Caffeine

What does caffeine's structure and IR spectrum look like?

You ask

What does its structure and IR spectrum look like?

Wobbl runs

Terminal
$ wobbl simulate caffeine "what does its structure and IR spectrum look like?" \  --no-open
The caffeine, its vibrations by hand, as the run left it at 5.4

Vibration mode, then its swing 5.4

peak partial charge

0.314 e

OK

Frequency 1,712 cm-1, IR strength 996 km/mol

Open the full report

Checks, numbers and charts

OKcaffeine (C8H10N4O2, 194.19 g/mol), by GFN2-xTB: its strongest IR band is at 1712 cm-1, the C=O stretch, its dipole is 3.98 D, dissolving in water is worth -43.9 kJ/mol. At 298 K its entropy is 470.8 J/(mol K) and its heat capacity 211.0 J/(mol K).

Dipole
3.98 D
Strongest IR band
1,712 cm-1
HOMO-LUMO gap
3.40 eV
Confidence
high
  • passgeometry convergence
  • passminimum
  • passSCF convergence
  • infothermochemistry
summary.txt
OK. caffeine (C8H10N4O2, 194.19 g/mol), by GFN2-xTB: its strongest IR band is at 1712 cm-1, the C=O stretch, its dipole is 3.98 D, dissolving in water is worth -43.9 kJ/mol. At 298 K its entropy is 470.8 J/(mol K) and its heat capacity 211.0 J/(mol K). Confidence: high. passed geometry convergence, minimum, SCF convergencedipole: 3.98 Dstrongest IR band: 1712 cm-1HOMO-LUMO gap: 3.40 eVentropy at 298 K: 470.8 J/(mol K) Checks:  [pass] geometry convergence: GFN2-xTB stopped after 18 steps with the largest force at 0.0029 eV/A, under the 0.003 eV/A it aims for  [pass] minimum: all 66 frequencies are real, so the structure is a true minimum, after following 1 imaginary mode downhill  [pass] SCF convergence: all 612 GFN2-xTB single points converged within 300 iterations  [info] thermochemistry: ideal gas at 298 K and 101 kPa with a rotational symmetry number of 1, and 4 soft modes under 100 cm-1 treated as hindered rotors Assumptions:  - caffeine as RDKit ETKDG distance geometry relaxed with MMFF94, the lowest of 1 conformers, refined with GFN2-xTB, good to about 0.02 A in bond lengths, 2.5 degrees in angles and 8% in frequencies.  - One molecule in the gas phase unless a solvent is named, and a solvent is a continuum (ALPB) rather than explicit molecules.  - Thermochemistry treats the molecule as an ideal gas, a rigid rotor and harmonic oscillators, with modes under 100 cm-1 blended into free rotors.  - The HOMO-LUMO gap from GFN2-xTB is a trend only, often off by over 1 eV, so it is not an optical gap. Images: ir: images/ir.pngReport: report.html

OKcaffeine (C8H10N4O2, 194.19 g/mol), by GFN2-xTB: its strongest IR band is at 1712 cm-1, the C=O stretch, its dipole is 3.98 D, dissolving in water is worth -43.9 kJ/mol. At 298 K its entropy is 470.8 J/(mol K) and its heat capacity 211.0 J/(mol K).

Dipole
3.98 D
Strongest IR band
1,712 cm-1
HOMO-LUMO gap
3.40 eV
Confidence
high
structure summary.txt
OK. caffeine (C8H10N4O2, 194.19 g/mol), by GFN2-xTB: its strongest IR band is at 1712 cm-1, the C=O stretch, its dipole is 3.98 D, dissolving in water is worth -43.9 kJ/mol. At 298 K its entropy is 470.8 J/(mol K) and its heat capacity 211.0 J/(mol K). Confidence: high. passed geometry convergence, minimum, SCF convergencedipole: 3.98 Dstrongest IR band: 1712 cm-1HOMO-LUMO gap: 3.40 eVentropy at 298 K: 470.8 J/(mol K) Checks:  [pass] geometry convergence: GFN2-xTB stopped after 18 steps with the largest force at 0.0029 eV/A, under the 0.003 eV/A it aims for  [pass] minimum: all 66 frequencies are real, so the structure is a true minimum, after following 1 imaginary mode downhill  [pass] SCF convergence: all 612 GFN2-xTB single points converged within 300 iterations  [info] thermochemistry: ideal gas at 298 K and 101 kPa with a rotational symmetry number of 1, and 4 soft modes under 100 cm-1 treated as hindered rotors Assumptions:  - Wobbl guessed: temperature not given: assumed 25 C (298.15 K).  - caffeine as RDKit ETKDG distance geometry relaxed with MMFF94, the lowest of 1 conformers, refined with GFN2-xTB, good to about 0.02 A in bond lengths, 2.5 degrees in angles and 8% in frequencies.  - One molecule in the gas phase unless a solvent is named, and a solvent is a continuum (ALPB) rather than explicit molecules.  - Thermochemistry treats the molecule as an ideal gas, a rigid rotor and harmonic oscillators, with modes under 100 cm-1 blended into free rotors.  - The HOMO-LUMO gap from GFN2-xTB is a trend only, often off by over 1 eV, so it is not an optical gap. Images: ir: images/ir.pngReport: report.html

About this example

Wobbl reads the name, builds the 3D structure with RDKit and relaxes it with GFN2-xTB. It then finds the vibrations, the IR spectrum and the thermochemistry. Nothing in the run is specific to caffeine, any name, formula or SMILES goes the same way.

Result Value
Molecule caffeine, C8H10N4O2, 194.19 g/mol
Strongest IR band 1712 cm-1, the C=O stretch
Dipole 3.98 D
HOMO-LUMO gap 3.40 eV
Entropy at 298 K 470.8 J/(mol K)
Vibrations all 66 real, so the structure is a true minimum

output/structure/playback.json and output/spectrum/playback.json hold the atoms, bonds and frames the viewer plays.

Run it yourself

Terminal
$ git clone https://github.com/sohan-bhat/wobbl$ cd wobbl/examples/caffeine && ./run.sh

The part, the script and every output are in examples/caffeine on GitHub.