スキル一覧に戻る

plan-spectroscopic-analysis

pjt222
更新日 2 days ago
7 閲覧
17
2
17
GitHubで表示
テストai

について

このスキルは、開発者が分光分析キャンペーンを計画する際に、意思決定マトリックスを用いて適切な分析技術を選択し、非破壊から破壊的分析へと順序立てて実施することを支援します。分析課題の定義、試料の評価、相互検証による成功基準の確立をユーザーに導きます。未知化合物の調査や、特定の科学的課題を解決するための分析シーケンスを最適化する際にご活用ください。

クイックインストール

Claude Code

推奨
メイン
npx skills add pjt222/agent-almanac -a claude-code
プラグインコマンド代替
/plugin add https://github.com/pjt222/agent-almanac
Git クローン代替
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/plan-spectroscopic-analysis

このコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします

ドキュメント

Plan Spectroscopic Analysis

Design spectroscopic campaign: pick right techniques, sequence efficiently, define success criteria → answer specific analytical question.

Use When

  • Investigate unknown compound → which spectroscopic techniques?
  • Optimize analysis sequence → preserve sample for destructive methods
  • Plan sample prep before instrument time
  • Cross-validate complementary techniques
  • Budget instrument time + prioritize when resources limited
  • Train new analysts in systematic planning

In

  • Required: Analytical question (structure ID, quantitation, purity, functional group screen, reaction monitoring)
  • Required: Sample desc (physical state, qty, known/suspected class)
  • Optional: Available instruments + capabilities
  • Optional: Budget + time constraints
  • Optional: Safety data (toxicity, reactivity, volatility, light)
  • Optional: Prior data (if any)

Do

Step 1: Define Analytical Question

Clarify info needed before picking technique.

  1. Classify question:

    • Structure ID: Full molecular structure of unknown. Broadest set.
    • Structure confirm: Known compound matches expected. Few, focused on diagnostics.
    • Quantitative: Concentration of known analyte. Calibration + good linearity (UV-Vis, NMR w/ internal std).
    • Purity: Impurities present? Identify? High sensitivity + separation.
    • Functional group screen: Which groups present, no full structure. IR often enough.
    • Reaction monitor: Track reaction over time. Speed + compatibility w/ conditions (in situ IR, Raman, UV-Vis).
  2. Success criteria: Explicit. Structure ID → "single proposal consistent w/ all data". Quantitation → "concentration w/ <5% rel error".

  3. Existing knowledge: Compile (elemental analysis, reaction scheme, expected product, lit precedent). Constrains problem, fewer techniques needed.

→ Clear analytical question w/ success criteria + existing knowledge summary.

If err: question vague ("characterize this") → narrow w/ requestor. Vague → unfocused → wasted instrument time.

Step 2: Assess Sample Characteristics

Eval sample → which techniques feasible.

  1. Physical state: Solid (crystalline, amorphous, powder), liquid, solution, gas, thin film, biological tissue. Each constrains prep + technique.
  2. Quantity: Total mass/vol. NMR needs mg, MS µg, SERS ng.
  3. Solubility: Test/estimate in common solvents (water, methanol, DMSO, chloroform, hexane). NMR → deuterated solvent. UV-Vis → transparent.
  4. Stability: Thermal (GC-MS needs volatilization), photo (Raman uses laser), air/moisture (KBr pellet), solution (time-dependent).
  5. Safety: Toxicity, flammability, reactivity, radioactivity. Affects handling, may exclude techniques (volatile toxics → no open-atmosphere Raman w/o containment).
  6. MW range: Small (<1000 Da) vs polymers/biomolecules (>1000 Da) → different MS ionization + NMR strategies.

→ Sample characterization summary: state, qty, solubility, stability, hazards, MW range.

If err: can't characterize adequately (qty too small to test solubility) → conservative: start non-destructive minimal-sample (Raman, ATR-IR), reassess after.

Step 3: Select Techniques via Decision Matrix

Pick most informative techniques based on question + sample.

TechniqueBest ForSample NeedsDestructive?SensitivityKey Limitations
1H NMRH connectivity, integration, coupling1--10 mg in deuterated solventNomgRequires solubility, insensitive
13C NMRCarbon skeleton, functional groups10--50 mg in deuterated solventNomgVery insensitive, long acquisition
2D NMRFull connectivity, stereochemistry5--20 mg in deuterated solventNomgHours of instrument time
IR (ATR)Functional group IDAny solid/liquid, minimal prepNougWater interference, fingerprint overlap
IR (KBr)Functional group ID, transmission1--2 mg solid in KBr pelletNo*ugMoisture sensitive, sample mixed
RamanSymmetric modes, aqueous samplesAny state, no prep for solidsNoug--mgFluorescence, photodegradation
EI-MSVolatile small molecules, fragmentationug, must be volatileYes (GC-MS)ng--ugRequires volatility
ESI-MSPolar/large molecules, MW determinationSolution in volatile solventYespg--ngAdduct complexity, ion suppression
MALDI-MSPolymers, proteins, large moleculesSolid + matrixYesfmolMatrix interference below 500 Da
UV-VisChromophores, quantitationSolution, ug--mgNougLimited structural information

*IR with KBr is non-destructive to the molecule but the sample cannot be easily recovered from the pellet.

  1. Match question to technique: Structure ID → NMR + MS + IR min. Functional group → IR only. Quantitation → UV-Vis or NMR best.
  2. Feasibility: Cross-ref candidates w/ Step 2 sample. Eliminate incompatible (GC-MS for non-volatile, NMR for paramagnetic).
  3. Prioritize by info density: Rank by info per question.
  4. Cost + availability: If equal info, prefer faster, cheaper, more available.

→ Ranked list of techniques w/ justification + excluded ones w/ reasons.

If err: no single sufficient (common for structure ID) → plan complementary techniques together. None suitable → note limitation, recommend alts (derivatization → GC-MS).

Step 4: Plan Sample Prep per Technique

Prep reqs per selected technique.

  1. NMR prep: Dissolve 1-50 mg in 0.5-0.7 mL deuterated solvent. Solvent by solubility + spectral window:
Solvent1H ResidualUse When
CDCl37.26 ppmNon-polar to moderately polar compounds
DMSO-d62.50 ppmPolar compounds, broad solubility
D2O4.79 ppmWater-soluble compounds, peptides
CD3OD3.31 ppmPolar organic compounds
C6D67.16 ppmAromatic region overlap avoidance
  1. IR prep: Method by sample state:

    • ATR: Solid/liquid direct on crystal. Fastest, minimal prep.
    • KBr pellet: Grind 1-2 mg w/ 100-200 mg dry KBr, press into transparent disk.
    • Solution cell: Dissolve in IR-transparent solvent (CCl4, CS2). Limited transparency windows.
    • Thin film: Cast from solution onto NaCl/KBr window. Polymers + oils.
  2. MS prep: Match ionization to sample:

    • EI (GC-MS): Sample volatile. Volatile solvent (DCM, hexane).
    • ESI (LC-MS): ESI-compatible solvent (methanol/water, acetonitrile/water w/ 0.1% formic acid).
    • MALDI: Mix w/ matrix (DHB, CHCA, sinapinic acid), dry on target.
  3. UV-Vis prep: UV-transparent solvent. Conc → absorbance at lambda-max 0.1-1.0. Matched cuvettes for sample + ref.

  4. Raman prep: Minimal. Solids neat. Liquids in glass vials (weak Raman). Avoid fluorescent containers. Aqueous solutions OK (water = weak Raman scatterer).

→ Prep protocol per technique: solvents, qtys, special handling.

If err: qty insufficient for all → prioritize by Step 3 hierarchy. Insoluble in all suitable → solid-state techniques (ATR-IR, Raman, solid-state NMR, MALDI-MS).

Step 5: Sequence + Cross-Validation Strategy

Order analyses → preserve sample, max info flow.

  1. Sequence by destructiveness: Non-destructive first, destructive last.

    • Tier 1 (non-destructive, no prep): Raman, ATR-IR
    • Tier 2 (non-destructive, requires prep): UV-Vis, NMR (sample often recoverable by evaporation)
    • Tier 3 (destructive or consumes sample): MS (ESI, EI/GC-MS, MALDI)
  2. Info flow: Early results refine later:

    • IR/Raman functional groups → choose NMR experiments (no carbonyl in IR → skip carbonyl-focused 13C).
    • MW from MS → interpret NMR (integration ratios, peak count).
    • NMR connectivity → interpret MS fragmentation.
  3. Cross-validation points: Where techniques must agree:

    • Molecular formula: MS (mol ion) = NMR (H + C count) = elemental analysis.
    • Functional groups: IR assignments consistent w/ NMR shifts + MS fragmentation.
    • Degree of unsaturation: From formula (MS) = observed rings + double bonds (NMR, UV-Vis).
  4. Contingencies: What if ambiguous:

    • NMR unexpected complexity → run 2D (COSY, HSQC, HMBC).
    • MS mol ion ambiguous → different ionization or HRMS.
    • IR dominated by one group → Raman for complementary.
  5. Document plan: Written plan w/ sequence, prep, turnaround, decision points.

→ Complete ordered plan w/ prep, cross-validation, contingencies doc'd.

If err: plan can't complete due to sample/instrument → doc limitations, propose best achievable subset.

Check

  • Analytical question clear w/ explicit success criteria
  • Sample characteristics assessed (state, qty, solubility, stability, hazards)
  • Techniques selected via decision matrix w/ justifications
  • Infeasible techniques excluded w/ reasons
  • Sample prep planned per technique
  • Analysis sequence non-destructive → destructive
  • Cross-validation points defined
  • Contingency experiments ID'd for ambiguous
  • Total sample consumption estimated vs available qty

Traps

  • Skip planning: Jumping to nearest instrument → wastes sample + time. 15 min planning saves hours of re-analysis.
  • Pick by habit not need: Not every analysis needs NMR. Functional group confirm → only IR. Match technique to question.
  • Underestimate sample reqs: Running out mid-sequence avoidable. Calc total upfront + 20% reserve.
  • Destructive methods first: GC-MS before NMR → NMR needs separate aliquot. Non-destructive first → max info per mg.
  • Neglect solvent compat: Sample in DMSO-d6 (NMR) → not easy for GC-MS (non-volatile). Plan solvents across all.
  • No cross-validation strategy: No checkpoints → contradictory results unnoticed until final interp.

  • interpret-nmr-spectrum — interpret NMR per this plan
  • interpret-ir-spectrum — interpret IR per this plan
  • interpret-mass-spectrum — interpret MS per this plan
  • interpret-uv-vis-spectrum — interpret UV-Vis per this plan
  • interpret-raman-spectrum — interpret Raman per this plan
  • validate-analytical-method — validate quantitative methods from this plan

GitHub リポジトリ

pjt222/agent-almanac
パス: i18n/caveman-ultra/skills/plan-spectroscopic-analysis
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

関連スキル

evaluating-llms-harness

テスト

このClaudeスキルは、lm-evaluation-harnessを実行し、MMLUやGSM8Kなど60以上の標準化学術タスクでLLMをベンチマークします。開発者がモデルの品質を比較し、トレーニングの進捗を追跡し、学術的な結果を報告するために設計されています。このツールはHuggingFaceやvLLMモデルを含む様々なバックエンドをサポートしています。

スキルを見る

cloudflare-cron-triggers

テスト

このスキルは、cron式を使用してWorkersをスケジュールするためのCloudflare Cron Triggersの実装に関する包括的な知識を提供します。定期的なタスクの設定、メンテナンスジョブ、自動化されたワークフローの構築を網羅し、無効なcron式やタイムゾーン問題といった一般的な課題への対処法も含みます。開発者はこれを使用して、スケジュールされたハンドラーの設定、cronトリガーのテスト、WorkflowsやGreen Computeとの連携を構成できます。

スキルを見る

webapp-testing

テスト

このClaude Skillは、Playwrightベースのツールキットを提供し、Pythonスクリプトを通じてローカルWebアプリケーションのテストを可能にします。フロントエンドの検証、UIデバッグ、スクリーンショット撮影、ログ表示を実現し、サーバーライフサイクルを管理します。ブラウザ自動化タスクにご利用いただけますが、コンテキストの汚染を避けるため、スクリプトのソースコードを読むのではなく直接実行してください。

スキルを見る

finishing-a-development-branch

テスト

このスキルは、開発者がテストの合格を確認し、構造化された統合オプションを提示することで、完成した作業を仕上げることを支援します。実装が完了した後のマージ、PR作成、ブランチの整理といったワークフローを案内します。コードが準備できてテスト済みの際に使用し、開発プロセスを体系的に完了させましょう。

スキルを見る