/95 % intervals? Sources: R138, the P60 report; name the conditionality (M/L, potential, rotation law) out loud.
幻灯片 7 · 温度：外升还是不升？——测不出来
Slide 7 · Temperature: outward rise or not? — cannot be measured
一个观点：ΔkT = kT(8–10′) − kT(2–4′) 在六种约定下从 −0.105 到 +0.310 keV，95% 区间全都跨越 0；谱拟合与 Fe-L 硬度两种方法其实是同一批数据的两遍投影，不是两次独立探测。PI 裁定现有数据在 +0.2 keV 处结案，只有 |ΔkT| ≳ 0.3 keV 才可探测（D146）。
One idea: ΔkT = kT(8–10′) − kT(2–4′) spans −0.105 to +0.310 keV across six conventions, and every 95 % interval straddles 0; the two methods (spectral fitting and Fe-L hardness) are one projection applied twice to the same data, not two independent detections. The PI has ruled the existing data closed at +0.2 keV — only |ΔkT| ≳ 0.3 keV is detectable (D146).
讲者备注：图：P34 六行 ΔkT 的森林图（或表）。来源：R091/R088、P34 REPORT.sol.md、D146。
Speaker notes: figure: a forest plot (or table) of P34's six ΔkT rows. Sources: R091/R088, the P34 report, D146.
幻灯片 8 · 简并：隐藏热相 vs 外升——为什么谱分不开
Slide 8 · Degeneracy: a hidden hot phase vs an outer rise — why the spectra cannot tell
一个观点：Fe-L 复合体的形状几乎只由发射加权的平均温度决定。一团藏起来的热相混合物与一个真实的外升温升，在 M104 的信噪比下给出几乎一样的谱：P43 的 56 组混合里 0/56 被现有可观测量在 3σ 排除，单温 ΔkT 偏差 −1.56…+1.67 keV、两种符号都能偏。
One idea: the Fe-L complex's shape depends almost entirely on the emission-weighted mean temperature. A hidden hot-phase mixture and a genuine outward rise give almost identical spectra at M104's S/N: of P43's 56 mixtures, 0/56 are excluded at 3σ by any existing observable, and the single-temperature ΔkT bias runs −1.56…+1.67 keV with both signs possible.
讲者备注：图：两条几乎重合的模拟谱（混合 vs 梯度）。来源：R119、P43 REPORT.sol.md、P37/R113。
Speaker notes: figure: two nearly identical mock spectra (mixture vs gradient). Sources: R119, the P43 report, P37/R113.
幻灯片 9 · 三条路都到此为止：仪器、投影、SZ
Slide 9 · Every route stops here: instruments, projection, SZ
一个观点：没有一条现有路线能分开简并——EPIC 10000 ks 只排除 15–20/56，Resolve（阀门关闭的现实响应）与 ACIS-S 都是 0/56，RGS 0/56；投影模型选择在 M104 信噪比下失效（8/9200 与 104/400，余量 ≥2 均 0 次）；热 SZ 最大期望 z=0.091、0/56 到 3σ，且单温平板可精确复现混合的 y。教训：P44 的"0/56 排除"是响应文件伪影（RT-110/R126）——预报必须点名每一个误差项和每一个仪器响应。
One idea: no existing route separates the degeneracy — EPIC at 10000 ks excludes only 15–20/56, Resolve (realistic gate-valve-closed response) and ACIS-S are both 0/56, RGS 0/56; projection-model selection fails at M104's S/N (8/9200 and 104/400, zero with margin ≥2); thermal SZ's maximum expected z = 0.091 with 0/56 reaching 3σ, and a single-temperature slab reproduces the mixture's y exactly. Lesson: P44's "0/56 excluded" was a response-file artefact (RT-110/R126) — a forecast must name every error term and every instrument response.
讲者备注：表：仪器/路线 → 排除数 → 关键限制。来源：R126–R130、P52/P50/P53 REPORT.sol.md、Synthesis 段。
Speaker notes: table: instrument/route → number excluded → key limitation. Sources: R126–R130, the P52/P50/P53 reports, the Synthesis paragraph.
幻灯片 10 · 下一步：写下来、等 PI、留给未来
Slide 10 · Next steps: write it down, await the PI, leave room for the future
一个观点：CoPI 提案 D157（待定）：(A) 把 Q-B 写成正式的简并（负面）结果——CoPI 推荐；(B) 留给未来的软 X 射线微量热计（LEM 类 1 Ms 排除 32 组危险混合中的 9 组）；(C) 多晕 SZ 叠加。新数据设计（P36/P36b/R118）：没有任何 XMM 设计能在 +0.2 keV 真值下到 2σ——前景模型的选择本身是未被跨越的系统项。论文草稿（P54，AASTeX，未经评审）已链接在论文页。
One idea: the CoPI proposal D157 (pending): (A) write Q-B up as a formal degeneracy (negative) result — the CoPI's recommendation; (B) leave it open for a future soft-X-ray microcalorimeter (a LEM-like 1 Ms excludes 9 of the 32 dangerous mixtures); (C) stack many haloes for SZ. New-data design (P36/P36b/R118): no XMM design reaches 2σ at a true +0.2 keV — the foreground-model choice itself is the unspanned systematic. The paper draft (P54, AASTeX, not reviewed) is linked on the paper page.
讲者备注：表：三个选项 + 各自的代价与收益。来源：D157、R118/R127、Synthesis 段、P54 results.tex。
Speaker notes: table: the three options + each one's cost and payoff. Sources: D157, R118/R127, the Synthesis paragraph, P54's results.tex.