interpret-uv-vis-spectrum
정보
이 스킬은 자외선-가시광선(UV-Vis) 스펙트럼을 분석하여 발색단을 식별하고 전자 전이를 분류하며, 공액 시스템에 대한 우드워드-피저 규칙을 적용합니다. 또한 비어-람베르트 법칙을 사용하여 정량적 농도 분석을 수행합니다. 방향족 고리와 같은 구조적 특징을 확인하거나 흡광도 변화를 통해 반응 동역학을 모니터링하는 데 사용하세요.
빠른 설치
Claude Code
추천npx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/interpret-uv-vis-spectrumClaude Code에서 이 명령을 복사하여 붙여넣어 스킬을 설치하세요
문서
Interpret UV-Vis Spectrum
Analyze UV-Vis absorption → id chromophores, classify electronic transitions, predict λ-max conjugated sys, apply Beer-Lambert for quant.
Use When
- ID chromophores + extent of conjugation in organic compound
- Confirm aromatic rings, conjugated dienes, enones
- Quant analysis (conc from absorbance)
- Monitor rxn kinetics via abs changes over time
- Characterize metal-ligand complexes (d-d + charge-transfer)
- Solvent effects on electronic transitions (solvatochromism)
In
- Req: UV-Vis data (λ nm vs abs / molar absorptivity)
- Req: Solvent
- Opt: Conc + path length (for Beer-Lambert)
- Opt: ε at λ-max
- Opt: Spectra in multi-solvents (solvatochromism)
- Opt: Structural info from other spectra
Do
Step 1: Verify Instrument Params + Quality
Ensure reliable data before interpret:
- λ range: Confirm relevant range. Standard UV-Vis 190-800 nm. Solvent cutoffs:
| Solvent | UV Cutoff (nm) | Notes |
|---|---|---|
| Water | 190 | Excellent UV transparency |
| Hexane | 195 | Non-polar, minimal solvent effects |
| Methanol | 205 | Protic, may cause blue shifts |
| Acetonitrile | 190 | Good general-purpose UV solvent |
| Dichloromethane | 230 | Absorbs below 230 nm |
| Chloroform | 245 | Absorbs below 245 nm |
| Acetone | 330 | Absorbs strongly, poor UV solvent |
- Absorbance range: Reliable A = 0.1-1.0. <0.1 → noise; >1.0 → stray light non-linear. Flag λ-max outside.
- Baseline + blank: Verify solvent blank subtracted. Residual solvent abs / cuvette artifacts → rising baseline at short λ.
- Slit width: Narrow → better res, lower S/N. Fine structure expected (vibrational progression) → confirm slit appropriate (typ 1-2 nm).
→ Instrument params documented, solvent cutoff respected, abs in linear range, baseline clean.
If err: A > 1.0 at λ-max → dilute + remeasure. Solvent absorbs in region → re-acquire in more transparent solvent.
Step 2: Locate λ-Max + Band Characteristics
Locate + characterize all abs bands:
- Locate λ-max: Per abs max → record λ (nm) + abs (or ε if known).
- Band shape: Broad featureless (typical soln-phase) or vibrational fine structure (rigid chromophores, polycyclic aromatics).
- Shoulders: Overlapping transitions → note approx λ + int.
- Classify by ε:
| epsilon (L mol-1 cm-1) | Transition Type | Example |
|---|---|---|
| < 100 | Forbidden (n -> pi*) | Ketone ~280 nm |
| 100--10,000 | Weakly allowed | Aromatic 250--270 nm |
| 10,000--100,000 | Fully allowed (pi -> pi*) | Conjugated diene ~220 nm |
| > 100,000 | Charge transfer | Metal complexes, dyes |
→ All abs maxima + shoulders tabulated w/ λ, abs/ε, qualitative shape.
If err: No distinct maxima (monotonic rise) → compound lacks chromophore in range, or conc too low. Increase conc / extend range.
Step 3: Classify Electronic Transitions
Assign each band → transition type:
- σ → σ* (<200 nm): Vacuum UV only. Saturated HCs + C-C/C-H. Not typically measured standard.
- n → σ* (150-250 nm): Lone pair → σ antibonding. Heteroatoms (O, N, S, halogens). Saturated amines ~190-200; alcohols/ethers ~175-185.
- π → π* (200-500 nm): Bonding π → antibonding π*. Strongest abs for organics. Int + λ increase w/ extended conjugation.
- n → π* (250-400 nm): Lone pair → π antibonding. Formally forbidden (low ε, 10-100). Characteristic C=O (270-280 simple ketones), N=O, C=S.
- Charge-transfer: e- transfer donor↔acceptor, or metal↔ligand. Very intense (ε > 10,000) + broad. Metal complexes + donor-acceptor organics.
- d-d (transition metal complexes): Weak broad in visible → crystal/ligand field splitting.
→ Each band assigned → transition type w/ rationale (pos, int, solvent sensitivity).
If err: Band unassignable → consider charge-transfer character / impurity abs. Multiple overlapping → deconvolution.
Step 4: Woodward-Fieser Rules for Conjugated Sys
Predict λ-max for conjugated dienes + enones, compare observed:
- Conjugated dienes (Woodward):
| Component | Increment (nm) |
|---|---|
| Base value (heteroannular diene) | 214 |
| Base value (homoannular diene) | 253 |
| Each additional conjugated C=C | +30 |
| Each exocyclic C=C | +5 |
| Each alkyl substituent on C=C | +5 |
| -OAcyl substituent | +0 |
| -OR substituent | +6 |
| -SR substituent | +30 |
| -Cl, -Br substituent | +5 |
| -NR2 substituent | +5 |
- α-β unsaturated carbonyls (Woodward-Fieser):
| Component | Increment (nm) |
|---|---|
| Base value (alpha-beta unsat. ketone, 6-ring or acyclic) | 215 |
| Base value (alpha-beta unsat. aldehyde) | 208 |
| Each additional conjugated C=C | +30 |
| Each exocyclic C=C | +5 |
| Homoannular diene component | +39 |
| Alpha substituent (alkyl) | +10 |
| Beta substituent (alkyl) | +12 |
| Gamma and higher substituent (alkyl) | +18 |
| -OH (alpha) | +35 |
| -OH (beta) | +30 |
| -OAc (alpha, beta, gamma) | +6 |
| -OR (alpha) | +35 |
| -OR (beta) | +30 |
| -Cl (alpha) | +15 |
| -Cl (beta) | +12 |
| -Br (beta) | +25 |
| -NR2 (beta) | +95 |
- Calc predicted λ-max: Sum base + all applicable increments.
- Compare observed: ±5 nm → supports proposed chromophore. Deviations > 10 nm → incorrect assignment / strong solvent+steric effects.
→ Predicted λ-max calc + compared observed → supports/refutes proposed chromophore.
If err: Disagreement → re-examine chromophore. Common errs: miscount substituents, overlook exocyclic double bond, wrong base val (homoannular vs heteroannular).
Step 5: Beer-Lambert for Quant
Absorbance → conc / ε characterization:
- Equation: A = ε * b * c, A = abs (dimensionless), ε = molar absorptivity (L mol-1 cm-1), b = path length (cm), c = conc (mol L-1).
- Determine ε: Conc + b known → calc ε from A at λ-max.
- Determine conc: ε known (lit / calibration) → calc c from A.
- Linearity: Valid in linear range (A = 0.1-1.0). Higher → deviations (stray light, mol interactions, instrumental).
- Solvent effects: Compare polar vs non-polar:
- Bathochromic (red) shift: λ-max → longer λ. π→π* red-shifts in more polar; n→π* in less polar.
- Hypsochromic (blue) shift: λ-max → shorter λ. n→π* blue-shifts in more polar/protic (H-bonding stabilizes lone pair ground state).
- Hyperchromic/hypochromic: Increase / decrease ε w/o λ change.
→ Quant results calc w/ appropriate sig figs, linearity verified, solvent effects documented if multi-solvent avail.
If err: Linearity fails → check sample degradation, aggregation at high conc, fluorescence interference. Dilute + remeasure to confirm.
Check
- Solvent cutoff respected + abs in linear range (0.1-1.0)
- All λ-max + shoulders tabulated w/ λ, abs, ε
- Each band → electronic transition type
- Woodward-Fieser calc where applicable + compared observed
- Beer-Lambert applied correctly w/ verified linearity
- Solvent effects characterized if multi-solvent
- Chromophore consistent w/ structure from other spectra
Traps
- Measure > A=1.0: Unreliable due to stray light. Always dilute + remeasure if λ-max abs > 1.0.
- Ignore solvent cutoff: Interpret abs below cutoff → artifacts, not real.
- Confuse transition types by intensity: Weak band ~280 could be n→π* carbonyl / forbidden π→π* aromatic. Context + solvent effects distinguish.
- Misapply Woodward-Fieser: Empirical rules → conjugated dienes + α-β unsat carbonyls only. Not for aromatic sys, isolated chromophores, metal complexes.
- Neglect impurity abs: Small amount of strongly-absorbing impurity → dominate spectrum. λ-max mismatch expectations → consider impurity.
- Assume 1 band = 1 transition: Broad bands often multi overlapping transitions. Deconvolution may be needed.
→
interpret-nmr-spectrum— mol connectivity → support chromophore IDinterpret-ir-spectrum— func groups contributing to chromophoreinterpret-mass-spectrum— formula + detect conjugation via fraginterpret-raman-spectrum— complementary vibrational → symmetric chromophoresplan-spectroscopic-analysis— select + sequence techniques pre-acquisition
GitHub 저장소
연관 스킬
llamaguard
기타LlamaGuard는 폭력 및 혐오 발언 등 6가지 안전 범주에서 LLM 입력과 출력을 조정하기 위한 Meta의 70-80억 파라미터 모델입니다. 94-95% 정확도를 제공하며 vLLM, Hugging Face 또는 Amazon SageMaker를 사용해 배포할 수 있습니다. 이 기술을 사용하여 AI 애플리케이션에 콘텐츠 필터링 및 안전 가드레일을 손쉽게 통합하세요.
cost-optimization
기타이 Claude Skill은 리소스 적정화, 태깅 전략, 지출 분석을 통해 개발자들이 클라우드 비용을 최적화할 수 있도록 지원합니다. AWS, Azure, GCP에서 클라우드 비용을 절감하고 비용 거버넌스를 구현하기 위한 프레임워크를 제공합니다. 인프라 비용을 분석하거나, 리소스를 적정화하거나, 예산 제약을 충족해야 할 때 사용하세요.
quantizing-models-bitsandbytes
기타이 스킬은 bitsandbytes를 사용하여 LLM을 8비트 또는 4비트 정밀도로 양자화하며, 최소한의 정확도 손실로 50-75%의 메모리 감소를 달성합니다. 제한된 GPU 메모리에서 더 큰 모델을 실행하거나 추론을 가속화하는 데 이상적이며, INT8, NF4, FP4와 같은 형식을 지원합니다. 이 스킬은 HuggingFace Transformers와 통합되어 QLoRA 학습 및 8비트 옵티마이저를 가능하게 합니다.
dispatching-parallel-agents
기타이 Claude Skill은 3개 이상의 독립적인 문제를 동시에 조사하고 해결하기 위해 다중 에이전트를 배치합니다. 공유 상태나 의존성 없이 해결 가능한 무관련 장애 시나리오에 맞게 설계되었습니다. 핵심 기능은 병렬 문제 해결로, 각 독립 문제 영역마다 하나의 에이전트를 할당하여 효율성을 극대화합니다.
