πŸ§ͺ SMILES β†’ TLC / Paper Chromatography Rf Predictor

Enter SMILES Β· choose solvent & solvent front Β· get predicted Rf values

⚠ Accuracy notice: Rf is not a fundamental molecular property β€” it depends on stationary phase, solvent, temperature, plate batch, and chamber saturation. This tool approximates Rf from molecular descriptors (logP, TPSA, HBD, aromaticity). Expect Β±0.10–0.15 in absolute Rf. It is very reliable for elution order and for choosing a solvent system, but not for identifying an unknown compound by Rf alone.
πŸ“˜ New to this predictor? Read the tutorial to understand what it can and cannot do.
Example: Caffeine: CN1C=NC2=C1C(=O)N(C(=O)N2C)C Aspirin: CC(=O)Oc1ccccc1C(=O)O Ibuprofen: CC(C)Cc1ccc(cc1)C(C)C(=O)O Paracetamol: CC(=O)Nc1ccc(O)cc1 Quercetin: O=c1c(O)c(-c2ccc(O)c(O)c2)oc2cc(O)cc(O)c12
Tips: β€’ SMILES is case-sensitive β€” c1ccccc1 is benzene, C1CCCCC1 is cyclohexane. β€’ Charges: [O-], [NH3+]. β€’ Aromatic nitrogens in brackets: [nH] for indole/pyrrole N-H. β€’ If parsing fails, check for balanced brackets and ring-closure digits.

No prediction yet

Go to Setup & Predict, paste SMILES strings, choose a solvent, then click Predict Rf Values.

πŸ“‘ On this page

  1. What this tool can and cannot do
  2. How Rf prediction works (QSRR)
  3. The descriptors used
  4. Reading the results table
  5. Choosing a solvent system
  6. Accuracy & validation
  7. When to trust β€” and when to test
  8. SMILES syntax cheat-sheet

⚠ 1. What this tool can and cannot do

TaskReliabilityReason
Predict elution orderHigh (β‰ˆ 90%)Order is driven by polarity (logP).
Predict polarity classHighDirectly follows logP / TPSA.
Choose a solvent systemGoodRelative polarity of solvents is well-known.
Flag co-elution riskGoodCompounds with similar logP tend to co-migrate.
Predict absolute Rf to Β±0.03Not possibleRequires experimental calibration.
Identify an unknown by Rf aloneNever do thisMany compounds share the same Rf.
Bottom line: Use this tool for planning, not for confirmation. Confirmation requires running a co-spotting standard.

πŸ”¬ 2. How Rf prediction works (QSRR)

QSRR = Quantitative Structure–Retention Relationship. Molecular structure β†’ descriptors β†’ retention.

Rf = sigmoid( k Β· (logP βˆ’ logP_solvent) ) + corrections

The physical intuition

  • logP > solvent logP β†’ prefers organic mobile phase β†’ travels far β†’ high Rf.
  • logP < solvent logP β†’ prefers polar stationary phase β†’ sticks β†’ low Rf.
  • They match β†’ distributes equally β†’ Rf β‰ˆ 0.5, ideal.

πŸ“Š 3. The descriptors used

DescriptorEffect on Rf (silica TLC)
logP↑ Rf β€” main driver
TPSA↓ Rf β€” polar surface retained
HBD↓ Rf β€” H-bonds to silanol
Aromatic ringsSlight ↓ Rf
Formal chargeStrong ↓ Rf
MWSmall ↓ for very large molecules

πŸ“‹ 4. Reading the results table

RfClassComment
0.00–0.10Very polarIncrease solvent polarity
0.10–0.30PolarAcceptable for very polar analytes
0.30–0.60ModerateIdeal range
0.60–0.85Non-polarWatch for tailing
0.85–1.00Highly non-polarDecrease solvent polarity

🧴 5. Choosing a solvent system

If predicted Rf is...Change solvent to...
All < 0.2More polar solvent
All > 0.8Less polar solvent
0.3–0.7 for most compoundsKeep this solvent
Two compounds with Ξ”Rf < 0.05Change solvent or use 2D chromatography

🎯 6. Accuracy & validation

Published QSRR models achieve RΒ² β‰ˆ 0.7–0.85, giving Β±0.10–0.15 in absolute Rf.

Sources of error

  • Plate batch / activation
  • Chamber saturation
  • Temperature β€” Rf changes by ~0.01 per Β°C
  • Sample load
  • Solvent composition

How to validate the prediction

  1. Run a reference standard of known Rf on the same plate.
  2. Compare your observed Rf to the reference β€” this is the calibration.
  3. Apply that calibration to your unknown compounds.
  4. If the predicted elution order matches your observed order, the model has done its job.

βœ… 7. When to trust β€” and when to test

SituationTrust the prediction?
Choosing which solvent to try firstYes β€” as guide
Comparing relative polarity of analogsYes
Predicting co-elutionYes β€” as warning
Reporting an Rf in a paperNo β€” measure it
Identifying an unknownNo β€” co-spot a standard
Ionic compoundsCaution
MW > 1000Caution

πŸ“ 8. SMILES syntax cheat-sheet

FeatureSyntaxExample
Aliphatic carbonCCC = ethane
Aromatic carboncc1ccccc1 = benzene
Double bond=C=O
Triple bond#C#N
Ring closuredigitc1ccccc1
Branch( )CC(C)C
Charged atom[ ][O-], [NH3+]
Aromatic N-H[nH]c1cc[nH]c1
Chlorine / BromineCl / BrCCl

Quick reference structures

  • Caffeine: CN1C=NC2=C1C(=O)N(C(=O)N2C)C
  • Aspirin: CC(=O)Oc1ccccc1C(=O)O
  • Quercetin: O=c1c(O)c(-c2ccc(O)c(O)c2)oc2cc(O)cc(O)c12
  • Glucose (open chain): OCC(O)C(O)C(O)C(O)C=O
  • Glycine: NCC(=O)O
If parsing fails: check (1) balanced parentheses, (2) paired ring-closure digits, (3) correct atom case (c vs C), (4) charges inside brackets with explicit H.
Home