Seminar — M31's 0.6–0.8 keV emission: detected in the halo fields, origin still open研讨会——M31 的 0.6–0.8 keV 发射：晕场已检出，起源仍然开放
Offline 16:9 slide deck for a 45-minute seminar + 10–15 minutes of questions, for graduate students and professors who do not know the project. About one slide per minute (30–40 main slides), plus backups. Every content slide carries a figure or table as evidence; the full talk script is below each slide with cumulative time. Assertion headlines throughout.
面向不了解项目的研究生和教授，离线 16:9 幻灯片，45 分钟研讨会 + 10–15 分钟提问。约一张幻灯片一分钟（30–40 张主幻灯片），加备用幻灯片。每张内容幻灯片都以图表为证据；完整的报告文稿在每张幻灯片下方，带累计时间。全程用断言式标题。
Backup slides备用幻灯片
Numbered B1–B6, expected in questions: systematics, alternative explanations, why not method X, how general the result is.
编号 B1–B6，预期会在提问中出现：系统误差、替代解释、为什么不用方法 X、结果的普遍性。
| Slide幻灯片 | Content内容 | Source来源 |
|---|---|---|
| B1B1 | Full run6 table: 33 ObsIDs, 490 bundles, 252 ObsID-region combinations完整 run6 表：33 个 ObsID，490 个包，252 个 ObsID-区域组合 | claims.yaml `run6-complete`; 02claims.yaml `run6-complete`；02 |
| B2B2 | nH: all 8 annuli preferred 0.0585e22 (\(\Delta\chi^2\) −27.8 to −218.6), but the preference is now Questioned — with the band below 0.5 keV removed the reduced refits prefer the frozen 0.03 (\(\Delta\chi^2\) 3.79 and 16.73), so it most likely reflects sub-0.5 keV calibration or soft-sky modelling rather than the column (R116, R126). Soft calibration is not excluded: a 2–3 % energy-dependent residual moves a fitted N_H by ~(0.9–2.5)e20 cm\(^{-2}\)nH：全部 8 个环带曾偏好 0.0585e22（\(\Delta\chi^2\) −27.8 到 −218.6），但该偏好现在受质疑——移除 0.5 keV 以下波段后，约化重拟合偏好冻结的 0.03（\(\Delta\chi^2\) 3.79 与 16.73），因此它最可能反映 0.5 keV 以下的校准或软天空建模，而非柱密度本身（R116、R126）。软波段校准并未被排除：2–3 % 的能量依赖残余会把拟合的 N_H 移动 ~(0.9–2.5)e20 cm\(^{-2}\) | R016, R116; Lit L02R016、R116；Lit L02 |
| B3B3 | WISE sky pedestal: W1 off-disk 6.87 DN = 53–97% of template; W4dust ~99% (B026)WISE 天空底座：W1 场外 6.87 DN = 模板的 53–97%；W4dust ~99%（B026） | B026B026 |
| B4B4 | Conditional vs profiled errors: \(\times\)2.84 / \(\times\)1.96 growth needed for chi^2/dof = 1条件性 vs profile 误差：为使 chi^2/dof = 1 需要 \(\times\)2.84 / \(\times\)1.96 增长 | R008 R008 |
| B5B5 | E033 branch B/C withdrawn: non-nested; B froze sky, C added 84 soft-proton params. Replaced by the registered halo-field detection (q = 22.4 under rule v2, 34.84 after warm restarts)E033 分支 B/C 撤回：非嵌套；B 冻结天空，C 加 84 个软质子参数。其后是已注册的晕场检出（规则 v2 下 q = 22.4，热重启后 34.84） | R003/R004, R071, R112R003/R004、R071、R112 |
| B6B6 | V11 blocked: no 90% stellar completeness or ECF closure; H1 stays source-selection-dependentV11 阻塞：无 90% 恒星完备度或 ECF 闭合；H1 仍依赖源选择 | E043; REPORT STATUS: BLOCKEDE043；REPORT STATUS: BLOCKED |
B8 — The age-resolved emissivity is still not established, and we now know why the first attempt looked wrong. The certified response is band-independent (no double vignetting) and free per-file floors return the old-age emissivity to zero, which confirms that the floors can absorb smooth old light — but the physical gate still fails because the QPB/SP degeneracy is unresolved (R125). The regression is now on its third generation and still not established: D1c fixed the QPB per file from the detector corners but its support rule removed 605/622 regions (13 regions left, 0/24 primary cases pass), so it is formal only — full-sample old \(\epsilon\) 1.99 [1.72, 2.25]e27 per formed M☉, ratio 82, dev/dof 1.74, QPB pull 8.4 — and D1d is running at pixel level with an absolute corner→FOV QPB (R129, R130). The soft QPB scale of 2.7 is not a calibration error: it is the 1.4–2.0 keV Al K\(\alpha\) + Si K\(\alpha\) line band missing from the run6 continuum template, 64.8 % of the corner counts there, with a corner/template ratio of 3.71 against 0.938 and 1.014 in the soft sub-bands (R127). The background practice for the next attempt is now written down: sky as one physical spectral model anchored by off-disk controls, QPB, soft protons and the Al/Si lines as separate detector terms, integer-count Poisson, deviance calibrated by simulation, and the old coefficient profiled jointly with sky/QPB/SP (R128). (Source: 07 R125, R127, R128.)
B8 —— 按年龄分辨的发射率仍未确立，而我们现已知道首次尝试为何看起来不对。已认证响应与波段无关（无双重消光），自由的逐文件底板把旧年龄发射率给回零，这确认了底板能够吸收平滑的旧星光——但物理门控仍失败，因为 QPB/SP 退化未解决（R125）。该回归现已到第三代且仍未确立：D1c 用探测器角区逐文件固定 QPB，但其支撑规则移除了 605/622 个区域（只剩 13 个区域、24 个主用例中 0 个通过），因此只是形式结果——全样本旧年龄 \(\epsilon\) 1.99 [1.72, 2.25]e27 每已形成 M☉、比值 82、dev/dof 1.74、QPB pull 8.4——而 D1d 正以像素级支撑、绝对角区→视场 QPB 运行（R129、R130）。软 QPB 尺度 2.7 不是校准误差：它是 run6 连续谱模板中缺失的 1.4–2.0 keV Al K\(\alpha\) + Si K\(\alpha\) 线带，该波段角区计数的 64.8% 正是这些线，角区/模板比值为 3.71，而软子波段为 0.938 与 1.014（R127）。下一次尝试的背景实践现已写定：天空作为一个由离盘对照锚定的物理谱模型，QPB、软质子与 Al/Si 线作为分开的探测器项，整数计数泊松，用模拟校准 deviance，并把 old 系数与 sky/QPB/SP 联合剖面（R128）。（来源：07 R125、R127、R128。）
B7 — The diffuse-gas limit is conditional, and the outer-annulus rule is undecided. With the stellar floor fitted rather than borrowed, only the HI-shaped gas term is separable (3.47 \(\pm\) 1.56\(\times\)10\(^{-17}\), 2.2\(\sigma\)), so the 3\(\sigma\) limit is L_X(0.5–2 keV) < 3.8e37 erg/s inside a = 56′ and < 4.4e37 outside — about six times below the first estimate; the separate SFR-shaped gas term of D1b is not detected: with the masked point-source count within 1′ as a covariate its amplitude drops 1.3603 → 0.3673 and the cluster-bootstrap interval [0, 0.97]e−16 includes zero, so at 90 % it is < 0.97e−16 ≈ L < 1.1e37 erg/s over the footprint (R131); the envelope 2.3–6.2e37 / 2.7–7.1e37 erg/s is quoted with its floor anchor named. The pre-registered outer rule came back undecided and power-limited, because the halo-field scatter is 2–4× the whole detected halo range, so no per-annulus test can separate a halo from a foreground; the independent foreground field is the next step. (Source: 07 R091, R098, R115, R122.)
B7 —— 弥散气体上限是有条件的，而外环带规则未决。在拟合而非借用恒星底板的情况下，只有 HI 形状的气体项可分离（3.47 \(\pm\) 1.56\(\times\)10\(^{-17}\)，2.2\(\sigma\)），因此 3\(\sigma\) 上限为 a = 56′ 以内 L_X(0.5–2 keV) < 3.8e37 erg/s、以外 < 4.4e37 erg/s——比最初的估计低约六倍；D1b 那个单独的 SFR 形状气体项未检出：把 1′ 内的被掩护点源计数作为协变量后，其幅度从 1.3603 降到 0.3673，簇自助法区间 [0, 0.97]e−16 包含零，因此 90% 下 < 0.97e−16 ≈ 在该足迹上 L < 1.1e37 erg/s（R131）；区间 2.3–6.2e37 / 2.7–7.1e37 erg/s 以其底板锚点注明引用。预注册外环带规则的裁定为未决且受功效限制，因为晕场散度是整个已检出晕范围的 2–4 倍，任何逐环带检验都无法区分晕与前景；独立前景视场是下一步。（来源：07 R091、R098、R115、R122。）
Anticipated Q&A预期问答
| Question问题 | Answer回答 | Backup备用 |
|---|---|---|
| Is the 0.7 keV component real, or an O–norm degeneracy?0.7 keV 成分是真实的，还是 O–norm 退化？ | Partly resolved: the profile is real (kT 0.71–0.90 keV, 9 annuli), the inner-disk hard excess survives point-spread and background tests (3.6–4.8 \(\sigma\)), and the halo-field 0.7 keV component is detected (q = 34.84; kT \(\le 0.804\) keV; norm [3.15, 5.82]\(\times\)10\(^{-8}\) arcmin\(^{-2}\)). One Galactic 0.7 keV foreground is the provisional reading, but its level is estimator-dependent, so "one common level" is Questioned; the pre-registered outer rule (V7) has come back undecided and power-limited. (Source: 07 R071, R093, R112, R115; I021.)部分解决：剖面是真实的（kT 0.71–0.90 keV，9 个环带），内圆盘硬超额在点扩散和背景检验后仍然存在（3.6–4.8 \(\sigma\)），晕场 0.7 keV 成分被检出（q = 34.84；kT \(\le 0.804\) keV；norm [3.15, 5.82]\(\times\)10\(^{-8}\) arcmin\(^{-2}\)）。单一银河系 0.7 keV 前景是临时性读法，但其水平依赖于估计器，因此“一个共同水平”受质疑；预注册外环带规则（V7）的裁定为未决且受功效限制。（来源：07 R071、R093、R112、R115；I021。） | B4B4 |
| Why foreground-free errors?为什么需要前景自由误差？ | Foreground and SrcApec trade flux with almost unchanged chi^2; conditional errors are far too small (\(\times\)2.84/\(\times\)1.96 growth needed).前景和 SrcApec 可以互换通量，chi^2 几乎不变；条件性误差远小（需要 \(\times\)2.84/\(\times\)1.96 增长）。 | B4B4 |
| Isn't it just the Milky Way foreground?不就是银河前景吗？ | The MW-only column toward M31 is \(\approx\)0.054–0.067e22, about twice the frozen disk nH=0.03 (R016); V1 tests the bracket 0.048/0.0585/0.076.指向 M31 的仅 MW 柱密度 \(\approx\)0.054–0.067e22，约为冻结圆盘 nH=0.03 的两倍（R016）；V1 测试括号 0.048/0.0585/0.076。 | B2B2 |
| Doesn't WISE W1 just trace old stars, so everything is stellar?WISE W1 不就是追踪旧恒星，所以一切都是恒星？ | No — obs/pred crosses unity at 60–68' and no 2-parameter Sigma_*-only model is acceptable. But old+C and old+FUV+C are identifiable (lower bound 3.3%).不是——obs/pred 在 60–68' 处穿过 unity，且没有任何 2 参数仅 Sigma_* 模板可接受。但 old+C 和 old+FUV+C 可辨识（下界 3.3%）。 | B3B3 |
| What about the halo fields — do they have 0.7 keV too?晕场也有 0.7 keV吗？ | They do — the halo-field component is detected (q = 34.84; kT \(\le 0.804\) keV at 90 %; norm interval [3.15, 5.82]\(\times\)10\(^{-8}\) arcmin\(^{-2}\)). But its level is estimator-dependent: two registered fits of the same data give 9.66 \(\pm\) 1.16\(\times\)10\(^{-8}\) and 5.0 [3.56, 7.43]\(\times\)10\(^{-8}\), 2.3–2.7\(\sigma\) apart, and registered levels span \(\times\)11.8. So "one common level" is Questioned, and the origin — Milky Way versus a cored M31 halo — is still open. (Source: 07 R071, R093, R112; I021.)有——晕场成分被检出（q = 34.84；90% 下 kT \(\le 0.804\) keV；norm 区间 [3.15, 5.82]\(\times\)10\(^{-8}\) arcmin\(^{-2}\)）。但其水平依赖于估计器：同一份数据的两个已注册拟合给出 9.66 \(\pm\) 1.16\(\times\)10\(^{-8}\) 与 5.0 [3.56, 7.43]\(\times\)10\(^{-8}\)，相差 2.3–2.7\(\sigma\)，已注册水平跨度为 \(\times\)11.8。因此“一个共同水平”受质疑，其起源——银河系还是带核 M31 晕——仍然开放。（来源：07 R071、R093、R112；I021。） | B5B5 |
| Is the result generalizable to other disk galaxies?结果能推广到其他圆盘星系吗？ | Pending — Q5 (fraction of diffuse soft X-ray that is unresolved stellar) is not started.待定 —— Q5（弥散软 X 射线中未解析恒星的比例）尚未开始。 | —— |
| Why MOS only, not pn?为什么只用 MOS，不用 pn？ | Project convention: run6 is MOS-only; pn was rejected in earlier pilots (E020 C-mode attempt).项目约定：run6 仅 MOS；pn 在早期试点中被拒（E020 C-mode 尝试）。 | B1B1 |
Newest source read: cognition-os/08_RUNNING.md revision 49, updated 2026-11-05 03:56 EST; 07_REVISION_LOG.md R054–R131 and 04_DECISION_LOG.md D26–D30 read in full.
最新读取的来源：cognition-os/08_RUNNING.md 修订版 49，更新于 2026-11-05 03:56 EST；07_REVISION_LOG.md R054–R131 与 04_DECISION_LOG.md D26–D30 已全文读取。