スキル一覧に戻る

develop-gc-method

pjt222
更新日 2 days ago
2 閲覧
17
2
17
GitHubで表示
開発general

について

このClaude Skillは、ガスクロマトグラフィー分析法を一から体系的に構築します。対象分析物質に応じたカラム選択、温度プログラム設定、検出器構成、初期性能検証までを包括的に扱います。揮発性または半揮発性化合物の分析を目的とした新規GC分析法の開発が必要な際にご利用ください。

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-gc-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 a GC Method

Systematic development of a gas chromatography method covering column selection, temperature program optimization, carrier gas and detector choice, and initial performance verification for volatile and semi-volatile analytes.

When to Use

  • Starting a new GC analysis for volatile or semi-volatile compounds
  • Adapting a published method to a different instrument or matrix
  • Replacing an existing method that no longer meets performance requirements
  • Developing a method for compounds with known boiling points and polarities
  • Transitioning from a packed-column method to a capillary method

Inputs

Required

  • Target analytes: List of compounds with CAS numbers, molecular weights, and boiling points
  • Sample matrix: Description of the sample type (e.g., air, water extract, solvent solution, biological fluid)
  • Detection limits: Required LOD/LOQ for each analyte

Optional

  • Reference method: Published method (EPA, ASTM, pharmacopeial) to use as a starting point
  • Available columns: Inventory of columns already on hand
  • Instrument configuration: GC model, available detectors, autosampler type
  • Throughput requirements: Maximum acceptable run time per sample
  • Regulatory framework: GLP, GMP, EPA, or other compliance context

Procedure

Step 1: Define Analytical Objectives

  1. List all target analytes with their physical properties (boiling point, polarity, molecular weight).
  2. Identify the sample matrix and any expected interferents or co-extractives.
  3. Specify required detection limits, quantitation range, and acceptable resolution between critical pairs.
  4. Determine whether the method must meet a regulatory standard (EPA 8260, USP, etc.).
  5. Document throughput needs: maximum run time, injection volume, sample preparation constraints.

Got: A written specification listing analytes, matrix, detection limits, resolution requirements, and any regulatory or throughput constraints.

If fail: If analyte volatility data is unavailable, estimate boiling points from structural analogs or use a scouting run on a mid-polarity column to establish elution order.

Step 2: Select the Column

Choose column dimensions and stationary phase based on analyte polarity and separation difficulty.

Column TypeStationary PhasePolarityTypical Use Cases
DB-1 / HP-1100% dimethylpolysiloxaneNon-polarHydrocarbons, solvents, general screening
DB-5 / HP-55% phenyl-methylpolysiloxaneLow polaritySemi-volatiles, EPA 8270, drugs of abuse
DB-170114% cyanopropylphenylMid polarityPesticides, herbicides
DB-WAX / HP-INNOWaxPolyethylene glycolPolarAlcohols, fatty acids, flavors, essential oils
DB-6246% cyanopropylphenylMid polarityVolatile organics, EPA 624/8260
DB-FFAPModified PEG (nitroterephthalic acid)Highly polarOrganic acids, free fatty acids
DB-3535% phenyl-methylpolysiloxaneMid-low polarityPolychlorinated biphenyls, confirmatory column
  1. Match analyte polarity to stationary phase: like dissolves like.
  2. Select column length (15-60 m): longer columns give more plates but longer run times.
  3. Select inner diameter (0.25-0.53 mm): narrower gives better efficiency, wider gives more capacity.
  4. Select film thickness (0.25-5.0 um): thicker films retain volatile analytes longer.
  5. For complex matrices, consider a guard column or retention gap.

Got: A column specification (phase, length, ID, film thickness) justified by analyte properties and separation requirements.

If fail: If no single column resolves all critical pairs, plan a confirmation column with orthogonal selectivity (e.g., DB-1 primary, DB-WAX confirmatory).

Step 3: Optimize the Temperature Program

  1. Set the initial oven temperature at or below the boiling point of the most volatile analyte (hold 1-2 min for solvent focusing).
  2. Apply a linear ramp. General starting points:
    • Simple mixtures: 10-20 C/min
    • Complex mixtures: 3-8 C/min for better resolution
    • Ultra-fast screening: 25-40 C/min on short thin-film columns
  3. Set the final temperature 10-20 C above the boiling point of the least volatile analyte.
  4. Add a final hold (2-5 min) to ensure complete elution and column bake-out.
  5. For critical pairs that co-elute, insert an isothermal hold at the temperature before their elution, or reduce the ramp rate in that region.
  6. Verify that the total run time meets throughput requirements.

Got: A temperature program (initial temp, hold, ramp rate(s), final temp, final hold) that separates all target analytes within the acceptable run time.

If fail: If critical pairs remain unresolved after ramp optimization, revisit column selection (Step 2) or consider a multi-ramp program with slower rates in the problem region.

Step 4: Select the Carrier Gas

PropertyHelium (He)Hydrogen (H2)Nitrogen (N2)
Optimal linear velocity20-40 cm/s30-60 cm/s10-20 cm/s
Efficiency at high flowGoodBest (flat van Deemter)Poor
Speed advantageBaseline1.5-2x faster than HeSlowest
SafetyInertFlammable (needs leak detection)Inert
Cost / availabilityExpensive, supply concernsLow cost, generator optionVery low cost
Detector compatibilityAll detectorsNot with ECD; caution with some MSAll detectors
  1. Default to helium for general-purpose work and regulatory methods specifying He.
  2. Consider hydrogen for faster analysis or when helium supply is constrained; install hydrogen-specific leak detection and safety interlocks.
  3. Use nitrogen only for simple separations or when cost is the primary driver.
  4. Set the carrier gas flow to the optimal linear velocity for the chosen gas and column ID.
  5. Measure actual linear velocity using an unretained compound (e.g., methane on FID).

Got: Carrier gas selected with flow rate set to optimal linear velocity, verified by unretained peak measurement.

If fail: If efficiency is lower than expected at the set flow, generate a van Deemter curve (plate height vs. linear velocity) using 5-7 flow rates to find the true optimum.

Step 5: Choose the Detector

DetectorSelectivitySensitivity (approx.)Linear RangeBest For
FIDC-H bonds (universal organic)Low pg C/s10^7Hydrocarbons, general organics, quantitation
TCDUniversal (all compounds)Low ng10^5Permanent gases, bulk analysis
ECDElectronegative groups (halogens, nitro)Low fg (Cl compounds)10^4Pesticides, PCBs, halogenated solvents
NPD/FPDN, P (NPD); S, P (FPD)Low pg10^4-10^5Organophosphorus pesticides, sulfur compounds
MS (EI)Structural identificationLow pg (scan), fg (SIM)10^5-10^6Unknowns, confirmation, trace analysis
MS/MSHighest selectivityfg range10^5Complex matrices, ultra-trace, forensic
  1. Match detector to analyte chemistry and required sensitivity.
  2. For quantitative work with simple matrices, FID is the default (robust, linear, low maintenance).
  3. For trace analysis in complex matrices, prefer MS in SIM mode or MS/MS in MRM mode.
  4. For halogenated compounds at trace levels, ECD provides the best sensitivity.
  5. Set detector temperature 20-50 C above the maximum oven temperature to prevent condensation.
  6. Optimize detector gas flows per manufacturer recommendations.

Got: Detector selected and configured with appropriate temperatures and gas flows for the target analytes.

If fail: If detector sensitivity is insufficient at the required detection limits, consider concentrating the sample (larger injection volume, solvent evaporation) or switching to a more sensitive/selective detector.

Step 6: Validate Initial Performance

  1. Prepare a system suitability standard containing all target analytes at mid-range concentration.
  2. Inject the standard 6 times consecutively.
  3. Evaluate:
    • Retention time RSD: must be < 1.0%
    • Peak area RSD: must be < 2.0% (< 5.0% for trace-level)
    • Resolution between critical pairs: Rs >= 1.5 (baseline) or >= 2.0 for regulated methods
    • Peak tailing factor: 0.8-1.5 (USP criteria T <= 2.0)
    • Theoretical plates (N): verify against column manufacturer specification
  4. Inject a blank to confirm absence of carryover or ghost peaks.
  5. Inject a matrix blank to identify potential interferents at target retention times.
  6. Document all parameters in a method summary sheet.

Got: System suitability criteria met for all analytes across replicate injections, with no carryover or matrix interferences at target retention windows.

If fail: If tailing is observed, check for active sites (re-condition column, trim 0.5 m from inlet end, replace liner). If RSD exceeds limits, investigate autosampler precision and injection technique. If resolution is insufficient, return to Step 3 to refine the temperature program.

Validation

  • All target analytes are separated with Rs >= 1.5 for critical pairs
  • Retention time RSD < 1.0% over 6 replicate injections
  • Peak area RSD < 2.0% over 6 replicate injections
  • Peak tailing factors within 0.8-1.5 for all analytes
  • Blank injection shows no carryover above 0.1% of working concentration
  • Matrix blank shows no interferents at target retention windows
  • Total run time meets throughput requirements
  • Method parameters are fully documented (column, temps, flows, detector settings)

Pitfalls

  • Ignoring column bleed temperature limits: Operating above the maximum isothermal temperature of the stationary phase causes elevated baseline, ghost peaks, and accelerated column degradation. Always check the column specification sheet.
  • Oversized injection volumes: Injecting too much solvent causes fronting peaks and poor resolution for early eluters. Match injection volume to column capacity (typically 0.5-2 uL for 0.25 mm ID columns in split mode).
  • Wrong liner for the injection mode: Splitless injections require a single-taper or double-taper deactivated liner; split injections use a liner with glass wool. Mismatched liners cause poor reproducibility.
  • Neglecting septum and liner maintenance: Septum coring and liner contamination are the most common sources of ghost peaks and tailing. Replace septa every 50-100 injections and liners on a documented schedule.
  • Skipping the van Deemter optimization: Running at the manufacturer's default flow rate instead of the measured optimum wastes efficiency, especially when switching carrier gases.
  • Insufficient column conditioning: New columns must be conditioned (ramped to maximum temperature under carrier gas flow, no detector) to remove manufacturing residues before analytical use.

Related Skills

  • develop-hplc-method -- liquid chromatography method development for non-volatile or thermally labile analytes
  • interpret-chromatogram -- reading and interpreting GC and HPLC chromatograms
  • troubleshoot-separation -- diagnosing and fixing peak shape, retention, and resolution problems
  • validate-analytical-method -- formal ICH Q2 validation of the developed GC method

GitHub リポジトリ

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

関連スキル

subagent-driven-development

開発

このスキルは、各独立したタスクに対して新規のサブエージェントを起動し、タスク間でコードレビューを実施しながら実装計画を実行します。レビュープロセスを通じて品質基準を維持しつつ、迅速な反復を可能にします。同一セッション内で主に独立したタスクに取り組む際に本スキルをご利用いただくことで、組み込まれた品質チェックを伴う継続的な進捗を確保できます。

スキルを見る

qmd

開発

qmdは、BM25、ベクトル埋め込み、およびリランキングを組み合わせたハイブリッド検索を用いて、ローカルファイルのインデックス作成と検索を可能にするローカル検索・インデックス作成CLIツールです。コマンドラインでの使用と、Claudeとの統合のためのMCP(Model Context Protocol)モードの両方をサポートしています。このツールは埋め込みにOllamaを使用し、インデックスをローカルに保存するため、ターミナルから直接ドキュメントやコードベースを検索するのに最適です。

スキルを見る

mcporter

開発

mcporterスキルは、開発者がClaudeから直接Model Context Protocol(MCP)サーバーを管理および呼び出せるようにします。このスキルは、利用可能なサーバーの一覧表示、引数を指定したツールの呼び出し、認証およびデーモンのライフサイクル管理を行うコマンドを提供します。開発ワークフローにおいてMCPサーバーの機能を統合およびテストする際に、このスキルをご利用ください。

スキルを見る

adk-deployment-specialist

開発

このスキルは、A2Aプロトコルを使用してVertex AI ADKエージェントをデプロイおよびオーケストレーションし、AgentCardの発見、タスク送信、およびコード実行サンドボックスやメモリバンクなどのサポートツールを管理します。Python、Java、またはGoで、順次、並列、またはループのオーケストレーションパターンを用いたマルチエージェントシステムの構築を可能にします。Google Cloud上でADKエージェントのデプロイやエージェントワークフローのオーケストレーションを求められた際にご利用ください。

スキルを見る