スキル一覧に戻る

develop-hplc-method

pjt222
更新日 2 days ago
8 閲覧
17
2
17
GitHubで表示
デザインdesign

について

このスキルは、高速液体クロマトグラフィー(HPLC)メソッドを体系的に構築する手順を開発者に示します。分離目標の定義、カラムの化学的特性と移動相の選択、グラジエントや流量などの条件の最適化を支援します。溶液中または複雑なマトリックス中の、揮発性が低く、熱的に不安定、あるいは極性を持つ対象分析物の分析にご利用ください。

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/develop-hplc-method

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

Develop an HPLC Method

Systematic build of HPLC method. Mode, column chemistry, mobile phase + gradient, flow + temp, detector, iterative refinement. For non-volatile, thermally labile, or polar analytes.

When Use

  • Analyze compounds non-volatile, thermally labile, or too polar for GC
  • Build new HPLC-UV, HPLC-fluorescence, or LC-MS method from scratch
  • Adapt literature or pharmacopeial HPLC method to different column or instrument
  • Fix existing method with poor resolution, long run times, or sensitivity issues
  • Pick right chromatographic mode (reversed-phase, HILIC, ion-exchange, SEC, chiral)

Inputs

Required

  • Target analytes: Compound names, structures, MW, pKa, logP/logD
  • Sample matrix: Formulation, biological fluid, environmental extract, or neat solution
  • Performance targets: Required resolution, detection limits, quantitation range

Optional

  • Reference method: Compendial or literature method as starting point
  • Available columns: HPLC column inventory
  • Instrument config: UHPLC vs conventional HPLC, available detectors, column oven range
  • Throughput: Max run time incl re-equilibration
  • Regulatory context: ICH, USP, EPA, or other framework

Steps

Step 1: Define Separation Goals

  1. Compile analyte properties: MW, pKa, logP (or logD at relevant pH), chromophores, fluorophores, ionizable groups.
  2. Identify matrix + expected interferents (excipients, endogenous compounds, degradation products).
  3. Set perf criteria:
    • Resolution between critical pairs (Rs >= 2.0 for regulated)
    • Detection limits (LOD/LOQ)
    • Acceptable run time incl gradient re-equilibration
  4. Method purpose: assay, impurity profiling, dissolution, content uniformity, cleaning verification — drives validation category.
  5. Isocratic vs gradient: use isocratic if all analytes elute within retention factor range 2 < k' < 10; else gradient.

Got: Spec doc lists analytes + physicochemical properties, matrix, perf criteria, isocratic vs gradient decision.

If fail: pKa or logP unknown? Estimate from structure via prediction tools (ChemAxon, ACD/Labs) or run scouting gradient on C18 at pH 3 + pH 7 to empirically assess retention.

Step 2: Pick Column Chemistry

Pick chromatographic mode + column by analyte properties.

ModeColumn ChemistryMobile PhaseBest For
Reversed-phase (RP)C18 (ODS)Water/ACN or water/MeOH + acid/bufferNon-polar to moderately polar, most small molecules
RP (extended)C8, phenyl-hexyl, biphenylWater/organic + modifierShape selectivity, aromatic compounds, positional isomers
RP (polar-embedded)Amide-C18, polar-endcapped C18Water/organic, compatible with high aqueousPolar analytes that elute too early on standard C18
HILICBare silica, amide, zwitterionicHigh organic (80-95% ACN) + aqueous bufferVery polar, hydrophilic compounds (sugars, amino acids, nucleotides)
Ion-exchange (IEX)SAX or SCXBuffer with ionic strength gradientPermanently charged species, proteins, oligonucleotides
Size-exclusion (SEC)Diol-bonded silica, polymerIsocratic aqueous or organic bufferProtein aggregates, polymers, molecular weight distribution
ChiralPolysaccharide (amylose/cellulose)Normal-phase or polar organic modeEnantiomeric separations, chiral purity
  1. Default to reversed-phase C18 for small molecules with logP > 0.
  2. Analytes with logP < 0? Evaluate HILIC or ion-exchange.
  3. Particle size: sub-2 um for UHPLC (higher efficiency, higher backpressure), 3-5 um for conventional HPLC.
  4. Column dimensions: 50-150 mm length, 2.1-4.6 mm ID. Narrower columns save solvent + improve MS sensitivity.
  5. Chiral separations? Screen min 3-4 chiral stationary phases with different selectors.

Got: Column chemistry, dimensions, particle size picked with justification by analyte properties.

If fail: Initial scouting shows poor retention on C18? Switch to more retentive phase (phenyl-hexyl for aromatics) or different mode (HILIC for polar).

Step 3: Design Mobile Phase + Gradient

  1. Pick organic modifier:
    • Acetonitrile (ACN): lower viscosity, sharper peaks, better UV transparency below 210 nm
    • Methanol (MeOH): different selectivity, sometimes better for polar analytes, higher viscosity
  2. Pick aqueous component + pH:
    • Neutral analytes: water with 0.1% formic acid (MS-compatible) or phosphate buffer (UV only)
    • Ionizable analytes: buffer mobile phase 2 pH units away from analyte pKa → single ionic form
    • pH 2-3 (formic/phosphoric acid): suppresses ionization of acids, good general starting point
    • pH 6-8 (ammonium formate/acetate): for basic analytes or when low-pH selectivity insufficient
    • pH 9-11 (ammonium bicarbonate, BEH columns): for very basic compounds on high-pH-stable columns
  3. Design gradient:
    • Start at 5-10% organic, ramp to 90-95% organic over 10-20 min for initial scouting
    • Evaluate scouting chromatogram → find useful organic range
    • Narrow gradient to only elution window of interest
    • Gradient slope: steeper = faster but lower resolution; shallower = better resolution, longer run
  4. Include column wash step (95% organic, 2-3 min) + re-equilibration (initial conditions, 5-10 column volumes).
  5. Isocratic methods: target k' = 3-8 for analytes of interest.

Got: Mobile phase composition (organic, aqueous, buffer/additive, pH) + gradient profile defined. Scouting run confirms analyte elution within programmed window.

If fail: Selectivity poor (analytes co-elute despite gradient opt)? Change organic modifier (ACN to MeOH or reverse), adjust pH by 2 units, or add ion-pair reagent for charged analytes.

Step 4: Optimize Flow Rate + Temperature

  1. Set initial flow rate by column dimensions:
    • 4.6 mm ID: 1.0 mL/min
    • 3.0 mm ID: 0.4-0.6 mL/min
    • 2.1 mm ID: 0.2-0.4 mL/min
  2. Verify backpressure within instrument + column limits (usually < 400 bar conventional, < 1200 bar UHPLC).
  3. Optimize column temp:
    • Start at 30 C for reproducibility (avoid ambient fluctuations)
    • Raise to 40-60 C to cut viscosity, lower backpressure, sharpen peaks
    • Chiral columns: temp often has strong effect on enantioselectivity — screen 15-45 C
  4. Evaluate flow rate effect on resolution: small flow increases can boost throughput without big resolution loss if near van Deemter minimum.
  5. Document optimal flow rate, column temp, backpressure.

Got: Flow rate + column temp optimized. Backpressure within limits. Resolution maintained or improved vs initial conditions.

If fail: Backpressure too high? Reduce flow rate, raise temp, or switch to wider-bore or larger-particle column. Resolution degrades at higher temp? Return to 30 C, accept longer run.

Step 5: Pick Detector

DetectorPrincipleSensitivitySelectivityKey Considerations
UV (single wavelength)Absorbance at fixed lambdang rangeCompounds with chromophoresSimple, robust, most common
DAD (diode array)Full UV-Vis spectrumng rangeChromophores + spectral IDPeak purity assessment, library matching
Fluorescence (FLD)Excitation/emissionpg range (10-100x more sensitive than UV)Native fluorophores or derivatizedExcellent selectivity, requires fluorescent analytes
Refractive index (RI)Bulk propertyug rangeUniversal (no chromophore needed)Temperature-sensitive, gradient-incompatible
Evaporative light scattering (ELSD)Nebulization + light scatteringng rangeUniversal, non-volatile analytesSemi-quantitative, non-linear response
Charged aerosol (CAD)Nebulization + corona dischargeng rangeUniversal, non-volatile analytesMore uniform response than ELSD
Mass spectrometry (MS)m/z detectionpg-fg rangeStructural, highest selectivityRequires MS-compatible mobile phases
  1. UV chromophores (aromatic rings, conjugated systems)? Start with DAD — gives both quantitation + peak purity.
  2. Trace analysis, complex matrices? Prefer MS (ESI or APCI) in SIM or MRM mode.
  3. Compounds without chromophores (sugars, lipids, polymers)? Use CAD, ELSD, or RI.
  4. Set detection wavelength at analyte's absorption max (lambda-max) for best sensitivity, or 210-220 nm for general screening.
  5. Fluorescence? Optimize excitation + emission wavelengths via spectral scan of analyte.
  6. Ensure mobile phase additives compatible: no phosphate buffers with MS, no UV-absorbing additives at low wavelengths.

Got: Detector picked + configured (wavelength, gain, acquisition rate). Right for analyte chemistry + sensitivity needs.

If fail: UV sensitivity insufficient at required LOQ? Consider fluorescence derivatization (e.g., OPA for amines, FMOC for amino acids) or switch to LC-MS/MS for max sensitivity + selectivity.

Step 6: Evaluate + Refine

  1. Inject system suitability standard 6x. Evaluate:
    • Retention time RSD < 1.0%
    • Peak area RSD < 2.0%
    • Resolution of critical pair >= 2.0
    • Tailing factor 0.8-1.5 for all peaks
    • Theoretical plates per column spec
  2. Inject placebo/matrix blank to check interference at analyte retention times.
  3. Inject stressed or spiked sample to verify method separates degradation products from main analyte(s).
  4. Any criterion fails? Adjust one var at a time:
    • Poor resolution: change pH, gradient slope, or column chemistry
    • Tailing: add amine modifier (TEA for basic analytes), change buffer, or try different bonded phase
    • Sensitivity: raise injection volume, concentrate sample, or switch detector
  5. Lock final method params + document all conditions.

Got: All system suitability criteria met. Method resolves target analytes from matrix interferents + known degradation products. Params documented for transfer.

If fail: Iterative adjustment doesn't fix issue? Try fundamentally different approach (change chromatographic mode, 2D-LC, or derivatization). Return to Step 2.

Checks

  • All target analytes resolved with Rs >= 2.0 for critical pairs
  • Retention time RSD < 1.0% across 6 replicate injections
  • Peak area RSD < 2.0% across 6 replicate injections
  • Tailing factors 0.8-1.5 for all analyte peaks
  • No matrix interference at analyte retention times
  • Degradation products resolved from main analyte(s)
  • Run time (incl re-equilibration) meets throughput
  • Mobile phase compatible with selected detector
  • Method params fully documented (column, mobile phase, gradient, flow, temperature, detector)

Pitfalls

  • Ignoring mobile phase pH for ionizable analytes: Running at pH near analyte's pKa → split peaks or poor reproducibility (compound exists in two ionic forms). Buffer at least 2 pH units away from pKa.
  • Phosphate buffers with MS detection: Phosphate = non-volatile, contaminates MS source. Use formate or acetate buffers for LC-MS.
  • Insufficient re-equilibration after gradient: Column must be flushed with at least 5-10 column volumes of initial mobile phase before next injection. Inadequate re-equilibration → retention time drift.
  • Too short a column for complex mixtures: Short columns (50 mm) fast but may not give enough plates for multi-component separation. Start 100-150 mm for method dev.
  • Neglecting system dwell volume: Dwell volume (mixer to column head) delays gradient reaching column. Differs between instruments → method transfer failures. Measure + document.
  • Running HILIC like reversed-phase: HILIC needs high organic (80-95% ACN) with small aqueous fraction. Raising aqueous content → raises elution strength — opposite of RP. Equilibration times also longer.

See Also

  • develop-gc-method -- gas chromatography for volatile + semi-volatile analytes
  • interpret-chromatogram -- reading + interpreting HPLC + GC chromatograms
  • troubleshoot-separation -- diagnose + fix peak shape, retention, resolution
  • validate-analytical-method -- formal ICH Q2 validation of developed HPLC method

GitHub リポジトリ

pjt222/agent-almanac
パス: i18n/caveman/skills/develop-hplc-method
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

関連スキル

executing-plans

デザイン

executing-plansスキルは、完全な実装計画があり、それを管理されたバッチでレビューチェックポイントを設けながら実行する場合に使用します。このスキルは計画を読み込んで批判的にレビューした後、小さなバッチ(デフォルトは3タスク)でタスクを実行し、各バッチの間に進捗状況を報告してアーキテクトのレビューを受けます。これにより、品質管理チェックポイントが組み込まれた体系的な実装が保証されます。

スキルを見る

requesting-code-review

デザイン

このスキルは、コードレビュアーサブエージェントを起動し、処理を進める前に要件に対してコード変更を分析します。タスク完了後、主要な機能の実装後、またはmainブランチへのマージ前などに使用すべきです。このレビューは、現在の実装と元の計画を比較することで、問題を早期に発見するのに役立ちます。

スキルを見る

connect-mcp-server

デザイン

このスキルは、開発者がHTTP、stdio、またはSSEトランスポートを使用してMCPサーバーをClaude Codeに接続するための包括的なガイドを提供します。GitHub、Notion、カスタムAPIなどの外部サービスを統合するためのインストール、設定、認証、セキュリティについて解説しています。MCP統合のセットアップ、外部ツールの設定、またはClaudeのModel Context Protocolを扱う際にご利用ください。

スキルを見る

web-cli-teleport

デザイン

このスキルは、タスク分析に基づいて開発者がClaude Code WebとCLIインターフェースの選択を支援し、これらの環境間でのシームレスなセッションテレポーテーションを可能にします。Web、CLI、モバイル環境を切り替える際のセッション状態とコンテキストを管理することで、ワークフローを最適化します。様々な段階で異なるツールを必要とする複雑なプロジェクトにご活用ください。

スキルを見る