Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› jacs-methods
jacs-methods
GitHub用于审计JACS稿件中合成与表征的严谨性，涵盖纯度、谱学数据、单晶X射线及催化控制实验。不撰写文本，仅核查数据是否充分支持结论。
触发场景
报告新化合物需确认完整表征
审稿人质疑纯度或缺失X-ray数据
声称反应机理但未提供测试证据
报道催化范围或选择性但对照实验不明确
安装
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill jacs-methods -g -y
SKILL.md
Frontmatter
{
"name": "jacs-methods",
"description": "Use when checking synthesis and characterization rigor for a Journal of the American Chemical Society (JACS) manuscript — purity, NMR\/HRMS\/IR\/UV-vis, elemental analysis or HPLC, single-crystal X-ray\/CCDC, and catalysis controls\/mechanism\/kinetics. Audits experimental rigor — it does not write narrative or polish prose."
}
Synthesis and Characterization Rigor (jacs-methods)
When to trigger
- New compounds are reported and you must confirm full characterization
- A reviewer could ask "is this compound pure?" or "where is the X-ray?"
- A mechanism is claimed but only proposed, not tested
- Catalysis scope/selectivity is reported and controls are unclear
Characterization rigor by claim type
JACS requires that claims are fully supported by characterization data and appropriate controls. Match the data package to what you assert.
Every new, isolable compound
- ¹H NMR — solvent, field; δ (ppm), multiplicity, J (Hz), integration
- ¹³C{¹H} NMR — all resolvable carbons (note coincident peaks)
- Heteronuclear NMR as relevant: ¹⁹F, ³¹P, ¹¹B, etc.; 2D for assignment
- HRMS — ionization mode, calcd vs found m/z for the assigned formula
- IR — diagnostic bands (C=O, N–H, C≡N, M–CO, etc.) when structurally informative
- Purity evidence — at least one of: CHN(S) elemental analysis (calcd vs found, typically within accepted tolerance), or HPLC/GC purity, plus clean NMR
- Melting point / optical rotation / UV-vis as appropriate to compound class
Stereochemistry / chirality
- ee by chiral HPLC/GC/SFC (column, eluent, flow, detector; report racemate trace)
- dr by NMR or HPLC; specify how measured
- Absolute configuration assigned (X-ray, VCD, comparison to known, or derivatization) — state the method
Solid-state structure
- Single-crystal X-ray for structurally pivotal claims; deposit CIF with CCDC and cite the number
- Crystal/refinement table (formula, space group, R1/wR2, GooF, completeness)
- checkCIF/PLATON run; A/B-level alerts resolved or justified
- PXRD for bulk-phase identity of solids/materials
Catalysis
- Scope table — representative, honest substrate range (not only easy cases); note failures
- Control experiments — no-catalyst, no-ligand, no-additive, radical traps/clocks as relevant
- Mechanism — evidence-based: kinetics (orders, rate law), KIE, labeling, intermediates (isolation/spectroscopy), Hammett, ee–ee(cat) correlation
- Catalyst loading, TON/TOF, and reproducibility (replicate runs) reported
- Metal/ligand source, purity, and any trace-metal control where relevant
Materials / physical
- Composition and phase: PXRD, EDX/XPS, ICP, TGA, BET as appropriate
- Morphology with scale bars and representative-vs-statistical imaging (SEM/TEM/AFM)
- Property measurement with method, conditions, and error/replication
- Spectroscopy (UV-vis/PL/Raman/EPR/electrochemistry) tied to the claimed property
Computation (if used to support mechanism/structure)
- Functional, basis set/pseudopotentials, solvation model, software + version
- Stationary points characterized (frequencies; TS has one imaginary mode)
- Optimized coordinates (xyz) deposited in SI; energies referenced consistently
The SI data string (write it once, correctly)
Every isolated compound gets a characterization block in the SI in the standard ACS order. A well-formed entry looks like:
(2R)-2-Fluoro-4-phenylbutan-1-ol (3a). Colorless oil, 158 mg (94%). [α]D²⁵ = +12.4
(c 1.0, CHCl₃). ¹H NMR (400 MHz, CDCl₃): δ 7.31–7.18 (m, 5H), 4.52 (dddd, J = 48.9,
8.1, 4.3, 3.9 Hz, 1H), ... ¹³C{¹H} NMR (101 MHz, CDCl₃): δ 141.6, 128.5 (d,
JCF = 5.2 Hz), ... ¹⁹F NMR (376 MHz, CDCl₃): δ −186.3. HRMS (ESI+) m/z: [M + Na]⁺
calcd for C₁₀H₁₃FONa 191.0843; found 191.0841. 97:3 er (Chiralpak IA, 95:5
hexanes/iPrOH, 1.0 mL/min, 210 nm; tR major 14.2 min, minor 16.8 min).
Audit each block for: mass and percentage yield, physical state, every nucleus relevant to the structure, coupling constants on diagnostic multiplets, calcd/found within instrument tolerance, and the full chiral method — not just the er value.
Referee rigor probes (pre-empt them)
- "Compound x shows a peak at δ 1.56 — water or impurity?" → annotate residual-solvent peaks in the SI spectra copies before a referee circles them.
- "Yields are single runs?" → state replicate counts for headline entries; give an average with the spread for the model reaction.
- "Could adventitious metal catalyze this?" → for base-metal or 'metal-free' claims, report reagent lot ICP data or a deliberate trace-metal spike control.
- "The cycle shows Ni(I)/Ni(III) — where is the evidence?" → every oxidation state drawn in a mechanism figure must trace to an EPR/UV-vis/electrochemical or kinetic observation, or be labeled as proposed.
Reproducibility test
Could an independent chemist reproduce each yield, ee, and spectrum from the written procedure alone? If not, the procedure is incomplete — fix before figures. The strongest SI procedures also state how the compound was purified (column eluent, distillation T/p), on what scale the reaction was validated, and any setup detail on which the outcome hinges (drying of solvent, exclusion of light, stirring rate for biphasic mixtures).
Anti-patterns
- New compound reported with no purity criterion (no EA, no HPLC, no clean ¹³C)
- "Pure by NMR" as the only purity statement for a key compound
- Mechanism drawn as a cycle with no kinetic/labeling/control support
- Crystal structure shown but CIF not deposited / no CCDC number
- Scope table that hides failed substrates; idealized yields not reproducible
- HRMS reported without ionization mode or calcd/found pair
Output format
【Compounds】N new; full characterization complete for [ ] / missing for [...]
【Purity basis】EA / HPLC / clean NMR — per compound
【Stereochem】ee/dr method + absolute config method (if chiral)
【X-ray】CCDC # [...]; checkCIF clean? Y/N
【Mechanism】claimed = [...]; evidence = [kinetics/KIE/labeling/controls]
【Reproducible from text?】Y/N — gaps: [...]
【Next】jacs-figures → jacs-supplementary
Related resources
../../resources/external_tools.md— instruments, CCDC/PDB deposition, crystallography and computational software
版本历史
- 1839142 当前 2026-07-05 13:59
